Human corneal fibroblasts were prepared from remnant donor eye tissue for penetrating keratoplasty within 8 hr post mortem. The cells were grown and maintained in Dulbecco's minimum essential medium-F12 (DMEM-F12, Gibco, NY, U.S.A.) containing 10% fetal bovine serum (FBS, Gibco, NY, U.S.A.), sodium bicarbonate (Sigma, St. Louis, MO, U.S.A.), penicillin and streptomycin of 100 units/mL (Gibco, NY, U.S.A.). The cultures were maintained in a 5% CO₂ in air, humidified atmosphere at 37℃ incubator. Approximately 3×10⁵ cells were plated in 6-well plates respectively.

The culture medium was changed to a serum-free medium pre-warmed to 37℃. The cells were divided into four groups. Group I was the control group, therefore it was not exposed to sub-lethal heat treatment. The cells of the other groups were transferred to an incubator and maintained at 43℃ for 1 hr. For recovery following stress, the cells were returned to a 37℃ incubator and allowed to recover for 1, 6 or 24 hr (Group II, III, and IV, respectively).

Following subsequent treatment, 10 µg of cell lysate (standardized by Bradford method) was dissolved into a gel sample buffer containing 0.63 M Tris-HCl pH 6.8, 8 M urea, 1% sodium dodecyl sulfate (SDS), 10% glycerol and 1% β-mercaptoethanol. 35 µL of each preparation was resolved with 10% SDS and electrophoretically transferred to a nitrocellulose membrane and processed for Western blot analysis. The monoclonal antibody to HSP 70 (StressGen Biotechnologies Co., Victoria, Canada) was diluted 1:1,000 in Trisbuffered saline (TBS) containing 1 mg/mL bovine serum albumin. Horseradish peroxidase-conjugated anti-mouse IgG (Bio-Rad, CA, U.S.A.) was used as the secondary antibody. Immunoreactive proteins were visualized as exposed bands on radiography film using the Immune-star chemiluminescent protein detection system (Pierce, Illinois, U.S.A.). Densitometric analysis was done by using the Image gauge 3.11 program in the Japanese FUJIFILM BAS-2500 system.

After each recovery time of 1, 3, and 6 hr in the reverse transcription polymerase chain reaction (RT-PCR) experiments, product #STM-507 (HSP 70B prime gene) by Stress-Gen, and modified protocol were used. 5 µL of synthesized cDNA, 5 µL of 10× polymerase buffer, 0.5 µL of Taq DNA polymerase, 2 µL of 10 mM dNTPs, 2 µL of HSP 70 sense primer, 33.5 µL of DEPC-DW and the primers necessary to amplify HSP 70 cDNA, 2 µL of HSP 70 sense primer (5'-CTCCAGCATCCGACAAGAAGC-3', 20 pmol), and 2 µL HSP 70 antisense primer (5'-ACGGTGTTGTGGGGGTTCAGG-3', 20 pmol) were added to make the total volume of 50 µL. After adding 40 µL of mineral oil, the reaction mixture was denatured at 95℃ for 3 min, put under 35 denaturing cycles at 94℃ for 1 min, annealed at 58℃ for 1 min elongated at 72℃ for 1 min, and lastly extended for 10 min at 72℃. At the same time, to compare HSP 70 cDNA amplification conditions, amplification was performed using a sense primer (5'-TGAAGGTCGGAGTCAACGGATTTCCGT-3') of G3PDH (glyceraldehyde-3-phosphatedehydrogenase) and antisense primer (5'-CATGTGGGCCATGAGGTCCAC-3'). After making 1.5 mL of agarose gel by placing 1.5 g of agarose powder per 100 cc of 1× TAE buffer and 50 µL of ethidium bromide, a 2 µL of sample buffer was created and electrophoresis was performed on the RT-PCR product at 100 V for 15 min.

S-nitroso-N-acetyl-D,L-penicillamine (SNAP) and xanthine/xanthine oxidase (X/XO) (Sigma, St. Louis, U.S.A.) were dissolved in phosphate buffered saline (Sigma, St. Louis, U.S.A.) and stored at 4℃. 15% cell growth inhibition concentration (IC₁₅ value), which was determined using MTT assay in a preliminary toxic test, was considered an intermediate concentration. We set up 1/2×IC₁₅, IC₁₅, and 2×IC₁₅, as experimental concentrations. After placing 2,500 cells in each well in a 96-well tissue culture plate, the plate was placed in a 43℃ incubator for 1 hr, and after each recovery time of 1, 6 and 24 hr, a 37℃, 5% CO₂ incubator. Each well was treated with SNAP (no treatment, 1 mM, 2 mM, 4 mM, respectively) and X/XO (no treatment, 0.005 mM/0.005 units, 0.01 mM/0.01 units, 0.02 mM/0.02 units, respectively). After 24 hr, 10 µL of MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) was added to each well and incubated at 37℃ at 5% CO₂ for 6 more hours. 90 µL of 10% SDS was added to the mixture to stop the reaction. Optical density (OD) was measured at 540 nm with a spectrophotometer (ELISA reader) for each 96-well plate. The measured OD was expressed in percent, determined cell survival rate in comparison to the control cell group, and calculated the statistical difference in the cell survival rate with one-way ANOVA. To avoid an induction of unwanted heat treatment response, media and phosphate buffered saline (PBS) were equilibrated at 37℃ before use.

In order to determine the effects of HSP expression on corneal fibroblasts, cultured corneal fibroblasts were treated with heat shock with the above method. After a certain period of recovery (1, 6, 24 hr), the cells were treated with SNAP 1 mM or X/XO 0.1/0.01 mM and the medium was discarded. Then, the cells were treated with trypsin and mixed with the medium again before being centrifuged at 4℃ for 5 min. After being centrifuged, the supernatant was discarded. The silanized microscope slide was placed on a cytospin, centrifuged at 500 rpm at room temperature for 5 min and dried. After washing the slide with a 0.05 M phosphate buffer solution for three times, the cells were treated in a balanced buffer solution for 15 min at room temperature, and under the presence of 11-digoxigenin dUTP, kept at 37℃ in TdT for 90 min. After placing the cells in a stop/wash buffer at 37℃ for 30 min, the cells were placed in bovine fetal serum to stop the immune reaction for 10 min, and again placed in anti-digoxigenin-peroxide for 30 min. After being stained with 2.24 mM diamonobenzidine, the cells were counterstained with hematoxylin and observed with an optical microscope. Under the low magnification of 100×, the number of positively stained cells was counted 5 times and the averaged. The results were compared with those from the control cells with no heat shock. ANOVA (analysis of variance) was performed.

Cells were collected by centrifuge at 1,000 rpm for 5 min, and the pellet was suspended in 50 µL 0.5% low-melting temperature agarose (type VII; Sigma) in phosphate buffered saline at 37℃. The cell mixture was pipetted onto a frosted glass slide precoated with a basal layer of 0.5% normal agarose (type III-a; Sigma) to promote cell adhesion. A coverslip was placed over the slide, and it was kept at 4℃ for 5 min. Once the coverslip was removed, an additional 200 µL 0.5% normal agarose was pipetted onto the cells. The slide was once again covered and kept at 4℃ for 5 min. After removing the coverslip, the slide was incubated in the dark at 4℃ and for 2 hr with a lysis buffer (2.5 M NaCl, 100 mM Na₂EDTA, 10 mM Tris, pH 10, 1% Triton X-100; Sigma). The slide was soaked in an electrophoresis buffer (300 mM NaOH, 10 mM Na₂EDTA) for 20 min. The buffer covered the slide to a 2-mm depth, and the slide was subjected to electrophoresis for 20 min at 12 V, 250 mA. The slide was gently washed with 3 changes of 0.4 M Tris (hydroxymethyl) aminomethane for 5 min to remove any excess alkali and was stained with 150 µL of ethidium bromide (2 µg/mL). The slide was then examined with a fluorescent microscope with an excitation filter of 515-560 nm and a barrier filter of 590 nm. This assay utilizes the principle that DNA containing more breaks will migrate a greater distance in an electrophoretic field due to a smaller mass. Tail moments (Fig. 1) were measured from a total of 50 cells per slide and two slides were viewed per sample using image analyzer software (Komet 3.1, Kinetic Imaging Co., Liverpool, U.K.). Descriptive data was expressed through means and standard deviation (mean±standard deviation). ANOVA (analysis of variance) was performed.

Fig. 2 demonstrates that subjecting cells to sub-lethal heat treatment leads to the rapid expression of HSP 70. The expression patterns of the control group and the heat-treated groups were as follows: A remarkable increase in the expression of HSP 70 (70.5 kDa) was observed after recovery from the sublethal heat shock. A slight expression of HSP 70 was revealed in group I (640 psl, in the Image Analyser 3.11 program), then increased substantially (group II, 34,660 psl), was expressed at maximum levels at 6 hr (group III, 62,370 psl), and persisted for 24 hr (group IV, 41,520 psl). These data demonstrate that heat shock can induce HSP 70 in human corneal fibroblasts prominently.

In RT-PCR, the expression of HSP 70 mRNA was not observed in the control group, but was slowly expressed after 1 hr of recovery from the sub-lethal heat treatment, reached a maximum level at 3 hr (prior to the elevation of protein level), and started to decrease following 6 hr of recovery time (Fig. 3). Protein synthesis was occurred after mRNA transcription, thus it may considered that maximum expression of HSP 70 mRNA transcription was showen at 3 hr.

Fig. 4 demonstrates that the treatment with NO and ROI decreased the viability of human corneal fibroblasts in a concentration dependant manner. Additionally, the cell viability, which was treated by NO and ROI following sub-lethal heat treatment, increased significantly in comparison to group I (control group), particularly after 6 hr of recovery time (group III, p<0.05).

Compared with the control group, the number of TUNEL positive corneal fibroblasts was significantly decreased at 6 hr after heat shock (p<0.05) (Table 1).

While the "comet tail" of fluorescent strands of DNA seen outside the cell is the characteristic feature of single strand DNA breaks, four other patterns of DNA distribution were seen using fluorescent microscopy (Fig. 5). These included a central core of DNA, a peripheral aggregation or clumping of DNA, a comet tail of DNA strands outside the cell, and diffusely dispersed DNA tending toward one pole of the nucleus. In the group not treated with SNAP or X/XO, the central core and peripheral pattern of DNA was dominant. When the concentration of SNAP or X/XO was increased, the length of the comet tail was elongated. And the pattern outside the cell and of diffusely dispersed DNA was more dominant (data not shown).

The fraction of DNA in the tail (DNA content in tail/DNA content in nucleus+tail) multiplied by the tail length is defined as the comet "tail moment" and is considered one of the most exact measures of DNA breakage making use of this assay. Accordingly, the mean (±standard deviation) tail moment was used to compare the data obtained from the different treatments (Fig. 6). The tail moments of cells without treatment of SNAP or X/XO in each group showed no significant difference (N). The tail moment increased in a dose-dependent manner in the SNAP and X/XO treated groups. However, the tail moments of group II, III, and IV showed a decrease as compared to group I. An especially significant decrease (p<0.05) in tail moment was observed in group III (6 hr of recovery time after sub-lethal heat treatment) (Fig. 6).