Key adaptations enabling mammals to cope with oxygen deficiency at high elevations relate to oxygen transfer into the blood. Among others, the efficiency of this mechanism depends on haematocrit (Hct, the volumetric fraction of red blood cells in blood). Although blood of high-elevation mammals is usually characterised by normal or slightly increased Hct, there are contradictory findings from studies along different elevational gradients. The aim of this study was to explore variability of Hct at both inter- and intraspecific levels in six rodent species from lower and higher elevations of Choke Mountain in Ethiopia. We found that Stenocephalemys sp. A from higher elevation had higher Hct than its congener Stenocephalemys albipes from lower elevation and a similar but weaker tendency was observed intraspecifically in Lophuromys simensis. Furthermore, Hct among four species occupying the high-elevation Afroalpine zone was comparable, and higher than in animals from lower elevations. Higher Hct in the three Afroalpine specialists probably contributes to local adaptations for life in high elevation environments under hypobaric hypoxia.
Steep elevational gradients have been attracting scientific attention mainly because climatic conditions and habitat characteristics change rapidly as a function of elevation on a relatively small geographic scale. These factors mutually influence survival of any organism living on a gradient. Although relatively consistent and predictable changes in conditions characterise elevational gradients, ecological transitions between successive habitats can be very rapid (Storz et al. 2010). Importantly, such transitions affect less mobile taxa occurring along a gradient, such as rodents, since they have restricted dispersal capabilities. Therefore, rodents are suitable candidates for studying the impact of elevation on various biological adaptations.
In the context of evolutionary physiology, two main selective pressures related to increasing elevation are cold stress and oxygen deficiency. The latter is caused by decreasing barometric pressure, the so-called hypobaric hypoxia (e.g. Barry 2008). Therefore, vertebrates living at high elevations have to physiologically compensate for relatively low ambient temperature and its substantial diurnal and seasonal variation, as well as limited access to oxygen (e.g. Tufts et al. 2013). As a result, various adaptations maximising gas exchange and thermogenesis have been commonly found in high-elevation mammals (Withers et al. 2016).
Briefly, key adaptations enabling mammals to cope with hypobaric hypoxia relate to oxygen transfer into the blood. The efficiency of this mechanism depends mainly on the quality and amount of haemoglobin (Hb), Hb-O2 affinity, the number and size of red blood cells (RBC), and haematocrit (Hct) – the volumetric fraction of RBC in the blood (e.g. Yousef et al. 1972, Lenfant 1973, Monge & León-Velarde 1991, Withers et al. 2016). These parameters mutually influence each other, thus forming relevant trade-offs affecting the efficiency of the whole mechanism. For instance, an increase in Hct results in a linear increase in the O2 carrying capacity of the blood and also an exponential increase in blood viscosity (Birchard 1997). As a result, while maximal oxygen transport occurs at certain levels of Hct, it decreases at higher and lower Hct due to either reduced oxygen carrying capacity or increased blood viscosity (Birchard 1997). Therefore, the blood of high-elevation species is usually characterised by high Hb-O2 affinity and normal or slightly increased Hct compared to species/populations living at lower elevations (Monge & León-Velarde 1991, Tufts et al. 2013).
In general, Hct usually ranges between 30 and 50% of RBC in whole blood in mammals (Windberger & Baskurt 2007). Whereas the highest average Hct of 63% was reported for the Weddell seal Leptonychotes weddellii (Guard & Murrish 1975), the alpaca Vicugna pacos adapted to high-elevation Andean environments has Hct around 27% (Sillau at al. 1976), the lowest mean Hct recorded for any healthy mammal. Although the effect of elevation on the change in Hct is a relatively well-studied phenomenon in some mammals including humans (e.g. Mortola & Wilfong 2017), rodents (apart from laboratory rats and mice) are rather overlooked in this context. Thus, Hct has been sufficiently studied only in a few rodents from different elevations in the South-American Andes (various species – Morrison et al. 1963a, b), the Rocky Mountains in North America (the North American deer mouse Peromyscus maniculatus – e.g. Tufts et al. 2013), and the Caucasus Mountains in Eurasia (the wood mouse Apodemus sylvaticus – Kalabuchov 1937). However, the findings from different studies are not consistent; for example, increased Hct levels were found in high-elevation deer mice but not in Andean rodents. Surprisingly, there are very limited data on Hct in rodents living along elevational gradients in Africa. Only Broekman et al. (2006) explored respiratory properties in the subterranean Mahali mole-rat Cryptomys hottentotus mahali from low and high elevations and found no significant variation in Hct.
The Ethiopian Highlands, reaching up to 4550 m a.s.l. and subdivided into the eastern and the western part by the Great Rift Valley, is the largest mountainous area in Africa. Due to its topography, some of the steepest elevational gradients on the planet can be found there. These gradients are known for extremely high habitat diversity, which in turn harbours diverse (and elevation-specific) assemblages of endemic rodents (e.g. Bryja et al. 2019a). Some rodent genera likely diverged parapatrically along elevational gradients, and different species now live in various habitats from low-elevation forests to high-elevation Afroalpine grasslands, with the genus Stenocephalemys being a prime example (Bryja et al. 2018). Due to their dynamic evolutionary history and parapatric distribution of sister species at different elevations of the same mountains (see Bryja et al. 2018), they represent an excellent model for studying the processes of adaptation and speciation along elevational gradients.
The objective of this study is to explore inter- and intraspecific variability of Hct in rodents from the Ethiopian Highlands by measuring this parameter in six rodent species from different elevations around Choke Mountain and surrounding areas of Ethiopia. At both inter- and intraspecific levels, we tested whether species or populations occupying the high-elevated Afroalpine zone differ in Hct compared to those from lower elevations. At the interspecific level, we focused on the genus Stenocephalemys where one species is primarily adapted to high and the other to lower elevation. At the intraspecific level we compared populations of Lophuromys simensis occupying either lower or higher elevations.
Material and Methods
Animals and collection
For this study, a total of 75 individuals belonging to six different species (the Ethiopian white-footed mouse Stenocephalemys albipes, the narrow-headed rat Stenocephalemys sp. A (sensu Lavrenchenko & Verheyen 2006), the Simien brush-furred rat Lophuromys simensis, the Abyssinian grass rat Arvicanthis abyssinicius, the African root-rat Tachyoryctes splendens (clade 2 sensu Šumbera et al. 2018), and the Natal multimammate mouse Mastomys natalensis) were captured at three different localities in Ethiopia (Fig. 1, for details of the number of females and males captured for each species or population see Table 1). Murids were captured using Sherman live traps and root-rats with a range of hand-made traps. The higher-elevation locality was close to the peak of Choke Mountain (3645-3970 m a.s.l.; 10.7058° N, 37.8432° E). For comparison, we sampled two lower-elevation localities, near Debre Markos (2342 m a.s.l.; 10.4206° N, 37.5634° E) and at the Sakala area (2360-2391 m a.s.l.; 11.0330° N, 37.1807° E) (Fig. 1). The genus Stenocephalemys is endemic to Ethiopia and Eritrea where its representatives often comprise the majority of small mammal communities in the highlands (Bryja et al. 2018, Mizerovská et al. 2020). In short, S. albipes is the most widespread species of the genus inhabiting a wide elevational range from 800 to 3800 m a.s.l., but its occurrence at high elevations is usually related to human settlements. Stenocephalemys sp. A (see Mizerovská et al. 2020, for more details and a formal description of this taxon) is a high-elevation specialist inhabiting only the northern part of the Ethiopian highlands. Similarly, brush-furred rats of the genus Lophuromys rank among the most abundant rodents in the Ethiopian highlands. Lophuromys simensis is mostly found above 3000 m a.s.l., but records from elevations as low as 1800 m a.s.l. are known (e.g. Lavrenchenko et al. 2004, 2007, Bryja et al. 2019a). Arvicanthis abyssinicius is a typical Afroalpine species inhabiting the moorland zone usually above 3500 m a.s.l. (Bryja et al. 2019b). The genus Tachyoryctes comprises relatively large highly fossorial rodents. Mastomys natalensis is a widespread sub-Saharan rodent living usually in a low-elevation savannah-type habitats. For more details about the genetic diversity and evolutionary relationships among different rodent species of Choke Mountain and surrounding areas see Kostin et al. (2020). Each captured individual was sexed, weighed and body measurements were taken and used for other research purposes.
Haematocrit (Hct) and its coefficient of variation (CV) in six rodent species from higher-elevation sites at Choke Mountain and lower-elevation sites in the Debre Markos and Sakala areas (nf and nm denotes to the number of females and males measured, respectively; for Stenocephalemys sp. A one individual was not sexed, numbers in parentheses denote range).
From each individual, approximately 130 µL of blood was drawn from the retro-orbital sinus with a heparinised capillary for measurement of Hct (inner diameter 1.5 mm, height 75 mm) immediately after sacrificing by cervical dislocation for other research purposes. Each capillary was sealed and centrifuged using a powered centrifuge at 5000 rpm for 10 min. The blood separated into two distinct phases. The upper transparent phase contained plasma and the dark red lower phase contained packed RBC. The lengths of both phases were measured with a Vernier calliper and served as a measure of their volume; total blood volume was calculated as the sum of the two phases. Haematocrit was calculated as the proportion of the volume of packed RBC to the total blood sample volume. All fieldwork complied with legal regulations in Ethiopia, and sampling was carried out with the permission of the Ethiopian Wildlife Conservation Authority and the Department of Biology of the Debre Markos University and the Environment, Forest and Wildlife Protection and Development Authority of the Amhara National Regional State.
DNA was extracted using commercial kits, and all samples were identified to species by DNAbarcoding using the mitochondrial gene for cytochrome b (cytb). The protocol used for PCR and Sanger sequencing is described in Bryja et al. (2014). Obtained sequences were aligned with available datasets from recent cytb-based studies (e.g. Bryja et al. 2018, 2019b, Šumbera et al. 2018). A single specimen of Mastomys was identified by the species-specific PCR-typing of cytb using the protocol from Martynov & Lavrenchenko (2018).
For each species/population and elevation, we calculated the mean, standard deviation, and coefficient of variation of Hct. Since all Hct values were expressed as proportions, they were arcsine transformed prior to analyses. There was only one individual M. natalensis captured at lower elevation (2342 m a.s.l.) and one individual S. albipes at higher elevation (3965 m a.s.l.). These two individuals were excluded from all analyses unless specified otherwise. First, we computed a linear model (LM) to test whether Hct differed between males and females for each of the tested species; in L. simensisthe populations from lower and higher elevations were treated as separate groups. Second, we used LM to test the effect of species on Hct. Thereafter, we used two-tailed t-tests to test a hypothesis of different Hct in species/populations originating from higher and lower elevation. Specifically, we compared two species from the genus Stenocephalemys (S. sp. A vs. S. albipes) and two populations of L. simensis. To assess whether the Hct value of a single S. albipes from higher elevation differed from Hct characterising the lower-elevation population of this species, the difference was considered statistically significant if the value fell outside the 0.025 and 0.975 quantiles of a normal distribution approximating arcsine-transformed Hct values of the lower-elevation population. All tests were performed using the R statistical software (R Core Team 2019). Throughout the text, means are given ± SD and with ranges in parentheses.
The highest Hct values of 70 and 71% were found in four individuals, two L. simensis, one S. albipes and one S. sp. A, all from higher elevations (3779-3965 m a.s.l.) (Table 1). The lowest Hct of 46% was detected in one individual of A. abyssinicius from elevation of 3965 m a.s.l. There was no difference in Hct between males and females in either of the studied species (LMs: A. abyssinicus: F = 0.4, p = 0.518; S. sp. A: F = 0.3, p = 0.729; S. albipes: F = 0.5, p = 0.504; L. simensis higher elevation: F = 2.8, p = 0.114; L. simensis lower elevation: F = 1.6, p = 0.265; T. splendens: F = 0.5, p = 0.491).
The variability of Hct tended to be higher for species (or population in the case of L. simensis) from higher elevation (Table 1). The highest Hct variability was found in A. abyssinicius, whereas L. simensis population from lower elevation showed the lowest variability (Table 1). Although being not significant, Hct tended to differ among the all studied species (LM: F4, 67 = 2.4, P = 0.06). The four species occupying the Afroalpine habitats (only the population from the higher elevation was included for L. simensis) of Choke Mountain had similar Hct (LM: F3, 53 = 1.0, P = 0.395). When we compared Hct in two species of the genus Stenocephalemys, S. sp. A living at higher elevation had significantly higher Hct than S. albipes living at lower elevation (Fig. 2; t-test: t = –3.7, df = 19, P = 0.002). The Hct value of the only individual of S. albipes from the higher elevation was above the 97.5% quantile of Hct values characterising the lower-elevation population of S. albipes. In L. simensis, the animals from higher elevation tended to have higher Hct than the animals from lower elevation (Fig. 2; t-test: t = 1.9, df = 23, P = 0.077).
Haematocrit (Hct) values for rodents occurring at different elevations from different continents. Hct values are given as mean ± SD (Hct of the fossorial African root-rat is presented in Table 3 with other fossorial species).
Compared to previously published data, the Hct values recorded in several Ethiopian rodent species in the present study are among the highest recorded for free-living rodents (Table 2) and for mammals generally (Windberger & Baskurt 2007). We found that species occupying Afroalpine habitats close to the peak of Choke Mountain had similar values of Hct (58-60%) which tended to be slightly higher compared to species or populations from lower elevations (S. albipes 54% and the lowland population of L. simensis 57%). Similar findings were obtained in a study on the North American deer mouse, Peromyscus maniculatus, which showed an increase of 11 percentage points in Hct in mice from Evans Mountain (4350 m a.s.l., Hct = 60%) compared to mice from a low-elevation locality at Nine Mile Prairie, Nebraska (430 m a.s.l., Hct = 49%) (Tufts et al. 2013). The near threefold difference in Hct between our study and that of Tufts et al. may be explained by the much greater elevational difference between low- and high-elevation sites in the latter (3965 vs. 2342 m a.s.l.). A gradual increase in Hct with increasing elevation was documented in another study on deer mice (Hammond et al. 1999). However, there may also be other factors that influence Hct levels. For example, in studies conducted on highland and lowland rodents from Peru and Chile the considerable variability in their Hct was not explained by elevation (Hct 47% for highland 3900-4700 m a.s.l. and 46% for lowland 0-650 m a.s.l. species), but less agile species from both elevations (43 and 41% for species from lower and higher elevation, respectively) had lower Hct than their more agile counterparts (51 and 52% for species from lower and higher elevation, respectively) (Morrison et al. 1963b). Although there are only limited behavioural data on most of the African species included in our study, it seems that for example S. albipes adapted to semi-arboreal life in various forest types is more agile than its congeners adapted to more terrestrial life in the grassy habitats of the Afroalpine zone (our unpublished data based on open field test observations). Nevertheless, the fact that S. sp. A and other species from higher elevation tended to have higher Hct than S. albipes from lower elevation implies that the differences in Hct are more likely an effect of elevation than agility.
Haematocrit (Hct) for fossorial rodents from various elevations and different continents (Hct values are given as mean ± SD).
Stenocephalemys congeners from low and high elevation, together with a single individual of S. albipes preferring generally lower elevations but captured in the natural Afroalpine habitat at high elevation give us an opportunity to explore the effect of elevation on Hct in more detail. Our findings showed that highland S. sp. A has Hct levels six percentage points higher than its sister species S. albipes from lower elevations. In this regard, it is important to mention that many high-elevation specialists have normal or only slightly increased Hct levels, which could be a result of the positive relationship between Hct and blood viscosity (e.g. Birchard 1997, Barve et al. 2016). This is a typical example of the cascade effect when high Hct increases blood viscosity that subsequently reduces oxygen delivery to the tissues. Nevertheless, it might still be possible that the six percentage-point increase in Hct in S. sp. A has an adaptive value for life in the Afroalpine environment under hypobaric hypoxia. Yet other adaptations may be helping to counter the hypobaric hypoxia. Among these, increased Hb concentration, higher Hb-O2 affinity and general adaptations related to hypoxia tolerance (see e.g. Barve et al. 2016) should be considered in future studies in rodents occurring along elevational gradients in the Ethiopian Highlands.
It is a well-known phenomenon that changes in Hct can be environmentally induced. For instance, Tufts et al. (2013) showed that Hct in deer mice native to high elevation decreased from 60 to 50% after six weeks acclimation in normoxia (i.e. 300 m a.s.l.), which was similar to the values of low-elevation natives (50% Hct). This indicates that Hct differences in deer mice are largely attributable to physiological plasticity rather than being genetically fixed (Tufts et al. 2013). A tendency for slightly lower Hct in L. simensis from lower elevation might represent such a decrease caused by physiological plasticity in this primarily highland species. An extremely high Hct value of one S. albipes (70%; 10 percentage points higher than the mean Hct of high-elevation adapted S. sp. A and 16 percentage points higher than its conspecifics from lower elevation) captured at almost 4000 m a.s.l. may represent another case of environmentally induced change in Hct, so-called acclimatization. During this process, individuals of species living at lower elevations usually respond to hypobaric hypoxia by increased Hct (e.g. Ou et al. 1985, Lui et al. 2015). Available data show that Hct increase during acclimatization varied greatly among species. For instance, whereas it was up to 11 percentage points in humans (5600 m a.s.l., Richalet et al. 1994), a 58% increase (from 47.9 to 75.6%) was documented in laboratory rats (acclimated for 28 days to elevation of 6580 m a.s.l., Shertzer & Cascarano 1972), 10 percentage points in deer mice (acclimated for one month to elevation of up to 3801 m a.s.l., Hammond et al. 1999), and 16 percentage points in S. albipes (this study). Because of its nature, a typical acclimatization response is temporary and Hct drops back after descending to lower elevation (e.g. Zubieta-Calleja et al. 2007).
The difference in Hct found between the two Stenocephalemys species might also indicate existence of genetically based differences in this haematological trait. If the differences in Hct have a genetic background, this might be one of the traits contributing to the ecological speciation suggested in Stenocephalemys (Bryja et al. 2018) and the ability of the genus to colonize new and extreme habitats, such as Afroalpine grasslands (see also possible adaptive mutations in one of the oxidative phosphorylation genes (cytb) in this genus in Kostin & Lavrenchenko (2018). Whether the differences in Hct between S. sp. A and S. albipes are genetically or environmentally based could be clarified by reciprocal-transplant experiments, i.e. those conducted at both low- and high-elevation study sites.
A specific case among rodents is fossorial species as conditions in their burrows are usually hypoxic (Buffenstein 2000). Thus, fossorial rodents from lower elevations might be adapted to equivalent oxygen conditions to those experienced by their non-fossorial counterparts from higher elevations. However, the fossorial rodents studied here do not seem to show any predictable differences in Hct when compared with other rodents (Tables 2 and 3), which corroborates the recent findings of Weber et al. (2017) who did not find any consistent patterns in Hct or intrinsic Hb-O2 affinity among African mole-rats. Further, when the Hct of lowland and highland populations of the same species are compared, neither pocket gophers Thomomys bottae nor common mole-rats C. h. mahali show any significant differences (Table 3). It should however be mentioned that the fossorial rodent from the highest elevation analysed, T. splendens from Choke Mountain, had the highest Hct of any fossorial rodent measured (Table 3). The effect of elevation on Hct in fossorial rodents is thus unclear and future studies should aim to clarify this issue. Certainly, other mechanisms than Hct may be involved. For instance, highland common mole-rats showed increased Hb concentration compared with lowland animals (Broekman et al. 2006) suggesting that high-elevation hypoxic adaptations might apply to other haematological traits. Likewise, there was a convergence in mechanisms regulating Hb-O2 affinity between African mole-rats and epigeic (active on the surface) rodents adapted for high elevations (Weber et al. 2017).
Conclusion and future perspectives
The results of the present study provide evidence that the species occupying the high-elevation Afroalpine zone have comparable Hct that is slightly higher than the Hct of animals from lower elevations. The same applies to two congeners of Stenocephalemys, as S. sp. A living at higher elevation had higher Hct than S. albipes from lower elevation. Future research should look at other of the diverse rodent communities along elevational gradients in Africa to establish whether slightly higher Hct might be a general phenomenon promoting the survival of Afroalpine specialists under environmental conditions of lower oxygen availability. It would also be interesting to study the genetic background of the evolution of Hct and Hb-O2 affinity, the latter trait being universally considered as a key characteristic of hypoxia-tolerant mammals (Withers et al. 2016), using reciprocal transplant experiments. Narrow-headed rats of the genus Stenocephalemys would be ideal candidates for this purpose, because parapatric ecological speciation along the elevational gradient has been proposed for them (Bryja et al. 2018).
We are grateful to Dr. A. Darkov (Joint Ethio-Russian Biological Expedition, Fourth Phase – JERBE IV) and Dr. S. Keskes (Ethiopian Ministry of Innovation and Technology) for management of the expedition in the field and in Addis Ababa. For help during the field work we acknowledge D. Yu. Alexandrov, M. Yihune and M. Kasso. We also thank two anonymous reviewers for their valuable comments on an earlier version of the manuscript. This study was supported by the Russian Foundation for Basic Research (project no. 18-04-00563-a) and the Czech Science Foundation (project no. 18-17398S). Author contributions: J. Bryja, R. Šumbera, L. Lavrenchenko, M. Lövy, J. Okrouhlík conceived and designed the study; L. Lavrenchenko, D. Kostin, A. Martynov collected the data; A. Martynov and J. Bryja analyzed the genetic data; M. Lövy performed statistical analyses and wrote the first version of the manuscript; all authors contributed, read and approved the final version of the manuscript.