Intrastromal Keratotomy with Femtosecond Laser Avoids Profibrotic TGF-β1 Induction

Abstract

purpose. To examine expression of the profibrotic cytokine TGF-β1 after selective intrastromal corneal injury with the use of a femtosecond laser.

methods. Rabbits underwent monocular intrastromal keratotomy at a preoperatively determined corneal depth of 160 to 200 μm with the use of a femtosecond laser. Femtosecond laser-induced TGF-β1 expression was compared in nonoperated control eyes and eyes treated with photorefractive keratectomy (PRK). Follow-up examinations were performed 1, 3, 7, and 28 days after surgery. TGF-β1 protein was identified by immunofluorescence labeling. With the use of laser-capture microdissection, epithelial, stromal, and endothelial cell layers were collected, and changes in TGF-β1 mRNA expression were quantified with quantitative RT-PCR.

results. TGF-β1 mRNA and protein expression did not significantly increase after intrastromal femtosecond laser keratotomy. In contrast, TGF-β1 was induced in corneal epithelial and stromal cells after PRK and showed up to 23-fold higher TGF-β1 mRNA levels compared with control corneas. The increase of TGF-β1 mRNA levels after PRK was accompanied by increased TGF-β1 protein production.

conclusions. Isolated stromal injury with a femtosecond laser does not result in induction of the profibrotic cytokine TGF-β1. Because TGF-β1 has been implicated in a fibrotic response of the corneal stroma to injury, absence of TGF-β1 induction argues for a favorable wound-healing response. These findings support highly selective intrastromal procedures in refractive surgery.

In wound healing after corneal surgery, minimal repair and replacement of stromal tissue are desirable to maintain transparency and a favorable refractive outcome. ¹ Recent studies support the idea that loss of corneal transparency can be caused by intense light scattering from repair fibroblasts (e.g., myofibroblasts) in corneal wounds. ^{2 3} Understanding the cellular and molecular interactions responsible for the differentiation of keratocytes to myofibroblasts is, therefore, essential for improving refractive surgical procedures. ⁴  

The differentiation of myofibroblasts from corneal fibroblasts requires TGF-β1. ⁵ Under normal circumstances (uninjured cornea), the epithelial basement membrane binds cytokines and may act as a barrier for TGF-β1 or other signaling molecules from the epithelium. ⁶ After injury, epithelial cells upregulate the production of TGF-β1, which diffuses into the stroma and contributes to paracrine fibroblast stimulation. ⁷ It has, therefore, been proposed by several investigators that the key to limiting corneal fibrosis may lie in minimizing proinflammatory epithelial-stromal interactions. ^{8 9}  

These insights into the pathomechanisms underlying the loss of corneal transparency after corneal damage have implications for refractive surgery and argue for a “no-touch” strategy concerning the corneal epithelium. Accordingly, considerable effort has been devoted to developing lasers that can be used for selective intrastromal cutting (subsurface surgery). ^{10 11 12} We have recently demonstrated that this goal can be achieved with the femtosecond laser under experimental conditions. With the help of this technology, epithelial injury could be avoided and selective intrastromal surgery could be performed. ¹³ Intrastromal cutting was associated with a favorable wound-healing response and the avoidance of differentiation of keratocytes into myofibroblasts. Other potentially sight-threatening risks of transepithelial refractive surgical procedures, such as malfunctions of the microkeratome, epithelial in-growth into the lamellar interface, and microbial keratitis, ^{14 15 16} were also avoided. 

In the present study, we extended our earlier investigations and analyzed the relationship between the type of corneal injury and the induction of the profibrotic cytokine TGF-β1. We hypothesized, based on the published pathophysiological data, that an epithelial lesion would cause a robust increase in TGF-β1 expression, whereas selective intrastromal tissue ablation would avoid the increased production of this central profibrotic cytokine. To address this issue, we compared TGF-β1 mRNA and protein production in different layers of injured rabbit corneas after epithelial and stromal injury induced by photorefractive keratectomy (PRK) and selective intrastromal injury induced by femtosecond laser keratotomy. Our data demonstrated that an isolated stromal injury avoids TGF-β1 induction in the cornea and provided further molecular evidence that femtosecond laser keratotomy is a potentially useful option for refractive surgery. 

Materials and Methods

Animals

New Zealand White rabbits (n = 24; 3–4 kg body weight) were obtained from Harlan Winkelmann Laboratories (Borchen, Germany) and were housed under standard laboratory conditions. One eye of each rabbit, selected at random, was subjected to surgery. Postoperative survival times were 1, 3, 7, and 28 days. Both femtosecond laser keratotomy and PRK were performed on three eyes for each time point. Contralateral eyes served as unoperated controls. 

Animals were anesthetized with xylazine hydrochloride (5 mg/kg intramuscularly; Rompun, Bayer, Leverkusen, Germany) and ketamine hydrochloride (50 mg/kg intramuscularly; Ketavet, Pharmacia, Erlangen, Germany). In addition, preservative-free oxybuprocain hydrochloride eyedrops (Benoxinat SE 0.4%; Alcon Pharma, Freiburg, Germany) were applied to each eye just before surgery. Animals treated with PRK received buprenorphine 0.05 mg/kg subcutaneously after surgery. Animals were killed under anesthesia by intracardiac injection of 5 mL embutramine 0.2 g/mL, mebezonium iodide 0.05 g/mL, and tetracaine hydrochloride 0.005 g/mL, (T61; Intervet, Unterschleissheim, Germany). All animals were treated in accordance with German law regarding the use of laboratory animals and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Intrastromal Femtosecond Laser Keratotomy

Preoperative corneal thickness was measured using an ultrasound pachymeter (SP 100; Tomey, Erlangen, Germany). A lamellar intrastromal cut that did not extend to the corneal surface was performed (Fig. 1)on 12 rabbits at a preoperatively determined depth of 50% of the central corneal thickness (160–200 μm; diameter, 6.5 mm) with a femtosecond laser (Femtec; 20/10 Perfect Vision, Heidelberg, Germany). This pulsed solid-body (Nd:glass) laser with a repetition-rate of 10 kHz emits light with a wavelength of 1059 nm and a pulse duration of 600 to 800 fs. Laser energy of 2.8 μJ, a spot size of 5 μm, and a spot separation of 8 μm were chosen. A contact lens (radius of concave curvature, 10.5 mm; 20/10 Perfect Vision) was aligned to the center of the pupil and placed in contact with the corneal surface without using a suction ring. After laser application, dexpanthenol ointment (Bepanthen; Roche, Grenzach-Wyhlen, Germany) was applied into the conjunctival sac. 

Photorefractive Keratectomy

A wire-lid speculum was positioned, and transepithelial laser ablation (Fig. 1)was performed on 12 rabbits. Tissue ablation was performed with a fluence of 160 mJ/cm² using an excimer laser (VISX Star S3; VISX Inc., Santa Clara, CA). An optical zone of 6 mm was selected. The ablation depth determined preoperatively was 150 μm, corresponding to a sphere correction of −13.50 D. At the end of the procedure, ofloxacin ointment (Floxal; Alcon Pharma, Freiburg, Germany) was applied. 

Tissue Fixation and Sectioning

For histologic analyses, whole globes were embedded in liquid tissue freezing medium (Leica Microsystems, Nussloch, Germany). Tissue was rapidly frozen in 2-methyl-butane at −80°C. Frozen tissue blocks were stored at −80°C until sectioning. Serial sagittal corneal sections (14 μm thick) were cut with a cryostat (Leica CM 3050S; Leica Microsystems, Wetzlar, Germany). For immunofluorescence microscopy studies, sections were placed on microscope slides (SuperFrost Plus; Menzel, Braunschweig, Germany) and air dried. For laser capture microdissection (LCM), cryostat sections were mounted on autoclaved polyethylene terephthalate membrane slides (Leica Microsystems). Sections were then fixed in ice-cold acetone for 3 minutes, dried on a heater at 40°C for 10 minutes, and subjected to histochemical staining. 

Immunofluorescence Microscopy Studies

Cryostat sections were incubated with primary antibodies overnight in a humidified chamber at room temperature. Mouse anti–human transforming growth factor beta 1 (TGF-β1, clone 9016; R&D Systems, Minneapolis, MN) antibody was used at a dilution of 1:50. This antibody was produced from a hybridoma resulting from the fusion of a mouse myeloma with B cells obtained from a mouse immunized with purified, Chinese hamster ovary cell-derived, recombinant TGF-β1 and latent TGF-β1. The secondary antibody (Alexa 568 goat anti-mouse, 1:1000; Molecular Probes, Eugene, OR) was applied for 90 minutes at room temperature. A counterstain for cell nuclei was performed using Hoechst 33258 (1 μg/mL, Sigma-Aldrich, Munich, Germany). Finally, sections were coverslipped using antifade mounting medium (Fluorescent Mounting Medium; Dako, Hamburg, Germany). To verify the specificity of the antibodies, separate incubations were performed with or without primary or secondary antibodies. Sections were investigated under a microscope (BX50; Olympus, Tokyo, Japan) equipped with a color digital camera (SPOT RT; Diagnostic Instruments, Sterling Heights, MI). Figures were prepared digitally using commercially available graphics software (Photoshop CS 8.0.1; Adobe, San Jose, CA). Single fluorescent images of the same section were digitally superimposed. Images were adjusted for contrast, brightness, and sharpness. 

Histochemistry

Toluidine blue 1% (Merck, Darmstadt, Germany) was dissolved in RNase-free DEPC-treated water and filtered through a 0.22-μm filter (Millex GP; Millipore, Eschborn, Germany). The solution was then directly applied onto the mounted section, incubated for 3 minutes at room temperature, and briefly rinsed in DEPC water. After differentiation in 75% ethanol in DEPC water for 3 minutes, sections were dried on a heater at 40°C for 10 minutes and immediately subjected to laser-capture microdissection. 

Noncontact Laser-Capture Microdissection, RNA Isolation, and Reverse Transcription

Membrane slides were mounted on a laser microdissection microscope (Leica AS LMD System; Leica Microsystems) with the section facing downward. After adjusting intensity, aperture, and cutting velocity, the pulsed UV laser beam was carefully directed along the borders of the central 5 mm of the stroma, epithelium, and endothelium. LCM was performed for the three different cell layers on three corneas for each time point (Fig. 2) . Six contralateral corneas served as unoperated controls. 

The area cut was transferred by gravity alone to a microcentrifuge tube cap placed directly under the section. The tube cap was filled with a guanidine isothiocyanate-containing buffer (Buffer RLT, RNeasy Micro Kit; Qiagen, Hilden, Germany) to ensure isolation of the intact RNA. Tissue collection was verified by inspection of the tube cap. Microcentrifuge tubes were immediately transferred on ice after tissue collection, followed by three freeze-thaw cycles in a dry ice/ethanol bath. Total RNA was isolated using a purification kit (RNeasy Mini; Qiagen) according to the manufacturer’s recommendations, including on-column DNase treatment. RNA was then reverse transcribed with reverse transcription reagents (TaqMan; Applied Biosystems, Darmstadt, Germany). 

Quantitative Reverse Transcription–PCR

cDNAs were subjected to PCR with a sequence detection system (ABI Prism 7000; Applied Biosystems) and PCR master mix (TaqMan Universal; Applied Biosystems). TGF-β1 primer and probe were selected with specialized software (Primer Express; Applied Biosystems) as follows: TGF-β1 forward, 5′-AAGGGCTACCACGCCAACTT-3′; TGF-β1 reverse, 5′-CGGGTTGTGCTGGTTGTACA-3′; TGF-β1 probe, 5′-TGCCCCTACATCTGGAGCCTGGAC-3′ (amplicon size, 102 bp). For normalization of Ct-values to an endogenous control, eukaryotic 18S ribosomal RNA primers and probe were used as follows: 18S forward, 5′-CGGCTACCACATCCAAGGAA-3′; 18S reverse, 5′-GCTGGAATTACCGCGGCT-3′; 18S probe, 5′-TGCTGGCACCAGACTTGCCCTC-3′ (amplicon size, 181 bp). Probes were labeled with 6-carboxyfluorescein (FAM) at the 3′-end and with 6-carbox-tetramethyl-rhodamine (TAMRA) at the 5′-end. For amplification, a standard amplification program was used (1 cycle of 50°C for 2 minutes, 1 cycle of 95°C for 10 minutes, 45 cycles of 95°C for 15 seconds, and 60°C for 1 minute). 

Statistical Analysis

Relative quantitation of target cDNA expressed in x-fold differences was performed using the 2^−ΔΔCt method, whereas the 2^−ΔCt method was used to calculate the fold change in TGF-β1 mRNA levels in stroma and in the endothelium compared with levels found in the epithelium. ¹⁷ All quantitations were normalized to an endogenous control (18S rRNA) to account for variability in the initial concentration and quality of total RNA and in the conversion efficiency of the reverse transcription reaction. The ΔCt data were used for statistical analysis. ¹⁸ Because of the sample size and the presumed nonhomogeneity of variance, the nonparametric Mann-Whitney U test was used for group comparisons. P < 0.05 was considered significant, and all computations were performed (SPSS release 15; SPSS Inc., Chicago, IL). 

Results

TGF-β1 Protein and mRNA Are Expressed in Uninjured Corneas

With the use of immunofluorescence staining, TGF-β1 expression was observed in epithelial and endothelial cells of unwounded corneas, whereas immunoreactivity for TGF-β1 was not detectable in the corneal stroma (Fig. 3A) . 

LCM and subsequent quantitative RT-PCR allowed layer-specific quantification of gene expression levels in all corneal layers. Using this technique, TGF-β1 mRNA expression was found in epithelial cells, stromal cells, and endothelial cells of unwounded corneas. TGF-β1 mRNA levels were 1.8-fold higher in the stroma and 9.0-fold higher in the endothelium compared with levels found in the epithelium (Fig. 3B) . 

TGF-β1 mRNA and Protein Expression Do Not Increase after Intrastromal Femtosecond Laser Keratotomy

One day after intrastromal femtosecond keratotomy, the keratotomy zone became visible as a distinct acellular band detectable by Hoechst staining (Fig. 4A) . After 28 days, specimens still revealed the acellular zone. TGF-β1–positive cells were not found around this femtosecond keratotomy zone at any time point, and staining patterns of epithelial and endothelial cells were comparable to those of controls (Figs. 4A 4B 4C) . In contrast to intrastromal keratotomy, TGF-β1 was markedly upregulated after PRK. By 1 day after treatment, clumpy TGF-β1 staining was associated with the epithelial defect (not shown). Three days after treatment, pronounced TGF-β1 expression was found in the subepithelial stromal layers (Fig. 4D) . Dense cell infiltration with polymorphonuclear leukocytes became detectable by Hoechst staining in the anterior half of the central cornea, as previously described. ¹³ These cells showed no immunoreactivity for TGF-β1 (Fig. 4D) . After 7 days, we observed an increase of TGF-β1 staining in the newly regenerated and partially thickened epithelium (Fig. 4E) . Up to 28 days after surgery, staining intensity of TGF-β1 remained high in the subepithelial stromal layers and the epithelium (Fig. 4F) . 

The detection of TGF-β1 protein using immunofluorescence staining correlated well with the kinetics of TGF-β1 mRNA expression. As seen in Figure 5 , relative TGF-β1 mRNA quantification demonstrated similar expression levels in corneal layers of control animals and corresponding layers of corneas treated with femtosecond laser keratotomy. Compared with control corneas, mRNA levels had increased or decreased 1, 3, 7 or 28 days after femtosecond-laser keratotomy by factors of 1.2, 1.0, 1.1, or 0.8 in epithelial cells, 2.0, 2.6, 1.8, or 2.0 in stromal cells, and 3.0, 1.3, 1.4, or 2.6 in endothelial cells, respectively (Fig. 5) . In contrast, epithelial and stromal cell layers showed markedly increased mRNA levels after PRK. We found increases by factors of 2.3, 12.5, 23.4, or 2.0 in epithelial cells and 6.3, 6.1, 10.7, or 1.5 in stromal cells, respectively, 1, 3, 7, or 28 days after surgery. However, endothelial cells showed similar expression levels after PRK compared with controls (increase by factors of 2.5, 2.9, 1.4, and 0.9, respectively). 

These data demonstrated that TGF-β1 mRNA and protein expression did not significantly increase after intrastromal femtosecond laser keratotomy. In contrast, TGF-β1 was induced in corneal epithelial and stromal cells after PRK and showed up to 23-fold higher TGF-β1 mRNA levels than control corneas (Fig. 5) . The increase of TGF-β1 mRNA levels after PRK was accompanied by increased TGF-β1 protein production, as shown by immunofluorescence staining (Figs. 4D 4E 4F) . 

Discussion

In the present study, we analyzed changes of the profibrotic cytokine TGF-β1 in microdissected cell layers of the cornea after isolated stromal injury (femtosecond laser keratotomy) and after a defined and reproducible epithelial and stromal lesion (PRK). We designed our study as a proof-of-principle study to test the hypothesis of whether the intactness of the corneal epithelium determines the stromal repair response. Comparison of an isolated intrastromal keratotomy with femtosecond LASIK or an isolated microkeratome cut might not have yielded as clear results as a comparison with PRK. After femtosecond-LASIK or microkeratome cut, the stromal repair response is inconsistent ^{19 20} and appears to be influenced by epithelial in-growth under the flap or intraoperative implantation of epithelial cells. ^{21 22} In addition, cytokines from the injured epithelium diffuse along the lamellar interface. ^{23 24}  

The main findings of our study can be summarized as follows: First, in uninjured corneas, TGF-β1 mRNA was found in all tissue layers. Levels of TGF-β1 mRNA were higher in the corneal endothelium and stroma than in the epithelium. Second, we could show that TGF-β1 mRNA and protein were strongly upregulated in the epithelium and stroma after epithelial and stromal damage (PRK) and that this upregulation persisted for several weeks. Third, we showed that an isolated stromal injury (femtosecond laser keratotomy) avoided increased expression of TGF-β1 mRNA and protein. 

We conclude from these findings that changes in TGF-β1 expression in the injured cornea depend primarily on damage to the corneal epithelium. Isolated damage to the corneal stroma is insufficient to elicit a clinically relevant TGF-β1 increase. Given that TGF-β1 is considered the major factor in myofibroblast formation, corneal scarring, and haze formation, surgical strategies aimed at avoiding epithelial damage appear to be warranted. 

Laser-Capture Microdissection and Subsequent Quantitative RT-PCR Reveal Layer-Specific Expression Levels of TGF-β1 mRNA in the Uninjured Cornea

LCM in combination with quantitative RT-PCR can be used to obtain gene expression profiles of corneal cell populations in their “natural” state. ²⁵ We first used this powerful technique to study expression levels of the profibrotic cytokine TGF-β1 in the epithelium, stroma, and endothelium of the uninjured cornea. In line with earlier in vivo and in vitro studies, ^{26 27 28} we could detect TGF-β1 mRNA in all three tissue layers. Furthermore, with use of the LCM/quantitative RT-PCR technique, we could compare expression levels of TGF-β1 between these layers. This sensitive approach revealed highest TGF-β1 mRNA levels in the endothelium, intermediate levels in the stroma, and lowest levels in the epithelium. 

Next, we used immunofluorescence labeling to visualize TGF-β1 in the uninjured cornea. In line with Ivarsen et al., ¹⁹ we observed weak immunoreactivity for TGF-β1 in the epithelium and endothelium but not in the stroma. The lack of staining for TGF-β1 in the stroma is probably caused by the low density of cells in this tissue layer. In tissues with few TGF-β1 synthesizing cells, the protein concentration is most likely below the detection threshold for an antibody. These immunolabeling data are compatible with our mRNA measurements, since quantitative RT-PCR compares expression of TGF-β1 mRNA in extracts with equal amounts of RNA. This explains why the cell-dense epithelium contained the strongest signal for TGF-β1 protein but not for TGF-β1 mRNA. In the epithelium, a large number of cells with moderate TGF-β1 mRNA produced more protein than a few stromal cells with slightly higher TGF-β1 mRNA expression. 

Taken together, we conclude from our data that cells in all three layers expressed TGF-β1 mRNA and could potentially upregulate TGF-β1 levels after refractive surgery. This prompted us to ask which tissue layers of the cornea do, in fact, contribute to the increase of TGF-β1 after combined epithelial and stromal injury (PRK) and whether stromal injury alone is sufficient to significantly upregulate TGF-β1 mRNA expression of stromal cells (femtosecond laser keratotomy). 

TGF-β1 mRNA and Protein Expression Are Strongly Increased after PRK

It has been demonstrated that damage to the corneal epithelium and stroma (e.g., by PRK) causes the upregulation of TGF-β1. ^{29 30} Our study confirmed and extended these data using a sensitive and highly layer-specific approach (i.e., LCM of corneal tissue layers in combination with quantitative RT-PCR). With both these techniques, we could demonstrate that TGF-β1 mRNA is strongly upregulated in the epithelium (23-fold) and in the stroma (11-fold), with a maximum around 7 days after treatment. At the protein level, robust increases in TGF-β1 immunofluorescence were also observed in the regenerated epithelium and in the stroma bordering the injury site. Thus, both TGF-β1 mRNA and protein are strongly upregulated in corneal tissues injured by PRK. 

Our finding that TGF-β1 mRNA is upregulated in the epithelium and in the stroma after PRK argues for a synthesis of TGF-β1 in both corneal layers. Thus, postlesional TGF-β1 production does not appear to be restricted to the epithelium; rather, it also involves stromal cells. Based on our findings and published data, we propose the following sequence of events: After PRK, TGF-β1 is abundantly produced by injured and regenerating epithelial cells and by fibroblasts in the injured stroma. TGF-β1 penetrates the perilesional tissue, where it stimulates corneal fibroblasts and regulates their differentiation into myofibroblasts. Because TGF-β1 can also upregulate its own synthesis, ^{31 32} its secretion into the injury site could increase TGF-β1 production even further in an autocrine or a paracrine fashion. Thus, high levels of TGF-β1 accumulate at the lesion site, ultimately promoting fibrosis and causing corneal haze. 

Selective Intrastromal Cutting Avoids Upregulation of TGF-β1 mRNA and Protein

We also studied changes in TGF-β1 mRNA and protein expression after selective intrastromal cutting. In contrast to PRK, this technique completely avoids injury of the corneal epithelium. With this approach, we did not observe significant increases in TGF-β1 mRNA in any of the three layers. Similarly, immunofluorescence for TGF-β1 protein failed to reveal changes in TGF-β1 immunoreactivity. Thus, selective intrastromal cutting effectively avoided the upregulation of TGF-β1 mRNA and protein. 

Given that the stroma expresses TGF-β1 mRNA under control conditions and upregulates TGF-β1 mRNA expression after PRK, we wondered why femtosecond laser keratotomy does not elicit a measurable increase in TGF-β1 mRNA production in the stroma. The most straightforward explanation is, again, the low density of cells in the stroma. Because the stroma is cell poor, localized damage to the stroma affects only a small number of cells and, thus, changes in stromal TGF-β1 levels may simply be below the threshold for detection. However, it is also possible that the upregulation of TGF-β1 mRNA in the stroma requires an “epithelial” signal, as has been repeatedly suggested. ^{33 34} In any case, our findings that femtosecond laser keratotomy avoids TGF-β1 upregulation in the cornea nicely explains why femtosecond laser treatment avoided the differentiation of keratocytes into myofibroblasts. ¹³  

Clinical Perspectives

In the present study, we analyzed TGF-β1 expression in the cornea after two types of injury: PRK, which injures the corneal epithelium and the stroma, and intrastromal femtosecond laser keratotomy, which selectively cuts into the stroma and avoids epithelial injury. Together with findings of our earlier study, ¹³ our data show that femtosecond laser keratotomy avoids TGF-β1 upregulation and corneal scarring. These findings are also consistent with previous reports indicating that intrastromal femtosecond tissue ablation without epithelial involvement results in central corneal thinning and preservation of corneal transparency. ^{11 35} Damage to the cell-dense epithelium appears to be critical because the large number of cells in this layer produce considerable amounts of TGF-β1 after injury, which could stimulate and amplify TGF-β1 production of epithelial and stromal cells. Although stromal cells are also able to upregulate TGF-β1, this is not clinically relevant if injury to the epithelium is avoided. These experimental data call for a no-touch strategy concerning the corneal epithelium and make it attractive to refine the femtosecond laser technique for clinical use. The amount of tissue damage required to alter refraction using intrastromal fs tissue ablation might be greater than that achieved in this study. However, recent publications have shown that isolated intrastromal keratotomy with minimal ablation effect using the femtosecond laser technique will be clinically useful for intrasomal correction of presbyopia (intraCOR). ³⁶  

³  Present affiliation: Department of Dermatology and Allergology, Charité Universitätsmedizin, Berlin, Germany.

 Supported by a Goethe-University Junior Research Group Fellowship grant (CM), a Young Investigator Award (GJB), and a Paul und Ursula Klein-Foundation grant (CM).

 Submitted for publication August 10, 2008; revised October 27, 2008, February 1 and March 9, 2009; accepted June 8, 2009.

 Disclosure: C. Meltendorf, 20/10 Perfect Vision Optical Devices, Ltd. (F); G.J. Burbach, None; C. Ohrloff, None; E. Ghebremedhin, None; T. Deller, 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.

 

Figure 1.

 

Surgical procedures. Intrastromal femtosecond laser keratotomy is generated by a circular pattern of femtosecond laser pulses; the surface of a concave curved contact lens serves as reference for determination of the cutting depth. Photorefractive keratectomy is performed transepithelially with a flying spot excimer laser.

Figure 2.

 

Light microscopy (toluidine blue staining) and selective LCM of corneal cell layers: (A, B) corneal stroma, (C, D) epithelium, and (E, F) endothelium. The rectangles indicate the cutting lines for the UV laser. Scale bar, 100 μm.

Figure 3.

 

Immunofluorescence labeling of TGF-β1 in the unwounded cornea, and layer-specific quantification of TGF-β1 gene expression. (A) TGF-β1 expression was observed in the epithelial and endothelial cells of unwounded corneas, but immunoreactivity for TGF-β1 was not detectable in the corneal stroma. Portions of the epithelium and the endothelium are shown at higher magnification in the insets. All sections were counterstained with Hoechst (blue) to visualize nuclei. Scale bar, 100 μm. (B) With the use of LCM and subsequent quantitative RT-PCR, TGF-β1 mRNA expression was found in epithelial, stromal, and endothelial cells of unwounded corneas. TGF-β1 mRNA levels per cell were 1.8-fold higher in the stroma and 9.0-fold higher in the endothelium than in the epithelium.

Figure 4.

 

Immunofluorescence labeling of TGF-β1 in the rabbit cornea after femtosecond laser keratotomy (A–C) and after PRK (D–F). (A–C) After femtosecond laser keratotomy, no increase in TGF-β1 labeling was observed in the corneal stroma compared with untreated controls. The keratotomy zone is detectable as a distinct acellular band (arrowheads) within the corneal stroma. (D–F) After PRK, TGF-β1 deposits were observed directly within the subepithelial stromal layers under the zone of ablation. Up to 28 days after surgery, TGF-β1 could be detected in the subepithelial stromal layers and the epithelium. All sections were counterstained with Hoechst (blue) to visualize nuclei. Scale bars, 100 μm.

Figure 5.

 

TGF-β1 mRNA expression after femtosecond laser keratotomy and photorefractive keratectomy in corneal epithelial, stromal, and endothelial cells. Relative quantification of TGF-β1 mRNA expression levels showed no differences between unoperated control eyes and eyes treated with femtosecond laser keratotomy. In contrast, TGF-β1 was induced in corneal epithelial and stromal cells and showed up to 23-fold higher TGF-β1 mRNA levels after PRK. *P ≤ 0.05 was considered statistically significant.

The authors thank Charlotte Nolte-Uhl and Lyudmyla Rudnyeva for excellent technical assistance.