Main

Bronchopulmonary dysplasia (BPD), a chronic lung disease in extremely low gestational–age neonates, is a major cause of long-term hospitalization, slow growth, recurrent respiratory illness, or death (1,2). The cause of BPD is complex and involves oxidant injury, barotrauma/volutrauma from mechanical ventilation, chronic inflammation, and disordered repair in immature lungs (3,4,5). The pathology of BPD includes inflammation, abnormal alveolarization, fibrosis, and vascular abnormalities (6,7,8,9,10,11). Pulmonary development occurs by two main processes, alveolarization and microvascular maturation. Microvascular development, which is essential for efficient gas exchange, is regulated by vascular endothelial growth factor (VEGF), which has been shown to be disrupted in BPD (12,13,14). Exposure to high concentrations of oxygen during lung development results in lung injury characterized by alterations in capillary density, endothelial cell destruction, pulmonary inflammation, and inhibition of the process of alveolarization (15,16,17,18,19).

VEGF is a potent, endothelial cell mitogen that stimulates vessel proliferation, migration, and tube formation leading to angiogenic growth of new blood vessels (15,16,17,18,19). It is essential for angiogenesis during development; the deletion of a single allele arrests angiogenesis and causes embryonic lethality (20,21,22). VEGF is highly regulated by oxygen. VEGF gene expression increases through several different mechanisms during hypoxia and decreases during hyperoxia. Tambunting et al. (23) have shown that VEGF and VEGF receptor mRNA expression are impaired in lungs of extremely preterm baboons developing BPD, contributing to the dysmorphic microvasculature and disrupted alveolarization. Furthermore, Hasan et al. (24) showed that VEGF levels are low and soluble VEGF receptor (sVEGFR)-1 levels are high at birth in the tracheal aspirates of preterm newborn infants who developed BPD later, suggesting that sVEGFR-1, the endogenous inhibitor of VEGF, may be a biological marker for BPD.

Infants dying from BPD have decreased VEGF levels in their lungs, particularly in the thickened septae (15). This finding suggests that exposure to oxygen prevents VEGF signaling in the lungs, leading to arrested microvascular maturation and BPD. Studies have examined many combinations of oxygen exposure and the effects of hyperoxia and hypoxia on vasculogenesis. Most commonly, a fraction of inspired O2 (FiO2) of >0.95 was used with “recovery” from hyperoxia at 0.21 to 0.5 to mimic the oxygen fluctuations experienced by extremely low gestational–age neonates with acute hyperoxia exposure (25,26). However, there are no studies examining the minimum FiO2 level that results in the suppression of VEGF. We therefore proposed the hypothesis that there exists a critical threshold of hyperoxia that suppresses regulators of angiogenesis in immature lungs. To test this hypothesis, we determined the response of systemic and pulmonary levels of VEGF and sVEGFR-1, as well as pulmonary angiogenesis gene profiles during graded oxygen exposure in newborn rat pups. Because male infants are at higher risk for respiratory complications (27), we examined gender differences in response to graded oxygen.

Results

Pooling of the pups for expanding litter sizes resulted in equal mean body weight and linear growth (crown to rump length) in all groups. There were no differences in mean lung wet weight or lung-to-body weight ratio after oxygen exposure. Table 1 lists the blood gas parameters (mixed venoarterial sample) in response to graded O2 exposure. As expected, PO2 levels increased significantly from 40% to 100% as compared with 21%, as did oxygen saturation (SaO2), which increased earlier from 30% to 100% and was lower at 10%. PCO2 was lower at 10%, 60%, 80%, 90%, and 100%. pH was lower at 10% and higher at 60%, 80%, and 100%. Base excess was lower at 10% and higher at 100%, and bicarbonate was lower only at 10%. The critical PO2 associated with decreased lung VEGF was 103 mm Hg achieved with inspired O2 of 60% ( Figure 1a ). PO2 was inversely correlated with lung VEGF levels (R = −0.377; P < 0.008).

Table 1 Blood gas parameters
Full size table
Figure 1
figure 1

Effects of graded inspired O2 on systemic VEGF levels (a) in all groups (n = 8 samples per group) and (b) by gender (n = 4 samples per group). In b, the solid square represents males and the open circle represents females. Data are expressed as mean ± SEM. §P < 0.01 vs. 21%; *P < 0.05; **P < 0.01 vs. females. VEGF, vascular endothelial growth factor.

Full size image

VEGF Levels

VEGF levels (pg/ml) in the serum are presented in Figure 1a (all groups) and Figure 1b (by gender). Serum VEGF was significantly increased at 70% inspired O2 vs. 60% (264.3 ± 30.8; P < 0.01) and vs. 21% (136.8 ± 4.5). A similar increase at 70% was noted in females only (298.4 ± 46.0; P < 0.01) as compared with 21% (144.8 ± 3.2). Males had higher VEGF levels at 40% (184.3 ± 44.0; P < 0.01) as compared with females (149.9 ± 6.7), but lower levels at 70% (230.2 ± 38.3 vs. 298.4 ± 46.0; P < 0.05), 80% (122.3 ± 1.4 vs. 148.4 ± 4.8; P < 0.01), and 100% (146.1 ± 7.7 vs. 177.3 ± 5.7; P < 0.01). VEGF levels (pg/mg protein) in the lungs are presented in Figure 2a (all groups) and Figure 2b (by gender). As compared with 21% (178.2 ± 55.8), lung VEGF levels were lower at 10% (42.0 ± 4.2; P < 0.01), 60% (27.7 ± 6.3; P < 0.01), 70% (32.5 ± 7.3; P < 0.01), 80% (22.6 ± 4.6; P < 0.01), 90% (29.6 ± 5.3; P < 0.01), and 100% (26.0 ± 5.1; P < 0.01). No gender effect was noted because both genders responded similarly, with increased lung VEGF levels at 30% (male: 282.9 ± 123.2, P < 0.01; and female: 288.5 ± 85.6, P < 0.01) as compared with 21% (male: 90.6 ± 25.0, and female: 165.8 ± 45.4), followed by decreases at 60%–100%.

Figure 2
figure 2

Effects of graded inspired O2 on pulmonary VEGF levels (a) in all groups (n = 8 samples per group) and (b) by gender (n = 4 samples per group). In b, the solid square represents males and the open circle represents females. Data are expressed as mean ± SEM. §P < 0.01 vs. 21%. VEGF, vascular endothelial growth factor.

Full size image

sVEGFR-1 Levels

In contrast to VEGF serum levels, sVEGFR-1 levels (pg/ml) in the serum are gender dependent, as shown in Figure 3a (all groups) and Figure 3b (by gender). Serum sVEGFR-1 levels declined at 30% (171.0 ± 61.4; P < 0.01) as compared with 21% (427.2 ± 19.7). A similar phenomenon occurred in males at 30% (146.6 ± 30.4; P < 0.01) as compared with 21% (443.3 ± 12.2), but not in females. Hypoxia (10% O2) resulted in higher sVEGFR-1 levels in males (463.8 ± 56.4; P < 0.05) than in females (244.1 ± 51.2). Similarly, hyperoxia at 50% resulted in higher sVEGFR-1 levels in males (506.2 ± 97.8; P < 0.05) than in females (142.5 ± 79.5). Lung sVEGFR-1 levels (pg/mg protein) are presented in Figure 4a (all groups) and Figure 4b (by gender). Lung sVEGFR-1 levels declined at 70% (3243.8 ± 230.0; P < 0.01) as compared with 21% (7571.1 ± 230.6). In males, lung sVEGFR-1 levels were lower at 10% (4840.3 ± 214.9; P < 0.01), 30% (5454.5 ± 449.8; P < 0.05), 50% (5602.5 ± 385.0; P < 0.05), 70% (2885.5 ± 326.7; P < 0.01), and 80% (4036.7 ± 214.6; P < 0.01) than at 21% (7431.3 ± 291.4). In females, lung sVEGFR-1 levels were lower at 10% (6201.7 ± 127.5; P < 0.05), 70% (3602.2 ± 309.5; P < 0.01), and 80% (4597.2 ± 321.0; P < 0.05) than at 21% (7710.9 ± 245.0). Consistent with serum levels, sVEGFR-1 levels were lower in males at 10% (P < 0.002), but higher at 60% (6561.3 ± 355 vs. 4769.4 ± 173.4; P < 0.003) than in females.

Figure 3
figure 3

Effects of graded inspired O2 on systemic sVEGFR-1 levels (a) in all groups (n = 8 samples per group) and (b) by gender (n = 4 samples per group). In b, the solid square represents males and the open circle represents females. Data are expressed as mean ± SEM. §P < 0.01 vs. 21%; *P < 0.05 vs. females. sVEGFR-1, soluble VEGF receptor-1.

Full size image
Figure 4
figure 4

Effects of graded inspired O2 on pulmonary sVEGFR-1 levels (a) in all groups (n = 8 samples per group) and (b) by gender (n = 4 samples per group). In b, the solid square represents males and the open circle represents females. Data are expressed as mean ± SEM. §P < 0.01 vs. 21%; **P < 0.01 vs. females. sVEGFR-1, soluble VEGF receptor-1.

Full size image

Angiogenesis-Related Gene mRNA Expression

Tables 2 – 4 show the fold difference in lung angiogenesis genes from 21%. Table 2 represents the combined male and female data (n = 4 samples per group), and Tables 3 and 4 depict male and female responses, respectively. For the combined data, the genes that responded to changes in inspired O2 from 10% to 100% with modest to robust upregulation were the antiangiogenesis-related collagen type XVIII α-1 (4- to 59-fold), endoglin (4- to 21-fold), matrix metalloproteinase-9 (2- to 5-fold), platelet-derived growth factor-α (6- to 18-fold), tissue inhibitor of metalloproteinase (TIMP)-3 (7- to 49-fold), and tumor necrosis factor-α (2- to 36-fold). Upregulation of most genes peaked at 70%, with the exception of endoglin and TIMP-3. Other genes that responded with upregulation at 70% were angiopoietin-1 (9-fold), angiopoietin-2 (15-fold), collagen type IV (9-fold), insulin-like growth factor-1 (4-fold), leptin (25-fold), and VEGF-B (15-fold). Interleukin-1β was upregulated 6-, 9-, and 5-fold at 70%, 80%, and 90%, respectively ( Table 2 ). Males exhibited a similar gene profile from 10% to 100%, but the effect was more robust. The most robust upregulation occurred at 70%, particularly for collagen type XVIII α-1 (1,068-fold), leptin (387-fold), and tumor necrosis factor-α (171-fold) ( Table 3 ). By contrast, the angiogenesis gene profile differed in females, with moderate upregulation of most genes at 80%. These genes included cyclooxygenase-1 (15-fold), epidermal growth factor (18-fold), insulin-like growth factor-1 (15-fold), interleukin-1β (47-fold), interleukin-6 (8-fold), leptin (13-fold), matrix metalloproteinase-2 (18-fold), platelet-derived growth factor-α (23-fold), TIMP-1 (19-fold), TIMP-2 (20-fold), TIMP-3 (14-fold), tumor necrosis factor-α (13-fold), and VEGFR-1 (27-fold) ( Table 4 ).

Table 2 Up- (+) or downregulation (−) of angiogenic growth factors in the lungs of neonatal rats exposed to incremental inspired O2 percentages on the first day of life for 2 h
Full size table
Table 4 Up- (+) or downregulation (−) of angiogenic growth factors in the lungs of female neonatal rats exposed to incremental inspired O2 percentages on the first day of life for 2 h
Full size table
Table 3 Up- (+) or downregulation (−) of angiogenic growth factors in the lungs of male neonatal rats exposed to incremental inspired O2 percentages on the first day of life for 2 h
Full size table

Discussion

The current study was conducted to test the hypothesis that there exists a critical threshold of hyperoxia that alters the factors regulating lung angiogenesis. Our study is the first to demonstrate a response of systemic and pulmonary angiogenesis biomarkers and genes to graded inspired O2 and PO2 levels. Despite the short (2 h) exposure, we demonstrated that PO2 and SaO2 were directly correlated with hyperoxia and inversely correlated with VEGF levels in the lungs but not in the systemic circulation. Nonlinear regression analysis revealed that the critical PO2 associated with decreased lung VEGF (57% from normoxia) was 66 mm Hg achieved with a FiO2 of 0.4. sVEGFR-1, the endogenous negative regulator of VEGF action, did not change as a function of PO2 or SaO2. We did not establish a relationship between systemic or pulmonary VEGF and sVEGFR-1. There was a robust elevation of mRNA expression for all angiogenesis genes at 70% O2, which was predominant in males. Taken together, these findings prove our hypothesis and enable us to conclude that exposure of immature lungs to even short periods of inspired O2 >40% will alter biomarkers and genes that regulate lung angiogenesis.

Rats born at term are in the late cannalicular to saccular stage of lung development (28). This corresponds to a gestational age of 24–30 wk in human lung development. BPD primarily occurs in this group of neonates (29). Therefore, the newborn rat is a good model to study the effects of graded hyperoxia on lung angiogenesis. The first test of our hypothesis demonstrated that serum levels tended to increase as a function of increasing inspired O2, with maximum levels at 70%. Based on the knowledge that VEGF is suppressed with hyperoxia, our results were surprising. It is important to note that because of the age and size of the pups, we collected mixed arterial–venous blood samples after decapitation. Therefore, three possible explanations could account for the VEGF response in serum. (i) First, short-term exposure to hyperoxia may cause an initial increase in systemic VEGF as a compensatory mechanism. This result has been demonstrated by Sheikh et al. (30) using intermittent hyperoxia for wound healing. A longer exposure may have resulted in lower levels. A duration of 2 h was chosen based on the phenotypic/posttranslational value of VEGF after hypoxia (25). (ii) A second explanation might involve cleavage of the higher VEGF membrane-bound isoforms to the lower soluble isoform, VEGF120, because our assay measures both VEGF164 and VEGF120. Cleavage of VEGF can be mediated by matrix metalloproteinase-9 (31). Indeed, we saw higher expression of matrix metalloproteinase-9, which peaked at 70%, coinciding with the peak of serum VEGF. (iii) A third possible explanation revolves around the developmental stage of the lungs. Studies by Bhandari et al. (32) showed that in tracheal aspirates of premature infants with respiratory distress syndrome, VEGF levels increased early during the first 12 h of life and followed a pattern in which there was an initial spike, followed by a decline, and then a subsequent rise, by days 21–28, in those neonates with an adverse outcome. Our serum data seem to have a similar biphasic response, with minor peaks at 30% and 40%, a decline at 50% and 60%, and a second peak at 70%.

The serum VEGF profile across the range of inspired O2 was consistent between genders. One interesting finding was the response of lung VEGF, which peaked at 30% and then declined at 40% to reach significance at 60%–100% O2. This finding demonstrates differences between the systemic and pulmonary responses to short-term oxygen exposure, with initial pulmonary compensatory elevations in VEGF occurring at a much lower FiO2. This result is important because it leads to speculation that biomarkers of angiogenesis in the systemic circulation may not be entirely representative of pulmonary responses. The differences in responses between the lungs and the serum may occur because the lungs are in direct contact with oxygen and, thus, the peaks occur earlier. Nevertheless, O2 exposure of >30%, even after only 2 h, leads to the suppression of pulmonary VEGF, which may ultimately result in aberrant angiogenesis and implies the need for further consideration of neonatal resuscitation.

VEGFR-1 is the respective tyrosine kinase receptor needed for VEGF to act physiologically (20,23,24,33). sVEGFR-1 is the splice variant of the VEGFR-1, which acts as a VEGF “trap” (21). The next test of our hypothesis involved the response of serum and pulmonary sVEGFR-1 to graded inspired O2. Of note, the systemic levels declined with 30%, whereas the pulmonary levels declined at 70%, although the response was not sustained. This result was in contrast to systemic and pulmonary VEGF, which increased at 70% and 30%, respectively. This finding strengthens the concept that sVEGFR-1 is antagonistic to VEGF. It is possible that a longer exposure to increased oxygen concentrations would have resulted in lower levels, as demonstrated by Feng et al. (34). Resumption of baseline levels would further inhibit the function of VEGF. The final test of our hypothesis was the response of pulmonary angiogenesis genes to inspired O2. The data revealed that a majority of genes were upregulated at 70%, but the most predominant effect was noted in males. At 70%, the antiangiogenic collagen type XVIII was robustly upregulated, whereas endoglin, an endothelial cell specification gene, was downregulated. In males, the negative regulators of angiogenesis were mostly upregulated. These included angiopoietin 1 and 2, which control blood vessel development and stability (35), collagen type XVIII, and TIMP-3, thus providing evidence for one of the mechanisms explaining why male infants are at higher risk for respiratory complications. In females, the response of angiogenesis genes occurred at a higher FiO2 of 0.8 and was significantly less robust, suggesting that females are more resistant to the effects of hyperoxia. Gender dimorphism may suggest a key role for gender hormones. Estrogen has been shown to be a potent stimulator of VEGF and angiogenesis, rendering a protective effect in females (36).

Despite the important clinical implications of this study, there were some limitations, including the use of mixed arterial–venous blood to determine blood gases. Arterial blood sampling in the neonatal rat is difficult; therefore, we opted to collect mixed venous–arterial blood immediately after killing the pups by decapitation. Better results would have been obtained with pure arterial sampling. Nevertheless, our study showed that short-term exposure to graded inspired O2 may result in compensatory increases in angiogenic and antiangiogenic factors, which are mostly associated with the suppression of vessel sprouting. Exposure of the immature lungs to even short periods of inspired O2 >40% will have a negative impact on lung angiogenesis and may increase the risk for the development of BPD.

Methods

This study was approved by the State University of New York, Downstate Medical Center Animal Care and Use Committee (Brooklyn, NY). Animals were cared for and handled according to the guidelines of the U.S. Department of Agriculture. The animals were killed according to the guidelines of the American Veterinary Medical Association Panel for Euthanasia.

Experimental Design

Certified infection-free timed-pregnant Sprague Dawley rats (200–300 g) carrying fetuses (9,10,11,12,13,14,15) of known gestational age (18 d) were purchased from Charles River Laboratories (Wilmington, MA). The pregnant rats were placed in cages compliant with the U.S. Department of Agriculture guidelines and allowed to stabilize for 48 h; they then remained undisturbed until the birth of their pups. The animals were housed in a 12-h day/12-h night cycle and were provided standard laboratory diet and water ad libitum. Within 2–6 h of birth, newborn pups delivered on the same day were pooled and randomly assigned to an expanded litter of 16 pups per litter to simulate the nutritional status of extremely low gestational–age neonates. Each pup was weighed and measured for linear growth (crown to rump length) and randomly assigned to 10%, 21%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100% oxygen exposure for 2 h (n = 16 pups per group). For hypoxia and hyperoxia exposure, the newborn rat pups and dams were placed in a specialized oxygen chamber, and the oxygen environment was maintained using an oxycycler (BioSpherix, Lacona, NY). At the end of the hypoxic, normoxic, or hyperoxic episode, the rat pups were immediately killed by decapitation. Blood samples (from 6 males and 6 females, respectively) were analyzed for blood gases and SaO2 using an i-Stat portable blood gas instrument (Abbot Laboratories, Abbot Park, IL). Blood samples (pooled for a total of 4 males and 4 females) and lung samples (from 6 males and 6 females, respectively) were analyzed for VEGF and sVEGFR-1 levels using enzyme immunoassay. Biopsy of the lung samples was carried out from the left lobe, rinsed in ice-cold phosphate-buffered saline, and snap-frozen in liquid nitrogen for mRNA expression of genes regulating angiogenesis using quantitative real-time PCR arrays (2 males and 2 females).

Assay of VEGF and sVEGFR-1

To determine the levels of VEGF and sVEGFR-1 in serum, blood samples were collected in sterile polypropylene tubes and allowed to clot on ice for 30 min before centrifugation at 3,000 rpm at 4 °C. The resulting serum was transferred to a clean, sterile polypropylene tube and frozen at −20 °C until assay. Immediately after the animals were killed, the whole lungs were removed and the wet weight was determined. Tissue samples were excised from the left lower lobe, rinsed several times in ice-cold phosphate-buffered saline until all evidence of blood elements was removed, snap-frozen in liquid nitrogen, and stored at −80 °C before assay. For VEGF and sVEGFR-1 levels, lung samples were placed in 2 ml of sterile normal saline and homogenized in microtubes containing ceramic beads using the FastPrep-24 instrument (MP Biomedicals, Solon, OH). The samples were centrifuged at 8,000 rpm for 20 min at 4 °C and filtered before assay. A portion of the filtrate (10 µl) was used for total cellular protein levels to standardize the levels of VEGF and sVEGFR-1 in lung homogenates. VEGF and sVEGFR-1 levels in serum and lung homogenates were determined using commercially available rat sandwich immunoassay kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s protocol. VEGF and sVEGFR-1 concentrations in the sample were determined from a linear standard curve ranging from 0 to 2,000 pg/ml. The coefficient of variation from inter- and intra-assay precision assessment was <10% for all assays.

Total Cellular Protein Assay

On the day of the assay, lung homogenates were assayed for total protein levels using the dye-binding Bio-Rad protein assay (Bio-Rad, Hercules, CA) with bovine serum albumin as a standard. The standard curve was linear from 0.05 to 1.45 mg/ml of protein.

Real-Time PCR

Total RNA was extracted from frozen lung samples using RNA Pro solution (MP Bio, Solon, OH) as previously described (37). Cleanup of the RNA was performed using RNEasy mini cleanup kits (Qiagen, Valencia, CA) followed by on-column treatment with DNase I (Qiagen). Reverse transcription was performed using a RT2 First Strand kit purchased from SABiosciences (Frederick, MD). The real-time PCR arrays were carried out with the resulting cDNA using the Rat Angiogenesis PCR array system (n = 4 per group) on 96-well plates precoated with 84 genes regulating angiogenesis using a Bio-Rad IQ5 real-time instrument (Bio-Rad). The PCR cycle and analyses were carried out as previously described (28).

Statistical Analyses

Two-way ANOVA was used to determine the differences among the oxygen groups for normally distributed data, and the Kruskal–Wallis test was used for non–normally distributed data following Bartlett’s test for equality of variances. Post hoc analysis was performed using Dunnett’s test for significance. An unpaired t test was used for comparisons between gender for normal distribution, and an unpaired t test with Welch correction was used for nonnormal data. Nonlinear regression analysis was used to determine the relationships between PO2 and VEGF or sVEGFR-1. Significance was set at P < 0.05, and data are reported as means ± SD. All analyses were two tailed and performed using SPSS (SPSS Chicago, IL).

Statement of Financial Support

This work was made possible through the Alicia and Madu Rao Family Foundation (Grant 26-2145011) and the Eunice Kennedy Shriver National Institute of Child Health & Human Development (Grant 1U54HD071594-01).

Disclosure

This paper was presented, in part, at the New York Academy of Medicine, Brooklyn Society for Pediatricians Resident Meeting, Brooklyn, New York.