Expression of HbC and HbS, but not HbA, results in activation of K-Cl cotransport activity in transgenic mouse red cells

K-Cl cotransport is elevated in humans with sickle cell disease that have homozygous hemoglobin (Hb) S (SS), homozygous HbC (CC), and in young human erythrocytes.^1-4  In SS disease, the density-isolated, reticulocyte-rich fraction has high K-Cl cotransport.⁵  This observation led to the hypothesis that the elevated K-Cl cotransport observed in SS results from the presence of young red cells.⁵  Elevated K-Cl cotransport correlates well with increased reticulocyte counts in SS and homozygous HbA (AA) but not in CC patients.⁶  K-Cl cotransport in CC red cells is comparable in magnitude to SS red cells while the reticulocyte counts are lower in CC than SS. Therefore, younger red cell age alone does not explain the elevation of K-Cl cotransport in CC erythrocytes and may not be a sufficient explanation in SS.

Elevated K-Cl cotransport in erythrocytes and particularly the reticulocytes of sickle cell disease patients has been implicated in the generation of very dense and dehydrated erythrocytes that contribute to the disease process by accelerating polymer formation. A subset of red blood cells from sickle cell anemia patients has been shown to have a very active volume- and pH-stimulated K-Cl cotransport cotransport activity.^7-9  Elevated K-Cl cotransport may account, at least in part, for the heterogeneous red cell density distribution observed in SS patients.

CC disease is characterized by uniformly elevated red cell mean corpuscular hemoglobin concentration (MCHC),^10-12  in contrast to SS disease in which dense cells represent only a fraction of all cells. The mechanism by which this occurs is still unclear, but elevated activity of K-Cl cotransport in these cells has long been recognized as a potential candidate. The elevation of MCHC caused by HbC has strong pathogenic consequences for patients who express both HbS and HbC and who therefore have SC disease. Sickle trait (AS) is clinically benign, but the elevation in MCHC in SC disease contributes to a clinical picture that is similar to that of sickle cell disease.^13,14 

The interaction of mutant hemoglobins with the red cell membrane and its components may serve as a model for regulation of K-Cl cotransport. It has long been recognized that HbA, HbS, and HbC bind to the red cell membrane and that the amount of Hb bound is in the order of C > S > A. Based on the effect of ionic strength on binding, it was proposed that the excess of HbS and C was due to the increasing positive charge in the series A < S < C.^15,16  In 1973, Steck¹⁷  demonstrated that the reversible binding of both HbA and HbS was competitive with that of glyceraldehyde-3-phosphate dehydrogenase (GAPD), an enzyme that binds to the N-terminal of the anion exchanger (AE1 or band 3), and that HbS can displace HbA. Reiss et al¹⁸  demonstrated that HbC also competes with GAPD when it binds to inside-out vesicles (IOVs) and that the same amount of GAPD can displace more HbA than HbC. These observations are consistent with known negative charge on the N-terminal of AE1 that has been shown to bind in the central cavity of hemoglobin.¹⁹ 

Olivieri et al²⁰  suggested that the positive charge at the β6 position found in both HbS and HbC activates K-Cl cotransport. In support of this hypothesis, Brugnara et al²¹  have shown evidence that suggests that hemoglobin S interacts with and activates a volume-dependent K-Cl cotransport cotransport in white ghosts prepared from either AA or SS red cells. More recently, Nagel et al²²  observed that erythrocytes containing HbO_Arab, which has an increased positive charge at the β121 position, have elevated volume-stimulated K-Cl cotransport, implying that the positive charge need not be at the β6 position. Furthermore, Gibson et al²³  have recently reviewed the evidence that the ligand state of hemoglobin affects K-Cl cotransport activity in many types of red cells. These results provide strong evidence to suggest that hemoglobin or a component thereof interacts with the cotransporter or one of its regulators. However, in humans, assessing the role of hemoglobin in K-Cl cotransport activation is complicated by the concomitant elevation of reticulocytes that often accompanies hemoglobinopathies. Hence, the extent to which hemoglobin alone may play a role in regulating erythrocyte potassium metabolism is unknown.

In the early 1990s, a series of first-generation transgenic mice were produced that express human α, β^S, and residual mouse globins.^24-28  A small volume-stimulated K-Cl cotransport was observed in erythrocytes from these mice. K-Cl cotransport in these sickle transgenic mice had a shorter delay time for activation than that observed in blood from the C57BL control mice.²⁹  This yielded an apparent increase in K-Cl cotransport maximum velocity (V_max) in transgenic mice over C57BL. However, other properties of K-Cl cotransport in these mice were significantly different from those seen in other mammalian erythrocytes: The magnitude of the volume-stimulated K-Cl cotransport was significantly smaller than that observed in SS patients even after correction for surface area. Furthermore, the putative mouse K-Cl cotransport showed anion characteristics that were distinctly different from those observed in human, rabbit, or sheep red cells. Isosmotic replacement of Cl^- by NO₃^- caused a significant stimulation of K⁺ efflux in mouse erythrocytes in contrast to the diminished efflux observed in human, sheep, or rabbit erythrocytes.^29,30  When Cl^- was replaced by the anion sulfamate, mouse red cells showed a very significant Cl^--dependent K⁺ efflux. However, although both the isotonic and hypotonic fluxes versus sulfamate were large, the volume-stimulated K⁺ efflux was small³⁰  and, as noted by Armsby et al, anion sensitivity was altered.³⁰ 

More recently, Fabry et al^31,32  developed mice expressing human HbC and HbS in the absence of mouse globins. These mice have many of the hematologic and pathophysiologic features found in their respective hemoglobinopathies. In the mice expressing HbS, reticulocytosis, anemia, loss of urine concentrating ability, low plasma arginine, and multiorgan pathology ameliorated by the presence of HbF were found.^32,33  In mice expressing HbC, dehydrated red cells (high MCHC), high-density reticulocytes, and intracellular crystals were found.³¹ 

The objective of the present work was to examine the effect of hemoglobin in situ on K-Cl cotransport in mouse red cells that express human hemoglobins. We examined K-Cl cotransport in 3 lines of knock-out (KO) mice expressing exclusively human hemoglobins: mice with HbA, HbC, and NY1KO mice that express HbS and 3 different levels of HbF.³²  Incorporation of human HbC or HbS into mouse red cells was associated with elevated K-Cl cotransport activity that had properties similar to those previously described in human red cells with HbC or HbS. These results suggest that in intact red cells HbC or HbS interacts through some unknown mechanism with the K-Cl cotransporter or its regulator(s). They also have implications for interpretation of protocols in which transporters are inserted into the membrane of unrelated species that may not have the same hemoglobins or other regulatory proteins.

The generation of NY1 mice expressing cointegrated miniLCRα2 and miniLCRβ^S constructs was previously described by Fabry et al³⁴  and in a more recent publication.³²  Briefly, NY1 mice were generated by coinjecting 2 constructs: (1) an 8 kilobase (kb) miniLCR (locus control region) and (2) the same miniLCR ligated to the SphI-XbaI fragment containing the β^S globin gene, which then cointegrated. The γH construct was generated by Gilman.³⁵  These constructs have been described in detail in a previous publication.³² 

NY1 mice were bred onto a background of C57BL/6J mice, and the mouse α- and β-globin KOs that have been backcrossed onto C57BL from 7 to 11 generations were bred in. The mouse α-globin KO was obtained from Pàszty and coworkers.³⁶  The mouse β-globin KO used for the NY1KO mice was obtained from Yang et al.³⁷ 

Generation of HbCKO mice was previously described.³¹  The miniLCRα2 construct previously described²⁸  was coinjected into fertilized CBA/B6 mouse eggs with the miniLCRβ^C construct that was exactly like the original β^S construct generated by Costantini.²⁸  The 2 transgenes cointegrated, resulting in mice that always express both human α and β^C. Three founders were obtained: 2 expressing high levels of human α and β^C (56% α^H, 34% β^C) and 1 expressing a lower level (21% α^H, 14% β^C). The low-expressing line did not transmit its gene to subsequent generations. Except where noted, all results presented here are from one of the lines of high-expressing animals that was designated the 500000 line. HbC mice were originally created on a CBA/B6 background and backcrossed onto C57BL for between 4 and 6 generations. K-Cl cotransport was measured at frequent intervals since creation of the mice (from founders to 4 to 6 generations of backcrossed mice), and no change in magnitude, anion sensitivity, or other properties was observed, which implies that migration from CBA to C57BL did not affect K-Cl cotransport properties and that the current cotransport properties are not modified by the original CBA background. Large variations in K-Cl cotransport for individual animals might be expected if a CBA gene were influencing K-Cl cotransport, but variability was not observed.

HbAKO mice were generated by R. Kumar and obtained from C. Pàszty. They express the α-globin knock-out generated by C. Pàszty, and the β-globin KO was generated by Ryan et al³⁸  and Pàszty et al.³⁹ 

All animals expressing either HbS or HbC with α KO and β KO were maintained on “sickle chow” developed by C. Pàszty without added arginine. It was obtained from Purina as diet no. 5740C. The mice had access to Nestlets nesting material. For this paper, we studied NY1KO, HbAKO, HbC-L, and HbCKO mice (Table 1). Approximately 10 different mice of each type were used for these experiments. No mouse was bled more than once a month. Blood samples were collected from a tail incision in heparinized mouse saline (330 mOsm). Table 1 lists a full description of each type of mouse.

Mice were bled from the tail (with a 2-hour recovery period under 40% oxygen) using protocols approved by the animal studies committee of the Albert Einstein College of Medicine. Blood samples were analyzed for reticulocytes and red cell indices using the Sysmex SE 9000 system (Toa, Kobe, Japan). Manual counts after staining with new methylene blue were used to validate the Sysmex reticulocyte counts in a limited number of cases, and good agreement was found. Blood smears were made from blood obtained from the tail and were dried, fixed, and stained with Giemsa. The mean corpuscular hemoglobin concentration was measured in plasma by measurement of hematocrit (MicroHematocrit; Damon/IEF Division, Needham Heights, MA) and hemoglobin concentration by diluting with Drabkins reagent and measuring the optical density at 540 nm.

K-Cl cotransport activity in mouse red cells was measured as described previously by us.²⁹  Briefly, we determined the volume-stimulated and Cl^--dependent K⁺ efflux from mouse red cells by incubating cells at 1% hematocrit (Hct) in isotonic (330 mOsm) and hypotonic (250 mOsm) media. Net K⁺ efflux was started by addition of red cells into prewarmed flux media. The media contained (mM) the following: (a) NaCl 150 (isotonic Cl^-); (b) NaCl 115 (hypotonic Cl^-); (c) NaNO₃ 115 (hypotonic NO₃^-); (d) NaSulfamate (SFM) 115 (hypotonic SFM). All media contained (mM) the following: 1 ouabain, 1 MgCl₂ or 1 Mg(NO₃)₂, 0.01 bumetanide, 10 glucose, 10 sucrose, and 10 Tris-MOPS (tris(hydroxymethyl)aminomethane– 3-[N-Morpholino]propanesulphonic acid) at pH 7.4 or pH 7.0 at 37°C. Samples in duplicates at 0, 5, 10, 25, 40, and 60 minutes were taken and pipetted into 1.5 mL ice-cold Eppendorf tubes containing 0.4 mL dibutylphthalate oil (d = 1.04 g/mL) and centrifuged for 10 seconds in a Fisher microcentrifuge (model 235C; Pittsburgh, PA). The supernatant was removed for K⁺ determination by atomic absorption spectrophotometry. The K⁺ efflux was calculated from the nonlinear regression analysis of the K⁺ concentration versus time and the Hct of the cells in the flux media. In some cases, hemolysis was measured at each time point by adding a 50 μL sample of the flux medium to 50 μL Drabkins reagent and measuring the optical density at 540 nm. With the exception of solutions containing N-ethylmaleimide (NEM), hemolysis was found to make a negligible contribution to extracellular K⁺. Therefore, hemolysis was not measured for all samples. Cl^--dependent K⁺ efflux (K-Cl cotransport activity) was estimated by subtracting the flux in NO₃^- or SFM media (as indicated) from that in Cl^- media. Volume-stimulated K⁺ efflux was estimated by subtracting the flux in isotonic media from that in hypotonic; pH-stimulated K⁺ efflux was estimated by subtracting the flux in pH 7.4 media from that in pH 7.0 media.

Reticulocyte depletion was accomplished as follows. Red cells were first depleted of white blood cells by passage through an α-cellulose column (Fisher Scientific, Suwannee, GA) and labeled with rat antimouse CD71-biotin, 0.1 μg/10⁶ cells, (Southern Biotech, Birmingham, AL) and incubated for 10 minutes on ice; they were then reacted with streptavidin magnetic microbeads (Miltenyi, Auburn, CA) and incubated in a refrigerator for 15 minutes. Labeled reticulocytes were removed by passage through several LS magnetic columns (Miltenyi). The initial reticulocyte count of approximately 15% was reduced to approximately 3% by this procedure.

Red cell densities were examined on Percoll (Sigma, St. Louis, MO)/Larex (Larex, White Bear Lake, MN) gradients as previously described.^9,40 

K-Cl cotransport has been shown to be volume and pH stimulated. It is experimentally defined as the difference between Cl^--dependent K⁺ efflux into hypotonic versus isotonic media or isotonic pH 7.0 versus isotonic pH 7.4. K-Cl cotransport activity was measured in KO mice expressing exclusively human hemoglobins: mice with HbA, HbC, and HbS+γ. These mice are described genetically and physiologically in Tables 1 and 2, respectively. The study of K-Cl cotransport suffers from the lack of a strong, specific inhibitor that has efficacy in all species. Because of this and, in contrast to our previous convention, all results are reported as total K⁺ efflux unless otherwise stated. The results of these experiments are as follows.

We measured K⁺ efflux into media that contained Cl^-, NO₃^-, or SFM as the anion in media buffered at either pH 7.0 or 7.4. The results of a representative experiment using combined red cells of 3 HbAKO mice are shown in Figure 1A. We find that HbAKO mice have a K⁺ efflux that is only slightly increased when isotonic Cl^- media are replaced by hypotonic Cl^- media (7.3 mmol/L cells × h [FU] to 10.5 FU, respectively) or when pH 7.4 media are replaced by pH 7.0. K⁺ efflux was higher when Cl^- was replaced by NO₃^- in hypotonic media as reported in previous studies of early sickle transgenic lines or C57BL. All characteristics of K-Cl cotransport are similar in C57BL, early sickle transgenic lines that express mouse α- and β-globins, and HbAKO mice that express exclusively HbA.

Figure 1B shows a representative experiment using combined red cells from 3 NY1KO-γH mice that express human HbS + 40% γ. K⁺ efflux into hypotonic media was large when compared with isotonic media (20.5 versus 11.3 FU). The hypotonic flux was reduced to 14.8 FU by NO₃^- and to 16.7 FU by dihydroindenyl-oxy-alkanoic acid (DIOA). This is in contrast to the previously reported findings for red cells from NY1DD mice expressing 56% human α, 75% β^S, and residual mouse globins.²⁹  The NY1DD mice, like the HbAKO mice shown in Figure 1A, had a small volume-stimulated K-Cl cotransport cotransport and were insensitive to DIOA and stimulated by NO₃^-. NY1DD mice were bled to increase the reticulocyte count to 50%, but the rate of K⁺ efflux in these high reticulocyte mice following volume stimulation was the same as that for low reticulocyte NY1DD mice; however, the previously reported delay time for induction of K⁺ efflux²⁹  was found to be shortened (data not shown) in high reticulocyte mice. In the NY1KO-γH mice, we also studied the effect of pH 7.0 on K⁺ efflux and observed a significant Cl^--dependent, pH-stimulated K⁺ efflux (Figure 1B) that was similar to that observed in human red cells.

To clarify the role of reticulocytes in NY1KO mice, we studied K-Cl cotransport activity in NY1KO mice with high γ (40%), medium γ (20%), or low γ (more than 3%) that have 12%, 30%, and 63% reticulocytes, respectively. We found strong volume dependence for all 3 types of mice (Table 2). Although there was a 5-fold variation in the percent reticulocytes between NY1KO-γL and NY1KO-γH mice (63% versus 12%, respectively), the change in volume-stimulated K-Cl cotransport was only 1.3-fold (11.4 versus 8.5 FU, respectively), which suggests that reticulocytes play a smaller role in K-Cl cotransport in the mouse than in humans. This is consistent with the shorter red cell life span found in mice. This evidence is strongly suggestive, but not conclusive, that the enhanced K-Cl cotransport can be attributed to the presence of HbS.

To more completely characterize the contribution of reticulocytes to K-Cl cotransport activity, we depleted whole blood from NY1KO-γH mice of CD71⁺ cells, which are primarily reticulocytes, measured volume-stimulated K-Cl cotransport, and estimated sensitivity to NO₃^- and sulfamate. The final reticulocyte count was reduced from about 12% to approximately 3%. The volume-stimulated K-Cl cotransport in this cell preparation was 4.8 FU. Isosmotic replacement of Cl^- by sulfamate completely eliminated volume-stimulated K⁺ efflux. In addition, replacement of Cl^- by NO₃^- reduced the volume-stimulated K⁺ efflux by 83%. Therefore, the red cell preparation devoid of CD71⁺ cells exhibits a volume-stimulated, NO₃^--sensitive, and Cl^--dependent K⁺ efflux.

Figure 1C shows a representative experiment using combined red cells from 2 HbCKO mice that express exclusively human HbC. K⁺ efflux into hypotonic media was large when compared with isotonic media (20.1 versus 10.2 FU). The hypotonic flux was reduced to 12.7 FU by NO₃^- and to 14.5 FU by DIOA. These mice had a strong volume-stimulated K-Cl cotransport activity that was similar to that observed in NY1KO mice. A similar and even larger effect was observed for K-Cl cotransport activity measured at pH 7.0.

To eliminate the possibility that elevated reticulocyte count contributes to these results, we studied K-Cl cotransport activity in partial knock-out mice (HbC-L) that have low (3% to 5%) reticulocyte counts and express 68% human α, 51% β^C, and residual mouse globins. Figure 2 summarizes the volume-stimulated component (Δ hypotonic Cl^- versus isotonic Cl^-) in 4 different mouse types. We found a strong volume-dependent K⁺ efflux in HbC-L mice that was not significantly different than that observed in HbCKO mice but was significantly greater than C57BL and HbAKO.

DIOA and okadaic acid have been used as inhibitors of K-Cl cotransport cotransport activity in human red cells. We measured K⁺ efflux into hypotonic Cl^- or isotonic pH 7.0 media in the presence or absence of either DIOA (0.01 mM) or okadaic acid (100 nM); 100 nM okadaic acid has been shown to give maximal inhibition in mouse red cells.³⁰ Figure 3A shows a summary of our results. DIOA and okadaic acid significantly blocked K⁺ efflux into hypotonic Cl^- media in HbC and HbS. A similar and more pronounced effect was seen when K⁺ efflux was stimulated by isotonic pH 7.0 media (Figure 3B).

In mammalian red cells, K-Cl cotransport is partially inhibited by DIOA and more strongly inhibited by okadaic acid. Figure 4 shows the total K⁺ efflux into hypotonic media and summarizes our results in C57BL and HbCKO mice. As seen in Figure 1 and “The HbAKO mouse has weak volume-stimulated K-Cl cotransport activity” for HbAKO mice, hypotonic NO₃^- supports a K⁺ efflux in C57BL red cells and a smaller K⁺ efflux in HbCKO that is similar to the flux we and others have previously reported in early sickle transgenic lines that express mouse α- and β-globins and in C57BL mice.^29,30  We further observed that K⁺ efflux in the presence of hypotonic NO₃^- was not significantly affected by either DIOA or okadaic acid in either C57BL or HbCKO mouse red cells. Isosmotic replacement of NO₃^- by sulfamate reduced K⁺ efflux in both cell types, and in both cell types the K⁺ efflux in hypotonic chloride in the presence of okadaic acid was significantly larger than the K⁺ efflux in hypotonic sulfamate (8.5 ± 1.2 and 4.5 ± 1.0 FU, respectively, P < .04). The latter observation is in agreement with that of Armsby et al³⁰  who previously reported that sulfamate inhibits more of the isotonic and hypotonic K⁺ efflux than okadaic acid in C57BL and CD1 mice.

The effect of mutant hemoglobins (such as HbC or HbS) on K-Cl cotransport activity has been complex to unravel, because the presence of mutant hemoglobins is often associated with shortened red cell life span and elevated reticulocyte counts. Reticulocytes clearly have elevated K-Cl cotransport,^5,43  and mature human red cells have extremely low K-Cl cotransport activity due to 2 factors: (1) K-Cl cotransport is lost as the red cell matures from the reticulocyte stage,^5,44  and (2) the lifetime of normal human red cells is longer than for most other mammalian red cells. This also results in a very low percent of reticulocytes and young red cells. Therefore, K-Cl cotransport in the human red cell is more sensitive to the presence of reticulocytes than is the case for most other mammalian red cells. Elucidation of the relative contribution of the mutant hemoglobin and shortened red cell life span is further complicated by the presence of young red cells that have lost the characteristic markers of reticulocytes but still have elevated K-Cl cotransport and other metabolic markers; furthermore, these cells cannot be easily enumerated.⁹ 

The overall results of the experiments described here show that red cells from transgenic mice that express exclusively human HbC or HbS+HbF have a K-Cl cotransport activity that closely resembles that observed in human red blood cells. In addition, the interaction of hemoglobin with K-Cl cotransport is much stronger in the case of HbC than HbS. We base these conclusions on the following considerations.

Early transgenic mice, such as the NY1DD mouse, that express human α and β^S as well as murine α- and β-globins showed evidence for the presence of a K-Cl cotransporter albeit much smaller in magnitude and with properties (such as altered anion sensitivity, insensitivity to DIOA, and a high basal flux under isotonic conditions) that differ from those observed in human and other mammalian red blood cells.^29,30  Bleeding NY1DD mice to induce reticulocytosis did not alter the properties of K-Cl cotransport. When human hemoglobin A (HbA) is introduced into a full knock-out mouse (HbAKO, a transgenic mouse in which both the mouse α- and β-globins have been knocked out) that expresses exclusively human HbA, K-Cl cotransport features are unaltered from those observed in control (C57BL) and NY1DD mice.

We can therefore conclude that the presence of murine globins is not required to elicit the characteristic properties of mouse K-Cl cotransport and that HbA does not alter the properties of mouse K-Cl cotransport.

When human HbC or HbS and HbF are introduced into full KO mice producing either HbCKO mice or NY1KO mice, the properties of K-Cl are dramatically altered: The volume-stimulated component increases to 7.8 and 8.9 FU, respectively; pH stimulation increases to 14.7 and 13.4 FU, respectively; and both become sensitive to NO₃^- and DIOA (Figure 1B-C). Not only is there a quantitative change in the magnitude of volume stimulation, but there is also a change in both anion and inhibitor sensitivity. In the case of HbC, this effect is clearly not dependent on the presence of reticulocytes. The case for HbS in whole blood of mice that express exclusively HbS and HbF is not as clear-cut because all of these mice have relatively high reticulocyte counts; however, in reticulocyte-depleted blood from NY1KO-γH mice, we found strong volume stimulation and sensitivity to NO₃^-, okadaic acid, and DIOA. These results support the contention that HbS as well as HbC modify the characteristics of mouse K-Cl cotransport.

We can conclude that the presence of HbC or HbS+HbF activates K-Cl cotransport possibly by interacting more strongly than the mouse globin either with the K-Cl cotransporter itself or with a regulatory protein.

HbCKO mice have a relatively high reticulocyte count; therefore, HbC-L mice that have lower reticulocyte counts and express about 50% β^C and 50% β^mouse were examined to definitively rule out a role for red cell age in the altered properties observed in mouse red cells expressing mutant hemoglobins. We have previously demonstrated that elevating the reticulocyte count in NY1DD mice does not restore the magnitude of K-Cl cotransport.²⁹ 

In contrast to NY1DD mice, partial knock-out HbC mice that still express mouse globins (HbC-L) have elevated volume-stimulated K-Cl cotransport and sensitivity to NO₃^- and DIOA. These observations are significant for 2 reasons: (1) HbC-L mice have a lower reticulocyte count than HbCKO mice and yet they still exhibit increased K-Cl cotransport and inhibitor sensitivity, and (2) these observations demonstrate that HbC can successfully alter the properties of mouse K-Cl cotransport in the presence of mouse globins, whereas HbS cannot.

We hypothesized that if K-Cl cotransport contributed in whole or in part to the NO₃^--stimulated K⁺ efflux into hypotonic media, this flux would be sensitive to either DIOA or okadaic acid, 2 known inhibitors of the K-Cl cotransport activity.^45,46  We found that the NO₃^--stimulated K⁺ efflux into hypotonic media in C57BL and HbCKO red cells was not inhibited by DIOA or okadaic acid (Figure 4). Armsby et al suggested that the differences in anion sensitivity seen in mouse might arise from intrinsic species differences in anion selectivity of the K-Cl cotransporter or from a mouse-specific alternate K⁺ efflux pathway activated by NO₃^-.³⁰  The ability of HbS and HbC to restore sensitivity to NO₃^- and the lack of sensitivity of the NO₃^--mediated flux to DIOA and okadaic acid favor the latter alternative.

The K⁺ efflux in response to volume and pH changes observed for C57BL mouse red cells and transgenic mice expressing HbS and mouse globins (small volume and pH response and insensitivity to DIOA and NO₃^-) differs from those observed in other mammalian red cells. The presence of human HbC or HbS activates K-Cl cotransport and restores these properties. This could originate in a number of ways: (1) Mouse K-Cl cotransport, or its regulatory elements, may differ from that found in other mammalian red cells preventing the characteristic response to volume and pH stimulation. This alternative is clearly ruled out by restoration of these properties by expression of HbS in the absence of mouse globins or HbC in their presence or absence. (2) The observed activity in C57BL red cells may be due to another isozyme or an entirely different transporter. Several isoforms of K-Cl cotransporters (KCCs) exist in other cell types,^47-49  and several isoforms, KCC1-4, have been reported in some erythrocytes as well.⁵⁰ (3) K-Cl cotransport may be present but inactive in the absence of a positively charged Hb, such as HbS and particularly HbC.

We speculate that hemoglobin binding to the N-terminal of AE1 can serve as a model for this last mechanism. A positively charged hemoglobin would have an activating effect, and hence HbA and HbF, which are more negatively charged than mouse hemoglobins, would have no effect. The positively charged, activating hemoglobin would first have to displace any competing mouse globins and, in this competition, the higher positive charge of HbC and subtle differences in the central cavity may render it a more efficient competitor than HbS.

We find that the presence of HbS+γ, in the absence of mouse globins, and HbC, even in the presence of mouse globins, has significant quantitative and qualitative effects on K-Cl cotransport in mouse red cells. We conclude that both mouse and human globins are able to affect activity and/or regulation of K-Cl cotransport in mouse red cells and that positively charged human hemoglobins can activate K-Cl cotransport in mouse red cells. These observations suggest that when red cell ion transporters are inserted into the membrane of other types of cells, there may be a lack of hemoglobins or other regulatory proteins necessary to elicit the full range of activity characteristic of the transporter in its native setting. Finally, the K⁺ efflux stimulated by NO₃^- probably represents a second, yet to be defined transporter that is particularly active in mouse red cells.

The presence of an active K-Cl cotransporter with humanlike qualities in cells from KO HbS and HbC mice allows the testing of hypotheses aimed at clarifying the role K-Cl cotransport plays in the pathophysiology of HbC and sickle cell disease.