selected ion monitoring: abnormalities in diabetics

Prostaglandins, Leukotrienes and Essential FattyAcids (1996) 54(6), 445--449

© Pearson ProfessionalLtd 1996

Analysis of the thromboxane/ prostacyclin balance in human urine by gas chromatography/selected ion monitoring: abnormalities in diabetics T. Hishinuma 1, Grace S. P. Yu 1, M. Takabatake 1, Y. Nakagawa 1, K. Ito 1, M. Nishikawa 1, M. Ishibashi 2, K. SuzukP, M. Matsumoto 4, T. Toyoda 4, M. Mizugak£ 1Department of Pharmaceutical Sciences, Tohoku University Hospital, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-77, Japan. 2Research Laboratories, Takata Seiyaku Co. Ltd, Ohmiya, Japan. 3Department of Internal Medicine, Tohoku-koseinenkin Hospital, 10 Aza-Takasago, Fukumuro, Miyagino-ku, Sendai 983, Japan. 4The Third Department of Internal Medicine, Tohoku University, 2-1 Seiryo-machi, Aoba-ku, Sendai 980-77, Japan

Summary We microanalyzed 2,3-dinor-6-keto-prostaglandin FI~ (2,3-dinor-6-keto-PGFl~ 1) and 11-dehydrothromboxane B2 (11-dehydro-TXB2, 2) in human urine. Samples containing a [2H4]-analogue as an internal standard were extracted by chromatography using Sep Pak tC18 and silica gel. The compounds were then analysed by means of the lactone ring opening reaction and dimethylisopropylsilylation. The conversion of 1 to 1-methyl ester (M E)-propylamide (PA)-9,12,15-dimethylisopropylsilyl (DMIPS) ether derivative and of 2 to 1-ME-6-methoxime (MO)-9,12,15-tris-DMIPS ether derivative was followed by gas chromatography/selected ion monitoring (GC/SIM). Interfering substances from the urine matrix were eliminated during GC/SIM analysis using a DB-5 column. We were able to detect 1 (222-1031 pg/mg creatinine) and 2 (18-155 pg/mg creatinine) in human urine. Furthermore, the thromboxane/prostacyclin (IX/PGI) ratio in the urine of diabetics was higher than that of healthy volunteers. This method can be used to determine the TX/PGI balance in human urine.

INTRODUCTION Arachidonic acid (AA), a fatty acid esterified to inward facing phospholipids on the cell membrane, can be converted into prostaglandins (PC), thromboxanes (TX) and leukotrienes (LT) and has various physiological functions. For example, the balance of TXA2, a potent vasoconstrictor

Received 10 July 1995 Accepted 5 October 1995 Correspondence to: M. Mizugaki, Tel. (81) 22 274 t111 Ext. 2830; Fax. (81)

22 274 1977.

and platelet activator, and PGI2, a platelet anti-aggregator, is important in blood coagulation. However, few reports have described the determination of these two forms at the same time and the relationship between the TX/PGI balance and diseases accompanying clotting. 1 Prostaglandins can be determined by immunoassay 2 or by gas chromatography/negative-ion chemical ionization mass spectrometry (GC/NICIMS)2 Cross reactivity of antibodies can decrease the sensitivity of immunoassays, whereas the resolution of GC/NICIMS is low and it cannot distinguish contaminants of the same molecular weight that have carboxylic acid groups. On the other 445

446

Hishinuma et al

hand, GC/electron impact ionization mass spectrometry (GC/EIMS) gives more information about the chemical structure and it can determine microquantities of PGs with high resolution. Diabetes mellitus induces platelet alterations such as hyperaggregability.4 Variations in PG production seem to be related to this phenomenon but the changes in PG levels remain unclear. For example, reports of prostacyclin levels in diabetics are conflicting; they are reported to be high,5-Zlow8-1° and the same as controls. 11,12These differences may have been due to the manner in which the PG measured was detected by the assay method employed. The major urinary metabolites of PGI2 and TXA2 are 2,3-dinor-6-PGG~ (1) and 11-dehydro-TXB2 (2). They are of interest in understanding the physiological role of the parent compounds. In this report, we determined the TX/PGI balance, by means of GC/HR-SIM using 1 and ~. as markers, in healthy volunteers and diabetics, and examined the levels in relation to diabetes mellitus.

MATERIALS AND METHODS Samples and reagents 2,3-Dinor-6-keto-PGFl~ (1), [3,3,4,4-2H412,3-dinor-6-ketoprostaglandin F~ 11-dehydro-TXB2 and [3,3,4,4-2H4] 11dehydro-TXB2 were purchased from Cayman Chemicals Co. (Ann Arbor, MI). [19,19,20,20-2H4]-1 (IS) was donated by Upjohn Pharmaceuticals (ibaragi, Japan). DMIPSimidazole was purchased from Tokyo Kasei Kogyo (Tokyo, Japan). The Sep-Pak tC18 cartridge was purchased from Waters Associates (Milford, USA). Bond Elut Silica cartridges were obtained from Analytichem International (Harbor City, CA). Diazomethane was prepared from Nmethyl-N-nitroso-p-toluenesulfonamide. Other solvents and reagents used were of analytical grade.

Gas chromatography/mass spectrometry A DX303 mass spectrometer (JEOL, Japan) was interfaced with a MS-GCG06 gas chromatograph (JEOL, Japan) which was equipped with an all glass VandenBerg-type solventless injector. The column was a 25 m x 0.25 m m I.D. fused silica capillary column cross-linked with methylsilicone (DB-5, J & W Scientific Co., Folsom, CA). The temperature of the column oven was maintained at 200°C for 1 min and increased to 320°C at 8°C/min. The carrier gas was helium with a linear velocity of about 30 cm/s. The temperature of the injection port and the transfer line was kept at 320°C and 280°C and that of the ion source at 250°C. The ionization energy and the trap current were 70 eV and 100 gA, respectively. The accelerating voltage was 3 kV.

GC/selected ion monitoring (GC/SIM) GC/SIM was performed under the same conditions as used in GC/MS but with a trap current of 250 mA. The ions at m/z 642 (1) and m/z 646 ([2H4]2,3-dinor-6-ketoprostaglandin FI~), rn/z 698 (2) and m/z 702 ([2H4]11dehydro-TXB2) were monitored. Sample preparation from human urine Human urine from healthy volunteers or non-insulindependent (Type 2) diabetic patients was investigated. After sampling, the urine was stored at -80°C until assayed. Determination of 2,3-dinor-6-keto-PGFl~ (1) Compound 1 was extracted and purified by a simplified method of Mizugaki et al. 13To h u m a n urine (10 ml), [2H4]1 (5 ng) was added as an internal standard (IS). Compound 1 can occur in a spirolactone form. Leading 1 to the lactone form, the sample was acidified to pH 2.5 with 1 N He1, allowed to stand for 30 min at room temperature, and passed through a Sep Pak tC 18 cartridge. The cartridge was washed with water (10 ml) and n-hexane (10ml). Compound I and IS were eluted with ethyl acetate (10 ml) and the eluate was evaporated to dryness. The residue was dissolved in n-hexane:ethyl acetate (2:1, 3 ml), then transferred onto a silica gel column (3 x 0.5 cm I.D.). The column was washed with n-hexane:ethyl acetate (2:1, 8 ml) and n-hexane:ethyl acetate (1:2, 1 0 m 1). The lactone form of 1 was eluted with n-hexane:ethyl acetate (1:3, 10 ml). After evaporating the solvent under reduced pressure, to open the lactone ring, the residue was dissolved with 1% O-methylhydroxylamine-HC1 pyridine solution (100 gl) and allowed to stand for 1 h at 60°C. The pyridine was evaporated under reduced pressure. The resulting MO derivative was treated with ethereal diazomethane (1 ml) for 30 min at room temperature. After evaporation, the residue was sflylated with DMIPS-imidazole (20 gl) for 30 min at 60°C. The reaction product was dissolved in n-hexane (2 ml), and passed through Bond Elut Silica with n-hexane (4 ml). The ME-MO-DMIPS ether derivative was eluted with nhexane:ethyl acetate (98:2, 10 ml). After evaporating the solvent, the residue was dissolved in n-hexane:pyridine (99:1, 100 gl) and used for GC/HR-SIM. Determination of 11-dehydro-TXB2 (2) Compound 2 was extracted and purified by modifying the method of Ishibashi et al.14 To the urine (10 ml), [2H4] an a~ logue (2.5 ng) was added as an IS. Compound 2 occursqn lactone and lactone ring-opened (acyclic) forms. Leading

Prostaglandins, Leukotrienes and Essential Fatty Acids (1996) 54(6), 445-449

© Pearson Professional Ltd 1996

447

Determination of TX/PGI balance in human urine

2 to the lactone form, the sample was acidified to pH 2.5 with 1 N HC1, then allowed to stand at room temperature for I h. The mixture was applied to a Sep Pak tC 18 cartridge, and washed with acidified water (10 ml) followed by n-hexane (10 ml). Compound 2 was eluted with ethyl acetate (10 ml). The eluate was evaporated to dryness and ?. in the residue was dissolved in methanol (0.1 ml). To this solution was added ethereal diazomethane, and the mixture was allowed to stand at room temperature for 15 rain. After evaporation of the solvent, the residue was dissolved in n-hexane:ethyl acetate (2:1, 3 rnl) and then transferred onto a silica gel column (3 × 0 5 cm i.d.). The column was washed with n-hexane:ethyl acetate (2:1, 10 ml), then the methyl ester of 2 was eluted with n-hexane:ethyl acetate (1:1, 10 ml). The eluate was evaporated to dryness and to open the lactone ring, npropylamine (0.3 ml) was added to the residue which was allowed to stand for 15 rain. After evaporating the excess of n-propylamine, the residue was silylated with DMIPSimidazole as described above.

Ma~

.0701

Z

4

5

Creatinine in h u m a n urine was determined using a Creatinine test kit (Wako Pure Chemical Industries, Osaka, Japan). The results are shown as corrected values. RESULTS Gas chromatography/selected

ion monitoring (GC/SIM)

Compounds 1 and 2, enzymatic metabolites of PGI2 and TXB2, respectively, were derivatized into ME-MO-DMIPS and ME-PA-DMIPS ether products. The derivatives of 1 and 2 were then measured by GC/SIM according to the intensity of their [M-43] + fragments, m/z 642 and m/z 698, respectively. Figure 1 shows typical tracings of 1 and 2 extracted from h u m a n urine in the presence of [d4]-I (m/z 646) and [d4]-2 (m/z 702) as IS, respectively. The retention times of IS were characterized using standard samples on the capillary column and interfering substances were eliminated during microanalysis. The calibration graphs for 1 and 2 were obtained by plotting the peak area ratio to IS against their weight ratios. The linearity was good in the range of 10 p g 100 ng/tube, which covered the concentrations found in 2.5-20 ml of normal h u m a n urine. Application to the analysis of human urine

lOOO

2000

© Pearson Professional Ltd 1996

0000

~a~ 4,4~59 ,

.

.

4 ,

.

11-dehydro-TXB2

u

(m/z

.

.

6

.

,

.

~ i

.

.

.

1o L

4ooo

.

18 '

.

..

ld ,

~c~n

R'T*ma5 .

"

698)

.._J internalstandard(mlz702)

"1.0 2800

3000

4000

Scan

Fig. 1 Selected ion recording from a 2,3-dinor-6-ketoprostaglandin FI~ (compound 1) methyl ester-methoximedimethylisopropylsilyl ether derivative and 11-dehydro-thromboxane B2 (compound 2) methyl ester-propylamide-dimethylisopropylsilyl ether derivative. Compound 1, d4-1 (IS), compound 2 and d4-2 (iS) were monitored at m/z 642, 646, 698 and 702, respectively.

studied. None had taken medications for at least two weeks prior to urine collection. Four of them were smokers. The results are plotted in Figure 2. Urine samples were collected after lunch. The excretion level of compound 1 was 78.6 + 30 (mean + SD, range 42-143) pg/mg A. 2,3-dinor-6-keto-PGFl~

a) t-

B. 11-dehydro-TXB2

250

1000

200

8OO

150

6OO

.m

@

.m

.=

o

0

o~

E

@

@ • 442

100

03

7 8 6 . 8 7 5 . 2

50

0

to

I

]

]

spot pooled

We investigated the production of prostacyclin and thromboxane in healthy volunteers to determine the levels of compounds 1 and 2 using GC/SIM analysis. 11 volunteers (10 males and 1 female, 31.8 + 8.3 years) were

~ . T~,

I~

internalstandard(m/z646)

1000

Creatinine contents

10

8

2,3-dinor-6-keto.PGFl=(m/z642)

urine urine (N=11) (N=10)

°

!

200

i

spot pooled urine urine (N=11) (N=10)

Fig. 2 Determination of 2,3-dinor-6-keto-PGFl~ (A) and 11dehydro-TXB2 (B) levels in the healthy adult urine.

Prostaglandins, Leukotrienes and Essential Fatty Acids (1996) 54(6), 445-449

448

Hishinuma et al

creatinine and of compound 2 was 442 + 94 (range 297-664) pg/mg creatinine (Fig. 2). The mean of the TX/ PGI ratio of each sample was 6.3. We determined the production of prostacyclin and thromboxane in healthy volunteers using pooled urine for 24 h. 10 volunteers (8 males and 2 females, 27.5 + 4.9 years) were studied (Fig. 2 pooled urine). The excretion levels of compound 1 were 75.2 + 39 (range 42-155) pg/ mg creatinine, 87.0 + 35 (range 39-150) n g / 2 4 h and compound 2 was 393 + 116 (range 268-664) pg/mg creatinine, 510 + 217 (range 241-920) ng/24 h. The mean of the thromboxane/prostacyclin ratio of each sample was 6.1. There was no difference in the range between the spot and pooled samples urine. In the following experiments, spot urine samples were determined and compared. The TX/PGI ratio was about 6.0 in healthy volunteers. We investigated the contents of compounds 1 and 2 in diabetics who had no kidney injury. Urine samples were collected after lunch. The excretions of 1 and 2 from 10 diabetics [8 males and 2 female out-patients, 64.9 + 5.4 years (53-72 years)] were studied. The results are plotted in Figure 3. The excretion level of compound 1 was 48.0 + 12 (range 32-63) pg/mg creatinine and that of compound 2 was 398 + 137 (range 209-564) pg/mg creatinine. The mean of TX/PGI ratio of each sample was 9.0. The amount of compound 1 was significantly lower and the TX/PGI ratio in diabetics seemed to be higher than that of healthy volunteers. Furthermore, 9 in-patients with diabetes [4 males and 5 females, 51.2 + 19 years (17-61 years)] were studied. The results are shown in Table 1. The excretion levels of compounds 1 and 2 were 78.4 + 51 (range 21-153) and 658 + 245 (range 439-1031) pg/mg creatinine. The mean A. 2,3-dinor-6-keto-PGFl~ 25O

O ec-

200

O O

150

B. 11-dehydro-TXB2 1000

p<0.05 I

800 I

600

100 O.

50 0

l

1 I

I •

398

400

78.6 1

1

I

I

I

200

64.9

•

Healthy Diabetics volunteers (N=10) (N=11) Fig. 3

442

•

E

I

0

I

I

Healthy Diabetics volunteers (N=10) (N=11)

Determination of 2,3-dinor-6-keto-PGFl~, (A) and 11healthy volunteers and diabetic out-patients.

dehydro-TXB2 (B) in

Table 1 Urinary excretion of prostacyclin and thromboxane metabolites in healthy volunteers and diabetics Groups

No & sex

Healthy 10M, 1F volunteers Diabetics 8M, 2F (out-patients) Diabetics 4M, 5F (in-patients)

2,3-Dinor11-Dehydro-TXB2 TX/PGI 6-keto-PGFl~ (TX. pg/mg cre.) (PG, pg/mg cre.) 78.6 + 30

442 _+94

6.3

48.0 + 12"

398 _+ 137

9.0

78.4_+ 51

658_+ 245**

8.4

M: males, F: females. Values are expressed as means _+S.D. * P < 0.05 vs. healthy volunteers and in-patients. ** P < 0.01 vs. healthy volunteers and out-patients.

of TX/PGI ratio of each sample was 8.4 and there was no difference from that of the out-patients. This value of 2 was significantly higher than that of healthy volunteers. DISCUSSIONS

The combination of the column-chromatographic sample preparation using silica gel and a chemically stable DMIPS ether derivative should be useful in the microanalysis of compounds I and 2 in human urine. The use of lactone ring opening reactions using methoxyamine and propylamine for compounds 1 and 2, respectively, were particularly advantageous. In addition, selection of the exact mass during GC/SIM enabled the selective detection of compounds I and 2 without serious interference from the urine matrix. In healthy volunteers, the content of compound 1 was 78.6 + 30 pg/mg creatinine in spot urine and 84.0 + 47.7 pg/mg creatinine in pooled urine for 24 h. By measuring the creatinine contents, there may be not so much difference between the results in spot urine and that pooled over 24 h. These results of compound 1 are similar to those reported. 15-~s Further, the content of compound 2 was 442 + 94 pg/ mg creatinine in spot urine and 393 + l l 4 p g / m g creatinine in pooled urine for 24 h. The levels of urinary compound 2 were reported to be 370 + 137 (n = 10) 29 or 382 + 147 pg/mg creatinine (n = 23, 27-70 year).2° Compared with these data, the results presented here were similar. Next, we determined the levels of urinary compounds 1 and 2 in diabetics, in whom the incidence of clotting is high, and compared the contents and TX/PGI balance with those of healthy volunteers. The mean content of compound 1 was lower in diabetics than in healthy volunteers and the TJUPGI ratio in diabetics was higher than that in healthy volunteers. These results suggested that the high ratio of thromboxane in diabetics is one reason for hypercoagulation.

Prostaglandins, Leukotrienes and Essential Fatty Acids (1996) $4(6), 445-449

© Pearson Professional Ltd 1996

Determination of TX/PGI balance in human urine

M a n y factors, for e x a m p l e , a g e a n d m i c r o a n g i o p a t h y , 21 affect t h e l e v e l s of t h r o m b o x a n e a n d p r o s t a c y c l i n . N u m e r i c a l u n d e r s t a n d i n g of c h a n g e s in t h e T X / P G I b a l a n c e is n e e d e d .

ACKNOWLEDGEMENTS The authors are grateful to Dr Y. Kurosaki, Upjohn Pharmaceuticals, for the gift of [2H4]2,3-dinor-6-keto-prostaglandin FI~. A part of this research was supported by Grants-in-Aid for Scientific Research on Priority Areas and Grants-in-Aid for Developmental Scientific Research from The Ministry of Education, Science and Culture, Japan.

REFERENCES 1. Needleman P., Raz A., Minkes M. S., Ferrendelli J. A., Sprecher H. Triene prostaglandins: prostacyclin and thromboxane biosynthesis and unique biological properties. Proc Natl Acad Sci USA 1979; 76: 944-948. 2. Vrbanac J. J., Eller T. D., Knapp D. R. Quantitative analysis of 6-keto-PGF~ using immunoaffinity purification and gas chromatography-mass spectrometry. J Chromatogr 1988; 425: 1-9.

3. FitzGerald G. A., Lawson J., Blair L A., Brash A. R. Analysis of urinary metabolites of TX and PGI by negative-ion chemicalionization GC/MS. Adv Prostaglandins Thromboxane Leukot Res 1985; 15: 87-90. 4. Dyerberg J., Bang H. O., Stoffersen E., Moncada S., Vane J. R. Eicosapentaenoic acid and prevention of thrombosis and atherosclerosis? Lancet 1978; ii: 117-119. 5. Harrison H. E., Reece A. H., Johnson M. Effect of insulin treatment on prostacyclin in experimental diabetes. Diabetologia 1980; 18: 65-68. 6. Harrison H. E., Johnson M., Raftery A. T. Vascular prostacyclin may be reduced in man. Lancet 1979; i: 325-326. Z Silberbauer K., Schernthaner G., Sinzinger H., Piza-Katzer H., Winter M. Decreased vascular prostacyclin in juvenile-onset diabetes. N Engl J Med 1979; 300: 366-36Z 8. Lane L. S., Jansen P. D., Lahav M. Circulating prostacydin and thromboxane levels in patients with diabetic retinopathy. Ophthalmology 1982; 89: 763-766.

© Pearson Professional Ltd 1996

449

9. R6sen P., Schr6r K. Increased prostacyclin released from perfused hearts of acutely diabetic rats. Diabetologia 1980; 18: 391-394. 10. Ylikorkala O., Kaila J., Viinikka L. Prostacyclin and thromboxane in diabetes. BrMedJ 1981; 283:1148-1150. 11. Davis T. M. E., Mitchell M. D., Turner R. C. Prostacyclin and thromboxane metabolites in diabetes. Lancet 1979; ii: 789-790. 12. Davis T. M. E., Bown E., Finch D. R. In-vitro venous prostacyclin production, plasma 6-keto-prostaglandin FI~ concentrations, and diabetic retinopathy. Br MedJ 1981; 282: 1259-1262. 13. Mizugaki M., Hishinuma T., Yu Grace S. P. et al. Microanalysis of 2,3-dinor-6-keto-prostaglandin FI~ in human urine using gas chromatography/high resolution-selected ion monitoring. J Chromatogr 1994; 658:11-19. 14. Ishibashi M., Nakagawa Y., Harima N., Hishinuma T., Mizugaki M. Analysis of 1 I-dehydrothromboxane t32 in human urine improved by use of gas chromatography/high-resolution selected ion monitoring with 1802-1abelled analogue as an internal standard. Biol Mass Spectrom 1994; 23: 612-620. 15. Fischer S., Bernutz C., Meier H., Weber P. C. Fmwnation of prostacyclin and thromboxane in man as measured by the main urinary metabolites. BiocMm Biophys Acta 1985; 8?6: 194-199. 16. FitzGerald G. A., Smith B., Pedersen A. K., Brash A. R. Increased prostacyclin biosynthesis in patients with severe atherosclerosis and platelet activation. NEnglJMed 1984; 310: 1065-1068. 1Z Wennmalm .~., Benthin G., Granstr6m E. F., Persson L., Winell S. 2,3-Dinor metabolites of thromboxane A2 and prostacyclin in urine from healthy human subjects: Diurnal variation and relation to 24 h excretion. Clfn Sci 1992; 83: 461-465. 18. Lorenz R., Helmer P., Uedelhoven W., Zimmer B., Weber P. C. A new method using simple solid phase extraction for the rapid gas-chromatographic mass-spectrometric determination of 11dehydro-thromboxane B2 in urine. Prostaglandins 1989; 38: 157-170. 19. Morrow J. D., Minton T. A. Improved assay for the quantification of 11-dehydrothromboxane B2 by gas chromatography-mass spectrometry. J Chromatogr 1993; 612:179-185. 20. Landolfi R., Ciabattoni G., Patrignani P. et al. Increased thromboxane biosynthesis in patients with poly cythemia vera: evidence for aspirin-suppressible platelet activation in vivo. Blood 1992; 80: 1965-1971. 21. Dollery C. T., Friedman L. A., Hensby C. N. et al. Circulating prostacyclin may be reduced in diabetes (Letter). Lancet 1979; ii: 1365.

Prostaglandins, Leukotrienes and Essential Fatty Acids (1996) 54(6), 445-449