Two human hemoglobin (Hb) variants, Hb C and Hb S, are known to protect against Plasmodium falciparum malaria and have evolved repeatedly in malaria endemic areas. Both aggregate to insoluble crystals (Hb C) or polymers (Hb S) under certain physiological conditions, impair parasite growth, and may facilitate retention of infected red blood cells (RBCs) in the spleen. Given the profound effects of parasites on host evolution in general, and that RBC Hb concentration is often close to its solubility limit throughout vertebrates, similar mechanisms may operate in nonhuman vertebrates. Here we show exercise-induced, profound in vivo Hb polymerization in RBCs of the Gulf toadfish. Hb aggregation was closely associated with the extent of plasma acidosis, fully reversible, and without any signs of hemolysis or anemia. Our literature analysis suggests that aggregation prone Hbs may be relatively old, evolved multiple times in nonhuman vertebrates, show enhanced aggregation during hemoparasite infections, and can be uncovered in vivo by splenectomy. We discuss the working hypothesis that widespread Hb aggregation within several vertebrate groups may be the result of ongoing or past selection pressure against RBC parasites. Further comparative studies of these evolutionary old systems may provide valuable insights into hemoparasite susceptibility and reservoir potential of livestock and companion animals but also into human malaria and sickle cell disease.
- sickle cell
- plasma acidosis
- Plasmodium falciparum
- Opsanus beta
blood o2 transport across vertebrates is optimized by high concentrations of hemoglobin (Hb) inside red blood cells (RBCs). Whereas Hb in nonmammalian vertebrates is even packed between the chromatin of the RBC nucleoplasm (16, 51), mammalian RBCs have dispensed with nuclei altogether, essentially containing cytoplasm filled with Hb close to the solubility limit (47). To avoid Hb aggregation inside RBCs, a strong selection pressure may be expected to act on the surface amino acid properties of Hb, as recently shown for myoglobin, another globin that is highly expressed in muscle cells of mammalian elite breath-hold divers (39). Nevertheless, two variants of human Hb A are known in which a single surface amino acid substitution causes intracellular Hb aggregation in vivo under certain physiological boundary conditions that affect Hb ionization, conformation, or concentration. Curiously, each variant confers protection against Plasmodium falciparum (P. falciparum) malaria. In the famous example of sickle cell disease, homozygous carriers of the Hb S (β6 Glu→Val) allele (with Hb SS RBCs) show intracellular Hb polymerization upon deoxygenation, which is enhanced at low pH and high mean corpuscular hemoglobin concentrations (MCHC) (7, 9). Formation of HbS polymers distorts RBC shape, increases their fragility, and reduces their deformability, leading to severe anemia and a reduced life expectancy (9). Heterozygote Hb S carriers (with sickle cell trait and Hb AS RBCs) require much stronger degrees of Hb deoxygenation, plasma acidosis, and MCHC elevation for RBC sickling to occur (7). Hb AS heterozygotes are therefore usually free of clinical symptoms, and, as the best known example of a balanced polymorphism in humans, they have a greatly reduced relative risk of clinical P. falciparum malaria (10, 20, 40).
In the case of Hb C (β6 Glu→Lys), homozygous carriers (with Hb CC RBCs) are prone to intracellular Hb crystallization, predominantly in oxygenated RBCs and at elevated MCHC (13, 14, 24). Patients show mild chronic hemolytic anemia but are even more strongly protected from clinical P. falciparum malaria than Hb AS heterozygotes. Hb AC heterozygotes show normal blood Hb levels and have some protection against clinical P. falciparum malaria (40). The mechanisms behind the protective effects of the Hb S and Hb C variants are incompletely understood but have been shown to include innate components at the RBC level (18, 19, 43, 49), as well as acquired components (10, 20). Each variant is thought to have arisen more than once in different malaria endemic regions (40). It is therefore not surprising that malaria has been named “the strongest known force for evolutionary selection in the recent history of the human genome” (32). Given the widespread occurrence of intracellular hemoparasites across vertebrates, including viruses, bacteria, and protozoa (15), and the dramatic effects of the single amino acid Hb S and Hb C variants in protection against P. falciparum in human RBCs, we suspected that aggregation-prone Hbs in other vertebrates may be much more common than previously thought.
Indeed, occasional reports about in vitro Hb aggregation as polymers or crystals inside RBCs of nonhuman vertebrates are scattered throughout the literature, including reports in, e.g., deer (52), sheep (12), and cats (2), but also lower vertebrates such as some reptiles (36, 51) and several groups of fishes (22, 28, 37, 53, 58). Thus the propensity of Hb to aggregate inside RBCs, and thereby often dramatically alter cell shape, is widespread among vertebrates. However, with the exception of our previous study on the teleost fish whiting (28), it is largely unknown to what extent and under what conditions Hb aggregation occurs in living animals.
The aim of the present study was therefore to test for the presence of in vivo Hb aggregation in RBCs of a nonhuman vertebrate, the marine Gulf toadfish (Opsanus beta). In contrast to previous studies on Hb aggregation in animals, we use chronically implanted catheters for blood sampling from unrestrained, unanesthetized animals and a standard exercise protocol (56) to vary blood gases and acid-base status in vivo within physiological limits. We demonstrate pronounced and prolonged, exercise-induced in vivo Hb polymerization and RBC sickling in Gulf toadfish, with no apparent signs of hemolysis or anemia. Further analysis of literature data on the occurrence of Hb aggregation in RBCs of other vertebrates, its differential prevalence in infected and uninfected RBCs and in intact and splenectomized animals, leads us to suggest the working hypothesis of a common physiological role for Hb aggregation in impairing RBC parasite propagation and/or in facilitating detection and retention of parasitized RBCs in the spleen.
MATERIALS AND METHODS
Animals and surgery.
Gulf toadfish (Opsanus beta; mean mass 82.0 ± 8.2 g and standard length 15.3 ± 2.3 cm; N = 15) were captured with a roller trawl by commercial shrimpers in Biscayne Bay, Florida in the winter of 2005. The toadfish were held in an outdoor tank at the shrimpers' holding facility with running seawater (ambient seasonal conditions) for no longer than 24 h following capture and then transferred to the laboratory at the Rosenstiel School of Marine and Atmospheric Science, University of Miami, where they were held for a minimum of 1 wk before use. Fish were treated with malachite green (final concentration 0.05 mg/l) in formalin (15 mg/l) (AquaVet, Hayward, CA) on the day of transfer to the laboratory to reduce infection by the ciliate, Cryptocaryon irritans. Fish were kept in 50-l glass aquaria with flowing, aerated seawater (salinity 33‰), at a temperature of 20°C, under a 12-h:12-h light/dark photoperiod regime. Fish were fed weekly with squid up until the time of surgery. Caudal vein catheterizations were performed on fish anaesthetized by immersion in a solution of MS-222 (1 g/l; Argent Chemical Laboratories, Redmond, WA) as outlined before (56). All procedures on toadfish were approved by the University of Miami Animal Care and Use Committee.
In vivo experiments.
During blood sampling, the heparinized saline that filled the catheter was discarded, and the initial 200 μl of blood was kept in a gas-tight syringe (Hamilton, Reno, NV). A 250-μl blood sample was then collected into a separate gas-tight syringe and immediately analyzed for blood pH (50-μl aliquot; Orion Research, Cambridge, MA), hematocrit (triplicate 45-μl samples; Micro-capillary Centrifuge model MB; International Equipment, Needham, MA), Hb concentration (duplicate 10-μl samples in Drabkin's solution measured at 540 nm using an extinction coefficient of 11.00 l·mmol−1·cm−1), total plasma CO2 (duplicate 30-μl samples, Corning 965 Carbon Dioxide Analyzer; Corning Medical and Scientific, Medfield, MA). In addition, a 30-μl blood sample was fixed in glutaraldehyde (see below). After removal of the blood sample, the blood stored in the syringe followed by 250 μl of 150 mM NaCl was injected through the catheter before the catheter was refilled with heparinized saline.
In a group of five fish, the sampling protocol was initiated immediately after surgery to study the effects of surgery on Hb aggregation and hematological parameters. Blood samples were taken at time 0 and also after 2, 24, and 48 h of recovery in individual 1.5-l tanks with running seawater.
All other fish were transferred to individual 1.5-l tanks as above to recover from surgery for 2 days, after which a control blood sample was taken. One experimental group (N = 5) was exposed to exercise, which reduces venous pH (56), a condition that induces Hb polymerization and RBC sickling in whiting (28). Individual fish were transferred to a 100-l tank with aerated seawater for the exercise regime. When prodded gently on the tail with the handle of a net, the fish turned around and attacked it aggressively. This behavior continued for 5 min, but often the fish appeared exhausted by the end. The fish were then returned to the recovery tanks, and a blood sample was taken immediately (t = 0 h) and after 0.5, 1, and 24 h. Parallel samples, studying effects of blood sampling per se, were taken from undisturbed resting fish (N = 5).
Plasma bicarbonate concentration, Pco2, and blood nonbicarbonate buffer value.
Plasma bicarbonate and Pco2 were calculated from total plasma CO2 content and blood pH using the Henderson-Hasselbalch equation and the apparent pKa and CO2 solubility coefficient for human plasma at 20°C, as obtained from Table 1 and Fig. 1, respectively, in reference (8).
A blood nonbicarbonate buffer value of 6.25 mmol·l−1·pH−1 was estimated from the bicarbonate-pH diagram of the closely related oyster toadfish obtained at 20°C and 0.69 mmol/l [Hb4] (based on blood iron content) (48). To assess the acid-base status of exercised Gulf toadfish with an initial mean [Hb4] of 0.45 mmol/l, it was assumed that the blood nonbicarbonate buffer value was proportional to blood [Hb4], and a value of 4.3 mmol·l−1·pH−1 was used.
Blood fixation and light microscopy.
All blood samples for light microscopy and transmission electron microscopy (TEM) were fixed in isoosmotic glutaraldehyde immediately after being sampled. The glutaraldehyde fixation solution was made from a 25% stock solution (Grade I; Sigma-Aldrich, Gillingham, UK), diluted with distilled water to obtain an osmolality of 340 mosmol/kg H2O and further diluted with isoosmotic saline to obtain a 2% glutaraldehyde solution. The composition of the saline was (in mmol/l) 160 NaCl, 3 KCl, 1.5 MgCl2, 1.5 CaCl2, 5 d-glucose, and 20 HEPES; adjusted to pH 7.97 at 20°C. Samples were fixed in an equal volume of this isoosmotic solution to yield a final glutaraldehyde concentration of 1%. They were stored at 4°C until counted or further treated for TEM. The samples were counted using an Axiovert 135 TV microscope (Carl Zeiss MicroImaging, Göttingen, Germany, using a ×100 oil immersion objective and differential interference contrast) fitted with a video camera (3-CCD Color Video Camera, KY-F55B, JVC; Alrad Instruments, Newbury, UK) and viewed with Scion Image software.
Fixed RBC samples were washed three times in pH 7.97 saline (see above) and embedded in 2% agarose before being stained with 1% OsO4 for 1 h, washed in 30% EtOH for 10 min, and incubated in 0.5% uranyl acetate (in 30% EtOH) for 1 h. This was followed by 10-min incubations in each of 30, 60, 70, 80, 90, and 100% EtOH and two washes in 100% acetone to dehydrate samples. The samples were then incubated in a 1:1 ratio of acetone:resin (Araldite) for 30 min followed by 30 min in pure resin, before the final incubation in molds for 24–48 h at 60°C in the oven. An ultramicrotome (Reichert Ultracut; Leica UK, Milton Keynes, UK) was used to cut sections, which were stained for 5 min each with 5% uranyl acetate and 2% Reynold's lead citrate. The samples were viewed on a Tecnai electron microscope (FEI 120kV Tecnai G 2 Spirit BioTWIN, equipped with a SIS Megaview III camera), and the pictures were analyzed using AnalySIS Pro [SIS] software.
Data analysis and statistics.
A minimum of 250 fixed RBCs were studied by light microscopy and assigned as normal (Fig. 1, A and F) or sickled (RBCs with bars of varying sizes in Fig. 1, B–E). The number of sickle cells was calculated as a fraction, Fsickle, of total cells. The mixed venous pH at which 50% of the RBCs were sickled (pKapp) was determined by nonlinear curve fitting of all data points from exercised fish, using Fsickle = Fsickle,max (10-pH)a [(10-pH)a + (10-pKapp)a]-1, where Fsickle,max is the maximal fraction of sickle cells and the exponent a indicates the apparent number of interacting proton-binding sites during the Hb polymerization process, analogous to the Hill number or cooperativity constant that is used to describe the apparent number of interacting O2-binding sites during Hb O2 binding. For comparison, literature data for [Hb] expressed in millgrams of Fe or grams of Hb were converted to molar quantities using the molecular mass of Fe and assuming 64,500 g per mole for Hb4.
Statistical differences in Fsickle and in hematological parameters were tested by a one-way ANOVA or, when appropriate, a one-way ANOVA for repeated measures (RM) followed by a Tukey test for pairwise comparisons (SigmaStat version 2.0, SPSS). All values shown are means ± SE. Statistical significance was accepted at P < 0.05.
Light microscopy showed that the vast majority of RBCs from resting Gulf toadfish with an indwelling venous catheter had a smooth surface and the flattened oval shape typical of nucleated RBCs (Fig. 1A). Immediately after exhaustive exercise, a large fraction of RBCs showed distinct cytoplasmic and nuclear bars of different numbers, shapes, and sizes (Fig. 1, B–E). These cells were referred to as sickle cells. The bars usually occupied only parts of the cytoplasm, except in RBCs with a granular appearance, and, despite occasional RBC projections, the general shape of the cells remained oval (Fig. 1, B–E).
TEM revealed that the bars in the cytoplasm and nucleoplasm consisted of arrays of ordered filaments (Fig. 2, C–E), which were generally absent in samples taken before exercise and in resting fish (Fig. 2, A and B). After 24 h of recovery from exercise, the appearance of the RBCs reversed and could not be distinguished from those of resting fish, either by light microscopy (Fig. 1, A and F) or by electron microscopy (Fig. 2, B and F).
Exercise caused an immediate and pronounced decrease in mixed venous blood pH from 7.91 ± 0.03 to 7.17 ± 0.06, which lasted for at least 1 h (Fig. 3A). This was accompanied by a significant increase in plasma Pco2 from 1.4 ± 0.1 to 9.4 ± 0.8 mmHg immediately after exercise, followed by a gradual decline to 5.7 ± 0.6 mmHg over the next hour (Fig. 3B). The pH-bicarbonate diagram revealed an acidosis of mixed respiratory and metabolic origin because the changes in plasma [HCO3−] with pH followed neither the estimated nonbicarbonate buffer line nor the Pco2 isopleth (Fig. 4). After 24 h, the acid-base status of mixed venous blood had returned to preexercise values (Fig. 3, A and B).
The change in the fraction of sickled RBCs before, immediately after, and during recovery from exhaustive exercise was a mirror image of mixed venous pH, thus it increased from 0.04 ± 0.02 to 0.84 ± 0.04 immediately after exercise followed by a gradual, but statistically insignificant, decline to 0.68 ± 0.11 within an hour of recovery (Fig. 3C). After 24 h, when mixed venous pH had returned to preexercise values, the recovery of sickled cells was complete (Fig. 3, A and C). In resting fish, mixed venous pH and plasma Pco2 remained at 7.91 ± 0.02 and 1.4 ± 0.1 mmHg, respectively, whereas the average fraction of sickled cells was 0.04 ± 0.01 throughout the experiment (Fig. 3).
One animal in the exercise group died from unknown causes in the recovery period before the last blood sample could be taken. Its blood acid-base values and the degree of sickling were well within the range of other individuals. The animal had a slightly lower blood [Hb4] than other individuals, but the reasons behind the slight anemia and whether it was causally related to its death are unknown.
Figure 5 shows a plot of sickle cell fraction against mixed venous pH for individual animals before, immediately after, and during recovery from exercise. A three-parameter sigmoid curve fit gave a maximal fraction of sickled RBCs (Fsickle,max) of 0.81 ± 0.04 (SE), an estimated number of interacting proton-binding sites during polymerization (a) of 5.2 ± 1.9 (SE), and a mixed venous pH value for half-maximal in vivo RBC sickling (pKapp) of 7.61 ± 0.04 (SE). The r2 value of 0.913 indicated that, using this sigmoid model, more than 90% of the observed variation in sickling could be explained by the changes in pH.
Average MCHC decreased at 0 and 0.5 h after exercise compared with resting control fish, indicative of RBC swelling. However, this was not statistically significant, and the mean MCHC values for exercised fish were close to resting values at 1 and 24 h of recovery (Fig. 6A).
Mean blood [Hb4] decreased during the course of the experiment, as expected with a serial blood sampling protocol, although values significantly lower than the starting values were only reached at the last sampling time point in resting and exercised fish (Fig. 6B). Despite this dilution effect, mean blood [Hb4] immediately after exercise was about 30% higher than the preexercise value and about 55% higher than the time-matched value for resting fish (Fig. 6B; P = 0.06), suggesting catecholamine-stimulated RBC release from the spleen (44).
Surgery caused a transient but significant acidosis of about 0.37 pH units, together with significant increases in Pco2 and plasma bicarbonate, compared with the combined pretreatment values of the rest and exercise groups 48 h after surgery (Table 1). At the same time, MCHC significantly decreased and [Hb4] significantly increased, consistent with RBC swelling and release of additional RBCs from the spleen. However, the fraction of sickle cells remained low and did not change significantly. Serial blood sampling after surgery indicated that the initial acid-base and hematological changes were relatively short lived. All values stabilized within 2 h of implanting the indwelling catheter, whereupon they no longer significantly deviated from the long-term 48-h recovery value (Table 1).
The key finding of this study is the pronounced in vivo polymerization of Hb in up to 84% of circulating RBCs in Gulf toadfish after exhaustive exercise (Figs. 1–3). Previous work by Harosi and coworkers (22) has clearly shown by optical methods that the very similar parallel arrays of fibers found in vitro in anoxic RBCs of the closely related oyster toadfish consist of ordered Hb filaments. Gulf toadfish RBCs mostly kept their normal oval shape in the present in vivo study. However, in accordance with previous work (22, 28), we refer to them as sickle cells because Hb polymerization was associated with intracellular RBC bodies of varying size that filled the cytoplasm and nucleus. Compared with resting fish, MCHC values tended to be reduced in the first 30 min after exercise (Fig. 6A), which may have limited the degree of Hb aggregation in individual cells and explain the absence of triangular and other more bizarrely deformed RBCs that have been found in vitro in the closely related oyster toadfish (22). Lowered MCHC values are indicative of RBC swelling, which, under the present experimental circumstances, may be a consequence of the strong exercise-induced plasma acidosis and associated changes in intracellular pH and in the Donnan distribution of permeable anions across the RBC membrane (4, 25). In addition, RBC swelling may due to the well-known β-adrenergic RBC stress response that involves activation of a sodium hydrogen exchanger in the RBC membrane, which tends to protect intracellular pH against acidosis and causes cell swelling (5, 41). The response varies among species and is usually most pronounced in active, visually oriented fishes (6). We have shown in vitro that the β-adrenergic RBC stress response in whiting is able to reduce the extent of RBC sickling (28); however, the magnitude of any potential adrenergic stress response in the RBCs of the sluggish, sedentary Gulf toadfish is unknown. An earlier study by one of us (P. Koldkjær) on the rainbow trout suggests that the response is most marked in young erythrocytes among the RBC population and declines with RBC age (29). Assuming this is similar in Gulf toadfish, this may explain the observation in the present study of a sizeable fraction of RBCs in which sickling was not observed even under the highest degrees of exercise-induced plasma acidosis (Figs. 3C and 5).
Nevertheless, any potential protective effect of an adrenergic RBC stress response would have been limited because the fraction of sickled RBCs remained above 70% for at least 1 h after exercise before it returned to the low values in resting control animals at 24 h (Fig. 3C, Table 1). The possibility of RBC sickling as an artifact was minimized by sampling blood anaerobically from unanesthetized, unrestrained animals via indwelling catheters and by showing that surgery involved in implanting the catheters did not significantly increase RBC sickling.
Sickling was innocuous, as there were no noticeable signs of hemolysis during RBC processing or of chronic anemia, as blood [Hb4] in the first of the serial samples (Fig. 6B) was similar to values in other relatively inactive fish species, whose RBCs are not known to sickle (0.39–0.69 mM Hb4) (21). The death of one fish from the exercise group toward the end of the recovery period may have been associated with the gradual hemodilution caused by serial blood sampling, rather than exercise-induced sickling per se. No fatalities were reported in a previous study on Gulf toadfish from the same area using the same exercise regime, when the number of serial samples was limited to just two (56).
RBC sickling in Gulf toadfish was closely associated with the degree of exercise-induced acidosis (Figs. 3 and 5). Half-maximal sickling occurred at pH 7.61, 0.3 pH units below resting mixed venous blood pH. We previously found very similar values for in vitro RBC sickling in the distantly related marine teleost fish whiting, namely pH 7.64 and 7.53 in air- and N2-equilibrated RBCs, respectively (28). Notably, in Atlantic cod, RBC sickling and aggregation of purified Hb are both enhanced at low pH values (22, 45). Physicochemical conditions promoting intracellular Hb aggregation in other vertebrates differ according to species, Hb phenotype, and form of aggregates (Table 2). In all known cases, elevated MCHC and temperature enhance Hb aggregation, whereas the effects of pH and oxygen tension vary as discussed earlier for human Hb C and Hb S. Preliminary in vitro studies indicate a slightly higher number of sickled cells in air-equilibrated toadfish RBC suspensions than under nitrogen and oxygen atmospheres and confirm enhanced sickling at elevated MCHC (P. Koldkjær and M. Berenbrink, unpublished observations).
The Hb system of Gulf toadfish has not yet been characterized and globin amino acid sequences, and structural data are presently unavailable. However, the steep pH dependence of sickling suggests that the molecular mechanism involves several interacting proton-binding sites, as also suggested for sickling in whiting RBCs (28).
All exercised Gulf toadfish in our study showed pronounced in vivo RBC sickling, and the sickling phenomenon has further been observed in vitro in two other species of toadfishes (22, 53). Although sickling appears to be absent in many fish species across several orders (22, 28), it was also found in all investigated individuals of the distantly related whiting (28) and several of its close relatives (22, 28) (Table 2), suggesting that it may have evolved independently in ancestors of the teleost fish orders Batrachoidiformes and Gadiformes, respectively.
The Cervidae (deer family) and Caprinae (sheep and goat subfamily of bovids) within the artiodactyls present two other taxonomic groups in which RBC sickling occurs in several closely related species (Table 2). In both groups, oxygenation and alkaline pH promote intracellular aggregation of certain Hb genotypes and RBC sickling, which is frequently observed in air-dried blood smears. Exposure to air oxygenates venous blood and alkalinizes it due to CO2 loss, so the in vivo relevance of the sickling phenomenon in these groups has been generally questioned (52, 57). Indeed, in a study on sika deer (42), RBC sickling was absent in venous blood with a normal pH of 7.37 obtained by an indwelling catheter from physically restrained animals. However, when venous pH was elevated above 7.4 by bicarbonate infusion, more than 80% of circulating RBCs became sickled (42). Such a modest blood alkalinization readily occurs in deer during exercise-induced hyperventilation (57), suggesting that sickling in deer, contrary to current opinion, may occur in vivo. This has previously been suggested for the Caprinae, where RBC sickling has been demonstrated in fresh, anaerobically obtained blood samples of Angora goats (27).
The present study on Gulf toadfish, together with previous reports (Table 2), suggests that in vivo Hb aggregation and RBC sickling have independently evolved and become prevalent in several diverse groups of vertebrates. As permanent Hb aggregation appears incompatible with function, it is not surprising that, when in vivo Hb aggregation is found, it is reversible and restricted to physiological boundary conditions that are infrequently encountered. The question arises as to why Hb aggregation should be tolerated at all in so many different animal groups.
A working hypothesis for the repeated occurrence of RBC sickling across vertebrates.
The answer may be related to the strong positive selection acting on the aggregating human β-chain variants Hb S and Hb C in populations exposed to P. falciparum malaria. The highly virulent nature of P. falciparum infection in humans is unusual for other Plasmodium infections among primates and thought to have evolved relative recently in humans some time after they diverged from their closest living chimpanzee relatives (30). This is in line with the absence of a genomic signature of strong selection pressure on the Hb β-chains of chimpanzees (34), which have the same amino acid sequence as in human Hb A (46) but in which P. falciparum infection appears asymptomatic (1). Based on the lower fitness cost of Hb C compared with Hb S, it has been estimated that, under continuous threat of P. falciparum malaria, the Hb C variant will become the dominant human Hb form in areas where these variants cooccur (40), and this is in as little as 50 generations (23). In such future human populations, the situation will, in a sense, appear very similar to what is currently observed in other vertebrates, such as white-tailed deer or whiting and toadfish, in which widespread Hb aggregation can be uncovered under certain physiological boundary conditions, without any apparent ill effect or obvious parasite association.
Indeed, when taken together, several seemingly unconnected reports throughout the literature are in line with a general role for Hb aggregation in protection against RBC parasites across vertebrates. Thus a variety of hemoparasites is known to increase the propensity for Hb aggregation in their RBC host; human Hb AS RBCs sickle up to eight times more readily under low oxygen tensions when infected with P. falciparum compared with uninfected Hb AS RBCs (33, 49). Cat RBCs experimentally infected with the bacterial parasite Mycoplasma haemofelis (formerly Haemobartonella felis) show a higher frequency of intracellular Hb crystallization than nonparasitized RBCs (50). In cattle and other artiodactyls, several protozoan Theileria species induce distinct parasite-associated Hb crystals in infected, but not uninfected, RBCs (54). In rhinoceros iguana, bacterial hemoparasite infection was associated with a fourfold higher number of sickled RBCs than in healthy control animals (51). Finally, viral erythrocytic necrosis in Atlantic cod was associated with intracellular Hb polymerization in infected, but not uninfected RBCs (3).
Parasite-induced formation of extensive Hb aggregates, which may fill most of the cytoplasm in the cat and cod (50, 53), may generally inhibit parasite growth and multiplication. Thus growth and multiplication of P. falciparum are impaired in human Hb AS and Hb SS RBCs exposed to low oxygen tensions in vitro (19, 43). This has been linked to a strong parasite-induced intracellular acidification that causes even Hb AS RBCs to sickle at low oxygen tensions (10) and has been termed “suicidal infection” (29). In human Hb CC RBCs, multiplication rates of P. falciparum are reduced at high and low oxygen tensions (18), consistent with formation of Hb C aggregates under both conditions (17, 24). Comparative parasite growth studies in RBCs of animals with aggregating Hbs are rare. Thus Holman and colleagues (26) noted that in vitro growth of Babesia bovis in RBCs of white-tailed deer partly depended on the sickling propensity of host RBCs (26) although no further details were given. More recently, it was shown that the susceptibility of sheep RBCs to experimental infection with Babesia divergens showed distinct intraspecific variation for as yet unknown reasons (35). Interestingly, an earlier study has shown in vitro sickling of sheep RBCs with distinct intraspecific differences: sheep with the Hb A and Hb AB phenotypes show in vitro RBC sickling, whereas those with the Hb B phenotype do not (12).
Clearly more studies are needed to decide whether Hb aggregation affects parasite growth and multiplication also in RBCs of nonhuman vertebrates. However, apart from this, parasite-triggered Hb aggregation may also generally facilitate recognition and removal of parasitized RBCs by decreasing their mechanical deformability and filterability in the spleen. Thus, in a mouse model of human malaria, the protection against mouse-specific Plasmodium infections afforded by the expression of human Hb S was reversed after splenectomy (29). In humans homozygous for Hb C, as well as in normal cats, the number of circulating RBCs with crystalline Hb bodies increases after splenectomy (2, 17). Similarly, the peculiarly deformed RBCs in cod fishes, later shown to be caused by intracellular Hb polymerization (22, 28, 53), have first been described in situ in the spleen of these fishes (58). Furthermore, in white-tailed deer, in which in vitro RBC sickling is widespread, splenectomy has been shown to uncover previously undetected infections with four different bacterial and protozoan hemoparasites, including a novel Plasmodium species (31).
Taken together, these apparently isolated findings in nonhuman vertebrates support a role of Hb aggregation as a general nonspecific, innate immune mechanism against RBC parasite infection. According to this model, natural selection has favored certain Hb mutations that increase the propensity for reversible Hb aggregation (polymerization or crystallization) under physiological boundary conditions. Upon infection, parasite-induced changes in the microenvironment of these meta-stabile Hbs, such as altered pH and oxygen tension, or increased MCHC, trigger Hb aggregation. This in turn impairs parasite growth and multiplication and/or alters mechanical deformability of infected RBCs, leading to their increased retention and phagocytosis in the rigorous RBC quality control system of the spleen.
Such a mechanism does not exclude significant contributions by other protective mechanisms affecting RBC invasion, cytoadherens, or acquired immune responses (10, 55). However, it accounts for the previously unexplained, widespread propensity for innocuous Hb crystallization and polymerization in many different vertebrate groups, as here shown in vivo for Gulf toadfish.
Perspectives and Significance
Finally, the comparative evolutionary approach to Hb aggregation, RBC sickling, and parasite protection may not only allow a glimpse into the possible fate of human populations under continuing malaria threat, but may also reveal novel mechanisms that have been established in different vertebrate groups over evolutionary time and that reduce the potentially adverse effects of Hb aggregation that are still seen in human sickle cell disease today, 100 years after it was first discovered. This is exemplified by the β-adrenergically activated sodium hydrogen exchanger in the RBC membrane of whiting, described above, that has been shown to reduce Hb aggregation in whiting RBCs in vitro (28). Conversely, applying concepts developed in human malaria research more widely to other vertebrates holds promise to understand currently unexplained variations in RBC parasite susceptibility of wildlife, pets, and livestock, with potentially important implications, such as the identification of parasite reservoirs and the management of emerging zoonotic diseases, such as human babesiosis (59).
This study was supported by a grant from the UK Biotechnology and Biological Sciences Research Council (BBSRC, Grant 26/S17991) to M. Berenbrink.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: P.K. and M.D.M. performed experiments; P.K., I.P., and M.B. analyzed data; P.K., I.P., and M.B. interpreted results of experiments; P.K., I.P., and M.B. prepared figures; P.K., M.D.M., I.P., and M.B. drafted manuscript; P.K., M.D.M., I.P., and M.B. edited and revised manuscript; P.K., M.D.M., I.P., and M.B. approved final version of manuscript; M.B. conception and design of research.
We thank Pat Walsh for discussions and for providing laboratory space at the University of Miami, David Montagnes for help with light microscopy, and Anne Lanevschi for commenting on an earlier draft of the manuscript.
- Copyright © 2013 the American Physiological Society