Ouabain-like Factor Quantification in Mammalian Tissues and Plasma
Comparison of Two Independent Assays
Abstract The resolution of controversies that concern the detectability of an endogenous ouabain-like factor (OLF) in mammalian tissues and plasma was approached by the application of a standardized method for its extraction and quantification. Two independent assays were used to quantify the OLF: (1) a radioimmunoassay, which used a polyclonal antiouabain antiserum, and (2) a radioenzymatic assay based on the inhibition of dog kidney Na+,K+-ATPase. Plasma and tissues were obtained from the Milan hypertensive strain (MHS) and the Milan normotensive strain (MNS) of rats and from healthy human volunteers. Results indicate that (1) a single high-performance liquid chromatography (HPLC) fraction identical to that of ouabain was identified by both assay methods in the rat hypothalamus and hypophysis and in both rat and human plasma; (2) dilution curves of OLF and standard ouabain were parallel and with a similar Kd, both in radioimmunoassay (3 nmol/L) and ATPase assay (14 nmol/L); (3) after HPLC, OLF was similarly quantified by the two methods in the hypothalamus, hypophysis, adrenals, and plasma of rats and in human plasma; (4) OLF was present in larger amounts in the hypothalamus, hypophysis, and plasma of MHS rats than that of MNS rats; (5) the HPLC fraction of human plasma was quantified similarly by both assays (range, 60 to 150 pmol/L); (6) recovery of standard ouabain in pre-HPLC plasma extracts was approximately 90%; and (7) pre-HPLC OLF concentrations in human plasma ranged between 0.05 and 0.75 nmol/L. Rat cerebral tissues and both rat and human plasma contained measurable amounts of OLF, which were quantified similarly by radioimmunoassay and ATPase assay, both before and after HPLC fractionation. The increased MHS tissue and plasma levels of OLF are in keeping with the pathogenetic role of this factor in MHS hypertension.
It has been suggested that the Na+-K+ pump is inhibited by OLF, which is present in mammalian circulation and tissues, and that this inhibition is involved in many physiological and pathological conditions related to the hydro-saline homeostasis of the body and particularly to the regulation of arterial pressure.1 2 3 4 5 Quite recently, it has been demonstrated that such an endogenous pump inhibitor is structurally similar to the cardiac glycoside ouabain,6 of which it is possibly a stereoisomer.7 8 These results have created a number of controversies, which concern both the detectability of this factor in human and animal plasma9 10 and the implications of its elevated levels in pathological states, such as essential or genetic hypertension.11 12 13 The major point of contention concerns the identification of OLF with plant ouabain and the variety of methodological approaches adopted for its extraction and quantification. Both the ELISA and an RIA, which use polyclonal antibodies raised against ouabain, have been developed recently to measure OLF levels.14 15 However, differences in the degree of specificity of the antiouabain antibodies, used by different laboratories, may be an additional source of inconsistency.
To assess the role of OLF in physiological and pathological conditions, a reliable method must be developed for its measurement, both in tissues and plasma. To avoid problems caused by the heterogeneity of human plasma samples and because of the possibility of contamination by exogenous ouabain, we used tissues and plasma from an animal model. The MHS and MNS rats are particularly suitable for these studies because they are reared under strictly controlled environmental and dietary conditions.16 Also, previous studies have demonstrated that the hypothalamic OLF content is 5 to 10 times greater in MHS than in MNS17 and that these levels are not influenced by exogenous sources.18 By comparing two independent assays in a single study—RIA, which uses a specific polyclonal antiouabain antiserum14 and a radioenzymatic assay based on inhibition of purified dog kidney Na+,K+-ATPase17 —we demonstrated that the two methodologies produce comparable results when applied to extracts obtained either before or after HPLC fractionation. Both assays were conducted on tissue and plasma extracts from adult MHS and MNS rats. Analyses of OLF from plasma of both Milan rats and healthy human volunteers were also performed to define the optimal experimental conditions that would provide reliable data about OLF plasma levels in humans.
The results demonstrate that, if appropriate and controlled extraction procedures are applied, OLF is always detectable in rat tissues and in both rat and human plasma, even after HPLC fractionation. The RIA and radioenzymatic assay quantified OLF similarly, both before and after HPLC fractionation.
Three-month-old male spontaneously hypertensive rats of the Milan strain (MHS/Gib) and their age-matched normotensive controls (MNS/Gib) were obtained from our original stock colony (Prassis-SigmaTau, Settimo M.se, Milan, Italy) and were maintained on a standard diet containing 2.5 g/kg NaCl (Altromin MT, Rieper) to the time of the experiments.
Indirect systolic blood pressure was measured by tail-cuff plethysmography on a W+W BP recorder (Basile) and averaged 170±1.5 and 132±1 mm Hg for MHS and MNS, respectively. The rats were decapitated with a hand guillotine.
All animals were maintained and treated in accordance with the European Community Directive 8/609/CEE of 24-11-1986 and the Italian Law DL n.116 of 27-1-1992. Strain and substrain symbols given in the text are in agreement with the Standardized Nomenclature of Inbred Laboratory Animals (Festing & Staats).
Tissues and Plasma
A total of 150 rats of both strains were killed; the following tissues were removed, pooled, and weighed: hypothalamus (22 to 25 g), hypophysis (1 to 1.5 g), and adrenals (4 to 5 g) for each strain. Tissues were stored at −20°C until use.
Arterial blood obtained from either decapitated or chronically cannulated MHS and MNS rats was collected in chilled, heparinized tubes and immediately centrifuged at 1500g for 10 minutes at 4°C. Plasma was separated and stored at −20°C until use. Venous blood samples from healthy normotensive human volunteers were treated in the same manner.
OLF Extraction and HPLC Fractionation Procedures
OLF was extracted from freshly thawed tissues following a previously described procedure.17 Tissues were homogenized with methanol (tissue/methanol, 1:10 [wt/vol]) that contained 2 mmol/L ascorbic acid and then stirred overnight at 4°C. The homogenate was centrifuged at 1500g for 30 minutes to pellet all the insoluble material. The supernatants were dried under vacuum, reconstituted in 0.1% distilled TFA acid (tissue/TFA, 1:5 [wt/vol]), and centrifuged at 3000g for 30 minutes at 4°C to discard the precipitated material. Different amounts of plasma were extracted for further purification and quantification as follows: 2 or 10 mL for pre-HPLC measurements and 40 mL for HPLC fractionation. Plasma samples were extracted either with methanol (plasma/methanol, 1:5 [vol/vol]), as in the procedure described for rat tissues, or directly mixed with 0.1% TFA (plasma/TFA,1:1 [vol/vol]) and centrifuged at 3000g for 30 minutes at 4°C. Supernatants from tissue and plasma extracts were passed by vacuum over prewashed C18 Mega Bond Elut columns (Varian). Several washes with water and one wash with 2.5% acetonitrile were performed to separate unbound material from compounds with less polarity. OLF was then eluted by the column with 25% acetonitrile, and the eluate was dried under vacuum. For pre-HPLC determinations, the dry residues from 2- or 10-mL plasma extracts were reconstituted with 200 μL bidistilled water and measured immediately by RIA or ATPase assay. For HPLC fractionation, plasma (40 mL), dry tissue extracts, or standard ouabain was dissolved in 500 μL bidistilled water and further chromatographed by injection onto a semipreparative reversed-phase PEP S C2/C18 high-performance column (Pharmacia, 39×250 mm, 4-μm particle size). A two-step linear acidified (0.05% distilled TFA) acetonitrile-to-water gradient was developed in 75 minutes at a 1-mL/min flow rate in accordance with the following gradient conditions: equilibration over 5 minutes in water, from 0% to 10% acetonitrile over 5 minutes; from 10% to 20% acetonitrile over 40 minutes; from 20% to 90% acetonitrile over 15 minutes; and from 90% to 0% acetonitrile over 10 minutes. One-minute fractions were collected, dried under vacuum, and assayed for the presence of OLF. HPLC was run in a Hitachi-Merck system controller (L6200A), and absorbance at 214 nm was recorded in parallel (L4000 UV detector). Blank runs were carried out with water and controlled immediately before sample injection.
Pre-HPLC Extraction and Exogenous Ouabain Recovery
Buffer and MNS rat plasma were spiked with concentrations of ouabain that increased from 125 to 500 pmol/L. These concentrations were chosen to span the physiological range of OLF concentrations, which is between 0.05 and 0.75 nmol/L (see “Results”). Buffer and plasma samples then were extracted with methanol and passed through a C18 cartridge as described previously. The basal plasma OLF concentration and the exogenous ouabain recovery were quantified by RIA.
OLF Quantification Methods
OLF or standard ouabain was measured by RIA as previously described.15 Tissue sample (25 μL) or commercial ouabain (Sigma) was incubated at room temperature for 15 hours with 25 μL [3H]ouabain (30 to 40 Ci/mmol, Amersham; 2 nmol/L final concentration, diluted in PBS containing 50 mmol/L sodium phosphate and 1% bovine albumin; pH 7.4) and 100 μL of a rabbit polyclonal antiouabain antiserum (R7) (1:20 000 final dilution in PBS), kindly provided by J.M. Hamlyn.14 In some experiments, the R7 antiserum was compared with another rabbit polyclonal antiouabain antiserum (SG), which was recently characterized19 and kindly provided by Dr S. Ghione (CNR Inst Clinical Physiology, Pisa, Italy). A goat anti-rabbit γ-globulin solution (1:800 final dilution in PBS) was added to be complexed with ouabain and ouabain antiserum. The incubation was stopped by the addition of 4 mL stop solution (5 mmol/L EDTA, 50 mmol/L sodium phosphate, pH 7.4) at room temperature; the separation of bound from free-labeled ouabain was achieved by the rapid filtration technique on Whatman GF/F glass fiber filters applied to a Brandel Cell Harvester apparatus (Biomedical Research and Development Laboratories). The filters were washed three times with a 4-mL stop solution and counted for radioactivity in a β counter (Beckman, LS 5000 CE) after the addition of 4 mL scintillation liquid (Filter Count, Beckman).
The R7 antiserum was characterized previously for its sensitivity and specificity.14 The immunocross-reactivity of R7 antiserum was 100% ouabain, 21% ouabagenin, 22% strophantidin-K, 9% digitoxin, 0.4% digoxin, 0.36% bufalin, <0.01% cinobufotalin and cinobufagin, 0.9% aldosterone, 0.05% β-estradiol, 0.012% progesterone, and not detectable for fatty acids. Furthermore, it was demonstrated that the threshold sensitivity of the R7 antiserum was approximately 0.02 nmol/L, and the estimated 50% displacement dose occurs at 2 to 3 nmol/L in typical ouabain dilution curves.14 The immunocross-reactivity of the SG antiserum was 100% ouabain, 34% ouabagenin, 32% strophantidin-K, 1.2% digitoxin, 1.2% digoxin, 2.7% bufalin, <0.01% cinobufotalin and cinobufagin, and not detectable for aldosterone, β-estradiol, progesterone, and fatty acids.
OLF levels in rat tissues and plasma extracts were calculated as percent displacement of the control sample, which was carried out in the absence of ouabain and OLF. The OLF concentration was expressed in nmol/L ouabain equivalents derived from the ouabain standard curve. With a molecular weight of 500 Da7 20 assumed for OLF, the tissue OLF content was calculated as nanograms per gram of wet weight. The intra- and interassay coefficient of variation was approximately 7% to 10%.
Radioenzymatic Assay (Na+,K+-ATPase Inhibition)
Renal Na+,K+-ATPase, from dog kidney outer medulla, was used to measure the ouabain and OLF inhibitory activity. Na+,K+-ATPase was purified as described by the Jorgensen procedure,21 and the protein concentration was estimated by the Lowry method, using bovine albumin as the standard.22 Na+,K+-ATPase activity was assayed by the 32Pi-ATP hydrolysis method as previously described.16
Aliquots of rat tissues and plasma extracts or standard ouabain and OLF, purified by HPLC from rat hypothalamus, were dried under vacuum and resuspended in 50-μL final volume of preincubation medium (140 mmol/L NaCl, 3 mmol/L MgCl2, 50 mmol/L Hepes-Tris, 3 mmol/L ATP, pH 7.5) that contained 0.3 μg of purified Na+,K+-ATPase and preincubated for 45 minutes at 37°C. Incubation solution (10 μL; 10 mmol/L KCl, 20 nCi 32Pi-ATP, Amersham, 0.5 to 3 Ci/mmol) was then added, and the reaction was allowed to continue for 15 minutes at 37°C. The reaction was stopped, and the free 32Pi was separated from the bound 32Pi by the addition of 100 μL cold perchloric acid (30%, vol/vol) and 500 μL activated charcoal. Free 32Pi was separated by centrifugation, and the radioactivity in the supernatant was measured in a liquid scintillation counter (Beckman LS 5000 CE) after the addition of 4 mL scintillation fluid (Instagel, Beckman). Under standardized conditions, the purified Na+,K+-ATPase activity from dog kidney was 99% ouabain inhibitable, and the specific activity ranged from 15 to 18 μmol · min−1 · mg−1 protein. OLF inhibitory activity was calculated as percent inhibition of the control sample and carried out in the absence of ouabain. OLF was expressed in units; one unit of OLF was defined as the amount of the sample that inhibited the Na+,K+-ATPase activity by 50%. From previously published data,16 the apparent molar concentration for 1 unit of OLF was estimated to be equivalent to 14 nmol/L. With a molecular weight of 50020 assumed for OLF, the tissue OLF content was expressed as nanograms per gram of wet weight.
Standard Ouabain and OLF Dilution Curves: Comparison Between RIA and Na+,K+-ATPase (ATPase) Assay
Fig 1⇓ shows the titration curves for ouabain and purified hypothalamic MHS OLF16 that was measured either by RIA (Fig 1A⇓) or the ATPase inhibition assay (Fig 1B⇓). The curves were fitted by a nonlinear regression program and showed a sigmoidal relation for ouabain, as well as for OLF, when determined by both assays. The calculated dose of ouabain that could produce a 50% displacement from the antiserum (Kd) was 3.2±0.2 nmol/L in the RIA (Fig 1A⇓), and the half- maximal inhibitory concentration (IC50) was 14.4±0.13 nmol/L in the ATPase assay (Fig 1B⇓). Hill coefficients for ouabain dose-response curves ranged from 0.95 (Fig 1A⇓) to 1.05 (Fig 1B⇓) in both assays, which indicate the presence of a single class of displacing (RIA) and binding (ATPase) sites. OLF dose-response curves showed a sigmoidal relation with a Hill coefficient of 1 in both assays and a Kd of 3.4±0.3 nmol/L (Fig 1A⇓) or IC50 of 14±0.3 nmol/L (Fig 1B⇓), which were identical to that of standard ouabain. With an OLF molecular weight of 500 Da20 assumed, the content of OLF in the sample of rat hypothalamic extract was similarly quantified by the two methods (RIA, 1.89 ng/g wet wt; ATPase, 1.66 ng/g wet wt).
Standard Ouabain and OLF Dilution Curves: Comparison of Two Different Antiouabain Polyclonal Antisera in the RIA Assay
Many of the discrepancies that concern OLF detectability and its concentration in plasma samples have been attributed to differences in the specificity of the antiouabain antisera used for RIA or ELISA. We compared two rabbit polyclonal antisera, R7 and SG, raised in two different laboratories, against plant ouabain, and already characterized.14 19 Fig 2⇓ shows the titration curves of the two antisera compared with ouabain and OLF purified from rat hypothalamus.17 Both antisera recognized with similar affinities both ouabain (Kd, 3.3±0.4 nmol/L) and OLF (Kd, 2.6±0.23 nmol/L) and similarly quantified an HPLC sample of human plasma (R7, 77 pmol/L; SG, 68 pmol/L).
HPLC Analysis of Standard Ouabain and Rat Tissue Extracts
To detect the HPLC elution profile of standard ouabain, 100 nmol of ouabain (500 μL) was injected into the C2/C18 HPLC column, and a gradient was developed as described in “Methods.” One-minute fractions were collected and assayed in parallel in RIA and ATPase tests. Blank injections were run before each HPLC sample injection and assayed in both tests. As shown in Fig 3A⇓, the HPLC elution profiles of standard ouabain obtained were the same in the two assays, both of which identified a single fraction containing ouabain with a retention time (t) of 21 minutes, corresponding to 15% acetonitrile. Also, in both assays the ouabain recovery determined in the HPLC active fraction accounted for 95% of the total injected ouabain. The parallel ultraviolet absorbance, recorded at 214 nm, identified a single fraction with t=21′ (not shown). As with “cold” ouabain, a similar elution profile (active fraction at t=21′) and recovery (90% to 95% of the total radioactivity injected) were also obtained when [3H]ouabain traces (Amersham, 35 Ci/mmol, 10 pmoles) were run on the same HPLC system (data not shown, 16).
Rat tissue extracts from C18 Bond Elut columns were run on the semipreparative reversed-phase C2/C18 column under elution conditions similar to those described for standard ouabain. The HPLC fractions were assayed in parallel in RIA and ATPase assay. Blanks were run immediately before each sample injection and were assayed with both tests.
Fig 3⇑ shows the HPLC elution profiles of rat tissue extracts from MHS and MNS hypothalamus (Fig 3B⇑), hypophysis (Fig 3C⇑), and adrenals (Fig 3D⇑). The figure shows that both RIA and ATPase identified the same active fraction, with a retention time identical to that of ouabain (t=21′), in all MHS and MNS tissue extracts. The HPLC elution profile of MHS and MNS adrenal extracts (Fig 3D⇑) showed, in addition to a fraction at t=21′, the presence of fractions that have minor polarity and retention times of approximately 35 to 45 minutes, as assayed by RIA. These fractions were not recognized by the ATPase assay. This result confirms similar previously reported findings,4 23 which concern the presence of less-polar fractions in adrenal extracts that contain some nonspecific material that is able to cross-react with the ouabain antibody but not to bind and inhibit the Na+,K+-ATPase. The nature of this material has not been further investigated.
Table 1⇓ summarizes OLF yields in MHS and MNS tissue extracts quantified by RIA and ATPase assay after HPLC fractionation (fraction at t=21′). The results indicate that rat tissues contained variable amounts of OLF, which were quantified similarly by the two methods. In particular, MHS showed in both hypothalamus and hypophysis 8- to 10- and 3- to 5-fold higher OLF yields, respectively, than MNS. On the contrary, no difference between MHS and MNS was detected in adrenal OLF content by either method.
HPLC Analysis of Rat and Human Plasma
Methanol extracts of plasma from MHS and MNS rats and three normotensive volunteers were run on HPLC with gradient conditions similar to those described for rat tissue extracts. Fig 4⇓ shows one representative HPLC plasma elution profile from MHS and MNS rats (Fig 4A⇓) and one from humans (Fig 4B⇓). These profiles were identical, both with the RIA and ATPase assay, showing a specific peak of elution at t=21′, which is similar to that observed for standard ouabain (Fig 3A⇑) or rat tissue extracts (Fig 3B⇑, C, D). No other active fractions were detected in either rat or human plasma with either assay method.
A second extraction procedure that consisted of a 0.1% TFA pretreatment followed by HPLC was investigated in which rat and human plasma were used (Fig 4C⇑ and 4D⇑). Under these experimental conditions, HPLC elution profiles were similar to those obtained with methanol pre-extraction (Fig 4A⇑ and 4B⇑). As before, both RIA and ATPase assay identified only one specific active fraction at t=21′ in all the plasma extracts. Therefore, the present results show that OLF is eluted in a single, specific, active fraction, both in rat and human plasma.
Plasma OLF concentrations in HPLC-active fractions were analyzed as a function of the pre-HPLC extraction procedures. As shown in Table 2⇓, the RIA and ATPase assay both quantified similar OLF concentrations in either rat or human plasma extracts. However, methanol extraction of MHS, MNS, and human plasma resulted in higher OLF concentrations than those found by the 0.1% TFA treatment (Table 2⇓). This result could be attributed to the higher density of the sample treated with TFA compared with that extracted with methanol. Therefore, a partial obstruction of the HPLC column may affect OLF recovery after HPLC. Notwithstanding this effect of TFA, the concordance between the RIA and ATPase assay for OLF quantification was verified by a linear regression analysis that showed a highly significant correlation coefficient between the two assays, both for OLF levels from rat tissues (Fig 5A⇓; r=.89, n=10, P<.001) and from rat and human plasma (Fig 5B⇓; r=.98, n=9, P<.001). Independent of the pre-HPLC extraction procedures, both with RIA and ATPase assay, plasma OLF concentration in MHS was approximately 2.8- to 3.8-fold higher than in MNS (Table 2⇓).
OLF Concentrations in Rat Tissues and Plasma Before HPLC
Table 3⇓ summarizes OLF yield from rat tissues measured by RIA in pre-HPLC extracts. The table shows that the amounts of OLF quantified before HPLC from hypothalamus and hypophysis in MHS were, respectively, approximately 7.5- and 3.5-fold higher than in MNS, which confirms the results shown previously on HPLC extracts (Table 1⇑). No differences were observed between the two strains for pre-HPLC adrenal extracts, as previously shown after HPLC (Table 1⇑). Moreover, it is very likely that the quantification of OLF in pre-HPLC adrenal extracts was overestimated because of the presence of nonspecific immunocross-reacting material demonstrated by the adrenal HPLC profile (Fig 3D⇑).
The recovery of exogenous ouabain, added either to 2 mL buffer or rat plasma and measured by RIA after the methanol pre-HPLC extraction procedure, is shown in Table 4⇓. It can be seen that ouabain added to the buffer samples was recovered on the average by 124% and that added to plasma by 87%.
MHS and MNS plasma-OLF concentrations were then quantified by RIA in pre-HPLC extracts to verify whether the difference observed in HPLC extracts between the two strains was maintained in this condition and whether it was influenced by the blood-sampling maneuver.
Pre-HPLC plasma-OLF concentrations, measured in trunk blood after methanol extraction, were 2-fold higher in MHS (0.565±0.06 nmol/L, n=17) than in MNS (0.284±0.05 nmol/L, n=15, P<.01), and similar results were obtained in samples withdrawn from chronically cannulated rats (MHS, 0.535±0.1 nmol/L; n=7; MNS, 0.225±0.05 nmol/L; n=6; P<.05). Therefore, similar differences in plasma-OLF concentrations were detectable between the two rat strains, both before and after HPLC fractionation (Table 2⇑), and are not affected by the sampling maneuver.
To verify if a sample pretreatment, different from methanol extraction, may influence the plasma-OLF quantification before HPLC, the same rat plasma sample was either extracted with methanol or pretreated with 0.1% TFA and assayed by RIA. The results showed that plasma-OLF was similarly quantified after either methanol (MHS, 0.338 nmol/L; MNS, 0.197 nmol/L) or TFA (MHS, 0.320 nmol/L; MNS, 0.198 nmol/L) pretreatment. Therefore, when plasma-OLF concentrations are assayed in pre-HPLC extracts by RIA, either methanol or TFA pretreatment can be used indifferently.
OLF Quantification in Human Plasma Pre-HPLC Extracts
To allow the quantification of OLF concentrations in plasma samples, with HPLC fractionation avoided, two conditions must be met: only one active fraction may be present in plasma extracts after HPLC fractionation, and thus the possibility of quantifying nonspecific cross-reacting material in pre-HPLC samples is excluded; quantitative differences between plasma samples (ie, between MHS and MNS rats) are maintained both before and after HPLC fractionation. In the current study, both of these conditions were met; therefore, we quantified OLF concentrations in pre-HPLC human plasma by comparing the RIA and ATPase assay. After methanol extraction, similar OLF concentrations were detected, both with RIA (309±19 pmol/L, n=20) and ATPase assay (349±18 pmol/L, n=12). On the contrary, TFA treatment resulted in a difference in quantification by RIA (354±29 pmol/L, n=12) and ATPase assay (115±9.6 pmol/L, n=5). The lower OLF concentrations, observed when samples were treated with TFA and quantified by ATPase compared with RIA, are probably caused by the different amounts of plasma needed to perform these two assays: 2 mL for RIA and 10 mL for ATPase assay, which are extracted volume to volume, dried, and reconstituted in 200 μL. The final density of the sample treated with TFA and assayed by the ATPase method was higher than that assayed by RIA; this may therefore affect the ability of OLF to interact with the Na+,K+-ATPase receptor.
The presence of an endogenous Na+,K+-pump inhibitor in mammals, which can affect renal Na+ excretion, vascular contractility, and nervous reactivity, has been suggested for many years and has been substantiated by considerable experimental evidence.1 2 5 24 25 Recently, Mathews et al6 and Tymiak et al7 have demonstrated that such an inhibitor is related structurally to ouabain and could represent an isomer of this natural cardiac glycoside. Moreover, it has also been suggested that other endogenous Na+,K+-pump inhibitors, structurally different from cardenolides, are present in mammalian plasma and may include bufadienolide derivatives, such as the marinobufagenin-like factor.26 Several studies support the existence and detectability of OLF, both in experimental animal models17 24 25 27 28 and healthy humans or patients affected by different cardiovascular disorders.5 29 30 31 32 Recently, however, considerable speculation and discordant data that concern the different identities of ouabain and OLF, its detectability, and its implications in physiological and pathological human disorders have been published.9 10 11 12 13
To resolve the problem of detectability and reliability of OLF measurements, we have measured this factor both in tissues and plasma of MHS and MNS rats, maintained under strictly controlled environmental and dietary conditions, by applying both ouabain RIA and Na+,K+-ATPase inhibition assay. By these methods, OLF has also been detected in human plasma samples, provided that, with appropriate preextraction procedures followed by HPLC, only one active fraction, which coelutes with plant ouabain, contains OLF.
First, we have verified that the two assays, RIA and ATPase, were able to quantify plant ouabain in the same way. The two dose-response curves obtained with RIA and ATPase assay are parallel and superimposable (Fig 1⇑), with a difference shown only in the calculated Kds, which was caused by the expected different affinities of the antiserum for ouabain (Kd, 3.4 nmol/L) and Na+,K+-ATPase (IC50, 14 nmol/L). When a sample of purified OLF from rat hypothalamus17 is processed in parallel with ouabain, the two substances are indistinguishable and show identical dose-response curves in both assays. These results show that ouabain and purified OLF are recognized and equally quantified by two tests, each of which is based on different biological criteria. Because OLF can inhibit Na+,K+-ATPase and cross-react with antiouabain antisera, it behaves as plant ouabain, even though it may be structurally different. The lower detection limit of RIA with the R7 antiserum is 0.07 nmol/L and that of the ATPase assay is 3 nmol/L. This suggests that to obtain reliable measurements of OLF levels with the ATPase assay, at least 5 to 10 times more tissue or plasma must be processed compared with the RIA.
The heterogeneity of the antisera used in different laboratories may be responsible for the variability of results reported for OLF quantification. In our work, we compared two rabbit polyclonal antiouabain antisera (raised by different laboratories [J. Hamlyn’s laboratory, R7; S. Ghione’s laboratory, SG]) that possessed very similar panels of cross reactivity with digitalis or steroidal molecules and showed that both of these antisera recognized (in addition to ouabain) two aglycones structurally related to ouabain, such as ouabagenin and strophantidin-K. With the use of these antisera in an RIA assay, we obtained completely superimposable ouabain and OLF titration curves (Fig 2⇑). Moreover, OLF concentration in a sample of human plasma, previously fractionated on HPLC, was quantified similarly by both antisera. This may indicate that antisera able to recognize ouabain and related aglycones are probably suitable for the recognition of the endogenous OLF, which is likely a close isomer of ouabain7 and whose isometry has been suggested recently to be more linked to one of the hydroxyl groups on the steroidal ring than to a different position of the sugar substitution.33 The possibility that the antisera used in the current study can recognize an endogenous inhibitor of Na+,K+-ATPase, structurally different from cardenolides and closer to bufodienolides,26 cannot be excluded completely, even though this seems unlikely because the two antisera show very weak cross reactivity with bufodienolides.
The extraction and HPLC-fractionation procedures and then the use of ELISA, instead of RIA, may also introduce other important factors of variability. In an effort to standardize these procedures, we analyzed each part of the entire procedure. With our experimental conditions (see “Methods”), approximately 100% of plant ouabain is recovered after HPLC fractionation. Only a single fraction that contains ouabain, which elutes at 15% acetonitrile with a retention time of 21 minutes, is equally recognized by the two assays (Fig 3A⇑). By this procedure, OLF from hypothalamus, hypophysis, and adrenals of MHS and MNS rats was quantified. These tissues were chosen because previous reports have demonstrated that they may be sites for OLF production15 17 20 34 and that MHS hypothalamus contains a higher amount of OLF than that of MNS.17 We have demonstrated that for all of these tissues a specific HPLC fraction, identical to that of ouabain, is similarly recognized as containing OLF by both RIA and ATPase methods (Fig 3B⇑ through 3D). Of more interest, however, is the observation that adrenal extracts, with the exception of the 21-minute active fraction, contain other material (with higher retention time) able to cross-react with the antiouabain antisera but unable to inhibit Na+,K+-ATPase. Considerable controversy has been raised about the origin of the endogenous ouabain in adrenals.23 35 Hamlyn’s group34 36 demonstrated that adrenals are the most highly enriched tissue source of endogenous ouabain. Doris et al,23 however, found a very diffused pattern of ouabain cross reactivity in conditioned medium from adrenal cells in culture, which suggests that even if adrenals contain and secrete ouabain, the levels are too low to play any physiological role in vivo.13 23 Because our data confirm that very low amounts of OLF and some nonspecific material, only detectable by RIA, is contained in adrenals, it would seem that very stringent methodological criteria, such as HPLC fractionation and double assay measurements, must be conducted to quantify OLF properly from this tissue. Moreover, by applying these criteria, no difference in adrenal OLF content was observed between MHS and MNS rats. On the contrary, hypothalamic and hypophyseal tissues are more enriched with OLF, whose levels are, respectively, nine and four times higher in MHS than in MNS. These data agree with those already published by our group17 and those obtained in other hypertensive rat strains, such as spontaneously hypertensive rats37 and Dahl rats.38
One of the most debated issues concerns OLF detectability in plasma. We have shown that, with both methanol extraction and TFA treatment, HPLC fractionation of human and rat plasma results in a single active fraction, which is indistinguishable from that revealed when standard ouabain is eluted, and that it contains material that is similarly detected by RIA and ATPase assay (Fig 4⇑). The amount of OLF quantified in this HPLC fraction depends on the preextraction procedure, being lower with TFA than with methanol (Table 2⇑). The reason for this difference may be linked to a “matrix” effect caused by the high density of the samples treated with TFA, which may cause an obstruction of the HPLC column. The possibility that a 0.1% TFA treatment can destroy OLF is unlikely because we showed that in pre-HPLC samples methanol extraction and TFA treatment give a similar quantification of OLF when measured by RIA. Moreover, we have already demonstrated that only extensive acid hydrolysis, with 6N HCl at 110°C for 2 hours, partially inactivates OLF, which is also not sensitive to 10% TCA treatment.17
The OLF concentrations we found in rat and human plasma after HPLC fractionation ranged from 25 to 100 pmol/L (Table 2⇑). These values are in agreement with those reported by other authors who used HPLC followed by RIA or ELISA.10 29 34 However, Lewis et al9 were unable to detect “endogenous ouabain” in HPLC-fractionated human plasma by ELISA that used an antiouabain antiserum. These authors observed only an immunoreactive fraction that contained a less-polar material than ouabain, which was not further characterized. There are several explanations for these discrepancies. First, Lewis’s antiouabain antiserum, as opposed to the two used in the current study, may not recognize OLF, even if the immunocross-reactivity profiles toward ouabain and related aglycones are very similar for all three. Second, different preextraction procedures may be involved. In the Lewis study, two different extraction methods were used: in the first, only 1 mL of plasma was passed through the C18 cartridge, eluted with 25% acetonitrile/0.1% TFA, dried, reconstituted, and applied to an HPLC column. The small plasma sample used may result in a low recovery of OLF, which is under the detection limit of the assay. In the second method, methanol was used as in our study,9 but contrary to our procedure, the methanol extract (from 20 mL of plasma) analyzed by Lewis et al was either allowed to pass through the C18 cartridge without the application of a vacuum or was simply precipitated, thus avoiding the cartridge passage. These two procedures may not extract sufficient ouabain-like material from the plasma matrix. Furthermore, we have shown that, provided the extraction procedures are well standardized, it is possible to detect OLF in both mammalian tissues and plasma samples by the use of the ATPase method alone, independent of the antibody used.
A practical result of this study was to standardize a method to quantify OLF in small samples of rat plasma without HPLC fractionation. This objective has been pursued after two criteria were verified as having been met: first, that under our experimental conditions, no other fractions except that which contained OLF were eluted in HPLC both from rat and human plasma; and second, that similar quantitative differences between MHS and MNS plasma samples are maintained both before and after HPCL fractionation. Plasma samples, spiked with exogenous ouabain at concentrations that span the physiological OLF range, showed after methanol extraction and C18 elution (Table 4⇑) a good recovery with a small degree of loss, probably caused by the binding of ouabain to plasma proteins. When plasma is extracted with methanol, OLF is similarly quantified by both RIA and ATPase assay in a range of concentrations between 0.05 and 0.7 nmol/L, both for rats and healthy humans. Therefore, the simple procedure described in this article, which utilizes small amounts of plasma (2 mL) and avoids the HPLC step, can be adopted for quantifying OLF in rats under different experimental, pharmacological, and physiopathological conditions. Preliminary data presented here on OLF quantification in pre-HPLC plasma from healthy human volunteers are in agreement with those obtained in rats. However, to validate this procedure for the measurement of OLF in the population at large, further studies should involve a sufficient number of patients suffering from different pathological conditions.
In conclusion, present findings demonstrate that an endogenous inhibitor of the Na+,K+ pump, recognized by antiouabain antisera, is detectable even after HPLC fractionation in mammalian tissues and plasma at picomolar concentrations. Two different assays, based on immune cross-reactivity (RIA) or enzyme inhibition (ATPase), quantify OLF similarly, provided that appropriate extraction procedures are applied. We suggest that all laboratories interested in OLF quantification should make uniform their experimental protocols and antisera and exchange samples to set up a unique standardized procedure, which could lead to the clarification of the functional role of OLF in various physiological and pathological conditions.
Selected Abbreviations and Acronyms
|ATPase||=||Na+,K+-ATPase inhibition assay|
|HPLC||=||high-performance liquid chromatography|
|MHS||=||Milan spontaneously hypertensive rat(s)|
|MNS||=||Milan normotensive rat(s)|
We thank Dr Sergio Ghione for the kind gift of the antiouabain antiserum and for fruitful discussions and collaboration.
- Received February 25, 1997.
- Revision received March 5, 1997.
- Accepted March 5, 1997.
Blaustein MP. Sodium ions, calcium ions, and blood pressure regulation: a reassessment and a hypothesis. Am J Physiol. 1977;232:C165-C173.
Hamlyn JM, Blaustein MP, Bova S, DuCharme DW, Harris DW, Mandel F, Mathews W, Ludens J. Identification and characterization of a ouabain-like compound from human plasma. Proc Natl Acad Sci U S A. 1991;88:6259-6263.
Tymiak AA, Norman JA, Bolgar M, Di Donato GC, Lee H, Parker WL, Lo LC, Berova N, Nakanishi K, Haber E, Haupert GT. Physicochemical characterization of a ouabain isomer isolated from bovine hypothalamus. Proc Natl Acad Sci U S A. 1993;90:8189-8193.
Lewis LK, Yandle TG, Lewis JG, Richards AM, Pidgeon GB, Kaaja RJ, Nicholls MG. Ouabain is not detectable in human plasma. Hypertension. 1994;24:549-555.
Kelly RA, Smith TW. Is ouabain the endogenous digitalis? Circulation. 1992;86:694-697.
Hansen O. Do putative endogenous digitalis-like factors have a physiological role? Hypertension. 1994;24:640-642.
Barber BR, Ferrari P, Bianchi G. The Milan hypertensive strain: a description of the model. In: Ganten D, de Jong W, eds. Handbook of Hypertension. Amsterdam, Netherlands: Elsevier; 1994:316-345.
Ferrandi M, Minotti E, Salardi S, Florio M, Bianchi G, Ferrari P. Ouabainlike factor in Milan hypertensive rats. Am J Physiol. 1992;263:F739-F748.
Haupert GT, Carilli CT, Cantley LC. Hypothalamic sodium transport inhibitor is a high affinity reversible inhibitor of Na-K ATPase. Am J Physiol. 1984;247:F919-F924.
Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem. 1951;193:265-275.
Doris PA, Jenkins LA, Stocco DM. Is ouabain an authentic endog- enous mammalian substance derived from the adrenal? Hypertension.. 1994;23:632-638.
Leenen FH, Harmse E, Yu H, Ou C. Effects of dietary sodium on central and peripheral ouabainlike activity in spontaneously hypertensive rats. Am J Physiol. 1993;264:H2051-H2055.
Doris PA. Ouabain in plasma from spontaneously hypertensive rats. Am J Physiol. 1994;266(Heart Circ Physiol 35):H360-H364.
Hout SJ, Pamnani MB, Clough DL, Buggy J, Bryant HJ, Harder DR, Haddy FJ. Sodium-potassium pump activity in reduced renal mass hypertension. Hypertension. 1983;5(part 2):I-94-I-100.
Goto A, Yamada K, Hazama H, Uehara Y, Atarashi K, Hirata Y, Kimura K, Omata M. Ouabainlike compound in hypertension associated with ectopic corticotropin syndrome. Hypertension. 1996;28:421-425.
Bagrov A, Fedorova O, Roukoyatkina N, Zhabko E. Effect of endogenous digoxin-like factor and digoxin antibody on myocardial Na-K pump activity and ventricular arrhythmias in acute myocardial ischaemia in rats. Cardiovasc Res. 1993;27:1045-1050.
Dong JG, Akritopoulou-Zanze I, Guo J, Berova N, Nakanishi K, Hauper GT. Cd theorical studies on ouabain pentanaphthoate analogs. In: Beaugé LA, Garrahan PJ, Gadsby DC, eds. Proceedings of the VIIIth International Conference on the Na/K-ATPase and Related Transport ATPases. Ann N Y Acad Sci. In press.
Hamilton BP, Manunta P, Laredo J, Hamilton JH, Hamlyn JM. The new adrenal steroid hormone ouabain. Curr Opin Endocrinol Diabetes. 1994:123-131.
Naruse K, Naruse M, Tanabe A, Yoshimoto T, Watanabe Y, Kurimoto F, Horiba N, Tamura M, Inagami T, Demura H. Does plasma immunoreactive ouabain originate from the adrenal gland? Hypertension. 1994;23(suppl 1):I-102-I-105.
Leenen FHH, Harmsen E, Yu H, Ou C. Effects of dietary sodium on central and peripheral ouabainlike activity in spontaneously hypertensive rats. Am J Physiol. 1993;264(Heart Circ Physiol 33):H2051-H2055.
Leenen FHH, Harmsen E, Yu H. Dietary sodium and central vs peripheral ouabain-like activity in Dahl salt-sensitive vs salt resistant rats. Am J Physiol. 1994;267:H1916-H1920.