Abstract
During capillary transit, red blood cells (RBCs) must exchange large quantities of CO2 and O2 in typically less than one second, but the degree to which this is rate-limited by diffusion through cytoplasm is not known. Gas diffusivity is intuitively assumed to be fast and this would imply that the intracellular path-length, defined by RBC shape, is not a factor that could meaningfully compromise physiology. Here, we evaluated CO2 diffusivity (DCO2) in RBCs and related our results to cell shape. DCO2 inside RBCs was determined by fluorescence imaging of [H+] dynamics in cells under superfusion. This method is based on the principle that H+ diffusion is facilitated by CO2/HCO3− buffer and thus provides a read-out of DCO2. By imaging the spread of H+ ions from a photochemically-activated source (6-nitroveratraldehyde), DCO2 in human RBCs was calculated to be only 5% of the rate in water. Measurements on RBCs containing different hemoglobin concentrations demonstrated a halving of DCO2 with every 75 g/L increase in mean corpuscular hemoglobin concentration (MCHC). Thus, to compensate for highly-restricted cytoplasmic diffusion, RBC thickness must be reduced as appropriate for its MCHC. This can explain the inverse relationship between MCHC and RBC thickness determined from >250 animal species.
Similar content being viewed by others
Introduction
The principal biological function of red blood cells (RBCs) is to exchange large volumes of O2 and CO2 during their brief (typically <1 s) transit through the microvasculature1. The speed of gas turnover by RBCs is therefore a measure of their physiological fitness, and highly-conserved biological adaptations are expected to relate to faster gas exchange. The steps in the gas exchange cascade include gas permeation across the cell membrane and binding onto hemoglobin. Additionally in the case of CO2, gas molecules are converted reversibly to HCO3− ions (which are then transported by anion exchanger 1, AE1) plus H+ ions (which are buffered by hemoglobin). These processes are coupled together by cytoplasmic diffusion. According to the prevailing consensus, efficiency of gas exchange is strongly dependent on protein-facilitated membrane transport, including AE1-assisted HCO3− transport and aquaporin1-assisted gas permeation2,3.
A conserved and characteristic feature of human and animal RBCs is their flattened shape. The physiological relevance of this geometry has largely been attributed to the mechanical benefits it offers to circulating blood. RBC flattening allows microvasculature to co-evolve smaller luminal diameters and therefore maximize capillary density. In the case of human RBCs, the biconcave shape supports laminar flow and shear-thinning4, minimizes platelet scatter and atherogenic risk5, permits cells to squeeze through microvasculature6,7 and is the lowest energy-level that the cell returns to following deformation in capillaries8,9,10,11. Elliptical or otherwise flattened RBCs of many non-human species may also manifest, to some degree, these mechanical benefits. Recently, cell geometry has been proposed as a stringent criterion by which splenic inter-endothelial slits select healthy RBCs for continued circulation12, but it is unclear if similar criteria apply in non-human species. These aforementioned mechanical considerations do not, however, offer a unifying explanation for the wide range of RBC thicknesses observed naturally in different animal species, and its inverse relationship with mean corpuscular hemoglobin concentration (MCHC)13,14,15. We postulate that RBC thickness and its relationship with MCHC may, instead, relate more closely to the efficiency of gas turnover.
Theoretically, the flattened RBC form may facilitate gas exchange in two ways. Firstly, it increases the cell’s surface area/volume ratio (ρ), which allows faster membrane transport. Secondly, it collapses the path-length for cytoplasmic diffusion, which reduces the time delays associated with intracellular gas transport. In the case of human RBCs (major radius, r, of 4 μm), adapting the shape from a hypothetical sphere (volume 268 fl; surface area 200 μm2) to a flattened form (volume 90 fl; surface area 136 μm2) doubles ρ and hence transmembrane flux13,14,15. The significance of this effect can be evaluated by considering the slowest membrane transport process, AE1-facilitated HCO3− transport, which has an apparent permeability constant3 (Pm,HCO3) of 18 μm/s and can be ascribed a time constant equal to 1/(ρ × Pm,HCO3). The AE1-related time constant for the spherical geometry is 0.074 s, which is reduced to 0.037 s for the flattened shape, yet both are compatible with equilibration during capillary transit. The biconcave shape of human RBCs reduces the mean cytoplasmic path-length to 0.9 μm (equal to the cell’s average half-thickness, h), which shortens intracellular diffusion time delays by a factor of seven, compared to a spherical RBC variant (r2/6 ÷ h2/2). However, this seven-fold acceleration will not translate into a meaningful improvement to gas exchange efficiency if gas diffusion coefficients are high, as they are in water. Paradoxically, gas diffusivity inside gas-carrying RBCs has not been measured, but is intuitively assumed to be rapid. Indeed, O2 and CO2 diffusivity measurements in hemoglobin solutions and hemolysates (~103 μm2/s) appear to support this assertion2,14,16. At this magnitude of diffusivity, cytoplasmic path-length and cell shape are not expected to be critical for determining the efficiency of gas exchange. In summary, our current understanding of RBC physiology predicts only a modest advantage of the flattened RBC shape for gas exchange, but this inference is based on an indirect characterization of gas transport inside the RBC.
In this study, we revisited the notion that gases diffuse rapidly in RBC cytoplasm. The required experimental approach must be capable of resolving diffusion on the scale of a single intact cell, and exclude contributions from permeation across extracellular unstirred layers and the surface membrane. To meet these criteria, we evaluated CO2 diffusivity (DCO2) in intact RBCs by measuring the ability of CO2/HCO3− to facilitate cytoplasmic H+ diffusion17,18,19. This measurement method is based on the observation that cytoplasmic H+ ions are heavily buffered20, and therefore diffuse only as fast as the buffers there are bound to (with essentially no free ion movement)21,22. Thus, by determining the CO2/HCO3−-dependent component of buffer-facilitated H+ diffusion17,18,23, it is possible to quantify DCO2. Our measurements on human RBCs demonstrate substantially restricted CO2 diffusivity, to 5% of the rate in water, which is an order of magnitude slower than estimates made previously in cell-free hemoglobin solutions. By imaging osmotically-swollen human RBCs (to dilute MCHC) and RBCs from species with different hemoglobin concentrations, we demonstrate that DCO2 decreases sharply with MCHC, consistent with hemoglobin-imposed tortuosity to small-molecule diffusion. We conclude that diffusion across cytoplasm is a hitherto unrecognized rate-limiting step for gas exchange which imposes a critical limit on the RBC thickness, beyond which physiological function becomes inefficient and incomplete. In particular, the full manifestation of the Bohr Effect24 (i.e. the process by which hemoglobin releases O2 at acidic vascular beds) is attainable only in RBCs that are adequately thin because the underlying H+ trigger is transmitted intracellularly by slow CO2/HCO3− diffusion. Highly restricted cytoplasmic diffusivity can explain the inverse RBC thickness/MCHC relationship observed amongst different animal species, and highlights the potential vulnerability of gas exchange efficiency in diseases that involve a change in RBC shape, such as in spherocytosis.
Results
Using measurements of spatio-temporal [H+] dynamics as a read-out of CO2 diffusion
CO2 cannot be imaged dynamically inside a single intact RBC, but its acidic chemistry allows pH-sensitive fluorescent dyes, such as cSNARF120, to monitor its diffusion. H+ ions are highly buffered in cells, therefore their apparent cytoplasmic diffusivity (DHapp) is determined exclusively by the mobility of H+-carrying buffers (Fig. 1A,B)21,22. Based on this biophysical principle, it is possible to describe the diffusive properties of cytoplasmic buffers from measurements of H+ dynamics in intact cells18,25. To determine DHapp, a local source of H+ ions was produced in cytoplasm by photolytic uncaging from the membrane-permeant H+-donor 6-nitroveratraldehyde (NVA)23 once every 0.13 s. This generates an acidic microdomain which dissipates at a rate determined by the buffers’ concentrations, reaction kinetics with H+ ions, and diffusion coefficients. To follow the progress of buffer-facilitated H+ diffusion, pH-sensitive cSNARF1 fluorescence was imaged confocally across the RBC’s horizontal plane at intervals between H+ uncaging (e.g. Fig. 1C,D)20. Provided that membrane transport of H+ and HCO3− ions is slow or inactivated, best-fitting these spatio-temporal [H+] data with a diffusion equation (derived previously18,23,26) gives a robust measure of DHapp.
During experiments, continuous superfusion of cells controls for temperature and CO2/HCO3− concentration20. In RBCs superfused with CO2/HCO3−-free solution, H+ diffusion is facilitated solely by hemoglobin (Fig. 1A). Under these conditions, DHapp reports H+ diffusion facilitated by the translation and rotation of hemoglobin molecules27. When RBCs are superfused with CO2/HCO3−-containing solution, cytoplasmic diffusion of H+ ions is additionally facilitated by CO2/HCO3−, thus DHapp reports the combined effects of hemoglobin and CO2/HCO3− (Fig. 1B). Since these two buffer-shuttles are additive, the effect on DHapp of introducing CO2/HCO3− into cytoplasm is a read-out of the turnover of the CO2/HCO3− buffer-shuttle. High carbonic anhydrase (CA) activity in RBCs ensures that the CO2/HCO3− buffer-shuttle is not meaningfully rate-limited by chemical reactions; instead, CO2/HCO3−-facilitated H+ diffusion is strongly dependent on the cytoplasmic diffusion coefficients of CO2 gas and HCO3− ions (DCO2, DHCO3). Thus, the first step in quantifying DCO2 and DHCO3 is to probe the CO2/HCO3−-dependent and independent components of DHapp.
CO2/HCO3− weakly facilitates cytoplasmic H+ diffusion in the cytoplasm of human RBCs
CO2/HCO3−-independent DHapp was measured in human RBCs superfused at 37 °C with CO2/HCO3−-free normal Tyrode (0NT) solution containing 1 mM NVA. Photolysis acidifies the cytoplasm at ~0.1 pH units/s, therefore DHapp measurements correspond to a mildly acid-shifted cytoplasmic pH (pHc). To probe DHapp over a range that encompasses physiological pHc, experiments were performed in superfusates at pH 7.4 or 7.8. Membrane transport of HCO3− by AE1 is largely inactivated in the absence of CO2/HCO3−, therefore uncaged H+ ions are retained in cytoplasm. Figure 1C shows the slow spread of H+ ions through the cytoplasm of a human RBC in the absence of CO2/HCO3−. Measured DHapp (Fig. 1E) was three orders of magnitude slower than in water (1.2 × 104 μm2/s)28. Immature RBCs, or reticulocytes, have been proposed to contain small histidine derivatives, such as carnosine29, which could facilitate H+ diffusion alongside hemoglobin. However, DHapp in cells identified positively by Thiazole Orange staining as reticulocytes was no different to DHapp in Thiazole Orange-negative (mature) RBCs (Fig. S1). Thus, the concentration of small-molecule histidine derivatives in reticulocytes is not sufficient to meaningfully influence H+ diffusivity.
Next, experiments were performed on RBCs superfused with CO2/HCO3−-buffered NT (BNT) containing 5% CO2 and either 22 mM HCO3− (for pH 7.4) or 55 mM HCO3− (for pH 7.8; Fig. 1D; solution osmolality was corrected to 296 mOsm/kg by balancing [NaCl]). If DCO2 were rapid, the CO2/HCO3− buffer-shuttle would accelerate DHapp substantially and collapse pH gradients; however, introducing CO2/HCO3− into cytoplasm increased DHapp by only 2–3 μm2/s (Fig. 1E), indicating that CO2/HCO3− has a limited capacity to facilitate H+ diffusion, comparable to that of hemoglobin. Activation of AE1 in CO2/HCO3−-containing superfusates may, in principle, compromise the accuracy of DHapp measurements. However, DHapp estimates were unaffected by blocking AE1 with 12.5 μM or 163 μM 4,4′-diisothiocyano-2,2′-stilbenedisulfonic acid (DIDS; Fig. 1E), indicating that the magnitude of membrane HCO3− transport is not sufficient to short-circuit the cytoplasmic CO2/HCO3− buffer-shuttle (NB: even the lower dose of DIDS was sufficient to inhibit AE1; Fig. S2). Cytoplasmic CA activity was not inactivated by DIDS, NVA or its photolytic derivative (Fig. S3), therefore low DHapp was not an erroneous result of inhibited CO2/HCO3− reaction kinetics.
Solute diffusion in RBC cytoplasm is restricted by the high concentration of hemoglobin
To investigate if slow cytoplasmic diffusivity is unique to CO2/HCO3−, the mobility of calcein (a fluorescent marker) was measured in intact human RBCs. Applying a high-intensity laser beam every 0.13 s to one end of the cell bleaches calcein locally, and drives a diffusive redistribution of fluorescence signal, which was imaged throughout the cell at intervals between bleaching events. Time delays in calcein fluorescence measured at different distances from the bleaching region provide a readout of cytoplasmic diffusivity (Dcalc), measured to be 47.8 ± 5.95 μm2/s, i.e. 13-fold lower than in water (~600 μm2/s)30 (Fig. 2A). This result indicates that RBC cytoplasm is a highly tortuous environment for diffusion.
Hemoglobin, which occupies a quarter of human RBC volume20, is likely to impose a substantial tortuosity to the movement of solutes in cytoplasm. Loosening hemoglobin density is expected to accelerate small-molecule diffusion, and this was tested in osmotically-swollen cells (Fig. 2B). RBCs were first pre-equilibrated in, and then superfused with 155 or 220 mOsm/kg solutions (prepared by reducing [NaCl]). In the absence of CO2/HCO3−, cytoplasmic dilution increased DHapp (Fig. 2C), presumably because of less restricted translational and rotational movements of H+-carrying hemoglobin molecules27. The relationship between DHapp and osmolality was steeper after introducing a constant concentration of CO2/HCO3− into cytoplasm (Fig. 2C). This finding indicates that at lower hemoglobin density, CO2/HCO3− is more effective in facilitating H+ diffusion. This occurs despite CA activity dilution, adding further evidence that H+ diffusion is not reaction-limited (i.e. even after 2-fold dilution, CA activity remains very high).
To explore the effect of naturally occurring differences in hemoglobin concentration on DHapp, measurements were performed on RBCs from chicken, alpaca and Xenopus (Fig. 3A–C). Previous studies have detected carnosine in nucleated erythrocytes31, which may augment DHapp in chicken and Xenopus RBCs by acting as a mobile buffer alongside hemoglobin and CO2/HCO3−. However, using a modification of Pauly’s assay, levels of small-molecule histidine derivatives were in the sub-millimolar range, with the highest levels detected in chicken RBCs (0.5 mM; Fig. S4). Over this low concentration range, histidine-containing small molecules cannot meaningfully increase DHapp, therefore hemoglobin and CO2/HCO3− remain the principal H+-carriers. To measure DHapp, alpaca and chicken RBCs were superfused in solution at 37 °C and pH 7.8, whereas Xenopus cells were superfused at room temperature and pH 7.5 to match their normal physiology. Nuclear regions in Xenopus and chicken RBCs were excluded in the analysis of pH (Hoechst 33342-positive nuclear areas also had higher intensity of cSNARF1 fluorescence, likely reflecting the ambient physicochemical environment of nucleoplasm; Fig. S5). Compared to human RBCs, chicken and Xenopus RBCs are larger and have a lower MCHC, whereas alpaca RBCs are smaller but with higher MCHC (Table S1). High CA activity was detected in hemolysates from all species studied, although it was lower in alpaca and Xenopus compared to humans (Fig. S6). Introducing CO2/HCO3− into cytoplasm increased DHapp in RBCs from chicken and Xenopus, but not alpaca (Fig. 3D). Hypotonic swelling (i.e. loosening hemoglobin density) increased DHapp further in chicken and alpaca RBCs (NB: Xenopus RBC are too fragile for this experiment). Figure 3E summarizes the relationship between DHapp and MCHC; with the exception of cold-blooded Xenopus, all data-points followed an exponentially-declining relationship. In summary, the facilitatory effect of CO2/HCO3− on DHapp was attenuated at higher MCHC, consistent with hemoglobin-imposed tortuosity.
Calculated CO2 diffusivity in cytoplasm is slow and dependent on hemoglobin concentration
Cytoplasmic H+ diffusivity is related mathematically21,22 to the concentration, mobility and protonation/deprotonation kinetics of hemoglobin and CO2/HCO3−. In this system, all variables except for CO2/HCO3− mobility (DCO2, DHCO3), are known, therefore an algorithm can be designed to calculate DCO2 and DHCO3 from DHapp measurements. Since CO2 diffuses 46% faster than HCO3− (a difference attributable to molecular size)2, DCO2 and DHCO3 are numerically constrained to one another, reducing the system of equations to one unknown. This algorithm, described in more detail in the Supplement, performs simulations for a range of different test-values of DCO2 (and DHCO3) and for each generates a predicted DHapp. Least-squares best-fitting this output to the experimentally-determined DHapp derives DCO2 (and DHCO3). The ‘known’ variables required to run these simulations were obtained as follows. Firstly, the concentration of hemoglobin was obtained from MCHC measurements20,32, and the cytoplasmic concentration of CO2/HCO3− was calculated using the Henderson-Hasselbalch equation from extracellular [CO2] and [HCO3−] and the measured transmembrane pH gradient. Secondly, hemoglobin reaction kinetics were assumed to be instantaneous, whereas the reactions of CO2/HCO3− were modelled kinetically using data for CA activity, hydration and dehydration rate constants. Finally, hemoglobin-facilitated H+ diffusivity is equal to DHapp measured in CO2/HCO3−-free media.
Cytoplasmic DCO2 and DHCO3 were found to be substantially lower than the rates measured in water (Fig. 4A) and were inversely correlated with MCHC (Figs 4B and S7). For example, CO2 diffusivity in human RBC cytoplasm was 5% of the rate in water. The extent to which hemoglobin restricts small-molecule diffusion is substantially greater than previous estimates (inset to Fig. 4B; e.g. for human RBCs, the difference is one order of magnitude). Thus, highly restricted diffusivity of CO2 gas and HCO3− anions in RBC cytoplasm is a hitherto unrecognized rate-limiting step in the process of gas exchange.
Discussion
Chemical reactions and membrane transport are recognized to be critically important in determining the rate of gas exchange, and their experimental characterization has informed our current model of RBC physiology. The results of this study indicate that diffusion inside RBCs is a strongly rate-limiting factor for gas turnover. We find that RBC cytoplasm is a highly tortuous environment that substantially restricts the movement of solutes, consistent with the uniquely high microviscosity inside RBCs33. In terms of resistances to overall gas transport34, the component attributable to RBC cytoplasm has hitherto been underestimated, and so it is plausible that the other resistances in series (imposed by membranes and extracellular unstirred layers) have been exaggerated in earlier analyses3.
The extent to which the cytoplasm of intact human RBCs restricts gas diffusion is greater than estimated previously from cell-free solutions12,15. This may relate to the significantly slower rotational diffusion of hemoglobin inside RBCs compared to solution33, which argues that encapsulation in a membrane produces a more rigid lattice of macromolecules. It is also possible that diffusivity measurements in cell-free solutions may have been over-estimated as a result of convective currents of the medium, which are much less likely to occur inside cells. Of note is that our measured effect of hemoglobin on CO2 diffusivity is quantitatively similar to the effect of an unrelated macromolecule, Ficoll70, on rhodamine green diffusion35,36. Thus, hemolysates and hemoglobin-solutions may have only a limited ability to fully recapitulate the biophysical properties of intact RBC cytoplasm.
The cytoplasm of human RBCs reduces CO2 diffusivity by a factor of 23 and calcein diffusivity by 13-fold. The difference in tortuosity imposed on CO2 and calcein may relate to the solute’s reactivity with hemoglobin. CO2 binds weakly and reversibly to hemoglobin2 (forming carbamino-hemoglobin), which will retard CO2 translational diffusion. In contrast, the highly negatively charged calcein will not interact with hemoglobin in this way. Also, because the CO2 molecule (radius 2 Å) is 50-fold smaller than the calcein molecule (radius 7.4 Å), CO2 may be able interact more intimately with the surface of the hemoglobin (radius 27.5 Å). Consequently, CO2 may experience a more tortuous path around hemoglobin, compared to calcein.
Since diffusivity, path-length and cell shape are inter-related, our measurements are an important addition to our understanding of RBC form and function. In light of highly restricted diffusivity, the advantage of flattening an RBC can be explained in terms of minimizing critical time delays, which are proportional to the square of distance. To explore this proposal, a mathematical model (described more fully in the Supplement) simulated the time course of CO2 penetration into human RBCs entering a hypercapnic (8% CO2) environment (Fig. 5Ai). Predictions for normal human RBCs (half-thickness 0.9 μm; diameter 8 μm) were compared to those for a hypothetical spherical variant (diameter 8 μm). Using DCO2 data derived previously from cell-free solutions (40% of diffusivity in water, labelled here as ‘fast’), the time-to-reach 90% of the equilibrium state (τ90) for CO2 partial pressure (pCO2) was adequately rapid in both the flat (0.03 s) and spherical (0.09 s) geometry (Fig. 5Aii). Simulations using the presently determined value of DCO2 (5% of diffusivity in water, i.e. ‘slow’) showed a dramatically prolonged τ90 for pCO2 equilibration: 0.09 s for a flat RBC and 0.38 s for its spherical variant (Fig. 5Aiii). The later delay is incompatible with efficient gas exchange with faster blood flows (e.g. in exercise). Thus, the critical importance of RBC thickness for efficient gas exchange becomes apparent only when slow cytoplasmic diffusivity is considered. The cytoplasmic restrictions imposed on CO2 movement may also apply to O2 diffusion, therefore RBC thickness could also be important in determining the rate of O2 exchange.
Another consequence of profoundly restricted CO2 and HCO3− diffusion is very slow H+ ion mobility in RBC cytoplasm. Remarkably, H+ diffusivity inside RBCs is the slowest among all ions studied in cells; this is somewhat surprising given that RBCs experience the highest acid-base turnover in the body26. H+ ions are a powerful signal that regulates O2 binding to hemoglobin through the Bohr24 and Root Effects37,38, but this response is only effective if the RBC’s mean cytoplasmic pH (pHRBC) is able to track changes in ambient pH (e.g. during the transit through acidic microvasculature of contracting muscles). Slow cytoplasmic H+ diffusion thus places an additional constraint on RBC thickness, which was analyzed, using the model, in terms of the pHRBC response to an ambient metabolic acidosis (Fig. 5Bi). Assuming the higher DCO2 value, a flattened RBC will acidify faster than a spherical cell, but in both cases, τ90 was under 1 s (0.48 s for flat; 0.75 s for spherical; Fig. 5Bii). Using the lower DCO2 value measured herein, τ90 would increase to 0.54 s and 0.95 s for flat and spherical cells, respectively (Fig. 5Biii). The delay associated with spherical geometry would result in only a partial manifestation of the Bohr Effect, and consequently suboptimal distribution of O2 to tissues. Thus, in order for O2 release from RBCs to respond adequately to changes in ambient pH, the cell must acquire a flattened form.
Circulating human RBCs undergo a dramatic velocity-dependent deformation (e.g. parachute-like shape) as they transit through the microvasculature39,40,41. As a result of cytoplasmic re-distribution, the intracellular diffusion distances may modestly increase towards the leading edge of the cell and decrease at rear. Nonetheless, the parachute RBC retains an overall flattened form that is compatible with efficient gas exchange. This ability of flattened human RBCs to attain a parachute shape in capillaries may be critical for maintaining minimal cytoplasmic path-lengths for efficient gas exchange.
Our work also quantified the effect of MCHC on gas diffusion. There is considerable inter-species variation in MCHC (Table S2) which, at least in part, is driven by demand for O2-carrying capacity, but nonetheless appears to be capped at ~550 g/L. High MCHC is associated with raised microviscosity and resistance to blood flow6,42, which could underlie the biological limit of MCHC, although this can be compensated for by reducing RBC count. Here, we demonstrate that raising MCHC by 75 g/L halves DCO2 (Fig. 4B), an effect that is quantitatively similar to that of other macromolecules (e.g. rhodamine green diffusivity halves with every 88 g/L rise in Ficoll70 concentration)36. Thus, a physiological need to pack more hemoglobin into a RBC (e.g. observed in some altitude-adapted species, such as llamas and alpacas) comes at the cost of more restricted gas diffusivity. Cells adapted in this manner would need to be appropriately thinner in order to support efficient gas exchange. To investigate this hypothesis, we compiled data on RBC half-thickness (i.e. the shortest cytoplasmic path-length) and MCHC from >250 cold- and warm-blooded species (see Table S2; Fig. S8). Figure 5C shows a two-dimensional frequency histogram for RBC half-thickness and MCHC, superimposed with simulations for the time-to-reach 90% pCO2 equilibrium (τ90), determined using the model, shown in Fig. 5A, with appropriately varied MCHC (hence DCO2, DHCO3) and half-thickness. The majority of naturally-occurring MCHC/half-thickness combinations fell in the τ90 range of 0.03−0.3 s, i.e. compatible with near-complete equilibration during typical capillary transit times. This model was also used to simulate τ90 for pHRBC equilibration in response to ambient acidosis. The majority of naturally-occurring MCHC/half-thickness combinations were constrained to meet the criterion of τ90 < 1 s (Fig. 5D). We postulate that thick RBCs with high MCHC (i.e. upper right quadrant in Fig. 5C,D) are not normally found in nature because these would be associated with unacceptably slow gas exchange (τ90 for pCO2 equilibration >0.3 s) and an incomplete manifestation of the Bohr Effect (τ90 for pHRBC equilibration >1 s). Membrane tension and, in some species, the presence of a nucleus limit the extent to which RBCs could be made thinner; we propose that this imposes a cap on MCHC. According to Fig. 4B, MCHC greater than 500 g/L would restrict CO2 diffusion by a factor of >100, requiring RBCs to be thinner than 1 μm; this may explain why such high hemoglobin levels are rare, despite pertinent selection pressures for increasing O2-carrying capacity. This line of reasoning may also explain why mature RBCs in mammals have lost their nucleus in a bid to become thinner.
The regulatory means by which normal RBCs meet the favorable MCHC/half-thickness criterion is likely to involve interactions within the set of gene loci (e.g. seventy-five in humans)43 that collectively determine RBC shape and hemoglobin content. Failure to attain a favorable combination of cell thickness and MCHC may impact on gas exchange efficiency. The analyses shown in Fig. 5C,D highlight a potential disadvantage of attaining spherical symmetry in diseases such hereditary spherocytosis44, where mean cytoplasmic path-length increases due to shape and also MCHC tends to be higher due to cellular dehydration45,46. These circumstances predict a less complete gas exchange and only a partial manifestation of the Bohr effect at higher blood flows, which may contribute towards the reduced exercise tolerance commonly observed in patients with hereditary spherocytosis46.
Restricted diffusion in RBC cytoplasm may have important physiological implications besides gas exchange at capillaries. Slow diffusion of nitric oxide in RBC cytoplasm, in addition to the immobilizing effect of hemoglobin nitrosylation, may be a requirement for confining its signaling cascade to a sub-membranous domain47. A dense network of proteins may restrict gas diffusion in non-erythroid cell types or subcellular structures48. The magnitude of macromolecule-restricted diffusion highlights the potential for cytoplasm to regulate gas fluxes, a feat considered thus far to be in the remit of cell membranes only.
In conclusion, we propose that the efficiency of gas exchange depends critically on RBC thickness because the high density of hemoglobin restricts the movement of small solutes. This restriction is substantially greater in intact RBCs, compared to cell-free hemoglobin solutions or hemolysates, which explains why the cytoplasmic mobility of gases, HCO3− and H+ ions has previously been overlooked as a rate-limiting step in the gas exchange cascade. In addition to increasing the cell’s surface area/volume ratio for greater trans-membrane solute traffic and providing mechanical advantages for blood flow, the evolutionarily-conserved flattening of RBCs is an essential adaptation to minimize delays owing to slow cytoplasmic diffusion.
Methods
Red blood cells
Alpaca (Vicugna pacos) and chicken (Gallus gallus) blood were collected by trained staff from Seralabs (U.K.) and delivered on ice within 24 hours of collection. Xenopus laevis blood was collected by ventricular puncture of animals that have been sacrificed humanely by pithing in accordance with Schedule I of Animal (Scientific Procedures) Act 1986 carried out in a licensed facility at Oxford University. Human blood was obtained from two volunteers who gave formal and informed consent, in accordance with Central University Research Ethics Committee (CUREC) guidelines (reference: R46540/RE001, approved procedure #24), and data were fully anonymized and not traceable to the donor. All methods involving the use of blood were performed in accordance with ethical guidelines set by Oxford University, with appropriate risk assessments in place. All experimental protocols performed on collected human red blood cells are performed in accordance with Oxford University’s Central University Research Ethics Committee (CUREC) guidelines (reference: R46540/RE001, approved procedure #24). Additionally, all experimental protocols perform on animal and human red blood cells were performed in accordance with Oxford University’s Health and Safety regulations (OHS Policy Documents 1/03, 1/01, OHS 2/03, UGN S1/95). Clinical tests confirmed that mean corpuscular volume (MCV) and mean corpuscular hemoglobin concentration (MCHC) of donor blood were in the normal range (MCV: 88.2–90.9 fL, MCHC: 325–334 g/L). Blood samples were collected in tubes treated with heparin (Xenopus) or EDTA (all other species), spun-down at 4 °C (10,000 for 5 min for mammalian blood; 5,000 for 5 min for chicken blood; 1,000 for 10 min for Xenopus blood) to remove the supernatant and buffy coat. A sample of the fraction containing packed red blood cells was re-suspended in Hepes-buffered NT and used for confocal imaging or flow cytometry; the remainder was freeze-thawed twice to break up cell membranes, and used for measurements of hemoglobin concentration (HemoCue Hb 201plus), histidyl-containing small molecules (Pauly’s assay) and carbonic anhydrase activity.
Confocal imaging
To image intracellular pH (pHi), cells were loaded with the acetoxymethyl (AM) ester of the pH-reported dye cSNARF1 for 60 minutes (Xenopus) or 10 minutes (all other species), and allowed to settle on a poly-L-lysine pretreated coverslip at the base of a superfusion chamber mounted on an inverted Zeiss Axiovert microscope that was coupled to a Zeiss LSM700 confocal system. Superfusates were delivered at 4 mL/min at 25 °C (Xenopus) or 37 °C (all other species). Cells were imaged with a x100 objective and stored as stacks of 256 × 256 pixel 8-bit bitmaps (53 nm pixel length). cSNARF1 was excited by a 555 nm laser line and fluorescence was collected at 580 ± 10 and 640 ± 10 nm. The fluorescence ratio was converted to pH using a calibration curve, obtained by a published method20. To image calcein, cells were loaded with calcein-AM for 10 minutes, and fluorescence (>515 nm) was excited with low intensity 488 nm laser line. Offline analysis was performed using ImageJ.
H+ uncaging and calcein bleaching
Photolytic uncaging of H+ ions from 6-nitroveratralhyde23 (NVA; added to solutions at 1 mM) was evoked by scanning a region of interest (ROI) at one end of the RBC (ROI width equal to 1/10th of RBC diameter) with 405 nm laser light, as shown in Fig. 1C,D. By alternating between regional H+ uncaging and whole-field pHi-imaging, the spatial dissipation of H+ ions from the uncaging site can be followed. Fluorescence maps were analyzed using ImageJ. Two stacks of images (580 nm and 640 nm fluorescence) were background-offset and fluorescence signal was averaged in ten ROIs of width equal to 1/10th of the RBC diameter along x-axis and height equal to the RBC diameter in the y-axis. For Xenopus and chicken, the fluorescence from the nucleus was excluded from analysis. ROI1 represented the uncaging region. The ratio of ROI fluorescence at 580 and 640 nm was converted to [H+] time courses, which were fitted with a diffusion equation to obtain the apparent diffusion coefficient (DHapp), according to an algorithm described previously26 and in the Supplement. Bleaching of calcein fluorescence was performed by a similar protocol to H+ uncaging. The excitation protocol alternated between high intensity laser for localized calcein bleaching and low intensity 488 nm laser for imaging cytoplasmic calcein. Time courses of calcein fluorescence, measured in ten ROIs, were fitted with a diffusion equation, similar to that described in the Supplement for analyzing DHapp.
Additional Information
How to cite this article: Richardson, S. L. and Swietach, P. Red blood cell thickness is evolutionarily constrained by slow, hemoglobin-restricted diffusion in cytoplasm. Sci. Rep. 6, 36018; doi: 10.1038/srep36018 (2016).
References
Wagner, P. D. Diffusion and chemical reaction in pulmonary gas exchange. Physiol Rev 57, 257–312 (1977).
Geers, C. & Gros, G. Carbon dioxide transport and carbonic anhydrase in blood and muscle. Physiol Rev 80, 681–715 (2000).
Endeward, V. & Gros, G. Extra- and intracellular unstirred layer effects in measurements of CO2 diffusion across membranes–a novel approach applied to the mass spectrometric 18O technique for red blood cells. J Physiol 587, 1153 1167 (2009).
Fedosov, D. A., Pan, W., Caswell, B., Gompper, G. & Karniadakis, G. E. Predicting human blood viscosity in silico. Proc Natl Acad Sci USA 108, 11772–11777 (2011).
Uzoigwe, C. The human erythrocyte has developed the biconcave disc shape to optimise the flow properties of the blood in the large vessels. Med Hypotheses 67, 1159–1163 (2006).
Chien, S. Red cell deformability and its relevance to blood flow. Annu Rev Physiol 49, 177–192 (1987).
Noguchi, H. & Gompper, G. Shape transitions of fluid vesicles and red blood cells in capillary flows. Proc Natl Acad Sci USA 102, 14159–14164 (2005).
Lim, H. W. G., Wortis, M. & Mukhopadhyay, R. Stomatocyte-discocyte-echinocyte sequence of the human red blood cell: evidence for the bilayer- couple hypothesis from membrane mechanics. Proc Natl Acad Sci USA 99, 16766–16769 (2002).
Grebe, R. & Zuckermann, M. J. Erythrocyte shape simulation by numerical optimization. Biorheology 27, 735–746 (1990).
Guest, M. M., Bond, T. P., Cooper, R. G. & Derrick, J. R. Red Blood Cells: Change in Shape in Capillaries. Science 142, 1319–1321 (1963).
Canham, P. B. The minimum energy of bending as a possible explanation of the biconcave shape of the human red blood cell. J Theor Biol 26, 61–81 (1970).
Pivkin, I. V. et al. Biomechanics of red blood cells in human spleen and consequences for physiology and disease. Proc Natl Acad Sci USA 113, 7804–7809 (2016).
Holland, R. A. & Forster, R. E. The effect of size of red cells on the kinetics of their oxygen uptake. J Gen Physiol 49, 727–742 (1966).
Vandegriff, K. D. & Olson, J. S. Morphological and physiological factors affecting oxygen uptake and release by red blood cells. J Biol Chem 259, 12619–12627 (1984).
Jones, D. A. The important of surface area/volume ratio to the rate of oxygen uptake by red cells. J Gen Physiol 74, 643–646 (1979).
Gros, G. & Moll, W. The diffusion of carbon dioxide in erythrocytes and hemoglobin solutions. Pflugers Arch 324, 249–266 (1971).
Spitzer, K. W., Skolnick, R. L., Peercy, B. E., Keener, J. P. & Vaughan-Jones, R. D. Facilitation of intracellular H(+) ion mobility by CO(2)/HCO(3)(-) in rabbit ventricular myocytes is regulated by carbonic anhydrase. J Physiol 541, 159–167 (2002).
Swietach, P., Spitzer, K. W. & Vaughan-Jones, R. D. pH-Dependence of extrinsic and intrinsic H(+)-ion mobility in the rat ventricular myocyte, investigated using flash photolysis of a caged-H(+) compound. Biophys J 92, 641–653 (2007).
Stewart, A. K., Boyd, C. A. & Vaughan-Jones, R. D. A novel role for carbonic anhydrase: cytoplasmic pH gradient dissipation in mouse small intestinal enterocytes. J Physiol 516 (Pt 1), 209–217 (1999).
Swietach, P. et al. Hydrogen ion dynamics in human red blood cells. J Physiol 588, 4995–5014 (2010).
Irving, M., Maylie, J., Sizto, N. L. & Chandler, W. K. Intracellular diffusion in the presence of mobile buffers. Application to proton movement in muscle. Biophys J 57, 717–721 (1990).
Junge, W. & McLaughlin, S. The role of fixed and mobile buffers in the kinetics of proton movement. Biochim Biophys Acta 890, 1–5 (1987).
Hulikova, A. et al. Stromal uptake and transmission of acid is a pathway for venting cancer cell-generated acid. Proc Natl Acad Sci USA (2016).
Bohr, C., Hasselbalch, K. & Krogh, A. Über einen in biologischer Beziehung wichtigen Einfluss, den die Kohlensäurespannung des Blutes auf dessen Sauerstoffbindung übt. Skand Arch Physiol 16, 401–412 (1904).
Swietach, P. & Vaughan-Jones, R. D. Relationship between intracellular pH and proton mobility in rat and guinea-pig ventricular myocytes. J Physiol 566, 793–806 (2005).
Swietach, P. et al. Modelling intracellular H(+) ion diffusion. Prog Biophys Mol Biol 83, 69–100 (2003).
Gros, G. et al. Evidence for rotational contribution to protein-facilitated proton transport. Proc Natl Acad Sci USA 81, 1710–1714 (1984).
Vanysek, P. In CRC Handbook of Chemistry and Physics (ed D. R. Linde ) Ch. 5, 93–95 (CRC Press, 1999).
Seely, J. E. & Marshall, F. D. Carnosine levels in blood. Experientia 37, 1256–1257 (1981).
Richards, M. et al. Intracellular tortuosity underlies slow cAMP diffusion in adult ventricular myocytes. Cardiovasc Res 110, 395–407 (2016).
Van Balgooy, J. N., Marshall, F. D. & Roberts, E. Carnosine in nucleated erythrocytes. Nature 247, 226–227 (1974).
Cass, A. & Dalmark, M. Equilibrium dialysis of ions in nystatin-treated red cells. Nat New Biol 244, 47–49 (1973).
Wang, D., Kreutzer, U., Chung, Y. & Jue, T. Myoglobin and hemoglobin rotational diffusion in the cell. Biophys J 73, 2764–2770 (1997).
Cooper, G. J., Occhipinti, R. & Boron, W. F. CrossTalk proposal: Physiological CO2 exchange can depend on membrane channels. J Physiol 593, 5025–5028 (2015).
Dauty, E. & Verkman, A. S. Molecular crowding reduces to a similar extent the diffusion of small solutes and macromolecules: measurement by fluorescence correlation spectroscopy. J Mol Recognit 17, 441–447 (2004).
Dix, J. A. & Verkman, A. S. Crowding effects on diffusion in solutions and cells. Annu Rev Biophys 37, 247–263 (2008).
Rummer, J. L., McKenzie, D. J., Innocenti, A., Supuran, C. T. & Brauner, C. J. Root effect hemoglobin may have evolved to enhance general tissue oxygen delivery. Science 340, 1327–1329 (2013).
Riggs, A. F. The Bohr effect. Annu Rev Physiol 50, 181–204 (1988).
Shelby, J. P., White, J., Ganesan, K., Rathod, P. K. & Chiu, D. T. A microfluidic model for single-cell capillary obstruction by Plasmodium falciparum-infected erythrocytes. Proc Natl Acad Sci USA 100, 14618–14622 (2003).
Tomaiuolo, G. & Guido, S. Start-up shape dynamics of red blood cells in microcapillary flow. Microvasc Res 82, 35–41 (2011).
Tomaiuolo, G. Biomechanical properties of red blood cells in health and disease towards microfluidics. Biomicrofluidics 8, 051501 (2014).
Fedosov, D. A., Noguchi, H. & Gompper, G. Multiscale modeling of blood flow: from single cells to blood rheology. Biomech Model Mechanobiol 13, 239–258 (2014).
van der Harst, P. et al. Seventy-five genetic loci influencing the human red blood cell. Nature 492, 369–375 (2012).
Reinhart, W. H. & Chien, S. Red cell rheology in stomatocyte-echinocyte transformation: roles of cell geometry and cell shape. Blood 67, 1110–1118 (1986).
Cynober, T., Mohandas, N. & Tchernia, G. Red cell abnormalities in hereditary spherocytosis: relevance to diagnosis and understanding of the variable expression of clinical severity. J Lab Clin Med 128, 259–269 (1996).
Perrotta, S., Gallagher, P. G. & Mohandas, N. Hereditary spherocytosis. Lancet 372, 1411–1426 (2008).
Pawloski, J. R., Hess, D. T. & Stamler, J. S. Export by red blood cells of nitric oxide bioactivity. Nature 409, 622–626 (2001).
Waisbren, S. J., Geibel, J. P., Modlin, I. M. & Boron, W. F. Unusual permeability properties of gastric gland cells. Nature 368, 332–335 (1994).
Acknowledgements
Supported by a Royal Society fellowship to P.S. We thank Colin Akerman and Mala Rohling for supplying Xenopus blood, and Alzbeta Hulikova for assistance with Pauly’s assay.
Author information
Authors and Affiliations
Contributions
S.L.R. performed and analyzed experiments. P.S. designed the research and wrote the manuscript.
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Electronic supplementary material
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Richardson, S., Swietach, P. Red blood cell thickness is evolutionarily constrained by slow, hemoglobin-restricted diffusion in cytoplasm. Sci Rep 6, 36018 (2016). https://doi.org/10.1038/srep36018
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/srep36018
- Springer Nature Limited
This article is cited by
-
Assessment of transient changes in oxygen diffusion of single red blood cells using a microfluidic analytical platform
Communications Biology (2021)