Anonic Silicon Hydrogels Affect the Concentration of Proteins in Tears during Wear

Abstract Purpose The objective of this study was to quantitatively assess the concentration of human tear proteins in patients wearing contact lenses of various ionicities and determine whether differences were related to the incidence of corneal infiltrative events (CIE). Methods 24 subjects (samples) were randomly selected for spectral count analysis to obtain protein concentrations using LCMS analysis. The subjects were neophyte and ametropic with ages between 18 and 40; 6 wore control lenses, 8 wore TestLens1, and 10 wore TestLens2. 16 subjects experienced CIEs during the study. Results A pairwise multiple hypothesis test identified 7 proteins that significantly differed in concentration between TestLens1 and control, and 11 proteins that differed between TestLens2 and control. Of the 12 unique proteins, 9 were at increased concentration and 3 were at lower concentration in the tears of test lens wearers compared to the control lens group. Bootstrap clustering confirmed these findings, showing 3 similar clusters to the original sample groups which separated people wearing control lenses from those wearing TestLens1 or TestLens2 with 83% accuracy and between TestLens1 and TestLens2 with 45% accuracy. Permutation testing identified 5 proteins that had significantly changed in concentration between people wearing TestLens2 and Control lenses. There was no difference in protein concentrations between those subjects who experienced a CIE and those who did not. Conclusion Wearing contact lenses of different ionicities can affect the concentration of proteins in the tear film. The current study did not find any associations of the concentration of proteins with CIEs. Future tests with increased sample size are needed to establish any relations between these changes and clinical performance.


Introduction
About 45 million people in the U.S. wear contact lenses; 8% are under 18 years, 17% are between 18 and 24 years, and 75% are 25 years and older; and 90% of contact lens wearers use soft contact lenses. 1 Contact lenses correct vision problems, including myopia, hyperopia, astigmatism, and presbyopia.Specially designed contact lenses can improve ocular vision experience. 2 Therefore, there is a continuous effort to improve the contact lens user experience by refining lens formulations.
However, there are also risks associated with contact lenses, from non-compliance to prescribed lens care and handling procedures or wear regimens to individual's incompatibility with particular lens materials.This risk could also be due to disproportional up-or down-regulation of immune and inflammatory responses 3 associated with tear proteins.The contact lens formulation could play a crucial role in regulating the immune and inflammatory responses.
In the context of contact lens wear, the interaction between the contact lens material and the tear film may happen quickly, resulting in the deposition of proteins and lipids within a short period of time.The uptake of these tear film components by the contact lens material is influenced by factors such as their charge, size, as well as the properties of the lens material itself, including its chemical composition, water content, ionic charge, and pore size.A more comprehensive explanation of these concepts can be found in other sources. 4he present study assessed two investigational contact lenses which were compared of the same based polymer material of the control silicone hydrogel lenses, but with increased anionicity.Silicone hydrogel contact lenses are known for enhanced oxygen permeability.However, these lenses have been associated with higher levels of corneal inflammatory events than regular hydrogel lenses 5, . 6This underscores a critical need to evaluate the mechanical, chemical, and inflammatory stressors imposed by lens materials on the cornea. 7Furthermore, the anionic hydrogel contact lens material etafilcon A has been used successfully by wearers for over 30 years and one study has shown that there is a decreased risk of developing corneal infection (one form of corneal infiltrative event [CIE]) when wearing etafilcon A lenses compared to certain non-ionic hydrogel lenses. 8This might imply that the anionicity of lens materials can influence the rate of CIEs 9, . 10o better understand how lens ionicity affects tear proteins or the rate of corneal inflammatory events, the differences between the study lenses were quantitively assessed by evaluating the concentration of proteins in the tears.Human tear proteins are objective, quantifiable characteristics of biological processes. 9Approximately 1500 tear proteins have been reported in humans. 10These proteins help modulate the normal ocular function and have been associated with ocular disease states. 11herefore, the current study aimed to determine whether there were reproducible robust changes in the concentration of proteins in tears of wearers of non-ionic and anionic silicone hydrogel lenses of different ionicity, and whether these lenses had different rates of CIEs.

Study design
The original clinical trial was a prospective, randomized, double-masked, 4-arm parallel, nine visits design with a 2week adaptation and 7-week post-adaptation period (ClinicalTrials.govIdentifier: NCT02543528).The trial was conducted in compliance with the study protocol, the International Conference on Harmonization Good Clinical Practice E6 (ICH-GCP), the Declaration of Helsinki, and all applicable regulatory requirements.
A total of 171 subjects were enrolled in the study, 142 subjects were dispensed contact lenses for the extended wear period (6-night/7-day replacement modality for 6 months), and 58 subjects completed all nine visits.All subjects participated in a 2-week adaptation period of contact lens wear before being randomized into one of the four extended-wear study lenses.The scheduled study visits were conducted over approximately 180 days.The lenses were replaced with a fresh pair after 6-nights/7-days of wear.
Additionally, the subject's spherical equivalent distance refraction was between −1.00 to −5.50 diopters, the refractive cylinder was less than 1.00 diopters and had the best-corrected visual acuity of 20/30 or better in each eye.The control lens was a commercially available silicone hydrogel lens, whereas the test lenses were modifications of the control lens's base polymer to increase its anionicity by 5% and 10%.
For this manuscript, tear films of 24 subjects (or 24 samples) were randomly collected from neophyte (no history of CL use) and ametropic subjects with normal eyes ages between 18 and 40 and used for the ex-vivo spectral analysis to obtain protein count.Each of these subjects wore the randomized study contact lens on an extended wear basis of a 6-night/7-day replacement modality for 6 months.Tear film samples were collected by capillary tubes and stored at −80 � C for ex-vivo analysis.Out of these 24 subjects, 6 wore commercially available control silicone hydrogel lenses, 8 wore TestLens1 (5% increased anionicity) lenses, and 10 wore TestLens2 (10% increased anionicity) lenses.Of the 24 subjects, 16 went on to experience a corneal infiltrative event (CIE) during the study period were labelled as symptomatic lens wearers.The subjects who did not experience a CIE (N ¼ 8) were labelled as asymptomatic subjects.All tears were collected before the CIE.
Bradford Assay for protein quantification and normalization: Prior to sample processing for LCMS analysis, the total protein content in the tear fluid was estimated in mg/ml by standard Bradford assay.A known concentration of the total protein from the Bradford assay was considered for further processing by LCMS. 10 ml of the extracted tear sample was added to 790 ml of double distilled water.To this solution, 200 ml of Bradford reagent was added to make a final volume of 1.0 ml.This solution was vortexed thoroughly and incubated in the dark at room temperature (RT) for 5 min allowing the Bradford reagent to interact with the basic amino acids in the tear sample resulting in a colorimetric reaction.The absorbance of the solution was measured in a UV-Spectrophotometer at a wavelength of 595 nm (k595) against a sample blank containing 10 ml of extraction buffer i.e. tri-ethyl ammonium bicarbonate.The absorbance was measured in triplicate and the values similar to the second decimal among the triplicates were only considered for further calculations.The total protein concentration (x) in the tear sample was calculated by incorporating the absorbance in the quadratic equation y ¼ mx þ c generated from the standard graph using exponentially increasing concentrations of bovine serum albumin (BSA), where y was the sample absorbance, m was the slope and c was the intercept for the standard BSA.The protein concentration determined for 10 ml of tear extract thereby was adjusted to a volume of 1 ml to achieve the protein concentration in mg/ml.

iTRAQ labeling
For LCMS analysis, the tear fluid proteins were reduced for 1 h at 37 � C and further alkylated for 10 min at RT prior to proteolysis by trypsin.For efficient digestion, the protein to enzyme ratio was maintained at 1:15 and incubated at 37 � C for 16 h.iTRAQ labeling of equal concentrations of the protein from paired samples were performed as per our experimental design and as described in the iTRAQ labeling kit (Sciex, Framingham, MA, USA).The digested peptide samples from control and patient samples were labeled with isobaric labels (114, 115, 116 etc-iTRAQ) and incubated at RT for 1h.The labeled samples were mixed in equal quantities and subjected to fractionation by strong cation exchange columns.The fractionated samples were purified using C18 desalting tips before subjecting to LCMS analysis.

LCMS analysis
The pooled labeled samples were then analyzed on Sciex Triple TOF 5600 þ mass spectrometer coupled with eksigent ekspert 425 nano LC.Approximately 1-2 mg of digested and labeled peptides were resolved on eksigent 3C18-CL-120, 3 mm, 120 Å, 350 mm � 0.5 mm 0.075 mm � 150 mm analytical column coupled with eksigent NanoLC Trap ChromXP CL-3 mm, 120 Å, 350 mm � 0.5 mm at a flow rate of 250 nl/ min over a gradient of 90 min using 95% water, 5% acetonitrile and 0.1% formic acid as aqueous phase/buffer A and 95% acetonitrile, 5% water and 0.1% formic acid as organic phase/buffer B. The full scan MS spectra were acquired within a mass range of 350-1200 m/z at a resolution 35,000 FWHM and the fragment ion spectra (MS/MS) were acquired in high sensitvity mode within a mass range of 100-1800 m/z.The resultant fragment spectra of proteins were analyzed against the uniprot human canonical database using protein pilot v 4.5 software.LCMS analysis of the iTRAQ labeled cell lysate digests were carried out in triplicates.The proteins with � 1.2-fold iTRAQ ratio and a pvalue of �.05 were considered as statistically significant.

Data exploration and visualization
For the count data, the expected variance grows with the mean.Thus, performing analysis directly on the counts may skew the results since some proteins with higher counts are substantially different from the other samples.Therefore, the counts data were transformed to perform exploratory analyses.There are various methods of transforming count data; however, regularized logarithm transformation 12 (rlog) was used to ensure variance stabilization across samples (details are in Supplementary Information [SI] Section a, Figure S1).

Sample-to-sample similarities
Heatmap (Figure 1) was constructed to assess the quality of the data and the similarity between the samples.The heatmap was constructed using rlog transformed data.The rows present the proteins, and the columns represent the samples grouped by lenses and strata (Symptomatic and Asymptomatic subjects).To produce the heatmap, the distances between the samples were computed, then the samples were grouped by the shorter distances.The amount by which each protein deviates from its average over the samples was displayed to observe the proteins' diversity.Additionally, principal components analysis (PCA) was conducted to quantitively observe the sample similarities and the data structure.

Statistical analysis
Data were analyzed using R (https://www.r-project.org/).In the dataset, there were two factors: strata with two levels (Symptomatic and Asymptomatic), and lenses with three levels (Control, TestLens1, and TestLens2).To identify significantly differed proteins, statistical models using factor levels and interaction between the lenses and strata was constructed.The statistically insignificant factors were sequentially removed using likelihood ratio tests to obtain parsimonious models for proteins.Consequently, only lens types were found statistically significant.Therefore, pairwise tests between the lenses were subsequently conducted to identify significant proteins.
For the statistical modeling, the distribution of the count data was assumed to be negative binomial to account for over-dispersion, which is inherent to the count data.Multiplicity correction was adjusted using covariate weighting [13][14][15] to identify significant proteins while ensuring the false discovery rate is controlled at 5%.The covariate weighting method uses independent information to boost the power of detectability of the true effects by reweighting statistical p-values when there are many correlated hypothesis tests involved while ensuring false discoveries are controlled.

Empirical reproducibility
The empirical reproducibility of the results obtained from the data analysis was verified using bootstrap clustering and permutation tests.Bootstrap clustering was used to verify sample classification and the permutation tests were used to verify significantly differed proteins.

Data summary
A numerical summary of the transformed data (rlog transformation, see Section 2.2) of the count by lenses and strata (Symptomatic and Asymptomatic subjects) of the samples is presented in Table 1.See Figure S2 and Table S1 in SI for additional information.The distribution patterns of the sample groups were almost the same.Means are higher than the median, and the data are mostly skewed toward the right.
In Table 1, the mean count and standard deviation (mean [Sd]) were calculated for all samples (24 subjects) and proteins (141 proteins), resulting in values of 9.97 [52.2].Within the Asymptomatic subjects' group, which consists of 8 samples across 141 proteins, the mean count was 10. 34 [54.92].Similarly, the Symptomatic subjects' group, comprising 16 samples across 141 proteins, had an average count of 9.79 [50.79].It's important to note that in the dataset, the columns represent the samples/subjects, the rows represent the proteins, and the counts correspond to the cell values.Thus, there are a total of 24 columns for samples and 141 rows for proteins in the dataset.
The groups Control, TestLens1, and TestLens2 each had 6, 8, and 10 samples respectively.The overall means for these groups were 13.37 [50.79], 9.09 [51.91], and 8.64 [53.19].These means were calculated using all 141 proteins.Importantly, there were no statistically significant differences observed among the corresponding sample groups.
Subject058, a Symptomatic sample, had the highest mean count of 23.72 [68.49] compared to the other samples, with a median count of 2 over the 141 proteins (Table S1 in SI).Only 3 subjects/samples exhibited a non-zero median count with means exceeding 5 across the 141 proteins.
Conversely, only 13 proteins had median counts above 5 across the 24 subjects [samples] (Table S2 in SI).Furthermore, it is worth mentioning that certain samples (subjects) exhibited strong positive correlations with other samples, with correlation values ranging from 80% to 99%, which were statistically significant (Figure S3 in SI).

Sample-to-sample similarities
The heatmap and principal component analysis identified three distinct sample groups that matched the original three lens types.The majority of proteins did not show any significant differences across the samples.It is conceivable that there may be variations in protein concentration between the control lens and test lenses, but no discernible discrepancies were observed among the test lenses or between Symptomatic and Asymptomatic subjects/samples within the lenses.The deviation of protein concentration from the average concentration ranged from −2 to 8. Since clustering is only relevant for proteins that carry signals (i.e. the proteins that demonstrate potential significant biochemical activities), only the top 40 proteins with the highest variance across the samples was displayed in Figure 1.Of these 40 proteins, 12 were found to be statistically significant.To see the heatmap for all 141 proteins, refer to Figure S4 in the SI.The PCA plot indicates little difference between the strata (different plotting shapes), but there are noticeable differences between the test lenses and the control lens (different colors).This difference suggests that the variation in the data is primarily explained by the lens groups.

Interaction between strata and lenses
Interactions between the strata (symptomatic and asymptomatic subjects) and lens types were also examined.The interactions of the top 12 proteins were presented in Figure 2. The concentration of proteins varied between subject/ sample groups and lenses.For example, the change of protein concentration of the Immunoglobulin Heavy Chain (A0A0A0MS08_HUMAN [5]) for the test lenses between the strata (Symptomatic vs. Asymptomatic) is almost zero; however, it differs for the control lens.On the other hand, for the protein Eosinophil Cationic Protein (ECP_HUMAN), the change of concentration for the control lens between strata is almost zero, whereas the changes for the test lenses differ.The change of protein concentration between the lenses and between the strata across proteins is also not consistent.These findings suggest a possible interaction between strata and lenses exists, and the interaction could be essential in interpreting the results.

Pairwise tests
The pairwise t-tests between the lenses with the covariate weighting [13][14][15] were applied to identify the significantly differed proteins between the lenses.Tests between Symptomatic vs. Asymptomatic were also conducted to quantify the protein concentration between the strata, although they were not  The multiple hypothesis tests between the lenses across all proteins identified 7 significant proteins between TestLens1 and Control (Table 2) and 11 significant proteins between TestLens2 and Control (Table 3), of which 12 were unique proteins.The comparison of TestLens2 vs. TestLens1 and Symptomatic vs. Asymptomatic subjects/samples did not reveal any significant proteins.However, it is worth mentioning that there were 4 additional significant proteins identified between TestLens2 vs.Control (Table 3) compared to TestLens1 vs.Control (Table 2).Furthermore, there were 3 proteins with small p-values [<.05] (Table S4 in the SI) between Symptomatic vs. Asymptomatic subjects prior to multiplicity correction.

Diagnosis of the models
Protein data often exhibit large counts, which can result in potential outliers.Outliers can arise from various factors, including technical issues, read mapping problems, and biological variations.Cook's distance was utilized to identify and detect outliers.As shown in Figure 4, Cook's distance was used on the transformed data (rlogtransformation of the protein count) to assess the TestLens2 vs.Control test.A few samples, such as subject058, subject107, and subject138, appeared to have outliers, accounting for approximately 13% of the 24 total samples.Therefore, the overall mean samples were trimmed for the large Cook's distance and scaled up by the size factor for that sample in the model.This approach replaced the outlier with the value predicted by the null hypothesis and does not lead to false positives.A similar procedure was adopted for the other comparisons as well.

Bootstrap clustering
Heatmap and PCA were conducted earlier (Section 2.3 and 3.2) to understand similarities between samples within and between groups (lenses and strata).The reproducibility of the results was verified by bootstrap clustering (for details, see section [f] in the SI).
The bootstrap cluster (BC) analysis separated the Control samples from the TestLens1 and TestLens2 samples with an accuracy of 83%.However, the accuracy in distinguishing between TestLens1 and TestLens2 was lower, at 45%.In Figure 5, five nodes (1, 2, 5, 13, and 14) were identified as reliable with respect to the original data, out of a total of 22 nodes.Furthermore, nodes 19, 20, and 22 appeared to differentiate the samples into three distinct groups, although it is possible that there could be more clusters.It is important to note that this finding does not perfectly match the original grouping depicted in Figure 1, but it provides evidence of similar grouping patterns.

Permutation test
Permutation tests were utilized to validate the reproducibility of the previously identified significant proteins between the lenses, employing multiple hypothesis tests (Section 2.4 and 3.4).Table S6 in the Supplementary Information (SI) presents the number of findings obtained through different methods.The permutation method specifically found significant results for the comparison between TestLens2 and Control, identifying 5 proteins that exceed the threshold of 3.6 (Figure S7 in SI) in terms of their absolute test statistics.These proteins are Albumin, Cathepsin G, Eosinophil Cationic Protein, Membrane Associated Phospholipase A2, and Vegf Coregulated Chemokine 1 and the corresponding test statistics are −3.69,4.29, 4.19, 3.69, and 3.68, respectively (Table 3).It is important to note that this discovery is particularly stringent due to the nature of the permutation test.Further details on this test can be found in section [g] of the SI, and specific test statistics can be referenced in Table 3. Permutation tests rely on the assumption of exchangeability of observations, which in turn requires equal variance for difference tests.Nonetheless, these findings confirm that there are significant variations between the test lenses and the control, likely stemming from higher ionicity in the test lenses.

Discussion
The objective of the study was to identify proteins that were either up-regulated or down-regulated as a result of using investigational lenses with higher ionicity.The purpose was to determine whether the anionicity of silicone hydrogel lenses affected protein deposits and consequently reduce the CIEs.The findings revealed that there were 7 significant proteins identified in the TestLens1 vs.Control comparison and 11 significant proteins identified in the TestLens2 vs.Control comparison, with 12 proteins being unique.It was observed that the proteins identified in the TestLens1 vs.Control comparison were also identified in the TestLens2 vs.Control comparison, except for Type I Cytoskeletal 9 Keratin.
Among the up-regulated proteins, Cathepsin G is known as inhibited trypsin-like serine proteases that activate protease-activated receptors and are involved in many inflammatory and extracellular matrix-degrading processes. 16athepsin G is associated with the neutrophil serine proteases family which helps kill pathogens.It affects inflammation and immune reaction and may work as pathogenesis of some autoimmune diseases. 17However, there is limited research on the specific role of Cathepsin G in dry eyes or discomfort.
Eosinophil cationic increases postoperative tear ECP levels and papillary excision.This protein comprises 30% of the eosinophil granule matrix.The principal eosinophil granule protein may have a negative impact on the corneal epithelium and contribute to corneal ulceration in severe cases of ocular allergy. 18The serum ECP was more influenced by associated major allergic manifestations such as eczema or asthma, which were nearly ubiquitous in atopics and close to absent in non-atopics, than by the conjunctival inflammation. 19inesin family member 20b has a cell-autonomous role in cytokinesis; however, the role of KIF20B in cytokinesis is not known. 20KIF20B is required for efficient cytokinetic furrowing and timely abscission in human cells, and thus its depletion affects the speed of both furrow ingression and abscission.It may accelerate or coordinate midbody maturation and may regulate late steps of maturation, including anillin dispersal and the formation of microtubule constriction sites. 20The role of the Kinesin-like protein KIF20B in dry eye and eye comfort is not well-studied.
Membrane Associated Phospholipase A2 is known to catalyze the hydrolysis of phospholipid in tears. 21The concentration increase of phospholipase A2 in the test lenses could promote tear film stability by increasing the abundance of phospholipids in tears.The changes indicate alterations in tear film stability and interactions with the corneal surface in lenses.
Angiogenin's concentration in tears suggests that it may act as a defensive agent in pro-inflammatory conditions induced by prolonged eye closure.This may indicate that the decrease in ANG levels with the increase in dry eye symptom severity reflects the progression of the inflammatory responses. 22Angiogenin is useful in measuring dry eye syndrome severity through proteomic analysis and might be negatively correlated with dry eye severity. 23-C Motif Chemokine 28 is correlated with meibomian gland yield secretion score (MGYSS).It significantly relieves both signs and symptoms of Meibomian gland dysfunction (MGD) due to a significant decrease in tears in patients with MGD. 24Serum levels of C-C motif chemokine were evaluated in patients with Sjogren's Syndrome, and there was a link between C-C motif chemokine and disease activity. 25CCL28 levels positively correlated with corneal staining, dendritic cell density and thus might play an important role in dry eye. 26ho GTPase-Activating Protein 7 (DLC1) was established as a regulator of yes-associated protein1 (YAP1) and deficiency of a different tumor suppressor gene, DLC1, enhances YAP stability and function, contributing to endothelial cell dysfunction. 279][30][31] In cancer cells, the DLC1 level is often low or lost due to genetic or epigenetic mechanisms, and reintroduction of DLC1 reduces cancer cell growth. 29While the specific role of Rho GTPase-Activating Protein 7 in dry eye and eye comfort is not well-documented, may have been implicated in the regulation of tear film homeostasis and ocular surface integrity.
Antileukoproteinase has been identified in secretory cells in lacrimal glands. 32Antileukoproteinase is a secretory leukocyte peptidase inhibitor with antimicrobial, antiviral, and antifungal properties 33, . 34Recently Cecilia Chao et al.. 35 identified it is significantly higher in adult orthokeratology wearers.It can suppress inflammatory responses and allergic reactions by inactivating Matrix metallopeptidase 9.36 Therefore, it further supports the clinical implication that Anonic Silicon Hydrogels lens wearers may have a lower risk of corneal infection or inflammation.
VEGF Coregulated Chemokine 1 (Vascular Endothelial Growth Factor [VEGF] coregulated chemokine [CXCL17]) concentration in tears can also be explained by the increased tear secretion.Although the level after one month of lens wear could play a role in immune signaling pathways, such level changes were not observed after three months of lens wear using pathway enrichment analysis. 37CXCL17 chemokine can recruit macrophages and dendritic cells and detect very high levels in the tears of patients with Primary Sjogren's syndrome (PSS). 38Tear VEGF levels are also significantly related to the severity of diabetic retinopathy patients. 39It was found that the concentration of VEGF is associated with ocular comfort.However, the change in VEGF concentration was more significant when contact lenses were not being worn, as opposed to when they were being worn. 40he 9 up-regulated proteins mentioned above have diverse functions and may be crucial for maintaining good eye health and combating diseases.These proteins are associated with various processes, including immune responses related to the neutrophil serine proteases family, conjunctival inflammation, and their role in cytokinesis, which is necessary for effective furrowing and abscission.Deficiency of these proteins can impact the speed and maturity of cytokinesis, potentially affecting tear film stability and interactions with the corneal surface.Additionally, reduced levels of ANG may indicate the progression of inflammatory reactions in dry eye syndrome.It is worth noting that this protein also exhibits tumor-suppressive functions, and its level is often low in cancer cells.Moreover, the up-regulation of these proteins in adult orthokeratology wearers may lower the risk of corneal infection or inflammation.
Among the down-regulated proteins, Immunoglobulin Heavy Chain (Ig gamma 1 chain C region [IGHG1]) was previously unreported by other proteomic studies of dry eye disease but has recently been shown by a study investigating dry eye subtypes. 41The study found reduced IGHG1 level in lipid-deficient dry eye and increased levels in aqueous-deficient dry eye.IGHG1 is involved in immune response, with roles in antigen binding, complement activation, and immune cell trafficking.
Albumin in tears increases with ocular disease and might serve as a biomarker for ocular insult. 42The presence of albumin in wounds reduces the antibacterial efficacy of various compounds.Serum albumin reduced the activity of highly protein-bound faropenem but not the activity of minimally bound imipenem; active serum restored the bactericidal activity of faropenem against E. coli. 43Down-regulated serum albumin using the test lenses over the control could suggest a reduction of corneal infiltrative events.
Type I Cytoskeletal 9 Keratin is known to occur only in the epidermis of palms and soles, and its presence in the normal tear band is unclear. 44However, similar keratins: type I cytoskeletal 16 and type II cytoskeletal 6 A have been detected in the ocular surface epithelia and may suggest increased shedding in contact lens-related dry eye 45, . 46This protein's significance was observed only in the comparison between TestLens1 and Control, but not in the comparison between TestLens2 and Control.It can be reiterated that the test lenses were modified versions of the control lens, with TestLens1 having a 5% increase and TestLens2 having a 10% increase in anionicity.Therefore, a balanced increment in ionicity of around 5% compared to the control may be necessary for this protein identification.
The down-regulation of these 3 proteins may potentially contribute to reducing corneal inflammation and adverse events.These proteins may also contribute to the development or progression of dry eye by impacting various cellular processes and pathways involved in ocular surface health and lubrication.These proteins could potentially affect tear film stability, inflammation, epithelial barrier function, and other factors related to ocular comfort.
The up-and down-regulation of proteins can be influenced by various factors beyond the ionic surface.We currently do not have knowledge about the specific impact of the ionic surface on protein levels.Moreover, the interpretation of the effect may be complicated by changes in the surface over time resulting from this interaction.To obtain a thorough understanding of these findings, it may be necessary to carry out extensive prospective longitudinal clinical trials on a large scale.
Furthermore, despite the fact that no significant proteins were detected when comparing TestLens2 to TestLens1, there were still more proteins found when comparing TestLens2 to the Control group in comparison with TestLens1 to the Control group.Additionally, there were 3 proteins with low p-values between Symptomatic and Asymptomatic subjects, indicating the possibility of significant proteins.However, these proteins were not identified due to the small sample size and low statistical power.

Conclusion
The study examines the tear proteins concentration of two investigational contact lenses with modified silicone hydrogel of varying ionicity compared to a control lens with a base silicone hydrogel formulation.The ionicity of TestLens1 is 5% higher and TestLens2 is 10% higher than the Control lens.
Quantitative evaluation suggests that these investigational lenses may have better immune and inflammatory responses, potentially leading to reduced corneal inflammation and adverse events compared to the control lens.These benefits could be attributed to the up-and down-regulated protein influenced by the test lenses' ionicity.The statistical tests' values reflect this finding quantitatively.

Figure 1 .
Figure 1.Heatmap of top proteins-concentration using the regularized logarithm transformation (rlogÞ values.The rows present the proteins, and the columns represent the samples grouped by lenses and strata. CURRENT EYE RESEARCHsignificant by the statistical model.The effect size estimates (log2FoldChange) of the proteins between the pairwise tests are provided in Figure3.The log2FoldChange demonstrates how much the protein concentration changed between the lenses or strata.The −log10(p-value) was plotted against log2FoldChange for better visualization.The log2FoldChange and the corresponding p-values greater than 0.6 and less than 0.05 before multiplicity correction are red.
Of 12 the unique proteins, Cathepsin G, Eosinophil Cationic Protein, Kinesin Family Member 20b, and Membrane Associated Phospholipase A2 were significantly up-regulated, and Immunoglobulin heavy chain and Albumin ware significantly down-regulated in both test lenses compared to the control.Additionally, Type I Cytoskeletal 9 Keratin was significantly down-regulated in TestLens1 only; and Angiogenin, C-C Motif Chemokine 28, Rho GTPase-Activating Protein, Antileukoproteinase, and Vegf Coregulated Chemokine 1 were significantly up-regulated in TestLens2 only compared to the Control.Among the 12 unique proteins, 9 were up-regulated and 3 were down-regulated between the test lenses and the Control.The up-and down-regulations are believed to be due to the different ionicity of the test lenses compared to the Control lens.

Figure 2 .
Figure 2. Interaction of the proteins between lens types and strata.The concentration level of some proteins is varied between strata and lens types.

Figure 3 .
Figure 3. Scatter plots of -log10(p-value) vs. log2FoldChange for strata and lenses.Red dots indicate fold changes of potentially significant proteins before multiplicity correction.Positive fold change favors the first group.

Figure 4 .
Figure 4. Box plots of the samples to detect outliers.

Figure 5 .
Figure 5. Hierarchical bootstrap clustering of 24 samples.The number at each node tells what percentage of the time that node appeared in the bootstrap replicate.It provides two types of percentages: AU (Approximately Unbiased) and BP (Bootstrap Probability).Clusters with AU larger than 95% indicate strong support from the data.

Table 1 .
Summary of the proteins counts of 24 samples.
NZ: number of non-zero counts out of N.