A Causative Role for Redox Cycling of Myoglobin and Its
Inhibition by Alkalinization in the Pathogenesis and Treatment of
Rhabdomyolysis-induced Renal Failure*
(Received for publication, June 11, 1998, and in revised form, August 27, 1998) Kevin P. Moore‡§, Steve G. Holt‡, Rakesh P. Patel¶, Dimitri A. Svistunenkoi, William Zackert**, David Goodier‡, Brandon. J. Reederi, Martine Clozel‡‡, Radhi Anand‡, Christopher E. Cooperi, Jason D. Morrow**, Michael T. Wilsoni, Victor Darley-Usmar¶, and L. Jackson Roberts II** From the ‡Joint Department of Medicine, Royal Free and University College Medical School, London NW3 2QG, United Kingdom, the¶Center for Free Radical Biology and Department of Pathology, Molecular and Cellular Division, University of Alabama, Birmingham, Alabama 35243, the **Departments of Pharmacology and Medicine, Vanderbilt Medical Center, Nashville, Tennessee 37232, ‡‡Preclinical Research, Hoffmann-La Roche, Basel 4070, Switzerland, and the iDepartment of Biological Sciences, University of Essex, Colchester, Essex CO4 3SQ, United Kingdom
Muscle injury (rhabdomyolysis) and subsequent dep-osition of myoglobin in the kidney causes renal vasocon-striction and renal failure. We tested the hypothesis that myoglobin induces oxidant injury to the kidney and the formation of F2-isoprostanes, potent renal vasoconstric-tors formed during lipid peroxidation. In low density lipoprotein (LDL), myoglobin induced a 30-fold increase in the formation of F2-isoprostanes by a mechanism in-volving redox cycling between ferric and ferryl forms of myoglobin. In an animal model of rhabdomyolysis, uri-nary excretion of F2-isoprostanes increased by 7.3-fold compared with controls. Administration of alkali, a treatment for rhabdomyolysis, improved renal function and significantly reduced the urinary excretion of F2 -isoprostanes by;80%. EPR and UV spectroscopy dem-onstrated that myoglobin was deposited in the kidneys as the redox competent ferric myoglobin and that it’s concentration was not decreased by alkalinization. Ki-netic studies demonstrated that the reactivity of ferryl myoglobin, which is responsible for inducing lipid per-oxidation, is markedly attenuated at alkaline pH. This was further supported by demonstrating that myoglo-bin-induced oxidation of LDL was inhibited at alkaline pH. These data strongly support a causative role for oxidative injury in the renal failure of rhabdomyolysis and suggest that the protective effect of alkalinization may be attributed to inhibition of myoglobin-induced lipid peroxidation.
In 1941 Bywaters and Beall (1) described four patients who developed acute renal failure, associated with dark brown uri-nary granular casts following crush injury. This syndrome has subsequently been termed rhabdomyolysis, since it occurs as a consequence of massive breakdown of muscle. It is usually associated with trauma, but may occur in a variety of clinical settings such as hyperthermia, seizures, muscle ischemia, or
exposure to toxins such as alcohol or drug overdose. The asso-ciated muscle breakdown releases myoglobin (Mb)1 into the circulation, which then becomes deposited in the kidney caus-ing renal failure in 30% of cases. Rhabdomyolysis accounts for 7% of cases of acute renal failure in the United States (2, 3).
The mechanism of the renal failure in rhabdomyolysis has been attributed to both intense renal vasoconstriction and re-nal tubular necrosis (4 – 6). Recent studies have suggested a potential role for free radicals in the pathogenesis of myoglo-binuria-induced renal failure based on the findings that in rats with rhabdomyolysis, levels of malondialdehyde in the kidney are increased (2–3-fold). There is an induction of antioxidant enzymes and consumption of antioxidants, and administration of antioxidants partially protects against the renal failure (7– 11). The modest increases of malondialdehyde levels are diffi-cult to interpret, since it is well recognized that measurements of aldehyde lack specificity as a marker of lipid peroxidation, and others have failed to replicate these findings (12). To ex-plain how myoglobinuria could induce an oxidant injury to the kidney, it has been suggested that Mb may release free ferrous (Fe21) iron, which could then induce lipid peroxidation as a result of the generation of hydroxyl radicals via the Fenton reaction (13, 14). However, it has been difficult to demonstrate hydroxyl radical production in the kidneys of animals with myoglobinuria and administration of desferrioxamine at doses that effectively chelate all free iron in the urine is only partially effective in preventing renal failure (15).
An alternative mechanism could involve the heme group in Mb itself, which can redox cycle between different oxidation states and promote lipid peroxidation reactions as a conse-quence of it’s ability to decompose lipid hydroperoxides to per-oxyl and alkper-oxyl radicals (16). This is achieved catalytically by a redox cycle between the ferric (Fe31) and ferryl ([Fe5O]21) Mb oxidation states, the latter of which can directly initiate lipid oxidation (17, 18) (Fig. 1). In muscle cells, the iron in Mb is maintained predominantly in the reduced ferrous (Fe21) state, which is unable to participate in such reactions. How-ever, the redox state of Mb in renal tubules has not been determined. We hypothesized that Mb in the extracellular mi-lieu, i.e. in renal tubules, would undergo autooxidation to the ferric form (MetMb), which is then catalytically competent to * This work was supported by the Medical Research Council and the
Wellcome Trust, London, United Kingdom and by National Institutes of Health Grants GM42056 and DK48831. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: Dept. of Medicine, Royal Free and University College Medical School, Rowland Hill St., London NW3 2QG, UK. Tel.: 44-171-830-2950; Fax: 44-171-794-3472; E-mail: [email protected].
1The abbreviations used are: Mb, myoglobin; MetMb, metmyoglobin;
IsoP, isoprostane; PG, prostaglandin; 13-HPODE, 13(S)-hydroperoxy-9(E),11(Z)-octadecadienoic acid; LDL, low density lipoprotein; DTPA, diethylenetriaminepentaacetic acid; REM, relative electrophoretic mobility.
© 1998 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.
This paper is available on line at http://www.jbc.org
31731
promote lipid peroxidation reactions. Importantly, unlike the Haber-Weiss/Fenton reactions involving free iron, this reaction is not dependent on the generation of superoxide or hydrogen peroxide; it can be driven solely by endogenous lipid hydroper-oxides (17–20). This hypothesis may also explain the partial efficacy of desferrioxamine to protect against renal failure in the rat model of rhabdomyolysis, since desferrioxamine is ca-pable of reducing ferryl-Mb and its associated globin radical to the ferric form (21, 22). Interestingly, nitric oxide also partially protects against the renal failure (23). In accord with our hy-pothesis, this might be explained by an effect of nitric oxide to inhibit heme-dependent lipid peroxidation (24).
Mb-induced lipid peroxidation could explain the occurrence of renal tubular necrosis in myoglobinuria, but how this could be causally linked to the reduction in renal blood flow that also characterizes the nephropathy of rhabdomyolysis is not imme-diately obvious. One hypothesis that could be advanced is that vasoactive products of lipid peroxidation are generated that cause renal vasoconstriction. In this regard, one of the most attractive candidates would be the isoprostanes (IsoPs), which we previously identified as a group of prostaglandin (PG)-like compounds that are formed nonenzymatically in vivo as prod-ucts of free radical-induced peroxidation of arachidonic acid (25). PGF2-, E2-, and D2-like compounds are formed in vivo by this process (25, 26). Importantly, two IsoPs formed in abun-dance, namely 15-F2t-IsoP (27) (8-iso-PGF2a) and 15-E2t-IsoP (8-iso-PGE2), are very potent renal vasoconstrictors that ap-pear to interact with a unique receptor (25, 28, 29).
An intriguing question then arises as to whether an effect on oxidant injury could be mechanistically linked with the obser-vation that alkalinization of the urine significantly attenuates the renal dysfunction (2, 12). A plausible and accepted expla-nation is that alkalinization enhances the solubility of Mb and thus its elimination from the kidney (12). However, we consid-ered an alternative explanation based on the recent demonstra-tion that Mb-induced lipid peroxidademonstra-tion reacdemonstra-tions are acceler-ated at acidic pH (30).
In this study, therefore, we examined the role of oxidative injury to the kidney in the pathogenesis of the renal failure associated with rhabdomyolysis in a rat model by measuring the specific marker of lipid peroxidation, F2-IsoPs, which has emerged as one of the most reliable approaches to assess oxida-tive stress status in vivo (31). For this purpose, F2-IsoPs were measured both in urine and in renal lipids. A novel aspect of the formation of IsoPs is that they are initially formed in situ ester-ified to tissue lipids and subsequently released (32). Since IsoPs generated in renal lipids would, in part, be directly released into the urine, measurement of urinary F2-IsoPs can provide an index of lipid peroxidation that occurs in the kidney (31). More direct
evidence localizing oxidant injury to the kidney can also be ob-tained by measurement of levels of F2-IsoPs esterified in renal lipids. Furthermore, we explored the hypothesis that the protec-tive effect of alkalinization operates by stabilizing ferryl-Mb, thus reducing its potential to induce lipid peroxidation.
MATERIALS AND METHODS
Preparation and Oxidation of Human LDL—Human LDL was
iso-lated by differential centrifugation (33). Horse heart Mb (determined to be MetMb) was purchased from Sigma, Poole, Dorset, United Kingdom. MetMb (0 –10mMfinal concentration) was added to LDL (50 or 100 mg/ml) in Hanks’ buffered salts solution in which the pH was adjusted as indicated and incubated at 37 °C for the time specified. Ferryl-Mb was prepared immediately before use by mixing 15mMhydrogen
per-oxide with 10mMMetMb. Control experiments involved incubation in phosphate-buffered saline alone or together with hydrogen peroxide (15 mM). The reaction was terminated by addition of butylated hydroxytolu-ene (100mM). All experiments were performed in the presence of the
iron chelator, DTPA (diethylenetriaminepentaacetic acid) (10mM), to exclude free iron-mediated oxidation. Electrophoretic mobility of LDL was determined using a lipoprotein electrophoresis system (Beckman Co). Gels were fixed, stained with Sudan black B, and results related to mobility of unoxidized LDL (33).
Animal Experiments—The rat model of glycerol-induced
rhabdomy-olysis used has been described previously (12). Animals were divided into three groups designated as controls, rhabdo, or rhabdo-alk. Control animals (n5 8) were untreated. Animals in the rhabdo group (n 5 6) were injected with glycerol (10 ml/kg of a 50% solution in saline), which produces a nonlethal form of renal failure associated with heavy myo-globinuria. The third group (rhabdo-alk, n5 6) was retreated for 24 h with water containing potassium bicarbonate (150 mM) and injected with potassium bicarbonate (2.5 ml/kg of 150 mMpotassium
bicarbon-ate) 1 h prior to injection with glycerol. Urine was collected from 0 to 24 h after injection of glycerol, after which the animals were sacrificed and kidneys snap-frozen in liquid N2for measurement of esterified
F2-IsoPs, or frozen in isopentane (at256 °C), and stored at 280 °C. For
the purpose of measuring renal esterfied F2-IsoPs a further three
ani-mals in the rhabdo group and a further four aniani-mals in the rhabdo-alk group were also studied. Kidney frozen in isopentane was embedded in OCT medium (Tissue-Tek, Miles Inc., IN) and 10-mm frozen sections cut onto glass slides for spectrophotometric analysis.
Extraction and Measurement of F2-IsoPs—Esterified F2-IsoPs in LDL
and kidney and free F2-IsoPs in urine were extracted and quantified by
stable isotope dilution mass spectrometric assay as described, except [2H
4]8-iso-prostaglandin F2a(Cayman Chemical Co., Ann Arbor, MI)
was used as the internal standard (34). Urinary excretion of F2-IsoPs
was normalized to creatinine clearance to correct for the widely differ-ing degrees of renal function, i.e. excretion rate/creatinine clearance. This is calculated as shown below and results expressed as pg/ml Cr.Cl. (pg of F2-IsoPs/ml of creatinine clearance):[urinary F2-IsoPs (pg/ml)]3
[plasma creatinine (mM)]/[urine creatinine (mM)].
Optical Spectroscopy Analysis of Mb in the Kidney—Examination,
under a microscope of thin (10mm) sections of rhabdomyolytic kidney showed, in contrast to normal kidney sections, a red/brown pigment to be located within the tubules. This pigment could easily be obtained from the mounted section by washing the 10-mm sections with 200 ml of sodium phosphate buffer, pH 7.4, containing 10mMDTPA. This extract was diluted to 1 ml with the same buffer. The spectrum of the extract from the kidney was compared with purified MetMb. Sodium dithionite (1 mM) was added to produce the deoxy form for comparison with
purified deoxy-Mb. Optical spectroscopy was performed using a Varian Cary 5E UV-visible spectrophotometer.
Electron Paramagnetic Resonance (EPR) Spectroscopy Analysis of Mb in the Kidney—Rat kidney was cut into pieces of similar size and frozen
in liquid nitrogen. One EPR sample, always made of one kidney, con-sisted of 20 –30 such pieces (;0.3 g). When placed directly into a Bruker finger Dewar (inner diameter5 4.5 mm) without using an EPR tube, such a sample occupied the volume of the Dewar, the length of which was always greater than the working zone of the cavity, 22 mm. This ensured that an EPR spectrum was always taken of the same volume (the cavity working zone) fully filled with the sample, the packing of this volume with tissue pieces being the only variable factor. This packing factor, however, was not a problem, since an EPR spectrum taken of the same sample after reloading into the Dewar (thus provid-ing a different packprovid-ing) differed by less than 5% in signal intensity. Furthermore EPR spectra of samples made from different parts of the same kidney differed by less than 10% in signal intensity. The tissue
FIG. 1. The mechanism for Mb-induced lipid peroxidation and
pieces were retained in the finger of the Dewar with a quartz rod to avoid movement of the pieces caused by liquid nitrogen boiling. Since sufficient kidney tissue was used to extend the sample beyond the working zone, the rod did not cause a problem with either frequency tuning or induce any background signals. All EPR measurements were performed at 77 K. The EPR spectra were measured on a Bruker EMX spectrometer with an ER 041XG microwave bridge (X-band). A 4103TM cavity was used. The experimental conditions comprised: microwave frequency, 9.2689 GHz; microwave power, 3.16 milliwatts; modulation frequency, 100 kHz; modulation amplitude, 7 G; time constant, 0.041 s; sweep time, 84 s; receiver gain, 104; number of scans, 4; data points, 2048/scan.
Where integration of the EPR signals was possible, the concentra-tions are quoted inmMby comparing the double integrals to that of
known standards separately measured in an EPR tube (35). Due to uncertainties in comparing signal intensities from different types of EPR sample (tissue slice placed directly in the Dewar versus frozen standard in EPR tube), we estimate that the systematic error of these absolute measurements (in terms ofmMspins) could be as high as 50%, but the relative intensities of the signals in controls versus rhabdomyo-lytic kidneys are accurate to within 5%. When integration of the EPR signal was not straightforward due to overlapping signals, the signal intensity was normalized to the average intensity seen in the control rats. It is important to note that it is not the peak intensity, but pure line shape second integral, that is proportional to concentration. The low spin ferric signals are much wider than the rhombic signal at g5 4.3; therefore, at equal concentrations of the two centers, the peak intensity of the low spin signal will be smaller. In addition the EPR absorbance “efficiency” of the paramagnetic centers (EPR signal second integral/concentration) depends on the respective g factors.
Assessment of the pH Dependence of Lipid Hydroperoxide Consump-tion by Mb and Decay of Ferryl-Mb—A soluConsump-tion of MetMb (10mM) was
prepared and the reaction initiated by addition of 40 mM 13(S)-hy-droperoxy-9(E),11(Z)-octadecadienoic acid (13-HPODE) (Cascade Bio-chem, Reading, United Kingdom). Consumption of 13-HPODE was followed by monitoring the loss of the conjugated diene chromophore at 234 nm, which occurs as 13-HPODE is converted to an epoxy peroxyl radical (Fig. 1). The single exponential progress curves were fitted by a regression analysis. The pH dependence of the decay of ferryl-Mb was determined by mixing 10mMMetMb with 20mMhydrogen peroxide to
form ferryl-Mb, and the decay of ferryl-Mb was followed by monitoring the increase in absorbance at 408 nm due to reformation of MetMb.
Statistics—Urinary F2-IsoP excretion in animals and esterified F2
-IsoPs in LDL were compared between groups using the nonparametric Mann Whitney U test and were considered significant when p, 0.05. Where mean values are reported, the S.E. is also indicated. For changes in creatinine clearance, the unpaired t test was used.
RESULTS
Formation of F2-IsoPs during Oxidation of LDL by
MetMb—We initially examined the capacity of Mb to induce the formation of F2-IsoPs using LDL as a model lipid containing system. MetMb (10 mM) caused marked lipid peroxidation of
LDL resulting in a;50-fold increase over base line in esterified levels of F2-IsoPs (p, 0.01). This was accompanied by a cor-responding increase in the relative electrophoretic mobility of the LDL, indicative of significant oxidative modification of ApoB-100 (Table I). In a separate experiment ferryl-Mb (10 mM), MetMb (10mM), or hydrogen peroxide (15mM) were
incu-bated with LDL for a period of 8 h to examine the relative efficacy of these redox forms of the heme protein in forming F2-IsoPs (17). Both ferryl-Mb and MetMb increased esterified F2-IsoP levels in LDL (Table II). However, ferryl-Mb treatment resulted in the formation of approximately 3-fold higher con-centrations of F2-isoprostanes than MetMb. Table II also shows that ferryl-Mb treatment resulted in a greater degree of ApoB-100 modification confirming pervious literature reports that ferryl-Mb is a more potent inducer of lipid peroxidation than MetMb (17). For comparison with Table I it should be noted that the greater amounts of F2-isoprostanes generated in the first experiment are due to a prolonged incubation period (24 h) compared with the data in Table II (8 h).
Studies in the Rat Model of Rhabdomyolysis—Following in-duction of rhabdomyolysis in the rat, there was a profound reduction (mean 76.4%) in creatinine clearance in the rhabdo group compared with the control group (p , 0.02, Fig. 2). Alkalinization of the urine significantly attenuated the fall in creatinine clearance (p, 0.02, rhabdo versus rhabdo-alk, Fig. 2A). The urinary excretion of F2-IsoPs was also markedly in-creased (mean 7.3-fold) in the rhabdo group compared with the control group (p, 0.01, Fig. 2B). Of considerable interest was the finding that the enhanced urinary excretion of F2-IsoPs in the rhabdo group was markedly suppressed (mean 78%) by alkalinization (p, 0.01, rhabdo versus rhabdo-alk) to an extent that the difference between urinary levels of F2-IsoPs in the control and rhabdo-alk groups were not significant (p5 0.17, Fig. 3B). Data for urinary F2-isoprostanes were corrected for creatinine clearance, since we have found that urinary excre-tion rate decreases in parallel to creatinine clearance.2 By comparison, if corrected for urinary creatinine alone, the mean urinary excretion was 336 6 pg/mg creatinine in the controls and 2036 42 and 26 6 12 pg/mg creatinine in the rhabdo and rhabdo-alk groups, respectively (p , 0.02 for rhabdo versus other groups). Urinary excretion rates of F2-IsoPs were 196 4, 196 6, and 14 6 2 ng/day in the control, rhabdo, and rhabdo-alk groups, respectively (no significant difference between groups). However, the data when presented as urinary excre-tion rate alone does not reflect the obvious increase in lipid peroxidation observed in the kidney. Thus, levels of F2-IsoPs esterified in renal lipids were also found to be strikingly ele-vated (mean 7.5-fold) in the rhabdo group compared with the control group (p, 0.01, Fig. 3) and was significantly attenu-ated by alkalinization (p, 0.05). These findings unequivocally
2K. P. Moore, S. G. Holt, R. P. Patel, D. A. Svistunenko, W. Zackert,
D. Goodier, B. J. Reeder, M. Clozel, R. Anand, C. E. Cooper, J. D. Morrow, M. T. Wilson, V. Darley-Usmar, and L. J. Roberts, II, unpub-lished observations.
TABLE I
Mb-induced formation of IsoPs in LDL and changes in relative electrophoretic mobility (REM)
Base line REM is 1.0 by definition. Formation of F2-IsoPs increased
significantly (p, 0.01) in the LDL following incubation with MetMb coincident with an increase in REM.
Preparation Base lineF
2-IsoPs 24-h Mb31 F2-IsoPs n-Fold Increase REM ng/mg LDL 1 0.9 49.8 55.3 4.8 2 0.4 13.4 30.4 2.9 3 0.9 38.0 42.2 3.8 4 0.6 12.7 21.9 4.0 5 0.3 30.7 102.3 3.8 Mean6 S.E. 0.66 0.1 296 7 506 14 3.86 0.3 TABLE II
Ferryl-Mb dependent oxidative modification of LDL
LDL (100mg/ml) was incubated with either MetMb (100 mM), fer-ryl-Mb (10mM), or H2O2(15mM) for 8 h at 37 °C in phosphate-buffered
saline containing 1 mMDTPA. Reactions were stopped by addition of butylated hydroxytoluene (100 mM) and F2-isoprostanes and REM
measured as described earlier. H2O2alone did not result in either
F2-isoprostane formation or an increase the REM, indicating that the
enhanced oxidative damage observed when H2O2was added to MetMb
is due formation of ferryl-Mb. For comparison with Table I it should be noted that the greater amounts of F2-isoprostanes generated in Table I
are due to a prolonged incubation period (24 h) compared with the data in Table II (8 h). Values represent means6 S.E. (n 5 3).
F2-isoprostanes REM ng/mg/LDL LDL (control) 0.726 0.05 1 MetMb1 LDL 8.56 2.1 1.536 0.07 Ferryl-Mb1 LDL 27.16 0.9 3.316 0.01 H2O21 LDL 0.76 0.03 1.066 0.04
established the occurrence of oxidant injury in the kidney in this animal model of rhabdomyolysis.
Following induction of rhabdomyolysis in the rat, deposition of Mb in the medulla and inner cortex of the kidney is apparent by visual inspection. The spectrum of an extract washed from a rhabdomyolytic kidney section is shown in Fig. 4A. The major feature is the Soret peak at 406 nm, typical of MetMb (36, 42). The visible region (inset) shows peaks at approximately 502 and 630 nm, also characteristic of MetMb. The peaks at 540 and 580 nm may be assigned to a small contribution from oxy-Mb (;10% of the total Mb) or to sulf-Mb, both of which have absorption bands at these wavelengths. Analysis following reduction and deoxygenation of the Mb extract with sodium dithionite is shown in Fig. 5B. Both a red shift of the Soret peak from 406 to 430 nm and the appearance of the peak at 550 nm in the visible region indicate the conversion of Mb to the deoxy ferrous form. However, the absorbance peak at 620 nm remains after the addition of dithionite. This absorbance band is consistent with the interac-tion of sulfide or sulfur-containing compounds, e.g. glutathione, with the porphyrin ring in ferryl-Mb forming sulf-Mb (37). An alternative explanation is that the band at 620 nm in the deoxy form is due, as reported by Catalano et al. (38), to cross-linking of the heme to the protein mediated by ferryl-Mb. Whether the presence of this band at 620 nm is due to sulf-Mb or to heme-protein cross-linking, it is highly likely that the Mb has passed through the ferryl state, since this intermediate is required for the formation of both species.
The ferric oxidation state of Mb (Fe31), but not the ferrous or ferryl forms, is detectable by EPR spectroscopy. Thus quanti-tative changes in the MetMb concentration can be measured using this technique. Significant changes were seen in EPR spectra from the kidneys of rhabdomyolytic rats (Fig. 5B) com-pared with controls (Fig. 5A). These data have been quantified and the results reported in Table III. Specifically, there was a large increase in the g5 6 high spin ferric heme signal, corre-sponding to metmyoglobin, and a low spin heme signal at gx5
2.58; gy5 2.28 (Table III). At the same time a new rhombic iron
signal with the same g5 4.3 but unknown ligands appears in the spectrum from the rhabdomyolysic kidneys, suggestive of an increase in low molecular weight ferric pools in the rhab-domyolytic kidneys.
Decreases are seen in the other cellular EPR signals, e.g. cytochrome P450 and iron-sulfur enzymes, which may be ex-plained as a result of general cellular damage and associated edema in the kidney. The low spin heme signal evident in the spectra B and C has different characteristics to cytochrome P450 with a different shape and slightly different g2. Since the high spin form of MetMb is very much enhanced in spectra B and C, it is likely to be the source of this new signal. Alkalin-ization had no significant effect on any of the components of the EPR spectra in the rhabdomyolytic kidneys, including the size of the new iron signal at g5 4.3 (Fig. 5, C and B, Table II). In particular there was no change in the concentration of MetMb following alkalinization (Table III).
Effect of Alkalinization on LDL Oxidation—The hypothesis that the oxidative potential of Mb to induce lipid peroxidation is pH-dependent was explored using MetMb-induced oxidation of LDL in vitro. LDL (three preparations) were incubated with 5 mM MetMb at a pH range of 5.5– 8.0. The results obtained indicate that Mb-induced formation of F2-IsoPs is highly pH-dependent, being markedly suppressed at alkaline pH (Fig. 6). Effect of Alkalinization on the Stability of Ferryl-Mb and Consumption of Lipid Hydroperoxides—Oxidation of lipids by Mb is a lipid hydroperoxide-dependent process (17). To further probe the mechanism for the pH dependence of Mb-induced lipid oxidation, consumption of the lipid hydroperoxide, 13-HPODE, by Mb was assessed as a function of pH. Calculated rate constants for consumption of 13-HPODE by MetMb at different pH values is shown in Table IV. The pH profile shows that consumption of 13-HPODE is low at pH values above 7 but increased 35-fold with a reduction in pH from 8.0 to 5.0. This profile is consistent with the pH dependence for the formation of F2-IsoPs in LDL (Fig. 6). To determine whether alkaliniza-tion stablilized the decay of ferryl-Mb, ferryl-Mb was incubated as above and the decay rate of ferryl-Mb followed spectropho-tometrically as a function of pH. The decay rate of ferryl-Mb FIG. 2. Effect of alkalinization on
renal function and urinary F2-IsoP
excretion in rats with rhabdomyoly-sis. A, creatinine clearance in normal rats
and in rats with glycerol-induced rhab-domyolysis (Rhabdo) and the effect of al-kalinization (Rhabdo-Alk). ** denotes p, 0.02 for rhabdo versus normal (Norm); * denotes p , 0.02 for rhabdo-alk versus rhabdo; # denotes p5 0.048 for rhabdo-alk versus normal (n 5 5 or 6 in each group). B, urinary excretion of F2-IsoPs in
normal rats and in rats with rhabdomy-olysis and the effect of alkalinization. The ** denotes p , 0.01 for rhabdo versus normal; * denotes p, 0.01 for rhabdo-alk
versus rhabdo; # denotes p5 0.17 for
rh-abdo-alk versus normal.
FIG. 3. Levels of F2-IsoPs esterified in renal lipids of rats with
rhabdomyolysis (Rhabdo), rhabdomyolysis and alkalinization (Rhabdo-Alk), and controls. Both rhabdo and rhabdo-alk groups
were significantly greater than controls (p, 0.01). ** denotes p , 0.05 compared with rhabdo alone.
was markedly reduced by 450-fold at pH 8.0 compared with that at pH 5.0 (Table IV).
DISCUSSION
The mechanistic basis for the pathogenesis of the renal fail-ure of rhabdomyolysis is poorly understood. The results ob-tained in this study demonstrating striking increases in the urinary excretion of F2-IsoPs in rats with rhabdomyolysis pro-vide compelling epro-vidence that lipid peroxidation is a feature of rhabdomyolysis. A finding of elevated levels of F2-IsoPs in the urine is highly suggestive, albeit not definitive, evidence that
this resulted from oxidant injury to the kidney. However, we found that levels of F2-IsoPs esterified in renal lipids of rats with rhabdomyolysis were also markedly increased, which un-equivocally localizes the occurrence of oxidant injury in the kidney. Importantly, we also found that alkalinization of the urine, which protects against renal failure, significantly sup-presses the urinary excretion of F2-IsoPs, supporting a causa-tive link between oxidant injury and the renal dysfunction.
The molecular mechanisms by which Mb deposited in the kidney induces oxidant injury could occur as a result of the FIG. 5. The EPR spectra of the kidney from rats with rhabdomyolysis and controls. The spectra represent the averaged spectra obtained from four to six samples (i.e. four to six animals) shown as a function of the magnetic field strength in Gauss (G). Spectrum A, normal rat kidneys (n5 4); spectrum B, kidneys from rats with rhabdomyolysis (n 5 6); spectrum C, kidneys from rats with rhabdomyolysis and alkalinization (n 5 6). The EPR spectrum of normal rat kidney (A) comprises a number of signals: the high spin ferric heme signal at g5 6, the signal at g 5 4.3 from the non-heme ferric iron with rhombic symmetry (serum transferrin and low molecular weight non-protein-bound iron coordinated with HCO32),
the signal arising from the low spin ferric heme in cytochrome P450 (gx5 2.42 and gy5 2.25), the Mn21signal (a six-line signal centered at g5 2 of which only the first three components are visible in this spectrum), the free radical (FR) signal (g5 2.003), the Mo51signal (g5 1.97), and the signals from the mitochondrial iron-sulfur (IS) enzymes NADH dehydrogenase and succinate dehydrogenase (g5 1.94–1.86). Two new signals appear In the rhabdomyolysis kidneys (spectra B and C), which are not apparent in the control kidneys (spectrum A). First there is a low spin ferric heme iron with gx5 2.58; gy5 2.28. This is most likely to be a low spin form of metmyoglobin and has similar g values to the MetMb-sulfide complex described in Ref. 37. The second new signal becomes apparent following subtraction of the control signal from that in the rhabdomyolytic kidneys in the g5 4.3 region (spectrum B-A) and is due to a new high spin rhombic ferric iron component. The g values for the EPR signals are indicated near corresponding vertical bars.
FIG. 4. Optical spectra of an extract
of Mb from rhabdomyolytic kidney before (A) and after reduction and deoxygenation with sodium dithion-ite (B). The inset shows the expanded
region from 450 to 700 nm. Spectrum A contains peaks characteristic of MetMb,
i.e. 406, 502, and 630 nm. Spectrum B is
typical of deoxy ferrous Mb exhibiting peaks at 430 and 550 nm. A peak at 620 nm, indicative of sulf-Mb, is present in both spectrum A and spectrum B, but partially masked by the MetMb peak in
release of free iron from Mb or, as suggested in the current study, from a direct pro-oxidant effect of the heme protein itself. In order for Mb to catalyze lipid peroxidation, ferrous Mb must be oxidized to the ferric form, which is competent to induce lipid peroxidation by redox cycling with ferryl-Mb. This is a highly reactive form of Mb, which can potently induce lipid peroxidation (Tables I and II). In support of the latter hypoth-esis, we found that the redox form of Mb deposited in the kidney is MetMb.
Importantly, in addition to explaining the occurrence of renal tubular necrosis, oxidant injury may also provide a plausible explanation for the occurrence of renal vasoconstriction in this disorder, which contributes to the renal dysfunction. The find-ing that IsoPs are markedly overproduced in myoglobinuria is potential highly relevant in that both 15-F2t-IsoP and 15-E2t -IsoP are very potent renal vaoconstrictors (25, 26, 28). Further-more, 15-F2t-IsoP has been shown to induce secretion of endo-thelin-1 (39), which is also a potent renal vasoconstrictor. Perhaps relevant to the latter finding is the recent report that bosentan, a combined Endothelin A and B receptor antagonist,
partially protects against renal failure in rhabdomyolysis (40). Thus, although it remains to be proven, it is attractive to speculate that overproduction of IsoPs may be responsible for the intense renal vasoconstriction that occurs in myoglobinuria.
Our findings also provide novel insights into mechanism by which alkalinization of the urine protects against the renal failure of rhabdomyolysis. The accepted explanation is that alkalinization enhances the solubility of Mb, and thus its ex-cretion from the kidney (12). Clearly, this could explain the finding that alkalinization reduced urinary excretion of F2 -IsoPs. However, EPR analysis failed to detect a reduction in the amount of high spin heme (MetMb) deposited in the kidney following treatment with alkali. Thus, an alternative explana-tion was considered based on the hypothesis for an effect of alkaline pH to reduce the capacity of Mb to catalyze lipid peroxidation. This hypothesis was supported by the finding that Mb-induced formation of F2-IsoPs in LDL was increasingly suppressed with increasing pH over the range of 5.5– 8.0. Fur-ther studies undertaken in an attempt to elucidate the mech-anistic basis for the pH dependence of the capacity of Mb to induce lipid peroxidation revealed that the rate of the ferryl-Mb to ferric Mb transition is strongly pH-dependent, being very low at pH.7.0, indicating stabilization, i.e. reduced reactivity, of ferryl-Mb at alkaline pH. The reason for this pH dependence may reside in the fact that the protonated form of ferryl heme (Fe(IV)OH2) is the reactive species, which is formally equiva-lent to a hydroxyl radical coordinated to ferric heme. It is important to note that renal tubular fluid is relatively acidic and the pH of the surrounding interstitium decreases further during low blood flow states (41). Thus, MetMb deposited in the renal tubules is in a milieu that greatly promotes the ability of Mb to induce lipid peroxidation.
In summary, a number of important new insights regarding the mechanism of the pathogenesis of the renal failure of rhabdomyolysis have emerged from these studies. First, com-pelling evidence was obtained for the occurrence of oxidant injury to the kidney in rhabomyolysis. Second, the observation that IsoP production is increased may explain the occurrence of intense renal vasoconstriction in this disorder. Third, the iden-tification of MetMb as the principal redox form of Mb deposited in the kidney provides a mechanistic basis to explain how myoglobinuria can cause lipid peroxidation independent of free iron and conventional Fenton reactions. Fourth, data obtained suggest that the mechanism by which alkalinization protects against renal failure is not by increasing solubility of Mb as previously proposed, but is consistent with stabilization of the highly reactive ferryl-Mb. These findings may have wider im-plications to the general functioning of heme proteins during
TABLE IV
Effect of pH on the consumption of 13-HPODE by MetMb and decay of ferryl-Mb
Consumption of 13-HPODE was followed by monitoring the loss of the conjugated diene chromophore at 234 nm, which occurs as 13-HPODE is converted to an epoxy peroxyl radical. The pH dependence of the decay of ferryl-Mb was determined by mixing 10mMMetMb with 20 mMhydrogen peroxide to form ferryl-Mb and the decay of ferryl-Mb was followed by monitoring the increase in absorbance at 408 nm due to reformation of MetMb.
pH k
13-HPODE consumption Decay of ferryl-Mb
s213 1023 5.0 2336 29 3200 6.0 876 2 310 6.5 206 2 100 7.0 96 0.3 77 8.0 66 0.5 7 TABLE III
Concentration of the paramagnetic species detected at 77 K in normal rat kidneys, kidneys from rats with glycerol-induced rhabdomyolysis,
and kidneys from rats with rhabdomyolysis with alkalinization
No significant difference was seen in any EPR signal between the rhabdomyolysis kidney and the rhabdomyolysis1 alkalinization kid-ney. The maximum concentration limit of signals that were undetect-able in the control kidney was determined by comparing the estimated signal size at different concentrations with the noise in the EPR spec-trum. Concentrations were established from the second integrals of pure line shapes of the signals obtained under non-saturating condi-tions. These second integrals were compared with those of standards at known concentrations.
Species Controlkidney Rhabdomyolysiskidney
Rhabdomyolysis 1 alkalinization
kidney
mM
High spin heme 226 5 2156 46a 3316 67a
Low spin heme Undetectable
(,5) 606 11
a 716 8a
Rhombic Fe(III) Undetectable
(,1) 7.56 2.5
a 76 3a
HCO3-coordinated Fe(III) 206 1 446 12 426 5
Cytochrome P450 1056 1.5 746 4a
766 11a a
Value significantly different from the control kidney.
FIG. 6. Effect of pH on MetMb-induced formation of F2-IsoPs in
LDL. Three LDL preparations were incubated with 5mMMetMb for
24 h at the pH specified. Data are expressed as the n-fold increase over base line.
changes of intracellular pH and provide a rationale to further explore whether the administration of antioxidants in conjunc-tion with alkaline therapy might have an additive effect in preventing renal injury as suggested by others (8). This would impact significantly on our ability to preserve renal function in patients with rhabdomyolysis.
REFERENCES
1. Bywaters, E. G., and Beall, D. (1941) Br. Med. J. 1, 427– 432 2. Better, O. S., and Stein, J. H. (1990) N. Engl. J. Med. 322, 825– 829 3. Grossman, R. A., Hamilton, R. W., Morse, B. M., Penn, A. S., and Goldberg, M.
(1974) N. Engl. J. Med. 291, 807– 811
4. Ayer, G., Grandchamp, A., Wyler, T., and Truniger, B. (1971) Circ. Res. 29, 128 –135
5. Kurtz, T. W., Maletz, R. M., and Hsu, C. H. (1976) Circ. Res. 38, 30 –35 6. Venkatachalam, M. A., Rennke, H. G., and Sandstrom, D. J. (1976) Circ. Res.
38, 267–279
7. Shah, S. V., and Walker, P. D. (1988) Am. J. Physiol. 255, F438 –F443 8. Salahudeen, A. K., Wang, C., Bigler, S. A., Dai, Z., and Tachikawa, H. (1996)
Nephrol. Dial. Transplant. 11, 635– 642
9. Paller, M. S. (1988) Am. J. Physiol. 255, F539 –F544
10. Nath, K. A., Dvergsten, J., Correa-Rotter, R., Hostetter, T. H., Manivel, J. C., and Rosenberg, M. E. (1994) Lab. Invest. 71, 209 –218
11. Abul-Ezz, S. R., Walker, P. D., and Shah, S. V. (1991) Proc. Natl. Acad. Sci.
U. S. A. 88, 9833–9837
12. Zager, R. A. (1989) Lab. Invest. 60, 619 – 629
13. Zager, R. A., and Foerder, C. A. (1992) J. Clin. Invest. 89, 989 –995 14. Zager, R. A., and Burkhart, K. (1997) Kidney Int. 51, 728 –738 15. Zager, R. A. (1992) J. Clin. Invest. 90, 711–719
16. Patel, R. P., Diczfalusy, U., Dzeletovic, S., Wilson, M. T., and Darley-Usmar, V. M. (1996) J. Lipid Res. 37, 2361–2371
17. Hogg, N., Rice-Evans, C., Darley-Usmar, V., Wilson, M. T., Paganga, G., and Bourne, L. (1994) Arch. Biochem. Biophys. 314, 39 – 44
18. Patel, R. P., Svistunenko, D. A., Darley-Usmar, V. M., Symons, M. C., and Wilson, M. T. (1996) Free Radic. Res. 25, 117–123
19. Harel, S., and Kanner, J. (1988) Free Radic. Res. Commun. 5, 21–33 20. Grisham, M. B. (1985) J. Free Radic. Biol. Med. 1, 227–232
21. Turner, J. J., Rice-Evans, C. A., Davies, M. J., and Newman, E. S. (1991)
Biochem. J. 277, 833– 837
22. Cooper, C. E., Green, E. S., Rice-Evans, C. A., Davies, M. J., and
Wrigglesworth, J. M. (1994) Free Radic. Res. 20, 219 –227
23. Maree, A., Peer, G., Schwartz, D. Serban, I., Blum, M., Wollman, Y., Cabili, S., and Laina, A. (1994) Nephrol. Dial. Transplant. 9, Suppl. 4, 78 – 81 24. Dee, G., Rice-Evans, C., Obeyesekera, S., Meraji, S. Jacobs, M., and Bruckdorfer,
K. R. (1991) FEBS Lett. 294, 38 – 42
25. Morrow, J. D., Hill, K. E., Burk, R. F., Nammour, T. M., Badr, K. F., and Roberts, L. J., II (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 9383–9387 26. Morrow, J. D., Minton, T. A., Mukundan, C. R., Campbell, M. D., Zackert,
W. E., Daniel, V. C., Badr, K. F., Blair, I. A., and Roberts, L. J., II (1994)
J. Biol. Chem. 269, 4317– 4326
27. Taber, D. F., Morrow, J. D., and Roberts, L. J., II (1997) Prostaglandins 53, 63– 67
28. Takahashi, K., Nammour, T. M., Fukunaga, M., Ebert, J., Morrow, J. D., Roberts, L. J., II, Hoover, R. L., and Badr, K. F. (1992) J. Clin. Invest. 90, 136 –141
29. Fukunaga, M., Makita, N., Roberts, L. J., II, Morrow, J. D., Takahashi, K., and Badr, K. F. (1993) Am. J. Physiol. 264, C1619 –C1624
30. Rodriguez-Malaver, A. J., Leake, D. S., and Rice-Evans, C. A. (1997) FEBS
Lett. 406, 37– 41
31. Roberts, L. J., II, and Morrow, J. D. (1997) Biochim. Biophys. Acta 1453, 121–135
32. Morrow, J. D., Awad, J. A., Boss, H. J., Blair, I. A., and Roberts, L. J., II (1992)
Proc. Natl. Acad. Sci. U. S. A. 89, 10721–10725
33. Moore, K. P., Darley-Usmar, V., Morrow, J., and Roberts, L. J., II (1995) Circ.
Res. 77, 335–341
34. Morrow, J. D., and Roberts, L. J., II. (1994) Methods Enzymol. 233, 163–174 35. Cammack, R., and Cooper, C. E. (1992) Methods Enzymol. 227, 353–384 36. Maiorino, M., Ursini, F., and Cadenas, E. (1994) Free Radic. Biol. Med. 16,
661– 667
37. Berzofsky, J. A., Peisach, J., and Blumberg, W. E. (1971) J. Biol. Chem. 246, 3367–3377
38. Catalano, C. E., Choe, Y. S., and Ortiz de Mortellano, P. R. (1989) J. Biol. Chem
264, 10534 –10541
39. Fukunaga, M., Yura, T., and Badr, K. F. (1995) J. Cardiovasc. Pharmacol. 26, Suppl 3, S51–S52
40. Karam, H., Bruneval, P., Clozel, J. P., Loffler, B. M., Bariety, J., and Clozel, M. (1995) J. Pharmacol. Exp. Ther. 274, 481– 486
41. DuBose, T. D., Jr., Pucacco, L. R., Lucci, M. S., and Carter, N. W. (1979)
J. Clin. Invest. 64, 476 – 482
42. Antonini, E., and Brunori, M. (1997) in Frontiers in Biology (Neuberger, A., and Tatum, E. L., eds) pp. 13–52, North-Holland, Amsterdam
