INTRODUCTION

Top Abstract Introduction Methods Results Discussion Appendix References

PLASMA LAYERS, 3-10 µm thick, may contribute up to 90% of the resistance to O₂ diffusion in erythrocyte perfused lung capillaries (7). In the kidney, resistance to diffusion within capillaries could limit O₂ delivery to support tubular function, particularly in the corticomedullary region. The corticomedullary region of the kidney normally functions under conditions of near maximal workload with limited O₂ supply and is particularly sensitive to changes in O₂ delivery (6).

Enhanced O₂ diffusivity may explain our observation that 5 g/l of purified hemoglobin (Hb), with a P₅₀ of 10 mmHg (P₅₀ is the partial pressure of O₂ at which the Hb is half saturated), greatly enhanced the function of isolated perfused rat kidneys (1). The improved function could not be explained simply by an increase in arterial O₂ content. An additional factor may have been enhancement of O₂ diffusivity by Hb in solution (13). Diffusion theory predicts O₂ diffusion in proportion to Hb concentration and O₂ affinity (1/P₅₀). Equations describing this relationship are developed in the APPENDIX. Experiments were undertaken to examine the impact of Hb concentration and P₅₀ on O₂ delivery at low PO₂. Cross-linked tetrameric hemoglobins (64 kDa) with P₅₀ of 11 or 35 mmHg were used to provide O₂ delivery to isolated perfused rat kidneys. The cross-linker in each case is a trimesic acid moiety positioned within the 2,3-diphosphoglycerate (DPG) binding site of the hemoglobin. The different P₅₀ values of the products are dictated by the amino acid residues within the DPG site that are involved in the linkage. Effective O₂ delivery was assessed from the reabsorption of Na, K, P_i, glucose, protein, and free water excretion.

	METHODS

Top Abstract Introduction Methods Results Discussion Appendix References

Materials. Hemosol (Etobicoke, ON, Canada) provided the cross-linked hemoglobins, prepared in lactated Ringer solution. Tm-Hb, 82-82'-Hb, and ethanolamine-Hb were prepared as previously described (15, 16, 18). Tm-Hb contains 33% ₂(Val¹)-Tm-(Lys⁸²) and 67% ₂(Val¹,Lys⁸²)-Tm-(Lys⁸²). Tm is the cross-linker trimesic acid. 82-82'-Hb contains 20% ₂(Lys⁸²,Lys¹⁴⁴)-Tm-(Lys⁸²) and 80% ₂(Lys⁸²)-Tm-(Lys⁸²). Ethanolamine-Hb contains 60% ₂(Lys⁸²)-Tm(ethanolamine)-(Lys⁸²). Ethanolamine is conjugated to the free trimesic acid carboxylate group via an amide linkage (15). The remainder is composed mainly of ₂(Lys⁸²,Lys¹⁴⁴)-Tm-(Lys⁸²) and ₂(Lys⁸²)-Tm-(Lys⁸²). O₂ affinity was measured with a Hemox analyzer. P₅₀ at 37°C, pH 7.4, was 11 mmHg for 82-82'Hb. The Hill coefficient was 1.8. Ethanolamine-Hb P₅₀ was 12 mmHg, and the Hill coefficient was 2.0. 82-82'Hb and ethanolamine-Hb produced similar renal function and will be referred to hereafter as Hb-P₁₁. For Tm-Hb, hereafter called Hb-P₃₅, the P₅₀ was 35 mmHg, and the Hill coefficient was 2.4. Solutions were adjusted to 22-24 mmHg oncotic pressure by dilution with appropriately constituted salt solutions containing bovine serum albumin (BSA, low endotoxin; Pentex) dialyzed against a modified Krebs-Henseleit salt solution for 36 h (3). The final perfusate contained (mmol/l) 142 Na, 5 K, 25 HCO₃, 2 Ca, 113 Cl, 14 lactate, 5 glucose, and a mixture of 20 amino acids totaling 6 mmol/l. The pH was 7.4 ± 0.1 at 37°C when equilibrated with 5% CO₂. [³H]inulin was dialyzed against distilled water for 24 h, and a fresh batch was prepared every 4 wk.

Isolated kidney perfusion. Male Wistar rats (Harlan Farms, 250-300 g) were fed Purina Laboratory Rodent Diet 5001 during acclimatization to the laboratory. They were starved overnight with free access to tap water prior to being anesthetized (Somnotol, 50 mg/kg ip) for isolated kidney perfusion, as we have previously described (1). After the right kidney and mesenteric artery were exposed through a mid-line incision, the right ureter was cannulated with PE-50 tubing, and 1 ml of 10% mannitol and 100 U heparin was injected intravenously. A double-lumen catheter was advanced through the mesenteric artery, the perfusion flow was started, and the catheter was inserted into the right renal artery and tied in place. The aorta and vena cava were rapidly cut to free the kidney. After several seconds to permit flushing of blood, the kidney was placed into the warmed cup of a recirculating perfusion apparatus. The surface of the kidney was covered with Parafilm to reduce dehydration. The perfusion system recirculated 160-180 ml of perfusate through a Hollow Fiber dialyzer (Fresenius F4; Fresenius, Bad Homburg, Germany). The perfusion fluid was continuously dialyzed against 1.5 liters of protein-free salt solution. To maintain a constant perfusate volume and protein concentration, the fluid level in the venous reservoir was monitored with an electronic sensor that activated a pump to inject appropriate volumes of protein-free salt solution. In all experiments, the dialyzing fluid was initially equilibrated with 95% air-5% CO₂. After a 40-min perfusion in some experiments, the gas was switched to 5% O₂-5% CO₂-90% N₂. The circuit included in-line borosilicate glass prefiltration filters and cellulose ester filters with 8-µm pore size (Millipore, Bedford, MA). Renal arterial pressure was measured through a no. 30 needle, which passed into the center of the 18-gauge perfusion catheter. Perfusate flow was adjusted to maintain constant arterial pressure of 80 mmHg. As we have done previously (1), we weighed the unperfused left kidney and used it as a reference for perfusion flow rate and inulin clearance, to compensate for any change in the perfused kidney volume and weight that might occur during perfusion with different solutions. At the end of the perfusion, some kidneys were flushed with 50 ml of isotonic salt solution, followed by 10 ml of 10% buffered Formalin. The kidney was then cut in 1-mm-thick slices and fixed in Formalin, and paraffin sections were cut at 5-20 µm for staining with periodic-acid Schiff/1% alcian blue and Perl's Prussian blue method for iron. In some experiments, the kidney was perfused for 5 min with 300 units/ml collagenase (Clostridium histolyticum; Sigma Chemical) with 5% BSA in perfusate solution. The kidney was removed, sliced with a Stadie-Riggs microtome, and incubated for 30 min at 37°C with 300 units/ml collagenase in perfusate solution with BSA and equilibrated with 95% O₂-5% CO₂. The tubule fragments were washed through a sieve with BSA-free perfusate solution, washed with centrifugation three times, and suspended in 43% Percoll, as previously described (2). The Percoll suspension was centrifuged for 30 min at 12,000 g. The fourth layer, which contains >85% proximal tubules, was washed with centrifugation three times. One aliquot of tubules was dissolved in concentrated nitric acid and analyzed for iron by atomic absorption in a Varian Spectra A300 Zeeman Graphite Furnace. Another aliquot was analyzed by the Lowry method for protein content.

O₂ content of arterial perfusate was varied by using different concentrations of cross-linked Hb and either 5% O₂-5% CO₂-90% N₂ or 95% air-5%CO₂. After an equilibrium period of 20 min, urine samples were collected at 20-min intervals up to 120 min. Perfusate samples were collected at the midpoint of each urine collection. Samples (1 ml) were drawn into airtight syringes from the venous outflow and the arterial bypass tube for measurement of pH, PO₂, and PCO₂ in a blood gas analyzer, and O₂ saturation of Hb and methemoglobin was measured by spectrophotometry (Co-oximeter). Hb modified the PO₂ readings from O₂ electrodes; therefore, solutions with various concentrations of Hb were equilibrated with analyzed gas mixtures to produce calibration curves. PO₂ was also measured with a Yellow Springs Instruments micro-oxygen probe in temperature-regulated, flow-through cells at 37°C and recorded on a YSI model 5300 Biological Oxygen Monitor (Yellow Springs Instruments). The detectors were incorporated into the bypass from the arterial inflow and in a catheter that collected part of the venous outflow. Total O₂ content was measured with a LexO₂con-K (Lexington Instruments, Waltham, MA). With samples containing low Hb concentration and low PO₂, we used two to four times the recommended sample volume of 20 µl for analysis; readings were linearly related to sample volume.

O₂ delivery was calculated as arterial perfusate O₂ (µmol/ml) multiplied by perfusate flow (ml/min). O₂ consumption was calculated as perfusate flow multiplied by the difference between arterial and venous O₂ content. Perfusate flow was measured with an electromagnetic flow transducer in the arterial line. Urine and perfusate samples were analyzed for Na, K, P_i, glucose, osmolality, protein, and [methoxy-³H]inulin (NEN Products, Boston, MA). Inulin clearance was calculated as urine ³H excretion (dpm/min)/³H in perfusate (dpm/ml). Perfusate oncotic pressure was measured at the beginning and end of each experiment (Refractometer and/or Weil Oncometer System IL-186, Instrument Laboratories). BSA concentrations in perfusate and urine were measured using an high-performance liquid chromatography (HPLC) assay incorporating an anion exchange column (Pharmacia Mono-Q, 1-ml capacity), using a Beckman System Gold and a salt gradient. Total Hb was determined after HPLC by optical density at 280 and 414 nm. Percent excretion of Na, K, P_i, glucose, BSA, and Hb were calculated as (urine excretion/filtered load) × 100, with filtered load equal to perfusate concentration of the relevant solute multiplied by inulin clearance. There was no urea in the perfusate; therefore, most of the urine osmolality was due to Na and K salts, and we calculated free water clearance as C_H2O/C_inulin = 1  (urine_Na+K/perfusate_Na + K)/(urine_inulin/perfusate_inulin).

Statistical analysis was by paired and unpaired t-tests and one-way ANOVA with Student-Newman-Keuls method for multiple comparisons and by two-way ANOVA. For nonparametric data, Kruskal-Wallis ANOVA was used with Dunn's method for pair-wise multiple comparisons. The SigmaStat program was used for these analyses.

	RESULTS

Top Abstract Introduction Methods Results Discussion Appendix References

The O₂ binding characteristics of Hb-P₃₅ and Hb-P₁₁ before and during perfusion are compared in Fig. 1. To create the lines in Fig. 1, we measured O₂ saturation before perfusion with a spectrophotometer (Hemox analyzer at 37°C, pH 7.4). The points for O₂ saturation in arterial and venous samples obtained after 100 min of perfusion were calculated from the maximum O₂ binding capacity of hemoglobin-methemoglobin and the measured O₂ content of each sample. Calculated O₂ bound to Hb at 100% saturation (mmol/l) = Hb (mg/ml) × [(1  %methemoglobin/100) × 0.062] (10). Methemoglobin concentration rose from 10% of the total Hb at the beginning of perfusion to between 25 and 35% after 100 min. Total O₂ content was measured by galvanometry with the LexO₂con, and O₂ bound to Hb was calculated by subtracting dissolved O₂ [PO₂ (mmHg) × 1.257 × 10³ (mmol/l)] (10). Perfusion for 100 min shifted P₅₀ to the left, with a more pronounced shift for Hb-P₃₅. The shift in P₅₀ is consistent with the known effect of methemoglobin to increase O₂ affinity (9).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Oxyhemoglobin dissociation curves for Hb-P35 ( and dashed line) and Hb-P11 ( and solid line). Lines were calculated using the Hill equation. A Hemox analyzer was used to obtain P50 and cooperativity values at 37°C, pH 7.4, for Hb solutions before perfusion. Methemoglobin concentration in these samples was 10.5%. Individual points represent results for arterial or venous samples obtained after 120 min of perfusion. %Saturation was calculated as described in RESULTS section, using the measured O2 content and the Hb-methemoglobin concentration. Calculated O2 bound to Hb at 100% saturation (mmol/l) = Hb (mg/ml) × [(1  %methemoglobin/100) × 0.625]. Total O2 content was measured by galvinometry with a LexO2con-K O2 analyzer, and O2 bound to Hb was calculated by subtracting dissolved O2 (PO2 × 1.257 × 103 mmol/l) (10) from total O2 content.

The relationship between Hb concentration and renal function was explored with Hb-P₃₅. Arterial O₂ content was proportional to Hb concentration and almost twice as high when 95% air rather than 5% O₂ was used (Fig. 2). Perfusate flow increased during the first 30 min of perfusion; thereafter, flow and renal vascular resistance were similar in all experiments and were not related to the concentration of Hb, arterial O₂ content, or arterial PO₂. Glomerular filtration rate (GFR) decreased and fractional Na excretion increased over the course of the 2-h experiments (data not shown). The data shown in the Figs. 1-7 and Tables 1-3 were obtained between 80 and 100 min. Filtration fraction, GFR, and total Na reabsorption (T_Na) increased with increasing O₂ delivery and plateaued above delivery rates >16 µmol · min¹ · g¹. O₂ consumption increased with O₂ delivery up to 30 µmol · min¹ · g¹, which was the highest delivery rate obtained (Fig. 2). Percent Na reabsorption was not significantly altered by changes in O₂ delivery (Fig. 2). In contrast, percent excretion of glucose, phosphate, protein, and hemoglobin decreased until O₂ delivery exceeded 20-25 µmol · min¹ · g¹ (Fig. 3).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   O2 delivery (µmol · min1 · g1, arterial O2 content × flow/g) by Hb-P35 (0.5-3%), equilibrated with 95% air or 5% O2, compared with the following. Top right: Na reabsorption (TNa) [(GFR/g × perfusate Na)  (urine flow/ g × urine Na)] (open circles) and % fractional Na excretion (solid squares). Bottom right: O2 consumption (µmol · min1 · g1) [flow/g × (arterial O2 content  venous O2 content)]. Top left: glomerular filtration rate (GFR, inulin clearance/g). Bottom left: %filtration fraction [(GFR/perfusate flow) × 100]. Values are means ± SE for the 80-100th min of perfusion. Arterial PO2, Hb concentration, and no. of experiments (in parentheses) are shown top left.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   O2 delivery (µmol · min1 · g1, arterial O2 content × flow/g) by Hb-P35 (0.5-3%), equilibrated with 95% air or 5% O2, compared with the following. Top right: %excretion BSA [(100 × urine BSA × urine flow)/(GFR × perfusate BSA)]. Bottom right: %excretion of Hb. Top left: %excretion of glucose [(100 × urine glucose × urine flow)/(GFR × perfusate glucose)]. Bottom left: %excretion of phosphate [(urine phosphate × urine flow)/(GFR × perfusate phosphate)]. Values are means ± SE for the 80-100th min of perfusion. Arterial PO2, Hb concentration, and no. of experiments are shown top left.

We used 1% Hb to examine the interactions between P₅₀ and PO₂. Venous PO₂ was significantly lower with Hb-P₁₁ (P < 0.001, Table 1). Neither P₅₀ nor PO₂ influenced perfusate flow (Table 1) or GFR (Fig. 4) during the period from 60 to 120 min of perfusion (Fig. 5). The relationship between O₂ consumption and Na reabsorption was linear and virtually identical for the two types of Hb (Fig. 6). However, the low P₅₀ reduced both phosphate and glucose excretion (Figs. 4 and 5). Phosphate excretion was also sensitive to changes in PO₂ and O₂ delivery, but the effect of PO₂ on glucose reabsorption was negligible, when the same concentration of Hb-P₁₁ and Hb-P₃₅ was used. Reduction of percent Na excretion by Hb-P₁₁ cannot be ruled out (P = 0.08). In other respects, renal function with high- and low-affinity Hb was indistinguishable: percent K excretion, percent protein excretion, urine osmolality, and free water clearance were not different (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Comparison of renal function in kidneys perfused with either 1% Hb-P11 or 1% Hb-P35, equilibrated with either 5% O2 or 95% air. Data were analyzed by two-way ANOVA for the effects of PO2 and P50. Top right: %Na excretion. Bottom left: %phosphate excretion. Top left: inulin clearance (GFR) (ml · min1 · g1). Bottom right: %glucose excretion. Data are means ± SE for the 80-100th min of perfusion. Number of experiments: 9 for Hb-P11 with 5% O2, 7 with 95% air; 11 for Hb-P35 with 5% O2, 7 with 95% air.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Comparison of inulin clearance (ml · min1 · g body wt1) and percent phosphate excretion between 60 and 120 min of perfusion with either 1% Hb-P11 or 1% Hb-P35 equilibrated with either 5% O2 or 95% air. Data are means ± SE and analyzed by unpaired t-test. No. of experiments: 9 for Hb-P11 with 5% O2, 7 with 95% air; 11 for Hb-P35 with 5% O2, 7 with 95% air. Inulin clearance was not significantly different. Percent phosphate excretion was lower in kidney perfused with Hb-P11 at all times (P < 0.05 to < 0.01).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   O2 consumption [(arterial O2 content  venous O2 content) × perfusate flow/g] and Na reabsorption [(GFR/g × perfusate Na)  (urine flow/g × urine Na)]. Solid circles, 1% Hb-P35 equilibrated with 95% air or 5% O2. Open circles, 1% Hb-P11 equilibrated with 95% air or 5% O2. Values were calculated for the 80-100th min of perfusion.

For comparison with the effects of different P₅₀ values, we examined the interactions between arterial PO₂ and concentration using 1 and 2% Hb-P₃₅. As expected, O₂ delivery to the kidney was doubled by using 2% Hb (Table 2). Venous PO₂ may have been slightly higher when kidneys were perfused with 2% Hb-P₃₅ (Table 2, P = 0.06). Increasing Hb concentration from 1 to 2% was associated with increased GFR (Fig. 7), but GFR did not increase further when the concentration was raised to 3% (Fig. 2). The decrease in phosphate excretion was a function of both higher PO₂ and higher Hb concentration. In contrast, the decrease of fractional glucose excretion was attributable to increased Hb concentration alone. If one compares Figs. 4 and 7, it appears that 1% Hb-P₁₁ had the same capacity as 2% Hb-P₃₅ to support phosphate and glucose reabsorption.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Comparison of renal function in kidneys perfused with either 1% Hb-P35 or 2% Hb-P35 equilibrated with either 5% O2 or 95% air. Data analyzed by two-way ANOVA for the effects of PO2 and P50. Top right, %Na excretion; bottom left, %phosphate excretion; top left, inulin clearance (GFR) (ml · min1 · g1); bottom right, %glucose excretion. Data are means ± SE for the 80-100th min of perfusion. No. of experiments: 11 for 1% Hb-P35 with 5% O2, 7 with 95% air; 4 for 2% Hb-P35 with 5% O2, 4 with 95% air.

Distal nephron function as reflected by percent K excretion was not influenced by P₅₀ (Tables 1 and 2) but was slightly reduced by increasing the concentration of Hb-P₃₅ (Table 2). Percent free water clearance (C_H2O/C_inulin) was not infleunced by either P₅₀ or Hb concentration.

Filtration and tubular uptake of Hb was examined by staining tissue sections for iron and by measuring the iron content of proximal tubules. Thick sections (20 µm) of paraffin-embedded tissue were examined for residual iron after the kidney had been flushed with 50 ml of saline and 10 ml of 10% buffered Formalin. There were very few scattered blue granules and no evidence of iron in the tubular cells, tubular lumens, or interstitium. Iron content of proximal tubules from kidneys perfused with 1-2% Hb-P₃₅ was similar to that in tubules from kidneys perfused with only BSA for 2 h (Table 3). There was no correlation between concentration of Hb-P₃₅ in the perfusate and the proximal tubular iron content. Brush border and cellular structure was intact by light-microscopic examination in kidneys perfused with either Hb-P₁₁ or Hb-P₃₅ at low or high PO₂.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion Appendix References

Is O₂ diffusivity within capillaries rate limiting for renal function? To examine this question, we exploited the fact that Hb in solution increases O₂ diffusivity in proportion to its concentration and O₂ affinity (1/P₅₀) (13). Adding two different cross-linked Hbs with a threefold difference in O₂ affinity to the kidney perfusate altered O₂ diffusivity separately from arterial O₂ content and delivery. The impact of P₅₀ on diffusivity increases as PO₂ approaches zero. Equation A16 (see APPENDIX) predicts that, at PO₂ levels of 5 mmHg, 1% Hb-P₁₁ would increase O₂ diffusivity by 28%, and 1% Hb-P₃₅ would increase O₂ diffusivity by 11%. Theory also predicts that doubling the Hb concentration will not only double the quantity of O₂ carried but also the effect on diffusivity.

Na reabsorption drives the bulk of renal O₂ consumption (12). Hb-O₂ affinity did not alter the linear relationship between O₂ consumption and Na reabsorption (Fig. 6) or between GFR and total Na reabsorption (Fig. 4); however, we cannot rule out a slightly higher fractional Na reabsorption (P = 0.084) in the Hb-P₁₁ perfused kidneys. GFR and T_Na were increased by changing from 1 to 2% Hb (Fig. 7), with no change in the ratio of Na reabsorption to O₂ consumption (T_Na/QO₂; ANOVA, P = 0.975). Increased GFR might have been due to efferent arterial vasoconstriction related to scavenging of NO (1), although there was no significant difference in total vascular resistance, perfusate flow (Table 2), or filtration fraction (P = 0.185, ANOVA). GFR and T_Na appeared to plateau above O₂ delivery rates of 16 µmol · min¹ · g¹ and were unaffected by increasing Hb-P₃₅ concentration to 3%.

Na reabsorption coupled to glucose and phosphate transport is sensitive to small decreases in O₂ supply (11) in proximal tubules. The late proximal tubule (S3 segment), in which Na transport is coupled 2:1 with glucose, is most sensitive to O₂ deprivation. This segment is also most susceptible to hypoxic damage (4, 5). High O₂ affinity (Hb-P₁₁) significantly reduced glucose and phosphate excretion (Figs. 4 and 5). To obtain similar low rates of phosphate and glucose excretion with Hb-P₃₅, it was necessary to increase O₂ delivery and diffusivity twofold by doubling Hb concentration in the perfusate (Fig. 7). Phosphate and glucose are largely reabsorbed by high-capacity, low-affinity transporters in the S1 and S2 segments. Low-capacity, high-affinity transporters in the S3 segment are responsible for producing low urinary concentration of glucose and phosphate (19). High-affinity 2:1 sodium-glucose cotransport in the S3 segment consumes twice as much ATP as 1:1-coupled reabsorption in the early proximal tubule (19). K_m for glucose reabsorption in S1-2 segments is 1.6 mM, and, in the S3 segment, it is 0.35 mM (19). Kidneys perfused with 1% Hb-P₁₁ produced urine glucose concentrations well below the K_m for the S3 segment (0.17 mM; 0.16-0.37 mmol/l; median, 25th-75th percentile). Kidneys perfused with 1% Hb-P₃₅ produced urine glucose concentrations that were twice the K_m value for the S3 segment (0.77 mM; 0.61-1.11 mmol/l) (P < 0.001). The S3 segment also contains a high-affinity, low-capacity phosphate transporter (17). These results strongly suggest that Na-coupled glucose and phosphate reabsorption in the S3 segment were significantly greater with Hb-P₁₁ than with Hb-P₃₅. Kidneys perfused with Hb-P₁₁ and Hb-P₃₅ had similar O₂ delivery and consumption rates, but O₂ delivery to S3 segments in the corticomedullary region was presumably enhanced with Hb-P₁₁, as indicated by the difference in glucose and phosphate reabsorption.

Delivery of O₂ for ATP production and Na transport depends on the steep O₂ gradient from capillaries to mitochondria and is influenced by resistance to O₂ diffusion (20). Inner cortical capillary PO₂ is unlikely to have been higher with Hb-P₁₁ than with Hb-P₃₅, because venous PO₂ was significantly lower (Table 1). Nonetheless, improved Na-coupled reabsorption strongly suggests that O₂ delivery to tubules in the corticomedullary region was increased, due to improved O₂ diffusivity. The effect of cross-linked Hb on O₂ diffusion probably occurred almost exclusively within vascular lumens, which is where up to 90% of the resistance to O₂ diffusion has been found in 3- to 10-µm layers adjacent to the walls of erythrocyte perfused capillaries (7). The uniform distribution of cross-linked Hb throughout the capillary lumen could facilitate diffusion of O₂ to capillary walls. It is unlikely that significant amounts of Hb passed from capillaries into the interstitial space, since no iron staining was seen in the renal interstitium, very little Hb passed through glomerular capillaries into the urine (Fig. 3), and there was no evidence of tubular iron uptake (Table 3).

Reduced resistance to O₂ diffusion is probably insufficient to completely account for improved O₂ delivery in Hb-P₁₁ perfusions. High Hb O₂ affinity might favor even distribution of PO₂ throughout the cortex and outer medulla by reducing release of O₂ in the usually well-oxygenated outer cortex and increasing availability of O₂ in the more hypoxic corticomedullary region. The combination of redistribution and facilitated diffusion enabled Hb-P₁₁ to deliver O₂ to the corticomedullary region without significantly depriving the outer cortex of O₂. This is shown by the similarity of GFR and T_Na obtained with 1% Hb-P₁₁ and 1% Hb-P₃₅ (Figs. 4 and 5).

If cross-linked hemoglobins are used clinically, it will be in patients at risk for ischemic acute tubular necrosis because of hemorrhagic shock and low arterial PO₂. Our observations indicate that Hb O₂ affinity will significantly influence kidney function under these conditions. The results demonstrate that the relationship between P₅₀ and O₂ delivery is different when Hb is free in solution, rather than encapsulated within erythrocytes. Hb in solution facilitates O₂ diffusion through plasma to the capillary wall, and Hb with a high O₂ affinity (low P₅₀) facilitates O₂ diffusion more than Hb with lower O₂ affinity. In contrast, lowering Hb P₅₀ within erythrocytes lowers the PO₂ at which O₂ is released into the plasma and decreases the gradient that drives O₂ diffusion to the mitochondria (14).

The corticomedullary region, which includes the S3 proximal tubules, is most at risk for hypoxic damage (5). Under hypoxemic conditions, Hb-P₁₁, with high O₂ affinity, improved O₂ delivery to tubular cells in the corticomedullary region, without evidence of an adverse effect on cortical or distal tubular function. This beneficial effect was probably related to enhanced O₂ diffusivity at low PO₂ and may also have been related to redistribution of O₂ from the cortex to the corticomedullary junction (13). We conclude that O₂ diffusivity can be rate limiting for O₂ delivery when PO₂ is low. Furthermore, cross-linked Hb with a low P₅₀ should be more effective in preserving kidney function under conditions of hypoxia than cross-linked Hb with P₅₀ similar to that found in erythrocytes.