Respiration During Chronic Hypoxia Andhyperoxia in Larval and Adult Bullfrogs (Rana Catesbeiana) I. Morphological Responses of Lungs, Skin and Gills

The great majority of studies on the effects of hypoxia on gas exchange processes in ectothermic vertebrates have been confined to responses to acute hypoxia. Respira tory homeostasis during acute hypoxic exposure is usually maintained by increases in the convective flow of air or water through gas exchange organs (see Shelton, 1970; Dejours, 1981, for reviews). However, given a sufficient length of hypoxic exposure, morphological and biochemical changes can complement or even supersede such initial physiological responses. Since chronic hypoxia in aquatic habitats may occur during periods of reproductive activity or developmental change in animals, an ability to evoke adjustments maintaining oxygen delivery to the tissues over the longterm may have important consequences for survival.

It is thus not surprising that frog or salamander larvae raised in chronic aquatic hypoxia develop branchial hypertrophy (Babák, 1907; Drastich, 1925; Bond, 1960; Guimond & Hutchinson, 1976). However, for some amphibian larvae the lungs and/or skin, not the gills, may be the most important sites of respiration (see Burggren & West, 1982). Yet changes in skin or lung morphology specifically related to chronic aquatic hypoxia (and hyperoxia) have yet to be investigated in amphibians. Larvae normally undergo major morphological change leading up to metamorphosis, whereas the potential for tissue differentiation is comparatively reduced in adults. Thus chron ic hypoxic or hyperoxic exposure might produce quite different morphological res ponses in the respiratory organs of larval compared with adult Rana catesbeiana.

This study, and a companion paper (Pinder & Burggren, 1983), begins an inves tigation of the morphological, physiological and biochemical responses to chronic hypoxia and hyperoxia in the larvae and adults of the bullfrog, Rana catesbeiana. In this particular report, morphological changes of the gas exchange organs induced by chronic hypoxia and hyperoxia are quantified.

Twenty-two adult bullfrogs (mean mass 97 ± 50g, mean ± 1 S.D.) and 28 tadpoles (mean mass 15·7 ± 4·6g) of developmental stages XV-XX (after Taylor & Kollros, 1946) were captured locally, and maintained on liver and fish (adults) or spinach and lettuce (tadpoles). Both populations experienced a maximum weight loss of about 5% body weight over the experimental period of 4 weeks.

Adults and larvae were each divided into three distinct populations, with body weight (and in the case of tadpoles, developmental stage) randomly distributed be tween populations. Each population was maintained in a separate gas-tight chamber (volume = 401) at 20–23 °C. The of water and gas was maintained at appropriate levels by constantly passing air/N₂ or air/O₂ mixtures through air stones inserted through the wall of each chamber.

The first population pair of adults and tadpoles, which served as the controls, were maintained under normoxic conditions, i.e. air-saturated water covered by an atmosphere of air (⁠ about 150 mmHg). The second pair of populations constituted the hypoxic group, in which both gas and water phases were maintained at a between 70 and 80 mmHg (Table 1). Acute exposure to these levels produces hyperventilation of gills and lungs in bullfrog tadpoles (Burggren & West, 1982) and the adults of other anurans (see Boutilier & Toews, 1977). The final pair of popula tions consisted of hyperoxic groups, in which both gas and water phases of the cham ber were maintained at a above 275 mmHg. Oxygen partial pressures of both gas and water phase were monitored daily, and varied less than 10 mmHg on a day to day basis. However, both frogs and tadpoles were ‘eased into’ chronic hypoxia or hyperoxia by steadily increasing or decreasing to the desired level over the course of the first 4 days of the experiment. This O₂ level was then maintained for a total elapsed exposure of 25 –28 days.

Animals were then killed, the heart immediately exposed, and a blood sample taken into a heparinized syringe for haematological analyses (see Pinder & Burggren, 1983). A catheter was tied into the ventricle and the atria were opened. Heparinized phosphate buffer (pH 7·7, sucrose added to produce 340 mosmol kg⁻¹ H₂O) contain ing India ink was then flushed through the circulation until the skin was uniformly darkened by the ink.

The following dissections/observations were then made, in the order presented.

The trachea of the undissected, freshly-killed frog was occlusively catheterized through the mouth. The catheter was connected to a T, with one arm attached to a glass, gas-tight syringe, and the other via non-compliant tubing to a calibrated Narco P-1000B pressure transducer attached to a Narco MK-IV chart recorder. Initially, the tap was opened to air and the body wall was compressed, causing collapse and emptying of the lungs. The tap was then closed and known volumes of air from the syringe were injected into the lungs until a static lung pressure of 6·0cmH₂O was achieved, which is a pressure common during buccal filling in adult Rana (West & Jones, 1975).

The lungs of larval Rana catesbeiana, and to a lesser extent the adults, are heavily pigmented, so even ink-filled capillaries in the lung walls are not distinguishable. However, when the lungs were dissected out, opened and flattened lumen side up under a coverslip, the primary lung septa were observed to form discrete invaginated cava approximately 40–70 μm in diameter. The number of these cava within a defined field of view was counted and expressed as lung wall cava 1 cm⁻². Pieces of lung tissue approximately 1 mm² were then fixed and embedded and sections made (see below). The mean minimum length of the gas diffusion pathway between capillary blood and lung gas was determined by using a calibrated monocular micrometer to measure the direct distance between the lung wall epithelial cells and the edge of the capillary lumen. An average of the five smallest diffusion pathways observed was calculated for much tissue specimen.

The internal gills of larval Rana consist of four pairs of gill arches. Each arch bears two rows of approximately 8–30 discrete gill filaments. The gill filaments are highly branched giving rise to a variable number of ‘finger-like’ projections, each 50–100 |im in length and 10–30μm in diameter (Figs 5, 6). These projections contain blood channels, and constitute the major gas exchange surface of the gills.

Excised gill arches from the left side were suspended in buffer for examination under a dissecting microscope (20–50× magnification). The following measurements were then made on each gill arch (III—VI) from the left side of the tadpole: arch length, number of filaments per arch (both rows counted), the height of each filament along one entire filament row, the cumulative filament height (the measured row ×2 to account for both rows on the arch) and the extreme height and extreme width of the filament occupying the centre position along the filament.

Individual branchial filaments were fixed and the mean minimum length of the diffusion pathway between water and capillary blood determined as described above.

Skin both from the tail and the lateral surface of the flank (tadpole), or lateral surface of the hindlimbs and lateral surface of the flank (adult), was dissected out and transferred to a saline bath. The skin pieces were trimmed to a 1 cm square, smoothed out on a microscope slide and covered with a glass coverslip. The ink-filled capillaries were usually clearly visible under 100–400× magnification. Capillary density was measured and expressed as the number of capillary ‘meshes ‘, i.e. the number of tissue spaces completely enclosed by interconnecting capillaries (see Czopek, 1965), coun ted in a defined field of view, and converted to number of capillary meshes mm⁻². After examination, the 1 cm² piece of skin was blotted dry and weighed. Pieces of tissue approximately 1 mm² were then fixed and skin thickness and the mean minimum diffusion pathway between the outer layer and capillary blood was measured as described above.

Excised tissues were fixed in 2% phosphate buffered glutaraldehyde, post-fixed in 1% phosphate buffered osmium tetroxide, and stained in uranyl acetate. Both fixative and post-fixative were adjusted with glucose to an osmolality of 325–340 mosmol kg⁻¹ H₂O to minimize tissue shrinkage. Some small amount of tissue shrinkage may have persisted, but in any event the data analysis focused on relative changes between experimental treatments. Standard ethanol dehydration was followed by Epon: Araldite resin embedding. Sections 1–4μm thick were stained toluidine blue and examined in a Bausch and Lomb Balplan phase contrast microscope.

Treatment effects (i.e. three oxygen levels) among larval populations and among adult populations were assessed initially by analysis of variance (ANOVA). Where significant (P<0·05) treatment effects existed, differences between specific means were subsequently assessed with Student ‘s t-test for independent means.

In all experimental groups, including controls, 5–25% of the animals died during the 3·5–4 week experimental period, most of them during the first week. The remain ing animals all appeared to be in a healthy condition. Analyses were based on the following numbers of individuals surviving hypoxia, normoxia and hyperoxia: adults – 9, 7 and 6, respectively; tadpoles ‐ 14, 7 and 7, respectively.

Lung morphometric data for both adults and tadpoles are presented in Fig. 1, while photomicrographs of lung tissue are presented in Figs 9A, B and 10.

Environmental had no significant treatment effect (P > 0·10, ANOVA) on any measured lung variable in adult bullfrogs.

In sharp contrast to those of adults, the lungs of tadpoles exhibited significant (P<0·05 or lower, ANOVA) morphometric differences related to environmental Weight-specific lung volume and lung wall cavum density were significantly elevated by 30% and 23%, respectively, above control levels after 4 weeks of hypoxic exposure (Fig. 1). However, the lung measurements of hyperoxic populations were not significantly different (P>0·l) from those of normoxic populations after this time, nor was the minimum blood-gas barrier in the tadpole lung significantly affected by environmental ⁠.

Environmental had no significant effect upon capillary mesh density or blood water diffusion distance of flank or hindlimb skin in adult bullfrogs (Figs 2, 3).

Again in sharp contrast to the situation in adults, highly significant changes in skin morphology were correlated with environmental in tadpoles. Tail skin thickness and weight were not greatly different between hypoxic and normoxic populations, but capillary mesh density was more than double, and the mean minimum diffusion pathway between blood and water fell by 58% in the hypoxic populations (Figs 2, 8). Tail skin variables in hyperoxic tadpoles were not significantly different from those of the normoxic populations.

The flank skin of hypoxic tadpoles was also profoundly influenced by environment al Flank skin of tadpoles exposed to 4 weeks of hypoxia was approximately one-half as thick and as heavy as the skin of those in normoxic conditions. In addition, the capillary density was elevated by 250% (Fig. 3). However, diffusion distance between blood and water remained unchanged. Hyperoxia produced no significant changes in flank skin from the normoxic condition.

Within any one of the three tadpole populations (i.e. independent of oxygen treatment effects), there was a significant correlation (P<0·05 for all correlation coef ficients, r) between body mass and the following branchial variables; gill arch length, number of gill filaments/arch, filament width and height, and cumulative filament height/arch. These variables, therefore, were expressed on a mass-specific basis to facilitate comparison between tadpole populations.

Highly significant changes associated with different environmental values occurred in every branchial variable (Fig. 4), with the exception of the mean minimum diffusion pathway between branchial blood and water. Considering gill arch V of hypoxic tadpoles, for example, the arch itself was 100% longer and bore 85% more gill filaments than gill arches of normoxic tadpoles. In addition, the centre gill filament was both 85% wider and taller than in normoxic tadpoles. The marked enlargement of individual gill filaments from the centre of arch V in hypoxic com pared with normoxic or hyperoxic tadpoles is shown in Fig. 5. Cumulative filament height for gill arch V was more than doubled by chronic hypoxic exposure. Changes of similar magnitude were observed in gill arches IV and VI, but were attenuated in arch III, which in all three populations was a much smaller branchial structure. While respiratory surface area per se was not measured, these data clearly indicate a marked branchial hypertrophy associated with hypoxic exposure.

No significant branchial changes (P<0·10, ANOVA) from the control tadpole population were associated with chronic hyperoxic exposure.

Important morphological differences emerge between the gills, skin and lungs of R. catesbeiana tadpoles which may affect oxygen transfer across these organs. In the gills and lungs, for example, minimum gas diffusion pathways range from 2 to 4 μm,whereas in the skin covering the tail and trunk, the diffusion pathway is ten times greater (Figs 1– 4). In the earlier developmental stages, however, the organs of oxygen uptake in decreasing order of importance are skin, gills and lungs (Burggren & West, 1982; Burggren, Feder & Pinder, 1983). Clearly, other factors such as respiratory surface area, the gradient between capillary blood and respiratory medium, and the convective supply of both water/air and blood, greatly influence the relative oxygen uptake performance of these respiratory structures.

Metamorphosis in R. catesbeiana tadpoles is accompanied by the total loss, reduction, or rearrangement of entire organ systems. The gills atrophy and disappear, as does the large tail, which in early developmental stages has a high surface area to mass ratio (Burggren & West, 1982), and probably makes a large contribution to cutaneous gas exchange. With respect to oxygen uptake, at least, there is thus a major shift of gas exchange from aquatic sources to aerial uptake via the enlarging lungs (see Burg gren & West, 1982), which undergo a 10-fold increase in mass-specific lung volume (Fig. 1). Although the degree of septation of the adult lung is about one-half that in larvae (Fig. 1), the net effect of metamorphosis must be a large increase in mass specific surface area. The minimum diffusion distance between pulmonary capillary blood and lung gas, however, was unaffected by metamorphosis. The lungs of lower vertebrates are generally perfused at much higher blood pressures with fluid of lower colloid osmotic pressure than in mammals, which may produce a large net plasma filtration into the lungs (Burggren, 1982). To prevent accumulation of excessive fluid in the alveolar spaces, the reductions in capillary and lung wall structure and thickness (which might alter gas diffusion path length) possible upon maturity may be restricted by a need to maintain a comparatively thick and ‘leak-free’ blood/gas barrier.

In R. catesbeiana tadpoles of developmental stages IV-XIX, about 60% of oxygen uptake occurs via the skin, while this falls to only 20% in the adult bullfrog (Burggren & West, 1982). This may result in part from the changing ventilation/perfusion relationships of the cutaneous and pulmonary vascular bed, but can also be partially explained by the three-fold increase in diffusion distance between the skin surface and cutaneous capillaries (Figs 2, 3).

In essence, the morphology of all the gas exchange organs of R. catesbeiana tad poles was profoundly adjusted during the course of 4 weeks of low oxygen exposure, while the respiratory structures of adult bullfrogs were essentially unaltered. Within a single amphibian species, then, the morphological response to chronic hypoxic stress clearly differs with developmental state.

The magnitude of hypoxic-induced changes in the skin, gills and lungs of tadpoles was extraordinary. For example, in tadpole skin (Figs 2, 3, 7, 8), the blood-water gas diffusion pathway as well as capillary mesh density doubled, reflecting a large increase in both number and surface area of cutaneous capillaries. All of these changes may help to maintain cutaneous oxygen uptake in the face of decreasing ambient ⁠.

Mass-specific lung volume and the degree of lung septation in tadpoles were also sig nificantly higher after chronic hypoxic exposure which, coupled with physiological res ponses of hyperventilation during hypoxic exposure (West & Burggren, 1982, 1983), would facilitate pulmonary oxygen uptake during aerial hypoxia.

A marked branchial hypertrophy also occurred, particularly in the largest ginarches (Figs 4, 5). A similar branchial hypertrophy after hypoxic exposure for 7–30 days has been reported in larval amphibians (Babák, 1907; Drastich, 1925; Bond, 1960; Guimond & Hutchinson, 1976) and larval fishes (see McDonald & McMahon, 1977). As with the urodele amphibians examined in these earlier studies, a large increase in the respiratory surface area of the internal gills of the bullfrog tadpole should facilitate aquatic oxygen uptake, particularly if any component of the profound branchial hyperventilation in response to acute hypoxia (West & Burggren, 1982) remains during chronic hypoxic exposure.

In contrast to the profound morphological adjustments in tadpoles, the respiratory structures (lungs, skin) of the adult bullfrog essentially remained unaltered after 4 weeks of hypoxic exposure (Figs 1–4).

What clearly emerges, then, are two quite different responses to longterm exposure to low oxygen levels by different developmental stages of R. catesbeiana. The tissues of larval amphibians obviously undergo tremendous differentiation and alteration during metamorphosis. This ‘plasticity’ of tissues and even whole organs, which usually manifests itself only during the normal metamorphic process, apparently can be stimulated by environmental factors unrelated to development, such as chronic hypoxic exposure, to produce morphological adjustments in respiratory structures which facilitate gas exchange under sub-optimum conditions.

Morphological changes to lungs and skin might be expected to aid gas exchange in adult bullfrogs, as well, but are completely absent (Figs 1–4). Changes in the convec tive delivery of lung gas and blood, along with large increases in blood oxygen affinity and oxygen capacity of adult bullfrogs in chronic hypoxia (Pinder & Burggren, 1983) few of which occurs in tadpoles -may suffice to maintain oxygen uptake without morphological adjustments in the gas exchange organs. Whether the ‘cost’ of changes in the morphology of respiration structures vs changes in blood properties and physiological responses is much greater in the terminal adult stage compared to the tadpole stages, or whether the adult simply lacks regulatory mechanisms for tissue growth which are stimulated by environmental hypoxia, remains to be demonstrated.

We thank Lourdes Algarin for assistance with tissue fixation, Alan Pinder for preparation of the figures and Juan Markin for comments during preparation of the manuscript. This project was supported by NSF grant PCM80–03752 and a University of Massachusetts Faculty Research Grant.