Oxygen Environment at birth

The transition to breathing oxygen is perhaps one of the most profound environmental changes one will ever experience. At term, the human fetal lung is bathed in amniotic fluid with a pO₂ of about 11 mm Hg or less than 1% oxygen (25). Focal hypoxia created by the lower oxygen levels in amniotic fluid relative to arterial levels may stimulate hypoxic signaling of hypoxia-inducible factor (HIF) and vascular endothelial growth factor (VEGF) necessary for promoting vascular growth and lung development (26). Indeed, expression of the transcription factor hypoxia inducible factor (HIF)-1α, HIF-1β, or HIF-2 are required for proper maturation of surfactant production and ability to survive at birth (27,28). Hydroxylation of prolines on HIF by HIF-dependent prolyl hydroxylases stimulates HIF degradation. Inhibition of HIF prolyl hydroxylases in preterm baboons increases HIF expression and improves lung growth and function (29, 30). Birth rapidly creates a hyperoxic environment in the airspace that may influence lung development by terminating HIF signaling. Maturational expression of antioxidant enzymes supports the idea that birth is an oxidizing event that requires protection against reactive oxygen species (ROS) produced during the transition to air (31). Although this implies the oxygen environment at birth is toxic, it may also serve to regulate normal postnatal lung growth. However, there are few studies suggesting that ambient levels of oxygen affect lung development. Perhaps the best example that birth may influence genes controlling postnatal lung development comes from a study showing how expression of the transcription factor Klf4 increases in mice two hours after birth (32). Mice lacking Klf4 display excessive proliferation and apoptosis in the respiratory epithelium and loss of α-smooth muscle actin in septal tips. This implies that the elevated expression of Klf4 at birth is required for the proper transition to air. Another example comes for a recent paper showing how oxygen-induced DNA damage signaling may influence proliferation of cardiomyocytes in the newborn mouse (33). Whether this is directly related to the oxygen environment in the lung or some form of oxidative stress created in the heart as the pAO₂ changes remains to be determined.

Although it remains unclear whether the transition to air regulates proper lung development, excessive levels clearly disrupt lung development. Excess levels are often given to preterm infants whose lungs are not adequately prepared to breathe air. Such infants often also have immature antioxidant defense that make them hypersensitive to oxygen-induced injury (34–36). Prolonged oxygen exposure can contribute to bronchopulmonary dysplasia (BPD), a chronic form of lung disease characterized by restrictive airways, inflammation, mild fibrosis, and reduced angiogenesis, increased elastogenesis, and alveolar simplification (34,37). Treatment strategies are generally supportive and designed to enhance oxygen delivery, improve monitoring of oxygen saturations, minimize tissue injury, and stimulate lung growth (for review see (38)). This includes the use of exogenous surfactant, milder ventilation strategies, inhaled nitric oxide, and retinoic acid. Such treatments have reduced infant mortality as well as increase survival of preterm infants born as young as twenty-three to twenty-four weeks gestation. However, these advances have changed the pathology of BPD from a scarring disease seen radiographically to one that requires continuous ventilation and supplemental oxygen treatment. This new BPD appears to be a form of disrupted lung development and impaired lung function that can be recapitulated in newborn animals exposed to high oxygen. Hence, there has been an increasing interest in understanding how oxygen influences lung development with a focus on the potential use of antioxidants and anti-inflammatory agents as therapeutic agents (24,39). But, recombinant human CuZn superoxide dismutase delivered intratracheally to very low-birth-weight preterm infants did not influence the incidence of BPD (40,41). It did, however, improve respiratory health at one year of corrected age and the incidence of retinopathy in preterm infants born at extremely low gestational age. It also improved oxygenation and development of the pulmonary vasculature in a preterm sheep model of prematurity (42). Unfortunately, additional studies testing the efficacy of antioxidant strategies have not gone forward, perhaps because there is some concern that antioxidants may adversely influence signaling pathways required for development of other organs such as brain.

Inflammation created by oxygen damage to the lung or by maternal and newborn infections may also contribute to BPD. But, a study in fetal sheep infected with Ureaplasma parvum, a common microorganism found in women with chorioamnionitis, showed that the inflammatory response to infection alone did not disrupt lung airspace or vascular development (43). On the other hand, maternal exposure of mice to lipopolysaccharide (LPS) and neonatal hyperoxia disrupts cardiac structure and function to the extent that it causes cardiac failure in adults (44). These studies suggest inflammation alone may not be sufficient to promote BPD. Inflammation may enhance oxygen-dependent changes in lung development and hence the pathogenesis of BPD. So, although there has been great interest in the role of oxidants and inflammation to mediate high oxygen-dependent changes in lung development, current antioxidant and anti-inflammatory strategies have yet to prove effective. Perhaps a combination therapy is needed.

Despite reduced mortality of preterm infants, survivors develop persistent pulmonary disease (PPD) later in life, and antioxidant therapy can improve some of these adverse outcomes. Children born preterm often display reduced lung function, are often rehospitalized following respiratory viral infections, and have increased risk for developing asthma (45–49). There is also a growing concern that premature infants will have neurodevelopmental delay and may be at risk for high blood pressure and heart disease as adults (50,51). These observations suggest that many infants born prematurely will require costly lifelong health care. Indeed, according to the National Heart Lung and Blood Institute (NHLBI) web site (http://www.nhlbi.nih.gov/new/press/06-07-26.htm), the annual costs of treating children born prematurely in 2005 were $26.2 billion dollars, of which 10% was just for treating infants with BPD. The high cost of treating prematurity is second to the treatment of asthma and far exceeds costs of treating cystic fibrosis. Hence, there is an urgent need to better understand how prematurity, and more specifically, how early-life exposure to oxygen, behaves as an environmental risk factor for promoting disease later in life (52).

A number of clinical trials have attempted to define an acceptable level of oxygen use in preterm infants that minimizes harm. Early clinical trials showed an increase in the incidence of retinopathy of prematurity (ROP) with the unrestricted use of oxygen (53). However, the Supplemental Therapeutic Oxygen for Pre-threshold Retinopathy of Prematurity (STOP-ROP) trial found high oxygen saturation levels did not increase the severity of ROP in those infants with prethreshold ROP (54). Subsequently, the Benefits of Oxygen Saturation Targeting (BOOST) trial showed how a higher oxygen-saturation range prolonged oxygen dependence (55). These and other studies identified the need to know whether the incidence of ROP was less in infants treated with a lower oxygen saturation target. The Surfactant, Positive Pressure, and Pulse Oximetry Randomized Trial (SUPPORT) compared the incidence of ROP in 1300 infants born between twenty-four and twenty-eight weeks gestation who were treated with the use of a low (85–89%) or high (91–95%) oxygen saturation target (56). While infants treated with a low oxygen saturation target had reduced ROP, mortality was increased when compared to infants treated with a high oxygen saturation target. Similar findings were recently reported in the BOOST II trial, which enrolled 2,400 infants at fifty-four hospitals in the United Kingdom, Australia, and New Zealand (57). While none of these clinical trials established an acceptable level of oxygen use in preterm infants, they collectively suggest preterm infants should be treated with oxygen saturations in the 90–95% range despite the increased risk for developing ROP. An alternative approach has been to define oxygen exposure as a cumulative dose over the first three days of life (58). Infants whose cumulative oxygen exposure was in the highest quartile were two to three times more likely to experience symptomatic airway dysfunction than infants in the lowest quartile. This implies the environmental load of inhaled oxygen is sufficient to predict risk of oxygen-related respiratory morbidity.

Because we recently wrote a review on animal models of oxygen-induced BPD (16), we will briefly summarize some of the main points on how high oxygen exposure at birth influences lung development in animal models. High oxygen exposure in newborn rodents promotes alveolar simplification and abnormal vascular development (59,60). Failure to produce vessels has been attributed to the oxygen-dependent loss of VEGF (61–63). Inhibiting VEGF signaling disrupts postnatal alveolar development (64–66), while overexpression of VEGF can partially protect the developing rat lung exposed to neonatal hyperoxia (67,68). In addition, neonatal hyperoxia suppresses the number of circulating endothelial cell precursors detected in blood and lung (69) and when restored can attenuate murine models of BPD (70,71). These changes may contribute to pulmonary hypertension and cardiovascular disease observed in human adolescents born preterm (72) and in adult rodents exposed to hyperoxia as neonates (73,74).

The effect of hyperoxia on the respiratory epithelium is equally complex. Alveolar epithelial type II cells express VEGF. Hyperoxia may directly inhibit expression of VEGF and indirectly by affecting the expansion of type II cells (59,75,76). Exposure of newborn rodents to hyperoxia for several days inhibits proliferation of type II cells (75). Growth arrest may be mediated by cell cycle checkpoints activated in response to oxidative DNA damage (77). The observation that overexpression of extracellular superoxide dismutase in type II cells attenuates oxygen-dependent changes in newborn lung development and DNA strand breaks supports this idea (78,79). Additional DNA damage may occur when leukocytes are recruited to the injured lung (80). Cell cycle checkpoints are thought to prevent replication of damaged DNA and also allow time for repair to take place. As seen in adult mice lacking the cell cycle inhibitor p21 (cdkn1), cell cycle arrest may limit the extent of oxidative stress and damage created by early-life oxygen exposure (81). On the other hand, we recently showed how short-term hyperoxia stimulates expansion of type II cells in newborn mice, which are subsequently growth arrested with continued exposure (82). Hyperplasia of type II cells has also been observed in a preterm baboon model of BPD and in adult rats exposed to sublethal doses of oxygen (4, 27). Although the relevance to lung disease remains to be determined, excessively expanded type II cells produced during neonatal hyperoxia are slowly pruned or depleted over several weeks when mice are returned to room air (82). Because type II cells produce surfactant necessary for reducing alveolar surface tension, express innate immune and other inflammatory genes, and function as transient amplifying precursors for type I cells, their depletion in adult mice exposed to neonatal hyperoxia may contribute to the altered lung function and host response to respiratory viral infection observed later in life (83,84).

Although early-life high oxygen exposure clearly perturbs normal lung development, sadly, a codified model of oxygen exposure in animal models has yet to be defined or accepted by the research community. This makes it difficult to compare outcomes between investigators who have used different doses and durations of hyperoxia on different developmental windows (16). Sex and strain of the animal can also influence how the lung responds to hyperoxia (85,86). We have taken two approaches to address this issue. First, preterm human infants are often born and exposed to excess oxygen during the saccular stage of lung. They typically go home breathing room air. Because the saccular stage of lung development in mice is between e17.5 and pnd 4, we have exposed mice to hyperoxia between birth and pnd4. The oxygen-exposed mice are then returned to room air so that we can study the long-term effects of oxygen exposure during the saccular phase of lung development. We found that exposure to 100% oxygen alters the balance of alveolar epithelial type I and II cells, promotes alveolar simplification, and alters the host response to influenza A virus infection (87). Exposure to 60% oxygen for four days altered lung development but not the host response to infection. Second, by exposing newborn mice to different levels of oxygen that produced the same cumulative dose, we were able to confirm the idea that quantifying neonatal cumulative excess oxygen can predict risk for respiratory morbidity later in life (88). Quantifying oxygen exposure as cumulative dose over different developmental windows may therefore help interpret findings from other studies that used different doses and durations of exposure. For example, adult mice or rats exposed to 100% oxygen as neonates develop cardiovascular disease (73,74). Perhaps the cumulative oxygen model can predict risk for high blood pressure such as seen in adolescent humans who were born preterm (72).

While preterm infants are often exposed to high oxygen at birth, people born at high altitude are born into low partial pressures of oxygen or hypoxia. It has been known for several decades that high altitude native Tibetans have a depressed response to hypoxia when at sea level (89). Whereas hypoxia stimulates erythropoietin and hemoglobin concentrations, natives born at high altitude develop larger lungs with increased functional capacity that is better suited for breathing lower partial pressures of oxygen (90). Hypoxia-induced lung growth is not seen after the age of ten, suggesting that the low oxygen-environment at birth affects postnatal lung development. Lower levels of hemoglobin have also been reported in high-altitude Tibetans and Bolivian Aymara people (91). Although lower levels of hemoglobin may seem counterproductive under high altitude conditions, it may be protective because elevated hemoglobin concentrations typically observed in hikers at high altitude are associated with mountain sickness and cardiovascular disease. In 2010, several studies were published showing how lower hemoglobin levels correlated positively with haplotypes in the HIF prolyl hydroxylase EGLN1, HIF-2α, and PPARA (92–95). Interestingly, the pattern of variation in EGLN1 differed between Tibetans and Andean natives (95). The Tibetan PHD2 haplotype (D4E/C127S) diminishes interaction of PHD2 with the heat shock protein chaperone p23, thereby reducing hydroxylation activity and downregulation of HIF pathway (96). Higher blood flow and circulating nitric oxide products may also serve to protect against high-altitude hypoxia (97). Exposure of newborn mice to low oxygen has revealed other pathways that may help clarify how a low oxygen environment influences development of the lung and other organs. For example, exposure of newborn mice to 12% oxygen causes lung simplification and intriguingly alters many of the same signaling pathways (such as TGF-β, PPAR-γ) typically activated by high oxygen exposure (98,99). As defined by elevated tumor necrosis factor-α and interleukin-6, neonatal hypoxia promotes an inflammatory state in rat heart and skeletal muscle (100). Thus, the low oxygen environment, such as seen in people living at high altitude, permanently influences postnatal lung development.

The successful adaptation of the lung to hypoxia at birth may come at an increased cost of altered long-term health of the high altitude native (for review see (90)). When compared to low landers, high-altitude natives have increased risk for cardiovascular disease particularly related to cardiac hypertrophy. A zip code study of children born at high altitude in Colorado suggests that birth at high altitude increases rehospitalization following infection with respiratory syncytial virus (RSV) (101). There is also some evidence that living at high altitude reduces brain activity (102). High-altitude natives may have lower rates of obesity (103), but are often born small for gestational age and have transient growth delay with compensatory catchup growth (90, 104). Rats living at high altitude also develop hypertension and have a shortened life span that can be reversed when birthed into sea-level amounts of oxygen (105). Not only do these data imply birth into low oxygen can be an early antecedent of disease later in life, but they also surprisingly sound like diseases attributed to prematurity and exposure to too much oxygen at birth. In other words, being born too early and exposed to too much oxygen or being born at high altitude and thus less oxygen alters lung development, host response to respiratory viral infections, cognition, and overall growth. This implies that complete development of the lung and perhaps other organs is heavily influenced by the amount of oxygen present in the environment at birth.

So, why would the lung and other organs respond to different levels of oxygen at birth? The geological record suggests that the oxygen environment ranged from as low as 15% to as high as 35% over the past 300 million years (106). These fluctuating changes in oxygen levels may have substantially influenced development and evolution of the mammalian lung and perhaps other organs (106,107). In fact, the transition from aquatic to terrestrial habitation by vertebrates likely occurred during a time when atmospheric levels of oxygen were rising (108). The cutaneous respiration and inadequate removal of carbon dioxide by aquatic species was incompatible with life on land, leading to the evolution of a more sophisticated vascular and respiratory tree (108). As organismal size and complexity increased over time, these two systems became critical for the efficient uptake and transport of oxygen to tissues and organs, thus increasing chances for survival (109). In Drosophila melanogaster, high oxygen increases body weight and wing size, whereas low oxygen promotes smaller body weight and wing size (106). Changes in oxygen concentrations negatively correlate with tracheal diameter and cell size in these insects (106, 110). Hence, the oxygen environment at birth is a modifier of evolutionary plasticity for the developing lung and other organs.

Additional Old and New World Perinatal Influences to Lung Development

The developing lung is constantly exposed to other environmental influences, both as a fetus and as a newborn. If the modern lung evolved in response to a changing oxygen environment, perhaps it evolved in response to other ancient environmental influences as well. Ozone, radiation, diet, and pathogens like viruses and bacteria are additional examples of old world perinatal influences. Because the mammalian lung evolved in an environment containing these agents, it presumably is programmed to respond to them. Consistent with this argument, expression profiling studies have defined a pattern of gene expression that defines a “time-to-birth” program wherein developmental genes are expressed first and genes involved in oxygen transport, protection against reactive oxygen species, and host defense are expressed near birth (111). This type of genetic programming ensures that the lung is ready to breathe air and defend against environmental toxins and pathogens at birth. Cigarette smoke and bisphenol A are examples of new world pollutants because they are products whose abundance in the environment is caused by human activity. The response to these agents may be more primitive or less programmed because the lung did not evolve when they were present in the environment. The following briefly describes the pulmonary response to an old world pollutant, ozone, and the new world pollutants, tobacco smoke and bisphenol A. Regardless of the exposure material, the net outcome is an effort to develop a lung that is best suited to function in an environment containing that material.

The photolysis of oxygen into ozone in the upper atmosphere by UV radiation allowed for the massive expansion of life. Normally there is little ozone in the lower atmosphere because UV radiation is weak. However, ozone has emerged as an urban pollutant created by chemical reactions between nitrogen oxides and volatile organic compounds typically produced from industries producing electricity, motor vehicle exhaust, gasoline vapors, and chemical solvents. Epidemiologic evidence suggests that ozone is an airway irritant that provokes asthma in both adults and children (18). Children may be at greatest risk because their lungs are still developing and likely to have greater exposure to ozone and other air pollutants as they play outdoors in polluted air. Indeed, early-life episodic ozone exposures of 0.5 parts per millions (ppm) in nonhuman primates and rats shorten growth of distal airways (112,113). A single acute exposure of 1 ppm for three hours is sufficient to inhibit alveolar cell proliferation in three-day-old mice (114). But, it is important to remember that ozone is not the only component of air pollution and that pollutant effects can originate from maternal exposures. For example, maternal diesel exhaust exposure enhances ozone-dependent airway hyperreactivity in newborn mice (115). Nonetheless, ozone is an oxidant gas that may influence distal airway and alveolar structure through similar mechanisms as hyperoxia, namely oxidant injury and inflammation (17). By shortening the airway and increasing alveolar size, the lung may be attempting to limit the amount of surface exposed to a highly reactive oxidant gas and yet still be able to efficiently exchange oxygen and carbon dioxide.

Tobacco smoke exposure is a significant cause of lung disease in the elderly population and is the primary cause of COPD and lung cancer. It has also been linked to airway hyperreactivity and asthma in children (18). Studies in animal models reveal maternal and neonatal exposures are sufficient to permanently affect host defense and lung development. Prenatal nicotine exposure reduces lung function in newborn rhesus monkeys presumably via the binding of nicotine to the nicotinic acetylcholine receptor (116). Neonatal mice exposed to cigarette smoke have reduced innate immunity and slight deficits in alveolar development as adults (117). Interestingly, high oxygen exposure in newborn mice enhances COPD in adult mice exposed to tobacco smoke products (118). Given the large number of pollutants present in tobacco smoke, it will be challenging to figure out which ones are responsible for altering lung development and how this happens. However, changes in innate immunity and increased airway hyperreactivity may represent an attempt to build a lung that is best suited to exclude particulate matter whose origins may be less molecularly defined than oxidant gases (i.e., antioxidants) or pathogens (i.e., toll-receptors).

Bisphenol A (BPA) is an organic chemical used to make polycarbonate plastics and epoxy resins. Because it is used to produce materials for packaging and storing food, its effects on children’s health are a concern. Fetal exposures in rhesus macaques stimulate expression of Scgb1a1 and MUC5AC/5B in the proximal conducting airways (119). Increased expression of these proteins suggests BPA may be inducing airway mucin defenses. Fetal exposure of pregnant mice to BPA reduced the extent of influenza A virus-associated pulmonary inflammation and antiviral gene expression in the offspring (120). This implies maternal BPA slightly modulates innate immunity. Similarly, maternal administration of BPA enhanced the response to ovalbumin challenge in offspring as defined by increased serum IgE, airway eosinophilia, and airway hyperresponsiveness (121). But those findings were not confirmed in a related study, which actually observed a subtle suppression of response to ovalbumin (122). This may reflect differences in experimental design and the developmental window when BPA was administered. Nonetheless, the observation that the developing lung is permanently affected by maternal exposure to an organic compound that is unlikely to reach the fetus is provocative. It suggests that the fetal lung is sampling what the mother is being exposed to and using that information to modify lung development. Once again, the response seems to be to build a lung designed to exclude materials that would negatively impact its ability to efficiently exchange oxidant gases. But, this example also raises the question of why the developing lung and not every organ is so heavily influenced by perinatal exposures.