Increased Cardiac Output and Preserved Gas Exchange Despite Decreased Alveolar Surface Area in Rats Exposed to Neonatal Hyperoxia and Adult Hypoxia

To the Editor:

Despite data indicating that adult and adolescent survivors of chronic lung disease associated with extreme prematurity exhibit deficits in ventilatory control, airflow, and exercise tolerance (1–4), there remains a paucity of preclinical data examining the long-term effects of neonatal lung injury on adult cardiopulmonary structure and function. Although the consequences remain unclear, consistent with the inhibition of alveolar–capillary development and increased need for oxygen supplementation observed in infants with chronic lung disease (5, 6), survivors of extreme prematurity have persistent mild to moderate decreases in lung diffusing capacity (2, 7, 8). In addition, increasing preclinical and clinical evidence indicates that disrupted alveolar–capillary membrane development associated with premature chronic lung disease is associated with an increased risk for pulmonary hypertension (9–13). It is unknown, however, whether pulmonary hypertension persists in these infants or places them at increased risk of developing future pulmonary vascular disease. In adults with pulmonary arterial hypertension, decreased diffusing capacity is an independent predictor of disease-related mortality and cardiac dysfunction (14), which raises concerns about ongoing cardiopulmonary dysfunction in survivors of prematurity. Furthermore, a recently published study in young adults born prematurely demonstrated increased right ventricular (RV) mass and elevated risk for RV dysfunction that correlated with the severity of prematurity (15). However, in the absence of a clear assessment of how RV function interacts with alveolar–capillary membrane disease and gas exchange, it is difficult to draw conclusions regarding the significance of RV hypertrophy (RVH) or its relation to diffusing capacity.

Using rodent models that closely resemble the pathological features of premature chronic lung disease, we recently demonstrated that briefly exposing the developing lung to hyperoxia was associated with alveolar simplification, reduced pulmonary arterial density, elevated RV systolic pressure (RVSP), and impaired gas exchange (16, 17). In addition, when adult rats were subsequently exposed to the secondary insult of hypoxia, previous exposure to neonatal hyperoxia significantly increased RVSP and RVH yet, surprisingly, maintained cardiac output, suggesting that animals exposed to neonatal hyperoxia respond to the secondary insult of hypoxia with a more adaptive RV phenotype. A better understanding of the mechanisms by which cardiopulmonary structure and function adapt to secondary insults may allow for the development of targeted therapeutics that promote optimal alveolar–capillary development and function after neonatal lung disease. Here, we sought to understand the role of diffusing capacity, as measured by the lung diffusion factor for carbon monoxide (Df_CO), in characterizing the relative contributions of alveolar surface area versus cardiac output to alveolar gas exchange in a model of premature lung disease. We demonstrate, first, that neonatal hyperoxia, but not adult hypoxia, significantly reduces alveolar surface area; second, that although gas diffusion efficiency is impaired similarly by neonatal hyperoxia and adult hypoxia alone, there is no additive effect of the combination of both exposures; and third, that in response to hypoxia, animals previously exposed to neonatal hyperoxia exhibit increased cardiac output in the face of a decreased diffusing capacity. We interpret this increase in cardiac output as a compensatory measure aimed at maintaining gas exchange.

Materials and Methods

All procedures were approved by the Indiana University Animal Care and Use Committee, as previously described (16, 17). Sprague-Dawley rat pups were exposed to either room air (RA) or continuous hyperoxia (≥90% oxygen) for the first 10 postnatal days, corresponding to the period of saccular and early alveolar lung development (18), and were then returned to RA. To understand how neonatal hyperoxic lung injury alters the response to a secondary insult, from 10 to 12 weeks of age, animals were subsequently exposed to either RA or hypobaric hypoxia (H) (P_atm = 362 mm Hg, equivalent to Fi_O2 of 0.1 at sea level). At the end of the 2-week hypoxia exposure, cardiac output was assessed by echocardiography (VisualSonics Inc., Toronto, ON, Canada), followed by assessment of functional gas exchange by Df_CO, as previously described (see Figure 1 for experimental timeline) (16, 17).

Figure 1.

Experimental groups and experimental timeline. Newborn pups were exposed to 10 days of postnatal hyperoxia (Fi_O2 ≥ 0.9, O₂) or room air (RA) and then recovered in RA. At 10 weeks of age, animals from each group were subjected to a 2-week hypobaric hypoxia (H) exposure (P_atm = 362 mm Hg; Fi_O2 = 0.1, H). All structural and functional endpoints were analyzed at 12 weeks of age. Df_CO, lung diffusion factor for carbon monoxide.

Determination of Df_CO was performed as previously described (17). Briefly, while the animals were anesthetized with isoflurane and following interruption of mechanical ventilation, the lungs were inflated rapidly with 3 ml test gas (0.5% carbon monoxide [CO], 0.5% neon [Ne] and the balance RA) via a tracheal catheter held for 9 seconds and withdrawn back into the syringe. The concentrations of CO and Ne were immediately determined by gas chromatography (INFICON, East Syracuse, NY). Df_CO was calculated as the ratio of CO uptake to Ne dilution measured in the sample after the 9-second breath hold, using the following equation:

$$\text{D}{f}_{\text{CO}}=1-\frac{{\left[\text{CO}\right]}_{9\text{s}}}{{\left[\text{Ne}\right]}_{9\text{s}}}.$$

Because CO is highly diffusible and binds to hemoglobin in the pulmonary capillary vessels, [CO]_9s represents its uptake across the alveolar–capillary membrane during the 9-second breath hold (19). Conversely, Ne does not diffuse across the alveolar–capillary membrane; therefore, [Ne]_9s represents the degree of Ne dilution by the alveolar gas and estimates alveolar volume. The degree of Ne dilution in alveolar volume strongly correlated with left lung volume measured by fluid displacement (r = 0.716; P < 0.0001; n = 24). Df_CO, thus, correlates to the clinically used diffusing capacity of the lung for carbon monoxide per unit of alveolar volume (Dl_CO/Va) (19). Increases in alveolar membrane surface area, capillary density, hematocrit, and cardiac output increase Df_CO, whereas alveolar–capillary simplification, anemia, pulmonary hypertension, and diminished cardiac output decrease Df_CO (19–21).

After determination of Df_CO, lungs were immediately inflation fixed. The left lung volume was determined by fluid displacement and then processed and stained for determination of mean linear intercept (Lm) and alveolar surface area (estimated as 4V/Lm), as previously described (17, 22).

Results are expressed as means ± SEM, with points representing individual animals. Comparisons between experimental groups were performed by one-way ANOVA, using Tukey’s multiple comparison post-test analysis (GraphPad Prism 5, La Jolla, CA). A P value smaller than 0.05 was considered statistically significant.

Results

Compared with RA controls (RA-RA), neither neonatal hyperoxia (O₂-RA) nor adult hypoxia (RA-H) alone significantly affected adult body weight or heart rate (Table 1). All animals tolerated the 2-week hypoxia exposure well, and no mortality occurred in any group. As expected, adult hypoxia increased hematocrit similarly, whether animals had been previously exposed to RA (RA-H) or neonatal hyperoxia (O₂-H) (Table 1). Structurally, although both RA-RA and RA-H animals had preserved alveolar complexity (Figures 2A and 2B), both O₂-RA and O₂-H animals demonstrated significant alveolar simplification (Figures 2C and 2D) and airspace enlargement, as measured by mean linear intercept (Figure 2E). Both neonatal hyperoxia and adult hypoxia exposure similarly increased left lung volume (Figure 2F). Compared with RA-RA controls, O₂-RA animals had significantly reduced alveolar surface area, whereas RA-H animals had similar alveolar surface area (Figure 2G). As a result of increased lung volume, despite having significantly enlarged airspaces, alveolar surface area in O₂-H animals was not statistically significantly decreased compared with RA-RA control animals. However, compared with RA-H animals, alveolar surface area in O₂-H animals was significantly decreased (Figure 2G).

Table 1.

Effects of Neonatal Hyperoxia and Adult Hypoxia on Growth, Heart Rate, Hematocrit and Cardiac Output

	RA-RA (n = 9)	O2-RA (n = 9)	RA-H (n = 6)	O2-H (n = 12)
Body weight, g	296.2 ± 21.6	295.0 ± 22.3	254.0 ± 16.2	279.4 ± 9.3
Heart rate	372.7 ± 27.1	352.3 ± 21.7	369.4 ± 7.2	364.7 ± 6.1
Hematocrit	35.8 ± 1.1	34.1 ± 1.1	57.0 ± 1.4*,†	59.0 ± 2.5*,†
Cardiac output, ml/min	102.2 ± 19.5	89.24 ± 7.9	64.18 ± 4.7‡	102.7 ± 7.0§

Results presented as mean ± SEM. Analysis by one-way ANOVA.

^*P < 0.0001 versus RA-RA.

^†P < 0.0001 versus O₂-RA.

^‡P < 0.001 versus RA-RA.

^§P < 0.001 versus RA-H.

Figure 2.

Lung histology and alveolar structural analysis of adult control rats compared with those exposed to neonatal hyperoxia, adult hypoxia, or the combination to the two. (A–D) Representative hematoxylin and eosin–stained alveolar lung sections at 12 weeks of age. Compared with room air controls (RA-RA, A), adult hypoxia (RA-H, B) does not affect alveolar structure significantly. However, both neonatal hyperoxia alone (O₂-RA, C) and in combination with adult hypoxia (O₂-H, D) results in significant alveolar simplification. (E) Determination of mean linear intercept reveals that neonatal hyperoxia results in a similar degree of airspace enlargement in both O₂-RA and O₂-H groups. (F) Left lung volume by fluid displacement demonstrates that both neonatal hyperoxia and adult hypoxia similarly increase lung volume. (G) Estimation of alveolar surface area confirms that neonatal hyperoxia exposure results in a significant loss of surface area for gas exchange in both O₂-RA and O₂-H groups. Results are expressed as means ± SEM, with points representing individual animals. Comparisons between experimental groups were performed by one-way ANOVA, using Tukey’s multiple comparison post-test analysis. *P < 0.0001, ^‡P < 0.001 versus RA-RA. ^†P < 0.0001. ^§P < 0.001 versus RA-H. Representative images taken at 10× magnification. Scale bars = 50 μm

Functionally, relative to RA-RA controls, exposure to neonatal hyperoxia (O₂-RA) for the first 10 days of life significantly impaired gas exchange efficiency at 12 weeks of age (Figure 3). Likewise, adult hypoxia exposure alone (RA-H) reduced Df_CO to a similar degree, but surprisingly, despite a significantly reduced alveolar surface area compared with RA-H animals, there was no additive effect of both neonatal hyperoxia and adult hypoxia (O₂-H) on Df_CO. Because Df_CO is directly correlated with cardiac output (20, 21), and given that we previously showed that, compared with RA-H animals, O₂-H animals had relative preservation of cardiac output in response to hypoxia (16), we more carefully examined the relationship between Df_CO and cardiac output (Figure 4). In RA-H animals, as expected, higher values for Df_CO correlated to higher values for cardiac output (r = 0.8541; P = 0.01). However, O₂-H animals demonstrated the opposite correlation, and lower Df_CO values were associated with higher cardiac outputs (r = −0.7100; P < 0.01), indicating that, in addition to maintaining higher cardiac outputs compared with RA-H animals, O₂-H animals demonstrate higher levels of cardiac output in the face of impaired gas diffusion. Taken together, our findings suggest that despite having a significantly reduced alveolar surface area for gas exchange, rats exposed to both neonatal hyperoxia and adult hypoxia exhibit significantly increased cardiac output and maintain a capacity for gas exchange similar to that of rats exposed to hypoxia alone.

Figure 3.

Gas exchange efficiency as measured by Df_CO. Compared with room air controls (RA-RA), both neonatal hyperoxia (O₂-RA) and adult hypoxia (RA-H) alone result in a 15% reduction in Df_CO (Df_CO: 0.877 ± 0.017 [O₂-RA] and 0.897 ± 0.020 [RA-H]). There is no additive effect observed in animals exposed to the combination of both neonatal hyperoxia and adult hypoxia (0.861 ± 0.020 [O₂-H]). Values are expressed relative to room air [RA-RA] controls. Results are expressed as means ± SEM, with points representing individual animals. Comparisons between experimental groups were performed by one-way ANOVA, using Tukey’s multiple comparison post-test analysis. *P < 0.0001 versus RA-RA.

Figure 4.

Correlation of cardiac output with Df_CO by linear regression in 12-week-old hypoxia-exposed animals. As expected, in the RA-H group, an increasing Df_CO is associated with increasing cardiac output (r = 0.8541; P = 0.01, dashed line). However, in the O₂-H group, a decreasing Df_CO is modestly and significantly associated with an increasing cardiac output (r = −0.7100; P < 0.01, solid line).

Discussion

We have previously shown that animals initially exposed to neonatal hyperoxia respond to adult hypoxia with increased RVH and RVSP, suggesting a more severe pulmonary hypertensive phenotype (16). In addition, we showed that compared with adult animals exposed to hypoxia alone, animals previously exposed to neonatal hyperoxia exhibit preservation of their cardiac output. Our present data expand these findings by suggesting that, despite having persistent and significant alveolar structural simplification, with regard to gas exchange efficiency as measured by Df_CO, animals exposed to neonatal hyperoxia are more capable of adapting to adult hypoxia.

Despite significantly different pathologies, we were intrigued to find that all three experimental groups experienced similar declines in Df_CO. Animals exposed to hyperoxia alone (O₂-RA) had significant alveolar simplification, resulting in loss of alveolar surface area for gas exchange, as well as an increase in lung volume, the combination of which would decrease Df_CO. Conversely, animals exposed to adult hypoxia alone (RA-H) had preservation of alveolar complexity and a modest increase in lung volume, resulting in a trend toward increased alveolar surface area for gas exchange. Therefore, the reduction in Df_CO was likely a result of decreased cardiac output secondary to pulmonary hypertension. Although we expected to see an additive effect in animals exposed to both neonatal hyperoxia and adult hypoxia (O₂-H), these animals actually exhibited similar deficits in Df_CO. Structural analysis indicated that O₂-H animals had a trend toward increased alveolar surface area compared with O₂-RA animals, as a result of an increase in lung volume. The increase in lung volume observed may have been caused by a combination of increased lung compliance (resulting from loss of alveolar tissue, an effect of neonatal hyperoxia), which would result in a larger lung volume under a given inflation pressure, and the hypoxia-induced lung growth, which has been well-described in rodents (23, 24) and humans (25–27). Thus, compared with O₂-RA animals, O₂-H animals may have displayed a similar Df_CO secondary to a relative increase in alveolar surface area and preservation of cardiac output. In fact, it is somewhat surprising that the Df_CO in the O₂-H rats was not better than in O₂-RA animals, given that the latter exhibited a lower hematocrit and trends toward a lower alveolar surface area and a lower cardiac output. One potential explanation is that animals exposed to neonatal hyperoxia may have a higher tendency to develop intrapulmonary shunts during hypoxia (28, 29). Intrapulmonary shunts have been associated with severe neonatal chronic lung disease (30), and such a phenomenon may lead to a decrease in Df_CO and, thus, would explain why O₂-H rats had no higher Df_CO than their O₂-RA counterparts. Alternatively, hypoxia-induced pulmonary vascular remodeling may have led to development of a substantial diffusion defect in O₂-H animals.

Perhaps most unpredicted was the finding that O₂-H and RA-H animals had similar Df_CO. Importantly, compared with RA-H animals, O₂-H animals had significantly increased degrees of alveolar–capillary simplification and pulmonary hypertension, both of which should have exacerbated declines in Df_CO. We did not find an increase in hematocrit or capillary density in O₂-H animals to otherwise account for the maintenance of Df_CO. Thus, we speculate that the effect of preserved Df_CO appears to be predominately via increased cardiac output from an adaptive RV remodeling response in the setting of neonatal hyperoxia, followed by adult hypoxia. Given the recent observation that increasing severity of prematurity is associated with increasing RV mass and an increased risk for RV dysfunction (15), our findings are clinically relevant and highlight the need for an improved understanding of the balance between adaptive and pathological RVH. It is likely that there are limitations to the inherent responses that can compensate for normal stress. This is suggested by the observation that adult survivors of chronic lung disease associated with extreme prematurity exhibit reduced exercise capacity and intolerance to maximal exercise, predominately because of an inability to compensate for deficits in pulmonary function (31, 32).

The mechanisms by which the RV may adapt and preserve cardiac output in this model of postnatal hyperoxia exposure remain undefined and are the subject of ongoing studies within our laboratories. Such long-term cardiac priming may be solely secondary to chronic RV afterloading, or it may be a result of unique molecular and metabolic alterations related to early hyperoxia exposure. Given that preservation of sufficient RV function is what ultimately determines morbidity and mortality in all forms of pulmonary vascular disease (33, 34), an improved understanding of the RV remodeling response is required. To optimize cardiopulmonary function after preterm birth, future efforts should be directed at understanding the mechanisms controlling the RV adaptive responses after neonatal lung injury.

In summary, in response to hypoxia, animals previously exposed to neonatal hyperoxia, despite having significant alveolar simplification, exhibited increasing cardiac output in the face of a decreasing Df_CO. This increase in cardiac output may serve as a compensatory measure aimed at maintaining gas exchange.