Laser Confocal Microscope Results
The number of wing cells in corneal epithelium and corneal endothelial cells showed no significant differences in the NW, 5W, 5–10W, and O10W groups before surgery (P = .30 and .80, respectively). The CNFL is shown in Figures 1A–1D and Figure 1M. The CNFL in the NW group was significantly higher than those of the other three groups (P < .01). The CNFL in the 5W group was not significantly different from that of the 5–10W group, but was higher than that of the O10W group (P < .01). The CNFL of the O10W group was also substantially less than that of the 5–10W group (P < .01). The value of CNFL showed a significantly downward trend with extension of wearing time.
NCASC in these four groups (Figures 1E–1H, Figure 1N) exhibited significant changes. The NCASC in the NW group was significantly higher than that of the other three groups (P < .01). The NCASC of the 5W group was significantly higher than that of the 5–10W (P = .01) and O10W (P < .01) groups. The NCASC in the O10W group was significantly less than that of the 5–10W group (P < .01). The NCASC also showed a significantly downward relation with the extension of wearing time. Figures 1I–1L and Figure 1O show the NCPSC in these four groups. NCPSC in the O10W group was significantly less than those of the other three groups (P = .01). However, there were no significant differences in the NW, 5W, and 5–10W groups.
Laser scanning confocal microscope data suggested that there were pathological changes that occurred in corneal stroma after wearing soft contact lenses. The decreased CNFL suggested that growth of nerve fibers was altered in the central corneal subcutaneous region. Longer wearing time correlated with more significant impact. Furthermore, the number of corneal anterior stromal cells and corneal posterior stromal cells were substantially altered in groups of soft contact lens wearers along with wearing time.
Multivariate Statistical Analysis of HPLC-TOFMS Data
Because 15,340 ions in both positive and negative modes were detected in the HPLC-TOF-MS analysis, it was difficult to identify similarities and differences among the NW, 5W, 5–10W, and O10W groups using traditional statistical methods. Multivariate statistical analysis such as PCA or PLS-DA are important tools for exhibiting patterns of metabolites in various corneal stroma samples. In this study, PCA was first performed using SIMCA-P software. The PCA plot of Figure A (available in the online version of this article) showed that the samples from the NW, 5W, 5–10W, and O10W group samples could be substantially distinguished. The samples of the 5–10W and O10W groups overlapped. The PCA result demonstrated that metabolites among the NW, 5W, 5–10W, and O10W groups were substantially different.
To identify those metabolites (biomarkers) that contributed to group differences, the supervised multivariate statistical analysis method PLS-DA was used for further analysis. First, to identify the various metabolites between the NW and soft contact lens wearing groups (5W, 5–10W, and O10W), a PLS-DA method was established (R2X = 0.848, R2Y = 0.986, and Q2Y = 0.831). As Figure BA (available in the online version of this article) shows, there was a distinguished classification between the clustering of the NW group samples and other groups' samples. In Figure BB, the corresponding loading plot shows several triangles, and each triangle represents an ion (variable). An ion away from the center indicates that the ion abundance was substantially altered between these two groups. The ability of contribution for the group classification was evaluated using variable importance projection (VIP) in Simca P software. When the VIP is 1.0 or greater, the ion could be considered a potential metabolite bio-marker between the NW group and all other groups.
A total of 20 ions (VIP of 1.0 or greater) of 15,340 were shown to contribute to the classification of the groups. The 20 substantially altered variables11 were presumed according to accurate MS and MS/MS fragments by searching in metabolite databases ( http://metlin.scripps.edu, http://www.hmdb.ca/) and then confirming using commercial standards (Table 4).
Using the same processes mentioned above, the classification of the 5W, 5–10W, and O10W groups are shown in Figure C (available in the online version of this article). In the end, six potential biomarkers were identified as contributors for group classification. Bio-marker information is shown in Table 5.
Related Pathological Process of the Identified Biomarkers and Their Functions
Table 4 displays the metabolites that were significantly altered between the NW group and the other groups. These metabolites are divided into three classes: short chain organic acids, long chain unsaturated fatty acids, and lipids, suggesting that short chain organic acids metabolism, fatty acid metabolism, and lipid metabolism in the corneal stroma were dysfunctional in soft contact lens wearers.
Citrate, oxaloacetate, succinate, pyruvate, and glutamate are all short chain organic acids that were significantly upregulated in our study. The related biological pathways of these five metabolites all participate in the citrate cycle process, suggesting that citrate cycle metabolism was severely disturbed by wearing soft contact lenses. Corneal health relies on a well-balanced avascular oxygen supply. Wearing soft contact lenses reduces the aerobic respiration of glucose. Therefore, the cornea resorts to anaerobic respiration for its energy needs.18 Dysfunction of aerobic respiration of the citrate cycle then occurs. The high levels of citrate, oxaloacetate, succinate, and pyruvate in this study further demonstrate that aerobic respiration of the cornea was inhibited, leading to citrate cycle metabolite accumulation.
Behenyl palmitate, cholesteryl oleate, oleyl palmitate, and oleyl oleate are all lipids. These lipids contribute to form the outermost layer of tear film and help to slow evaporation of the aqueous layer in tear film. After wearing soft contact lenses, lipid metabolism is affected. The low levels of behenyl palmitate, cholesteryl oleate, oleyl palmitate, and oleyl oleate suggest that the synthesis of lipids after wearing soft contact lenses was inhibited. The resulting reduction in the protective function of the lipid layer leads to eye discomfort.
Taurine is the most abundant short chain organic acid found in ocular tissue.19 It inhibits the proliferation and migration of corneal stromal cells and protects against retinal and optic nerve damage.20,21 In this study, the level of taurine was significantly down-regulated, thereby enhancing the proliferation and migration ability of corneal stromal cells. Conversely, low levels of taurine in soft contact lens wearers may reduce an important nutrient substrate for the corneal nerve. This will be verified in a future study.
Arachidonic acid is an inflammatory factor. It is converted to downstream mediators such as prostaglandins and leukotrienes that further fuel the inflammatory cycle.22 The high levels of arachidonic acid in this study suggest that inflammatory activity existed in the corneal stroma of soft contact lens wearers.
Table 5 shows that metabolites were significantly altered between the 5W, 5–10W, and O10W groups. L-carnitine plays an important role in maintaining the ocular surface microenvironment. Khandekar et al23 reported that L-carnitine regulated human corneal epithelial cell volume and ameliorated apoptosis under hyperosmotic stress. Hua et al24 demonstrated that L-carnitine protected human corneal epithelial cells from oxidative stress by reducing declines in antioxidant enzymes and suppressing reactive oxygen species production. The inhibitory effects further reduce membrane lipid oxidative damage and protect the integrity of the tear film lipid layer. The substantially lower levels of L-carnitine in this study suggest that the corneal stroma or ocular surface lacked protective substances, possibly resulting in ocular surface diseases.
In the ocular surface, the glucose content is approximately 40%. High glucose levels in the ocular surface may facilitate the growth of pathogenic micro-organisms and alter the stability of the precorneal tear film.25 In this study, high levels of glucose in the 5W, 5–10W, and O10W groups suggested that pathogen invasion risk was greater.
Prolonged use of soft contact lenses can alter corneal innervations.26 Neuroprotectin D1, biosynthesized from docosahexaenoic acid, which was downregulated in this study, has anti-inflammatory and neuroprotective actions.27,28 In this study, with decreased docosahexaenoic acid levels, the synthesis of neuroprotectin D1 will be affected. Therefore, the anti-inflammatory and neuroprotective effects would be weakened in those wearing contact lenses for more than 5 years.
Sphinganine participates in sphingolipid metabolism. Sphingosine 1-phosphate is a metabolite of sphinganine that participates in neuroactive ligand-receptor interactions. The low levels of sphinganine in this study further suggest that the neuroprotective effect was weakened after wearing soft contact lenses.
Palmitoleic acid and linoleic acid are both long chain unsaturated fatty acids. Linoleic acid is the precursor of arachidonic acid. Because high levels of arachidonic acid are markers of inflammation,22 low levels of linoleic acid will help reduce the development of inflammation.