High-Performance liquid Chromatographic Analysis of Catecholamines in Biological Samples by liquid/liquid Extraction Prepurification

HIRONORI

TSUCHIYA AND TOKISHI HAYASHI

A prepurification procedure for the determination of catecholamines in biological samples was studied to simplify the procedure and increase selectivity and recovery. The procedure is based on liquid/liquid extraction of catecholamine-borate complexes. Under optimal conditions, catecholamines formed complexes with diphenylborate in an alkaline NH4Cl/NH40H buffer that were extracted with n-heptanol. Catecholamines were reextracted into an HCI-acidified solution, followed by reversed-phase ion-pair high-performance liquid chromatographic separation with fluorometric detection. Application of the method to various samples of rat tissues and human urine showed procedural simplicity, high selectivity, and high recovery (about 90% or more for tissue samples). This method can be used for pharmacological studies on catecholamines. Key Words: uid/liquid

Catecholamines;

extraction;

Borate

High-performance

liquid

chromatography;

Liq-

complex

INTRODUCTION The catecholamines possess important

norepinephrine

functions

(NE),

epinephrine

as neurotransmitters

(E), and dopamine

in physiological

(DA)

states and relate

to various pathoses. Pharmacological studies on catecholamines require simple and selective analytical methods for biological samples. Such a requirement has provided, fluorometric, radioenzymatic, immunologic assays, gas chromatographic and high-performance In these lectivity,

purification

(HPLC) assay methods

(Krstulovic,

1986).

is essential

before

with regard to procedural

time) and low recovery from biological (Gerlo and Malfait, 1985; Shellenberger

purification

the actual

HPLC

separation.

sepreAl-

(Anton and Sayre, 1962; Shellenberger and Gordon, 1971) has been used for the prepurification. However, major problems have re-

in such a method

analytical and 70%) line

chromatographic

of catecholamines

umina treatment most frequently mained

liquid

methods, HPLC has been widely used because of its high resolution, and detection sensitivity. When one is analyzing biological samples,

methods

were

reported

simplicity

(tediousness

and long

samples (ranging between 50% and Gordon, 1971). Several on-

for procedural

simplification

and

auto-

From the Department of Dental Pharmacology (H.T.), Asahi University School of Dentistry, Motosu, Gifu, Japan, and the National Institute of Neuroscience (T.H.), National Center of Neurology and Psychiatry, Kodaira, Tokyo, Japan. Address reprint requests to: Dr. Tokishi Hayashi, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4-l-l Ogawahigashi, Kodaira, Tokyo 187, Japan. Received April 12, 1989; revised and accepted July 6, 1989. 21 journalof

Pharmacological

0 1990 Elsevier

Science

Methods Publishing

23, 21-30 Co.,

Inc.,

(1990) 655 Avenue

0160.5402/90/$3.50 of the Americas,

New

York,

NY 10010

22

H. Tsuchiya and T. Hayashi

mation (De Jong et al., 1987; Tsuchiya et al., 1989a). However, such methods require additional and complex systems. Compounds with 1,2-diol moieties readily form complexes with borate compounds. When forming complexes of catechol compounds with appropriate borates, such complexes can be efficiently trapped on solid-phase (Higa et al., 1977) or extracted into organic liquid-phase (Herregodts et al., 1985; Smedes et al., 1982; Tsuchiya and Hayashi, 1989). Liquid/liquid extraction based on complex formation with the borate is suitable for purification of catecholamines in biological samples because it is possible to carry out a simple procedure that offers high selectivity and recovery. In the present study, a prepurification method by solvent extraction of the cate~holamine-borate complexes using HPLC with fluorometric detection was developed. The utility of this method was evaluated in various biological samples, including rat tissues and human urine. MATERIALS AND METHODS NE hydrochloride, E bitartrate, and 3,4-dihydroxybenzylamine (DHBA) hydrobromide were purchased from Sigma (St. Louis, MO), and DA hydrochloride from Wako (Osaka, Japan). Their standard solutions (2.0 mol/mL of each) were prepared by dissolving them in 10 mM HCI containing 5 mM sodium pyrosulphite as a preservative. These solutions were stored at 4°C in the dark. They were diluted daily to appropriate concentrations with 1 mM HCi or water. Several borate compounds were obtained from Tokyokasei (Tokyo, lapan). A diphenylborate (DPB) solution was prepared by dissolving diphenylboric acidlethanolamine (Aldrich, Milwaukee, WI) in 2.0 M NH&I/NH40H buffer (pH 8.2, containing 10 mM EDTA-2Na) at a concentration of 0.3% (w/v). Sodium octanesulphonate used as a counter ion for the HPLC separation was of ion-pair chromatographic grade (Nakarai, Kyoto, Japan). Acetonitrile and methanol used in the mobile phase were of liquid chromatographic grade. All other reagents were of analytical reagent grade. Water was redistilled by an all-glass apparatus after purification by a Milli Q water purification system (Nihon Millipore, Tokyo, Japan). All glass apparatus used in the present study was silanized by treating with 5% (v/v) dichlorodimethylsilane solution in toluene and washing with methanol. Male Wistar strain rats, 7-wk old and weighing 200-250 g, were sacrificed by exsanguination under light ether anesthesia. All tissues were immediately removed, placed on ice, washed with ice-cooled saline, blotted with filter paper, and weighed. The tissue was homogenized in a glass homogenizer maintained in an ice bath in 4.0-100.0 mL (depending on catecholamine content in the tissues) of ice-cooled 0.4 M perchloric acid solution containing DHBA (200.0 pmollml) as an internal standard (IS), 1 mM EDTA-2Na and 2 mM sodium pyrosulphite as preservatives. The homogenate was centrifuged at 10,000 g for IO min at 4°C and the supernatant was stored at -80°C until analysis. After the pH of the supernatant (0.5 mL) was adjusted at 6 to 7 with 3.0 M NaOH, 3.0 mL of n-heptanol and 1.0 mL of the DPB solution were added to the supernatant. The mixture was vortex-mixed for 1.5 min and

HPLC Analysis of Catecholamines

centrifuged at about 1,000 g for 1 min. To the organic phase, 0.4 mL of 1.5 m HCI was added, and the mixture was vortex-mixed for 1.5 min. After the mixture was centrifuged at about 1,000 g for 1 min, 0.2 mL of the aqueous phase was subjected to HPLC analysis. Urine was collected from healthy male subjects, aged 30-37 yr, who were drug free. To 0.5 mL of urine, 0.05 mL of the DHBA solution (2.0 nmol/mL) as the IS, 3.0 mL of n-heptanol, and 1 .O mL of the DPB solution were added, and then extraction was carried out in a similar manner to that used for the tissue sample. The organic phase was washed three times by vortex-mixing with 3.0 mL of 20 mM sodium phosphate buffer (pH 8.2) containing 1 mM EDTA-2Na for 30 set, followed by reextraction with 0.4 mL of 0.15 M HCI. An aliquot of 0.2 mL of the aqueous phase was subjected to HPLC analysis. The HPLC separation was performed using a similar procedure to one previously reported (Tsuchiya et al., 1989b). A 150 x 4.6-mm i.d. stainless steel column laboratory-packed with NS Gel C-18 (particle size: 5 km; Sakata, Tokyo, Japan) was used for the HPLC separation. A mobile phase consisting of 20 mM sodium octanesulphonate, 4% (v/v) acetonitrile, 4% (v/v) methanol, and 92% (v/v) 0.25 M sodium phosphate buffer (pH 3.0, containing 0.1 mM EDTA-2Na) was pumped into the HPLC column at a flow rate of 1.8 mUmin and at a column temperature of 45°C. After the separation, catecholamines were detected by means of native fluorescence (280 nm for excitation and 325 nm for emission wavelength). Catecholamines in tissue and urine samples were determined based on calibration graphs, which were prepared by treating standard catecholamine solutions duplicating those used for biological samples, and plotting the peak height ratios of catecholamines to the IS. Prepurification conditions were optimized by analyzing peak heights of catecholamines in duplicate, and the results are shown as their means. For the estimation of analytical recovery and reproducibility, rat tissue and human urine samples were analyzed repeatedly according to the methods described above after spiked with (n = 8) and without (n = 2) standard catecholamines of the amounts corresponding to their actual contents in such samples. Total recoveries were obtained by comparing differential peak heights between the spiked sample and the nonspiked sample with those of nonextracted standard solutions. Relative recoveries were also obtained by comparing the differential peak height ratios with those of the nonextracted standard solutions. RESULTS AND

DISCUSSION

Four catecholamines, including the IS, were extracted into organic solvents using several borates such as DPB, phenylborate, m-aminophenylborate, ferroceneborate, and n-butylborate. Extractability of the catecholamines increased depending on the DPB concentrations in NH4CI/NH40H buffer and reached a plateau at concentrations over 0.3% (w/v) as shown in Figure 1. The effect of the DPB concentration was greater for more hydrophilic catecholamines such as NE and E. Increasing concentrations of the other borates showed no effect on extraction of any catechola-

23

24

H. Tsuchiya and T. Hayashi

0

a2 Diphenylborate

FIGURE 1.

concentration

0.4 in

buffer

(%I

Effect of diphenylborate

concentration on extraction of cartecholamines. Catecholamines (200.0 pmollml of each, 0.5 ml) were extracted with 2.0 M NH&WNHeOH buffer (pH 8.6,l.O ml) containing diphenylborate of different concentrations (%, w/v) and n-hexanol (3.0 ml), followed by HPLC separation. Number: (1) NE; (2) E; (3) DHBA; (4) DA.

mines, which is similar to the extraction of L-3,4_dihydroxyphenylalanine (L-dopa)DPB complex (Tsuchiya and Hayashi, 1989). These results are due to higher hydrophobic substituents in DPB than in the other borates. Maximum extraction of the catecholamine-DPB complexes was achieved at pH 7.8-8.8 of NH4CI/NH40H buffer and remarkably decreased above pH 8.8, as shown in Figure 2. Kamperman and Kraak (1985) reported that normetanephrine and metanephrine were partly converted into NE and E when extracting over pH 8. Such methylated catecholamines were vortex-mixed with NH4CI/NH40H buffers of pH 7.8-9.0 for 1 J-5.0 min. However, no conversion into NE and E was detected at pH values lower than 8.2 during 1.5-min mixing. Small unknown peaks appeared near the retention times of NE and E when mixing for 5 min at pH values higher than 8.6. The stability of both normetanephrine and metanephrine during extraction

HPLC Analysis of Catecholamines

4

8

6

10

pH of buffer FIGURE 2. Effect of pH on extraction of catecholamines. Catecholamines (200.0 pmol/ml of each, 0.5 ml) were extracted with 2.0 M NH4CI/NHIOH buffer (containing 0.3% (w/v) diphenyl-borate, 1.0 ml) of different pH and n-hexanol (3.0 ml), followed by HPLC separation. Number: (1) NE; (2) E; (3) DHBA; (4) DA.

observed in this study agree with the report of Herregodts et al. (1985). A pH of 8.2 was chosen as optimum in the present study. The catecholamines could be extracted from NH,CI/NH,OH, potassium phosphate, NH,OH/phosphate, or trimethylamine/HCI buffers, but could not be recovered from a Tris/HCI buffer. Since DPB forms negatively charged complexes with catecholamines (Smedes et al., 19821, the extractability of the complexes can be increased by forming ion-pairs with cationic counter ions. The lack of recovery from the Tris/HCI buffer may be ascribed to the hydrophilic moiety in cationic trisfhydroxymethyh-aminomethane. Tetraalkylammonium cations with larger hydrophobic moieties are effective for the increased extractability of the complexes (Smedes et al., 1982; Tsuchiya and Hayashi, 1989). In the present study, however,

25

26

H. Tsuchiya and T. Hayashi

they were not added to prevent simultaneous extraction of fluorescent anionic contaminants in biological samples, although the recovery was slightly inferior to those obtained from the use of tetraalkylammonium bromide. Catecholamines were extracted with various organic solvents as shown in Table 1. However, two major interfering peaks were also extracted from such extraction solvents. From these results, n-heptanol was chosen in the present study. All catecholamines, including L-dopa, were simultaneously separated within 11 min as shown in Figure 3. Results of the other tissue samples revealed similar chromatographic profiles. Although one peak was derived from the DPB reagent, it did not interfere with the determination of the catecholamines, indicating satisfactory selectivity of this method. Unknown large peaks sometimes appeared at retention times longer than 45 min, and their appearance depended on the source of the samples. Washing of the HPLC column wtih 40% (v/v) aqueous acetonitrile solution for about 8 min solved the problem of interfering peaks. Calibration graphs showed a good linearity for all catecholamines in a range of 5.0-1000.0 pmol/mL of sample solution. Table 2 summarizes total recovery and relative recovery together with the reproducibility of this method. Recoveries from tissue samples were higher than conventional alumina treatments, which are typiEffect of Extraction Solvent on Peak Heights of Catecholamines

TABLE 1

PEAKHEIGHT (ARBITRARYUNIT) NE

SOLVENT

E

n-Butanol

IPb

ND=

n-Pentanol

43.58 40.32 45.11 44.94 IP

38.53 35.90 35.81 33.96 32.30

n-Hexanol n-Heptanol n-Octanol n-Nonanol n-Pentane n-Hexane n-Heptane c-Hexane

ND

ND

ND

38.20

ND

ND

ND

ND

524.11

ND ND

ND ND 9.37

ND

35.75

48.31

ND

ND 10.06

Ethyl acetate

IP

Chloroform

6.96

0.3%

(200.0

pmol/mL

of each,

(w/v) diphenyl-borate,

9.28 ND 3.01

0.5 mL) were

pH 8.2,

extracted

ND ND 20.91

interfering

’ Overlapped

peaks.

with interfering

’ Could

not be detected.

’ Buffer

containing

0.2%

ND

ND

38.33

4.79

ND

12.29

ND

ND

ND

ND

ND