ANALYTICAL
196,84-88
BIOCHEMISTRY
(19%)
Fluorometric Determination of Plasma Adenosine Concentrations Using High-Performance Liquid Chromatography Katsuyuki
Miura,
Department
Received
Michiaki
of Pharmacology,
November
Okumura, Tokihito Yukimura, and Kenjiro Yamamoto Osaka City University Medical School, Abenu-ku, Osaka 545 Japan
1,199O
Methods are described for the fluorometric determination of plasma adenosine concentrations, using HPLC. Plasma obtained from blood of dogs treated with erythro-(2-hydroxy-3-nonyl)adenine hydrochloride and dipyridamole was deproteinized with perchloric acid and the neutralized sample was put sequentially onto a SepPak Cl8 and boronic acid affinity column. Subsequently, adenosine in the final elution was converted to 1 ,p-ethenoadenosine and was quantitated by HPLC with a fluorescence detector. The percentage recovery of adenosine added to the deproteinized plasma was nearly 100%. In the adenosine deaminase treated plasma, the increase in adenosine concentration of even 4 nM can be accurately determined. The control renal venous plasma concentrations of adenosine in anesthetized dogs were 19.9 + 1.9 nM, a significantly higher value than the corresponding arterial concentrations (12.7 -t- 1 .l nM), thereby suggesting the renal release of adenosine. This release was markedly enhanced following the removal of the renal arterial occlusion. Thus, taken together with the in uiuo results, the present method is sensitive, hence most useful for the determination of plasma adenosine concentrations. 0 1991 Academic Press, Inc.
Adenosine has an important role in the regulation of the cardiovascular and central and peripheral nervous systems (1). To elucidate the involvement of adenosine in the related process, methods have been developed to determine concentrations of adenosine in tissue extracts, artificial perfusate, plasma, and urine (2-10). Included are HPLC separation combined with uv detection (2,9) or fluorescence detection as l,N6ethenoadenosine (5-8,lO) and radioligand binding assay (3), including radioimmunoassay (4). However, the latter method requires specific binding protein (such as
antibody) and the use of radioactive materials; hence, applicability is limited. During the development of the present method, we found the plasma concentrations of adenosine in dogs to be extremely low (-20 nM). This range of concentration is hardly detectable by a conventional uv detection following HPLC separation, thus a highly sensitive and accurate method is needed. With the conversion of adenosine to fluorescent 1,iV6-ethenoadenosine, the sensitivity can be increased approximately lo- to 20-fold compared to findings in case of uv detection (6). Despite the high sensitivity of fluorometric determination with HPLC, the application of this method to plasma samples has not been established and use was limited to tissue extracts (5,6,8), incubation medium (7) and artificial perfusate (10). In the present paper, we used this method to determine plasma adenosine concentrations and verified the assay procedures, including the handling process and extraction methods. MATERIALS
AND METHODS
[U-14C]Adenosine (554 mCi/mmol) came from Amersham. Dipyridamole and adenosine were from Sigma Chemical Co. Adenosine deaminase was purchased from Boehringer Mannheim. erythro-(2Hydroxy-3nonyl)adenine hydrochloride (EHNA)’ was obtained from Burroughs Wellcome Co. 2’-Deoxycoformycin was a kind gift from Yamasa Shoyu Co. (Chiba, Japan). A prepacked boronic acid affinity column (Affipak) was obtained from Pierce Chemical Co. and 40% aqueous solution of chloroacetaldehyde (Practical Grade) was purchased from Wako Pure Chemical Industries (Osaka, Japan). Other reagents were of analytical grade and were from commercial sources.
1 Abbreviation hydrochloride.
used:
EHNA,
erythro-(2-hydroxy-3-nonyl)adenine
84 All
Copyright 0 1991 rights of reproduction
0003-2697/91$3.00 hy Academic Press, Inc. in any form reserved.
FLUOROMETRIC
DETERMINATION
Blood Collection A 5ml disposable syringe prewetted with heparin was placed on ice. Three milliliters of blood were gently drawn into the syringe from a polyethylene catheter cannulating the blood vessel of the anesthetized dog. The blood sample was immediately placed in a chilled plastic tube containing stopping solution and then was inversion-mixed. The stopping solution consisted of 50 ~1 of 8.2 mM dipyridamole (dissolved in 0.01 N HCl saline) and 30 ~1 of 1 mM EHNA to give a final concentration of 135 and 10 pM, respectively. Preparation
of Adenosine
Extract
from Plasma
Blood samples were immediately centrifuged at 4°C (2OOOg, 5 min) and about 1.5 ml of plasma was mixed with 70 ~1 of 60% perchloric acid during vigorous vortexing. After centrifugation at 15,000g (4°C 20 min), the supernatant was neutralized by mixing with twice the volume of the Alamine/Freon reagent (0.5 N trioctylamine in 1,1,2-trichloro-1,2,2-trifluoroethane) (11) and then was centrifuged. One milliliter of the upper aqueous phase (approximately pH 5.5) was loaded onto a SepPak Cl8 cartridge (Waters) prewashed with 5 ml of methanol and 10 ml of distilled water. This procedure was followed by a serial washing twice with 1 ml of water and twice with 0.5 ml of 3% methanol, and the adenosine-containing fraction was eluted with 4 ml of 60% methanol, under negative pressure using a SepPak rack (Waters). The aliquot was evaporated to dryness under reduced pressure and resuspended in 2 ml of 0.25 M ammonium acetate buffer (pH 8.8). A 1.8-ml sample of the aliquot was then applied onto a boronic acid affinity column (Affipak, Pierce). As described by Uziel et al. (12), this column binds the cis-diol containing molecule, including adenosine, at alkaline pH. The column was washed with 10 ml of 0.25 M ammonium acetate buffer (pH 8.8). Adenosine was then eluted with 8 ml of 0.1 N acetic acid and the aliquots were evaporated under reduced pressure. Sample Analysis The resulting residue was dissolved in 500 ~1 of distilled water with sonication. Two hundred microliters of the aliquot were transferred into a microcentrifuge tube and 40 ~1 of 1 M acetate buffer (pH 4.5) and 9 ~1 of 40% chloroacetaldehyde were added. The preparation was incubated at 60°C for 4 h for conversion of adenosine to a fluorescent l,N’-ethenoadenosine (6,8). As a standard, 100 nM adenosine was incubated, under the same condition. After the incubation, the sample was mixed with twice the volume of Alamine/Freon reagent and centrifuged at 1OOOg for 3 min and the upper aqueous phase collected. For HPLC, a reverse-phase HPLC column (Nucleosil 7C18, 4.6 mm X 25 cm, Nagel) was
OF
ADENOSINE
IN
PLASMA
85
maintained at 30°C with a column oven (655A-52, Hitachi). An isocratic elution with 7.5% acetonitrile in 20 mM potassium phosphate buffer (pH 5.7) at a flow rate of 1.32 ml/min was performed with a HPLC pump (L600, Hitachi). A hundred microliters of the sample were injected with an autosampler (655A-40, Hitachi) and elution was monitored using a fluorescence spectrometer (F-1000, Hitachi) at an excitation wavelength of 280 nm and an emission wavelength of 410 nm. A chromatointegrator (D-2000, Hitachi) was used for recording. In every 10 samples, a standard sample was injected. Adenosine was quantitated by comparing the peak height with that for standard adenosine. Recovery
Study
Recovery of adenosine between blood collection and deproteinization. Since adenosine can be rapidly eliminated from the plasma compartment by uptake by blood cells and by enzymatic catabolism by adenosine deaminase (13), adenosine might be lost before deproteinization, a step where the plasma adenosine becomes stable. To assure the recovery of adenosine before deproteinization, heparinized canine blood withdrawn through the aortic catheter was placed in 1.5-ml microcentrifuge tubes and preincubation was carried out at 37°C. [ 14C] adenosine was added to each tube (1.75 X lo5 dpm/ml, 143 nM) at time zero. Three kinds of stopping solutions or saline were added before and at 30 s and 3 min after the addition of [14C]adenosine during incubation at 37°C. The stopping solutions used were MnCl, plus dipyridamole (4), EHNA plus dipyridamole (9), or deoxycoformycin (14) plus dipyridamole. Immediately after [14C]adenosine came into contact with stopping solution or saline in the whole blood, the tubes were cooled on ice and spun at 4°C at 2000g for 5 min. The obtained plasma was deproteinized with perchloric acid, centrifuged, and neutralized with Alamine/Freon reagent, as described above. One hundred microliters of the upper aqueous phase was injected onto a reverse-phase HPLC column (7C18 Nucleosil, 4.6 mm X 25 cm, Nagel) and eluted isocratically with 7.5% acetonitrile in 20 mM potassium phosphate (pH 5.7), while monitoring the radioactivity, (TRI-CARB RAM 7500, Packard). The fractions containing adenosine and inosine were collected and mixed with 10 ml of Aquasol(New England Nuclear), and the radioactivity was determined in a liquid scintillation counter (Model 460, Packard). Recovery of adenosine following deproteinization. We made use of two steps: that is, SepPak Cl8 and boronic acid affinity column and a significant loss of adenosine can occur during these processes. Therefore, we examined the recovery of adenosine following deproteinization. Adenosine deaminase (Boehringer Mannheim) was added to dog plasma at a final concentration of 1 U/ml and the preparation was incubated for 1 h at 40°C
86
MIURA
ET
AL.
to eliminate any remaining endogenous adenosine. The adenosine deaminase treated plasma was deproteinized and neutralized as described. The plasma extract was put into eight tubes (3.5 ml) and adenosine was added to each tube to reach the final concentration of 0, 2, 4, 8, 16, 32, 64 and 128 nM. Triplicate l-ml samples from each tube were serially put onto a SepPak Cl8 and boronic acid affinity column and adenosine concentrations were determined twice, on different days, as described. Adenosine Concentration Plasma in Dogs
in Arterial
and Renal Venous
Arterial and renal venous blood samples (3 ml each) were simultaneously collected from anesthetized dogs (n = 29). In three dogs, the left renal artery was completely occluded with a bulldog clamp for 45 min. Blood samples were withdrawn from the renal vein just before the occlusion and 0 (within 30 s after reflow), 2,5,15,30, and 60 min after the release of the occlusion. All blood samples were processed as described and plasma concentrations of adenosine were determined. All data are shown as mean _+ SE, unless otherwise stated. RESULTS
Validation
AND
DISCUSSION
of the Assay Procedure
Recovery of adenosine between blood collection and deproteinization. Figure 1 shows the effect of delaying the addition of each stopping solution, MnCl, plus dipyridamole (4), EHNA plus dipyridamole (9), or deoxycoformycin (14) plus dipyridamole, and the effect of no stopping solution. When whole blood was pretreated with each stopping solution, the recovery of 14C-radioactivity as adenosine was almost 100%. Even with a 30-s delay in adding the stopping solution, over 90% of the radioactivity was recovered as adenosine and the remaining radioactivity corresponded to inosine (Fig. la) whereas 60-70% of the radioactivity remained as adenosine and the radioactivity of inosine increased in the absence of stopping solution (Fig. lb). When [14C]adenosine was incubated in the whole blood for 3 min, over 20% of the radioactivity disappeared from the plasma compartment and the radioactivity of inosine increased with a decline in the radioactivity of [“Cladenosine in the plasma compartment. These results are in marked contrast to the rapid elimination of plasma adenosine from the pig (15) and human (13) blood but are consistent to the reported half-life of 3 min in case of canine blood (13). As about 20-30 s are needed for blood collection, the loss of endogenous plasma adenosine is minimal before deproteinization when using any combination of stopping solution described for the present experiment.
x*
0
,.I,
"
0
5
10
0
5
10 (min)
F
a
IncubfWon in whole
0
a0 set period blood
3 mln
of 14C-adenosine~
Disappearance of [“Cladenosine from the plasma comof canine whole blood and its inhibition by stopping solutions. [“Cladenosine was added to tubes containing whole blood preincubated at 37°C at time zero. Three kinds of stopping solutions were put in each tube before (0), and at 30 s and 3 min after the addition, as shown on the abscissa of the right graph. On the left are shown typical chromatographs of neutralized acid extracts of plasma. Closed arrow indicates the peak for [“Cladenosine (retention time, 6.4 min). Open arrows indicate the peak for [“Clinosine (retention time, 3.6 min) that was assessed by the peak shift of [“Cladenosine following reaction with adenosine deaminase. (a) EHNA and dipyridamole was added 30 s after the addition of [“Cladenosine. (b) No stopping solution, but saline was added 30 s after addition of [“Cladenosine. The right graph shows the recovery of radioactivity as adenosine in the plasma compartment (shown as percentage of initially added radioactivity of [“Cladenosine) plotted against time of incubation of whole blood with [“C!]adenosine before adding stopping solution. Eluates corresponding to adenosine (5.7-7.7 min) and inosine (2.5-4.5 min) were collected, respectively, and the radioactivity was determined. Stopping solutions used were: l , deoxycoformycin (1 pM) and dipyridamole (136 @M), A, EHNA (10 pM) and dipyridamole (136 pM), W, MnCl, (10 mM) and dipyridamole (136 PM), 0, no stopping solution (saline). FIG.
1.
partment
Recovery of adenosine following deproteinization and the detection limit. Although frequent use is a mode of uv detection and fluorometric detection of 1,N6-ethenoadenosine with HPLC separation, a number of interfering substances preclude the analysis without a cleanup of the biological samples. In addition, despite the benefit of the fluorometric determination of adenosine (higher sensitivity compared to uv detection), we found no documentation of application of the methods for the determination of plasma adenosine. In our preliminary experiments, deproteinized plasma was cleaned up by either SepPak Cl8 or boronic acid affinity column alone, an approach often used before adenosine detection (2,9,16). As the huge peaks of fluorescence during HPLC precluded the analysis in either case (data not shown), we combined these two steps. Figure 2 shows typical recordings of the chromatograph for plasma samples and standard adenosine. The peak of l,NG-ethenoadenosine appeared at about 7.5-8 min after injection of the sample and this peak was clearly separated
FLUOROMETRIC
DETERMINATION
OF
ADENOSINE
IN
87
PLASMA
t
Ii 4
0
5
I
10
0
5
10 Onin)
FIG. 2. Representative chromatographic recording of 100 nM adenosine following incubation with chloroacetaldehyde (left) and plasma extract reacted in the same manner (right). The arrow indicates the peak corresponding to l,N6-ethenoadenosine. The concentration of adenosine in the final aliquot of plasma extract was 33.6 nM, a value corresponding to the plasma adenosine concentration of 18.7 nM.
from the preceding or following unknown peaks. The standard calibration curve was linear over a wide range (Fig. 3). Addition of authentic ethenoadenosine raised the peak height, without any additional peak (data not shown). When adenosine was added to the deproteinized plasma at a final concentration of 0, 2, 4, 8, 16, 32, 64, and 128 nM, the calculated adenosine concentrations
1
10
100
1000
10000
FIG. 3. Standard calibration curve for quantitation of adenosine. Adenosine was dissolved at final concentrations of 1,2,4,8,16,32,64, 128,256,512,1024, and 2048 nM and reacted with chloroacetaldehyde to form l,Np-ethenoadenosine (see Materials and Methods).
2 Chcentratii
4
1
8 of adenosine
16 added
1
32
I
64
to plasma
128 (nM)
FIG. 4. Relationship between the concentrations of adenosine added to plasma extract and the increase in adenosine concentration. Known concentrations (0,2,4,8,16,32,64, and 128 UM) of adenosine were added to the acid extract of plasma pretreated with adenosine deaminase. The concentrations of plasma adenosine were determined as described in the text and the increase in plasma adenosine concentration (Y, ordinate) was plotted against the concentration of adenosine added to plasma extract (X, abscissa). A simple linear regression analysis shows Y = 0.995X + 0.076 and r = 0.9996. Since linear regression analysis depends on the assumption of normality and homoscedasticity, a logarithmic transformation of both variable was made. Even in this case, the correlation coefficient was 0.9988, thereby indicating a straight-line relationship over a wide range (log Y = 1.021og x - 0.031)
were 13.0 + 0.5, 14.7 2 0.4, 16.9 + 0.3, 21.0 + 0.1, 28.8 + 1.0, 44.7 I 1.4, 78.3 +- 0.9, and 139.9 _t 2.6 nM (mean +- SD), respectively. No additional peak appeared with the spike of adenosine and only the enhancement of the peak for l,N6-ethenoadenosine was seen, thereby indicating that the peak corresponding to lp-ethenoadenosine (13.0 nM equivalent in plasma) in the nonspiked sample derived from adenosine, even in adenosine deaminase-treated plasma. Since an excess amount of adenosine deaminase was added (1 U/ml,; 1 unit will deaminate 1.0 pmol of adenosine per minute at pH 7.5 at 25”C), we interpreted this residual adenosine to the high K, of adenosine deaminase (-35 PM) (17) compared to the concentration of adenosine present in the plasma. The added adenosine was quantitatively recovered in the final extract (Fig. 4). The slope is close to 1 and the y intercept (=0.076) is cIose to 0 nM, thereby suggesting a complete recovery of adenosine added to the deproteinized plasma, even with two steps of cleanup. The coefficient of variance (CV) of the calculated increase with the addition of 2 nM adenosine was 16.2%. The CVs were decreased to 8.0,1.7,6.4,4.4,1.4, and 2.6% with the addition of 4, 8, 16, 32, 64, and 128 nM adenosine. The recovery of each concentration of adenosine added was
MIURA
!
a C Time
0 after
release
2 of renal
r7rlF-l 5
15
occlusion
30
60
(min)
FIG. 5.
Renal venous adenosine concentrations following the release of occlusion of the renal artery. The left renal artery of the anesthetized dog was completely occluded for 45 min. Renal venous blood samples were collected before occlusion (C) and after various times after reflow, as shown on the abscissa. Plasma adenosine concentrations were determined as shown on the ordinate. Time 0 means the adenosine concentration in the renal venous plasma obtained immediately after the reperfusion. Data represent mean +- SE.
91.0, 97.6, 100.2, 99.0, 99.2, 101.9, and 99.1%, respectively. These results clearly show that an increase in the plasma adenosine concentration exceeding 4 nM can be detected using the present method. Due to complete recovery of the added adenosine, values obtained were not corrected by the recovery rate. Practically speaking, since obvious fluorescent peaks disappeared within 30 min after injection of the sample onto the HPLC, adenosine measurements could be made every 30 min. Adenosine Concentration in Arterial Plasma in Dogs
and Renal Venous
Plasma adenosine concentrations were determined using arterial and renal venous blood obtained simultaneously from the aorta and the renal vein of anesthetized dogs (n = 29). Renal venous plasma concentration of adenosine was 19.9 +- 1.9 nM, a value significantly higher than corresponding arterial concentration (12.7 5 1.1 nM) (P < 0.001 by paired t test). This result clearly shows that adenosine is added to the renal venous effluent during even a single passage of arterial blood through the kidney. In addition, the difference of 7.2 nM was evident in plasma samples obtained from an in uivo study. The extremely high sensitivity of the present method is thereby given support. Postocclusive vasoconstriction is a unique characteristic of renal vasculature and was attributed to the renal vascular action of adenosine (18,19). Therefore we ex-
ET
AL.
amined whether in viuo reperfusion of the kidney would elicit an increase in adenosine release from the kidney. The renal venous concentrations of adenosine before renal ischemia were 16.6 -+ 1.0 nM and the values were remarkably elevated immediately after the removal of renal arterial occlusion to be followed by a rapid reversion to control values (n = 3) (Fig. 5). Thus, we obtained evidence that renal ischemia-reperfusion enhances the release of adenosine into the renal venous effluent. In summary, using methods described herein, a very low concentration of plasma adenosine concentration could be determined. As there was an efficient blockade of the elimination of plasma adenosine from the whole blood when stopping solution was added immediately after blood collection and that complete recovery of adenosine was attained with the entire procedure, the present method can be used to examine physiological and pathophysiological roles of adenosine. ACKNOWLEDGMENTS This work was supported thank M. Ohara for editorial
by a grant assistance.
from
Chichibu
Cement.
We
REFERENCES 1. Ohisalo, J. J. (1987) Med. Biol. 66, 181-191. 2. Gehrke, H. W., Kuo, K. C., Davis, G. E., and Suits, Chromatogr. 150,455-476. 3. Olsson, R. A., Davis, (1978) Anal. Biochem.
C. J., Gentry, 85,132-138.
R. D. (1978)
M. K., and Vomacka,
J.
R. B.
4. Sato, T., Kuninaka, A., Yoshino, H., and Ui, M. (1982) Anal. Bio&em. 121,409-420. 5. Wojcik, W. J., and Neff, N. H. (1982) J. Neurochem. 39,280-282. 6. Jacobson, M. K., Hemingway, L. M., Farrell, C. E. (1983) Am. J. Physiol. 245, H887-H890. 7. MacDonald, W. F., and White, T. D. (1985) 791-797.
T. A., and Jones, J. Neurochem.
8. Ramos-Salazar, A., and Baines, A. D. (1985) Anal. 9-13. 9. Jackson, E. K., and Ohnishi, A. (1987) Hypertension
Biochem.
45,
145,
10,189-197.
10. Slegel, P., Kitagawa, H., and Maguire, M. H. (1988) Anal. Biothem. 171,124-134. 11. Khym, J. X. (1975) Clin. Chem. 21.1245-1252. 12. Uziel, M., Smith, L. H., and Taylor, S. A. (1976) Clin. Chem. 22, 1451-1455. 13. Moser,
G. H., Schrader,
J., and Deussen,
A. (1989)
Am. J. Physiol.
256,C799-C806. 14. Capogrossi,
M. C., Holdiness, M. R., and Israili, Chromatogr. 227, 168-173. 15. Gewirtz, H., Brown, P., and Most, A. S. (1987) Biol. Med. 185, 93-100.
Z. H. (1982) Proc.
J.
Sot. Exp.
16. Harmenberg, J. (1983) J. Liquid Chromatogr. 6,655-666. 17. Bergmeyer, H. U., Grassl, M., and Walter, H-E. (1983) in Methods in Enzymatic Analysis (Bergmeyer, H.U. Ed.), 3rd ed., Vol. 2, pp. 126-328, Verlag Chemie, Wienheim. 18. Osswald, H., Schmitz, H.-J., and Kemper, R. (1977) Arch. 37 1,45-49. 19. Osswald, H. (1984) Trends Pharmacol. Sci. 5, 94-97.
P/Ii&em
196,84-88
BIOCHEMISTRY
(19%)
Fluorometric Determination of Plasma Adenosine Concentrations Using High-Performance Liquid Chromatography Katsuyuki
Miura,
Department
Received
Michiaki
of Pharmacology,
November
Okumura, Tokihito Yukimura, and Kenjiro Yamamoto Osaka City University Medical School, Abenu-ku, Osaka 545 Japan
1,199O
Methods are described for the fluorometric determination of plasma adenosine concentrations, using HPLC. Plasma obtained from blood of dogs treated with erythro-(2-hydroxy-3-nonyl)adenine hydrochloride and dipyridamole was deproteinized with perchloric acid and the neutralized sample was put sequentially onto a SepPak Cl8 and boronic acid affinity column. Subsequently, adenosine in the final elution was converted to 1 ,p-ethenoadenosine and was quantitated by HPLC with a fluorescence detector. The percentage recovery of adenosine added to the deproteinized plasma was nearly 100%. In the adenosine deaminase treated plasma, the increase in adenosine concentration of even 4 nM can be accurately determined. The control renal venous plasma concentrations of adenosine in anesthetized dogs were 19.9 + 1.9 nM, a significantly higher value than the corresponding arterial concentrations (12.7 -t- 1 .l nM), thereby suggesting the renal release of adenosine. This release was markedly enhanced following the removal of the renal arterial occlusion. Thus, taken together with the in uiuo results, the present method is sensitive, hence most useful for the determination of plasma adenosine concentrations. 0 1991 Academic Press, Inc.
Adenosine has an important role in the regulation of the cardiovascular and central and peripheral nervous systems (1). To elucidate the involvement of adenosine in the related process, methods have been developed to determine concentrations of adenosine in tissue extracts, artificial perfusate, plasma, and urine (2-10). Included are HPLC separation combined with uv detection (2,9) or fluorescence detection as l,N6ethenoadenosine (5-8,lO) and radioligand binding assay (3), including radioimmunoassay (4). However, the latter method requires specific binding protein (such as
antibody) and the use of radioactive materials; hence, applicability is limited. During the development of the present method, we found the plasma concentrations of adenosine in dogs to be extremely low (-20 nM). This range of concentration is hardly detectable by a conventional uv detection following HPLC separation, thus a highly sensitive and accurate method is needed. With the conversion of adenosine to fluorescent 1,iV6-ethenoadenosine, the sensitivity can be increased approximately lo- to 20-fold compared to findings in case of uv detection (6). Despite the high sensitivity of fluorometric determination with HPLC, the application of this method to plasma samples has not been established and use was limited to tissue extracts (5,6,8), incubation medium (7) and artificial perfusate (10). In the present paper, we used this method to determine plasma adenosine concentrations and verified the assay procedures, including the handling process and extraction methods. MATERIALS
AND METHODS
[U-14C]Adenosine (554 mCi/mmol) came from Amersham. Dipyridamole and adenosine were from Sigma Chemical Co. Adenosine deaminase was purchased from Boehringer Mannheim. erythro-(2Hydroxy-3nonyl)adenine hydrochloride (EHNA)’ was obtained from Burroughs Wellcome Co. 2’-Deoxycoformycin was a kind gift from Yamasa Shoyu Co. (Chiba, Japan). A prepacked boronic acid affinity column (Affipak) was obtained from Pierce Chemical Co. and 40% aqueous solution of chloroacetaldehyde (Practical Grade) was purchased from Wako Pure Chemical Industries (Osaka, Japan). Other reagents were of analytical grade and were from commercial sources.
1 Abbreviation hydrochloride.
used:
EHNA,
erythro-(2-hydroxy-3-nonyl)adenine
84 All
Copyright 0 1991 rights of reproduction
0003-2697/91$3.00 hy Academic Press, Inc. in any form reserved.
FLUOROMETRIC
DETERMINATION
Blood Collection A 5ml disposable syringe prewetted with heparin was placed on ice. Three milliliters of blood were gently drawn into the syringe from a polyethylene catheter cannulating the blood vessel of the anesthetized dog. The blood sample was immediately placed in a chilled plastic tube containing stopping solution and then was inversion-mixed. The stopping solution consisted of 50 ~1 of 8.2 mM dipyridamole (dissolved in 0.01 N HCl saline) and 30 ~1 of 1 mM EHNA to give a final concentration of 135 and 10 pM, respectively. Preparation
of Adenosine
Extract
from Plasma
Blood samples were immediately centrifuged at 4°C (2OOOg, 5 min) and about 1.5 ml of plasma was mixed with 70 ~1 of 60% perchloric acid during vigorous vortexing. After centrifugation at 15,000g (4°C 20 min), the supernatant was neutralized by mixing with twice the volume of the Alamine/Freon reagent (0.5 N trioctylamine in 1,1,2-trichloro-1,2,2-trifluoroethane) (11) and then was centrifuged. One milliliter of the upper aqueous phase (approximately pH 5.5) was loaded onto a SepPak Cl8 cartridge (Waters) prewashed with 5 ml of methanol and 10 ml of distilled water. This procedure was followed by a serial washing twice with 1 ml of water and twice with 0.5 ml of 3% methanol, and the adenosine-containing fraction was eluted with 4 ml of 60% methanol, under negative pressure using a SepPak rack (Waters). The aliquot was evaporated to dryness under reduced pressure and resuspended in 2 ml of 0.25 M ammonium acetate buffer (pH 8.8). A 1.8-ml sample of the aliquot was then applied onto a boronic acid affinity column (Affipak, Pierce). As described by Uziel et al. (12), this column binds the cis-diol containing molecule, including adenosine, at alkaline pH. The column was washed with 10 ml of 0.25 M ammonium acetate buffer (pH 8.8). Adenosine was then eluted with 8 ml of 0.1 N acetic acid and the aliquots were evaporated under reduced pressure. Sample Analysis The resulting residue was dissolved in 500 ~1 of distilled water with sonication. Two hundred microliters of the aliquot were transferred into a microcentrifuge tube and 40 ~1 of 1 M acetate buffer (pH 4.5) and 9 ~1 of 40% chloroacetaldehyde were added. The preparation was incubated at 60°C for 4 h for conversion of adenosine to a fluorescent l,N’-ethenoadenosine (6,8). As a standard, 100 nM adenosine was incubated, under the same condition. After the incubation, the sample was mixed with twice the volume of Alamine/Freon reagent and centrifuged at 1OOOg for 3 min and the upper aqueous phase collected. For HPLC, a reverse-phase HPLC column (Nucleosil 7C18, 4.6 mm X 25 cm, Nagel) was
OF
ADENOSINE
IN
PLASMA
85
maintained at 30°C with a column oven (655A-52, Hitachi). An isocratic elution with 7.5% acetonitrile in 20 mM potassium phosphate buffer (pH 5.7) at a flow rate of 1.32 ml/min was performed with a HPLC pump (L600, Hitachi). A hundred microliters of the sample were injected with an autosampler (655A-40, Hitachi) and elution was monitored using a fluorescence spectrometer (F-1000, Hitachi) at an excitation wavelength of 280 nm and an emission wavelength of 410 nm. A chromatointegrator (D-2000, Hitachi) was used for recording. In every 10 samples, a standard sample was injected. Adenosine was quantitated by comparing the peak height with that for standard adenosine. Recovery
Study
Recovery of adenosine between blood collection and deproteinization. Since adenosine can be rapidly eliminated from the plasma compartment by uptake by blood cells and by enzymatic catabolism by adenosine deaminase (13), adenosine might be lost before deproteinization, a step where the plasma adenosine becomes stable. To assure the recovery of adenosine before deproteinization, heparinized canine blood withdrawn through the aortic catheter was placed in 1.5-ml microcentrifuge tubes and preincubation was carried out at 37°C. [ 14C] adenosine was added to each tube (1.75 X lo5 dpm/ml, 143 nM) at time zero. Three kinds of stopping solutions or saline were added before and at 30 s and 3 min after the addition of [14C]adenosine during incubation at 37°C. The stopping solutions used were MnCl, plus dipyridamole (4), EHNA plus dipyridamole (9), or deoxycoformycin (14) plus dipyridamole. Immediately after [14C]adenosine came into contact with stopping solution or saline in the whole blood, the tubes were cooled on ice and spun at 4°C at 2000g for 5 min. The obtained plasma was deproteinized with perchloric acid, centrifuged, and neutralized with Alamine/Freon reagent, as described above. One hundred microliters of the upper aqueous phase was injected onto a reverse-phase HPLC column (7C18 Nucleosil, 4.6 mm X 25 cm, Nagel) and eluted isocratically with 7.5% acetonitrile in 20 mM potassium phosphate (pH 5.7), while monitoring the radioactivity, (TRI-CARB RAM 7500, Packard). The fractions containing adenosine and inosine were collected and mixed with 10 ml of Aquasol(New England Nuclear), and the radioactivity was determined in a liquid scintillation counter (Model 460, Packard). Recovery of adenosine following deproteinization. We made use of two steps: that is, SepPak Cl8 and boronic acid affinity column and a significant loss of adenosine can occur during these processes. Therefore, we examined the recovery of adenosine following deproteinization. Adenosine deaminase (Boehringer Mannheim) was added to dog plasma at a final concentration of 1 U/ml and the preparation was incubated for 1 h at 40°C
86
MIURA
ET
AL.
to eliminate any remaining endogenous adenosine. The adenosine deaminase treated plasma was deproteinized and neutralized as described. The plasma extract was put into eight tubes (3.5 ml) and adenosine was added to each tube to reach the final concentration of 0, 2, 4, 8, 16, 32, 64 and 128 nM. Triplicate l-ml samples from each tube were serially put onto a SepPak Cl8 and boronic acid affinity column and adenosine concentrations were determined twice, on different days, as described. Adenosine Concentration Plasma in Dogs
in Arterial
and Renal Venous
Arterial and renal venous blood samples (3 ml each) were simultaneously collected from anesthetized dogs (n = 29). In three dogs, the left renal artery was completely occluded with a bulldog clamp for 45 min. Blood samples were withdrawn from the renal vein just before the occlusion and 0 (within 30 s after reflow), 2,5,15,30, and 60 min after the release of the occlusion. All blood samples were processed as described and plasma concentrations of adenosine were determined. All data are shown as mean _+ SE, unless otherwise stated. RESULTS
Validation
AND
DISCUSSION
of the Assay Procedure
Recovery of adenosine between blood collection and deproteinization. Figure 1 shows the effect of delaying the addition of each stopping solution, MnCl, plus dipyridamole (4), EHNA plus dipyridamole (9), or deoxycoformycin (14) plus dipyridamole, and the effect of no stopping solution. When whole blood was pretreated with each stopping solution, the recovery of 14C-radioactivity as adenosine was almost 100%. Even with a 30-s delay in adding the stopping solution, over 90% of the radioactivity was recovered as adenosine and the remaining radioactivity corresponded to inosine (Fig. la) whereas 60-70% of the radioactivity remained as adenosine and the radioactivity of inosine increased in the absence of stopping solution (Fig. lb). When [14C]adenosine was incubated in the whole blood for 3 min, over 20% of the radioactivity disappeared from the plasma compartment and the radioactivity of inosine increased with a decline in the radioactivity of [“Cladenosine in the plasma compartment. These results are in marked contrast to the rapid elimination of plasma adenosine from the pig (15) and human (13) blood but are consistent to the reported half-life of 3 min in case of canine blood (13). As about 20-30 s are needed for blood collection, the loss of endogenous plasma adenosine is minimal before deproteinization when using any combination of stopping solution described for the present experiment.
x*
0
,.I,
"
0
5
10
0
5
10 (min)
F
a
IncubfWon in whole
0
a0 set period blood
3 mln
of 14C-adenosine~
Disappearance of [“Cladenosine from the plasma comof canine whole blood and its inhibition by stopping solutions. [“Cladenosine was added to tubes containing whole blood preincubated at 37°C at time zero. Three kinds of stopping solutions were put in each tube before (0), and at 30 s and 3 min after the addition, as shown on the abscissa of the right graph. On the left are shown typical chromatographs of neutralized acid extracts of plasma. Closed arrow indicates the peak for [“Cladenosine (retention time, 6.4 min). Open arrows indicate the peak for [“Clinosine (retention time, 3.6 min) that was assessed by the peak shift of [“Cladenosine following reaction with adenosine deaminase. (a) EHNA and dipyridamole was added 30 s after the addition of [“Cladenosine. (b) No stopping solution, but saline was added 30 s after addition of [“Cladenosine. The right graph shows the recovery of radioactivity as adenosine in the plasma compartment (shown as percentage of initially added radioactivity of [“Cladenosine) plotted against time of incubation of whole blood with [“C!]adenosine before adding stopping solution. Eluates corresponding to adenosine (5.7-7.7 min) and inosine (2.5-4.5 min) were collected, respectively, and the radioactivity was determined. Stopping solutions used were: l , deoxycoformycin (1 pM) and dipyridamole (136 @M), A, EHNA (10 pM) and dipyridamole (136 pM), W, MnCl, (10 mM) and dipyridamole (136 PM), 0, no stopping solution (saline). FIG.
1.
partment
Recovery of adenosine following deproteinization and the detection limit. Although frequent use is a mode of uv detection and fluorometric detection of 1,N6-ethenoadenosine with HPLC separation, a number of interfering substances preclude the analysis without a cleanup of the biological samples. In addition, despite the benefit of the fluorometric determination of adenosine (higher sensitivity compared to uv detection), we found no documentation of application of the methods for the determination of plasma adenosine. In our preliminary experiments, deproteinized plasma was cleaned up by either SepPak Cl8 or boronic acid affinity column alone, an approach often used before adenosine detection (2,9,16). As the huge peaks of fluorescence during HPLC precluded the analysis in either case (data not shown), we combined these two steps. Figure 2 shows typical recordings of the chromatograph for plasma samples and standard adenosine. The peak of l,NG-ethenoadenosine appeared at about 7.5-8 min after injection of the sample and this peak was clearly separated
FLUOROMETRIC
DETERMINATION
OF
ADENOSINE
IN
87
PLASMA
t
Ii 4
0
5
I
10
0
5
10 Onin)
FIG. 2. Representative chromatographic recording of 100 nM adenosine following incubation with chloroacetaldehyde (left) and plasma extract reacted in the same manner (right). The arrow indicates the peak corresponding to l,N6-ethenoadenosine. The concentration of adenosine in the final aliquot of plasma extract was 33.6 nM, a value corresponding to the plasma adenosine concentration of 18.7 nM.
from the preceding or following unknown peaks. The standard calibration curve was linear over a wide range (Fig. 3). Addition of authentic ethenoadenosine raised the peak height, without any additional peak (data not shown). When adenosine was added to the deproteinized plasma at a final concentration of 0, 2, 4, 8, 16, 32, 64, and 128 nM, the calculated adenosine concentrations
1
10
100
1000
10000
FIG. 3. Standard calibration curve for quantitation of adenosine. Adenosine was dissolved at final concentrations of 1,2,4,8,16,32,64, 128,256,512,1024, and 2048 nM and reacted with chloroacetaldehyde to form l,Np-ethenoadenosine (see Materials and Methods).
2 Chcentratii
4
1
8 of adenosine
16 added
1
32
I
64
to plasma
128 (nM)
FIG. 4. Relationship between the concentrations of adenosine added to plasma extract and the increase in adenosine concentration. Known concentrations (0,2,4,8,16,32,64, and 128 UM) of adenosine were added to the acid extract of plasma pretreated with adenosine deaminase. The concentrations of plasma adenosine were determined as described in the text and the increase in plasma adenosine concentration (Y, ordinate) was plotted against the concentration of adenosine added to plasma extract (X, abscissa). A simple linear regression analysis shows Y = 0.995X + 0.076 and r = 0.9996. Since linear regression analysis depends on the assumption of normality and homoscedasticity, a logarithmic transformation of both variable was made. Even in this case, the correlation coefficient was 0.9988, thereby indicating a straight-line relationship over a wide range (log Y = 1.021og x - 0.031)
were 13.0 + 0.5, 14.7 2 0.4, 16.9 + 0.3, 21.0 + 0.1, 28.8 + 1.0, 44.7 I 1.4, 78.3 +- 0.9, and 139.9 _t 2.6 nM (mean +- SD), respectively. No additional peak appeared with the spike of adenosine and only the enhancement of the peak for l,N6-ethenoadenosine was seen, thereby indicating that the peak corresponding to lp-ethenoadenosine (13.0 nM equivalent in plasma) in the nonspiked sample derived from adenosine, even in adenosine deaminase-treated plasma. Since an excess amount of adenosine deaminase was added (1 U/ml,; 1 unit will deaminate 1.0 pmol of adenosine per minute at pH 7.5 at 25”C), we interpreted this residual adenosine to the high K, of adenosine deaminase (-35 PM) (17) compared to the concentration of adenosine present in the plasma. The added adenosine was quantitatively recovered in the final extract (Fig. 4). The slope is close to 1 and the y intercept (=0.076) is cIose to 0 nM, thereby suggesting a complete recovery of adenosine added to the deproteinized plasma, even with two steps of cleanup. The coefficient of variance (CV) of the calculated increase with the addition of 2 nM adenosine was 16.2%. The CVs were decreased to 8.0,1.7,6.4,4.4,1.4, and 2.6% with the addition of 4, 8, 16, 32, 64, and 128 nM adenosine. The recovery of each concentration of adenosine added was
MIURA
!
a C Time
0 after
release
2 of renal
r7rlF-l 5
15
occlusion
30
60
(min)
FIG. 5.
Renal venous adenosine concentrations following the release of occlusion of the renal artery. The left renal artery of the anesthetized dog was completely occluded for 45 min. Renal venous blood samples were collected before occlusion (C) and after various times after reflow, as shown on the abscissa. Plasma adenosine concentrations were determined as shown on the ordinate. Time 0 means the adenosine concentration in the renal venous plasma obtained immediately after the reperfusion. Data represent mean +- SE.
91.0, 97.6, 100.2, 99.0, 99.2, 101.9, and 99.1%, respectively. These results clearly show that an increase in the plasma adenosine concentration exceeding 4 nM can be detected using the present method. Due to complete recovery of the added adenosine, values obtained were not corrected by the recovery rate. Practically speaking, since obvious fluorescent peaks disappeared within 30 min after injection of the sample onto the HPLC, adenosine measurements could be made every 30 min. Adenosine Concentration in Arterial Plasma in Dogs
and Renal Venous
Plasma adenosine concentrations were determined using arterial and renal venous blood obtained simultaneously from the aorta and the renal vein of anesthetized dogs (n = 29). Renal venous plasma concentration of adenosine was 19.9 +- 1.9 nM, a value significantly higher than corresponding arterial concentration (12.7 5 1.1 nM) (P < 0.001 by paired t test). This result clearly shows that adenosine is added to the renal venous effluent during even a single passage of arterial blood through the kidney. In addition, the difference of 7.2 nM was evident in plasma samples obtained from an in uivo study. The extremely high sensitivity of the present method is thereby given support. Postocclusive vasoconstriction is a unique characteristic of renal vasculature and was attributed to the renal vascular action of adenosine (18,19). Therefore we ex-
ET
AL.
amined whether in viuo reperfusion of the kidney would elicit an increase in adenosine release from the kidney. The renal venous concentrations of adenosine before renal ischemia were 16.6 -+ 1.0 nM and the values were remarkably elevated immediately after the removal of renal arterial occlusion to be followed by a rapid reversion to control values (n = 3) (Fig. 5). Thus, we obtained evidence that renal ischemia-reperfusion enhances the release of adenosine into the renal venous effluent. In summary, using methods described herein, a very low concentration of plasma adenosine concentration could be determined. As there was an efficient blockade of the elimination of plasma adenosine from the whole blood when stopping solution was added immediately after blood collection and that complete recovery of adenosine was attained with the entire procedure, the present method can be used to examine physiological and pathophysiological roles of adenosine. ACKNOWLEDGMENTS This work was supported thank M. Ohara for editorial
by a grant assistance.
from
Chichibu
Cement.
We
REFERENCES 1. Ohisalo, J. J. (1987) Med. Biol. 66, 181-191. 2. Gehrke, H. W., Kuo, K. C., Davis, G. E., and Suits, Chromatogr. 150,455-476. 3. Olsson, R. A., Davis, (1978) Anal. Biochem.
C. J., Gentry, 85,132-138.
R. D. (1978)
M. K., and Vomacka,
J.
R. B.
4. Sato, T., Kuninaka, A., Yoshino, H., and Ui, M. (1982) Anal. Bio&em. 121,409-420. 5. Wojcik, W. J., and Neff, N. H. (1982) J. Neurochem. 39,280-282. 6. Jacobson, M. K., Hemingway, L. M., Farrell, C. E. (1983) Am. J. Physiol. 245, H887-H890. 7. MacDonald, W. F., and White, T. D. (1985) 791-797.
T. A., and Jones, J. Neurochem.
8. Ramos-Salazar, A., and Baines, A. D. (1985) Anal. 9-13. 9. Jackson, E. K., and Ohnishi, A. (1987) Hypertension
Biochem.
45,
145,
10,189-197.
10. Slegel, P., Kitagawa, H., and Maguire, M. H. (1988) Anal. Biothem. 171,124-134. 11. Khym, J. X. (1975) Clin. Chem. 21.1245-1252. 12. Uziel, M., Smith, L. H., and Taylor, S. A. (1976) Clin. Chem. 22, 1451-1455. 13. Moser,
G. H., Schrader,
J., and Deussen,
A. (1989)
Am. J. Physiol.
256,C799-C806. 14. Capogrossi,
M. C., Holdiness, M. R., and Israili, Chromatogr. 227, 168-173. 15. Gewirtz, H., Brown, P., and Most, A. S. (1987) Biol. Med. 185, 93-100.
Z. H. (1982) Proc.
J.
Sot. Exp.
16. Harmenberg, J. (1983) J. Liquid Chromatogr. 6,655-666. 17. Bergmeyer, H. U., Grassl, M., and Walter, H-E. (1983) in Methods in Enzymatic Analysis (Bergmeyer, H.U. Ed.), 3rd ed., Vol. 2, pp. 126-328, Verlag Chemie, Wienheim. 18. Osswald, H., Schmitz, H.-J., and Kemper, R. (1977) Arch. 37 1,45-49. 19. Osswald, H. (1984) Trends Pharmacol. Sci. 5, 94-97.
P/Ii&em
