Male, 8~10-week-old Sprague Dawley rats (n=99) were divided randomly into five groups based on irradiation intensity: shamtreated controls (n=9), 3 Gy (n=18), 10 Gy (n=27), 20 Gy (n=18), and 30 Gy (n=18). All animal experiments were approved by the Institutional Animal Care and Use Committee at Seoul National University College of Medicine.

Prior to irradiation, rats were firmly immobilized in a custommade acrylic mold and anesthetized by an intraperitoneal injection of ketamine (Ketalar, Yuhan corporation, Korea, 60 mg/kg) and xylazine (Rompun, Bayer, US, 2.5 mg/kg). and were firmly immobilized in a custom-made acrylic mold. Irradiation was carried out with 6 MV X-rays from a Varian 21 EX linac linear accelerator (Varian Medical Systems, Palo Alto, CA) at a dose rate of 300 cGy/min. The irradiation field was circular with a diameter of 25 cm and six rats were irradiated simultaneously. The rats were locally irradiated in the region of the head and neck with a single dose of 3, 10, 20, or 30 Gy. Sham-treated control rats were subjected to the same procedure but were not irradiated.

Rats were sacrificed at the early post-irradiation phase (1 and 8 h, and 1, 2, 3 and 5 days after irradiation) and late post-irradiation phase (10, 20, 30 and 60 days after irradiation). Three animals per group were sacrificed at each time point in each irradiated groups; one animal was sacrificed at each time in the control group. The SMGs in all irradiated groups were analyzed up to 5 days after irradiation. Only the 10 Gy irradiated rats were analyzed from 10 days after irradiation. Small pieces (0.2~1.0 mm³) were separated from resected glands for electron microscopy. SMGs were fixed in 4% paraformaldehyde immediately after removal. After a minimum fixation of 48 h, the tissue was trimmed and was processed by standard paraffin-embedding methods.

Sections 4 µm in thickness were cut and deparaffinized, and the sections were stained with hematoxylin and eosin (H & E) to obtain conventional histological sections. For morphological analysis, the sections were coded and were randomly examined by the use of light microscopy with a model BX51 microscope (Olympus, Japan). For each examination, after the entire section was examined, 10 representative non-contiguous and non-overlapping fields were systematically selected and analyzed. After an initial qualitative assessment of the morphological changes, the number of acinar and granular convoluted tubule (GCT) cells were scored in five randomly chosen fields per section (field size: 145×215 µm² at ×400 magnification) and the number of scored cells was averaged for each rat. The values were expressed as a percentage of the control value and the number of scored cells for a sham-treated control rat was set to 100%.

To detect apoptotic cells, the terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) method was performed using an ApopTaq peroxidase In Situ apoptosis detection kit (Oncor, Gaithersburg, MD). Sections were deparaffinized, rehydrated in a decreasingly graded alcohol series, and incubated with proteinase K. Endogenous peroxidase in the sections was inhibited with 3% hydrogen peroxide, and slides were incubated at 37℃ for l h with terminal deoxynucleotidyl transferase (TdT) to catalyze the addition of digoxigenin-labeled nucleotides to the 3'-OH ends of fragmented DNA. Next, slides were incubated with horseradish peroxidase (HRP)-conjugated antidigoxigenin antibodies, and DNA fragmentation was detected by staining with 3'-3'diaminobenzidine (DAB). Finally, sections were counterstained with hematoxylin. Samples from negative controls omitted the addition of TdT. The sections stained by the use of the TUNEL assay were randomly chosen from each animal. The number of acinar, GCT and striated duct (SD) cells were scored. Approximately 1,000 cells from each cell population were counted at a magnification of ×400 using a BX51 light microscope (Olympus) and the percentage of TUNEL positive cells was calculated. The level of apoptosis for each group was obtained by averaging the percentages of all of the rats.

Small pieces of tissue (0.2~1.0 mm³) were fixed with 2% glutaraldehyde in 0.1 M Na cacodylate buffer (pH 7.2) at room temperature. These samples were postfixed with 1% OsO4 at 4℃ for 1 h, dehydrated and embedded in epoxy resin. Ultra-thin sections were observed using a JEM-100CX transmission electron microscope (JOEL, Japan).

Immunohistochemistry was performed to detect the expression of AQP5 and TGF-β. Analysis was performed only for 10 Gy- and 30 Gy-irradiated animals. The primary antibodies used were rabbit anti-rat AQP5 (2 µL/mL, Chemicon, Temecula, CA) and anti-mouse TGF-β(1 mg/mL, Chemicon). After dewaxing, rehydration and blocking of endogenous peroxidase, slides were subjected to microwave antigen retrieval (800 W, 3×5 min) in 0.01 M sodium citrate buffer (pH 6.0) to enhance staining. Primary antibody diluted 1 : 200 was applied overnight in a humidified chamber at 4℃, followed by the addition of a 1 : 200 dilution of rabbit-anti-mouse secondary antibody and streptavidin-horseradish peroxidase for 30 min each. The site of immunoreaction was made visible by DAB in the presence of hydrogen peroxide. Sections were finally counterstained with hematoxylin. The sections were randomly evaluated by a certified histologist who was blinded as to the origin of all the sections. After an initial qualitative assessment of the morphological changes, cells that showed definite staining after exposure to the antibodies were evaluated. To evaluate the degree of reduced protein expression on a per cell basis, a semiquantitative score was expressed as a percentage of the control value, and the number of scored cells for a sham-treated control rat was set to 100%. The median score of the measurement from the 10 fields was taken for each rat, and the median score of for each group was obtained.

Data are expressed as means±SEM. The comparison between experimental and control data was determined by use of two-way analysis of variance, followed by Turkey's honestly significant difference post hoc test, with a value of p<0.05 as statistically significant. SPSS software (SPSS, Chicago, IK) was used for evaluation.

Morphological changes induced by SMG irradiation were assessed using conventional H & E staining. The SMG of the rat is composed of acinar, GCT, SD, intercalated duct (ID) and excretory duct (ED) cells. One hour post-irradiation, mild vacuolization of the acinar cells and some pyknotic nuclei in the acinar and GCT cells were observed. At day 3 post-irradiation, the salivary glands showed increased numbers of cytoplasmic vacuoles in the acinar cells and some degenerative changes that included frequent hyperstaining in the mucosal acini, hypostaining in the nuclei, edematic alteration and inter-acinar spaces with unclear outlines of the acini. At day 5 post-irradiation, SMGs showed histopathological changes that were similar but less prominent than those observed at day 3. These degenerative changes were detected more frequently at a high dose and a dose-related effect was evident. Tissue regeneration, evident as the re-appearance of normal histologic characteristics, was detected at day 10 post-irradiation. However, signs of late destruction in the irradiated glands including strong vacuolization, presence of many pyknotic nuclei and lysis of acinar cells were observed after day 30 post-irradiation. Large numbers of GCT and SD epithelial cells were degenerated and desquamated, and some duct cells were lysed. Interstitial edema and a slight elevation of fibrotic tissue were observed (Fig. 1). Based on morphometric analyses, irradiation did not induce changes in the number of acinar and GCT cells at any dose in early post-irradiation phase. However, the number of GCT and acinar cells decreased at late postirradiation phase (day 30 and 60, respectively; Fig. 2).

In the SMGs of control rats, apoptosis was scant in acinar, GCT and SD cells. During the early post-irradiation phase, slightly increased apoptotic indices were observed in all cell types at all radiation doses, but not more than 5% of the cells became apoptotic even after a single dose as high as 30 Gy. At the late post-irradiation phase, an increased number of apoptotic cells were seen in irradiated rats. There were differences in the apoptotic percentage of acinar cells, but most apoptosis were found in GCT and SD cells (Fig. 3A, B). Electron microscopy also showed the apoptotic cells with a particular chromatin margination and condensation from irradiated animals. More apoptotic cells were observed at days 30 and 60 post-irradiation, and apoptotic nuclei were frequent in GCT and SD cells (Fig. 3C).

Light microscopic examination of SMGs of control rats detected AQP5 expression in the apical membrane. AQP5 expression in SMGs of irradiated rats was similar to the controls for up to 2 days after exposure to X-rays. But, by day 5 post-irradiation, declined AQP5 expression was evident. The loss of AQP5 expression in 10 Gy-irradiated rats was comparable to the loss of AQP5 expression in 30 Gy-irradiated rats. Loss of AQP5 expression was observed throughout the late post irradiation phase, and was more pronounced than in the early post irradiation phase. In control SMGs, GCT, SD and ED cells had intense TGF-β expression. In contrast, no TGF-β expression was observed in acinar cells, ID cells or connective tissue. In most GCT cells, portions of the apical cytoplasm representing the secretory granules were more intensely immunoreactive than the rest of the cytoplasm. In SD and ED cells, TGF-β expression was diffuse throughout the entire cytoplasm, with substantially higher intensity in the apical portion. After irradiation, the pattern of SMG TGF-β expression was the same, but loss of TGF-β expression were observed in cells from irradiated animals. The loss was not conspicuous during the early irradiation phase. However, there was a statistically significant loss of TGF-β expression in cells from the irradiated groups at the late irradiation phase (Fig. 4, Table 1).