The experimental protocols described herein were reviewed and approved by the Animal Care and Use Committee of the Samsung Biomedical Research Institute, Seoul, Korea. This study also followed the institutional and National Institutes of Health guidelines for laboratory animal care. Timed pregnant Sprague Dawley rats were housed in individual cages with free access to water and laboratory chow, and the rat pups were delivered spontaneously. The experiment began within 10 hr after birth and continued through postnatal d 14. Ninety rat pups were divided into four groups: normoxia control group (NC, n=8); normoxia with EPO treatment group (NE, n=8); hyperoxia control group (HC, n=37); and hyperoxia with EPO treatment group (HE, n=37). Throughout the experiment, rat pups of NC and NE were kept in the standard cage with nursing mother rat at room air, and pups of HC and HE were kept in the standard cage with mother rat within 50 liter Plexiglas chamber in which hyperoxia (>95% oxygen concentration) was maintained. Humidity was maintained at 50%, and environmental temperature was maintained at 24℃. To avoid oxygen toxicity of nursing mother rat, mother rats were rotated daily between litters in normoxia groups and hyperoxia groups. Pups in NE and HE were injected with 3 unit/gm of rhEPO (Epokine prefilled: Erythropoietin alpha, CJ Corp., Seoul, Korea) intraperitoneally at 4th, 5th, and 6th postnatal day. Pups in NC and HC received normal saline administered in the same way. Survival and body weight of rat pups in each group were checked daily throughout the experiment, and sacrificed at postnatal d 14 after intraperitoneal injection of ketamine (Yuhan corp., Seoul, Korea, 50 mg/kg).

The lungs were resected after perfusion of the heart with normal saline. Right lung was saved for later biochemical analyses. Left lung was fixed in situ at a constant inflation of 20 cmH₂O pressure, and then fixed overnight in 10% buffered formalin. The fixed lung tissue was embedded in paraffin wax after tissue processing. Five-micrometer thick sections were cut from the paraffin blocks, and stained with hematoxylin and eosin. Images of each section were captured with a Magnafire digital camera through an Olympus BX40 microscope (Olympus optical co. Ltd., Tokyo, Japan) and were saved JPEG files. Analysis of each section was carried out in a blind fashion by a single observer.

Alveolarization was assessed by performing radial alveolar count (RAC), according to the method of Emery and Mithal (17). From the center of the respiratory bronchiole, a perpendicular line was drawn to the edge of the acinus (as defined by a connective tissue septum or the pleura), and the number of septa intersected by this line was counted.

The inter-alveolar distance was assessed by measuring mean linear intercept (MLI), obtained by dividing the total length of lines drawn across the lung section by the number of intercepts encountered, as described by Cooney and Thurlbeck (18).

Each successive lung field was individually assessed for the severity of interstitial fibrosis and allotted a score 0 and 8 as described by Ashcroft et al. (19) (Fig. 1). A minimum of 10 lung fields per each section were examined for each analysis.

The activity of myeloperoxidase (MPO), an indicator of neutrophil accumulation, was determined by a modification of the method by Gray et al. (20). The lung tissues were homogenized in a phosphate buffer (pH 7.4) and centrifuged at 30,000 g for 30 min. The pellet was resuspended in another phosphate buffer (50 mM, pH 6.0) with 0.5% hexadecyltrimethyl ammonium bromide. MPO activity in the resuspended pellet was assayed by measuring absorbance changes spectrophotometrically at 460 nm, using 0.167 mg/mL O-dianisidine hydrochloride and 0.0005% hydrogen peroxide. One unit of MPO activity was defined as the quantity of enzyme degrading 1 µM of peroxide/min,

Total RNA was isolated from lung tissues using the Trizol reagent (Invitrogen, Carlsbad, CA, U.S.A.) according to the manufacturer's protocol. First-strand complementary DNA (cDNA) was synthesized using random primers and the AMV reverse transcriptase (Promega Corp., Madison, WI, U.S.A.). RT-generated cDNA encoding TNF-a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified using PCR. The housekeeping gene GAPDH was used as an internal standard. The primers used were for rat-specific TNF-α and GAPDH. PCR reactions were initiated at 94℃ for 3 min, followed by 35 cycles of amplication (denaturation at 94℃ for 1 min, annealing at 55℃ for 1 min, and extension at 72℃ for 1 min) with a final primer extension at 72℃ for 5 min. The PCR products were separated on 1.5% agarose gel and stained with ethidium bromide. The intensity of each mRNA band was quantified by densitometry using a gel documentation and analysis system and normalized to values for GAPDH.

All data were presented as the mean±standard deviation. Statistical comparison between each group was performed by one-way analysis of variance with Bonferroni's correction. For comparison of survival curves, Kaplan-Meiyer analysis was performed. A p value of <0.05 was considered significant.

Exposure to oxygen (HC) significantly decreased the survival rate at the end of experiment (postnatal d 14) compared to normoxia groups (NC and NE) (27% vs. 88%, p<0.05), and this hyperoxia-induced decreased survival rate was increased to 41% with EPO treatment (HE) (Fig. 2).

Although birth weight was not different between the four groups, pups nursed in hyperoxia (HC) showed a significantly reduced body weight gain compared to normoxia groups (NC, NE) both at postnatal d 7 and 14, and this hyperoxia-induced decrease in body weight was increased at postnatal d 14 with EPO treatment (HE) (Table 1).

In morphometric analyses, significantly decreased RAC and increased MLI were observed in HC compared to normoxia groups (NC and NE) (Fig. 4A, B). These hyperoxia-induced morphometric abnormalities were significantly improved with rhEPO treatment (HE).

Representative photomicrographs showing histopathologic differences in each are given in Fig. 3. The significantly increased fibrosis score observed in HC compared to normoxia goups (NC and NE) was also improved with rhEPO treatment (HE) (Fig. 4C).

The significantly increased MPO activity and TNF-α expression observed in HC compared to normoxia groups (NC and NE) were significantly improved with rhEPO treatment (HE) (Fig. 5).