The role of endothelin (ET) receptors was tested in volume-stimulated atrial natriuretic factor (ANF) secretion in conscious rats. Mean ANF responses to slow infusions (3 × 3.3 ml/8 min) were dose dependently reduced (P < 0.05) by bosentan (nonselective ET-receptor antagonist) from 64.1 ± 18.1 (SE) pg/ml (control) to 52.6 ± 16.1 (0.033 mg bosentan/rat), 16.1 ± 7.6 (0.33 mg/rat), and 11.6 ± 6.5 pg/ml (3.3 mg/rat). The ET-A-receptor antagonist BQ-123 (1 mg/rat) had no effect relative to DMSO controls, whereas the putative ET-B antagonist IRL-1038 (0.1 mg/rat) abolished the response. In a second protocol, BQ-123 (≥0.5 mg/rat) nonsignificantly reduced the peak ANF response (106.1 ± 23.0 pg/ml) to 74.0 ± 20.5 pg/ml for slow infusions (3.5 ml/8.5 min) but reduced the peak response (425.3 ± 58.1 pg/ml) for fast infusions (6.6 ml/1 min) by 49.9% (P < 0.001) and for 340 pmoles ET-1 (328.8 ± 69.5 pg/ml) by 83.5% (P < 0.0001). BQ-123 abolished the ET-1-induced increase in arterial pressure (21.8 ± 5.2 mmHg at 1 min). Changes in central venous pressure were similar for DMSO and BQ-123 (slow: 0.91 and 1.14 mmHg; fast: 4.50 and 4.13 mmHg). The results suggest 1) ET-B receptors mainly mediate the ANF secretion to slow volume expansions of <1.6%/min; and 2) ET-A receptors mainly mediate the ANF response to acute volume overloads.
- atrial natriuretic peptides
- blood volume
- endothelin receptor antagonists
- atrial natriuretic factor
atrial natriuretic factor (ANF), a potent diuretic, natriuretic, and vasorelaxant hormone (5), has pathophysiological importance in the regulation of pulmonary and systemic arterial blood pressure (20). Accordingly, both hypoxia and hypervolemia stimulate the secretion of ANF in animals and humans (discussed in Ref. 1). The mechanisms of ANF secretion are still incompletely understood, although cardiac sympathetic activation partly mediates the hypoxia-stimulated ANF secretion (1, 11-13), whereas atrial stretch is often invoked to explain the volume-stimulated ANF release. In isolated hearts or atria, atrial stretch causes only a short-lived ANF secretion lasting a few minutes (summarized in Ref. 11). Thus stretch of the atrial myocytes does not fully account for the sustained increase in plasma ANF concentrations commonly observed in hypervolemia. Cofactors may sustain the increase in stretch-stimulated ANF secretion. Some evidence points to ATP-sensitive potassium (KATP) channels (24) and angiotensin II receptors (15), and there is considerable data supporting a role for the endothelium-derived peptide endothelin.
Endothelin has potent ANF-releasing activity in all types of in vivo or in vitro preparations, and it is a prime candidate as a paracrine or endocrine regulator of stretch-induced ANF release (20, 21). Endothelin increases the ANF release in culture (9, 14, 22) and isolated atria (21), presumably via an A-type endothelin receptor (e.g., 9, 14, 21, 23). In vivo, anti-endothelin antiserum, the wide-spectrum endothelin receptor antagonist bosentan, and the A-selective endothelin receptor antagonist BQ-123 significantly inhibit volume-stimulated ANF release (6, 15). Despite these advances, there are three concerns: 1) the stimulus used in some of these studies is supraphysiological, corresponding to a volume expansion of 25% in 1 min; 2) the endothelin receptor antagonists have only partial effects; and3) the role of ET-B-receptors was not tested in vivo.
The first aim of this paper, therefore, was to reexamine the role of endothelin receptor subtypes in ANF regulation in vivo at low rates of infusion, corresponding to a volume expansion of <1.6% per minute. The second aim was to test the antagonist BQ-123 at low and high rates of infusion in the same animal to minimize the confounding effects of different preparations and experimental procedures. The results suggest a major role of ET-B receptors at low rates and of ET-A receptors at high rates of volume expansion.
Two series of experiments were performed on male Wistar rats (325 ± 25 g body wt) that were housed in clear plastic cages in a quiet room and separate from other rats (22°C, lights on from 7 AM to 8 PM). In series 1, a tail artery catheter was inserted in experimental and donor rats under methoxyflurane anesthesia as described previously (2). Arterial catheters were maintained patent by automatically injecting heparinized saline (0.4 ml, 150 U/ml) every 5 h over 1 min. In series 2, the experimental rats were anesthetized 1 day before the experiment using 2.5% halothane in oxygen. Segments of PE-10 tubing filled with phosphate-buffered saline (PBS) were inserted into the right femoral vein (for the fast infusions and endothelin injection), the right femoral artery (for arterial blood pressure measurements, blood sample collection, and slow infusions), and the jugular vein (for central venous blood pressure measurements). In the latter case, the catheter was pushed into the vena cava close to the right atrium. Catheters were then tunneled subcutaneously to exit at the back of the neck. Donor rats received only the femoral arterial catheter.
On the day of the experiments, 5–8 ml of donor blood were obtained from each donor rat, centrifuged, cleared of peptides by passage of the donor plasma through C18 reverse-phase cartridges (Waters, Milford, MA or Amersham Switzerland, Zurich), recombined with the donor red blood cells, stored on ice, and warmed to 37°C before use. This procedure was successfully used in several studies to obtain blood samples for stable hormone measurements, while minimizing sampling-induced changes in blood pressure (e.g., Ref. 2).
Groups of Rats
Experiments were performed on a total of 67 conscious rats split into 11 groups (7 tests, 4 controls), using another 67 donor rats as described above. Group sizes, tests, and controls are indicated in Table 1 (series 1) and Table2 (series 2).
In series 1, the wide-spectrum endothelin receptor antagonist bosentan (4) from Roche (Basel, Switzerland) was tested in 15 rats at three doses (0.033, 0.33, 3.3 mg/rat), using 10 vehicle (0.5 ml H2O)-treated controls. The specific ET-A receptor antagonists BQ-123 (1 mg/rat) and the putative ET-B receptor antagonist IRL-1038 (0.1 mg/rat) (both from American Peptide) were tested in 16 rats, using 12 vehicle-treated rats as controls (50 μl DMSO in 0.45 ml 0.9% saline, and 20 μl DMSO in 0.28 ml 0.9% saline, respectively). DMSO was used because both IRL-1038 and BQ-123 were sold in a water-insoluble form at the time the experiments were begun. Vehicle or antagonist was injected intra-arterially over 1 min betweensamples 1 and 2.
In series 2, BQ-123 (0.5 mg/rat, dissolved in 25 μl DMSO and 0.45 ml 0.9% saline) or vehicle was tested in 14 rats and administered intra-arterially over 1 min before each stimulus (see below).
Series 1. The rats were freely moving in their cage while the arterial catheter was allowed to turn in a swivel connector. Plasma ANF concentrations were measured before and during a slow blood volume expansion. Arterial blood samples (1.0 ml) were obtained over 2 min, the midpoint of the collection times being −8 min and −2 min (controls) and 10, 26, 38, and 52 min (tests). The blood lost was immediately replaced by infusing 1.0 ml of donor blood over 1 min. A slow blood volume expansion was produced by infusing via the tail artery catheter three times with 3.3 ml of saline-dextran (Abbott) at a rate of 0.41 ml/min (8 min each period), starting at times 0, 14, and 28 min. A slow rate and intra-arterial route of infusion were chosen to better approximate physiological conditions. Three 3.3-ml infusions rather than a single 9.9-ml infusion was adopted for taking arterial blood samples and arterial pressure measurements, because only a single catheter was suited for freely moving rats. Test substances and appropriate vehicles were applied in the same series of experiments to minimize the effect of group or week of testing. This procedure yields reproducible, robust ANF responses over a given series of experiments.
Series 2. Because the ANF responses subside quite quickly (see Fig. 1), a protocol with repeated stimuli could be adopted forseries 2. The blood sampling, slow infusion protocol, and drug injections were similar to the first part of the protocol of series 1. The fast infusion protocol was similar to the procedures described by Leskinen et al. (15). Although the volumes and routes of infusion were not equal (3.5 ml intra-arterial vs. 6.6 ml intravenous, see below), the advantage of this protocol was that the results could be compared with those of series 1 and of Leskinen et al. (15). The rats were placed in a towel-covered plastic cylinder to have access to the three catheters. A rest period of 30–60 min was allowed while physiological parameters were monitored (see below). The rats quickly adapted to their confined surroundings and were desensitized during this period to the sampling procedure by repeatedly connecting the sampling syringe to the catheter. The protocol started at time 0 and lasted 80 min. Either vehicle or BQ-123 was injected intra-arterially in each rat before each stimulus (at times 2 and 34 min, in 9 rats also at time 61 min). Repeated application was required, because it was noted that the efficacy of BQ-123 diminished over a period of 20 min. The slow infusion was applied (3.5 ml PBS in 8.5 min, once) via the femoral artery catheter starting at time 11 min. The fast infusion was applied (6.6 ml PBS over 1 min) via the intravenous catheter starting at time 40 min. Human endothelin-1 (340 pmol, American Peptide) was injected in 0.85 ml of PBS via the intravenous catheter at time 65 min. Drugs were all infused over 1 min to avoid an unwanted stimulation of the atria. Four blood samples were taken via the arterial catheter over 2 min with midpoints at times 7, 22, 43, and 68 min. Physiological parameters were measured (seeMeasurements) up to time 80 min.
The ANF extraction and radioimmunoassay are identical to methods published previously (13). Samples were extracted from 0.4 ml of plasma with C18 reverse-phase columns (see Donor Blood), eluted with 1–1.5 ml of an ethanol-water-acetic acid (90:6:4 vol/vol/vol) mixture, dried at 60°C with air or nitrogen, and reconstituted in 200–250 μl of assay buffer. Recovery of 20–64 pg unlabeled rat α-ANF-(1–28) (Peninsula) added to 0.4 ml of peptide-cleared plasma was 73.2 ± 5.7 (SE) % (n = 8 assay series). ANF values were corrected for recovery. The radioimmunoassays were performed exactly as described previously, using 125I-labeled rat α-ANF-(1–28) as a tracer, an NH2-terminal-sensitive anti-rat ANF antibody (no. IM1871, Amersham), and a goat anti-rabbit antiserum (Calbiochem) for precipitation. Sensitivity (5% displacement, 1.2 pg/tube; 50% displacement, 24 pg/tube) correlation coefficients (r > 0.95) of the log-logit regressions and coefficients of variation (intra-assay, <4%; interassay, <20%) were as described previously (13).
Justification for Concentrations of Test Substances
The doses of bosentan spanned a wide range (0.033–3.3 mg/rat), with a maximum exceeding the dose used in vivo to prevent postischemic renal vasoconstriction (4). The dose of BQ-123 (series 1: 1 mg/rat; series 2: 0.5 mg/rat, 2 or 3 times) was chosen to attain a concentration >5–10 μmol/l, assuming a distribution volume of 100 ml. It effectively blocked the volume-stimulated increase in arterial pressure (see Table 1) and the endothelin-1-induced increase in arterial pressure and ANF (see Fig. 3). This concentration exceeds the effective concentration used by others to inhibit stretch-induced ANF release in vitro (3 μmol/l; see Ref. 21) and the cardiac hypertrophy by volume overload in vivo (0.5 μmol/l; see Ref.10). Similarly, the dose of the ET-B receptor antagonist IRL-1038 (0.1 mg/rat) was selected to exceed the effective concentration necessary (0.3 μmol/l) for reducing by 87% the coronary vasodilatation evoked by an ET-B receptor agonist (23).
Changes in arterial plasma ANF concentrations were measured (seeAssays) as an indicator of changes in ANF secretion. Inseries 1, mean arterial blood pressure was measured by connecting the tail artery catheter to a Statham P23 Db pressure transducer. Heart rate was measured with a tachometer that was triggered by the rising phase of the arterial pressure pulse. Central venous pressure was monitored in some animals, but the artifacts induced by the shifts in position of the freely moving rats made the measurements of small pressure changes (≅1 mmHg) difficult, and this measurement was discontinued. Mean arterial pressure and heart rate were plotted on a chart recorder (model 67RP5-A, Grass, Quincy, MA). Hematocrit was measured in some groups as a control, showing that it decreased similarly by <2.5%; that is, 1.5 ± 0.9% after the first 3.3-ml infusion and <0.6% after the second and third 3.3-ml infusion.
In series 2, blood pressures including central venous pressure and heart rate were recorded in the conscious rats by connecting the catheters to pressure transducers (model BLPR, World Precision Instruments, Sarasota, FL) using a computerized data acquisition system (HEM 3.1 software, Notocord, Croissy, France). Because the rats were lying quietly in the plastic cylinder, accurate pulsatile central venous pressure measurements could be obtained.
In series 1, the baseline (average of samples 1 and2) and the change from baseline (Δ) at each time point (samples 3–6) were calculated for plasma ANF, mean arterial blood pressure, and heart rate for each rat. The mean change (Δ over samples 3–6) was also calculated for each rat and averaged over each group. In series 2, sample 1represented the baseline, and mean central venous pressure was included in the analysis. Means ± SE were then calculated at each time point for each particular test or control group. Changes within groups and differences between groups at each time point were analyzed for statistical significance by two-way ANOVA for repeated measures. Differences in baselines or mean changes from baseline were analyzed for statistical significance by ANOVA and by Duncan's multiple range tests. Calculations were performed with computer programs from the SAS Institute (Cary, NC) on a personal computer.
Role of Endothelin Receptors During Slow Blood Volume Expansion
The slow infusion for the H2O vehicle control caused robust changes in plasma ANF concentrations (Fig.1, closed circles), a significant increase (7.5 mmHg) in mean arterial blood pressure (Table 1), and no significant change in heart rate (Table 1). The wide-spectrum endothelin receptor antagonist bosentan, injected between samples 1 and 2, dose dependently inhibited the ANF response (Fig.1, open circles), the mean change decreasing from 64.1 to 52.6, 16.1, and 11.6 pg/ml with increasing dose (Table 1). Baselines of blood pressure and heart rate were similar to the control for all doses of bosentan (Table 1), as were the changes in hematocrit (not shown), though the increases in blood pressure were dose dependently reduced up to a dose of 0.33 mg/rat.
To better define the role of endothelin receptors, BQ-123 (1 mg/rat), a specific ET-A receptor antagonist, was injected in series 1between samples 1 and 2. Although the mean ANF response was decreased (39.7 vs. 64.1 pg/ml for the H2O control, Table 1), the appropriate DMSO control (50 μl DMSO in 450 μl saline) also elicited a lower mean ANF response, and the time course of plasma ANF for BQ-123 was quite similar to that of its DMSO control (Fig. 2 A). The increase in mean arterial blood pressure was abolished by BQ-123 relative to the DMSO control (P < 0.05). In contradistinction, the putative ET-B receptor antagonist IRL-1038 (0.1 mg) injected between samples 1 and 2 strongly attenuated the ANF response with respect to its DMSO control (20 μl DMSO in 280 μl saline) (Fig. 2 B; Table 1).
The results of series 1 did not support the view that ET-A receptors mediate the ANF responses to blood volume expansion as shown by others (15); the main difference in our study was the low rate of volume infusion. Thus a new protocol was applied in series 2 to compare the effects of BQ-123 on ANF responses to slow and fast infusions in the same animal.
ANF Responses to Slow and Fast Blood Volume Expansion: Effect of BQ-123
In series 2, BQ-123 had no effect on baseline plasma ANF nor on the plasma ANF reached after the slow blood volume expansion (Fig.3 A). The increase in plasma ANF (ΔANF) was similar to the peak increase observed in series 1(Fig. 2) and was not significantly reduced by BQ-123 (Fig. 3 C). However, for fast infusions, BQ-123 significantly inhibited the plasma ANF, the ΔANF being reduced by 49.9%. The BQ-123 was effective because the ANF response to endothelin-1 was clearly blunted, the ΔANF being reduced by 83.5%, and the arterial blood pressure increase measured 1 min after endothelin was totally abolished (Fig. 3,B and D). A decrease in heart rate (Fig. 3 F) correlated with the increased arterial pressure in the controls, as also shown by the absolute heart rates in Table2. Fifteen minutes after endothelin (not shown), arterial pressure was increased by 50.8 ± 3.2 mmHg in the controls (P < 0.0001) and by 13.5 ± 2.7 mmHg (P < 0.001) in the BQ-123 group, indicating that the BQ-123 had lost some of it efficacy within 20 min of its injection.
Mean baseline central venous pressure, measured at the heart level, was close to zero (Table 2). The increases in central venous pressure relative to baseline were similar in the absence or presence of BQ-123 (Fig. 3 E), and the absolute central venous pressures were also similar (Table 2). The slow infusion caused a significant, small increase (0.9–1.1 mmHg), as anticipated from preliminary measurements in series 1. The fast infusion caused an increase of 4.1–4.4 mmHg, whereas endothelin-1 tended to decrease central venous pressure by <0.2 mmHg. Thus BQ-123 did not change the atrial stretch imposed by slow and fast infusions nor by endothelin-1.
This study in conscious rats shows for the first time the role of endothelin receptors in ANF responses to a physiologically relevant, slow saline infusion calculated to expand the blood volume by <1.6% per minute. The new findings are that a combined ET-A/ET-B receptor blockade almost abolishes the ANF response to a slow volume expansion, and that the ANF response is blocked by an ET-B but not by an ET-A receptor antagonist. As previously reported (15), this study also shows that the ET-A receptor antagonist significantly inhibits the ANF response to an acute overload, calculated to expand the blood volume by 25% per minute.
Endothelin Receptors Mediate Volume-Stimulated ANF Secretion
Consistent with previous studies (6, 9, 14, 15, 21, 22), the results indicate that an endothelin receptor blockade by the nonselective antagonist bosentan dose dependently inhibits volume-stimulated ANF secretion (Fig. 1, Table 1). In contrast to previous studies (15, 21), this inhibition was almost total rather than partial. Furthermore, the ET-A antagonist BQ-123 had no effect on the ANF response, whereas the putative ET-B antagonist IRL-1038 abolished it (Fig. 2 and Table 1). Thibault et al. (22) and Leite et al. (14) detected mainly ET-A receptors in primary cultures of atrial cells and blocked endothelin-stimulated ANF secretion with BQ-123. Irons et al. (9) showed that atrial myocytes had 50 times more mRNA encoding ET-A receptors relative to the ET-B type. Skvorak et al. (21) found that BQ-123 attenuated ANF release from stretched isolated atria. Leskinen et al. (15) diminished the ANF response to acute volume loading in conscious rats by one-half with 1 mg/kg BQ-123. Thus two essential differences emerge with previously published studies: the total dependence of the ANF response on endothelin receptors, and the lack of significant effects of an ET-A receptor blockade.
Possible Factors Explaining Differences With Previous Studies
Several factors were considered: 1) the vehicle for dissolving the antagonists, the doses used, their efficacy, and their selectivity; and 2) the rate (or magnitude) of volume loading.
First, DMSO was used in our study to dissolve the BQ-123, whereas water or saline was used in other studies. DMSO itself partially inhibited the ANF response to volume loading. Yet IRL-1038, which was also dissolved in DMSO, completely abolished the response at a dose of 0.1 mg/rat. Furthermore, the dose of BQ-123 (1 mg/rat, Fig. 2) was three times higher than in the study by Leskinen et al. (15). Because the arterial blood pressure increase normally seen with volume loading was abolished (Table 1), BQ-123 did have biological activity in vivo. Thus although the selectivity of IRL-1038 for ET-B receptors may be questioned (8), the lack of effect of BQ-123 in Fig. 2 A is unlikely to be due to the vehicle used, the dose of the antagonist, or to a potential loss of its activity.
Second, the rate and magnitude of volume loading was lower than that in other studies. The peak of the ANF response was usually obtained following a slow infusion of 3.3 ml of saline-dextran over 8 min. Leskinen et al. (15) applied 6.2 ml of saline over 1 min, a rate of volume-loading 15 times and a total volume 2 times higher than that in our study. This explains why Leskinen et al. (15) measured an increase in right atrial pressure of 3–4 mmHg, whereas central venous pressure changes were approximately 1 mmHg in our study. It also explains why plasma ANF reached three to four times baseline for high rates of volume loading (15) and less than two times baseline in Figs.1 and 2. Similarly, Skvorak et al. (21) applied a rather large atrial pressure increase (5–7 mmHg) in vitro to demonstrate the effectiveness of BQ-123. For these reasons, we checked in series 2 whether the effects of BQ-123 might differ depending on the rate (and magnitude) of volume loading.
Role for ET-A Receptors in ANF Release to an Acute Volume Overload
The results of Fig. 3 confirm the study by Leskinen et al. (15) that high rates of volume loading evoke a huge increase in plasma ANF that attains three to four times baseline and that this response is reduced in half by BQ-123. Hemodynamic factors are not involved in the effect of BQ-123, because central venous pressure changes and arterial pressures and heart rates were the same (Fig. 3, Table 2). These results are compatible with the hypothesis that a large volume load induces endothelin release from nonmyocytes and that endothelin acts on myocyte ET-A receptors (9, 14, 20-22) to release ANF.
Although not directly tested in our experiments, the ET-B receptors probably play a minor role in the large ANF response to an acute volume overload. The ANF response to fast infusions is already strongly reduced by the ET-A receptor blockade, and bosentan, a blocker of both ET-A and ET-B receptors, does not block this ANF response (15). In addition, other factors such as mechanisms intrinsic to the atrial myocytes, angiotensin II (15), and prostacyclins are known to be involved (see Ref. 20 for review).
In the same experiments where BQ-123 strongly diminished the ANF response to fast infusions (Fig. 3), BQ-123 did not significantly inhibit the ANF response to slow infusions. The dose of BQ-123 applied (0.5 mg/rat = 1.5 mg/kg body wt) before the slow infusions, although lower than in Fig. 2, still exceeded the dose used by Leskinen et al. (15) (1 mg/kg body wt). The same dose applied before injection of endothelin-1 reduced the ANF release by 83.5% (Fig. 3). Thus the lack of effectiveness of BQ-123 during slow infusions cannot be attributed to an insufficient dose. Because the effectiveness of BQ-123 decreased gradually over 20 min (see results), BQ-123 was applied repeatedly over the 80-min experimental period to maintain equal test conditions. In conclusion, the lack of effect of BQ-123 on ANF responses to slow infusions in Fig. 3 cannot be explained by a lack of effectiveness of BQ-123, by an insufficient dose, or by hemodynamic factors. This suggests that the ET-B receptor subtype may be involved.
Role for ET-B Receptors in ANF Release to a Slow Blood Volume Expansion
Taken together, three lines of evidence in this study support the hypothesis that ET-B receptors mediate the ANF secretion to a slow blood volume expansion: 1) an almost total inhibition of the ANF response by the combined ET-A/ET-B receptor antagonist bosentan,2) a total inhibition of the ANF response by the ET-B receptor antagonist IRL-1038, and 3) a lack of significant effects of the ET-A receptor antagonist BQ-123. Others have shown that stretch stimulates the release of endothelin from nonmyocytes (16), ET-B receptors are present on nonmyocytes, including endothelial cells (17), and ET-B activation may lead to the release of an ANF-releasing substance such as prostacyclin (3, 18, 19). The fraction of ET-B receptors, 30% vs. 70% for ET-A, in whole (human) atria (7) far exceeds the small proportion (2%) anticipated from measurements of mRNA in atrial myocytes alone (9) and appears to confirm the presence of significant numbers of ET-B receptors on nonmyocytes.
Thus we propose a new model for ANF stimulation that is compatible with many published reports. A physiologically relevant blood volume expansion should stimulate the release of endothelin from nonmyocytes and activate endothelial ET-B receptors, resulting in the secretion of an ANF-releasing substance. The amounts of ANF released by this mechanism are modest, causing at most a doubling of the plasma ANF concentration. To check this hypothesis, further studies are needed to test the role of other factors such as prostacyclins and angiotensin II in ANF responses to slow volume expansion. Increasing rates of volume expansion would activate, in addition, the myocyte ET-A receptors and mechanisms intrinsic to the atrial myocytes to cause a large increase in plasma ANF.
This study was supported by National Heart, Lung, and Blood Institute Grant RO1 HL-40553, the American Heart Association Grant VA-94-G-25, and the Swiss National Science Foundation Grant FN 31–49798.96 (to A. J. Baertschi) and Grant FN 31–53860.98 (to T. Pedrazzini).
Address for reprint requests and other correspondence: A. J. Baertschi, Dept. of Physiology, Univ. of Geneva School of Medicine, 1 rue Michel Servet, 1211 Geneva 4, Switzerland (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society