Sergey V. Rudenko, PhD, Igor A. Zupanets, PhD, ScD, Sergey K. Shebeko, PhD
Human red blood cells (RBCs) when suspended in a Low Ionic Strength medium (LIS) demonstrate characteristic triphasic shape changes (morphological response, MR) and become reduced in volume. Tahitian Tabari Noni juice (Tb), after being given during the terminal phase of MR, was shown to initiate an unusual cell response. This response can be described as a volumetric four-phasic response including frst a shrinking phase, attributed to the initial sucrose-induced shrinkage during typical MR, a rapid frst swelling phase, induced by application of the juice, followed spontaneously by the occurrence of a more prolonged second shrinking phase, which culminated in the swelling and hemolysis of the cells. All the phases of volumetric response can be independently regulated by chloride, DIDS, cations Ca2+, Ag+, Hg2+ or plasma. The second shrinking phase is not inhibited by clotrimazole, a known inhibitor of Gardos channels, and can be replicated by a mixture of two ionophores (valinomycin and CCCP), suggesting the involvement of the putative K+/H+ exchanger as a mechanism of this phase. We suggest that the erythrocyte membrane is equipped with additional molecular systems, poorly characterized at present, that regulate the cell shape and volume. The cell should, therefore, be considered as an “active” responsive system instead of a “passive” osmometer-like structure.
HBS, HEPES buffered saline; PS, physiological saline; RBC, red blood cell; MR, morphological response; SI, shape index; SSS, standard sucrose solution; LIS, low ionic strength; NSVDC, non-selective voltage dependent channel; OUW, osmotically unresponsive water; AcZA, acetozolamide; BzA, benzalkonium chloride; CCCP, carbonyl cyanide m-chlorophenylhydrazone; Tb, Tabari Noni juice; VR, volumetric response; VRS, volume responsive system; SRS, shape responsive system; RVD, regulatory volume decrease; RVI, regulatory volume increase; CAII, carboanhydrase II; AE1, anion exchanger 1 (band 3); FSHF, frst shrinking phase; FSWF, frst swelling phase; SSHF, second shrinking phase; SSWF, second swelling phase.
Under regular physiological conditions, normal RBCs have a biconcave discoid shape. The shape corresponds to broad variations in the cell volume ranging from 70 to 110 ﬂ in humans. The frequency distribution of the volume and surface area of the circulating erythrocytes of healthy donors, as well as the distribution of the intracellular hemoglobin concentration, are more or less normal (Gaussian) [1,2]. The cells tend to maintain a constant area to volume ratio that is likely to conserve the ﬂexibility and ability to adequately perform their oxygen-transport function, while moving through the blood vessels. However, in order to do that the cell must control its original isotonic volume. Mathematical models  and other considerations  show that the coordinated work of some membrane systems including pumps like Na+-K+-ATPase and Ca2+-ATPase, Cl-/HCO3- exchanger (AE1) and Ca2+-activated K+ channel (Gardos channel) prevent cell swelling caused by an increase in the permeability of the main osmoregulating cations, K+ and Na+.
Many cell types were shown to regulate the volume (regulatory volume decrease (RVD) and regulatory volume increase (RVI) ) and intracellular osmolarity through the activation of several mechanisms including some transporters [6-8] and molecular crowding [9-12]. However, unlike other cells, including the nucleated red cells, mature human red cells do not demonstrate classical regulatory volume responses. If they are swollen or shrunken osmotically, they do not necessarily regain their initial volume and behave like nonideal osmometers . Only the reticulocytes and young erythrocytes exhibit RVD in the hypotonic media through the activation of KCl-cotransport [14-16]. This is interpreted as a lack (or inactivation) of the effector and sensing systems responsible for the RVD in the mature erythrocytes .
However, because the RBCs do not restore their original isotonic volume in an anisosmotic medium it does not imply that the volume is not regulated, because complete or incomplete recovery depends upon the operation of the putative volume regulatory systems involved, as was described for the regulation of the intracellular pH in some cells . A different mode of operation of the acid loaders and extruders can lead to complete or partial recovery of intracellular pH after extracellular or intracellular pH interventions, depending upon suitable conditions (for reviewer see ). As suggested in , the regulation of intracellular pH and RBC volume are mutually interdependent. It is possible to assume, that the absence or incomplete volume recovery under anisosmotic conditions, could involve a molecular mechanism similar to that which controls the intracellular pH.
The best way to investigate RBC volume regulation is to use a LIS solution, where NaCl is substituted isosmotically by sucrose, and where the nature of the cell reaction is not well understood. In the preliminary experiments we found that in the LIS medium Tb initiated a reproducible unusual volumetric response of the RBCs. Morinda citrifolia (Noni) is a natural product containing microelements, 3-Methyl-1,3-butanediol, iridoid glucosides, scopoletin, rutin, fatty acid glucosides, and anthraquinones . It has been reported to have broad therapeutic effects, including anticancer activity , liver protective activity , and immunomodulatory effects . The action of Tb on human RBCs has never been explored earlier.
The aim of the present study was to give new insights into the basic processes of the regulation of RBC shape and volume in the LIS medium. We explored the RBC shape and volume changes during the morphological response in the standard LIS induced by different conditions. We conclude that the RBC volume responses cannot be explained as responses of a passive osmometer-like structure. Therefore, the existence of an “active” volume regulatory or responsive system (VRS) is assumed.
N-(2-Hydroxyethyl)piperazine-N’-2-ethanesulfonic acid (HEPES) and 4,4’-diisothiocyanato-2,2'-stilbene-disulphonic acid (DIDS) was purchased from Serva, and sucrose from Merck. All other reagents, AgNO3, CaCl2, HgCl2, AlCl3, were obtained from major suppliers and were of the highest purity available. The stock solution of benzalkonium chloride (BzA; pharmaceutical mixture, consisting of a mixture of alkylbenzyldimethylammonium chlorides of various evennumbered alkyl chain lengths, Fluka, Germany) was prepared with concentration of 0.4 mg/ml. Pure Tahitian Tabari Noni juice Morinda Citrifolia (Joy Products, S.A., Costa Rica) was fltered through 0.22 mm millipore flter and concentration of solids (50 mg/ml) was determined by dry weight after evaporation. In standard experiment, when 50 ml of juice was added to 2 ml of cell suspension, the fnal concentration of juice at weight basis was 1.25 mg/ml.
Preparation of Erythrocytes
Human blood was withdrawn in vacuum vacutainer tubes with EDTA as anticoagulants on the day of the experiment from informed volunteers. After free blood sedimentation 0.225 ml of the red blood cell pellet was transferred into and washed once in 5 ml of physiological saline (PS) (150 mM NaCl, 3,000 g, 2 min) and then once again in 5 ml of HEPES buffered saline HBS (150 mM NaCl, 5 mM HEPES, pH 7.4, 3000 g, 1 min). Excess supernatant after the centrifugation was removed to leave ~1.5 ml fnal volume of red blood cell stock suspension with hematocrit about 17% which, if not indicated otherwise, was allowed to stay for 1-2 hours at room temperature before measurements to avoid the initial drift in cell properties .
Measurement of dynamics of RBC shape and volume changes
Morphological response of the cells suspended in a standard sucrose solution (0.3 M, pH 5.8-6.2 without buffer, SSS) was monitored using homemade SA-01 shape-meter/aggregometer which allows, besides measuring optical density or light transmission, assessment of light ﬂuctuations carrying information on cell shape . The shape index (SI) was calculated following the previously described protocol [25,26] from the equation: SI= k·D, where k is a constant depending on amplifcation factor and meter calibration and D is a mean value of light amplitude ﬂuctuations, averaged in 1 s intervals. The calibration factor k allows formation of SI scale reﬂecting the erythrocyte discoidal–spherical shape factor (1.0 for discs and 0.06 for spheres). Erythrocytes (6-8 µl of stock suspension) were placed into a cylindrical (diameter 10 mm) glass cell containing 2 ml of HBS so that the initial optical density was 0.30 ± 0.01, which corresponds to cell concentration of 6·106/ml. The cell suspension was stirred on a magnetic stirrer at 600 rpm.
Dynamics of volume and shape changes in the SSS induced by Tb
Despite the fact that Tb is a complex mixture, in the present study we used Tb as an agent inducing reproducible volume changes in the RBCs. The data shown in Fig. 1 illustrate typical examples of complex volume and shape changes in the cells induced by Tb during the terminal phase of MR. This unusual response can be simply described as a volumetric four-phasic response (VR) with the frst shrinking phase (FSHF) attributed to the initial sucrose-induced shrinkage during typical MR in SSS, then a rapid frst swelling phase (FSWF) due to Tb application followed by a spontaneously occurring more prolonged second shrinking phase (SSHF) and then a second swelling phase (SSWF) ultimately leading to hemolysis of the cells.
Similar to MR, VR represents a variable process. It depends on the blood sample, composition of the medium and exhibits day-to-day variations. Variations relate largely to the rates and magnitudes of the different phases of VR (compare tracks 1 and 2 in Fig. 1), although some phases could be totally absent in particular cases. As seen in Fig. 1, a typical sequence of volume changes is accompanied by different dynamics of shape changes. It shows that volume- and shape-regulating mechanisms in the RBCs are largely separated.
The effect of chloride, calcium, and DIDS on volumetric response
Fig. 2A shows that an increase in the chloride concentration in the SSS eliminates the FSHF and reduces the rate of the subsequent phases of the VR, ultimately inhibiting them. The process does not depend on the cations present (Na+ or K+) suggesting the role of the anions in the effect. Accordingly, in the physiological salines (PS or HBS), VR becomes reduced to Tbinduced FSWF, while the other phases are absent.
The data presented in Fig. 2B show that VR depends on the Tb concentration and it can be signifcantly stimulated by Ca2+. At high Tb concentration the Ca2+ does not affect the FSWF; however, both the SSHF and SSWF are greatly accelerated when the Ca2+concentration is increased. Relatively high (5 mM) concentrations of the extracellular Ca2+ requires more than one order of magnitude lower concentration of Tb to initiate the highly visible SSHF and SSWF (Fig. 2B, V3). Tb is unable to initiate FSWF indicating that this phase directly depends upon its concentration. In the absence of the extracellular Ca2+ low concentrations of Tb too did not initiate FSWF, but inhibited the FSHF and MR similar to the Ca2+ effect (data not shown). Interestingly, the VR similar to MR is not identical in solutions prepared from different sucrose batches. As illustrated in Fig. 3, VR in sucrose A does not contain SSHF and SSWF, whereas VR in sucrose B does (Fig. 3, V1).
However, the missing phases can be initiated by DIDS in sucrose A (Fig. 3, V2). If they do exist, these phases are greatly stimulated, just as in sucrose B (Fig. 3, V1, track 2). Therefore, an activation effect of DIDS is similar to that of Ca2+(compare Fig. 2B, V2, section 2 and Fig. 3, V1, section 2, track 2). The dynamics of the corresponding shape changes visualized by the SI tracks (the latter accompanied these VRs) in sucroses A and B was different. In sucrose A the cells reached the minimal volume in the disks, whereas in sucrose B they were converted into spheres. A pronounced difference between the time courses of volume and shape changes that are reported here, strongly suggest that the cell volume and shape in LIS are regulated by independent mechanisms. The data confrm the previous observations in physiological salt solutions .
The effect of water channel inhibitors on volumetric response
As changes in cell volume are related to water transport, we tested the effect of two water channel inhibitors Ag+ and Hg2+  on VR. These data are listed in Fig. 4.
Synergistic effect of calcium, mercury, and the ionophores valinomycin and CCCP on volumetric response
Data presented in Fig. 5 demonstrate the synergistic effect of the Ca2+ and Hg2+ in the activation of VR. In a mixture with Tb, these cations induced rapid VR including all the phases. The minimal volume in the SSHF approaches 0.6, whereas typically the value is greater than 0.7 (Figs. 2 and 3).
It is obvious, that an increase in the shrinkage is due to the activation effect of the Hg2+. As pronounced shrinkage was observed during the SSHF, we assume that the process is accompanied by the K+ efﬂux, contrary to the FSHF where the cell shrinkage is attributed presumably to the Cl-/OH- exchange, with minimal release of K+. Indeed, the addition of the two ionophores, valinomycin and CCCP induced the same effect as Ca2+ and Hg2+(Fig. 5, V2, section 1). When the cells were injected in the valinomycin and CCCP-containing medium, the rate and extent of the FSHF were enhanced and the effect of Tb was modifed which in turn resulted in the elimination of the FSWF and SSHF (but not SSWF). These results show that the SSHF can be mimicked by a combination of valinomycin and CCCP (which imitate K+/H+ exchanger). We speculate that the components of the Tb in the presence of Ca2+ and Hg2+ stimulate a putative internal K+/H+ exchanger in the red blood cell membrane. Experimental pathways shown in Fig. 5 (V3) demonstrate that the components of the plasma can be used as an inhibitor of the SSWF. A minimal quantity of the plasma signifcantly reduced the rate of SSWF, induced by a combination of Tb, valinomycin and CCCP (Fig. 5, V3, section 1). When added prior to valinomycin and CCCP, the plasma signifcantly reduced the SSHF rate and completely eliminated the SSWF within 10 min (Fig. 5, V3, section 2). Taken together the data reveal that the different phases of the VR can be regulated (activated or inhibited) by specifc chemicals.
Osmotic response of red blood cells
Common view is that human red cells respond to changes in the tonicity of the external media like nonideal osmometer that follows modifed Van’t Hoff law in terms of the cell volume asfollows:
where V 0 and C0 are initial cell volume and osmolarity of the medium, W – volume fraction of cell water in normal isosmotic conditions (0.7 ) and R – an empirical coeffcient to ft to experimental results . For ideal perfect osmometer R=1, for human erythrocyte the magnitude for R is expected to be ~0.5-0.7 [13,33,34]. Some reports confrm linearity of osmotic behavior of red cells [33-35]. Others suggest the overall nonlinear osmotic behavior [13,36-38]. We previously reported that rehydrated swollen erythrocytes can lose partially or completely a nonideal osmotic behavior depending on conditions [39,40]. Osmotic behavior of red cells described by formula (1) depends on normal water fraction W which is varied in different blood samples. It is important to access: i) how these variations modify the relationship between cell volume and inverse osmolarity and ii) whether experimentally measured volume is different from that predicted by non-linear osmotic model if normal variation in water fraction are taking into account. This comparison is shown in Fig. 6.
We assume that statistically the normal water content in the red cells can vary from 0.65 to 0.75 with mean of 0.7, which corresponds to much of the published data [2,31,41,42]. Two lines in this plot (corresponding to W=0.65 and W=0.75) defne an expected volume range at a given external osmolarity. The data show that the osmotic response of the native RBC deviates from a theoretical expectation, as some of the experimental values are statistically different from the theoretical ones (p<0.05). However, the cells adopted their volume to the exact theoretical values in the same medium containing 2 mM DIDS (Fig. 6, closed symbols). The conclusion, therefore, is that a genuine property of the human red cells is to be an ideal perfect osmometer with R=1 similar to a hemoglobin-flled giant vesicle. The fact that the osmotic behavior of the human red cell, in principle, can be well described as R=1, suggests that the osmotically unresponsive water (OUW) content in the RBC is very low and all the water participates as a solute to dissolve the intracellular solids, hemoglobin, metabolites and ions. Theoretically, one cannot rule out the possibility that under different experimental conditions part of free water could become bound to proteins, thus producing asymmetric water ﬂows [36,38] in response to hypertonic or hypotonic conditions. If that is so, then the explanation of the DIDS effect is that the DIDS somehow (perhaps utilizing a crowding principle) prevents any changes in the amount of OUW during an osmotic challenge.
Osmotic response of red cells in sucrose low chloride solution is more complicated compared with physiological salt saline. At these conditions the cells exhibit spontaneous shrinkage (FSHF) and re-swelling after addition of hyperosmotic or isosmotic NaCl . Due to this complex response it is clear that the fnal steady-state volume will not correspond to that of ideal osmometer. However, if RBCs are indeed osmometers, it is expected that the initial fast shrinkage caused by an abrupt increase in tonicity is close to reaction of giant vesicle. To verify this we compared measured and predicted values of osmoticallyinduced volume excursions. Excursion values were calculated using expression which takes into account a variation of cell water fraction, if it took place before osmotic treatment in the form:
where V 0 and C0 are initial cell volume and osmolarity of the medium, V1 – new cell volume in isosmotic conditions before osmolarity changes, C – current osmolarity, and W – volume fraction of cell water in normal isosmotic conditions. If V 1=Vo, this expression converts into formula (1). Fig. 7 shows
dependencies of predicted cell volume under given fnal osmolarity C as a function of initial relative volume.
As seen in the Fig.7, much of the data do not correspond to theoretical predictions. Depending upon the experimental conditions, true volume can be more or lesser than expected. For example, the resultant volume of the cells is smaller than expected in DIDS or the Tb-containing sucrose solution and is larger in the presence of plasma or BzA. These opposite deviations cannot be explained assuming R is less than unity because this parameter may correct the data only in one direction. Taken together, the data demonstrate that the hypertonic osmotic response of the RBC cannot be described by a nonideal osmotic model. Thus, our data extend the prior observations that the actual shape and volume of the red cells critically depend upon the external environment [43-48]. In the LIS medium this effect is even more highly pronounced and can shed light on the mechanism that causes it to happen.
Obviously, the modifed Van’t Hoff equation (formula 1) can be applied only for cells which do not demonstrate active regulatory volume responses like RVD and RVI. Our data indicate that the human red blood cells undergo pronounced re-swelling, resembling classical RVI in the hyperosmotic LIS medium and slightly reduce the volume in the hyposmotic ones. The mechanism of these responses is expected to be different from that of RVD and RVI because they also occur in an isosmotic medium .
The complex nature of the volume regulation mechanisms in the RBC is illustrated in the phenomenon described here, namely, the multi-phasic Tb-induced volumetric response (Fig. 1). Each phase in this response is a subject for regulation. For example, the SSHF and SSWF are inhibited by the chloride (Fig. 2A) or plasma (Fig. 5, V3) and stimulated by Ca2+, Ag+, low doses of Hg2+ and DIDS (Figs. 2-5).
The most conservative phase in this response (FSWF) is inhibited by the higher doses of Hg2+ (Figs. 2-5) and by the mixture of the CCCP and valinomycin (Fig. 5, V2, section 2). This is probably due to fact that the Hg2+ as well as the CCCP plus valinomycin per se stimulate cell shrinkage . The VR modulators can inhibit either both the adjacent phases or inhibit the one and activate the other phase in this sequence (Fig. 5). These observations suggest that the different phases in the VR are being regulated independently. The different pharmacology (sensitivity to DIDS) and kinetics of the FSWF and Cl--induced re-swelling  prove that they have a distinct molecular background. The weak DIDS sensitivity of the FSWF suggests that this process does not involve AE1 as part of the underlying mechanism. In contrast, the re-swelling is strongly inhibited by the DIDS indicating its possible involvement. The participation of the aquaporins is not excluded in view of the inhibitory action of the Hg2+; however, a stimulatory action of the other water channel inhibitor, Ag+ , makes this suggestion controversial. Overall, it is not clear how the RBC can swell in the medium containing only the impermeant sucrose and less than 1 mM NaCl in a manner comparable to that in the physiological salt saline (Fig. 2A). Probably, the cells utilize a special mechanism mainly related to the water transport and to a lesser extent to the ion transport.
Our data show that the SSHF can be mimicked by a combination of the CCCP and valinomycin (Fig. 5). This forms the basis to suggest that the cell shrinkage during the SSHF is due to the operation of the putative K+/H+ exchanger in the RBC membrane. The SSHF could be also explained as a result of the activation of the K+-selective Gardos channel demonstrated for E.coli a-hemolysin . We found that clotrimazole (inhibitor of this channel) [17,51] had no effect on the SSHF. Clotrimazole was reported to be an activator of NSVDC . On the other hand, there is no clear indication whether or not this channel, if it exists, is DIDS sensitive . Thus, the involvement of the Gardos and NSVDC channels in the SSHF appears to be less likely. The pharmacology of this exchanger is different from the K+/H+ exchanger described earlier by Bernhardt and colleagues [53-56]. For sake of simplicity we designate their exchanger as BKHE and our putative exchanger as RKHE. The BKHE is inhibited by the DIDS  and operates rather slowly, releasing the K+ efﬂux, to about 60 mmol/lcells/h [55,57]. In contrast, the RKHE is strongly activated by the DIDS, Ca2+ and Hg2+ and proceeds much faster leading to a complete K+ release within 20-40 s, as indicated by the volume dynamics. These major differences suggest that the BKHE and RKHE may represent a different molecular structure in the RBC membrane. Whether the BKHE and RKHE are identical or different warrants further examination.
The VR is terminated by the last SSWF leading to cell hemolysis. This can be due to the nonspecifc damage of the RBC membrane, now permeable for the ions and nonelectrolytes like sucrose. An expected characteristic of such type of damage is a lack of inhibitors for colloid osmotic swelling and hemolysis, particularly if the process is already in progress. On the other hand, if such inhibitors exist, this will imply that the process is a regulated one and is not totally related to the nonspecifc membrane damage. Data obtained show a modulatory effect of the cations, DIDS, and plasma (Fig. 3 and Fig. 5, V3). This indicates that the SSWF can be regulated as well, and hence does not appear to be a nonspecifc damage of the RBC membrane.
All the data given above demonstrate that the multi-phasic VR is a precisely regulated process. This conclusion is highly evident from a comparison of the VRs in the two types of sucrose in which the cells demonstrate different dynamic volumetric trajectories. On mixing sucrose A and B in different proportions we got sequential transitions between the pattern characteristic for sucrose A alone into the pattern characteristic for sucrose B through many intermediate patterns (not shown). This illustrates that the actual VR dynamics (trajectory) critically depends upon the amount and proportion of each constituent (including inhibitors and activators) affecting one or the other phase of the VR.
Taken together, our results can uniformly explain why measured volume changes in the red blood cells in an anisosmotic medium in some cases are well described but in many other cases are different from that predicted for the ideal osmometer with an R value of unity [13,34,35]. In a simple case of mixed NaCl/ sucrose media, the inconsistency arises from the re-swelling phase –a clearly visible volume and shape restoring process which is activated immediately after the osmotic challenge . We speculated earlier that the native red blood cell is equipped with mechanisms which can be preliminarily designated as the Volume Responsive System (VRS) and Shape Responsive System (SRS). These systems are able to sense and respond to the changes in the cell environment making it “active” in this sense . It is clear that plotting this data in the Van’t Hoff coordinates without considering the operation of the VRS, helps us obtain a huge number of dependencies. Depending on different conditions, these plots can be ideal or non-ideal, linear or nonlinear. The reason for this lies in the different modes of operation in the responsive systems under the given circumstances. In light of this fnding, the broad scattering of published results related to the connection between cell volume and tonicity [11,33-35,38-40] is not surprising. Our data suggest that the shape and volume of the red blood cells after transition in the new environment should be considered as a unique state, critically dependent upon the prior maneuvers and operation mode of the putative responsive systems. This notion is also confrmed by the data reported in  and  showing the possible existence of an adenosine-dependent regulatory volume decrease in the red cells in the hypotonic salt saline, as well as our own result presented in Fig. 7.
In summary, the envelope of data indicates that the erythrocyte membrane possesses responsive systems, poorly characterized at the moment, regulating the shape and volume responses during changes in the cell environment. The cell, therefore, should be considered rather as an “active” responsive object instead of a “passive” osmometer-like structure. The challenge is to understand the functional principles and molecular arrangement of these systems and their role in the red blood cell physiology.
S. V. Rudenko et al. / International Journal of BioMedicine 3(2) (2013) 104-111