Control of breathing and respiratory gas exchange in high-altitude ducks native to the Andes

The air at high altitude is both cold and hypoxic. These conditions challenge the ability of birds and mammals to adequately match O₂ supply and O₂ demand, because cold increases the demand for O₂ while hypoxia restricts O₂ availability (Monge and Leon-Velarde, 1991; Storz et al., 2010). The first step in obtaining O₂ is pulmonary ventilation, and thus the hypoxic ventilatory response (HVR) is critical for O₂ uptake in the thin air at high altitude (Birchard and Tenney, 1986; Brutsaert, 2007).

Ventilation is modulated by changes in blood gas levels and metabolism. Acute exposure to hypoxia leads to a drop in arterial partial pressure of O₂ (P_aO2), which stimulates an increase in ventilation (the HVR) that helps offset the fall in P_aO2 (Powell et al., 1998). This reflex is initiated primarily by the carotid bodies, peripheral chemoreceptors that are sensitive to changes in arterial P_O2 and P_CO2/pH located in the carotid arteries supplying the brain in mammals and birds (Gonzalez et al., 1994). Prolonged exposure (days to weeks) to hypoxia leads to further increases in breathing by increasing the ventilatory sensitivity to hypoxia through the process of ventilatory acclimatization to hypoxia (VAH) (Powell et al., 1998). Ventilation is also modulated by changes in metabolism, which helps match O₂ supply to tissue O₂ demand during exercise, thermogenesis, metabolic depression or changes in body temperature (Barros et al., 2006; Chappell, 1992; Eldridge, 1994).

Birds and mammals that live at high altitude have been shown to differ from low-altitude taxa in their ventilatory responses to hypoxia. Some species/populations native to high altitude in the Himalayas and on the Tibetan Plateau, such as bar-headed geese (Anser indicus), plateau pika (Ochotona curzoniae) and Tibetan people, breathe more and exhibit HVRs that are equivalent or greater in magnitude than species/populations native to low altitude (Beall et al., 1997; Brutsaert, 2007; Lague et al., 2016; Moore, 2000; Pichon et al., 2009; Scott and Milsom, 2007). In contrast, in some high-altitude residents in the Andes, such as Andean goose (Chloephaga melanoptera), guinea pigs (Cavia porcellus) and Andean people, breathing and the ventilatory response to hypoxia are reduced compared with their low-altitude counterparts (Beall, 2000; Brutsaert et al., 2005; Ivy et al., 2018; Lague et al., 2017; Schwenke et al., 2007). It has been difficult to determine whether these patterns of variation result from evolved differences or from environmentally induced plasticity (acclimatization, developmental plasticity, etc.) (Brutsaert, 2016; Laguë, 2017; Moore, 2017). Nevertheless, these intriguing results suggest that there may be convergent mechanisms for coping with hypoxia at high altitude within a given geographic region, but divergent mechanisms between species inhabiting different geographic regions. These findings, however, arise from a small number of very different species, native to different habitats, with different lifestyles, activity levels and respiratory physiologies. It thus remains unclear whether convergent and/or divergent responses will also be observed in similar closely related species that independently colonized high altitude in one geographic region.

The objective of this study was to investigate whether the HVR has been altered across multiple duck species from the high Andes of Peru, and if so, whether there are similar or distinct changes in each high-altitude species. Waterfowl (Order Anseriformes) native to the Andes are a powerful taxonomic group for examining the general patterns of variation across high-altitude taxa, because many species have independently colonized similar aquatic habitats at high altitude (McCracken et al., 2009a; Natarajan et al., 2015). Previous studies of haemoglobin evolution in Andean waterfowl showed that genetically based increases in haemoglobin (Hb)–O₂ binding affinity have arisen in most high-altitude waterfowl examined to date (McCracken et al., 2009a; Natarajan et al., 2015), but other aspects of respiratory physiology have not been comprehensively examined in this group. Here, we examined the HVR of six species of ducks in their native high-altitude environment in the Andes, five of which were compared with a closely related population of the same species or with a sister species in their native environment at low altitude (Fig. 1).

Ducks were captured and studied in July and August of 2014 and 2015. Five species were captured and tested at high altitude (3812 m above sea level) at the Lake Titicaca National Reserve (Puno, Peru) in August 2014: speckled teal (Anas flavirostris oxyptera Vieillot 1816; n=12, 4 males and 8 females), Andean ruddy duck (Oxyura jamaicensis ferruginea; n=12, 5 males and 7 females), yellow-billed pintail (Anas georgica Gmelin 1789; n=13, 10 males and 3 females), Andean cinnamon teal (Anas cyanoptera orinomus Vieillot 1816; n=12, 8 males and 4 females) and puna teal [Anas puna (Tschudi 1844); n=12, 7 males and 5 females; body mass of 404±11 g]. Four species, representing closely related populations of the same species or sister species of four of these high-altitude taxa, were captured at low altitude in Oregon, USA (at either Summer Lake Wildlife Management Area at 1260 m or Malheur National Wildlife Refuge at 1256 m) in July 2015, and were tested at Summer Lake: green-winged teal (Anas crecca Linnaeus 1758; n=10, 5 males and 5 females), ruddy duck [Oxyura jamaicensis jamaicensis (Gmelin 1789); n=8, 4 males and 4 females], northern pintail (Anas acuta Linnaeus 1758; n=10, 7 males and 3 females) and northern cinnamon teal (Anas cyanoptera septentrionalium Vieillot 1816; n=11, 6 males and 5 females). Torrent ducks (Merganetta armata Gould 1841) were also captured and tested in August 2015, both at high altitude (3000–4086 m above sea level; n=8, all males) on the Chancay River Valley near Vichaycocha, Lima, Perú, and at low altitudes (1092–1665 m above sea level; n=14, all males) on the Chillón River in Santa Rosa de Quives, Lima, Perú. Ducks were allowed to recover overnight from capture for at least 6–12 h, with unlimited access to water, before responses to acute hypoxia were measured. During this time, birds were held in large animal kennels with dry bedding. All experiments were performed within 2 days of capture, and birds were tube fed commercial duck chow if held for longer than 1 day in captivity, but food was always withheld for 6–12 h before measurements took place. Ducks were collected in accordance with permits issued by the Ministerio del Ambiente del Peru (004-2014-SERNANP-DGANP-RNT/J), the Ministerio de Agricultura del Peru (RD 169-2014-MIN AGRI-DGFFS/DGEFFS and 190-2015-SERFOR/ DGGSPFFS), the US Fish and Wildlife Service Region 1 Migratory Bird Permit Office MB68890B-0 (MB68890B-0) and the Oregon Department of Fish and Wildlife (Scientific Taking Permit 101-15). All experimental procedures followed guidelines established by the Canadian Council on Animal Care, and were approved by institutional animal care committees.

We measured the respiratory and metabolic responses to acute hypoxia using plethysmography and respirometry techniques similar to those used previously for Andean goose (Ivy et al., 2018; Lague et al., 2017). Ducks were held in a cradle that permitted unrestricted breathing, with their head in a 4 l opaque chamber that was sealed around the neck with a latex collar. Ducks were given 60–90 min to adjust to the apparatus (when they exhibited a noticeably relaxed and stable breathing pattern) before measurements began, with ambient air supplied to the head chamber at a flow rate of 5 l min⁻¹ (volume in standard temperature and pressure of dry air, STPD). Measurements of breathing and metabolism were then recorded at several inspired O₂ tensions (P_O2), starting in ambient conditions (18 and 13 kPa at low and high altitude, respectively) and then during step-wise decreases in P_O2, for 25 min at each of the following steps: 18 (low-altitude ducks only), 13, 12, 9, 7 and 6 kPa. Dry incurrent air and nitrogen were mixed using pre-calibrated rotameters (Matheson Model 7400 Gas Mixer, E700 and E500 flowtubes, Oakville, ON, Canada) to achieve each level of hypoxia.

Metabolism, breathing, arterial O₂ saturation and body temperature were measured continuously during the above exposures, and we report the average values across the last 10 min at each inspired P_O2. The excurrent air leaving the head chamber was subsampled at 200 ml min⁻¹, dried with silica gel (MLA6024, ADInstruments, Colorado Springs, CO, USA), and passed through CO₂ and O₂ analyzers (FOXBOX, Sable Systems, Las Vegas, NV, USA). These data, together with the flow of air through the head chamber, were used to calculate rates of O₂ consumption (V̇_O2), as described by Lighton (2008), which we express here in volume units at STPD. Tidal volume (V_T) and breathing frequency (f_R) were determined from the flow oscillations of the biased outflow from the head chamber, measured using a pneumotachograph (8311A series, Hans Rudolph Inc., Shawnee, KS, USA) and differential pressure transducer (Validyne DP45, Cancoppas, Mississauga, ON, Canada) zeroed to baseline flow through the chamber. Body temperature (T_b) was measured continuously using a rectal probe (RET-1, Physitemp, NJ, USA). All of the above data were acquired using a PowerLab 16/32 and LabChart 8 Pro software (ADInstruments). Total ventilation (V̇_E) was determined as the product of f_R and V_T. Both V_T and V̇_E are reported in volumes expressed at body temperature and pressure of water-saturated air (BTPS), which best reflects the air volumes moved by the animal. Air convection requirement (ACR) was calculated as the quotient of V̇_E and V̇_O2. Pulmonary O₂ extraction (%) was calculated as V̇_O2 divided by the product of V̇_E and the concentration of O₂ in inspired air (i.e. ml O₂ in STPD per ml air in BTPS). Arterial O₂ saturation (S_aO2) was measured using the MouseOx Plus pulse oximetry system and software (Starr Life Sciences, PA, USA) with neck collar sensors, which was enabled by plucking a small number of feathers from around the neck. We have previously demonstrated that the MouseOx Plus system is able to provide accurate measurements of S_aO2 in waterfowl (Ivy et al., 2018). Arterial O₂ saturation was measured in all low-altitude ducks and all torrent ducks, but in only a subset of high-altitude ducks (speckled teal: n=6; Andean ruddy duck: n=5; yellow-billed pintail: n=6; Andean cinnamon teal: n=5).

Two-factor ANOVA was generally used for most comparisons to examine the main effects and interactions of population/species altitude and acute inspired P_O2 (repeated measures) within each independent pair of closely related high-altitude and low-altitude populations/species, and we used Holm–Šidák post hoc tests to test for pairwise differences between populations/species within each inspired P_O2. However, for puna teal, a species for which we do not have data for a close low-altitude relative, we used one-factor ANOVA to examine the main effect of inspired P_O2. For body mass data, we used two-factor ANOVA and Holm–Šidák post hoc tests to test for the main effect of altitude and the pairwise differences between high- and low-altitude pairs. Values are reported as means±s.e.m. All statistical analysis was conducted with SigmaStat software (v. 3.5) with a significance level of P<0.05.

Breathing and the HVR were elevated in some high-altitude ducks compared with their close relatives from low altitude (Fig. 2, Table 1). All ducks increased V̇_E by up to ∼36–174% in response to acute hypoxia challenge (Fig. 2A), as reflected by significant main effects of inspired P_O2 for all high–low pairs in two-factor ANOVAs (Table 1). Two high-altitude taxa, torrent duck and yellow-billed pintail, exhibited higher V̇_E than their low-altitude counterparts across a range of inspired P_O2 (Fig. 2A), and there was a population main effect or population×P_O2 interaction for these species (Table 1). The remaining high-altitude taxa – Andean ruddy duck, Andean cinnamon teal and speckled teal – had similar or slightly reduced V̇_E when compared with closely related low-altitude taxa (Fig. 2A, Table 1). Increases in f_R (∼17–170%) were the main contributor to increases in V̇_E in response to reductions in inspired P_O2 in nearly all species (Fig. 2B), with only modest changes in V_T (∼4–50%) in some species (Fig. 2C, Table 1). The exception to this pattern was the puna teal, the only species for which we do not have data for a close lowland relative; this species increased V̇_E in acute hypoxia primarily by increasing V_T (Table 2). However, f_R was similar or lower in high-altitude ducks than in low-altitude ducks for all high–low comparisons (Fig. 2B, Table 1). High-altitude torrent ducks and yellow-billed pintails breathed with deeper V_T than their lowland counterparts across a range of inspired P_O2, which was the dominant contributor to the increases in V̇_E in these high-altitude taxa (Fig. 2C). However, this was not the case in high-altitude Andean ruddy duck, Andean cinnamon teal or speckled teal, which had V_T values similar to those of their low-altitude counterparts (Fig. 2C, Table 1).

Metabolic rates (as reflected by V̇_O2) were stable or increased under hypoxic conditions across species (Fig. 3A, Table 1). V̇_O2 was largely similar between high and low altitude, with the exception that high-altitude yellow-billed pintail had higher V̇_O2 than closely related low-altitude northern pintail. This was reflected by a main effect of population in two-factor ANOVA (Table 1), and the differences were particularly evident as hypoxia became more severe (Fig. 3A). Increases in both V̇_E and pulmonary O₂ extraction with deepening hypoxia likely helped avoid any falls in V̇_O2 across species. This was reflected by main effects of inspired P_O2 on ACR (Fig. 3B) and pulmonary O₂ extraction (Fig. 3C) for nearly all high–low pairs (Table 1). The increases in ACR during hypoxia were similar in high-altitude Andean ruddy duck, Andean cinnamon teal and speckled teal when compared with their close low-altitude relatives, and there were no main effects of population on this variable. However, ACR was lower in the deepest levels of hypoxia (but not at the intermediate levels) in the other two high-altitude taxa, torrent duck and yellow-billed pintail, when compared with their close lowland relatives, and there were significant population×P_O2 interactions for these high–low comparisons (Table 1). High-altitude yellow-billed pintail appeared to counterbalance this decline in ACR in the most severe levels of hypoxia with an increase in pulmonary O₂ extraction compared with low-altitude northern pintails (Fig. 3C, Table 1).

T_b declined by up to ∼1–2°C in response to acute hypoxia in most duck species (Fig. 4), as previously observed in many other birds (Kilgore et al., 2008; Novoa et al., 1991; Scott et al., 2008). The exception to this pattern were the two diving ducks from low altitude – torrent duck and ruddy duck – but these low-altitude populations differed from their conspecific high-altitude populations, which did exhibit a reduction in T_b in hypoxia (Fig. 4, Table 1). The variation in T_b did not appear to be clearly associated with any comparable variation in V̇_O2 (Fig. 3A).

There were surprisingly few differences in S_aO2 in hypoxia between high-altitude and low-altitude ducks. As expected, S_aO2 decreased progressively with increasing severity of acute hypoxia, and the main effect of inspired P_O2 was seen across species (Fig. 5, Table 1). However, yellow-billed pintail was the only high-altitude duck that had a higher S_aO2 than their low-altitude counterpart, which was reflected by higher saturations than low-altitude northern pintails at levels of moderate hypoxia that are environmentally realistic in the high Andes (Fig. 5, Table 1). S_aO2 was similar between high- and low-altitude taxa for all other high–low comparisons. These data suggest that the differences in S_aO2 between the high-altitude yellow-billed pintails and low-altitude northern pintails may be a product of differences in both breathing and Hb–O₂ affinity, as will be examined in more detail in the Discussion.

Body mass (M_b) often differed between high- and low-altitude taxa (Fig. 6). There was variation between species pairs, with teals tending to be smaller than ruddy ducks and pintails. M_b was greater in high-altitude speckled teal (30%), Andean cinnamon teal (43%) and Andean ruddy duck (62%) compared with their close relatives from low altitude (Fig. 6A). In contrast, yellow-billed pintail was 27% smaller than northern pintail, and torrent duck was of similar M_b between high and low altitudes (Fig. 6A). The variation in M_b across pairs accounted for some of the variation in metabolic rate, particularly when excluding torrent ducks, in which case mass-specific V̇_O2 (measured at a common P_O2 of 13 kPa) was related to M_b with a scaling exponent of −0.33 (i.e. V̇_O2∝M_b^−0.33) (Fig. 6B). However, there was no significant relationship between M_b and ACR (data not shown), suggesting that M_b had no effect on breathing that was independent of its effect on metabolic rate.

Previous work has shown that breathing and the HVR of high-altitude natives differ from that of their low-altitude counterparts in distinct ways (see Introduction). Here, we show that breathing and pulmonary gas exchange in different lineages of high-altitude ducks can differ in distinct ways from their close relatives from low altitude, even among similar species inhabiting similar high-altitude habitats. Based on these and previous findings, high-altitude ducks differ from those from low altitude in at least two possible ways: (i) increases in V̇_E and the HVR and/or (ii) an increased Hb–O₂ binding affinity (Natarajan et al., 2015). However, only when both occurred together, as in the yellow-billed pintail, did these differences lead to any net benefit in increasing S_aO2.

All ducks, from both low and high altitude, responded to acute hypoxia with robust increases in V̇_E and pulmonary O₂ extraction. The magnitude of these responses appears to be similar to that found in previous studies of other waterfowl species, being somewhat intermediate between bar-headed geese (which exhibit pronounced increases in V̇_E but only modest increases in extraction in hypoxia) and Andean geese (which exhibit very modest increases in V̇_E but large increases in pulmonary O₂ extraction in hypoxia) (Scott and Milsom, 2007; Lague et al., 2017). Increases in f_R contributed the majority to increases in V̇_E in hypoxia, as is typical of many species of bird and mammal, with smaller to negligible increases in V_T (Fig. 2) (Scott and Milsom, 2007; Lague et al., 2017). In contrast, puna teal increased V̇_E primarily by increasing tidal volume in hypoxia (Table 2). This hypoxia response of puna teal may be particularly beneficial for gas exchange, because increases in V_T are more effective at increasing parabronchial ventilation than are increases in f_R (the former reduces the ratio of dead space gas in the air-ventilating parabronchioles). Puna teal is endemic to high altitude and has likely been established in the high Andes for over one million years (McCracken et al., 2009b), and further examination of high-altitude adaptations in this species may be a fruitful direction for future research.

In comparison to close relatives from low altitude, the increase in V̇_E was elevated in only two duck taxa native to high altitude: torrent ducks and yellow-billed pintails (Fig. 2). One possible explanation for the differences in these two high-altitude groups is that chronic exposure to hypobaric hypoxia led to plasticity in the underlying neural networks controlling breathing, increasing V̇_E and enhancing the HVR, as has been observed in some (Black and Tenney, 1980; Lague et al., 2016) but not all (Powell et al., 2004) previous studies. A second possible explanation is that increases in V̇_E and the HVR arose from evolutionary adaptation to high altitude, as appears to have occurred in the bar-headed goose (Scott and Milsom, 2007). For this latter possibility to occur, the effects of selection would have to be very strong, because gene flow still occurs between high- and low-altitude populations of most of the species we examined (McCracken et al., 2009a,c; Natarajan et al., 2015).

The increases in V̇_E and the HVR in the yellow-billed pintail (but not torrent duck) from high altitude could also be a consequence of elevated metabolic rate (Fig. 3). Yellow-billed pintail are ∼25% smaller than northern pintail (Fig. 6A), so the allometric effects of body mass on resting metabolic rate could explain some of this difference between species (Fig. 6B) (White et al., 2006), but the more substantial ∼2-fold differences in metabolic rate at lower P_O2 cannot be explained by these more subtle effects of allometry. Nevertheless, the observation that yellow-billed pintail from high altitude had higher V̇_E (Fig. 2) but not higher ACR (Fig. 3) than their relatives from low altitude supports the suggestion that differences in metabolic rate could drive the differences in total ventilation. In fact, ACR was slightly less in this high-altitude taxon during severe hypoxia. Under these conditions, the higher metabolic rates in high-altitude pintails are matched by increases in pulmonary O₂-extraction in severe hypoxia as compared with low-altitude pintails (Fig. 3). This could reflect a high O₂-diffusing capacity in the lungs of high-altitude pintails, as previously observed for Andean geese (Maina et al., 2017).

However, it is also possible that the increases in metabolic rate in yellow-billed pintail from high altitude arise as a consequence of the increases in V̇_E. Metabolic savings have been suggested as one possible advantage of blunting V̇_E and the HVR in some other high-altitude taxa, so long as O₂ demands can still be met (Powell, 2007). By a similar rationale, it is possible that increases in V̇_O2 owing to respiratory muscle activity could consume any added O₂ taken up into the blood as a result of breathing more in hypoxia. However, we have previously shown that the metabolic cost of breathing is quite low (∼1–4% of resting metabolic rate) in all of the species studied here (York et al., 2017). We also found that the dynamic compliance of the respiratory system is greater in high-altitude yellow-billed pintail and in some other high-altitude ducks than in their low-altitude counterparts, such that many high-altitude ducks have higher breathing efficacy and can move a greater volume of air for a given power output of breathing (York et al., 2017).

Three high-altitude taxa – ruddy duck, Andean cinnamon teal and speckled teal – in contrast, had similar levels of V̇_E as their low-altitude relatives (Fig. 2). In all three comparisons between high and low-altitude pairs, increases in ACR were similar during progressive hypoxia. Although this lack of variation could reflect a lack of any plasticity or evolved changes in these high-altitude taxa, it could also reflect concurrent but opposing effects of plasticity and evolution on the control of breathing. The latter scenario could reflect counter-gradient variation, a term that describes situations in which the effects of phenotypic plasticity on a trait are opposed by local adaptation, thus minimizing phenotypic change along an environmental gradient (Conover and Schultz, 1995). Although increases in breathing improve O₂ uptake, they also result in excessive rates of CO₂ excretion and/or respiratory water loss (Powell, 2007). In this case, there could be opposing selective forces at high altitude for enhanced O₂ delivery versus preservation of acid–base homeostasis and/or water balance. This may also explain why some other species of high-altitude waterfowl, such as the Andean goose and crested duck, have evolved a blunted HVR (Ivy et al., 2018; Lague et al., 2017).

Some of the high-altitude duck species studied here have evolved an increased Hb–O₂ binding affinity compared with their close relatives from low altitude (McCracken et al., 2009a; Natarajan et al., 2015). Birds co-express two major Hb isoforms – HbA and HbD – that express different α-chain subunits, encoded by two distinct α-globin genes (Storz, 2016a). HbA is the major isoform, comprising ∼70–80% of all blood Hb, and this isoform has a higher affinity (lower P₅₀, the P_O2 at 50% saturation) in high-altitude yellow-billed pintail, Andean cinnamon teal, speckled teal and puna teal when measured in the presence of physiologically relevant concentrations of allosteric modifiers (Natarajan et al., 2015). In contrast, Hb–O₂ affinity is very similar in high-altitude torrent duck and Andean ruddy duck as compared with their low-altitude relatives. However, the Hb–O₂ affinity of the low-altitude populations of each of these two diving species is already characteristic of the high-altitude populations, not low-altitude populations, of non-diving taxa (Natarajan et al., 2015). It is possible that selection for traits that improve breath holding underwater increased Hb–O₂ affinity in these diving duck species, as observed in some other diving birds (Meir and Ponganis, 2009), such that they may have been ‘pre-adapted’ to life at high-altitude. Then upon colonizing high altitude, these diving species may have since been constrained in increasing Hb–O₂ affinity any further, or there may have been weaker selective pressure for doing so. Nevertheless, in general, there is very strong evidence for convergent evolution of increased Hb–O₂ affinity in high-altitude taxa, making a strong case for the adaptive value of this trait for life in hypoxic environments (Storz, 2016b). However, the physiological implications of this evolved trait for S_aO2 and respiratory gas exchange have seldom been examined in high-altitude natives.

Our current findings suggest that the relative advantage of increasing Hb–O₂ affinity for improving arterial O₂ saturation is contingent upon the relative levels of ventilation. Of the three species of high-altitude ducks that have evolved an increased Hb–O₂ affinity that we examined (yellow-billed pintail, Andean cinnamon teal and speckled teal), only one – the yellow-billed pintail – maintained higher S_aO2 in hypoxia than its close relative from low altitude, and only in moderate hypoxia (Fig. 5). The yellow-billed pintail was also the only one of this group of high-altitude ducks in which V̇_E was elevated compared with its low-altitude counterpart during hypoxia. The other two high-altitude duck species that have evolved higher Hb–O₂ affinities – Andean cinnamon teal and speckled teal – had similar or lower V̇_E as compared with their close low-altitude relatives, and S_aO2 in hypoxia was not enhanced. In high-altitude torrent duck, where V̇_E was higher but Hb–O₂ affinity was similar to that of its low-altitude relative, S_aO2 in hypoxia was also not enhanced. These findings suggest that evolved increases in Hb–O₂ affinity may only be expected to improve S_aO2 in high-altitude taxa that also exhibit higher V̇_E than their low-altitude relatives. Therefore, interactions between multiple respiratory traits in the O₂ transport cascade affect the integrated systems-level function of animals native to high altitude.