J Appl Physiol 95: 563–570, 2003. First published April 25, 2003; 10.1152/japplphysiol.00223.2003.
O2 affinity of cross-linked hemoglobins modifies O2 metabolism in proximal tubules A. D. Baines and P. Ho Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario M5G 1L5, Canada Submitted 3 March 2003; accepted in final form 22 April 2003
Baines, A. D., and P. Ho. O2 affinity of cross-linked hemoglobins modifies O2 metabolism in proximal tubules. J Appl Physiol 95: 563–570, 2003. First published April 25, 2003; 10.1152/japplphysiol.00223.2003.—Previous experiments using cross-linked tetrameric hemoglobins (XLHb) to perfuse isolated rat kidneys showed that high-O2-affinity XLHb improved proximal tubule function more effectively than low-O2-affinity XLHb. To determine how function was improved, proximal tubule fragments were incubated with albumin, Hb34 [half-saturation point (P50) 34 Torr], or Hb13 (P50 13 Torr) with PO2 values ranging from 22 to 147 Torr. ATP content reflected O2 delivery to mitochondria. Both XLHb increased ATP, Hb34 with PO2 ⱖ 47 Torr and Hb13 with PO2 ⱕ 47 Torr. XLHb increased Na-K-ATPase activity (86Rb uptake) in similar PO2-dependent patterns. O2 con˙ O2) was measured in a closed, well-stirred chamsumption (Q ˙ O2, reflecting Na-Kber. Ouabain- and oligomycin-inhibited Q ATPase activity and oxidative phosphorylation, respectively, mirrored the PO2-dependent patterns of ATP and 86Rb uptake. As PO2 fell below the midpoint of XLHb desaturation, ˙ O2, uncoupled from oxidative phosphorylation, transiently Q increased. The increase was most pronounced with Hb34. ˙ O2. Inhibitors Nitro-L-arginine methyl ester had no effect on Q of NAD(P)H oxidases and diamine oxidase partially pre˙ O2 surge with Hb34. In conclusion, facilitated vented the Q diffusion accounts for PO2-dependent XLHb effects on ATP content and Na-K-ATPase and for Hb13’s effectiveness in hypoxic perfused kidneys. NO scavenging was not a factor. O2-binding characteristics influence XLHb effects on mitochondria and O2-sensitive enzymes such as oxidases. ATP; Na-K-ATPase; Rb uptake; NAD(P)H oxidases; diphenyleneiodonium; aminoguanidine; nitro-L-arginine methyl ester; deferroxamine; oxidative phosphorylation HEMOGLOBIN-BASED BLOOD SUBSTITUTES are being developed to improve O2 delivery in conditions that require blood transfusion. Blood substitutes not only carry O2 but also influence O2 transfer by “facilitated diffusion” through plasma to O2-consuming cells. Facilitated diffusion depends on the diffusion constant for oxygenated-hemoglobin and chemical reaction rates for O2 binding and release (26). The diffusion constant is inversely proportional to molecular size. The amount of O2 carried is directly proportional to hemoglobin concentration, PO2, and O2-binding characteristics (4, 11, 26). The half-saturation point (P50) and cooperativity (Hill
Address for reprint requests and other correspondence: A. D. Baines, Rm. 408 Banting Institute, 100 College St., Toronto, ON M5G 1L5, Canada (E-mail:
[email protected]). http://www.jap.org
coefficient) determine the range of PO2 over which O2 is released from hemoglobin. In previous experiments, we perfused isolated rat kidneys with cross-linked tetrameric hemoglobins (XLHb) that differed only in ˙ O2) was their P50 and Hill coefficients. O2 consumption (Q similar with the two XLHb but venous PO2 was 40–60% lower when the perfusate contained high-affinity XLHb (4). High-affinity XLHB also maintained tubular glucose and phosphate reabsorption more effectively. That observation suggested that a high-affinity XLHb, which delivered O2 at very low PO2, maintained kidney function most effectively. The XLHb with O2 affinity similar to that of erythrocytes was less effective. Both XLHb facilitate O2 diffusion; the only difference is the PO2 at which O2 is released. Oxygenation of proximal tubules increases Na-K-ATPase activity and converts the response to protein kinase C and protein kinase A from inhibition to stimulation (19). PO2 influences the activity of phospholipase A2, NAD(P)H oxidases that produce reactive O2 species (24), and oxygenases that produce cytokines from arachidonic acid (18). The products of these various O2-dependent processes could modulate Na-K-ATPase and other aspects of proximal tubule function. In the following experiments, we investigated the biological mechanisms responsible for the beneficial effect of high-affinity hemoglobin on proximal tubule function. Another attribute of high-affinity XLHb may contribute to its superior performance. It has a high affinity for nitric oxide (NO) (29). NO inhibits oxidative phosphorylation when PO2 is low (1, 22). Under hypoxic conditions that exist in the renal inner cortex and outer medulla, ATP production could be inhibited by NO or limited by the availability of O2. Reduced ATP production could limit Na-K-ATPase activity and Na cotransport of glucose and phosphate (7). Therefore, high-affinity hemoglobin could improve proximal tubular function by its effects on O2 transfer or by scavenging NO. To examine these possibilities, we measured ˙ O2, ATP content, and Na-K-ATPase activity (86Rb Q uptake) in proximal tubule fragments incubated with high- or low-affinity XLHb or bovine serum albumin (BSA).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
8750-7587/03 $5.00 Copyright © 2003 the American Physiological Society
563
564
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
Results obtained with tubule fragments exposed directly to XLHb in vitro may be extrapolated to tubules in vivo because of the close proximity between tubule cells and capillary lumens in the renal cortex. Twothirds of epithelial cells’ basal surface is only 0.15–0.4 m distant from a capillary lumen. Capillary endothelium is thin and contains many fenestrae closed by 40–60 Å thick diaphragms (3). With this close approximation, XLHb circulating in capillary blood interacts almost directly with epithelial cells. Tetrameric XLHb, polymerized hemoglobin, and albumin all escape from capillaries and contact tubular epithelium before appearing in renal lymph (25). In vivo, some XLHb enters the tubule lumen and is reabsorbed in an O2-dependent manner (4). METHODS
Hemosol (Etobicoke, ON, Canada) provided the XLHb, prepared in lactated Ringer solution. Hb34 [trimesoyl-Hb (1–82⬘); P50 34, Hill 2.5] or Hb13 [trimesoyl-Hb (82–82⬘); P50 13, Hill 1.9] was prepared as previously described (20, 21, 30). Hb34 contains 33% ␣2(Val1)-Tm-(Lys82) and 67% ␣2(Val1,Lys82)-Tm-(Lys82). Tm is the cross-linker trimesic acid. Hb13 (82-82⬘-Hb) contains 20% ␣2(Lys82,Lys144)Tm-(Lys82) and 80% ␣2(Lys82)-Tm-(Lys82⬘). O2 affinity was measured with a Hemox analyzer at 37°C; the parameters measured in these lots of XLHb differ slightly from those described in our previous publication in which P50 were 11 and 35. Rohlfs et al. (29) report values of 15 and 2 for Hb13, which they called 82-Hb. Their values for Hb34, which they called Tm-Hb, were 39 and 2.8. Tubules were prepared from outer cortical slices of Wistar rats as previously described (5). The kidneys were cleared of blood by infusing 60 ml of isotonic saline through the aorta. Slices were removed from the outer cortex with a StadieRiggs microtome placed in ice-cold saline and minced finely with a razor blade. The minceate was incubated at 37°C for 30 min in 6 ml of Krebs-Henseleit buffer containing 7.2 mg of collagenase (Sigma Chemical, St. Louis, MO) and 30 mg of BSA. The reaction was stopped with ice-cold buffer solution. The tissue was passed through a tea strainer, washed four times with buffer, and suspended in 30 ml of 45% Percoll in Krebs-Henseleit buffer. Tubules were separated from glomeruli by centrifugation in a 60TI rotor at 20,000 rpm for 20 min. Microscopic examination showed that the bottom layer containing 80–90% proximal tubule fragments with virtually no glomeruli. This layer was removed and washed four times with buffer and passed once through a 100-m sieve. Tubules were kept on ice in a modified Krebs-Henseleit solution, which contained (in mM) 136 Na, 5 K, 111 Cl, 25 HCO3, 0.5 Mg, 1 Ca, 5 or 25 glucose, 2 lactate, 0.2 pyruvate, 2 glutamine, 1 arginine, 1 alanine, 1 heptanoic acid, and 10 g/l BSA. The pH was 7.4 at 37°C when equilibrated with 95% air-5% CO2. For measurement of ATP concentration and 86Rb uptake, tubule fragments (0.8–1.6 mg protein) were incubated in 1.5–2 ml containing 1.1% BSA, 1% Hb13 ⫹ 0.1% BSA or 1% Hb34 ⫹ 0.1% BSA in a 50-ml Erlenmeyer flask at 37°C for 10 min with 95% air-5% CO2. The gas was then changed to 3–10% O2-5% CO2 for a further 10 min. ATP (ATP bioluminescent assay kit, Sigma Chemical) was measured with a Luminometer after addition of 30 l of 70% perchloric acid. Ouabain-sensitive 86Rb uptake by proximal tubule fragments was used to measure Na-K-ATPase activity (5) 86Rb (Perkin-Elmer Life Sciences, Boston, MA) was added to proJ Appl Physiol • VOL
duce an activity of ⬃1 Ci/ml. Uptake was measured with and without 2.5 mM ouabain. Uptake was terminated after 1 min by layering the tubule suspension onto 0.5 ml of a 2:1 mixture of dibutyl-dioctyl phthalate in a 1.5-ml centrifuge tube and centrifuging for 10 s in an Eppendorf 5414 centrifuge. The medium above the oil layer was removed, and the tubule was rinsed five times with distilled water without disturbing the oil. The pellet of tubules was dissolved in 1 ml of 0.1 N NaOH, and 200 l of the solution were added to 10 ml of liquid scintillation cocktail (Ready Safe, Beckman Coulter, Fullerton, CA) for counting in a liquid scintillation counter. In preliminary experiments, [3H]inulin was added with 86Rb. Less than 1% of the 3H passed through the oil with the tubules; therefore, in subsequent experiments, we did not include [3H]inulin and did not correct for trapped extracellular fluid. ˙ O2, tubules were incubated at Before measurement of Q 37°C with 5% CO2-95% air for 30 min. In some experiments, oligomycin (10 mol/l), ouabain (2.5 mmol/l), aminoguanidine (5 mM), or deferoxamine mesylate (1 mM) was added to the incubation solution 5–10 min before measurement of ˙ O2. Nitro-L-arginine methyl ester (L-NAME, 100 mol/l) Q ˙ O2 measurement. Diphenylenewas added 30 min before Q iodonium (DPI; 10 M) was added directly to the measurement chamber without preincubation. Chemicals were sup˙ O2 plied by Sigma-Aldrich Canada unless otherwise noted. Q measurement was started by injecting tubules (0.8–1.6 mg of protein) into the 0.6-ml analytic chamber (Diamond General Development). The chamber contained either 1% BSA or XLHb with sufficient albumin to maintain a constant 1% protein concentration. Albumin and hemoglobin were dissolved in Krebs-Henseleit solution and preequilibrated with 95% air-5% CO2 at 37°C. PO2 was recorded polarographically with a YSI model 5300 biological oxygen monitor. Comparisons were done in duplicate or triplicate in most experiments. Many investigators have used the linear rate of change in ˙ O2 by a variety of cells O2 partial pressure (PO2) to measure Q and mitochondria. When tubules were incubated with 1% BSA alone, O2 content was calculated from PO2 and the solubility of O2 in plasma at 37°C: O2 mol ⫽ PO2 (Torr) ⫻ 0.6 ml ⫻ 0.001257 mol 䡠 ml⫺1 䡠 Torr⫺1 (12). A simple linear relationship between PO2 and O2 content does not exist when the solution contains hemoglobin. Instead, O2 content can be described by a function that includes PO2, the concentration of hemoglobin, hemoglobin O2 affinity (P50), and the Hill coefficient (Hill). For solutions containing hemoglobin, the content was calculated as follows: O2 mol ⫽ PO2 (Torr) ⫻ 0.6 ml ⫻ 0.001257 mol 䡠 ml⫺1 䡠 Torr⫺1 ⫹ {Hb g/dl ⫻ 0.62 ⫻ 0.6 ml ⫻ (1-methemoglobin) ⫻ [PO2Hill/(PO2Hill ⫹ P50Hill)]}. Output from the YSI O2 monitor was recorded with a Biopac Systems MP100 recorder at 2 samples/s. The data ˙ O2 were transferred to an Excel program, which calculated Q ˙ O2 over consecutive from an average of the O2 content and Q 15-s intervals. The program also made corrections for atmospheric pressure and changes in methemoglobin. For the latter calculations, it was assumed that methemoglobin increased linearly with time during each run. Methemoglobin content of the hemoglobin solution was measured before addition of tubules and at the end of each run (13). The accuracy of the equation was confirmed by equilibrating XLHb solutions with various O2 concentrations and comparing calculated O2 content with O2 content measured by a LexO2Con analyzer (Lexington Instruments, Waltham, MA). Sensitivity analysis of the predicted O2 content was done by varying P50 and Hill coefficients used in the calculations.
95 • AUGUST 2003 •
www.jap.org
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
Fig. 1. A: ATP content of proximal tubule fragments incubated under 95% air-5% CO2, 10% O2-5% CO2, 6.5% O2-5% CO2, or 3% O2-5% CO2. Incubation solution contained 10 g/l bovine serum albumin (BSA), 5 g/l Hb34 with 5 g/l BSA, or 5 g/l Hb13 with 5 g/l BSA. * and ⫹P ⬍ 0.05 by repeated-measures ANOVA comparing Hb34 with BSA and Hb13 with BSA, respectively. B: ouabain-sensitive 86Rb uptake by proximal tubule fragments incubated as indicated above. Values are means ⫾ SE.
565
To analyze the relationship between ATP concentration ([ATP]) and Na-K-ATPase activity, [ATP] was calculated by assuming that tubule cells contained 2 l water/mg protein (Fig. 2), which is within the range found by others (1.6–2.2 l/mg protein) (10, 31, 32). During incubation with BSA, Na-K-ATPase activity was independent of [ATP] above 2 mM. We assumed that the relationship between [ATP] and 86Rb uptake is described by Michaelis-Menten kinetics. Km was 0.88 ⫾ 0.37 mM, and Vmax was 147 ⫾ 8 mol 䡠 min⫺1 䡠 mg⫺1 for the rectangular hyperbola fitted to the BSA data using Sigma Plot 4.01. This Km is in the range found by direct measurement of Na-K-ATPase in cell membranes (17, 31). The XLHb produced O2-dependent increases in Na-K-ATPase relative to [ATP], which suggests that factors in addition to [ATP] were responsible for increased activity. ˙ O2. To examine the effects of XLHb on various Q ˙ O2, we used several inhibitors. Oligocomponents of Q ˙ O2 for ATP production by oxidative mycin blocks Q phosphorylation, and ouabain blocks O2 and ATP consumption by Na-K-ATPase. Other inhibitors were used to block O2 consumed by oxidases. After addition of tubules to a BSA solution, PO2 decreased linearly down to 20 Torr or less (8). O2 content of a BSA solution is directly related to PO2; ˙ O2 was calculated from the rate of decrease therefore, Q in PO2. After addition of tubules to a XLHb solution, PO2 decreased in a nonlinear fashion (inset in Fig. 3). O2 content was calculated by using the XLHb concentration and the Hill coefficients and P50 values previously measured for Hb34 with a Hemox analyzer at 37°C (central thick line). A range of values for the Hill coefficient and P50 were used to show sensitivity of the
RESULTS
ATP and Rb uptake. The effects of O2 concentration on Na-K-ATPase and ATP content were examined by incubating tubules with 95% air (147 Torr) or 10% (72 Torr), 6.5% (47 Torr), or 3% O2 (22 Torr). Ouabain inhibited 86Rb uptake by 75 ⫾ 1% . Ouabain-sensitive 86 Rb uptake into tubules incubated with BSA decreased slightly over the range of PO2 from 147 down to 47 Torr (Fig. 1B). Hb34 increased uptake relative to BSA by 3% at 147 Torr and 11% at 72 Torr PO2. Hb13 increased Rb uptake by 13% at 47 Torr and 38% at 22 Torr PO2. When tubules were incubated with BSA, ATP content increased in a sigmoidal fashion relative to PO2 (Fig. 1A). Incubation with XLHb at 47 and 72 Torr PO2 increased ATP content by 22–33%. Hb13 also sustained 10% higher ATP at 22 Torr PO2. Thus the effects of Hb13 on both ATP and Rb uptake were left-shifted relative to Hb34. J Appl Physiol • VOL
Fig. 2. ATP concentration and ouabain-sensitive 86Rb uptake in proximal tubule fragments incubated under 95% air-5% CO2, 10% O2-5% CO2, 6.5% O2-5% CO2, or 3% O2-5% CO2. Incubation solution contained 10 g/l BSA, 5 g/l Hb34 with 5 g/l BSA, or 5 g/l Hb13 with 5 g/l BSA. * and ⫹P ⬍ 0.05 by repeated-measures ANOVA comparing Hb34 with BSA and Hb13 with BSA, respectively. The curve was fitted to the results obtained with BSA by using the formula y ⫽ ax/(b ⫹ x), with a ⫽ 147 and b ⫽ 0.88. ATP concentration was calculated by assuming 2 l water/mg protein. Values are means ⫾ SE.
95 • AUGUST 2003 •
www.jap.org
566
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
˙ O2 to PO2; therefore, we used 5 g/l for all relating Q subsequent experiments. The initial methemoglobin content of hemoglobin solutions was 7.5 ⫾ 0.5% for Hb13 and 12.6 ⫾ 0.1 for Hb34. Methemoglobin increased by 0.28 ⫾ 0.03%/min with Hb13 and 0.20 ⫾ 0.03%/min with Hb34 reaching 10.6 ⫾ 0.3 and 14.1 ⫾ 0.2%, respectively, by the end of the measurement. In the calculation of O2 content, we assumed that methemoglobin increased at a constant rate. ˙ O2 curves sampled at intervals of Figures 4–7 show Q 5 Torr PO2. Duplicate or triplicate measurements with a tubule preparation were averaged before carrying out statistical analyses based on the number of tubule preparations. O2 content fell most rapidly when PO2 decreased below the midpoint of the XLHb O2 dissociation curves. Consequently, as PO2 fell from 65 to 25 ˙ O2 surged up briefly. With Hb13, Q ˙ O2 Torr with Hb34, Q increased below 15 Torr PO2, but to a much smaller
˙ O2) calculation for a Fig. 3. Sensitivity analysis of O2 consumption (Q representative experiment using 5 g/l Hb34 with 1.2 g of proximal tubule fragments. Inset: decrease in PO2 after addition of tubules to ˙ O2 was the sealed chamber containing Hb34. Central heavy line for Q calculated by using a Hill coefficient of 2.5 and a half-saturation point (P50) of 34 Torr (see text for the formula). A: effect of changing the P50 while keeping the Hill coefficient at 2.5. Bottom: effect of changing the Hill coefficient by using a P50 of 34.
˙ O2 calculation. The P50 values and Hill coefficients Q reported by Rohlfs et al. (29) yield slightly lower peak ˙ O2 consumption but do not produce major changes in Q the overall pattern. In preliminary experiments using 5 and 10 g/l of the XLHb, concentration had no signif˙ O2 nor on the shape of the curves icant effect on total Q J Appl Physiol • VOL
˙ O2 (B) as a function of Fig. 4. Total (A) and oligomycin-insensitive Q PO2 for proximal tubule fragments incubated with 10 g/l BSA, 5 g/l Hb34 with 5 g/l BSA, or 5 g/l Hb13 with 5 g/l BSA. Oligomycin (10 M) ˙ O2 measurewas added to the incubation solutions 5–10 min before Q ˙ O2 curves of ment. The points were obtained by sampling the Q individual experiments at intervals of 5 Torr PO2 (means ⫾ SE). Duplicate or triplicate measurements with the same tubule preparation were averaged to obtain a value for each preparation. Numbers of tubule preparations are shown in parentheses. * and ⫹P ⬍ 0.05 comparing BSA with Hb34 and Hb13, respectively (ANOVA).
95 • AUGUST 2003 •
www.jap.org
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
˙ O2 remained when oxiextent (Fig. 4). The surge in Q dative phosphorylation was inhibited by oligomycin (Fig. 4 below). ˙ O2 with Both XLHb increased oligomycin-sensitive Q Hb13 by up to 30% when PO2 was between 5 and 10 Torr (P ⫽ 0.05–0.01). With Hb34, the increase was up to 20% when PO2 was between 20 and 40 Torr, but no single point reached statistical significance (P ⫽ 0.15–0.2) (Fig. 5A). Nystatin (10 M), which raises Na-K-ATPase activity by increasing Na entry into the cell, stimulated ˙ O2 with Hb34. The effect was oligomycin-sensitive Q most notable between 20 and 40 Torr, which confirms the PO2-dependent stimulation of Na-K-ATPase by Hb34 (P ⫽ 0.02–0.005 for 4 paired measurements in 2 tubule preparations). ˙ O2 also appeared to With XLHb, ouabain-sensitive Q be sensitive to PO2, but none of the points reached statistical significance (Fig. 5B). The insensitivity of rat tubules to ouabain makes oligomycin a more reli-
567
able index of mitochondrial ATP production than ouabain. O2 affinity of the XLHb influenced the patterns of increased ouabain- and oligomycin-sensitive ˙ O2 (Fig. 5). Likewise, O2 affinity influenced the PO2Q dependent increases in 86Rb uptake (Fig. 1). Rb uptake was measured in a steady state with constant O2 con˙ O2 measurements were made uncentration, whereas Q der non-steady-state conditions as O2 concentration decreased over 3–8 min. Both methods indicated increased Na-K-ATPase activity with XLHb. XLHb did not significantly alter the ˙ O2. The ratio ratio of Rb uptake to ouabain-sensitive Q was calculated for Rb uptake from 47 to 147 Torr and ˙ O2 from 30 to 110 Torr. The averages were BSA for Q 9.2, Hb34 9.8, and Hb13 10.9. Under optimum conditions with oxidation of NADH-linked substrates, mitochondria should synthesize 4–6 ATP for each O2 that is consumed (16). Assuming that Na-K-ATPase transports 2 Rb (or K) for each ATP consumed, then the ratio ˙ O2 should be 8–12. of Rb uptake to ouabain-sensitive Q Several inhibitors of oxidases were tested. Hb34 increased sensitivity to aminoguanidine, an inhibitor of diamine oxidase and nitric oxide synthase activity (NOS) (Fig. 6B). L-NAME, which blocks NOS (Fig. 6, ˙ O2, but inset), tended to stimulate rather than inhibit Q the change was not significant; therefore, the response to aminoguanidine was not due to inhibition of NOS. DPI, an NAD(P)H oxidase inhibitor, decreased the ˙ O2 caused by Hb34. DPI-sensitive Q ˙ O2 rose surge of Q significantly as PO2 fell to 20 Torr (Fig. 6A). Oligomycin added during DPI inhibition exerted its full effect; therefore, DPI and oligomycin acted additively (Fig. 7). ˙ O2. In this respect, it DPI only inhibited the burst of Q ˙ O2 over the differed from oligomycin, which reduced Q entire range of PO2 (Fig. 4). ˙ O2 with Deferoxamine (1 mM) tended to reduce Q Hb34 or BSA, but the 4 nmol 䡠 min⫺1 䡠 mg⫺1 decrease in ˙ O2 was not statistically significant (paired t-test, peak Q P ⫽ 0.10, n ⫽ 5). Deferoxamine also tended to decrease the production of methemoglobin (95% confidence interval 0–6% ). DISCUSSION
˙ O2 means ⫾ SE. Fig. 5. Oligomycin-sensitive and ouabain-sensitive Q Proximal tubule preparations were incubated with 10 gl/l BSA, 5 g/l Hb34, or 5 g/l Hb13. Hemoglobin solutions also contained 5 g/l BSA. ˙ O2 from total Q ˙ O2 Values were obtained by subtracting inhibited Q obtained with the same tubule preparation and averaging the values for each tubule preparation (means ⫾ SE). Numbers of tubule preparations are shown in parentheses. In 2 experiments with paired measurements, 10 M nystatin was used with Hb34. J Appl Physiol • VOL
The effects of XLHb on isolated perfused rat kidneys prompted these experiments with proximal tubules. Perfusing kidneys with Hb13 increased glucose and phosphate reabsorption more effectively than perfusing with Hb34 (4). Because glucose and phosphate are cotransported with Na in the proximal tubule, their reabsorption is very sensitive to Na-K-ATPase activity and ATP production (15). Virtually all the ATP in proximal tubules comes from mitochondrial oxidative phosphorylation. Thus XLHb could improve phosphate and glucose reabsorption by influencing Na cotransport, Na-K-ATPase, or oxidative phosphorylation. Given the ability of XLHb to facilitate O2 diffusion, oxidative phosphorylation is the most likely point at which they could act. Hypoxic conditions in the inner renal cortex may limit oxidative phosphorylation by reducing O2 deliv-
95 • AUGUST 2003 •
www.jap.org
568
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
Fig. 6. Diphenyleneiodonium-sensitive (A) and aminoguanidine-sen˙ O2 (B) by proximal tubule fragment preparations incubated sitive Q with 10 g/l BSA, 5 g/l Hb34 with 5 g/l BSA, or 5 g/l Hb13 with 5 g/l BSA (means ⫾ SE). Numbers of tubule preparations are shown in paren˙ O2 theses. Inset: nitro-L-arginine methyl ester (L-NAME)-sensitive Q was calculated for tubules preincubated with 100 mM L-NAME.
ery. Hypoxia also favors competitive inhibition of oxidative phosphorylation by NO (22). Hb13 and Hb34 could support oxidative phosphorylation by facilitating O2 diffusion to cells and by scavenging NO. Both XLHb are NO scavengers, but Hb13 has a higher affinity for NO (29), which might account for its beneficial effects in the isolated perfused kidney. However, if NO scavenging is a factor, then blocking NO synthesis with L-NAME should have reduced the effects of Hb34 on ˙ O2. This did not happen (Fig. 6). Furthermore, in Q previous experiments, L-NAME did not alter 86Rb uptake by proximal tubules incubated with BSA (6). Therefore, it is unlikely that XLHb stimulated Na-K˙ O2 by removing extracellular NO. XLHb is ATPase or Q not likely to have entered tubule cells and altered intracellular NO during the short duration of these experiments. Improved O2 delivery through facilitated diffusion is a more likely explanation for the beneficial effects of Hb13. J Appl Physiol • VOL
XLHb facilitate O2 diffusion by binding O2 at the high side of a PO2 gradient and releasing it at the low side. O2 transfer depends on the product of the diffusion constant and the amount of O2-hemoglobin in solution. Free O2 diffuses much more rapidly than O2 bound to hemoglobin, but the solubility of O2 is low, so the net transfer rate is low. XLHb diffuse more slowly but can hold much more O2 in solution. O2-binding equilibria determine the PO2 levels at which O2 is bound and released from the XLHb. To examine facilitated diffusion, McCarthy et al. (26) perfused an O2permeable capillary surrounded by N2 gas with tetrameric Hb with low and high O2 affinity (P50 33 and 15 Torr, respectively). The high-affinity XLHb facilitated diffusion more effectively when PO2 was low. Hb13 and Hb34 are hemoglobin tetramers like those used by McCarthy et al.; therefore, we expected that Hb13 would deliver O2 most effectively at low PO2. In our experiments, ATP content provides a biological index of facilitated diffusion. ATP content decreased in a sigmoidal fashion as the equilibrating O2 concentration was lowered (Fig. 1). ATP production by oxidative phosphorylation, indicated by oligomycin˙ O2, and ATP consumption by Na-K-ATPase, sensitive Q ˙ O2 and 86Rb uptake, indicated by ouabain-sensitive Q were insensitive to PO2 until it fell below 20 Torr (Figs. 1 and 5). Decreasing [ATP] with relatively constant oxidative phosphorylation is consistent with a kinetic model of oxidative phosphorylation developed by Korzeniewski (23). His model predicts that, when respiration is 70–80% of maximum, lowering mitochondrial O2 concentration from 30 to 3 M can decrease [ATP] ˙ O2 and ATP by 80% with only a slight decrease of Q production (B. Korzeniewski, personal communication). The PO2-dependent effects of Hb13 and Hb34 on [ATP] are consistent with their effects on facilitated diffusion as a function of PO2. Na-K-ATPase activity was independent of [ATP] above 2.5 mM during incubation with BSA (Fig. 2). Yet
˙ O2 by proximal tubule fragFig. 7. Diphenyleneiodonium-sensitive Q ment preparations incubated with 5 g/l Hb34 with 5 g/l BSA. Oligomycin was added to the chamber partway through the recording. Mean of 2 paired runs with a single tubule preparation.
95 • AUGUST 2003 •
www.jap.org
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
569
Fig. 8. Summary flow diagram showing effects of Hb13 and Hb34. HIF, hypoxia inducible factor; ROS, reactive oxygen species.
Hb13 and Hb34 produced PO2-dependent patterns of increased Na-K-ATPase activity relative to [ATP], which suggests that increased O2 transfer stimulated Na-K-ATPase activity by factors in addition to [ATP]. There is precedent for this suggestion (14). Oxygenation with 100% O2 for 15 min increased total Na-KATPase hydrolytic activity in proximal tubule fragments by 25% when compared with tubules incubated with air. In those experiments, Na-K-ATPase activity was measured with exogenous saturating concentrations of ATP. Increased oxygenation also converted the Na-K-ATPase response to protein kinase C or protein kinase A from inhibition to stimulation (14, 19). Fe´raille et al. (14) attributed the Na-K-ATPase stimulation to reduced phospholipase A2 activity (28). Phospholipase A2 is activated by hypoxia and releases arachidonate. Lipoxygenases act on arachidonate to produce derivatives that inhibit Na-K-ATPase (14, 19). Oxygenation blocks this inhibitory pathway by preventing phospholipase A2 activation. ˙ O2, As PO2 fell and O2 began to escape from XLHb, Q unrelated to oxidative phosphorylation, increased rapidly, reaching a peak at the midpoint of Hb13 desaturation (P50 13 Torr) and 10–15 Torr below the midpoint ˙ O2 was for Hb34 (P50 34 Torr) (Fig. 4B). The burst of Q most evident with Hb34; therefore, we examined it in more detail. A nonspecific inhibitor of NAD(P)H oxidases inhibited two-thirds of the brief oligomycin-insensitive increase (Fig. 6A). NAD(P)H oxidase activity is found in plasma membranes and is a source of reactive O2 species (ROS) in the proximal tubule (24). NAD(P)H oxidase in plasma membranes activates hypoxia-inducible factor-1 and may be a factor in the cellular response to hypoxia (27). Diamine oxidase, inhibited by aminoguanidine, was also stimulated by Hb34. Absence of a response to L-NAME indicates that NOS, which is also inhibited by aminoguanidine, was not involved (Fig. 6). A PO2-dependent pattern of inJ Appl Physiol • VOL
˙ O2 similar to that shown in creased then decreased Q Fig. 4 could be produced by O2-consuming processes with non-Michaelis-Menten kinetics due to substrate (O2) inhibition. Juranek et al. (18) show such a relationship between O2 concentration and cyclooxygenase-1. Uncoupling of oxidative phosphorylation triggered by increased ROS production in the mitochondria or by redox reactions with hemoglobin (2, 9) might also ˙ O2 surge. However, ROS have contributed to the Q production by oxidation of Hemoglobin to methemoglobin was low and similar for Hb34 and Hb13 and is unlikely to have been a major factor accounting for the difference in responses to the two XLHb. Furthermore, blocking ROS with deferoxamine de˙ O2 with Hb34 by a statistically insigcreased peak Q nificant 4 nmol/min. The effects of Hb13 and Hb34 on proximal tubular function are summarized in Fig. 8. Hb13 facilitates O2 diffusion more effectively than Hb34 under hypoxic conditions. This is manifested by increased [ATP]. Stimulation of ATP production and Na-K-ATPase activity by increase O2 delivery at low PO2 could account for improved reabsorption of phosphate and glucose by proximal tubules in perfused kidneys (4). Additional factors such as reduced phospholipase activity may have contributed to O2-dependent increase in Na-KATPase activity above that due simply to increased [ATP]. NOS activity and ROS production do not appear to have been major factors in the responses to XLHb. Within a narrow window of O2 concentrations around its P50, Hb34 stimulated oxidase activity and other ˙ O2 unrelated to oxidative phosphorylation. forms of Q Increased XLHb-induced oxidase activity in vivo could produce cytokines that influence the function of proximal tubules and surrounding vascular tissue and contribute to the vasoconstriction that follows XLHb transfusion (29).
95 • AUGUST 2003 •
www.jap.org
570
CROSS-LINKED HEMOGLOBINS MODIFY O2 METABOLISM
Drs. R. Kluger, G. Adamson, and S. Pang provided valuable advice and encouragement. 16. DISCLOSURES This research was supported by grants from the National Research Council of Canada and the Canadian Blood Services/Canadian Institutes of Health Research Partnership. Hemosol provided the Hb34 and Hb13.
17. 18.
REFERENCES 1. Adler S, Huang H, Loke KE, Xu X, Tada H, Laumas A, and Hintze TH. Endothelial nitric oxide synthase plays an essential role in regulation of renal oxygen consumption by NO. Am J Physiol Renal Physiol 280: F838–F843, 2001. 2. Alayash AI. Hemoglobin-based blood substitutes and the hazards of blood radicals. Free Radic Res 33: 341–348, 2000. 3. Aukland K, Bogusky RT, and Renkin EM. Renal cortical interstitium and fluid absorption by peritubular capillaries. Am J Physiol Renal Fluid Electrolyte Physiol 266: F175–F184, 1994. 4. Baines AD, Adamson G, Wojciechowski P, Pliura D, Ho P, and Kluger R. Effect of modifying O2 diffusivity and delivery on glomerular and tubular function in hypoxic perfused kidney. Am J Physiol Renal Physiol 274: F744–F752, 1998. 5. Baines AD, Drangova R, and Ho P. Role of diacylglycerol in adrenergic-stimulated 86Rb uptake by proximal tubules. Am J Physiol Renal Fluid Electrolyte Physiol 258: F1133–F1138, 1990. 6. Baines A and Ho P. Glucose stimulates O2 consumption, NOS, and Na/H exchange in diabetic rat proximal tubules. Am J Physiol Renal Physiol 283: F286–F293, 2002. 7. Balaban RS, Mandel LJ, Soltoff SP, and Storey JM. Coupling of active ion transport and aerobic respiratory rate in isolated renal tubules. Proc Natl Acad Sci USA 77: 447–451, 1980. 8. Balaban RS, Soltoff SP, Storey JM, and Mandel LJ. Improved renal cortical tubule suspension: spectrophotometric study of O2 delivery. Am J Physiol Renal Fluid Electrolyte Physiol 238: F50–F59, 1980. 9. Balagopalakrishna C, Manoharan PT, Abugo OO, and Rifkind JM. Production of superoxide from hemoglobin-bound oxygen under hypoxic conditions. Biochemistry 35: 6393–6398, 1996. 10. Blumenthal SS, Ware RA, and Kleinman JG. Sodium gradient-driven transport processes in ATP-depleted renal tubules. Am J Physiol Cell Physiol 244: C331–C335, 1983. 11. Bouwer ST, Hoofd L, and Kreuzer F. Reaction rates of oxygen with hemoglobin measured by non-equilibrium facilitated oxygen diffusion through hemoglobin solutions. Biochim Biophys Acta 1525: 108–117, 2001. 12. Endre ZH, Ratcliffe PJ, Tange JD, Ferguson DJP, Radda GK, and Ledingham JGG. Erythrocytes alter the pattern of renal hypoxic injury: predominance of proximal tubular injury with moderate hypoxia. Clin Sci (Lond) 76: 19–29, 1989. 13. Evelyn KA and Malloy HT. Microdetermination of oxyhemoglobin, methemoglobin and sulfhemoglobin in a single sample of blood. J Biol Chem 126: 655–662, 1938. 14. Fe´raille E, Carranza ML, Rousselet F, Doucet A, and Favre H. Protein kinase C-dependent stimulation of Na-KATPase in rat proximal convoluted tubule. Am J Physiol Cell Physiol 268: C1277–C1283, 1995. 15. Gullans SR, Brazy PC, Soltoff SP, Dennis VW, and Mandel L. Metabolic inhibitors: effects on metabolism and transport in
J Appl Physiol • VOL
19.
20.
21.
22. 23. 24.
25. 26.
27.
28.
29.
30. 31. 32.
the proximal tubule. Am J Physiol Renal Fluid Electrolyte Physiol 243: F133–F140, 1982. Gullans SR and Mandel LJ. Coupling of energy to transport in proximal and distal nephron. In: The Kidney. Physiology and Pathophysiology, edited by Seldin DW and Giebisch G. New York: Raven, 1992, p. 1291–1337. Jorgensen PL. Sodium and potassium ion pump in kidney tubules. Physiol Rev 60: 864–917, 1980. Juranek I, Suzuki H, and Yamamoto S. Affinities of various mammalian arachidonate lipoxygenases and cyclooxygenases for molecular oxygen as substrate. Biochim Biophys Acta 1436: 509–518, 1999. Kiroytcheva M, Cheval L, Carranza ML, Martin PY, Favre H, Doucet A, and Feraille E. Effect of cAMP on the activity and the phosphorylation of Na⫹,K⫹-ATPase in rat thick ascending limb of Henle. Kidney Int 55: 1819–1831, 1999. Kluger R and Song Y. Changing a protein into a generalized acylating reagent. Reaction of nucleophiles 3,5-dibromosalicyl trimesyl-((Lys82)-(Lys82))-hemoglobin. J Org Chem 59: 733– 736, 1994. Kluger R, Wodzinska J, Jones RT, Head C, Fujita TS, and Shih TD. Three-point cross-linking: potential red cell substitutes from the reaction of trimesoyl tris(methyl phosphate) with hemoglobin. Biochemistry 31: 7551–7559, 1992. Koivisto A, Pittner J, Froelich M, and Persson AEG. Oxygen-dependent inhibition of respiration in isolated renal tubules by nitric oxide. Kidney Int 55: 2368–2375, 1999. Korzeniewski B. Regulation of ATP supply in mammalian skeletal muscle during resting state 3 intensive work transition. Biophys Chem 83: 19–34, 2000. Li N, Yi FX, Spurrier JL, Bobrowitz CA, and Zou AP. Production of superoxide through NADH oxidase in thick ascending limb of Henle’s loop in rat kidney. Am J Physiol Renal Physiol 282: F1111–F1119, 2002. Matheson B, Razynska A, Kwansa H, and Bucci E. Appearance of dissociable and cross-linked hemoglobins in the renal hilar lymph. J Lab Clin Med 135: 459–464, 2000. McCarthy MR, Vandegriff KD, and Winslow RM. The role of facilitated diffusion in oxygen transport by cell-free hemoglobins: implications for the design of hemoglobin-based oxygen carriers. Biophys Chem 92: 103–117, 2001. Osada M, Imaoka S, Sugimoto T, Hiroi T, and Funae Y. NADPH-cytochrome P-450 reductase in the plasma membrane modulates the activation of hypoxia-inducible factor 1. J Biol Chem 277: 23367–23373, 2002. Portilla D, Mandel LJ, Bar-Sagi D, and Millington DS. Anoxia induces phospholipase A2 activation in rabbit renal proximal tubules. Am J Physiol Renal Fluid Electrolyte Physiol 262: F354–F360, 1992. Rohlfs RJ, Bruner E, Chiu A, Gonzales A, Gonzales ML, Magde D, Magde MD, JR, Vandegriff KD, and Winslow RM. Arterial blood pressure responses to cell-free hemoglobin solutions and the reaction with nitric oxide. J Biol Chem 273: 12128–12134, 1998. Schumacker MA, Dixon MM, Kluger R, Jones RI, and Brennan RG. Allosteric transition intermediates modelled by crosslinked haemoglobins. Nature 375: 84–87, 1995. Soltoff SP and Mandel LJ. Active ion transport in the renal proximal tubule. III. The ATP dependence of the Na pump. J Gen Physiol 84: 643–662, 1984. Tessitore N, Sakhrani LM, and Massry SG. Quantitative requirement for ATP for active transport in isolated renal cells. Am J Physiol Cell Physiol 251: C120–C127, 1986.
95 • AUGUST 2003 •
www.jap.org
