Effect of Corneal Collagen Cross-Linking on Subsequent Corneal Fungal Infection in Rats

Abstract

Purpose: The purpose of this study was to determine whether corneal collagen cross-linking (CXL) alters fungal susceptibility and increases the severity of keratitis through macrophage activation in rats.

Methods: Four weeks following CXL pretreatment, the corneal epithelium of adult rats was removed and inoculated with Candida albicans (C. albicans; CXL + inoculation group). The non-CXL-pretreated corneas were also inoculated with C. albicans (inoculation group). Clinical scoring and histopathological examination were performed to determine the severity of fungal keratitis. Immunofluorescence and confocal microscopy imaging were applied to determine the effects of CXL treatment on corneal local macrophage content. Real-time polymerase chain reaction (RT-PCR) and Western blots were used to evaluate mRNA and protein expression. Flow cytometry assays were performed to detect M1- and M2-type macrophages.

Results: CXL pretreatment (CXL + inoculation) resulted in higher infection success rate and more severe fungal keratitis than inoculation alone (inoculation group). On days 1, 3, and 7 following fungal infection, the increase in macrophage infiltration and IL-1β, MMP-9, and VEGFA expression was greater in the CXL + inoculation group than in the inoculation group. Number of M1- and M2-type macrophages, M1 to M2 ratio, M1-type macrophage genes, inducible nitric oxide synthase (iNOS), and tumor necrosis factor (TNFα) expression were higher in the CXL + inoculation group compared with the inoculation group.

Conclusions: Our data demonstrate that CXL may increase the colonization of macrophages and activate more M1-type macrophages to increase fungal susceptibility and severity of keratitis.

Translational Relevance: This study may aid long-term risk assessment and treatment of the complications of CXL.

Introduction

Corneal collagen cross-linking (CXL) is the first and only procedure currently applied to prevent the progression of keratoconus.^1,2 CXL has become a major nontransplant therapy for the prevention and treatment of keratoconus and other corneal ectatic diseases.^3,4 

In addition to initial indications for use of CXL to stabilize the ectatic disease process, several studies have used the principles of CXL to treat infectious keratitis.^5–7 The number of CXL operations worldwide has increased dramatically.⁴ However, it is well known that during CXL treatment, corneal stromal cells suffer severe damage due to UV radiation and the production of reactive oxygen species.⁸ Whether CXL will alter the postoperative immune microenvironment with consequently increased complications requires further study. 

Interestingly, recent studies have revealed that CXL may lead to adverse events, including serious infectious keratitis.^9–12 Incomplete corneal epithelium, contact lenses, and ocular surface disease may be potential risk factors for secondary keratitis following CXL.^13–15 Nonetheless, the mechanism of fungal infection after CXL remains unclear. Faramarzi et al. considered that CXL may activate immune cells, such as macrophages, which in turn stimulate matrix metalloproteinase (MMP) production.¹⁶ It has been demonstrated that human macrophages respond to bacteria, viruses,^17,18 and fungus.¹⁹ Macrophages can be divided into M1-type and M2-type macrophages according to their functions. Specifically, M1-type macrophages mainly produce proinflammatory factors and reactive oxygen species, including inducible nitric oxide synthase (iNOS) and tumor necrosis factor α (TNFα),^20,21 whereas M2-type macrophages mainly secrete anti-inflammatory factors. Activation of macrophages in the cornea may play a critical role in regulating the immune response to bacterial corneal infection with F. solani.¹⁹ Jiang et al. demonstrated that macrophages are polarized in fungal keratitis due to A. fumigatus.²² However, the alteration of corneal immune cells has been rarely studied in CXL and subsequent microbial infection. 

In this study, we established a rat model of CXL inoculation with C. albicans and compared fungal keratitis in non-CXL-treated controls to analyze corneal antifungal immunity, in particular, the potential role of macrophages in CXL. Study of the persistent changes in the corneal immune microenvironment after CXL might aid long-term risk assessment and treatment of complications of CXL. 

Materials and Methods

Animals and Fungi

A total of 120 wild-type SD rats (female rats, 8 to 10 weeks of age, 200 to 300 grams in weight) were used. Rats with corneal disease were excluded by slit lamp examination. All animals were treated in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research. Groups: control (normal): untreated normal rats; CXL: rats treated with CXL for 4 weeks; inoculation: rats inoculated with C. albicans; and CXL + inoculation: rats pretreated with CXL for 4 weeks and then inoculated with C. albicans. 

The strain of C. albicans used for these studies was ATCC 90029. The yeast was cultured on Sabouraud dextrose agar for 2 days at 30°C, harvested and diluted with sterile saline to yield approximately 10⁵ colony-forming units (CFUs) per 5 µL inoculum using a predetermined conversion factor of 0.5 McFarland = approximately 1 to 5 × 10⁶ CFU/mL. 

CXL Procedure

Following removal of corneal epithelium, isotopic 0.22% riboflavin solution eye drops (10 mg riboflavin in 4.5 mL of 20% dextran T-500 solution) were administered to rats every 3 minutes for 30 minutes. Eyes were then irradiated at a surface intensity of 9 mW/cm² for 240 seconds (UV-360, Medical Photon Corp., Xiamen, China). All protocols were performed according to previous CXL procedures in our laboratory.^23,24 Following CXL, tobramycin eyedrops (Tobramycin; Alcon Laboratories, Inc.) were administered four times daily until the corneal epithelial wound was healed. None had clinical signs of corneal inflammation 1 month after treatment. 

Infection Procedure

One month after treatment with CXL, rats were treated with 0.1% fluoromethalone eyedrops 3 times a day for 3 days prior to inoculation with C. albicans. Specifically, corneas were mechanically de-epithelialized over a 4 mm diameter area and corneal surface smeared with 5 µL inoculum of C. albicans. Subsequently, a corneal contact lens with a diameter of 6 mm was cut by trephine and applied to cover the cornea. The eyelids were sutured shut with 5-0 black silk. Contact lenses were removed after 24 hours. Untreated rats received the same fungal infection progression as the control group. 

Corneal Scoring

The severity of keratomycosis was observed under a dissecting microscope and slit lamp. The following 3 criteria were scored from 0 to 4: opacity area, opacity density, and surface regularity (Table 1). If the clinical score of fungal keratitis was less than 3 on day 3 and C. albicans was not found in a corneal smear, the inoculation was considered to have failed and the infection unsuccessful. The number of failed cases was recorded. 

Table 1.

 

Visual Scoring System for Murine Fungal Keratitis

 

Histopathological Examinations

Candida albicans-infected eyes were fixed in formalin, embedded in paraffin, and sectioned at a thickness of 5 µm for histologic study. Sections were dewaxed and stained with hematoxylin and eosin, then examined by light microscopy. 

Immunofluorescence Staining

Rat eyes were enucleated and embedded in optimal cutting temperature (OCT) compound and frozen in liquid nitrogen. Ten-micrometer-thick sections were cut and mounted on glass slides. After a 20-minute fixation in 4% paraformaldehyde, slides were blocked with 20% GS for 1 hour at room temperature. Sections were then incubated overnight with CD68 antibody (ab125212; Abcam, Cambridge, MA, USA) at 4°C, followed by Alexa Fluor594-labeled Goat antirabbit IgG (ab150080, Abcam) at 37°C for 1 hour. The slides were mounted with 4′,6-diamidino-2-phenylindole (DAPI) anti-fluorescence quencher. Sections were photographed with a confocal microscope (LSM710, ZEISS) and analyzed with ImageJ software. 

Real-Time Polymerase Chain Reaction

Total RNA was purified from rat corneas using an RNA Extraction Kit (QIAGEN). The cDNA was synthesized using a cDNA synthesis kit (TaKaRa, China). Real-time polymerase chain reaction (RT-PCR) was performed using Power SYBR Green PCR Master Mix (Applied Biosystems, Paisley, UK) and the 7500 Real-Time PCR System (Thermo Fisher Scientific). The relative expression of target mRNA was normalized using GAPDH as an internal control. The sequences of the primers are shown in Table 2. 

Table 2.

 

Sequences of Rat Primers Using For RT–PCR

 

Western Blot Analysis

Rat corneas were harvested and total protein was extracted from the lysed samples in RIPA buffer. The samples were separated by SDS–PAGE gels and then transferred to NC nitrocellulose membranes. Nonspecific proteins on the membrane were blocked with 5% nonfat milk for 2 hours and incubated with primary antibodies against MMP-9 (1:1000, ab38898, Abcam) and IL-1β (1:1000, ab9722, Abcam). After overnight incubation, the blots were incubated with the corresponding secondary antibodies for 2 hours. Finally, the blots were visualized using the chemiluminescence method. The density of the blots was obtained and analyzed by ImageJ software. 

Flow Cytometry Assay

Rat corneas were harvested on day 3 after inoculation and cut into pieces in 1640 medium containing 5% fetal bovine serum (FBS) and penicillin–streptomycin. Type IV collagenase was added at 37°C for digestion for 30 minutes, and the supernatant was centrifuged and resuspended in red blood cell lysis buffer, then filtered through a screen filter of 40 µm. The cell suspension was collected in phosphate-buffered saline (PBS) and incubated with fluorescence-bound antibodies at 4°C for 30 minutes. The antibodies used were as follows: CD45-APC-CY7, D68-FITC, CD86-PE-CF594, and CD206-AF647 (Santa Cruz, Dallas, TX, USA.) Flow cytometry analyses were performed using DxFLEX (Beckman Coulter, Inc., Brea, CA, USA). The gate was set on the CD45⁺ population. Data were analyzed using Flow Jo software. 

Statistical Analysis

Data are expressed as mean ± SD. All statistical analyses were carried out using GraphPad Prism 7 (GraphPad Software Inc., La Jolla, CA, USA). Significant differences were determined by unpaired Student's t-test and 1-way analysis of variance (ANOVA) at confidence levels of 95% (*P < 0.05). 

Results

CXL Pretreatment Increases the Success of C. albicans Keratitis Development

The results obtained from the analysis of success rates are presented in Table 3. Data were collected from 46 rats in each group. The success rate of inoculation on the third day in the CXL + inoculation group and the inoculation group was 97.83% and 82.61%, respectively. Chi-square testing revealed that the success rate of the CXL + inoculation group was significantly higher (P < 0.05) than that of the inoculation group. Our results show that the cornea was more likely to be infected with fungal keratitis following CXL treatment. 

Table 3.

 

Development of Fungal Keratitis 3 Days After C. albicans Inoculation.

 

CXL Pretreatment Increases the Inflammatory Response of C. albicans Keratitis

Four weeks after CXL, the inflammation associated with CXL subsided in rats. At this stage, the CXL treatment group and the non-CXL-treated group were inoculated with C. albicans (Fig. 1A). Figure 1B shows corneal photographs after C. albicans infection. The corneas were examined under a slit lamp and clinically scored on days 1, 3, and 7 post-infection. Typical lesions of fungal keratitis, such as nebular lesions, pseudopodia, and hypopyon, were present on days 1 and 3 post-infection. The inflammatory response in the CXL + inoculation group was significantly worse (P < 0.01, N = 46) than in the inoculation group (see Figs. 1B, 1C). The lesions gradually healed and corneal transparency increased during later stages of infection (7 day; P < 0.05; see Figs. 1B, 1C). In the normal group, the corneal stroma was regular with no inflammatory cell infiltration. In the CXL group, the corneal stroma showed a reticular collagen structure with corneal thickness and infiltration of inflammatory cells almost returned to baseline level compared with that in normal corneas (Fig. 1D). In the inoculation group, the cornea exhibited local inflammatory cell infiltration. In the CXL + inoculation group, severe corneal edema occurred with massive inflammatory cell infiltration. Corneal edema was more severe in the CXL + inoculation group (****P < 0.0001, N = 6; Fig. 1E). 

Figure 1.

 

Effects of CXL pretreatment on the inflammatory response of C. albicans keratitis. (A) The corneas of rats were inoculated with C. albicans after CXL treatment. (B) Photographs of the cornea under a slit lamp microscope at 1, 3, and 7 days after infection in the CXL + inoculation and the inoculation groups. Corneal edema, nebular lesions, pseudopodia, and hypopyon were present in the inoculation and CXL + inoculation groups on days 1 and 3 post-infection. Corneal scar left in the Inoculation and CXL + inoculation groups 7 days after infection. (C) Clinical score of C. albicans–induced keratomycosis in rats (N = 46). (D) Pathological results of corneal inflammation and edema in the normal and CXL-treated corneas with or without C. albicans inoculation for 3 days. In the normal group, the corneal stroma was regular with no inflammatory cell infiltration; in the CXL group, the corneal collagen network was formed and inflammatory cell infiltration returned to normal; in the inoculation group, the cornea exhibited local inflammatory cell infiltration; in the CXL + inoculation group, severe corneal edema occurred with massive inflammatory cell infiltration. Scale Bar = 20 µm (400 ×). (E) Corneal thickness in normal and CXL-treated corneas with or without C. albicans inoculation. There were six rats in each group. Results are presented as mean ± SD. *P < 0.05, **P < 0.01, ****P < 0.0001.

CXL Pretreatment Increases Macrophage Content

The CD68 antibody is a monoclonal antibody directed against macrophages. Four weeks after CXL treatment, CD68+ cell level in the central corneas returned almost to baseline compared with that in the normal corneas. After inoculation, corneal lesions were imbued with CD68+ cells in both the CXL + inoculation group and inoculation group (Fig. 2A). The CXL + inoculation group had more macrophages than the inoculation group, especially on day 3 and day 7 (Fig. 2B). It is worth mentioning that on day 7, the epithelium of the Inoculation group and CXL + inoculation group both gradually recovered. The number of positive macrophages in the inoculation group decreased compared with that on the third day, whereas a great number of positive macrophages in the CXL + inoculation group were distributed in the anterior stromal layer of the cornea (see Fig. 2A). The immunofluorescence results were consistent with the histopathologic features. 

Figure 2.

 

Effects of CXL pretreatment on corneal macrophage content. With the use of immunofluorescence staining, macrophages (red) were labeled with the CD68+ marker. (A) Distribution of CD68+ cells in corneas of normal and CXL-treated groups with or without fungal inoculation at different time points. a: Normal cornea with a low distribution of macrophages; a’: Macrophages in the cornea after 4 weeks of CXL treatment almost recovered to normal; b: inoculation (day 1): A large number of macrophages are concentrated in the infected corneal stroma and not in the surrounding uninfected tissue, with the absence of epithelium, which was removed previously; b’: CXL + inoculation group (day 1): Diffuse distribution of macrophages in the corneal stroma with loss of epithelium; c: the inoculation group (day 3): Macrophages in the corneal stroma decreased compared with the inoculation group (day 1) group, and the epithelium gradually repaired; c’: CXL + inoculation group (day 3): Macrophages in the corneal stroma increased, and the corneal epithelium was not repaired fully; d: the inoculation group (day 7): Macrophages content in the corneal stroma did not change significantly compared with those at 3 days; d’: CXL + inoculation group (day 7): Massive macrophages are diffusely distributed in the corneal stroma, and the epithelium repaired. Scale Bar = 20 µm. (B) CD68⁺ cell population in the central cornea of CXL-treated and normal corneas with or without C. albicans inoculation on days 1, 3, and 7. Data are presented from 4 to 6 rat corneas. ANOVA was used for statistical analysis (*P < 0.05, ****P < 0.0001, ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001).

CXL Pretreatment Increases the Secretion of Cytokines in Fungal Keratitis

MMP-9, IL-1β, and VEGFA mRNA levels were quantitatively analyzed in corneal tissue at different times (0, 1, 3, and 7 days) after fungal inoculation using RTPCR. All cytokines were increased substantially on days 1 and 3, and reduced on day 7 in the infected groups. The gene expression of MMP-9, IL-1β, and VEGFA in the CXL + inoculation group were higher than that in the inoculation group (Figs. 3A, 3B, 3C). Figures 3D and 3E show that the gene and protein level of MMP-9 and IL-1β was clearly increased after C. albicans inoculation compared with the normal and CXL-treated corneas. All cytokines were higher in the CXL + inoculation group than the inoculation group, indicating that corneal inflammation caused by fungal infection was more serious following CXL. 

Figure 3.

 

Expression of inflammatory factors in C. albicans keratitis after CXL. Expression of MMP-9 (A), IL-1β (B), and VEGFA (C) mRNA in C. albicans-infected corneas (N = 5). Expression of inflammatory factors was higher in the CXL + inoculation group than in the inoculation group on days 1, 3, and 7 post-infection. (D) Western blot was performed to examine the expression of MMP-9 and IL-1β. Noninfected corneas were used as the control with β-actin as the loading control. I = inoculation group, C = CXL + inoculation group. (E) Quantitative results of MMP9 and IL-1β protein expression (N = 3). The expression of MMP-9 and IL-1β protein was higher in the CXL + inoculation group than in the inoculation group on days 1 and 3 post-infection. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

CXL Pretreatment Activates More M1-Type Macrophages

To further explore the mechanism of fungal susceptibility after CXL, we performed flow cytometry of rat corneas to evaluate M1-type (CD68+ and CD 86+) and M2-type (CD68+ and CD 206+) macrophages (Fig. 4A). Although the increased number of M1- and M2-type macrophages returned to near normal after 4 weeks of CXL treatment, the ratio of M1 to M2 was higher in the CXL group than the normal control group (see Figs. 4A, 4B). Moreover, M1- and M2-type macrophages increased simultaneously in the CXL + inoculation group (see Fig. 4A). Additionally, M1/M2 was higher in the CXL + inoculation group than in the inoculation group (N = 4, P < 0.05; see Fig. 4B). RT–PCR was performed to examine the gene expression of iNOS (Fig. 4C) and TNFα (Fig. 4D), which were secreted by M1-type macrophages in normal and CXL rats with or without inoculation at 7 days. CXL increased the gene expression of iNOS and TNFα after fungal infection (N = 3, P < 0.01). These results indicate that CXL pretreatment may have increased the severity of keratitis by activating more M1-type macrophages. 

Figure 4.

 

Changes to macrophage subtypes in C. albicans keratitis after CXL. (A) Flow cytometry of M1 (CD68+ and CD 86+) and M2 (CD68+ and CD 206+) type macrophages in normal and CXL rats with or without inoculation for 3 days. N: Low number of both M1 and M2-type macrophages; CXL: M1 and M2 -type macrophage count return to almost normal; Inoculation: The number of M1-type macrophages increased significantly, whereas the number of M2-type macrophages showed no significant change; CXL + inoculation group: Both M1 and M2-type macrophages count increased significantly. (B) Quantitative results of M1/M2 type macrophages (N = 4). RT–PCR was performed to examine the gene expression of iNOS (C) and TNFα (D), which are secreted by M1-type macrophages, in healthy and CXL-treated rats with or without inoculation for 7 days (N = 3). ANOVA was used for statistical analysis (* compared with the control group, # compared with inoculation group, ^#P < 0.05, ** or ^##P < 0.01, ***P < 0.001).

Discussion

Our results showed that the cornea was more likely to be infected with fungal keratitis following CXL treatment and was more severe. In addition, macrophages might play a potential role in the pathology of fungal keratitis following CXL treatment. Interestingly, abnormal activity of M1 macrophages may be one of the components of fungal susceptibility following CXL. 

Although in many studies infectious keratitis occurred mostly just a few days following CXL,^9,11,25 Madhu et al.²⁶ found that CXL could exacerbate the rate of late perforation in fungal keratitis. Many patients have long-term CXL, so it is important to understand its effects on the ocular surface and to understand the mechanisms of ocular surface resistance to infection by foreign microorganisms following CXL. Our previous studies indicated that inflammatory cytokines recovered and that corneal inflammation fully subsided by day 28 after CXL.²⁴ However, some few CXL treated-corneas for 1 month developed mild stromal edema, which did not fully recover to the normal level. Thus, in this study, the corneas of rats inoculated with C. albicans 1 month after CXL treatment more realistically reflected the level of inflammation and macrophage changes in keratitis. The results demonstrated that the success rate of fungal infection in the CXL group was significantly higher than in the normal controls, and the corneal score of fungal keratitis and pathological staining results showed corresponding changes. The results demonstrate that the cornea was more likely to be infected with fungal keratitis after CXL treatment. After C. albicans inoculation, fungal keratitis was more severe in the CXL-treated group than in the non-CXL-treated group. Moreover, hematoxylin and eosin (H&E) pathological staining and immunofluorescence of CD68 macrophages showed that the infiltrating infective foci of the CXL + inoculation group were relatively located in the anterior stromal layer of the cornea. There are few studies of corneal pathogen infection after CXL. Kymionis et al. indicated that UV-A light could be a stimulus to trigger reactivation of latent HSV infection and contribute to herpes simplex virus keratitis after CXL.²⁷ Mazzotta et al. found that CXL led to apoptosis of keratocytes and keratocyte repopulation began after 2 to 3 months and was complete after 6 months.⁸ These results indicate that the increased susceptibility to keratitis following CXL is related to changes in corneal stromal cells and the immune microenvironment during the CXL process. Corneal edema after CXL may be related to corneal thickness and UV energy.²⁸ We found no reports of pathological structures of corneal edema after CXL. Therefore, we cannot rule out the possibility that the microstructure and function of the cornea may not have fully returned to normal 1 month after CXL, resulting in a higher and more severe susceptibility to the fungus in the CXL pretreated cornea. 

Our results demonstrated that the expression of cytokines (IL-1β, MMP-9, and VEGFA) in the CXL + inoculation group was significantly higher than that in the Inoculation group. Studies have proven that IL-1β is a proinflammatory cytokine that is increased in the circulation of keratitis patients and induces the recruitment of neutrophils.^29–31 Dong et al.³² found that MMP-9 is probably attributed to corneal inflammation through degradation of the basement membrane. In addition, neovascularization and inflammatory infiltration in keratitis affect each other, and the cytokines VEGFA and MMP-9 are both angiogenic factors, so they also indirectly promote corneal inflammatory infiltration.^33,34 Therefore, the abnormally increased cytokine levels of IL-1β, MMP-9, and VEGFA in rat corneas after CXL and fungal infection may be related to the susceptibility to and severity of infectious keratitis. 

Research has found that IL-1β, MMP-9, and VEGFA in the cornea are produced by macrophages.^35–37 Hence, increased fungal susceptibility following CXL may be associated with innate immune cells — macrophages. Our results indicated that the disproportionate M1- and M2-type macrophages were evenly distributed within the anterior corneal stroma one month after CXL treatment. In particular, M1/M2-type macrophages increased after CXL treatment. This suggests that M1-type macrophages play a crucial role in the rapid immune response after CXL treatment, leading to increased fungal susceptibility and more severe fungal keratitis. Luke found that macrophages have alternative functions.³⁸ As well as participating in responses to tissue remodeling and angiogenesis, they play a central role in the immune inflammatory response, with dynamic interactions between the functions.³⁸ CXL can damage corneal cells and immunocytes while inducing collagen fiber reactions and enhancing corneal mechanical strength. 

Macrophages participate in the repair of corneal injury after CXL and then colonize the corneal stroma over a long time. Interestingly, Polarization of M1 to M2 macrophages can modulate inflammatory responses and accelerate diabetic wound repair.³⁹ Topical calcitriol application promoted corneal wound healing and nerve regeneration by ameliorating neutrophil infiltration and promoting the M1-to-M2 macrophage transition in diabetic mice.⁴⁰ However, when the cornea is infected by microbes, macrophages transform and play an important role in the immune inflammatory response, leading to an excessive immune response and exacerbation of corneal inflammation.⁴¹ Hu et al.¹⁹ demonstrated that the activation of macrophages in the cornea may cause an excessive immune response. We also noted that M1- and M2-type macrophages increased simultaneously in the CXL + inoculation group. It is known that M1-type macrophages are pro-inflammatory cells and M2-type macrophages are anti-inflammatory. Nonetheless more M1-type than M2-type macrophages were activated resulting in an increased M1/M2 ratio.^42,43 An imbalance between the pro-inflammatory and anti-inflammatory systems of macrophages may contribute to more severe fungal keratitis. From the above discussion, it can be concluded that CXL treatment may change fungal susceptibility and increase the severity of keratitis by activating more M1 type macrophages. 

In conclusion, we demonstrate that the rat cornea is more likely to develop fungal keratitis after CXL treatment and that the fungal keratitis is more severe. In addition, M1 inflammatory factor chemotaxis macrophages might be one of the components of fungal susceptibility after CXL. Additional studies are required to determine whether interfering with M1 macrophages can reduce or alleviate corneal fungal infection after CXL. 

Acknowledgments

Supported by the National Natural Science Foundation of China (81970770 to Wei Chen). 

Disclosure: F. Peng, None; Q. Xie, None; J. Chen, None; Y. Fang, None; W. Xu, None; D. Jiang, None; W. Chen, None