Comparative Analysis of Two Femtosecond LASIK Platforms Using iTRAQ Quantitative Proteomics

Sharon D’Souza,1 Andrea Petznick,2 Louis Tong,2–5 Reece C. Hall,3 Mohamad Rosman,2,3 Cordelia Chan,3 Siew Kwan Koh,2 Roger W. Beuerman,2,4,6 Lei Zhou,2,4,6 and Jodhbir S. Mehta2,3,5

1Narayana Nethralaya, Bangalore, India 2Singapore Eye Research Institute, Singapore 3Singapore National Eye Centre, Singapore 4Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 5Office of Clinical, Academic, and Faculty Affairs, Duke–National University of Singapore Graduate Medical School, Singapore 6Signature Research Program (SRP) Neuroscience and Behavioural Disorder, DUKE–National University of Singapore Graduate Medical School, Singapore

Correspondence: Lei Zhou, Singa- PURPOSE. New femtosecond laser platforms may reduce ocular surface interference and LASIK- pore Eye Research Institute, 11 Third associated dry eye. This study investigated tear protein profiles in subjects who underwent Hospital Avenue, Singapore 168751; LASIK using two femtosecond lasers to assess differences in protein expression. [email protected]. Louis Tong, Singapore Eye Research METHODS. This was a randomized interventional clinical trial involving 22 patients who Institute, 11 Third Hospital Avenue, underwent femtosecond laser refractive surgery with a contralateral paired eye design. Singapore 168751; Corneal flaps of 22 subjects were created by either Visumax or Intralase laser. Tear samples [email protected]. were collected preoperatively, and at 1 week and 3 months postoperatively using Schirmer’s SD and AP contributed equally to the strips. Tear protein ratios were calculated relative to preoperative protein levels at baseline. work presented here and should The main outcome measures were the levels of a panel of dry eye protein markers analyzed therefore be regarded as equivalent using isobaric tagging for relative and absolute quantitation (iTRAQ) mass spectrometry. authors. RESULTS. A total of 824 unique proteins were quantifiable. Tear protein ratios were Submitted: February 7, 2014 differentially regulated between the eyes treated with different lasers. The secretoglobulins Accepted: April 24, 2014 Lipophilin A (1.80-fold) and Lipophilin C (1.77) were significantly upregulated (P < 0.05) at 1 Citation: D’Souza S, Petznick A, Tong week postoperatively in Visumax but not in Intralase-treated eyes. At 1 week, orosomucoid1 L, et al. Comparative analysis of two was upregulated (1.78) in Intralase but not Visumax-treated eyes. In the same eyes, lysozyme, femtosecond LASIK platforms using , and lipo-oxygenase were downregulated at 0.44-, 0.64-, and 0.64-folds, iTRAQ quantitative proteomics. Invest respectively. Transglutaminase-2 was downregulated in both groups of eyes. Ophthalmol Vis Sci. 2014;55:3396– 3402. DOI:10.1167/iovs.14-14113 CONCLUSIONS. Different laser platforms induce distinct biological responses in the cornea and ocular surface, which manifests as different levels of tear proteins. This study has implications for surgical technology and modulation of wound healing responses. (ClinicalTrials.gov number, NCT01252654.) Keywords: refractive surgery, laser, , randomized controlled trial

ASIK is the most common refractive surgery procedure. Our laboratory has used tear proteomic analysis to discover L Although it has been shown to be safe, effective, and highly biomarkers for objective diagnosis of dry eye.7 Dry-eye patients predictable,1 dry-eye complaints or tear dysfunction are major displayed an altered tear proteome profile with upregulated side effects.2 The truncation of the corneal afferent nerves inflammatory markers and decreased levels of protective during flap creation results in corneal hypoesthesia and proteins.7,10,11 Specifically, a panel consisting of a-enolase, 3 disruption of the cornea-lacrimal gland unit. This, in turn, prolactin-inducible protein (PIP), lipocalin-1, S100A9 (calgra- leads to a reduced blink rate and changes in tear quantity as nulin B) has been shown to produce a high diagnostic accuracy well as tear hyperosmolarity and ocular surface inflamma- for presence and severity of dry eye.7 In a rabbit model of tion.4,5 It has not been clear if patients with LASIK-associated Sjogren’s Syndrome–associated dry eye,12 similar dry eye dry eye have inflammatory ocular surface changes as in pathological dry eye,6,7 or if these changes are more related biomarkers such as the S100 A9 and PIP were found. Refractive to the regenerative changes in the corneal nerves, which give surgery is expected to cause major changes in the tear protein rise to only sensations of pain and sensory irritation.8 The tear profile due to disruption of the cornea-lacrimal gland functional 13,14 film is an integral part of the ocular surface and represents the unit. Previously, it has been shown that differences in extracellular matrix for ocular surface epithelial cells. Here, surgical technique can result in differential tear protein analysis of the human tear proteome can provide both levels.9,15,16 These studies are, however, limited to a few qualitative and quantitative information that is useful in proteins such as nerve growth factor, or IL-6, -8, matrix assessing the health of the ocular surface.9 metalloproteinase 9, and epidermal growth factor only.15,16

Copyright 2014 The Association for Research in Vision and Ophthalmology, Inc. www.iovs.org j ISSN: 1552-5783 3396

Downloaded from tvst.arvojournals.org on 09/26/2021 Tear Protein Profile in LASIK IOVS j June 2014 j Vol. 55 j No. 6 j 3397

Newer femtosecond lasers create smaller flaps, use higher flap diameter of 8.4 6 0.0 mm for Visumax (nonconverted 7.9 repetition rates and lower laser pulse energies and use 6 0.1 mm) and 8.9 6 0.2 mm for Intralase. Laser parameters different corneal stabilization methods.17 These advances for Visumax flaps were: small (S) cone, superior hinge, 858- may result in fewer adverse consequences for the ocular sidecut angle, 708-hinge angle, laser bed energy 0.16 to 0.165 surface and may affect the expression of tear protein lJ, spot separation 1.5 (rim) and 4.8 lm (lamellar), and line differently. It has been shown that different femtosecond spot separation 1.5 (rim) and 4.8 lm (lamellar). Laser lasers can induce different wound healing responses in rabbit parameters for Intralase flaps were: standard suction, superior corneal tissue.18,19 hinge, 708-sidecut angle, 508-hinge angle, laser bed energy 0.95 A comparison between Intralase (Abott Medical Optics, lJ (range 0.92–1.08 lJ), pocket enable on, 0.25-mm pocket Inc., Santa Ana, CA, USA) and the newer femtosecond laser width, 230-lm pocket start depth, 7-lm pocket tangent, and 6- Visumax (Carl Zeiss Meditec, Jena, Germany) has shown that lm radian spot separation. both platforms have similar efficacy and predictability of visual Stromal tissue laser ablation in all eyes was performed with outcomes.17,20 The Intralase laser uses a lower repetition rate, the Wavelight Allegretto Eye-Q 400Hz excimer laser system higher energy and conjunctival suction, while the Visumax (Alcon Laboratories, Inc., Fort Worth, TX, USA) using the laser platform uses higher repetition rate, lower energy, and wavefront-optimized treatment profile. The attempted optical corneal suction.17 In a contralateral paired-eye study with 45 zone was 6.50 mm and target refraction was plano spherical myopic patients, no statistically significant differences in equivalent. corneal sensitivity, Schirmer’s test, tear break-up time (TBUT), All eyes were treated with nonpreserved moxifloxacin and corneal staining were identified between the Intralase and hydrochloride 0.5% eyedrops (Vigamox; Alcon Laboratories, Visumax platform over a 3-month period.20 In this comparison Inc.) and dexamethasone 0.1% eyedrops (Maxidex; Alcon the clinical signs, such as Schirmer’s, TBUT, corneal fluorescein Laboratories, Inc.) four times daily for 1 week. Nonpreserved staining, corneal sensitivity, and visual outcomes including artificial tears (Tears Naturale Free; Alcon Laboratories, Inc.) contrast sensitivity, were not sensitive enough to differentiate were prescribed postoperatively with a dosage adjusted to the laser technologies. The corneal sensitivity however, subject’s symptoms and needs. revealed slightly faster recovery rates in Visumax-operated eyes as compared with Intralase-operated eyes, which may be Tear Sample Collection attributable to the technical differences.17 We hypothesize that the tear proteome may be altered after Tear samples were collected using a Schirmer’s Type I tear test LASIK in a different way from previously reported profiles in without local anesthesia, as employed in previous studies in idiopathic dry eye, and the Visumax femtosecond laser our laboratory.7,11,21 No other drops were administered prior platform with different technical specifications may induce to the Schirmer’s test. Briefly, a precalibrated filter paper strip distinct protein changes from the Intralase platform. To address (Sno strips; Bausch & Lomb, Rochester, NY, USA) was gently these hypotheses, a proteomic study of tears was carried out placed in the inferior temporal fornix for 5 minutes, the eye with patients who underwent a randomized contralateral eye shut, and the patient told not to move the eye under the closed treatment using the Visumax and Intralase laser platforms and lids. After collection, the wetted area of the Schirmer’s strips evaluated at 1 week and 3 months postoperatively. were measured and the strip immediately frozen at 808C until analysis. To start the analysis procedure, Schirmer’s strips were cut into small pieces and soaked in 150 Lof50mM METHODS l ammonium bicarbonate and 1.33 protease inhibitor (Pierce; Subject Selection Thermo Scientific, Rockford, IL, USA) for 3 hours to elute tear proteins. Total tear protein concentrations of each tear sample Institutional review board approval for this study was obtained were measured using the RC DC Protein Assay (Bio-Rad from the Singapore Eye Research Institute ethics committee. Laboratories, Hercules, CA, USA). Tear samples were lyophi- This prospective contralateral paired eye clinical study adhered lized and stored at 808C until further analysis. to the tenets of the Declaration of Helsinki and was registered with the National Institutes of Health database (http://www. clinicaltrials.gov, record number: NCT01252654). Informed Quantitative Tear Protein Analysis written consent was obtained from 45 consecutive subjects Tear samples collected at the visit prior to surgery (day 0), and after explanation of the nature and possible consequences of then at 1 week and 3 months postoperatively were analyzed the study. For the purpose of this publication, tear protein using the proteomics study design shown in Figure 1. Samples profiles of 22 subjects were analyzed. were quantitatively analyzed using isobaric tagging for relative Inclusion criteria were as follows: greater or equal to 21- and absolute quantitation (iTRAQ)-based technology combined years old, stable myopic prescription, best corrected visual with one-dimensional nano liquid chromatography (LC)-nano- acuity of at least 20/20, minimum corneal thickness of 500 lm, electrospray ionization (ESI)-mass spectrometry (MS)/MS in preoperative Schirmer’s Type I test value greater than 5.0 mm triplicates. in 5 minutes, and cessation of soft contact lens wear for at least Data consisted of protein ratios for the 1 week postoper- 2 weeks or rigid gas-permeable lenses for at least 3 weeks. ative visit (preoperative to 1 week) and the 3-month Subjects were excluded if they had any anterior or posterior postoperative visit (preoperative to 3 months). Protein ratios segment pathology, previous ocular surgery, metabolic or larger than 1.300 were considered upregulated and below autoimmune disease, severe dry eye, connective tissue disease 0.769 downregulated. or significant atopic syndrome, and were undergoing chronic systemic corticosteroid or immunosuppressive therapy. Analysis of Tear Fluid Using Mass Spectrometry Surgical Technique and Liquid Chromatography Corneal flaps were created with a 500-kHz Visumax femtosec- Lyophilized tear proteins (25-lg total protein) were reconsti- ond laser (Carl Zeiss Meditec) or a 60-kHz Intralase femtosec- tuted, denatured, and reduced in 50 mM ammonium bicar- ond laser (Abott Medical Optics, Inc.). The targeted flap bonate solution (filter-aided sample preparation [FASP] kit; thickness was between 110 to 115 lm with a mean targeted Expedeon, Inc., Harston, UK) 2% SDS solution (Vivantis

Downloaded from tvst.arvojournals.org on 09/26/2021 Tear Protein Profile in LASIK IOVS j June 2014 j Vol. 55 j No. 6 j 3398

1250 with charge state of 2 to 5, and an abundance threshold of greater than 120 counts per second. Former target ions were excluded for 12 seconds and former ions were excluded after one repeat. Maximum number of candidate ions per cycle was 30 spectra. Information-dependent acquisition advanced ‘‘roll- ing collision energy (CE)’’ and ‘‘adjust CE when using iTRAQ reagent’’ were required. The data was processed and searched against the IPI Human version 3.77 protein database (115194 proteins searched) using ProteinPilot software 4.1 (AB Sciex). Some important settings in ProteinPilot were configured as follows: (1) sample FIGURE 1. Design of the iTRAQ experiments for the analysis of pre- and postoperatively collected tears. type: iTRAQ 4plex (peptide labeled), (2) Cys alkylation: MMTS; (3) digestion: , (4) instrument: TripleTOF 5600, (5) special factors: none, (6) identification focus: biological Technologies, Selangor, Malaysia), and tris(2-carboxyethyl)- modifications, (7) search effort: thorough identification, and phosphine (TCEP) from iTRAQ kit (AB Sciex, Framingham, (8) bias correction and background correction were applied. MA, USA) for 1 hour at 608C. The reduced protein samples Ninety-five percent confidence level was used at the peptide were then transferred to 30 kDa cut-off membrane cartridge level. False discovery rate (FDR) analysis in the ProteinPilot (FASP kit; Expedeon, Inc.) to remove the excess reagent using software was performed and FDR less than 1% was set for 75% urea solution. The reduced samples were then alkylated protein identification. Reverse database search strategy was using methyl methane thiosulfonate (MMTS; AB Sciex) for 20 used to calculate FDR for peptide identification. For relative minutes at room temperature. The alkylated-samples were quantification, ProteinPilot software uses Pro Group algorithm further washed with 75% urea solution and 50 mM ammonium to calculate the reporter ions’ (iTRAQ114, 115, and 116) peak bicarbonte solution prior to trypsin digestion overnight with areas. Auto bias correction was applied to eliminate possible substrate: enzyme ratio of 1:25 at 378C. Peptides were then pipetting error during sample preparation. eluted with 50 mM ammonium bicarbonate solution and sodium chloride (provided in FASP kit; Expedeon, Inc.). The Statistical Analysis eluted samples were lyophilized and tear samples were labeled with iTRAQ reagents for 3 hours at room temperature as Each sample was analyzed in triplicates by iTRAQ, and only follows: preoperatively collected tears with iTRAQ 114, 1 week data points that were within the 30% coefficient of variance postoperatively collected tears with iTRAQ 115 and 3 month value were used for the analysis, while outliers were filtered postoperatively collected tears with iTRAQ 117 (Fig. 1). out. Data was sorted according to upregulated proteins (ratio Labeled samples (iTRAQ 114, iTRAQ 115, iTRAQ 117) were > 1.300) and downregulated proteins (ratio < 0.769). A paired then pooled together, dried and desalted using ultramicro spin Wilcoxon test for nonparametric data was done on the log- columns (The Nest Group, Inc., Southboro, MA, USA) prior to transformed median values to identify significantly upregulated nano-LC/MSMS analysis. and downregulated proteins at 1 week and 3 months postoperatively for both femtosecond laser platforms using The one dimensional nano LC-MS/MS system (Dionex the R software (R Development Core Team, Vienna, Austria). A Ultimate 3000 Nano LC system; Thermo Fisher Scientific, further comparison with the paired Wilcoxon test was Sunnyvale, CA, USA), coupled with AB Sciex TripleTOF 5600 performed to compare differences between the log-trans- system (AB Sciex), was used for the proteomic analysis. A 50 formed median of tear protein ratios of Visumax-operated cm 3 75 lm (internal diameter) Dionex Acclaim PepMap RSLC and Intralase-operated eyes at the 1 week and 3 month C18 packed column was employed (Thermo Fisher Scientific). postoperative visits using the R software. The P value was This column was connected to a spray tip (New Objective, adjusted for multiple comparisons. A significance level (alpha) Inc., Woburn, MA, USA), which was directly coupled with the of 0.05 was used. nanospray interface of AB Sciex TripleTOF 5600 MS. Samples were loaded onto a trap column (Dionex Acclaim PepMap 100 C18, 2 cm 3 75 lm i.d.; Thermo Fisher Scientific) at a flow rate RESULTS of 5 lL/min. After a 3 minute wash with loading buffer (2/98 vol/vol of acetonitrile [ACN]/water with 0.1% formic acid) the Subject demographics including age, sex, and mean spherical system was switched into line with the C18 analytical capillary equivalent refraction are displayed in Table 1. Twenty-two column. A step linear gradient of mobile phase B (2/98 vol/vol patients completed all three visits and were included in this of water/ACN with 0.1% formic acid) starting from 7% to 24% analysis. A total of 1594 unique proteins (908 unique proteins for 57 minutes, to 24% to 40% for 27 minutes, to 40% to 60% with ‡2 peptides) with a false detection rate of less than 1% for 7 minutes, and 60% to 95% for 1 minute at a flow rate of were identified in tears by iTRAQ analysis. Among them, 1229 300 nL/min was used for this analysis. The typical parameters unique proteins (824 unique proteins with ‡2 peptides) were for TripleTOF 5600-MS were as follows: ionspray voltage quantifiable. The tears collected from Visumax-operated eyes floating (ISVF) ¼ 2400 V, curtain gas (CUR) ¼ 30, ion source gas showed greater number of upregulated proteins (eight 1 (GS1) ¼ 12, interface heater temperature (IHT) ¼ 1258C, proteins in Visumax versus two proteins in Intralase), while declustering potential (DP) ¼ 100 V. All data was acquired using the Intralase-operated eyes had a greater number of downreg- information-dependent acquisition (IDA) mode with Analyst TF ulated proteins (nine proteins in Intralase versus seven 1.5 software (AB Sciex). Time-of-flight mass spectrometry proteins in Visumax; Tables 2 and 3). (TOF-MS) scan (experiment 1) parameters were set as follows: Tears from Visumax eyes exhibited eight upregulated 0.25 seconds TOF-MS accumulation time in the mass range of proteins (protein ratio range of 1.3–3.0) after the first week 350 to approximately 1250 Da followed by product ion scan of surgery, with two proteins that remained upregulated (experiment 2) of 0.05 seconds accumulation time in the mass (protein ratio range of 1.7–2.2) at 3 months postsurgery (Table range of 100 to approximately 1500 Da. Switching criteria 2). Visumax showed four downregulated proteins (protein were set to ions greater than m/z ¼ 350 and smaller than m/z ¼ ratio range of 0.7–0.8) at 1 week postoperatively and four

Downloaded from tvst.arvojournals.org on 09/26/2021 Tear Protein Profile in LASIK IOVS j June 2014 j Vol. 55 j No. 6 j 3399

TABLE 1. Subject Data Demographics DISCUSSION Sex (n) This study carried out detailed tear protein profiling following Mean Age 6 SD, LASIK using two different femtosecond platforms, Visumax and Range, y Male Female Eyes (n) Mean SER 6 SD, D Intralase and identified differences in the relative quantity of 28.45 6 4.45 9 13 22 (Visumax) 4.47 6 1.71 proteins between the two femtosecond laser technologies. 22–39 22 (Intralase) 4.34 6 1.70 Numerous proteins were up- and downregulated in both platforms. However, the profile of dysregulated proteins was SER, spherical equivalent refraction; D, diopters. dissimilar between the two platforms, and somewhat different from some proteins reported previously in dry eye.7 The iTRAQ tagging technology coupled with nano-LC/MS downregulated proteins (protein ratio range of 0.6–0.8) after 3 has been shown previously to be useful for documenting months of surgery (Table 2). In tears from Intralase-operated changes in tear proteins in disease.7,9,11,21 Unlike previous eyes, two upregulated proteins (protein ratio range of 1.5–2.0) studies, we employed iTRAQ to compare three different were identified at 1 week postoperatively, while no proteins samples, one collected prior and two following LASIK surgery were found to be upregulated at 3 months postoperatively. In with either Visumax and Intralase lasers. This enabled a the same group, six proteins were downregulated (protein comparison of the two femtosecond laser platforms and their ratio of 0.4–0.7) at 1 week following surgery, and three impact on selected dry-eye proteins at two time points. To our proteins were downregulated (protein ratio of 0.5–0.7) at 3 knowledge, no previous studies that evaluated changes in tear months following surgery (Table 3). proteins after LASIK have been published. The present study A comparison between Visumax-operated and Intralase- identified upregulated tear proteins such as the inflammatory operated eyes was performed to detect proteins that were protein alpha1-acid 1 (ORM1) and the cystatin SN distinctly different between the two femtosecond laser precursor (CST1) that had previously been found to be platforms at 1 week and 3 months of surgery. The lacrimal dysregulated in dry eye. We also detected downregulation of gland protein (LACRT) had increased in tears collected proteins LYZ, TF, and secretoglobin family 1D member 1 from Visumax-operated eyes by more than two times, while in (SCGB1D1). Furthermore, the current tear analysis detected Intralase-operated eyes LACRT remained unchanged at 1 week novel changes of ALB, calgranulin A (S100A8), S100A9, and postoperatively (protein ratio of 2.228 in Visumax and 0.847 in mammaglobin B precursor (SCGB2A1). Tear LACRT was Intralase, P ¼ 0.008; Fig. 2). However, LACRT was not significantly higher in the Visumax group compared with the significantly upregulated in Visumax-operated eyes due its Intralase groups at 1 week and 3 months postoperatively. We identified upregulation of CST1 1 week postoperatively relatively large variability (Fig. 2). in the Visumax-operated eyes. Cystatin SN precursor is At 3 months following surgery, the protein ratios of LACRT, supportive of the innate immunity system,22 and has antimi- Clusterin (CLU), and Pre–pro-megakaryocyte potentiating crobial functions.23 It also has a protective function, acting as a factor (MSLN) were statistically greater in Visumax-operated natural inhibitor of the highly abundant cysteine proteinases. eyes as compared with Intralase-operated eyes (LARCT with Another cystatin, CST4, has been detected in elevated levels in 1.432 vs. 0.803, P ¼ 0.019; CLU with 1.343 vs. 1.028, P ¼ dry eye and dysfunction.24 The CST1 protein 0.008; and MSLN with 1.486 vs. 1.033, P ¼ 0.016 for Visumax was downregulated in tears collected from patients with fungal and Intralase, respectively; Fig. 3). The comparison also keratitis25 and keratoconus.26 We observed other evidence of showed that Intralase-operated eyes had significantly lower activated immune system, such as the elevation of immuno- amounts of Lactotransferrin (LTF), Alpha-2-glycoprotein 1 zinc globulins, polymeric immunoglobulin, Ig mu chain C region (AZGP1), Lysozyme (LYZ), and Lactoperoxidase (LPO) than IGHM protein (IGHM), and IG alpha 1 protein (IGHA1) in the Visumax-operated eyes (LTF with 0.973 vs. 0.686, P ¼ 0.023; Visumax-operated eyes. AZGP1 with 1.248 vs. 1.887, P ¼ 0.013; LYZ with 1.112 vs. A higher level of ORM1 was noted in Intralase-operated eyes 0.783, P ¼ 0.024; and LPO with 1.225 vs. 0.964, P ¼ 0.018 for at 1 week postoperatively. The inflammatory protein alpha1- Visumax and Intralase, respectively; Fig. 4). acid glycoprotein 1 is a constitutively expressed binding/

TABLE 2. Upregulated and Downregulated Proteins in Visumax-Operated Eyes at 1 Week and 3 Months Postoperatively

1wk 3mo

Gene Symbol Median (Q1–Q3) P Value Symbol Median (Q1–Q3) P Value

Upregulated proteins PIGR 3.011 (1.337–6.1) 0.001 PIGR 2.239 (1.1–3.251) 0.004 IGHM 2.599 (1.372–3.856) 0.001 IGHA1 1.706 (0.938–2.173) 0.022 IGHA1 2.403 (1.186–2.902) 0.001 CST1 2.065 (0.958–2.559) 0.012 SCGB1D1 1.804 (1.063–4.207) 0.017 IGLV3–19 1.776 (1.275–2.831) 0.001 SCGB2A1 1.765 (1.006–3.587) 0.005 CLU 1.343 (1.052–2.213) 0.002 Downregulated proteins S100A9 0.773 (0.633–1.054) 0.018 S100A8 0.78 (0.46–0.96) 0.015 ZG16B 0.756 (0.248–0.997) 0.021 S100A9 0.673 (0.362–0.904) 0.009 PSME2 0.735 (0.621–0.971) 0.004 TF 0.607 (0.2020–1.019) 0.017 TGM2 0.705 (0.479–0.914) 0.029 ALB 0.563 (0.191–0.932) 0.017 Upregulated protein with a ratio of >1.300; downregulated protein with a ratio of <0.769. PIGR, polymeric immunoglobulin receptor; IGLV3- 19, immunoglobulin lambda variable 3-19; ZG16B, zymogen granule protein 16 homolog B; PSME2, proteasome activator complex subunit 2; TGM2, transglutaminase-2.

Downloaded from tvst.arvojournals.org on 09/26/2021 Tear Protein Profile in LASIK IOVS j June 2014 j Vol. 55 j No. 6 j 3400

TABLE 3. Upregulated and Downregulated Proteins in Intralase-Operated Eyes at 1 Week and 3 Months Postoperatively.

1wk 3mo

Gene Symbol Median (Q1–Q3) P Value Gene Symbol Median (Q1–Q3) P Value

Upregulated proteins IGHM 1.983 (0.938–3.192) 0.014 ORM1 1.477 (1.072–3.622) 0.048 Downregulated proteins LGALS3BP 0.692 (0.308–0.979) 0.018 HP 0.67 (0.246–0.947) 0.04 CTSB 0.643 (0.424–0.931) 0.016 SERPINB1 0.577 (0.338–1.103) 0.024 LPO 0.637 (0.231–1.078) 0.018 SCGB1D1 0.536 (0.238–1.132) 0.034 TGM2 0.617 (0.423–0.997) 0.001 ZG16B 0.452 (0.256–1.202) 0.03 LYZ 0.435 (0.131–1.159) 0.05 Upregulated protein with a ratio of >1.300; downregulated protein with a ratio of <0.769. LGALS3BP, galectin-3 binding protein precursor; CTSB, cathepsin B precursor; HP, haptoglobin precursor; SERPINB1, leukocyte elastase inhibitor.

transport protein that modulates inflammatory responses, with long-term glaucoma medication,21 and pterygium.30 Since highest levels found during acute inflammation.27 It is secreted these proteins may also play a protective role31,32 and because by hepatocytes, other epithelial cells, endothelial cells, LASIK induces reactive oxygen radical formation,33 a reduction macrophages, and polymorphonuclear leukocytes. The degree of S100A8 and S100A9 in Visumax-operated eyes may signify of ORM1 fucosylation has been identified as a possible immune dysfunction and impaired ocular surface protection. prognostic measure.27 In the present study, the upregulation Lacritin, a lacrimal functional unit specific protein, has been of ORM1 probably indicates an increased early inflammatory previously found to be downregulated in dry eye.34 It is mainly response to the Intralase procedure. produced in the lacrimal gland, moderately in the Following 3 months of surgery, the plasma-derived proteins and the meibomian gland, and only weakly in corneal serum serotransferrin (SF) albumin (ALB) were downregulated . In the current study, the increase in LACRT in tears of Visumax-operated eyes. Serum serotransferrin is an postoperatively may be a compensatory measure. Since it has iron-binding transport protein that is secreted by the liver and a role in the maintenance of the ocular surface,35 the response epithelia, including lacrimal gland acini. Reduced levels of SF in the Visumax patients may be more beneficial than that in the 28 were detected in mild to moderate dry eye patients. The Intralase patients. Some proteins downregulated in dry eye lacrimal gland secreted protein lysozyme (LYZ) was also have not been found to be affected post refractive surgery in downregulated at 1 week postoperatively. This may suggest a our study: tear lipocalin-1 (LIPOC-1) and zinc-alpha-2-glycopro- greater disruption of the corneal-lacrimal gland unit as a result tein (ZAG-2).36 of a slightly larger flap diameter of the Intralase platform Are there differences in the Visumax and Intralase consequently damaging a larger number of corneal nerves. femtosecond lasers that may translate into distinctive effects Similarly, the protein SCGB1D1 (lipophilin A) was lowered at 3 to the ocular surface and tear protein composition? The months postoperatively, perhaps reflecting an alteration of Visumax laser uses a higher repetition rate and lower laser bed lacrimal . Albumin is considered an indirect sign of energy, while the Intralase laser employs a lower repetition subclinical inflammation and is usually detected at low levels in 28 rate and higher laser bed energy. Another major difference normal tears. In dry eye, however, serum albumin is between the two platforms lays in the suction system, which significantly increased.28 The calcium-binding proteins positions the eyeball and facilitates creation of a flap. Visumax S100A8 and S100A9 (also called calgranulins), are often found generates suction on the cornea via a curved contact glass, in as heterodimers and have mainly been considered to be pro- contrast to Intralase, which applies suction on the conjunctiva inflammatory proteins.29 Previous studies have reported that and sclera via a suction ring. If suction is placed directly onto S100A8 and S100A9 were detected with increased levels in the the conjunctival epithelium and goblet cells, there may be tears of patients with inflammatory ocular surface conditions, such as dry eye,7 meibomian gland dysfunction,11 the use of more damage to these tear producing elements and to the

FIGURE 2. Upregulation of proteins in tears collected from Visumax- operated eyes versus Intralase-operated eyes at 1 week (P < 0.05). FIGURE 3. Upregulation of proteins in tears collected from Visumax- Ratio: Ratio of tear lacritin (LACRT) at 1 week relative to baseline. operated eyes versus Intralase-operated eyes at 3 months (P < 0.05).

Downloaded from tvst.arvojournals.org on 09/26/2021 Tear Protein Profile in LASIK IOVS j June 2014 j Vol. 55 j No. 6 j 3401

Supported by grants from the National Medical Research Council, National Research Foundation’s Translational Clinical Research Program for Translational Research Innovations in Ocular Surgery, SingHealth Foundation, and National Medical Research Council Centre Grant 2010, R738, NMRC\CSA\045\2012. Disclosure: S. D’Souza, None; A. Petznick, None; L. Tong, None; R.C. Hall, None; M. Rosman, None; C. Chan, None; S.K. Koh, None; R.W. Beuerman, None; L. Zhou, None; J.S. Mehta, None

References

1. Yuen LH, Chan WK, Koh J, Mehta JS, Tan DTA. 10-year prospective audit of LASIK outcomes for myopia in 37,932 eyes at a single institution in Asia. Ophthalmology. 2010;117: 1236–1244, e1231. Figure 4. Downregulated proteins in tears collected from Intralase- 2. Solomon KD, Fernandez de Castro LE, Sandoval HP, et al. operated eyes versus Visumax-operated eyes at 3 months postopera- LASIK world literature review: quality of life and patient tively (P < 0.05). satisfaction. Ophthalmology. 2009;116:691–701. 3. Lee BH, McLaren JW, Erie JC, Hodge DO, Bourne WM. conjunctiva nerves.37–39 and these may result in tear film Reinnervation in the cornea after LASIK. Invest Ophthalmol . 2002;43:3660–3664. instability aqueous-deficiency and increased tear hyperosmo- Vis Sci larity.4,40 4. Dooley I, D’Arcy F, O’Keefe M. Comparison of dry- Femtosecond laser flaps sever the sensory corneal nerves, severity after laser in situ keratomileusis and laser-assisted which form the afferent branch of a reflex arc at the flap subepithelial keratectomy. J Cataract Refract Surg. 2012;38: margin.41 An anticipated effect of this corneal nerve truncation 1058–1064. would be a reduced expression of lacrimal gland proteins into 5. Sambursky R, O’Brien TP. MMP-9 and the perioperative tears.42–44 A downregulation of lacrimal gland proteins has management of LASIK surgery. Curr Opin Ophthalmol. been noted in studies in dry eye with lower levels of LYZ and 2011;22:294–303. TF in patients with more severe dry eye.6,7,45 It is possible that 6. Srinivasan S, Thangavelu M, Zhang L, Green KB, Nichols KK. Intralase has a greater impact on the neural arc reflex than iTRAQ quantitative proteomics in the analysis of tears in dry Visumax,6,42 since Intralase flaps have a 0.5-mm larger eye patients. Invest Ophthalmol Vis Sci. 2012;53:5052–5059. diameter leading to greater nerve damage. 7. Zhou L, Beuerman RW, Chan CM, et al. Identification of tear Femtosecond lasers also induce inflammation46 via necrotic fluid biomarkers in using iTRAQ quantita- cell death46,47 and infiltration of inflammatory cells into the tive proteomics. J Proteome Res. 2009;8:4889–4905. corneal stroma.46,47 The degree of necrosis is directly related to 8. Wilson SE. Laser in situ keratomileusis-induced (presumed) energy levels needed for flap creation. Intralase lasers with neurotrophic epitheliopathy. Ophthalmology. 2001;108:1082– higher pulse energies stimulated a far greater level of 1087. keratocyte cell death, inflammation, and healing responses at 9. Zhou L, Beuerman RW. Tear analysis in ocular surface diseases. the flap margins than laser settings with lower pulse Prog Retin Eye Res. 2012;31:527–550. 47 energies. This may explain the inflammatory proteins in 10. Grus FH, Podust VN, Bruns K, et al. SELDI-TOF-MS ProteinChip Intralase-operated eyes. array profiling of tears from patients with dry eye. Invest The present study established relative quantitative ratios but Ophthalmol Vis Sci. 2005;46:863–876. not actual concentrations. The methodological approach 11. Tong L, Zhou L, Beuerman RW, Zhao SZ, Li XR. Association of excludes important proteins such as cytokines involved in tear proteins with Meibomian gland disease and dry eye the inflammatory/healing cascade that are expressed at lower symptoms. Br J Ophthalmol. 2011;95:848–852. levels. Furthermore, a global control that would have allowed 12. Zhou L, Wei R, Zhao P, Koh SK, Beuerman RW, Ding C. us to perform direct comparisons with other eye conditions, Proteomic analysis revealed the altered tear protein profile in a such as idiopathic dry-eye, was not incorporated into the rabbit model of Sjogren’s syndrome-associated dry eye. proteome analysis. Our laboratory is currently developing Proteomics. 2013;13:2469–2481. standards to establish absolute quantitative concentrations to 13. Stern ME, Beuerman RW, Fox RI, Gao J, Mircheff AK, enable comparisons of disease conditions in the future.9 Pflugfelder SC. The pathology of dry eye: the interaction In summary, tear protein comparison between two between the ocular surface and lacrimal glands. Cornea. 1998; platforms revealed that femtosecond laser surgery induce 17:584–589. changes of tear proteins with a profile different from idiopathic dry eye. Higher levels of inflammation may be encountered in 14. Stern ME, Beuerman RW, Fox RI, Gao J, Mircheff AK, Pflugfelder SC. A unified theory of the role of the ocular the Intralase eyes compared with Visumax-treated eyes, as surface in dry eye. Adv Exp Med Biol. 1998;438:643–651. shown by the relatively higher levels of ORM1 in the Intralase eyes. Additionally, lower levels of lacrimal gland proteins in 15. Lee HK, Lee KS, Kim HC, Lee SH, Kim EK. Nerve growth factor Intralase suggested that a recovery of the neural arc reflex with concentration and implications in photorefractive keratecto- innervation of the lacrimal gland may require more time as my vs laser in situ keratomileusis. Am J Ophthalmol. 2005; compared with Visumax-operated eye. 139:965–971. 16. Gonzalez-Perez J, Villa-Collar C, Gonzalez-Meijome JM, Porta NG, Parafita MA. Long-term changes in corneal structure and Acknowledgments tear inflammatory mediators after orthokeratology and LASIK. The authors thank Tian Dechao for organizing the data, Hla Myint Invest Ophthalmol Vis Sci. 2012;53:5301–5311. Htoon, MD, PhD, for statistical advice, and Sharon Yeo for editorial 17. Rosman M, Hall RC, Chan C, et al. Comparison of efficacy and and proofreading work. safety of laser in situ keratomileusis using 2 femtosecond laser

Downloaded from tvst.arvojournals.org on 09/26/2021 Tear Protein Profile in LASIK IOVS j June 2014 j Vol. 55 j No. 6 j 3402

platforms in contralateral eyes. J Cataract Refract Surg. 2013; 32. Sun Y, Lu Y, Engeland CG, Gordon SC, Sroussi HY. The anti- 39:1066–1073. oxidative, anti-inflammatory, and protective effect of S100A8 18. Angunawela RI, Riau A, Chaurasia SS, Tan DT, Mehta JS. in endotoxemic mice. Mol Immunol. 2013;53:443–449. Manual suction versus femtosecond laser trephination for 33. Bilgihan K, Bilgihan A, Adiguzel U, et al. Keratocyte apoptosis penetrating keratoplasty: intraocular pressure, endothelial cell and corneal antioxidant enzyme activities after refractive damage, incision geometry, and wound healing responses. corneal surgery. Eye. 2002;16:63–68. Invest Ophthalmol Vis Sci. 2012;53:2571–2579. 34. McKown RL, Wang N, Raab RW, et al. Lacritin and other new 19. Riau AK, Angunawela RI, Chaurasia SS, Tan DT, Mehta JS. proteins of the lacrimal functional unit. Exp Eye Res. 2009;88: Effect of different femtosecond laser-firing patterns on 848–858. collagen disruption during refractive lenticule extraction. J 35. Nakajima T, Walkup RD, Tochigi A, Shearer TR, Azuma M. Cataract Refract Surg. 2012;38:1467–1475. Establishment of an appropriate animal model for lacritin 20. Petznick A, Chew A, Hall RC, et al. Comparison of corneal studies: cloning and characterization of lacritin in monkey sensitivity, tear function and corneal staining following laser in eyes. Exp Eye Res. 2007;85:651–658. situ keratomileusis with two femtosecond laser platforms. Clin 36. Versura P, Bavelloni A, Blalock W, Fresina M, Campos EC. A Ophthalmol. 2013;7:591–598. rapid standardized quantitative microfluidic system approach 21. Wong TT, Zhou L, Li J, et al. Proteomic profiling of for evaluating human tear proteins. Mol Vis. 2012;18:2526– inflammatory signaling molecules in the tears of patients on 2537. chronic glaucoma medication. . Invest Ophthalmol Vis Sci 37. Jumblatt MM, McKenzie RW, Jumblatt JE. MUC5AC mucin is a 2011;52:7385–7391. component of the human precorneal tear film. Invest 22. Fabian TK, Hermann P, Beck A, Fejerdy P, Fabian G. Salivary Ophthalmol Vis Sci. 1999;40:43–49. defense proteins: their network and role in innate and 38. Diebold Y, Rios JD, Hodges RR, Rawe I, Dartt DA. Presence of acquired oral immunity. Int J Mol Sci. 2012;13:4295–4320. nerves and their receptors in mouse and human conjunctival 23. Choi HM, Lee YA, Yang HI, Yoo MC, Kim KS. Increased levels goblet cells. Invest Ophthalmol Vis Sci. 2001;42:2270–2282. of thymosin beta4 in synovial fluid of patients with rheumatoid arthritis: association of thymosin beta4 with other 39. Rodriguez AE, Rodriguez-Prats JL, Hamdi IM, Galal A, Awadalla factors that are involved in inflammation and bone erosion in M, Alio JL. Comparison of goblet cell density after femtosec- joints. Int J Rheum Dis. 2011;14:320–324. ond laser and mechanical microkeratome in LASIK. Invest Ophthalmol Vis Sci. 2007;48:2570–2575. 24. Soria J, Duran JA, Etxebarria J, et al. Tear proteome and protein network analyses reveal a novel pentamarker panel for tear 40. Lee JB, Ryu CH, Kim J, Kim EK, Kim HB. Comparison of tear film characterization in dry eye and meibomian gland secretion and tear film instability after photorefractive dysfunction. J Proteomics. 2013;78:94–112. keratectomy and laser in situ keratomileusis. J Cataract Refract Surg. 2000;26:1326–1331. 25. Ananthi S, Chitra T, Bini R, Prajna NV, Lalitha P, Dharmalingam K. Comparative analysis of the tear protein profile in mycotic 41. Dartt DA. Dysfunctional neural regulation of lacrimal gland keratitis patients. Mol Vis. 2008;14:500–507. secretion and its role in the pathogenesis of dry eye syndromes. Ocul Surf. 2004;2:76–91. 26. Acera A, Vecino E, Rodriguez-Agirretxe I, et al. Changes in tear protein profile in keratoconus disease. Eye (Lond). 2011;25: 42. Dartt DA. Neural regulation of lacrimal gland secretory 1225–1233. processes: relevance in dry eye diseases. Prog Retin Eye Res. 2009;28:155–177. 27. Ceciliani F, Pocacqua V. The acute phase protein alpha1-acid glycoprotein: a model for altered during 43. Salvatore MF, Pedroza L, Beuerman RW. Denervation of rabbit diseases. Curr Protein Peptide Sci. 2007;8:91–108. lacrimal gland increases levels of transferrin and unidentified 28. Versura P, Bavelloni A, Grillini M, Fresina M, Campos EC. tear proteins of 44 and 36 kDa. Curr Eye Res. 1999;18:455– Diagnostic performance of a tear protein panel in early dry 466. eye. Mol Vis. 2013;19:1247–1257. 44. Glasgow BJ, Gasymov OK. Focus on molecules: tear lipocalin. 29. Foell D, Wittkowski H, Vogl T, Roth J. S100 proteins expressed Exp Eye Res. 2011;92:242–243. in phagocytes: a novel group of damage-associated molecular 45. Versura P, Nanni P, Bavelloni A, et al. Tear proteomics in pattern molecules. J Leukoc Biol. 2007;81:28–37. evaporative dry eye disease. Eye (Lond). 2010;24:1396–1402. 30. Zhou L, Beuerman RW, Ang LP, et al. Elevation of human alpha- 46. Netto MV, Mohan RR, Medeiros FW, et al. Femtosecond laser defensins and S100 calcium-binding proteins A8 and A9 in tear and microkeratome corneal flaps: comparison of stromal fluid of patients with pterygium. Invest Ophthalmol Vis Sci. wound healing and inflammation. J Refract Surg. 2007;23: 2009;50:2077–2086. 667–676. 31. Gomes LH, Raftery MJ, Yan WX, Goyette JD, Thomas PS, Geczy 47. de Medeiros FW, Kaur H, Agrawal V, et al. Effect of CL. S100A8 and S100A9-oxidant scavengers in inflammation. femtosecond laser energy level on corneal stromal cell death Free Radic Biol Med. 2013;58:170–186. and inflammation. J Refract Surg. 2009;25:869–874.

Downloaded from tvst.arvojournals.org on 09/26/2021