Bilateral Reduction of Corneal Sub-basal Nerves in Eyes with Recurrent Unilateral Posner-Schlossman Syndrome: An In Vivo Laser Scanning Confocal Microscopy Study

PURPOSE. To evaluate the circumstances of corneal sub-basal nerves (CSNs) in the affected and contralateral unaffected eyes of patients with unilateral Posner-Schlossman Syndrome (PSS) through in vivo laser scanning confocal microscopy (LSCM) analysis. METHODS. Total 74 eyes from 37 unilateral PSS patients and 38 eyes of control subjects were recruited. Sub-basal corneal nerve parameters in the in vivo LSCM, including the number and density of total nerves, main nerve trunks, and nerve branches were evaluated. Corneal epithelial and endothelial cell densities were also calculated. RESULTS. The affected eyes of the patients with unilateral PSS demonstrated a signicant decrease in all CSN parameters (P<0.005), including the number and density of total nerves, main nerve trunks and nerve branches, as compared to the control eyes. Lower CSN value was observed in contralateral unaffected eyes for total nerve density as well as the number and density of main nerve trunks (P<0.03) as compared to the control eyes. Moreover, corneal endothelial cell density was signicantly reduced in affected eyes with PSS as compared to the control eyes (P<0.015). Yet, signicant reduction could only be observed in the affected and contralateral unaffected eyes of the patients with recurrent unilateral PSS. In addition, a positive correlation was found between endothelial cell density and the density of main nerve trunks in PSS-affected eyes (r = 0.33, P<0.05). CONCLUSIONS. In vivo LSCM analysis revealed the reduction of CSNs in the affected and contralateral unaffected eyes of the patients with recurrent unilateral PSS. The positive correlation between corneal endothelial cell density and CSNs could suggest a potential link between corneal innervation and corneal endothelial cells. In vivo LSCM analysis (Heidelberg Retina Tomograph 3 with the Rostock Cornea Module; Heidelberg Engineering GmbH, Heidelberg, Germany) was performed in all study subjects [3, 19] . A helium neon diode laser source with a wavelength of 670 nm was adopted. Two-dimensional images of the in vivo LSCM were captured as 384×384 pixels over an area of 400×400 μm 2 (160,000 μm 2 ) at a selectable corneal depth. One drop of topical anesthesia (0.5% proparacaine hydrochloride eye drops; s.a. ALCON-COUVREUR n.v) was instilled in each examined eye. A disposable sterile cap (Tomo-Cap, Heidelberg, Germany) was mounted over the lens after covering the lens with 0.2% Carbomer ere gel (Dr. Gerhard Mann Chem -Pharm Fabrik GmbH) below. Corneal images were obtained by asking subjects to x their gaze at a red xation light, and the lens was advanced manually until the images were obtained at specic corneal locations. Sequential scans were captured from the corneas of both eyes in the patients with unilateral PSS, and only one randomly selected eye in the control subjects. The operator adjusted the position of the objective lens until a bright image of the endothelium was visible in the center of the eld.


Introduction
Posner-Schlossman syndrome (PSS), also known as glaucomatocyclitic crisis, is usually characterized by the unilateral self-limited recurrent episodes of remarkable intraocular pressure (IOP) elevation accompanied by mild anterior chamber in ammation and ne white keratic precipitates (KP) [1,2]. With the repeated attacks, 26.4% of PSS eyes would develop glaucoma, and 17% of eyes are progressively damaged and eventually need to receive glaucoma ltering surgery [3].
Recurrent anterior chamber in ammation and the appearance of KP could make an impact on the cornea [4]. Multiple studies have reported a decrease in corneal endothelial cell number in PSS eyes [5][6][7]. A recent study demonstrates an increase in corneal Langerhans cells and activation of corneal stromal cells in eyes with PSS through the in vivo laser scanning confocal microscopy (LSCM) analysis [8]. Although the etiology of PSS remains elusive, previous research studies suggest the possible involvement of viral infection in the pathogenesis of PSS [9]. Current in vivo LSCM studies demonstrated that both affected and contralateral unaffected eyes of patients with unilateral acute microbial keratitis have a signi cant bilateral loss of corneal sub-basal nerves (CSNs) as compared to the control subjects[6, 10,11]. Cornea is an innervated tissue from the supply by the ophthalmic division of the trigeminal nerves, which function to maintain the integrity of the epithelium and wound healing through trophic functions of the nerves [12,13]. The CSN parameters, including total nerves, main nerve trunks, and nerve branches, and the endothelial layers have been shown to be the important indicators for corneal health and diseases [14,15].
In vivo LSCM is a non-invasive technique allowing the in vivo examination of human cornea visualized at all cellular levels by continuous confocal scanning [16]. Using LSCM, the density and morphology of corneal nerves could be evaluated in various corneal and systemic conditions, such as PSS, neurotrophic keratopathy, corneal infections and after corneal surgeries [8,14,17,18]. Moreover, LSCM has also been applied in viral keratitis, keratoconus and corneal dystrophy to the disease [14]. Even though the involvement of corneal Langerhans cells, stromal cells and endothelial cell impactions has been shown in PSS, whether CSN parameters would be altered in PSS patients is still unclear. Herein, this study aimed to evaluate the number and density of CSNs in both affected and contralateral unaffected eyes of patients with unilateral PSS using the in vivo LSCM. In addition, the correlations of CSNs with corneal epithelial and endothelial cell densities were also analyzed.

Study subjects
This was a retrospective cross-sectional study, which was approved by the Ethics Committee for Human Medical Research at Joint Shantou International Eye Center of Shantou University and the Chinese University of Hong Kong and conducted in accordance with the tenets of the Declaration of Helsinki.
Total 37 patients with diagnosis of unilateral PSS patients (74 eyes) and 38 control subjects were recruited at Joint Shantou International Eye Center of Shantou University and the Chinese University of Hong Kong between 2018 and 2020.
The eligible diagnostic criteria of Posner-Schlossman syndrome include [3,19]: (1) Recurrent transient episodes of mild nongranulomatous cyclitis; (2) Symptoms of mild discomfort, halos, and slight blurring of vision; (3) Findings of corneal edema, elevated IOP, open angles, few cells and minimal are, mydriasis, no iris posterior synechia and small-to-medium, discrete, round, white KP accumulating in the lower half of the cornea in the affected eye; (4) Attacks last for few hours to few weeks. The exclusion criteria include: (1) Bilaterally affected PSS; (2) Subjects with a history of ocular surgery; (3) other ocular diseases, except senile cataract, conjunctivitis, dry eye or refractive error; (4) spherical equivalent below -6.0 diopters. The control group included individuals without any history of eye diseases (except senile cataract), conjunctivitis, dry eye, refractive error with spherical equivalent below -6.0 diopters, a history of ocular surgery, current use of contact lenses or long-term use of eye drops. All included subjects received complete ophthalmic examinations, including measurement of refraction, best corrected visual acuity, IOP, anterior segment evaluation by slit-lamp biomicroscope and in vivo LSCM.

In Vivo Laser Scanning Confocal Microscopy
In vivo LSCM analysis (Heidelberg Retina Tomograph 3 with the Rostock Cornea Module; Heidelberg Engineering GmbH, Heidelberg, Germany) was performed in all study subjects [3,19] . A helium neon diode laser source with a wavelength of 670 nm was adopted. Two-dimensional images of the in vivo LSCM were captured as 384×384 pixels over an area of 400×400 μm 2 (160,000 μm 2 ) at a selectable corneal depth. One drop of topical anesthesia (0.5% proparacaine hydrochloride eye drops; s.a. ALCON-COUVREUR n.v) was instilled in each examined eye. A disposable sterile cap (Tomo-Cap, Heidelberg, Germany) was mounted over the lens after covering the lens with 0.2% Carbomer ere gel (Dr. Gerhard Mann Chem -Pharm Fabrik GmbH) below. Corneal images were obtained by asking subjects to x their gaze at a red xation light, and the lens was advanced manually until the images were obtained at speci c corneal locations. Sequential scans were captured from the corneas of both eyes in the patients with unilateral PSS, and only one randomly selected eye in the control subjects. The operator adjusted the position of the objective lens until a bright image of the endothelium was visible in the center of the eld.

Image Analysis
Digital images were stored in a workstation with the Heidelberg built-in software (Heidelberg Eye Explorer version 1.5.10.0, Heidelberg Engineering GmbH 2005, Heidelberg, Germany). An experienced masked observer selected the three best-focused representative images of the CSN plexus, epithelial and endothelial cell layer analysis from the central cornea of all subjects, who satis ed the criteria for selection. The criteria to select the images were the best-focused complete images, with the whole image in the same layer and good contrast, and at least 50 good-quality images. The CSN plexus was chosen from the basal epithelial layer and anterior to the Bowman's layer. Epithelial cell and the endothelial cell densities were quanti ed by the computer built-in in vivo LSCM corresponding software. A semiautomated tracing program NeuronJ [20], a plug-in for ImageJ distributed by the National Institutes of Health (http://rsb.info.nih.gov/ij/http://rsb.info.nih.gov/ij/;http://www.imagescience.org/ meijering/software/neuronj/; accessed July 12, 2020), was adopted for the CSN analysis. Main nerve trunk was de ned as the nerve trunk in 1 image that did not branch from other nerve and nerve branch was de ned as branch emerging from the main nerve trunk in the same image. The total nerves were de ned as all nerves, including main nerve and branches in 1 image [21].The density and number of nerves were analyzed by tracing all visible nerve bers in the image, and the length of the nerve bers was calculated as millimeters (mm)/mm 2 and number (n)/frame.

Statistical Analysis
Statistical analysis was performed with SPSS version 25.0 (SPSS Inc, Chicago, IL, USA). Distribution of data was analyzed by the Kolmogorov-Smirnov test. As most of the nerve parameters were non-normally distributed, comparison between two groups was analyzed by the Kruskal-Wallis H test and adjusted by Bonferroni correction for multiple testing correction. Correlation analysis was performed using Spearman's rank correlation test. Categorical variables were compared using the Fisher's exact test. All the data are presented as mean ± standard error of mean. P less than 0.05 was considered as statistically signi cant.

Demographics of the study subjects
Total 37 affected and 37 contralateral unaffected eyes from 37 patient diagnosed with unilateral PSS (mean age: 40.6±15.3 years; 22 males and 15 females) and 38 eyes from 38 control subjects (mean age: 40.3±14.9 years; 19 males and 19 females) were studied. (Table 1). No statistically signi cant differences were found in age (P = 0.80) and gender (P = 0.50) between the PSS patients and the control subjects. Moreover, there was also no statistically signi cant differences in the numbers of left and right eyes studied between the PSS patients and the control subjects (P = 0.50). However, the IOP in the affected eyes of the PSS patients was signi cantly smaller than that of the control subjects (P = 0.0001).

Reduction of corneal endothelial cells in the affected eyes of patients with unilateral Posner-Schlossman Syndrome
The in vivo LSCM analysis on the unilateral PSS patients revealed that the epithelial cell density of the affected eyes (5237 ± 809 cells/mm 2 ) showed no statistically signi cant differences as compared to the unaffected eyes (5441 ± 820 cells/mm 2 , P = 0.690) and the control eyes (5665 ± 614 cells/mm 2 , P = 0.690; Table 2). Instead, the endothelial cell density of the affected eyes of the unilateral PSS patients (2627 ± 322 cells/mm 2 ) was signi cantly reduced as compared to the control eyes (2862 ± 298 cells/mm 2 , P = 0.014; Fig.1D and 1F). Yet, the endothelial cell density of the unaffected eyes of the unilateral PSS patients (2714 ± 362 cells/mm 2 ) showed no statistically signi cant differences as compared to the affected eyes of the unilateral PSS patients (P = 0.380) and the control eyes (P = 0.590).
Reduction of corneal sub-basal nerves in the affected and unaffected eyes of patients with unilateral Posner-Schlossman Syndrome For the affected eyes of the patients with unilateral PSS, the number of total nerves (9.7 ± 3.4/frame, P = 0.0001), main nerve trunks (4.8 ± 1.6/frame, P = 0.001), and nerve branches (4.9 ± 2.9/frame, P = 0.005) as well as the density of total nerves (2.3 ± 0.7 mm/mm 2 , P = 0.0001), main nerve trunks (1.6 ± 0.5 mm/mm 2 , P = 0.0001) and nerve branches (0.6 ± 0.4 mm/mm 2 , P = 0.0001) were signi cantly lower than the control eyes (number: total nerves 13.8 ± 4.9/frame, main nerve trunks 6.2 ± 1.6/frame, nerve branches 7.6 ± 4.6/frame; density: total nerves 3.3 ± 0.8 mm/mm 2 , main nerve trunks 2.3 ± 0.6 mm/mm 2 , nerve branches 1.1 ± 0.6 mm/mm 2 ; Table 2 and Fig. 2). Notably, the number of main nerve trunks (5.1 ± 1.6/frame, P = 0.027) as well as the density of total nerves (2.8 ± 0.6 mm/mm 2 , P = 0.021) and main nerve trunks (1.8±0.4 mm/mm 2 , P = 0.010) in the contralateral unaffected eyes of the patients with unilateral PSS were also signi cantly lower than the control eyes. Besides, the number of total nerves (P = 0.006) and nerve branches (P = 0.004) as well as the density of total nerves (P = 0.015) and nerve branches (P = 0.010) in the affected eyes of the patients with unilateral PSS were also signi cantly lower than the contralateral unaffected eyes. In addition, a weak but signi cant positive correlation was observed between the endothelial cell density and the density of main nerve trunks in the affected eyes of the patients with unilateral PSS (r = 0.33, P = 0.048; Fig. 3).

Reduction of endothelial cell density and corneal sub-basal nerves in patients with recurrent unilateral Posner-Schlossman Syndrome
The 37 PSS patients were subdivided into 24 recurrent PSS and 13 non-recurrent patients as proven to be recurrent at the follow-up visits. For the affected eyes of the patients with recurrent PSS, the endothelial cell density (2578 ± 381 cells/mm 2 , P < 0.0001), the number of total nerves (9.3 ± 2.9/frame, P < 0.0001), main nerve trunks (4.8 ± 1.7/frame, P < 0.0001), and nerve branches (4.5 ± 2.4/frame, P < 0.0001) as well as the density of total nerves (2.1 ± 0.7 mm/mm 2 , P < 0.050), main nerve trunks (1.5 ± 0.5 mm/mm 2 , P < 0.050) and nerve branches (0.6 ± 0.3 mm/mm 2 , P < 0.050) were signi cantly lower than the control eyes (Table 3 Fig. 4). For the contralateral unaffected eyes of the patients with recurrent PSS, also showed statistically signi cant reduction in the total nerve density (2.8 ± 0.7 mm/mm 2 , P = 0.034) as compared to the control eyes. On the contrary, the affected and unaffected eyes of the patients with non-recurrent PSS did not statistically signi cant differences as compared to the control eyes.

Discussion
In the current study, we demonstrated that not only the endothelial cell density showed a signi cant reduction in the affected eyes of the patients with unilateral PSS, but also the number and density of CSNs, especially the contralateral unaffected eyes, as compared to the age-matched controls. We, for the rst time, evaluated the in vivo LSCM images of corneal nerves and endothelial cells together in the patients with unilateral PSS. Our results indicated that unilateral corneal endothelial cells and bilateral corneal nerves were signi cantly reduced in patients with unilateral PSS.
In the current study, we observed a signi cant reduction of endothelial cell density only in the affected eyes of the patients with unilateral PSS. Endothelial cell density decease in the eyes affected with PSS has been reported in previous study [22][23][24] , but not compared between the unaffected eyes and the normal control eyes. Corneal endothelial cell loss has been suggested to be positive correlated with the cytomegalovirus viral load of aqueous humor [25]. Signi cant reduction of endothelial cell density was only shown in the PSS-affected eyes instead of both eyes, potentially attributed to different viral load between the PSS-affected and unaffected eyes.
Using the in vivo LSCM, we observed a signi cant lower number and density of CSNs in the affected eyes of the patients with unilateral PSS, as compared to the control eyes. Notably, in the contralateral unaffected eyes of all patients with unilateral PSS, we also observed a signi cant reduction in the number of main nerve trunks as well as the density of total nerves and main nerve trunks as compared to the control eyes. We further showed a positive correlation between endothelial cell density and the density of main nerve trunks in the affected eyes of the patients with unilateral PSS. Although bilateral changes of CSN has been described in unilateral virus keratitis[6, 26] and after unilateral cataract surgery [27], we, for the rst time, demonstrated the bilateral corneal nerve reduction in the patients with unilateral PSS. However, the exact mechanisms involved in corneal nerve degeneration in eyes with PSS were not well characterized. The possible cause of PSS could be viral infection, especially cytomegalovirus and herpes simplex virus [24,28]. In contrast, we postulated that PSS could be related to HLA-B27 anterior uveitis and acute attack of primary angle closure with the involvement of central Langerhans cells and keratocyte activation[8] as the CSN reduction in patients with PSS are located in the same layer of the Langerhans cells.
The bilateral nerve reduction in the patients with unilateral PSS described here could be explained by the CNS-mediated contralateral effects although we cannot exclude subclinical infection in the contralateral eyes. The nervous interdependence in different regions of human body affecting the opposite side of the unaffected structures have been described [29]. There are also well-documented events that peripheral trigeminal nerve bers directly project to bilateral areas of brainstem nuclei and caudal medulla that affects the opposite side of the unaffected structures [30,31]. The anatomical pathway could involve an extensive bilateral central projection sending to the contralateral peripheral Gasserian ganglion [32]. A previous study reported that the patients undergoing unilateral microincisional cataract surgery show bilateral reduction of CSN plexus at 1 month postoperatively [27]. Those ndings con rm the evidence that corneal nervous interdependence between the affected and unaffected eyes. The changes in the contralateral unaffected eyes could be secondary to the affected eyes as the CSNs of unaffected eyes were higher than the affected eyes. Neuropeptides and neurotrophins, such as substance P, VIP, nerve growth factor, brain-derived neurotrophic factor, and glial cell-derived neurotrophic factor are found in the cornea with anti-in ammatory and antiapoptotic effect [33,34]. The presence of contralateral corneal nerve impairment was also believed to be related to the sympathetic immune response through the local release of proin ammatory neuropeptides by afferent dysfunctional neurons [10]. Furthermore, this study demonstrated a weak positive correlation between endothelial cell density and the density of main nerve trunks in the affected eyes of the patients with unilateral PSS. The reduction of endothelial cell density could potentially be attributed to the reduction of corneal nerves, which in turn could lead to the decreased levels of neuropeptides, subsequently resulting in endothelial cell loss [35].
When we compared the recurrent and non-recurrent patients with PSS, the endothelial cell density as well as the number and density of total nerves, main nerve trunks and nerve branches of the affected eyes of the patients with recurrent PSS were signi cantly lower as compared to the control eyes, but the reduction was not observed in the non-recurrent patients. A Chinese study on predominant PSS reported that one fourth of the PSS eyes shows glaucomatous damage development because of repeated attacks [3]. These imply that recurrent episode could exacerbate the damage of PSS. However, a previous study indicated that corneal nerve density of 1064 μm/frame is su cient for normal sensation [6]. This could explain why the nerve damage in PSS patients did have the clinical symptoms.
The current study has several limitations. First, there could be unnoticed changes because the range of in vivo LSCM has its own limit. Second, some morphological parameters of cornea nerves have not been evaluated, such as tortuosity. Third, the retrospective study design of this study was not able to determine if the nerve reduction and the endothelial cell loss were progressive and if there were other in uencing factors, such as treatment and IOP level during each episode affecting the nerve and the endothelial cells. The follow-up visits were limited preventing us to assess the persistence or resolution of these changes over time as well as the long-term implications of the contralateral sympathetic nervous response.
In summary, our results, based on the in vivo LSCM analysis, revealed the reduction of endothelial cell density and CSNs in the affected and contralateral unaffected eyes of the patients with unilateral PSS, especially the recurrently affected eyes.     The density of the total nerves, the main nerve trunks, and the nerve branches. (B)The number of total nerves, main nerve trunks, and nerve branches. Error bars represents standard deviation from the mean. **P<0.03, ***P<0.006, and ****P<0. 0001 by analysis of Kruskal-Wallis H test.

Figure 3
The correlation between endothelial cell density and the density of main nerve trunks in the affected eyes of patients with unilateral PSS.