INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

TWO FAMILIES of Na⁺/Ca²⁺ exchange proteins have been described in mammalian tissues (16, 19): the cardiac-type Na⁺/Ca²⁺ exchanger, which has a stoichiometry of 3 Na:1 Ca (12, 18), and a K⁺-dependent exchanger, expressed principally in retinal rods, which has a stoichiometry of 4 Na⁺:(1 Ca²⁺ + 1 K⁺) (4, 23). Three different genes for the cardiac-type exchanger (NCX1, NCX2, and NCX3) have been described. Cardiac Na⁺/Ca²⁺ exchangers are abundantly expressed in the heart and are also found in lower amounts in brain, kidney, smooth muscle, and skeletal muscle (15, 20). In contrast, the tissue distribution of mRNA for the retinal rod exchanger (NCKX1) was originally thought to be limited to the retinal photoreceptors (19). Subsequently, however, transport studies with synaptosomal membranes provided evidence for both K⁺-dependent and K⁺-independent exchange activity (6). Recently, a second member of the K⁺-dependent exchanger family (NCKX2), 55% identical in amino acid sequence to NCKX1, was cloned from rat brain (28).

We have previously reported that human platelets showed K⁺-dependent Na⁺/Ca²⁺ exchange activity (13). Repeated attempts to identify the exchanger in platelets by immunologic or PCR techniques were unsuccessful. In this report, we show that K⁺-dependent Na⁺/Ca²⁺ exchange activity is also present in a human megakaryocytic cell line. We have utilized PCR techniques with mRNA from this cell line to show that the exchanger expressed in megakaryocytes is identical to the human retinal rod exchanger. mRNA encoding this exchanger was also shown to be present in human platelets. Finally, we present evidence that the rate of Na⁺/Ca²⁺ exchange in human platelets is dependent on the electrochemical gradient for not only Na⁺ but also K⁺ across the plasma membrane.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell culture. Cells from a human megakaryocytic cell line, CHRF-288 (31), were generously provided by Drs. G. W. Dorn II and M. Lieberman (Department of Internal Medicine, Molecular Genetics, Biochemistry and Microbiology, University of Cincinnati, Cincinnati, OH). The cells were cultivated at 37°C, 5% CO₂-95% air, in either RPMI medium supplemented with 10% fetal bovine serum or Fisher's medium supplemented with 20% horse serum. Both growth media contained 100 U/ml penicillin G plus 100 µg/ml streptomycin. Cells were passed by dilution to ~0.5 × 10⁸ cells/ml in fresh medium every 3-4 days.

Platelet isolation. Platelets were isolated by differential centrifugation as described previously (13, 14). Briefly, platelet-rich plasma was obtained by centrifugation at 200 g for 10 min at room temperature. After treatment with 0.1 mmol/l aspirin for 20 min, platelet-rich plasma was centrifuged at 1,000 g and platelets were washed two times in a buffer consisting of (in mmol/l) 140 NaCl, 5 KCl, 10 glucose, 0.3 EGTA, and 10 HEPES and one time in the same buffer plus 0.1% BSA. Washed platelets were kept in Ca²⁺-free buffer.

Monitoring of cytosolic Ca²⁺ and Na⁺. In preparation for cytosolic Ca²⁺ (Ca²⁺_i) and Na⁺ (Na⁺_i) monitoring, CHRF-288 cells were washed twice with HEPES buffer containing (in mmol/l) 140 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 10 HEPES, and 10 glucose plus 0.1% BSA. Cells were incubated for 30 min at 37°C with 5 µmol/l fura 2-AM (Molecular Probes, Eugene, OR) in HEPES buffer. The extracellular dye was removed by centrifugation. Ca²⁺_i measurements were performed in HEPES buffer at 37°C under constant stirring in SPEX Fluoromax (SPEX Industries, Edison, NJ). Excitation wavelengths were set at 340 and 380 nm, and emission wavelength was set at 505 nm. Calibration of the fura 2 traces in terms of Ca²⁺_i were carried out as described by Grynkiewics et al. (9). Ratios of minimum and maximum fluorescence intensities were determined by the addition of 1 mmol/l CaCl₂ and 20 µmol/l digitonin followed by 10 mmol/l EGTA (final pH 8.5). Autofluorescence was determined at the end of each experiment by the addition of 1 mmol/l MnCl₂ and 20 µmol/l digitonin. Calculations of Ca²⁺_i were performed as previously described (13, 14).

Na⁺/Ca²⁺ exchange activity measurements using Ba²⁺ influx in platelets. To measure exchange activity in platelets, Ba²⁺ was employed as a substrate for Ca²⁺ using a modification of the procedures described by Condrescu et al. (5). Washed platelets were loaded with fura 2 and 1 mmol/l Ca²⁺ for 30 min. Platelets were then treated with 5 µmol/l monensin, 5 µmol/l nigericin, and 0.1 mmol/l ouabain for 5 min in Ca²⁺-free buffer containing different concentrations of Na⁺ and K⁺. Treatment was terminated by the addition of 1% BSA for 1 min. Cells were washed and resuspended in Ca²⁺-free HEPES buffer containing different concentrations of Na⁺, K⁺, and Li⁺. BaCl₂ (1 mmol/l) was then added and fura 2 fluorescence monitored (excitation 350 and 390 nm; emission 510 nm). Autofluorescence was measured at the end of each experiment by the addition of 2 mmol/l MnCl₂ and 20 µmol/l digitonin. Because absolute values of the 350/390 ratios tended to vary with different platelet preparations, data are presented as percentages of the maximal increase in the 350/390 ratio observed in each experiment.

RT-PCR. Two pairs of primers were used in the present study. The first pair was based on the partial identity between the bovine retinal rod exchanger (NCKX1) and the rat brain K⁺-dependent exchanger (NCKX2): AACGTGGGCATTGGCACC (F1) and GAGCTKGACACAGCCATGTC (R1), where K = G or T. The sequences correspond to the positions 1,756-1,773 and 3,528-3,509 in the bovine retinal rod exchanger. A second pair of primers was specific for the human retinal rod exchanger [ATCTACCAGCTCATGCTCC (F2) and TCCATCTTCTACACCTTTCTC (R2)] and corresponded to positions 1,957-1,975 and 2,499-2,479, respectively. mRNA was isolated from CHRF-288 cells using MessageMaker Reagent Assembly (Life Technology, Gaithersburg, MD). Total RNA from human platelets and CHRF-288 cells was isolated using Trizol (Life Technology). RT-PCR was performed using the SuperScript One-Step RT-PCR System (Life Technology). Products were analyzed on 1% agarose gels. Bands of expected size (1,200-1,600 bp) were extracted (QIAquick Spin, Qiagen, Valencia, CA), subcloned (TOPO TA Cloning, Invitrogen, Carlsbad, CA), and sequenced (ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit, Perkin-Elmer, Foster City, CA).

Northern blot analysis. mRNA from CHRF-288 cells was electrophoresed on 1% agarose-formamide gels and transferred overnight to nylon membranes by capillary diffusion. The ultraviolet-cross-linked membranes were hybridized with digoxigenin-UTP-labeled riboprobe according to the manufacturer's instructions (Boehringer Mannheim, Indianapolis, IN). The probe was made using 544-1,074 cDNA of human NCKX1.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

K⁺-dependent Na⁺/Ca²⁺ exchange activity in CHRF-288 cells. Cells (grown in RPMI with 10% fetal bovine serum) were treated with 0.2 mmol/l ouabain for 3 h. Ouabain treatment increased basal Na⁺_i and Ca²⁺_i (Table 1). Ouabain treatment also increased the ionomycin-induced Ca²⁺_i rise in Ca²⁺-free medium (reflecting Ca²⁺ release from intracellular Ca²⁺ stores) (Fig. 1A).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effect of acute (A) and chronic (B) treatment with ouabain on ionomycin-induced Ca2+ release from Ca2+ stores in CHRF-288 cells. A: CHRF-288 cells were treated with 0.2 mmol/l ouabain for 3 h. B: CHRF-288 cells were cultured in Fisher's medium with or without 25 nmol/l ouabain for 72 h. Fura 2-loaded cells were suspended in Ca2+-free HEPES buffer, and 5 µmol/l ionomycin was added as indicated by arrow. Con, control cells; Oua, ouabain-treated cells; Ca2+i, cytosolic Ca2+. Values are means ± SE (SE is denoted by vertical deflections in composite of ionomycin-evoked Ca2+ signals); n = 3.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Composites of K+-dependent changes in cytosolic Ca2+ (Ca2+i) in fura 2-loaded CHRF-288 cells without (A) or with (B) ouabain treatment. A: traces represent 4 separate protocols in which cells were suspended in 140 mmol/l LiCl + 5 mmol/l KCl, 145 mmol/l LiCl, 140 mmol/l NaCl + 5 mmol/l KCl, or 140 mmol/l NaCl + 5 mmol/l LiCl. Ca2+ (1 mmol/l) was added as indicated by arrow. The 4 traces were essentially indistinguishable from one another. B: cells were treated with 0.2 mmol/l ouabain for 3 h and suspended in 140 mmol/l LiCl + 5 mmol/l KCl [Li,K(+)], 145 mmol/l LiCl [Li,K()], 140 mmol/L NaCl + 5 mmol/L KCl [Na,K(+)], or 140 mmol/l NaCl + 5 mmol/l LiCl [Na,K()]. Ca2+ (1 mmol/l) was added as indicated by arrow. Traces for Na,K(+), Na,K(), and Li,K() were essentially indistinguishable from one another. Values are means ± SE (SE is denoted by vertical deflections); n = 3.

Identification of the exchanger in CHRF-288 cells and platelets. With the use of poly(A)⁺ RNA from CHRF-288 cells, RT-PCR with primers F1 and R1 yielded several bands in the 1.2- to 1.6-kb region. The DNA in this region was subcloned as described in METHODS, and two products at 1.45 and 1.6 kb were sequenced. The 1.6-kb band did not display an open reading frame and was not characterized further. The 1.45-kb product was found to be 100% identical to the sequence of a splice variant of the human retinal rod NCKX1 (AF062921). This fragment codes for amino acids extending approximately from transmembrane segment 2 to transmembrane segment 8 and thus includes the entire central hydrophilic domain of the exchanger. The splice variant for the human retinal rod exchanger has an 18-amino acid insertion (KPGDGAIAVDELQDNKKL) at amino acid residue 629 of the originally published sequence (AF026132) (29); the insertion occurs shortly after the fifth transmembrane segment of the exchanger at the beginning of the large hydrophilic domain. A similar sequence, with 14 identical amino acids, is found in a similar position (amino acids 634-641) in the bovine retinal rod exchanger (19). To determine whether the retinal rod exchanger is also expressed in human platelets, we performed RT-PCR using total RNA from platelets and primers F2 and R2, based on the human retinal rod sequence (see METHODS). As shown in Fig. 3, a band of the expected size (543 bp) was obtained, confirming that human platelets express mRNA for the human retinal rod exchanger.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   RT-PCR analysis of mRNA for retinal rod exchanger (NCKX1) in platelets and CHRF-288 cells. Lane 1, DNA ladder; lane 2, 500 ng of total RNA from platelets; lane 3, 250 ng of total RNA from CHRF-288 cells.

Northern blot results of mRNA from CHRF-288 cells. A riboprobe was constructed on the basis of the cDNA sequence for the human retinal rod exchanger between positions 544 and 1,074. Using this probe, we showed a major transcript band of ~5.8 kb and two minor bands of ~4.9 and ~6.8 kb (Fig. 4). The two minor bands, as well as the 5.8-kb band, were also found using another riboprobe (positions 1,564-3,017 in the cDNA sequence). The presence of the two bands suggests alternative splicing, incomplete processing, or multiple polyadenylation signals.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Northern blot analysis of NCKX1 mRNA in CHRF-288 cells. mRNA (2 µg) was loaded onto the well.

Effect of external and cytosolic Na⁺ and K⁺ concentrations on Na⁺/Ca²⁺ exchange activity in human platelets. Because the Michaelis-Menten constant (K_m) for K⁺ is ~1 mmol/l in platelets (13) and retinal rods (23), the K⁺-binding site on the exchanger is likely to be saturated at both the cytoplasmic and extracellular membrane surfaces under physiological conditions. In experiments with retinal rods (22, 24) and proteoliposomes containing the purified exchanger (7), Ca²⁺ movements were shown to be directly linked to the K⁺ gradient despite the use of K⁺ concentrations that greatly exceeded the K_m. To examine this issue in platelets, we monitored exchange activity after using ionophores to establish various transmembrane gradients of Na⁺ and K⁺ (see METHODS). For these experiments we used Ba²⁺ as a substitute for Ca²⁺ to minimize possible complications caused by organellar sequestration of the transported cation (5).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Effect of changing external and cytosolic Na+ and K+ concentrations on Na+/Ca2+ exchange activity in human platelets (expressed as a percentage of Na+/Ca2+ exchange activity at the largest Na+/K+ gradient). Composites in A and C depict Ba2+ uptake by platelets. Ba2+ was added at time 0. A: cytosolic Na+ (Na+i) and K+ (K+i) concentrations were clamped at 140 and 5 mmol/l, respectively. External Na+ (Na+o) and K+ (K+o) concentrations (in mmol/l) were as follows: a, 0 Na+, 145 K+; b, 10 Na+, 135 K+; c, 35 Na+, 110 K+; d, 70 Na+, 75 K+; e, 105 Na+, 40 K+; and f, 140 Na+, 5 K+. B: rates of Ba2+ influx in A plotted vs. Na+o concentrations. C: Na+o was maintained at 10 mmol/l and K+o at 135 mmol/l while Na+i and K+i concentrations were alternated. Na+i and K+i concentrations (in mmol/l) were as follows: a, 145 Na+, 0 K+; b, 140 Na+, 5 K+; c, 125 Na+, 20 K+; d, 85 Na+, 60 K+; and e, 10 Na+, 135 K+. D: rates of Ba2+ influx in C plotted vs. Na+i concentrations. Note that K+ gradient was equal but opposite to Na+ gradient in these experiments. Values are means ± SE (SE is denoted by vertical deflections in A and C and error bars in B and D); n = 3.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Effect of changing K+ concentration on Na+/Ca2+ exchange activity in human platelets (expressed as percentage of Na+/Ca2+ exchange activity at largest gradients). Composite in A depicts Ba2+ uptake by platelets. Ba2+ was added at time 0. Na+i and K+i concentrations were clamped at 140 and 5 mmol/l, respectively. A: Na+o concentrations were maintained at 0 mmol/l (except for f). K+o and Li+o concentrations (in mmol/l) were as follows: a, 145 K+, 0 Li+; b, 110 K+, 35 Li+; c, 75 K+, 70 Li+; d, 40 K+, 105 Li+; e, 5 K+, 140 Li+; and f, 140 Na+, 5 K+. B: rates of Ba2+ influx in A plotted vs. K+o concentrations. Values are means ± SE (SE denoted by vertical deflections in A and error bars in B); n = 3.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In our previous studies with human platelets, we showed that Na⁺/Ca²⁺ exchange activity is K⁺ dependent with a K_m for K⁺ of ~1 mmol/l (13). The present work, showing the presence of K⁺-dependent Na⁺/Ca²⁺ exchange activity in platelet precursors, i.e., megakaryocytes, complements our previous finding in platelets. Demonstrating the K⁺ dependence of the Na⁺/Ca²⁺ exchanger in CHRF-288 cells was an essential step in identifying the exchanger type. Platelets are anucleated cells with rudimentary RNA content. CHRF-288 cells thus provided the RNA for our initial screening for, and eventual identification of, the retinal rod Na⁺/Ca²⁺ exchanger in human platelets. To our knowledge, only human platelets and retinal rods express the retinal rod Na⁺/Ca²⁺ exchanger (NCKX1). Although K⁺-dependent Na⁺/Ca²⁺ exchange activity has been shown in rat brain, the Na⁺/Ca²⁺ exchanger in this tissue (NCKX2) does not share complete identity with the retinal rod Na⁺/Ca²⁺ exchanger (28).

Our findings suggest a role of the Na⁺/Ca²⁺ exchanger in overall Ca²⁺ homeostasis in both human platelets and megakaryocytes. This is shown by the acute and chronic inhibitions of the Na⁺ pump, which substantially raised the Ca²⁺_i and particularly the freely exchangeable Ca²⁺ stored in intracellular organelles of megakaryocytes and platelets. We have explored the possible roles of a number of biologic variables in regulating K⁺-dependent Na⁺/Ca²⁺ exchange activity in platelets (M. Kimura and A. Aviv, unpublished data) and found no effect of protein kinase C (PKC) stimulation (by phorbol 12-myristate 13-acetate) or inhibition (by staurosporine) and no effect of rising cAMP levels (by iloprost) or rising cGMP levels (by sodium nitroprusside) on K⁺-dependent Na⁺/Ca²⁺ exchange activity. Ni²⁺ (4 mmol/l) and La³⁺ (30 µmol/l) did inhibit the activity of K⁺-dependent Na⁺/Ca²⁺ exchange activity in platelets.

The K⁺ dependence of the retinal rod/platelet exchanger implies that the transmembrane K⁺ gradient contributes to the driving force for Na⁺/Ca²⁺-K⁺ exchange. However, the K_m for K⁺ for the activation of Ca²⁺-K⁺ cotransport is ~1 mmol/l (13, 23), suggesting that under physiological conditions the activation site for K⁺ (the K⁺ site; Ref. 21) would be saturated at both the intra- and extracellular membrane surfaces. Nevertheless, both retinal rods (22, 24) and proteoliposomes containing the purified exchanger protein (7) show a graded influence of K⁺ concentrations in the range of 10-140 mmol/l on the Ca²⁺ movements via the exchanger. The results presented in Fig. 6 show similar effects of high K⁺ concentrations on exchange-mediated Ba²⁺ influx in platelets.

The effects of K⁺ and other cations on exchange activity in retinal rods are complex and involve interactions with multiple sites on the exchanger (reviewed in Ref. 21). For example, when K⁺ is present on the same side of the membrane as Na⁺, it inhibits Na⁺-dependent Ca²⁺ movements in retinal rods (25, 26). This appears to reflect a competition between K⁺ and Na⁺ for binding to one of the Na⁺ transport sites. Such competitive interactions probably contributed to the results shown in Fig. 5B, in which Ba²⁺ influx increased sharply as Na⁺_i increased and K⁺_i decreased.

The effects of high concentrations of K⁺_o on Ba²⁺ influx in the absence of Na⁺_o (Fig. 6) appear to be inconsistent with previous measurements of a K_m of 1 mmol/l for activation of Ca²⁺ influx (13). At present, we cannot fully explain this discrepancy, although it seems possible that the use of Ba²⁺ instead of Ca²⁺ in the present experiments might be a contributing factor. Experiments with retinal rods demonstrate that Ca²⁺ increases the affinity of the K⁺ site for K⁺, and competing cations such as Mg²⁺ decrease the K⁺-site affinity (22). It is not known how Ba²⁺ affects the K⁺-site affinity, so the K_m for K⁺ activation of Ba²⁺ influx is not necessarily predictable from its effects on Ca²⁺ influx. Moreover, there may be additional effects of high K⁺ concentrations that are not yet well characterized. For example, we have observed in preliminary experiments that exchange-mediated Ca²⁺ influx in platelets at 145 mmol/l K⁺_o is increased 2.7-fold over that in 140 mmol/l LiCl plus 5 mmol/l KCl (a nominally saturating concentration of K⁺; M. Kimura, unpublished observations). On the basis of the results in Figs. 5 and 6 and the considerations discussed above, we propose that changes in the intracellular and extracellular K⁺ concentrations could have important modulatory effects on the rate, and possibly the direction, of Na⁺/Ca²⁺ exchanger in human platelets.

An additional consideration involves possible effects of K⁺_o on the membrane potential. Experiments with potential-sensitive dyes suggest that K⁺ gradients do not strongly affect the membrane potential of platelets, unless a K⁺ ionophore such as valinomycin is added (M. Kimura, unpublished observations). Moreover, any effects of membrane depolarization by K⁺ would be more pronounced at low K⁺ concentrations and would not be expected to yield the linear dependence on K⁺_o observed in Fig. 6. Thus the effects of K⁺_o on exchange activity probably reflect multiple interactions between K⁺ and cation binding sites on the exchanger rather than alterations in membrane potential.

In the unique ionic milieu of the retinal rod, the K⁺ dependence of the Na⁺/Ca²⁺-K⁺ exchanger appears to be critical for maintaining Ca²⁺ efflux activity. In the dark, the cGMP-gated channels of the retinal rod are open, resulting in membrane depolarization and high intracellular concentrations of both Na⁺ and Ca²⁺. The thermodynamic driving force generated by coupling Ca²⁺ movements to both the K⁺ and Na⁺ gradients allows Ca²⁺ to be extruded under these unusual ionic conditions. The teleological rationale for the presence of the retinal rod exchanger in platelets is less clear. Perhaps the increased driving force for Ca²⁺ efflux provided by the K⁺ gradient is necessary to maintain the resting Ca²⁺_i at a very low level and prevent accidental platelet activation in the circulation. On the other hand, the local rise in extracellular K⁺ and the rise in Na⁺_i in response to local thrombotic factors (3) at sites of biologic trauma could retard the rate of Ca²⁺ efflux by the exchanger, thereby facilitating a better response to local factors that activate platelets by raising Ca²⁺_i.

It is noteworthy, however, that dependence of platelet Na⁺/Ca²⁺ exchange activity on both the Na⁺ and K⁺ gradients would increase the sensitivity of platelets to factors that inhibit the Na⁺ pump, because such factors would concomitantly raise Na⁺_i and lower K⁺_i, thereby diminishing both the Na⁺ and K⁺ gradients and retarding Ca²⁺ efflux through Na⁺/Ca²⁺ exchange. A number of hypotheses linking the high intake of salt (NaCl) to high blood pressure in human beings are based on the concept that circulating Na⁺-pump inhibitors raise Na⁺_i and consequently retard Ca²⁺ extrusion by the Na⁺/Ca²⁺ exchanger, thereby increasing the Ca²⁺ load in tissue such as vascular smooth muscle cells (for opposing views on this subject, see Refs. 2, 10, and 17). Platelets would be highly sensitive to fluctuating levels of these putative factors for two reasons. First, because of the dependency of the Na⁺/Ca²⁺ exchanger in platelets on both the Na⁺ and K⁺ gradients across the plasma membrane, inhibition of the Na⁺ pump is likely to increase the Ca²⁺ load in platelets to a greater extent than in other cells that express the more common cardiac-type Na⁺/Ca²⁺ exchanger. Second, circulating platelets are anucleated cells without an appreciable protein synthesis machinery. Platelets would therefore adapt poorly to the inhibition of the Na⁺ pump and the consequent increase in the Ca²⁺ load because they would be unable to increase the synthesis of new Na⁺ pumps, Na⁺/Ca²⁺ exchangers, or other Ca²⁺ transport systems such as the plasma membrane Ca²⁺-ATPase. It follows then that platelets would be more sensitive to daily fluctuations in salt intake if these were mirrored by fluctuations in levels of circulating Na⁺-pump inhibitors.

Because platelet activity plays a central role in thromboembolic events and atherosclerosis, it is conceivable that the link between salt intake and cardiovascular diseases is exerted not only via a Na⁺-mediated increase in blood pressure, although such an effect in the general population is controversial (27, 30), but also through a Na⁺-mediated increase in platelet Ca²⁺ stores and platelet activity. In this context, a high K⁺ intake would exert the opposite effect on platelet Ca²⁺ and platelet activity, because K⁺ is a potent natriuretic factor (8). This concept is in line with observations in human beings of an inverse relationship between K⁺ intake and the frequency of occlusive stroke, independent of the effect of K⁺ intake on the systemic blood pressure (1, 11).