Keratan Sulfate and Chondroitin/Dermatan Sulfate in Maximally Recovered Hypocellular Stromal Interface Scars of Postmortem Human LASIK Corneas

Abstract

purpose. To analyze the amounts and distributions of nonsulfated and sulfated keratan sulfate (KS) and chondroitin/dermatan sulfate (CS/DS) disaccharides in the interface wound of human postmortem LASIK corneas in comparison with normal control corneas.

methods. Corneal stromal tissue samples from central and paracentral hypocellular primitive stromal interface scars of human LASIK corneas and from similar regions of normal control corneas were collected by laser capture microdissection (LCM) and subsequently were digested with specific glycosidase enzymes. Digests were directly analyzed by electrospray ionization tandem mass spectrometry (ESI-MS/MS).

results. Concentrations of both monosulfated GlcNAc(6S)-β-1,3-Gal (MSD2) and disulfated Gal (6S)-β-1,4-GlcNAc(6S) (DSD) KS disaccharides from the LASIK interface scars were significantly lower than in normal control corneal stromas. No significant difference was found for the concentration of nonsulfated (NSD) KS disaccharides in LASIK interface scars compared with normal controls. The concentration of ΔUA-β-1,3-GalNAc(6S) (Δdi-6S) CS/DS disaccharides from the LASIK interface scar was significantly higher than normal corneal stroma, whereas concentrations of ΔUA-β-1,3-GalNAc(4S) (Δdi-4S) and nonsulfated Δdi-0S CS/DS disaccharides demonstrated no significant differences from normal corneas.

conclusions. The profiles of KS and CS/DS disaccharides in LASIK interface scars are significantly different from those in normal cornea stromal tissue, as revealed by LCM and ESI-MS/MS.

The human cornea is avascular. Therefore, wound healing in the human corneal stroma is much slower and less complete in repairing normal stromal tissue when compared with most other tissues in the body. ¹ In human corneas, wound healing after LASIK surgery does not stabilize until 3.5 years after surgery. ² Whereas animal corneas heal completely after microkeratome cuts and LASIK procedures, ^{3 4} human corneas undergo only primitive and limited wound healing. Moreover, LASIK wound healing has been associated with numerous postoperative complications (e.g., trauma flap wound dehiscence, diffuse lamellar keratitis, interface fluid syndrome, and infections). ⁵ Recent human LASIK studies have demonstrated that corneal stromal wounds heal in one of two reparative fashions: hypercellular fibrotic stromal scarring or hypocellular primitive stromal scarring. ^{6 7 8} Corneal stromal wound healing at the LASIK flap margin is influenced by epithelial–stromal interactions and results in the formation of a hypercellular fibrotic stromal scar. The hypercellular fibrotic scar is functionally abnormal, in that it is more hazy than normal corneal stroma because of myofibroblast formation and, to a lesser extent, abnormal extracellular (ECM) matrix deposition. Yet, it is relatively strong, attaining 30% to 40% of normal stromal tensile strength. ^{2 9} In contrast, the central and paracentral regions of the LASIK interface wound heal by producing a transparent hypocellular primitive stromal scar with a reduced tensile strength, attaining only 2% to 3% of normal stromal tensile strength. ^{2 9} Ultrastructural and immunofluorescent analyses of the hypocellular primitive stromal scar have shown that it is mainly composed of electron-dense granular material that contains abnormally large (300 × 20 nm), noncollagen-fibril–bound proteoglycans (PGs), as well as keratocytes and very few interspersed collagen fibrils that are smaller than normal in diameter. ^{6 7 8}  

The principal PGs found in the human corneal stroma are those of keratan sulfate (KSPG) and chondroitin/dermatan sulfate (CS/DSPG). ¹⁰ In the cornea, these PGs consist of small leucine-rich core proteins (SLRPs) covalently bearing one or more long sugar side chains, glycosaminoglycans (GAGs), that project from the convex surfaces of the core proteins. ^{11 12} KS GAG chains are composed of repeating disaccharides of residues of d-galactose (Gal) and N-acetyl-d-glucosamine (GlcNAc)–linked β-1,4 and β-1,3, respectively, ^{13 14 15 16} whereas CS/DS GAG chains are composed of repeating disaccharides of N-acetyl-d-galactosamine (GalNAc) residues alternating in glycosidic linkages with glucuronic acid (in CS domains) or iduronic acid (in DS domains) residues. ^{17 18} Specific sulfotransferases attach sulfate groups to some sugars as they are incorporated into GAG side chains during biosynthesis. ¹⁹ Many factors influence sulfation, leading to the appearance of both nonsulfated and variously sulfated domains along individual chains. In KS, sulfation of the hydroxyl groups at the C-6 positions of either Gal or GlcNAc residues, or in some cases of both Gal and GlcNAc residues, can be demonstrated. In CS/DS, GalNAc residues are predominantly sulfated in the C-4- (chondroitin A) or C-6-hydroxyl (chondroitin C) positions, interspersed with a few nonsulfated residues. ¹⁶ The SLRP core proteins are thought to wrap around collagen fibrils and constrain growth of their diameters, whereas the projecting heterogeneously sulfated GAG side chains repel one another, resulting in even spacing between collagen fibrils. ¹¹ It has been established that GAGs are responsible for hydration dynamics in the cornea. ^{20 21}  

Because the hypocellular primitive LASIK scar has been found to display increased fluid retention and abnormally large PGs, it was determined that more detailed information about the proteoglycan composition of this scar was needed. Therefore, tissue samples from the central and paracentral hypocellular primitive stromal LASIK scar were precisely excised from human LASIK eye bank donors using laser capture microdissection (LCM), a recent technique developed to dissect homogenous tissue samples from heterogeneous tissue under microscopic observation. Subsequently, enzymatic digests of the extracted scar tissue were analyzed by electrospray ionization tandem mass spectrometry (ESI-MS/MS), to identify and quantify the amount and distribution of KS and CS/DS nonsulfated and sulfated disaccharides in the hypocellular primitive stromal LASIK scar compared with control corneas. 

Materials and Methods

Tissue

All human cornea samples were collected in accordance with the guidelines in the Declaration of Helsinki for research involving human tissue. After approval by the Emory University Institutional Review Board (IRB), five corneoscleral specimens from four donors (mean age ± SD, 46.3 ± 2.9 years) with a prior history of LASIK (range, 4.0–6.5 years after surgery) and five control normal corneoscleral specimens from four donors (mean age ± SD, 53.0 ± 19.0 years) were obtained from various eye banks in North America. Thus, all LASIK corneas used in this study had undergone maximum wound healing at the time of their collection. ² All specimens were stored in preservative (Optisol-GS; Bausch & Lomb Surgical, Irvine, CA) at 4°C within 17 hours of death and were evaluated in our laboratory within 5 days of death. Review was performed of the donor’s ophthalmic history, when available. LASIK and control corneoscleral specimens were randomly paired to form five groups of sample pairs, with one LASIK and one normal control cornea specimen per pair. Detailed information about characteristics of the donors and corneal specimens in each group is provided in Table 1 . 

All 10 corneoscleral specimens (LASIK [n = 5]; controls [n = 5]) were evaluated for gross abnormalities with a dissecting microscope (SZ-40; Olympus Tokyo, Japan) using direct illumination and indirect retroillumination techniques, and subsequently were bisected (Fig. 1) . The bisected specimens were placed in 70% ethanol for 24-hour fixation, dehydrated through a graded series of alcohol solutions (80%, 95%, and 100%), cleared with xylene solution, infiltrated with paraffin, and embedded in paraffin blocks at 60°C for 2 hours. Six-micrometer-thick sections were cut from the blocks and placed on special PEN-membrane–coated histology glass slides (Arcturus Bioscience, Mountain View, CA). The unstained sections were dewaxed with a 120-second wash of xylene, air-dried in a laminar flow hood, and transferred for LCM. 

Laser Capture Microdissection Technique

The protocol for laser capture microdissection (LCM) has been well described previously. ²² Our procedure, which was optimized for ESI-MS/MS analysis, was performed on a commercial LCM system (PixCell II; Arcturus Engineering). The protocol entailed microdissecting and capturing the hypocellular primitive stromal LASIK scar (central and paracentral stromal scars) from each LASIK corneal section on LCM caps (CapSure HS LCM caps; Arcturus) by using 200 contiguous laser spots (Figs. 2A 2B) . Because of the Health Insurance Portability and Accountability Act (HIPAA) guidelines, it was difficult to obtain refractive error and ablation depth from the donors’ ophthalmic records. However, the measured thickness of the LASIK flap in all the patients averaged 140 ± 12.6 μm (range, 120–160 μm). Corneal tissue was microdissected by LCM at the same depth in the control corneas (Fig. 2C) . The infrared laser settings were as follows: 55-mW power, 900-ms duration, and 7.5-μm diameter laser spot size. Replicate sets of each sample were taken, so that totals of 10 LASIK and 6 to 10 normal control sets were collected from sample pairs (five LASIK and three to five normal LCM caps per glycosaminoglycan analysis group). 

Reagents

KS disaccharides were obtained from Seikagaku Corp. (Tokyo, Japan) and V-LABS, Inc. (Covington, LA). CS/DS disaccharides were obtained from Sigma-Aldrich (St. Louis, MO), EMD Biosciences, Inc. (San Diego, CA), and V-LABS, Inc. ΔUA-β-1,4-GlcNS, also called IV-S and used as the internal standard for quantification of CS/DS sulfated disaccharides, was purchased from EMD Biosciences, Inc. Gal-α-1,3-Gal, used as the internal standard for quantification of KS and CS/DS nonsulfated disaccharides, was purchased from V-LABS, Inc. Chondroitinase ABC (from Proteus vulgaris, protease-free, EC 4.2.2.4), chondroitinase AC (from Flavobacterium heparinum, protease-free, EC 4.2.2.5), ammonium acetate, and ammonium sulfate were obtained from Sigma-Aldrich. Keratanase II (from Bacillus sp.) and endo-β-galactosidase (from Escherichia freundii) were purchased from Seikagaku America (East Falmouth, MA). High-pressure liquid chromatography (HPLC)–grade solvents and ultrafree-MC centrifugal filter units (5000 NMWL; Millipore, Bedford, MA) were purchased from Fisher (Santa Clara, CA). 

Preparation of Internal Standard Solutions

All disaccharide standards were diluted in a series of 10 nmol/μL, 1 nmol/μL, 100 pmol/μL, and 10 pmol/μL of sterile water, and aliquots were stored at −20°C. For quantitative analysis, 10 μL of the 1-nmol/μL solution of each standard was diluted by adding 70 μL MeOH, 5 μL of 5 mM ammonium acetate buffer (pH 7.5), 5 μL of 2 mM ammonium sulfate, and 10 μL water, which made the solution 7:3 MeOH:H₂O, containing 100 pmol/μL disaccharide. 

Enzymatic Digestion of KS and CS/DS

The microdissected and captured tissue samples (five sets of LASIK and three to five sets of normal control human corneas) of each group were enzymatically digested and processed at the same time. Single LASIK or normal control samples of each pair were exposed to 10 μL digestion solution, wrapped with film, and incubated in a moist chamber at 37°C for 24 hours. For keratanase II digestion (releases Gal-β-1,4-GlcNAc(6S) [MSD1] and Gal (6S)-β-1,4-GlcNAc(6S) [DSD] KS disaccharides), the digest solution was composed of 0.1 M ammonium acetate buffer (pH 6.0), 0.1 mU/μL keratanase II, and 50 pmol/μL ΔUA-β-1,3-GalNAc(2S) [Δdi-2S] (internal standard). For endo-β-galactosidase digestion (releases GlcNAc(6S)-β-1,3-Gal [MSD2] and GlcNAc-β-1,3-Gal [NSD] KS disaccharides), the digest solution was composed of 0.05 M ammonium acetate buffer (pH 5.8), 0.1 mU/μL endo-β-galactosidase, and 50 pmol/μL Gal-α-1,3-Gal (internal standard). For analysis of CS/DS disaccharides, the digest solution was composed of 50 mM ammonium acetate buffer (pH 7.5), 1 mU/μL chondroitinase ABC or 1 mU/μL chondroitinase ABC plus 1 mU/μL chondroitinase AC (releases ΔUA-β-1,3-GalNAc [Δdi-0S], Δdi-4S, and Δdi-6S disaccharides), and 50 pmol/μL ΔUA-β-1,4-GlcNS or Gal-α-1,3-Gal (internal standards). The digest solutions of each set were collected and diluted by adding 70 μL of MeOH, 5 μL of 5 mM ammonium acetate buffer (pH 7.5), 5 μL of 2 mM (NH₄)₂SO₄, and 10 μL of water. The mixtures were centrifuged at 3800 rcf for 30 minutes at 4°C (Ultra free-MC Centrifugal Filter unit, 5000 NMWL; Millipore) with subsequent analysis of the filtrate by ESI-MS/MS. 

ESI-MS/MS

Mass spectra were obtained with an electron spray ionization source on a quadrupole ion trap instrument (Esquire 3000; Bruker Daltonics, Billerica, MA). Parameters used for analysis of all internal standards, and corneal stromal tissue digests were as follows: spray voltage, 3.5 kV; dry gas, nitrogen, flow: 5.0 L/min; and drying temperature, 180°C. The mass range scanned was m/z 50 to 1000, and data were aquired (Data Analysis 3.0 acquisition software; Bruker Daltonics). 

Quantitative Analysis of KS Disaccharides

Representative mass spectrometry traces of keratanase II digests of one control sample and one LASIK wound tissue sample, plus the internal standard, Δdi-2S, are shown in Figures 3A and 3B . The intensities of the MSD1 and DSD peaks were compared with the intensity of the internal standard peak, and a single-point normalization factor method was used to determine the amounts of MSD1 and DSD KS disaccharides as described previously. ²³ The same methods were used to determine the amounts of NSD and MSD2 from endo-β-galactosidase digests of control and LASIK wound tissue. ²³  

Quantitative Analysis of CS/DS Disaccharides

The molar concentrations of Δdi-0S, Δdi-4S, and Δdi-6S CS/DS disaccharides were determined using single-point normalization factors. A system of two equations that corrects for the presence of additional ions in the MS ² spectrum allows the relative molar percentages of Δdi-4S and Δdi-6S isomers in a mixture to be calculated from the observed relative intensities of the diagnostic ions for each isomer in the MS ² spectrum of that mixture. ²⁴  

Results

Content of the KS Disaccharides

KS disaccharide sulfation profiles are summarized in Table 2 . Concentrations of the sulfated KS disaccharides MSD2 and DSD were significantly lower in LASIK interface scars than in normal human corneal stromal tissue (P < 0.01). Concentrations of MSD2 for LASIK interface scars averaged 2.25 ± 0.18 pmol/μm³× 10⁻³ (range, 2.09–2.35 pmol/μm³× 10⁻³) compared with normal control corneal samples, which averaged 2.96 ± 0.36 pmol/μm³× 10⁻³ (range, 2.7–3.21 pmol/μm³× 10⁻³). Concentrations of DSD for LASIK interface scars averaged 2.73 ± 0.69 pmol/μm³× 10⁻³ (range, 2.2–3.6 pmol/μm³× 10⁻³) compared with normal corneal stromal samples, which averaged 3.95 ± 0.79 pmol/μm³× 10⁻³ (range, 3.4–5.3 pmol/μm³× 10⁻³). In contrast, the concentrations of KS NSD and MSD1 disaccharides in the LASIK scar were not significantly different from those in normal stroma (P > 0.05). 

Content of CS/DS Disaccharides

CS/DS disaccharide sulfation profiles are summarized in Table 3 . The tissue concentration of the sulfated CS/DS disaccharide Δdi-6S was significantly increased in the LASIK stromal scar (mean, 0.23 ± 0.04 pmol/μm³× 10⁻³) compared with normal control corneal samples (mean, 0.07 ± 0.02 pmol/μm³× 10⁻³; P < 0.01). In contrast, the concentrations of Δdi-4S and of Δdi-0S in LASIK interface scars were not significantly different from corresponding areas of stromal tissue of the normal control corneas (P < 0.05). This resulted in an average 3.65-fold increase in the molar ratio of Δdi-6S to Δdi-4S CS/DS in LASIK scars compared with controls. Moreover, in the normal cornea, Δdi-6S represented 18.39% molar percentage of the total sulfated monosulfated CS/DS disaccharides, Δdi-6S + Δdi-4S, whereas in the LASIK stromal scars Δdi-6S represented 44.36%. 

Discussion

In this study, LCM and ESI-MS/MS techniques were used to quantify directly the KS and CS/DS disaccharides in human corneal tissue along the central and paracentral hypocellular primitive LASIK stromal interface scar and from similar areas of nonwounded normal control corneas. The recently developed LCM technique allows microdissection and capture of very small circumscribed regions of a tissue and thus permits previously unobtainable biochemical analyses of these very small regions. In this study, LCM coupled with ESI-MS/MS made it possible to determine unambiguously the KS and CS/DS disaccharide composition of the extremely thin LASIK interface scar. Without LCM, the KS and CS/DS disaccharide composition of the much larger stroma mass surrounding the scar would have overwhelmed resolution of the unique scar GAG composition. 

Concentrations of KS MSD2 and DSD in LASIK interface scars were significantly lower than in normal corneal stroma. Although MSD1 concentration was not significantly lower in the LASIK scar compared with normal control stroma, it was the same as the concentration of MSD2. This finding supports the conclusion that the total concentration of mono- and disulfated KS disaccharides is slightly but significantly lower in the LASIK scar than in the normal corneal stroma. In contrast, the concentration of KS NSD in the LASIK scar was not significantly different from that in the normal corneal stroma. These LCM/ESI-MS/MS quantitative results are consistent with a previous human immunofluorescence study that used a primary monoclonal antibody directed against sulfated KS epitopes, which showed that the degree of KS fluorescence in the central and paracentral hypocellular primitive interface scar was slightly less than in normal corneal stroma. ⁷ It is also consistent with the animal studies that show both a reduced amount of KS and transiently undersulfated KS in opaque corneal scars. ^{25 26 27} These data suggest that sulfation of both Gal and GlcNAc in KS is reduced in the hypocellular primitive LASIK scar, presumably from reduced and altered KS biosynthesis. The data presented herein or in previous studies do not indicate, however, whether the LASIK corneal interface scar contains altered concentrations of any of the KS proteoglycan core proteins (e.g., keratocan, lumican, mimecan/osteoglycin). ^{7 25 26 27 28} Their concentrations may be normal, despite the changes in the posttranslational modifications of attached KS side chains. Alternatively, the reduced concentration of sulfated KS in LASIK tissue may be caused by reduced synthesis of KS core proteins, thus providing fewer glycosylation sites on which KS chains could be assembled. 

An increase in the CS/DS content was also found in the central and paracentral hypocellular primitive LASIK interface scar, predominantly due to an increase in the concentration of Δdi-6S, which was only minimally present in normal corneal stroma. In contrast, no significant changes in the LASIK scar concentrations of nonsulfated disaccharides and 4S-sulfated disaccharides were found compared with those in the normal corneal stroma. The disaccharide ratios found in the LASIK scar may reflect changes in specific activities of chondroitin sulfotransferases, specifically chondroitin 6-O-sulfotransferase. ^{16 17} CS/DS that is enriched with Δdi-6S may be associated with the electron-dense granular material repeatedly detected in the hypocellular primitive LASIK stromal interface scar. ^{6 7 8} Animal studies have shown that other types of corneal scars contain abnormally large chondroitin and dermatan sulfate PGs. ^{25 26 27 28 29 30 31} Major alterations in the biochemical structure of PGs were seen in dermatan sulfate, which had a twofold increase in iduronic acid content and was in a more highly sulfated form than normal. ²⁷ A previous human immunofluorescence evaluation of the central and paracentral hypocellular primitive LASIK interface scar detected no DS proteoglycan, which is normally the major CS/DS proteoglycan of the corneal stoma. ⁷ That study, however, used a primary monoclonal antibody directed against the CS/DS proteoglycan core protein, decorin, rather than specifically against the DS glycosaminoglycan side chains. Further study is needed to determine why decorin immunoreactivity could not be demonstrated in the hypocellular primitive LASIK scar, and whether expressions of any new CS/DS proteoglycan core proteins (e.g., serglycin, biglycan, and versican) are activated in resident keratocytes or invading inflammatory cells in LASIK scars. The current data highlight the striking increase in the concentration of Δdi-6S in the hypocellular primitive LASIK stromal interface scar, but do not resolve whether those specific disaccharides were originally derived from hybrid chains of chondroitin sulfate C or from dermatan sulfate. These data also do not identify to which core protein these CS/DS chains enriched in Δdi-6S are covalently bound. Enhanced presence of chondroitin sulfates, including Δdi-6S, has also been detected in glial scars and correlated with inhibition of neurite extension in these scars. ³² Perhaps, a similar phenomenon occurs in LASIK corneas, where recovery of preoperative subbasal nerve density in the corneal LASIK flap is delayed by up to 5 years or more. ³³  

The causes of the alterations in PG and GAG distribution observed in hypocellular central and paracentral LASIK primitive scars are likely to be complex. The LASIK procedure includes two main steps: creation of a thin LASIK flap composed of the overlying corneal epithelium and anterior stroma with mechanical or laser microkeratome and subsequent laser ablation of the exposed underlying corneal stromal tissue. Therefore, the measured changes in KS and CS/DS are most probably the result of both microkeratome cut and LASER ablation. 

Decorin-, lumican-, and mimecan-null mice all show increased skin collagen fibril diameter heterogeneity and increased skin fragility due to reduced tensile strength, suggesting that loss of any of these SLRP proteoglycan core proteins in the LASIK scar could lead to reduced scar tensile strength. ^{12 34 35} Thus, if the decrease in sulfated KS in the LASIK stromal interface scar tissue were caused by reduced synthesis of any of the core proteins to which KS chains are attached, it may also partially explain the decrease in tensile strength of the LASIK scar. More information about the concentration of PG core proteins in LASIK scars would be helpful, and newer mass spectrometry techniques are currently being developed to obtain these proteomic data. 

Recent studies have shown that the proteoglycans in the interface of the LASIK wound, between the flap and bed, swell, with high IOP and corneal endothelial decompensation. ³⁶ Interface swelling also causes a high degree of reflectivity, as seen by confocal microscopy. Therefore, changes in KS and CS/DS may be responsible for this swelling. KS and CS/DS GAGs have different water-binding properties. ²¹ Under normal hydration of the cornea, CS/DS is fully hydrated and holds its bound water very tightly, whereas KS is only partially hydrated and releases its bound water more readily. ¹⁴ Greatly increased 6-sulfated CS/DS holding water very tightly in the hypocellular scar could explain the fluid accumulation often observed along the interface in human LASIK corneas. ^{37 38 39} Previous animal cell culture work has shown that corneal wound healing also results in the production of hyaluronic acid. ^{27 40} The hyaluronic acid content of LASIK wounds should be assessed in future studies, because hyaluronic acid is very hydrophilic and thus may contribute to the hydrophilic properties of the hypocellular primitive LASIK scar. 

In summary, profiles of the concentrations of KS and CS/DS nonsulfated and sulfated disaccharides released from the hypocellular primitive LASIK scar and from normal control stroma were successfully determined with LCM and ESI-MS/MS. The concentration of sulfated KS disaccharides was reduced in the hypocellular primitive LASIK scar, whereas the concentration of Δdi-6S, a CS/DS-derived disaccharide, was enriched significantly along the hypocellular primitive LASIK scar, indicating that these ECM components may have an important role in corneal wound healing after refractive surgery. 

²  Contributed equally to the work and therefore should be considered equivalent authors.

 Presented in part at the annual meeting of the of the American Academy of Ophthalmology, Chicago, Illinois, 2005.

 Supported by National Eye Institute Grants EY 000933 (HFE), EY 000952 (GWC), P30-EY06360 (HFE), and T32-EY 07092 (HFE); an unrestricted departmental grant from Research to Prevent Blindness (HFE); and the Gertrud Kusen Foundation, Hamburg, Germany (IS).

 Submitted for publication December 7, 2005; revised January 25, 2006; accepted March 20, 2006.

 Disclosure: Y. Zhang, None; I. Schmack, None; D.G. Dawson, None; H.E. Grossniklaus, None; A.H. Conrad, None; Y. Kariya, Seikagaku Corporation (E); K. Suzuki, Seikagaku Corporation (E); H.F. Edelhauser, None; G.W. Conrad, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Demographics of LASIK and Control Corneas

Group	Cornea Type	Location* (OD/OS)	Donor (Age/Gender)	Time Interval (LASIK–Death)	Time interval (Death–Fixation)	Central LASIK Flap Thickness, † (μm)
1	LASIK	OD	45/F	6.5 y	4 d	120
	Control	OD	62/M	—	3 d	—
2	LASIK	OD	49/M	4.0 y	2 d	140
	Control	OD	73/F	—	5 d	—
3	LASIK	OD	49/F	6.5 y	2 d	160
	Control	OS	73/F	—	5 d	—
4	LASIK	OS	49/M	4.0 y	3 d	140
	Control	OD	22/F	—	5 d	—
5	LASIK	OD	42/M	5.0 y	3 d	140
	Control	OS	55/M	—	5 d	—

*  OD = right eye; OS = left eye.

†  Measured from the histological sections of the corneas.

Figure 1.

 

Diagram demonstrating how the LASIK (A) and normal control (B) corneas were bisected. Gray ring: sclera; white circle: cornea; blue-dashed semicircle: LASIK flap wound margin.

Figure 2.

 

(A) The hypocellular primitive LASIK stromal interface scar (blue line) and the area of the interface wound, which was microdissected from the remaining LASIK corneal stroma (black line). P, paracentral scar; C, central scar. Two sections were used per specimen (LASIK and normal control cornea) to collect a sufficient amount of tissue. Overall, the final LCM cap contained tissue captured from 200 contiguous LCM laser spots. Photomicrographs show the areas of the microdissected LASIK scars (B) and normal stromal control tissue (C). (B) The thermal-responsive film is still in place in this LASIK corneal section, which highlights the 100 contiguous LCM laser spots and the underlying silhouette of the extracted tissue. (C) Section of normal control cornea that shows the silhouette of the microdissected tissue and the remaining corneal stromal tissue.

Figure 3.

 

Mass spectra of products released by keratanase II digestion of the normal and LASIK corneas by laser capture microdissection (LCM) in the negative ionization mode. (A) MS spectrum of the normal cornea sample. Molecular ions at m/z 462.0 and m/z 270.4 correspond to Gal-β-1,4-GlcNAc(6S) (MSD1) and Gal(6S)-β-1,4-GlcNAc(6S) (DSD), respectively. The ion at m/z 457.9 corresponds to the KS sulfated disaccharide internal standard ΔDi-2S. (B) MS spectrum of the LASIK cornea sample. Molecular ions at m/z 462.0 and m/z 270.4 correspond to Gal-β-1,4-GlcNAc(6S) (MSD1) and Gal(6S)-β-1,4-GlcNAc(6S) (DSD), respectively. The ion at m/z 457.9 corresponds to the KS sulfated disaccharide internal standard ΔDi-2S.

Table 2.

 

Concentrations of KS Nonsulfated and Sulfated Disaccharides Released by Digestion with Endo-β-Galactosidase and Keratanase-II

Group	Corneal Type	Samples (n)	NSD*	MSD1, †	MSD2*	DSD, †	Total KSD MSD1 + DSD
1	LASIK	5		1.6 ± 0.4a , ‡		2.2 ± 0.2a	3.9 ± 0.33a
	Control	5		2.2 ± 0.2b		3.7 ± 0.2c	5.9 ± 0.22c
2	LASIK	5		1.8 ± 0.3a		2.3 ± 0.3a	4.1 ± 0.26a
	Control	5		2.1 ± 0.7a		3.4 ± 0.3b	5.6 ± 0.80c
3	LASIK	5		3.26 ± 0.29a		3.62 ± 0.22a	6.87 ± 0.41a
	Control	3		4.91 ± 0.19c		5.27 ± 0.34c	10.18 ± 0.29c
4	LASIK	5	0.96 ± 0.19a , ‡		2.35 ± 0.09a
	Control	5	0.89 ± 0.02a		3.21 ± 0.27c
5	LASIK	5	0.90 ± 0.17a		2.09 ± 0.19a
	Control	3	0.81 ± 0.10a		2.70 ± 0.23b
Mean ± SD	LASIK	25	0.93 ± 0.18a	2.21 ± 0.83a	2.25 ± 0.18a	2.73 ± 0.69a	4.94 ± 1.45a
Mean ± SD	Control	21	0.85 ± 0.08a	2.82 ± 1.26a	2.96 ± 0.36c	3.95 ± 0.79c	6.77 ± 2.01b

 Data are expressed as picomoles per cubic micrometer × 10⁻³. NSD, nonsulfated disaccharide, GlcNAc-β-1,3-Gal; MSD1, monosulfated disaccharide, Gal-β-1,4-GlcNAc(6S); MSD2, mono-sulfated disaccharide, GlcNAc(6S)-β-1,3-Gal; DSD, disulfated disaccharide, Gal(6S)-β-1,4-GlcNAc(6S); KSD, keratan sulfate disaccharides.

*  Released by digestion with endo-β-galactosidase

†  Released by digestion with keratanase-II

‡   t-Test, for the same-group column data, a and a: P > 0.05; a and b: P < 0.05; a and c: P < 0.01.

Table 3.

 

Concentrations of CS/DS Nonsulfated and Sulfated Disaccharides Released by Digestion with Chondroitinase ABC and Chondroitinase AC

Group	Cornea Type	Sample (n)	Δdi-0S (pmol/μm3 × 10−4)	Δdi-4S (pmol/μm3 × 10−3)	Δdi-6S (pmol/μm3 × 10−3)	Molar Percentage of Δdi-4S*	Molar Percentage of Δdi-6S*
1	LASIK	5		0.24 ± 0.026a	0.20 ± 0.024a	53.86 ± 2.51a	46.14 ± 2.51a
	Normal	5		0.27 ± 0.026b	0.05 ± 0.004c	84.65 ± 2.35c	15.35 ± 2.35c
2	LASIK	5		0.23 ± 0.04a	0.19 ± 0.02a	55.89 ± 3.24a	44.11 ± 3.24a
	Normal	5		0.26 ± 0.038a	0.05 ± 0.008c	83.03 ± 3.08c	16.97 ± 3.08c
3	LASIK	5	0.86 ± 0.10, †	0.37 ± 0.03a	0.28 ± 0.06a	58.81 ± 2.59a	41.19 ± 2.59a
	Normal	3	0.77 ± 0.04a	0.31 ± 0.01b	0.08 ± 0.02c	78.49 ± 3.54c	21.51 ± 3.54c
4	LASIK	5	0.81 ± 0.12a	0.32 ± 0.03a	0.26 ± 0.03a	55.13 ± 2.48a	44.87 ± 2.48a
	Normal	5	0.68 ± 0.12a	0.36 ± 0.03a	0.09 ± 0.01c	81.24 ± 1.81c	18.76 ± 1.81c
5	LASIK	5	0.76 ± 0.06a	0.28 ± 0.03a	0.24 ± 0.02a	54.50 ± 2.99a	45.50 ± 2.99a
	Normal	3	0.66 ± 0.07a	0.34 ± 0.03b	0.08 ± 0.01c	80.62 ± 2.11c	19.38 ± 2.11c
Mean ± SD	LASIK	25	0.81 ± 0.05a	0.29 ± 0.06a	0.23 ± 0.04a	55.64 ± 1.93a	44.36 ± 1.93a
Mean ± SD	Normal	21	0.70 ± 0.06a	0.31 ± 0.04a	0.07 ± 0.02c	81.61 ± 2.35c	18.39 ± 2.35c

*  [Δdi-4S or Δdi-6S/(Δdi-4S + Δdi-6S)] × 100.

†   t-Test, for the same-group column data, a and a: P > 0.05; a and b: P < 0.05; a and c: P < 0.01.