Parabronchial remodeling in chicks in response to embryonic hypoxia

The adequate supply of oxygen is an absolute requirement for the survival of aerobic tissues. Birds may face low oxygen levels (hypoxia) in natural environments (such as at high altitude and in burrows), during diving, and under pathological conditions (Ramirez et al., 2007). Even within the egg at the end of incubation, a deficit in oxygen delivery is normally observed when the chicken embryo increases its metabolic rate by up to 60% but the gas exchange through the egg shell is not sufficiently increased to supply enough oxygen (Visschedijk, 1968; Freeman and Misson, 1970; Szdzuy et al., 2008).

Chronic hypoxia is related to morphological changes affecting the pulmonary system, which may lead to diseases such as ascites (pulmonary hypertension) (Hislop and Reid, 1976; Rabinovitch et al., 1979; Stenmark and McMurtry, 2005; Rabinovitch, 2012). In contrast, transcriptional activators involved in oxygen homeostasis can induce gene responses that might promote cellular modifications in order to increase oxygen delivery or facilitate metabolic adjustments to hypoxia (Semenza, 2000; Semenza, 2012).

During embryonic development, contrasting morpho-physiological changes induced by chronic hypoxia have been demonstrated in chickens. On the one hand, there are data on decreased organ mass proportional to reduced embryonic body mass (Azzam and Mortola, 2007; Chan and Burggren, 2005; Zhang and Burggren, 2012), reduced metabolic rate (Bjønnes et al., 1987; Mortola and Labbè, 2005; Azzam et al., 2007; Lourens et al., 2007; Mortola and Cooney, 2008) and delayed ductus arteriosus maturation (Copeland and Dzialowski, 2008). On the other hand, hypoxia can lead to increased O₂ transport in the blood (Baumann et al., 1983), intensified chorioallantoic membrane vascularization (Xu and Mortola, 1989; Tazawa et al., 1971; Ruijtenbeek et al., 2000; Strick et al., 1991) and greater lung mass (Xu and Mortola, 1989).

The consequences of hypoxia on the morphological and physiological development is, however, dependent on the specific embryonic stage of exposure. This is because there are critical windows when environmental stressors can have lasting effects and alter the trajectory of ontogenetic development (Chan and Burggren, 2005). The first half of embryonic development is considered a critical window for the embryo's survival to hypoxia, whereas the second half of incubation is related to compensatory responses of key organs to low O₂ (Zhang and Burggren, 2012). In this context, the main development and differentiation of the parabronchi, the organ for gas exchange in post hatching birds, occurs during the second half of incubation in chickens (Maina, 2003). The lumen of the parabronchi starts to develop and proliferate, connecting the secondary bronchi, on the eighth day of incubation. The atria are visible not earlier than day 15 and the infundibula on day 16; the air capillaries form and connect with blood capillaries on day 18 (Maina, 2003; Mortola, 2009). On the day of hatching (day 21), the lungs are fully formed and functional, and the blood–gas tissue barrier becomes extremely thin (Maina, 2003). Based on the ontogeny of the parabronchial structures, a reduced O₂ supply specifically during the main phase of their development and maturation may affect their morphology. In addition, this effect would last longer than the incubation phase, even under normoxic conditions after hatching. This latter idea is based on evidence of long-lasting physiological and morphological effects of hypoxia in hatchlings that returned to normoxia (Dzialowski et al., 2002; Hassanzadeh et al., 2004; Szdzuy and Mortola, 2007; Mortola, 2009; Ferner and Mortola, 2009; Amaral-Silva et al., 2017). Thus, we hypothesized that hypoxia exposure during the embryonic development of parabronchial structures affects their morphology, and that this effect endures after hatching. To test this hypothesis, chicken embryos were exposed to normoxia during the entire incubation or to 15% O₂ during days 12–18 of incubation. The parabronchial structures of normoxic 1- and 10-day-old chicks were analyzed via light and transmission electron microscopy using stereological methods.

Fertile eggs of the domestic fowl [Gallus gallus domesticus (Linnaeus 1758); Cobb500^®] were incubated (incubator Premium Ecologica^®, Belo Horizonte, Minas Gerais, Brazil) at 37.5°C and a relative humidity of 60%, and turned every 2 h. The eggs were subjected to two treatments: normoxia (Nx; 21% O₂ throughout incubation) and late hypoxia (Hx; 21% O₂ until day 11 and 15% O₂ on days 12–18). For the exposure to hypoxia on days 12–18, the Hx eggs were transferred to a hypoxic incubator kept at 14.5–15.5% O₂ by leaking a small stream of N₂ (0.2–0.4 ml min⁻¹) from a pressurized tank controlled by a flowmeter (White Martins, Osasco, São Paulo, Brazil, pure N₂). The O₂ concentration and temperature inside the hypoxic incubator were monitored continuously by an O₂ analyzer (Sensepoint XCD, Honeywell, Morris Plains, NJ, USA) and three thermometers in different places inside the incubator, respectively. At day 19, all eggs (Nx and Hx) were taken to the same hatcher and hatched together under the same environmental conditions in normoxia. After hatching, animals remained in normoxic ambient conditions until day 1 (at ∼32°C) or day 10 (at ∼29°C) for lung sampling. The protocol was performed in agreement with the guidelines of the National Council of Control in Animal Experimentation (CONCEA-MCTIC-Brazil), and was approved by the local Animal Care and Use Committee (CEUA, FCAV-UNESP, protocol number 024166/13).

Five 1- and five 10-day-old chicks (both sexes) were used for each treatment. After induction of death by overdose of anesthetic (90 mg kg⁻¹ ketamine+4.5 mg kg⁻¹ xylazine), a solution of 2.5% glutaraldehyde (Sigma-Aldrich, St Louis, MO, USA) in 0.1 mol l⁻¹ sodium phosphate buffer (pH 7.5; Synth, Diadema, São Paulo, Brazil) was driven by gravity into the trachea for fixation of the lungs. This method (adapted from Maina et al., 1989) consisted of inserting, in supine animals, an intratracheal cannula (1 and 2 mm diameter for 1- and 10-day-old chicks, respectively) connected to another cannula (4.5 mm diameter) attached to a funnel positioned 20 cm above the insertion point. The fixative solution was gravity-fed to air sacs and lungs through the funnel and cannulas. After this, the trachea was ligated and the respiratory system remained immersed in the same fixative solution for 24 h inside the coelomic cavity at 4°C. Following fixation, the coelomic cavity was opened, both whole lungs were removed from the rib cage, and the lungs were 2-mm parallel sectioned in random orientation (independent uniform random) to measure volumes by the Cavalieri principle (Michel and Cruz-Orive, 1988; Nielsen et al., 2001; Howard and Reed, 2005). A total of 12 small lung pieces were collected from both whole lungs using systematic random sampling by covering all slices with a translucent sheet marked with Cavalieri points and used as a biopsy plate.

The samples remained immersed in the same fixative solution for 15–20 h, rinsed three times in the same buffer, and post-fixed in phosphate buffered 1% osmium tetroxide (EMS, Hatfield, PA, USA) for 2 h. Dehydration was carried out in an ascending series battery through increasing concentrations of acetone (30–100%; Synth, Diadema), and the samples were embedded and included in Epon/Araldite resin (EMS).

For volume density and absolute volume estimation of lung components, 12 semi-thin sections (500 nm), one from each lung sample, were taken per animal using an ultramicrotome (Leica UC7, Wetzlar, Hessen, Germany), stained with 4% Toluidine Blue and photographed at a magnitude of 20× (Zeiss, Axio Image Z2 and camera AxioCam MRC5, Oberkochen, Baden-Württemberg, Germany). Micrographs were assembled (Mosai X function of the microscope's software) and were overlapped with frames, and point counting was used to estimate the volume densities (V%; relative volume in %) of the lung structures (Howard and Reed, 2005; Weibel et al., 2007). This method is based on point counting of each structure and its relation to point counting of the total lung volume. We also estimated the absolute volume (cm³) of each structure, which was calculated from the relationship between the structures' volume densities (V%) and the total lung volume (V_L; in cm³), i.e. the absolute volume of a certain structure was V%×V_L (Maina, 1984). The absolute volumes of each lung structure have also been standardized by body mass (M_b; cm³ g⁻¹).

The lung structures quantified were air capillaries, blood capillaries, atria, infundibula, parabronchial lumen, interparabronchial tissue, bronchi (primary and secondary) and blood vessels larger than capillaries (lung arteries, arterioles, veins and venules; Makanya and Djonov, 2009) (Fig. 1).

Five resin blocks containing a randomly selected lung sample from each animal were used for this analysis. Each block was trimmed and cut into ultrathin sections (50–70 nm) using an ultramicrotome (Leica UC7) with a diamond knife. The copper grids with sections were stained with aqueous 2% uranyl acetate and 0.4% lead citrate (Reynolds, 1963; Venable and Coggeshall, 1965) and examined using a JEOL JEM 1010 (Akishima, Tokyo, Japan) transmission electron microscope at 80 kV electron beam. For each block, 10 micrographs were taken at a magnification of 7500×, totaling 50 photomicrographs per specimen, 250 photomicrographs per treatment.

The anatomically measurable component of the diffusing capacity is the ‘anatomical diffusion factor’ (ADF), which is defined as the ratio of the respiratory surface area and the harmonic mean thickness of the blood–gas barrier (Perry, 1978, 1981). We separately estimated the ADF for the lungs of each of the four groups of samples (1 and 10 days after hatching, both incubated under normoxia and hypoxia). We randomly selected 100 electron micrographs per group (20 samples per individual) and determined the respiratory surface area of the air capillaries by counting all intersections with the surface, as well as the points that fell over the reference volume, to obtain the surface-to-volume ratio. In order to do so, we superimposed a squared grid over the electron micrographs in Gimp (version 2.8.16). The grid was randomly positioned for each of the electron micrographs. Surface area was calculated based on standard procedures (Howard and Reed, 2005), using the estimate of air capillary volume (see above) as the reference volume. Blood–gas tissue barrier thickness was calculated as two-thirds of the harmonic mean of the measured distances at each of the intersections (Weibel and Knight, 1964). In addition to a specimen-based approach leading to mean values for each group that could be subjected to statistical analyses, we also conducted a population-based approach using the pooled data for each group, yielding only a single value per age and treatment that could not be evaluated statistically (Table 4).

Data are presented as means±s.d. The results were subjected to two-way ANOVA (factors: age and incubation treatment) and, in the case of statistically significant differences (P≤0.05), means were compared by Tukey's test (5% probability). Data were tested for unequal variance and normality and appropriate transformations were performed when necessary. To compare lung volume data obtained in the present study with values given in the literature, we extracted total lung volume and body mass results from other studies, log₁₀-transformed the values and performed linear ordinary least squares (OLS) regression analyses.

General characteristics of chicks and lungs, such as number of chicks, incubation time, M_b, V_L and V_L/M_b, are presented in Table 1. No significant differences between treatments were observed regarding the above structures. M_b and V_L were greater, whereas V_L/M_b were lower, in the older animals independent of treatment.

All the structures analyzed in the present study are identified in a representative photomicrograph of a 1-day-old chick lung in Fig. 1 and a more detailed view of air and blood capillaries of 1- and 10-day-old chick lungs is shown in Fig. 2.

Air capillaries of normoxic animals occupied 13.6% of the lung in 1-day-old and 15.3% in 10-day-old animals. Hypoxia caused an increase of 27.4% and 42.7% in air capillary volume densities in 1- and 10-day-old chicks, respectively (Tukey's test, P<0.001; Table 2). In addition, air capillary volume was significantly greater in 10-day-old than in 1-day-old chicks incubated in hypoxia, but no change was observed between those incubated in normoxia (Tukey's test, P<0.001; Table 2).

Normoxic incubated animals had 21.6% of their lungs occupied by blood capillaries at 1 day of age; this value was 21.3% in 10-day-old chicks. Hypoxia during incubation increased blood capillary volume density in both 1-day-old (23.5% increase) and 10-day-old (13.8% increase) chicks (Tukey's test, P<0.001; Table 2).

The atrial volume densities of 1- and 10-day-old animals incubated in hypoxia were significantly lower than those of animals incubated in normoxia (Tukey's test: 1-day-old, P<0.001; 10-day-old, P=0.041; Table 2); the atrial volumes decreased with age in chicks incubated in normoxia, but not in hypoxia (Tukey's test, P<0.001). In chicks incubated under normoxia, the parabronchial lumen volume density was significantly higher in older animals when compared with younger animals, an effect that was inhibited by hypoxic incubation (Tukey's test, P<0.025; Table 2). Regarding the other structures assessed, such as the infundibula, bronchi, blood vessels larger than capillaries and interparabronchial tissue, no significant effect was observed to be caused by the conditions during incubation.

The absolute volumes and volumes/M_b of lung structures are shown in Table 3. The absolute volumes of air capillaries, blood capillaries, atria, infundibula, parabronchial lumen and blood vessels larger than capillaries were greater in the older chicks (Tukey's test, P<0.001 to 0.004) whereas the volume/M_b of air capillaries, blood capillaries, atria, parabronchial lumen and normoxic incubated chicks' infundibula and bronchi were lower (Tukey's test, P<0.001). Hypoxia during incubation increased absolute volumes of air and blood capillaries (Tukey's test, both P<0.001) in 1-day-old chicks, but decreased atria volume at both ages (Tukey's test, P<0.001 for both), and reduced parabronchial lumen in 10-day-old chicks (Tukey's test, P<0.001; Table 3). Similar results were observed for volume/M_b, except that air capillaries were significantly increased after hypoxia incubation in 10-day-old chicks and infundibula were decreased in 1-day-old chicks.

Table 4 summarizes the functional anatomical parameters of the lung analyzed by electron microscopy for the specimen-based approach and also the population-based approach (pooled data of all the individuals in a group). The population-based estimates correspond well with the means of the individual-based approach. Surface-to-volume ratio increased under hypoxia incubation in 1- and 10-day-old chicks (treatment: F_1,16=5.84, P=0.028; age: F_1,16=0.20, P=0.661; no interaction: F_1,16=0.00006, P=0.99). Surface area of the air capillaries also increased under hypoxia incubation in both 1- and 10-day-old chicks (treatment: F_1,16=6.01, P=0.026; age: F_1,16=4.74, P=0.045; no interaction: F_1,16=0.174, P=0.68). The harmonic mean thickness of the blood–gas tissue barrier was nearly identical in all four groups (treatment: F_1,16=0.059, P=0.810; age: F_1,16=0.014, P=0.908; no interaction: F_1,16=0.448, P=0.513). The ADFs of 1- and 10-day-old chicks exposed to hypoxia increased by 49.47% and 27.57%, respectively, compared with normoxic chicks (treatment: F_1,16=4.496, P=0.050; age: F_1,16=3.983, P=0.063; no interaction: F_1,16=0.659, P=0.429). Similarly, following the population-based approach, the ADFs were increased by approximately 48.53% and 31.72%, respectively, in the hypoxic groups.

Our results indicate that exposure to hypoxia during a main phase of embryonic lung development can induce a morphological remodeling of parabronchial tissue in chicks up to 10 days after hatching. To our knowledge, this is the first demonstration of a direct effect of hypoxia during embryonic development resulting in post-natal lung morphological remodeling. This response was characterized by an increase in the volume densities of air and blood capillaries together with an increase in respiratory surface area, involved in gas exchange, over a decrease in volume densities of air-conducting structures (atria and parabronquial lumen).

The effect of hypoxia favoring structural changes related to gas exchange seems to occur in many species. Young and adult hypoxic rats show heavier lungs (Cunningham et al., 1974) and increased lung volume (Bartlett and Remmers, 1971; Mortola et al., 1986; Sekhon and Thurlbeck, 1996), surface area (Bartlett and Remmers, 1971; Cunningham et al., 1974), number of alveoli (Sekhon and Thurlbeck, 1996), alveolar density (Bartlett and Remmers, 1971; Cunningham et al., 1974; Sekhon and Thurlbeck, 1996; Burri and Weibel, 1971) and diffusion capacity (Howell et al., 2003). Guinea pigs also show greater alveolar surface and lung volume in hypoxia (Lechner and Banchero, 1980). In addition, dogs exposed to hypoxia from a young age possess increased air volume, surface area and pulmonary diffusion capacity (Johnson et al., 1985). Hypoxic chicken embryos show increased parabronchial tissue (Lewallen and Burggren, 2015), and birds evolutionarily adapted to hypoxic exposures (Anser indicus and Chloephaga melanoptera) show thin blood–gas diffusion barriers (Maina and King, 1982; Maina et al., 2017). Even ectothermic vertebrates, such as tadpoles, show gill hypertrophy, thinner skin (blood–water barrier) and an increased capillary bed in response to hypoxia (Burggren and Mwalukoma, 1983). Moreover, in a rare case of unilateral lung aplasia, an adult one-lunged snapping turtle (Chelydra serpentina) specimen showed a marked increase in compensatory pulmonary volume and surface area, as well as a larger surface-area-to-volume ratio in the remaining lung (Schachner et al., 2017).

It is known that hypoxia affects the expression of genes associated with metabolic adaptation to low oxygen (Semenza, 2000). In this context, reduction in tissue oxygen supply can often lead to neovascularization to match the needs of the hypoxic tissues (Adair et al., 1990), as can be observed in tissues of mammals and birds, such as the brain, heart (Miller and Hale, 1970), skeletal muscles (Miller and Hale, 1970; Cassin et al., 1971; Snyder et al., 1984), and lungs (Howell et al., 2003; Burri and Weibel, 1971). An angiogenic response is observed in the pulmonary circulation of adult rats exposed to hypoxia for 2 weeks, where lower O₂ concentrations increase both the length of pulmonary vessels and the surface area of the capillary endothelium (Howell et al., 2003). High-altitude birds (Fulica americana peruviana), which live under hypobaric hypoxia, show a greater capillarity in flight muscles compared with lowland species (León-Velarde et al., 1993). It is interesting to note that this does not seem to be a response to longtime genetic selection, because Snyder et al. (1984), studying Canadian geese (Branta canadensis) and bar-headed geese (A. indicus), showed that increases in capillarity in flight muscles can also be stimulated in birds from low altitudes exposed to hypoxia during embryonic/fetal development. Regarding embryos, an increase in blood vessel number likely improves the functional surface for gas exchange under exposure to hypoxia, especially during the late stages of incubation. A 40–60% mass increment and angiogenesis of the chorioallantoic membrane, a critical structure for O₂ uptake and CO₂ release, was in fact demonstrated to occur in avian embryos (Strick et al., 1991; Wagner-Amos and Seymour, 2003; Chan and Burggren, 2005; Azzam and Mortola, 2007; Zhang and Burggren, 2012). Interestingly, an increased expression of vascular endothelial growth factor (VEGF), a protein that stimulates angiogenesis and vasculogenesis, was identified in the lungs of chicken embryos incubated in hypoxia (Lewallen and Burggren, 2015). A similar increase in VEGF may explain the greater volume density of blood capillaries observed by us in 1- and 10-day-old chicks incubated under hypoxia.

What is interesting regarding our results is that even at 10 days after hatching, under normoxic conditions, the morphological remodeling of lungs induced by hypoxia during incubation had been preserved. This indicates that lung development was affected by the low oxygen supply and that these morphological changes remained after termination of the hypoxic exposure. This agrees with the results of Ravikumar et al. (2015), who raised guinea pigs at high altitude for 4 months followed by an intermediary altitude for 7–8 months. In their study, at the end of that phase, the guinea pigs still preserved some lung structure rearrangement, such as increased type-2 cell and interstitial volumes and the reduction of alveolar ducts, without changes in type-1 epithelium volume and alveolar surface. Moreover, Okubo and Mortola (1989) exposed neonate rats to 10% hypoxia from the first to the sixth day of post-natal life, and then returned them to normoxia until 50 days of age, resulting in animals having a larger chest, anteroposterior diameter and diaphragm surface area, as well as heavier and expanded lungs. Taken together, these studies suggest that some structural changes induced during neonatal hypoxia are not reversed after the animal is returned to normoxia. Our study adds information on the direct effect of embryonic hypoxia on lung structures in post-natal life in a bird, which is difficult to demonstrate in mammals as the embryonic development occurs inside the mother's body.

Regardless of age, the respiratory surface area of our normoxia-incubated groups fell within the range of previous estimates for the domestic fowl (10.09 cm² g⁻¹ according to Abdalla et al., 1982, and 12.97 cm² g⁻¹ according to Vidyadaran et al., 1990; cf. Table 4). Incubation in hypoxia induced comparable increases in respiratory surface area in both ages (39.85% higher than 1-day-old Nx chicks; 39.13% higher than 10-day-old Nx chicks), whereas mass-specific respiratory surface area was always lower in the older groups.

In contrast to the increase in air and blood capillary volumes and respiratory surface area, the volumes of atria and the parabronquial lumen decreased. After normoxic incubation, the volume densities of atria decreased in older animals, a pattern that was not observed in chicks incubated in hypoxia, because they already showed reduced atria on the first day after hatching. As for the parabronchial lumen, it occupied a larger relative space in the lung in older animals incubated in normoxia, but this was not observed in the hypoxia group. Age did not affect air or blood capillaries in chicks incubated in normoxia, but hypoxia changed this pattern, inducing a greater volume of air capillaries in older animals. Thus, it seems that hypoxia not only changes lung structure, but also affects the anatomical design during early development after hatching.

From our results, as the total lung volume was not changed by hypoxia (Table 1), we suggest that the increase in the structures involved in gas exchange, i.e. air and blood capillary volumes (Tables 2 and 3), and air capillary surface (Table 4), are related to the reduction in volumes of the atria and parabronchial lumen in the chicks incubated in hypoxia (Tables 2 and 3). Actually, the volume densities of air and blood capillaries increased by approximately 50% and 56% in 1- and 10-day-old chicks, respectively, whereas atria and parabronchial lumen volume densities decreased by approximately 52% and 50% in the 1- and 10-day-old animals, respectively. This scenario may corroborate the idea of a remodeling of lung tissues based on an anatomical limitation related to body size (keeping the same total lung volume), similar to what happens with post-natal guinea pigs acclimated to hypoxia (Lechner and Banchero, 1980). Moreover, it would indicate a functional tissue rearrangement in the lung favoring gas exchange structures at the expense of airflow conduction regions. Indeed, a reduced atrial volume, as found in efficient flying birds, particularly in small forms (Duncker, 1972), is considered of functional importance for reducing dead space. Moreover, the changes observed in our hatchlings became more pronounced in 10-day-old chicks, which might be related to physiological changes induced by hypoxia during embryonic development. Increased metabolic rate, possibly related to catch-up growth, is indeed observed in 10-day-old chicks submitted to the same hypoxia treatment as that of the present study (Amaral-Silva et al., 2017).

In contrast to the differences found in parabronchial structure volumes and respiratory surface area, the thickness of the tissue barrier did not differ between treatments at either age (Table 4). The morphological diffusing capacity of the tissue barrier for oxygen correlates inversely with its thickness, but directly with the surface area (e.g. Hills, 1974). Birds present the thinnest tissue barrier among vertebrates (Maina and West, 2005), and the values found in our chicks (0.33–0.36 µm) are well within the range of variation found in chickens (Maina and West, 2005). Regarding the chicks incubated in hypoxia, it seems that the reduced O₂ supply to the embryonic lung tissue at days 12–18 was not enough to affect the thickness of the tissue barrier. Interestingly, the blood–gas (tissue) barrier of the Andean goose, a bird that is adapted to fly at high altitudes, is not considered remarkably thin compared with that of other birds (Maina et al., 2017). The thinnest barriers are generally found in small, metabolically demanding birds (Maina and West, 2005; Maina et al., 2017).

The relative volumes of air and blood capillaries in our chicks (Table 3) were somewhat smaller than those (55.5–64.8% for air capillaries and 21.3–33.8% for blood capillaries) demonstrated in other studies in chickens (Abdalla, 1977, as cited in Vidyadaran et al., 1988; Abdalla et al., 1982; Duncker, 1972; Duncker, 1973, as cited in Maina, 2005; Vidyadaran et al., 1987, as cited in Maina, 2005; Abdalla and Maina, 1981, as cited in Maina, 2005). These differences are because the relative volumes were calculated in the earlier studies considering only the gas exchange structures (air capillaries, blood capillaries, blood–gas barrier) as a reference volume, whereas in the present work, the volumes are all relative to the total lung volume.

One could argue that the morphological changes in the lung induced by hypoxia could be caused by differences in the time for air capillaries to fill with air during the first breaths after internal piping. Although we have not analyzed the internal and external piping time, the total duration of incubation was not different between the Nx and Hx groups (Table 1). In addition, exposure to 15% O₂ from day 12 to 18 of incubation did not change internal (Nx=476.1 h versus Hx=476.7 h) and external (Nx=488.8 h versus Hx=490.8 h) piping times in another group of embryos (N=45 and 48 for Nx and Hx, respectively; K.C.B. and P. A. Toro-Velasquez, unpublished data).

Hypoxia during incubation did not affect body mass in either group of chicks (Table 1), corroborating results from previous studies on hatchlings (Azzam et al., 2007; Lourens et al., 2007; Ferner and Mortola, 2009), but differing from others showing smaller (Hassanzadeh et al., 2004; Dzialowski et al., 2002; Amaral-Silva et al., 2017) or larger chicks (Bahadoran et al., 2010). It is known that hypoxia decreases yolk consumption (Chan and Burggren, 2005), which may be the reason for the absence of changes in body mass in our hatchlings.

Finally, lung volume was not affected by hypoxic incubation, but by age. Total lung volume was greater in older chicks, regardless of treatment, but V_L/M_b was significantly greater in younger chicks. When comparing our results with data from the literature (Fig. 3), we observed that non-galliform birds increase lung volume proportionally with body mass, following the scaling relationship of V_L=0.0199M_b^1.077. On a smaller number of species, Lasiewski and Calder (1971) found a similar relationship with lung volume, scaling as M_b^0.94, and Maina et al. (1989) found a regression for birds, including Galliformes, of V_L=0.0218M_b^1.048. Galliform birds, in contrast, show an allometric relationship of V_L=0.0251.M_b^0.9005, the slope being significantly different from that of non-galliform birds. Therefore, the reduction in relative lung volume in 10-day-old chicks (Table 1) can be explained by the non-proportionality of lung volume to body mass in Galliformes (Fig. 3).

In conclusion, our results indicate that a reduction in O₂ supply during embryonic lung development and maturation induces a morphological adjustment of parabronchi that persists through post-hatching life, for at least 10 days, in chickens. The changes indicate a parabronchial remodeling favoring pulmonary structures involved in gas exchange (increased blood and air capillary volume densities and respiratory surface area) at the expense of air-conducting structures (reduced atria and parabronchial lumen volume densities). Suitable follow-up studies could be designed to evaluate the physiological role of this remodeling in a possible increase in lung oxygen extraction in young birds that have been exposed to low O₂ concentrations during embryonic development.

This study was part of the activities developed by L.A.-S. during her Master's at the Animal Science Graduate Program, FCAV-UNESP, Jaboticabal, SP, Brazil. We would like to thank Marcia F. Mataqueiro for support with light and electron microscopy preparations. We would also like to thank the Electron Microscopy Laboratory from the FCAV-UNESP, Jaboticabal Facility for technical support. We thank Steven F. Perry for discussions about estimating ADF. M.L. would like to thank Michael H. Hofmann for supporting his research. Finally, we appreciated all the comments by the two anonymous reviewers, which resulted in the improvement of the manuscript.