applications and limitations of measurement of 15-keto,13,14-dihydro prostaglandin e2 in human blood...

PROSTAGLANDINS

APPLICATIONS AND LIMITATIONS OF MRASURRMHNT OF 15-ICRTO,13,14-DIHYDRO PROSTAGLANDIN E2 IN HUMAN

BLOOD BY RADIOIMMUNOASSAY

Stewart A. Mets, M.D. Maureen G. Rice, Ph.D. R. Paul Robertson, M.D.

From the Division of Clinical Pharmacology, Veterans Administration Medical Center, 4435 Beacon Avenue South, Seattle, Washington 98108, and the Departments of Medicine and Pharmacology, University of Washington, Seattle, Washington 98195.

This workwas supported by the Medical Research Service of the Veterans Administration.

Send reprint requests to: Stewart A. Metz, M.D. Division of Clinical Pharmacology Veterans Administration Medical Center 4435 Beacon Avenue South Seattle, Washington 98108

Key Words: Prostaglandin Metabolism Prostaglandin E 15-keto-13,14-dghydro Prostaglandin E2 Lipolysis Free Fatty Acids Prostaglandin Radioimmunoassay

JUNE 1979 VOL. 17 NO.6 839

PROSTAGLANDINS

ABSTRACT

It has been anticipated that the inherent limitations of radioimmunoassays for prostaglandin E (PGE) would be obviated by assays for its major circulating metabolite, 15-keto, 13,14- dihydro PGE2 (KHz-PGE2) which has a longer half-life in blood. We examined the effects of PGE2 infusion and alterations in lipolysis in vivo, and of clotting, prolonged storage and hemolysis in -- vitro, on KI$-PGE2 immunoreactivity in unextracted human pi&ma and serum samples. Indeed IU-I~-PGEZ levels rose several hundred fold during infusions of PGE2 at doses which cause little or no increment in peripheral PGE levels. During stimulation of lipolysis by infusions of epinephrine, apparent KHz-PGE2 levels rose five- fold. However, the dilution curve of plasma obtained during stimulation of lipolysis was not parallel to the standard curve; furthermore, apparent KHz-PGE2 levels were correlated strongly with free fatty acid (FFA) levels, suggesting that FFA's cross- reacted in the RIA weakly but significantly due to their very high molar concentration in blood. Clotting and prolonged storage of samples, but not hemolysis, also caused marked apparent increments in KH2-PGE2 levels. Competition curves using dilutions of such samples were again not parallel to the standard curves in plasma or buffer, but resembled dilution curves of samples containing high levels of FFA. These results suggest that handling of human blood samples for KHz-PGE2 measurement must be carefully standardized to avoid significant artifacts which presumably are due in'part to fatty acids released from triglyceride stores in vivo or from -- disrupted membrane phospholipids in vitro. Unextracted plasma appears to be unsatisfactory for use in this RIA.

840 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

Considerable interest exists in clinical research in the measurement of the primary prostaglandins in blood and tissue. For example, recent evidence has suggested a possible involvement of prostaglandins, particularly of the E series (PGE), in the hypercalcemia of cancer (l-3), in diabetes mellitus (4), in Bartter's Syndrome (5,6) and in tumor-induced diarrhea (7). However, major difficulties exist in the application of available techniques to the measurement of the native prostaglandins in biologic samples. The most accurate tool for the measurement of prostaglandins has been by gas chromatography/mass spectrometry. However, this technique is laborious, requires large samples and is not suitable for routine and repetitive clinical use.

Although several radioimmunoassays (RIA) for PGE have been developed, a number of problems have prevented their universal acceptance. First, specificity is a problem, since investigators using different radioimmunoassay systems have reported widely divergent levels of plasma PGE. Furthermore, such levels often are greater than those measured by mass spectrometry. Second, a number of artifacts attributable to the handling of blood or tissue samples have been reported which render the use of the RIA for native PGE potentially inaccurate (for reviews of prostaglandin RIA's see ref. 8 and 9). These include generation of PGE in vitro -- during frequent freeze-thawing (lo), prolonged storage (ll), clotting (12,13), and in the traumatization of skin or mechanical agitation or disruption of blood elements while obtaining or processing samples. Finally the practical value of PGE measurements in blood has been questioned since PGE is extensively degraded by the lung and liver, and thus it may not circulate in measurable levels.

The recent development by Levine (14) of a radioimmunoassay for the major circulating metabolite of PGE2, 15-keto-13,14- dihydro-PGE2 (EH2-PGE2), offers considerable promise of a tool which might obivate these problems. KHz-PGE2 has a longer half- 'ife in vivo than native PGE2 (8-10 minutes vs. ~1 minute; -- ef 15) and would be expected to accumulate in vivo during gen- -- ration and metabolism of PGE2; furthermore, it has been expected hat this metabolite would not be generated in vitro from blood -- lements during the obtaining of blood specimens or during pro- onged storage. Furthermore, recent studies have suggested that he levels of XH2-PGE2 measured by radioimmunoassay in normal umans (14,16) and in rabbits bearing the PGE2-producing VX2 tumor 17) are in reasonably good agreement with levels measured by ass spectrometry (18-20).

In the current paper we have applied the RIA for XH2-PGE2 ublished by Levine to examine its usefulness in human investiga- ion. Our results suggest that although measurement of XH2-PGE2 s considerably more sensitive than assays for PGE to reflect

JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

total body PGE production and degradation, there remains a number of limitations and artifacts. The purpose of this manuscript is to bring these problems to the attention of other investigators in the field.

METHODS A. Materials and methods

Unlabelled KH2-PGE2 was kindly supplied by Dr. John Pike of the Upjohn Company. Tritiated KH2-PGE2 (64-81 Ci/mmol; 98% pure by thin layer chromatography on silica gel) was purchased from Amersham Corp. (Arlington Heights, IL). Normal rabbit serum and goat anti-rabbit globulin for phase separation were purchased from Clinical Assays, Inc. (Cambridge, MA). The antiserum against KH2- PGE2and the details of the RIA were generously supplied by its developer Dr. Lawrence Levine (Brandeis University, Waltham, MA). The characteristics of this antiserum have been previously described (14). Undegraded total PGE was measured by RIA (1,21). The sodium salt of oleic acid and fatty-acid-free human serum albumin (HSA) were purchased from Sigma Chemicals (St. Louis, MO). The sodium oleate was bound to the HSA in a molar ratio of approximately 6:l by rapid stirring at 4O'C. Free fatty acids (FFA) were measured by a modification of the method of Dole (22); in studies where sodium salicylate had been infused, all samples were extracted twice with 0.02 N sulfuric acid to remove the interference by salicylate in the titration procedure observed by us and others (23-24).

B. Radioimmunoassay system

The RIA employed was basically that described by Levine (14) but with several minor modifications. These include the use of an antiserum dilution of 1:2000 and use of sufficient tracer to obtain a total of 8500 to 11,000 CPM per tube. Under such conditions, initial binding of tracer usually approximates 25%. The normal rabbXt serum was used in a dilution of 1:lOO. The goat anti-rabbit- gamma-globulin was preassayed to insure that the concentration used was in the region of antibody excess with respect to the gamma globulin carrier provided by the normal rabbit serum in the phase separation procedure. 500A of human plasma were added per tube, except in studies where dilution curves of plasma were performed. Standard curves were plotted on semilogarithmic paper after the values for non-specific precipitation of counts (i.e., in the absence of antibody against KH2-PGE2) in buffer and for each plasma source are subtracted to yield specific CPM. The CPM reported in the figures have been corrected for quenching caused by the NaOH us9d to dissolve the precipitate via a standard quench curve calibrated specifically for use in this system; efficiency of the scintillation counter is constant to within 1% and thus is not corrected for. In such a system, the use of a semilogarithmic scale results in an inverse linear relationship between the amount of unlabelled KH2-PGE2 added and the CPM (or B/Be) in the precipitate,

JUNE1979 VOL.17 NO.6

PROSTAGLANDINS

with a correlation coefficient of 0.97 to 1.00 over the range of

10 to 10,000 pg KIi2-PGE2 added per tube. Such a system is capable of detecting displacement of tracer by as little as 1 pg of KH2-

PGE2 which is statistically different from a repetitively analyzed BO value (no added standard); however, the region from l-10 pg per tube does not fall on the linear portion of standard curve. Therefore, the effective lower limit of detection in the assay is considered to be 10 pg per tube. Intra-assay variation in our laboratory has been 10-15X in the region of 100 to 250 pg/ml in plasma. In studies in which varying amounts of pure KHz-PGE2 standard (from 5 to 2500 pg) were added to human plasma pools, 100 ? 4% (2 + SE) of the added KHz-PGE2 was detected in the RIA.

C. Method of obtaining human blood samples

All samples for measurement of KHz-PGE2 were from plasma except where indicated to be serum. Plasma was obtained by placing 10 ml of human blood into pre-chilled test tubes contain-

ing EDTA. These samples were kept on ice and spun twice within 30 minutes (at 2500 rpm x 15 minutes at 4'C), taking care to avoid hemolysis, clotting or disturbing the buffy coat or platelets. The plasma obtained was then frozen at -10' to -2O'C and assayed without prior thawing within two weeks of obtaining samples.

From a total of 13 normal subjects (10 male, 3 female; ages 24-55 years), l-6 samples for determination of basal KH2PGE2 levels were obtained either by venipuncture or via indwelling venous catheter. Prostaglandin E2 (kindly provided by Dr. John Pike of the Upjohn Company) was infused in two normal subjects at a rate of O.O3ug/kg/min for 60 minutes and in one subject at 0.15 to 0.30pg/kg/min for a total of 75 minutes. Blood samples were obtained for an additional hour at the end of the infusions.

In the studies of KHz-PGE2 levels during stimulation or inhibition of lipolysis, samples were obtained from indwelling venous scalp-vein needles from normal fasting male volunteers. Two samples for determination of basal KHZ-PGE2 levels were obtained at 30 and 45 minutes after insertion of the needles. During the ensuing hour, subjects received infusions either of normal saline (n=6), sodium salicylate (40mg/min; n=6) or propranolol (5mg IV stat and 80 ugfmin; n=3). After this first hour all subjects received infusions of epinephrine (6ug/min) for 120 minutes by constant infusion. Sodium salicylate, where used, was discontinued after a total of 120 minutes of infusion since previous studies had shown that salicylate levels at this time

reached the therapeutic range of 25-30mg/dl and remained at that

level throughout the duration of the study. In studies using propranolol the infusion was continued through the two hour infu; sion of epinephrine as well as the 30 minute period after discontinuation of epinephrine.

JUNE 1979 VOL. 17 NO. 6 843

PROSTAGLANDINS

In studies designed to ascertain the effect of prolonged storage and/or repeated thawing on KH2-PGE2 levels, samples from earlier studies which had been frozen at -10“ to -2O'C (but previously thawed at least once) were selected for measurement. These samples had been collected in EDTA from normal, fasting humans.

In studies designed to ascertain directly the effects of various handling procedures on measured KH2-PGE levels, samples were subjected either to (a) vigorous mechanica z shaking by hand for one to two minutes and then heating in a 37°C water bath x 30'; (b) extensive osmotic hemolysis caused by adding 2 ml of distilled water to 5 ml of fresh human blood and inverting gently several times; or (c) clotting at room temperature for 20 to 90 minutes.

RESULTS

A. Normal "basal" values and values during PGE, infusion

Apparent basal KH2-PGE levels in the plasma from 13 normal volunteers were 57 + 25 (SD)pgjml. During infusion of PGE2 at a rate of O.O3ug/kg/min (Fig. la) for 60 minutes, increments of KH2-PGE2 levels from normal basal levels could be detected by 5 minutes and reached peak levels of 3500 and 2750 pg/ml by the end of the 60 minute infusion. After discontinuation of the PGE2 infusion, fZH2- PGE2 levels fell toward normal. In one of these studies, measurements of undegraded PGE did not show any rise (Fig. la). In a third study (Fig. lb) a 60 minute infusion of PGE2 at a higher rate of O.l5ug/kg/min led to increases in KH2-PGE2 from a basal level of 23 pg/ml to a level of 740 pg/ml within 5 minutes. A peak of 7500 pg/ml was reached at 45 minutes; increasing the rate of PGE2 infusion to 0.3ug/kg/min for a final 15 minutes led to a further increase of KH2-PGE2 level to 10,000 pg/ml. After discontinuation of the PGE2 infusion, KH2-PGE2 levels declined with an approximate half-life of 16 minutes. Plasma from one of the low dose infusions was tested in serial dilutions in the radioimmunoassay, which revealed (Fig. 2a) a curve nearly but not fully parallel to the standard curve (see Discussion). When pure standard was added to a plasma pool, dilution curve was obtained (Fig. 2b) that was nearly identical to the curve obtained from dilutions of a sample obtained during PGE2 infusion.

844 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDJNS

$0

MINUTES

Fig la: Response of plasma KHz-PGE z

in two patients and plasma PGE in one patient to a low- ose infusion of PGE2.

Fig lb: Response of plasma KHz-PGEZ in one patient to a high-dose infusion of PGEZ.

JUNE 1979 VOL. 17 NO.6 845

PROSTAGLANDINS

1000 -

600 -

200 ’ IO

1 I 100 1000 10.000

KHz-P% PP

PLASMA, pl

Fig 2a: Standard displacement curve of KH2-PGE in TRIS-EDTA buffer compared to multiple dilutions of plasma z rom a patient with a high plasma level of KH2-PGE2 achieved by an infusion of PGE2.

2600 - - 0”ffrr 1’.965.5 I f3738

I. 0.99

2200 - .---. PlDIrnO “‘-1320.9 x +4122.9

f.1.00

I800 -

z 0

1400-

.

100 1 1000 10.000

KH, -PGE,. PO

PLASMA, pl

Fig 2b: Standard displacement curve of KH -PGE in buffer compared to multiple dilutions of plasma to w g* zH2-PGE2 standard ich had been added.

846 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

B. Relationship between lipolysis and KH2-PGE2-immunoreactivity

During infusion of epinephrine into 6 subjects, there was a consi- stant increase in KH2-PGE -like

1 immunoreactivity (Fig. 3a), with le-

vels rising from basal va ues of 54 * 17 pg/ml (ji f SE) to a peak of 236 * 43 (p<.O2) at 60 minutes of infusion. However, such increments were not seen during simultaneous infusion of epinephrine and either of two antilipolytic substances, sodium salicylate (Fig. 3b) and propranolol (Fig. 3~). During sodium salicylate infusion,

300,

1 I 1 r 350 -30 0 30 WI0 30 60 90 I2OIO 30

MINUTES

Fig 3a: KH -PGE2 levels (solid lines) and free fatty acid (FFA) levels (das ed lines) during infusion of epinephrine and saline i (control).

Fig 3b: KH -PGE2 levels (solid lines) and free fatty acid (FFA) . levels (das ed lanes) during infusionsof epinephrine and sodium t

salicylate.

JUNE 1979 VOL. 17 NO. 6 847

PROSTAGLANDINS

EPINEPWRINE Sp9lmin

200 I PROPRANOLOL ,m9 IV AND 00~9h’”

1 1

Fig 3c: KHZ-PGE levels (solid lines) and free fatty acid (FFA) levels (dashed lines) during infusions of epinephrine and propranolol.

basal KH -PGE levels of 55 * 15 rose insignificantly to peak levels of 83 * 219 p&ml at 30 minutes of infusion of epinephrine (p=NS). Furthermore, in 3 subjects beta adrenergic blockade with propranolol totally abolished the rise in measured KHZ-PGE2 during epinephrine infusion (Fig. 3~). In one additional subject treatment with nicotinic acid (another anti-lipolytic agent) also obliterated any measured rise in KH -PGE2 and FFA levels during infusion of epinephrine (data not shown P. Finally, in one subject pretreated with indomethacin (100 mg/day for 2 days and 100 mg the morning of the study) apparent KH -PGE2 775

levels rose (basal: 68, peak 172 pg/ml) and PFA rose (basal: , peak 1545 pEq/L) during epinephrine infusion despite pharmaco-

logic inhibition of prostaglandin synthesis. Dilution of a plasma sample taken during the peak rise in apparent KH2-PGE2 level during epinephrine infusion yielded a displacement curve which was not parallel to the standard curve (Fig. 4b). There was a striking correlation between changes in free fatty acids and apparent KHZ-PGE2 in all studies (Fig. 3a-3c). For example, in the studies involving epinephrine, apparent KH -PGE levels correlated highly with FFA levels (y = 0.15 x -46.87; r2= 0.?'3, n=66, p<.OOl). Furthermore, the correlation analysis of basal KH2-PGE2 levels (y) with free fatty acids (x) in 29 basal samples yielded an equation of y = 0.07 x +7.42 (r = 0.51; n=29, p<.Ol), suggesting that approximately one quarter of the measured basal levelsas well could be ascribed to FFA in the samples. In order to directly assess the problem of cross-reaction with FFA's, dilutions of a solution of oleic acid bound to HSA were added to the

848 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

RIA. A displacement curve was obtained that was nearly identical to the standard curve (Fig. 4~). However, dilutions of FFA-free HSA (the conjugate for KHZ-PGE in the immunization procedure) alone, in buffer, also displaced ? or bound) tracer in the assay (data not shown), thus preventing accurate assessment of the cross-reaction of FFA's.

C. Effect of clotting

In 9 samples from subjects who had normal basal plasma KH - PGE values (Table 1) clotting at room temperature caused a hun red- fol2

2

increase of KH2-PGE placement curve using mu 1

to 5348 2 1179 pg/ml. However, the dis- tiple dilutions of such a sample was not

parallel to the standard curve (Fig. 4a) but resembled more those

performed on samples obtained during epinephrine-stimulated lipolysis (Fig. 4b) and on samples measured after prolonged storage (Fig. 4c; see below). The apparent rise in KH2-PGE2 during clotting was not due to prolonged exposure to room temperature since plasma levels of anticoagulated blood from 3 subjects left at room temperature for 90 minutes (Table 1) yielded values slightly lower than those of chilled plasma samples.

600 -

200 1 0 IO 100 IO00 10,000

KHz-PGE,. PI

PLASMA, /bl

Fig 4a: Standard displacement curve of KH2-PGE to serial dilutions of serum. 2

in buffer compared

JUNE 1979 VOL. 17 NO. 6 a49

TABLE 1

Effects of clotting, hemolysis and exposure to room temperature on apparent

KH2-PGE2 levels (pg/ml)

CHILLED

SAMPLE*

PLASMA

SERUM

PLASMA FROM BLOOD

INCUBATED 90" AT ROOM

HEMOLYSIS

TEMPERATURE

26

3,800

<20

10,000

<20

10,000

<20

8,000

101

1,840

100

5,600

73

2,996

84

4,820

-_-

1,080

<20;

___

<20

--_

<20°

--_

<20°

---

---

-_-

---

98

36+

62

<20+

45

---

___

* 500~1 of sample were used for each measurement in the RIA

0 mechanical hemolysis, followed by 37' water bath x 30 min.

+ osmotic hemolysis (2cc H20 to 5 ml blood)

PROSTAGLANDINS

I00 1 I I IO 100 IO00 10,000

KHz-PGEZ, ~9

PLASMA. pl

Fig 4b: Standard displacement curve of KHz-PGE2 in buffer compared to serial dilutions of plasma during llpolysis induced by epinephrine.

-Buffer p.964.0 I + 3655.3

r = 0.99

3000 - .---.P00kd woclls (1976.1977,

“..I6354 I + 5582.2

, =1.00 2600 -

a_--. Pooh* Plosmo (19781

“..1557.1 I +59w I

2200 - r * 0.99

z 0

I800 -

1000 - \

‘\

600 1 \ I IO 100 1000 10,000

KHz-PGE,, P9

PLASMA, /J”

Fig 4c: Standard displacement curve of KHz-PGE in buffer compared to serial dilutions of two pools of plasma whit ?I had been stored for prolonged periods.

JUNE 1979 VOL. 17 NO. 6 851

PROSTAGLANDINS

100 I 1 1 IO 100 ItlOO 10.000

K&-P% p9

OLEIC ACID-MA. &I

Fig 4d: Standard displacement curve of KH,-PGE, in buffer compared to serial dilutions bumin, in buffer.

of 4mM oleic acid $oundLto human serum al-

D. Effect of storage

Multiple samples from 4 intervals (range: 5 to 1063 immunoreactivity. There was

subjects which had been stored for varying days) were measured for KHz-PGE2 like no apparent difference in results whether

samples had been initially collected for prostaglandin determinations or less carefully for routine determination of hormones (e.g., insulin). In all subjects examined there were marked increases in KHz- -PGE2 immunoreactivity with increasing duration between collection date of the sample and measurement date (cf. Fig. 5). Two pooled samples (1976 to 1977, and 1978) were serially diluted and yielded a displacement curve not parallel to the standard curve (Fig. 4c), but similar in its increased slope to curves derived from serum or FFA-rich plasma (Fig. 4a-b).

E. Effect of hemolysis (Table 1)

Hemolysis, whether induced by vigorous shaking of the samples or by addition of distilled water, did not lead to any increase in measured KHz-PGE2 levels. The tendency for a decrease in some samples may be due in part to the effect of dilution due to the distilled water added.

852 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

0’ I I I 100 200 300

DAYS STORED

Fig 5: Linear regression analysis correlating apparent KH -PGE2 levels (y) with duration of storage of various plasma samp es ? (x) from one subject.

DISCUSSION

This RIA is sensitive enought to measure apparent basal KH2-PGE2 levels in human subjects. These levels (57 * 25 pg/ml) are not dis- similar from values (18,19) measured by mass spectrometry (30-33 pg/ml). The slightly greater values in our RIA are probably attributable to free fatty acids in the plasma (see Results and vide infra). This similarity between the two methods is consistent with the observations of Tashjian, et al. (17) and Seyberth, et al. (20) of a fairly good correlation between levels of KH -PGE in the VK2-bearing rabbit measured by these two independen 6 methods. Furthermore, this study confirms that there are marked advantages to the use of this radioimmunoassay for estimation of PGE infusion of a low dose of PGE2, t ere were increases of 35 to 125 fold i

production. In studies involving

in measured KH2-PGE2 levels despite absence of any systemic side effects or any detectable increment in native PGE levels measured by radioimmunoassay. Furthermore, a five-fold higher rate of infusion produced a rise of approximately 300-fold in KH2-PGE2 although we have previously shown that a comparable rate of infusion causes only 2-3 fold increments in immunoreactive PGE (25). These studies complement observations of Tashjian, et al. in the VK2 carcinoma-bearing rabbit, where there was only a 50% increase in immunoreactive PGE2 levels (3) at a time when KH -PGE levels had risen by 75-fold (17). Taken to- gether, these datz suggest that measurements of KH -PGE than a hundred times more sensitive than measuremegts $

may be more o native PGE as

an indicator of total body production of prostaglandin E.

JUNE 1979 VOL. 17 NO;6 853

PROSTAGLANDINS

However, we also found several important limitations to this method. Our data suggest that there persist certain artifacts created during handling of blood specimens; thus all difficulties seen with the measurement of undegraded PGE may not be completely obviated by the use of this radioimmunoassay for the major PGR2 metabolite. The first clue to these findings was the observation that high levels of KH2-PGE2 in plasma did not yield a displacement curve fully parallel to the standard curve in buffer. Although hemolysis and exposure to room temperature appeared not to generate any KH2-PGE2 in vitro, clotting produced an apparent marked increase of "KH -PGE 'I. Dilution of serum samples showed non-parallelism to KH -PGi bug rather a steeper slope strikingly similar to that seen wi h di *2 *1 ution of other samples which also contained apparent marked increases in KHZ-PGE2 --stored plasma or plasma containing high levels of FFA. It is likely that fragmentation of cellular elements of blood during the process of clotting releases considerable amounts of arachidonic acid from membrane.phospholipids which cross-reacts in the RIA. However, we cannot exclude the possibility that some true KH2-PGE2 is generated during clotting of human blood. Levine has previously reported that clotting of non-human blood samples had a mini- mal effect on measured KHZ-PGE2 levels (14). However, the same author also reported that formed elements in human blood appear to have at least the initial enzyme (15-hydroxy prostaglandin dehydrogenase) required for the metabolism of PGE (26,27). Furthermore, other studies have also confirmed that some isolated tissues have the ability not only to make PGE2 but also to convert in situ to its metabolites (28). We also recognize that our data do not exclude a contribution due to non-specific-interference in the double-antibody phase separation step or other "non-specific" effects of serum proteins due to the use of large volumes of plasma.

We also found that storage of plasma samples creates an artifact in apparent "KH2-PGE2" levels which increases linearly with time. The steepness and non-parallelism of the dilution curves of stored samples resembled that of clotted blood (serum) and samples with high free fatty acid content. FFA levels have been observed to rise with time in stored samples (29); thus this artifact may again be due to an interference caused by long-chain fatty acids, as was suggested by studies involving epinephrine infusion. This artifact appears to be related to duration of storage, not to methods of initially handling the samples or to the frequency of times that the samples had been subjected to freezing and thawing.

Although there are data to suggest that adrenergic agents stimulate prostaglandin synthesis (30), our findings of a rise in KH2-PGE -like immunoreactivity during infusions of epinephrine in man are pro ably i? also due to an artifact. Since three structurally different antilipolytic substances (sodium salicylate , propranolol and nicotinic acid) all block this rise, it is not likely that they are all acting as general inhibitors of prostaglandin synthesis- but rather that they

JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

are blocking increases of free fatty acids which interfere in the radioimmunoassay. The apparent displacement of tracer caused by addition of one fatty acid (oleic acid) in the RIA supports this formulation. This concept is also supported by the significant corre- lations between free fatty acids and KH2-PGE2 levels in the basal state and during epinephrine-induced lipolysis with or without con- comitant infusions of antilipolytic substances and even in the face of prostaglandin synthesis blockade with indomethacin. The correlation analyses between plasma free fatty acids and apparent "KH2-PGE2" levels suggest that the antiserum used cross-reacts extremely weakly with circulating free fatty acids (i.e., 5-10 uEq/L of FFA appear to be read as approximately 1 pg/ml in the RIA). However, this cross-reaction is an important problem because free fatty acids normally circulate in blood in molar concentrations greater than lo6 times those of KH -PGE and rise even further during lipolysis induced by stress, ca$echofamines, etc.

We cannot totally exclude the possibility that the three antilipolytic agents may act indirectly as prostaglandin synthesis inhibitors by blocking the release of fatty-acid prostaglandin precursors (e.g., arachidonic acid). There is little data measuring arachidonic acid and prostaglandin release during lipolysis. For several reasons one could question whether activation of hormone-sensitive lipase would substantially augment prostaglandin release as a direct consequence of triglyceride hydrolysis (although lipolytic agents such as epinephrine could augment arachidonic acid release and consequent prostaglandin synthesis by some other mechanism) (37). For example, it is known that there is very little arachidonic acid (the major precursor of PGE2) in triglyceride stores in adipose tissue (31) and that the rates of release of individual fatty acids are proportionate to their concentration in adipose tissue (32). Indeed during lipolysis the per- centage of arachidonic acid in blood varies independently of or inver- sely with that of shorter chain more saturated FFA's (33,34). Stimula- tion of lipolysis with norepinephrine does not increase (35), and inhibition of lipolysis with nicotinic acid does not reduce arachidonic acid levels (36). Furthermore, although KH2-PGE2 levels paralleled free fatty acid levels during the entire two hour epinephrine infusion (both showing a decline to basal levels by the end of the infusion) it is known that this decline of free fatty acids is not due to inhibition of lipolysis but rather to re-uptake of free fatty acids with subsequent re-esterification (38). Thus, during epinephrine infusion, lipolysis continues (at least in adipose tissue); however, there is no continued rise in "KHZ-PGE2". This suggests that lipolysis does not consistantly augment KH2-PGE levels detectably.. Finally, the correlation analysis of FFA vs. I+% *- 1 H PGE during stimulation of lipolysis (r2 = 0.53), and the non-paral el s ope of the displacement curve of FFA-rich samples compared to standard curves in buffer, suggest that much of the measured changes in the RIA are not due to changes in true KHZ-PGE2 levels. Nonetheless, it should be emphasazied that these

JUNE 1979 VOL. 17 NO. 6 855

PROSTAGLANDINS

studies do not exclude the possibility that epinephrine or stimulation of ‘lipolysis do stimulate prostaglandin synthesis in man. These hypotheses cannot be examined currently in vjvo due to the limitations of this assay. The addition of extraction and chromatography to this method may permit these difficulties to be overcome. In pre- liminary studies we have found that extraction of plasma samples in diethyl ether allows good recovery of KH FFA levels, suggesting that ether extrac t*

-PGE2 and markedly reduces ion may be helpful.

These observations have important ramifications since many agents which inhibit prostaglandin synthesis are also agents which affect lipolysis. For instance, in the current studies sodium salicylate, a known inhibitor of prostaglandin synthesis (39) and of lipolysis

(40), could have been interpreted, probably incorrectly, as blocking prostaglandin synthesis which had been stimulated during epinephrine infusion. Although definitive studies of the effect of indomethacin on lipolysis in vivo in humans have not been performed, it is reason- able to expect that any inhibitor of prostaglandin synthesis may also aiter circulating free fatty acid levels, since endogenous prostaglandins themselves appear to be potent modulators of lipolysis (41). Thus, interpretation of previous studies using RIA's for prostaglandins or PG metabolites before or during treatment with known PG- synthesis inhibitors require re-examination. in the context of the free fatty acid levels in the state being measured (i.e., starvation, stress). Even if extrrction and chromatography were employed, the assays still require verification that all fatty acids which interfere with the assay have been removed. Few reports of RIA's for prostaglandins have provided these controls, and some other RIA's for PG metabolites are currently performed on unextracted plasma (e.g. refs. 42,431.

In summary, it is clear that an RIA for KHz-PGE2 offers marked advantages as compared to RIA's for native PGE . However, it is also apparent that methods of collecting samples wi 1 need to be rigo- -1 rously standardized with care to avoid clotting. Assays should be run within one to two weeks after obtaining blood samples to minimize the storage-induced artifact. Studies should be performed to verify that apparent changes of KH2-PGE2 levels are not artifactual. Finally, it appears that extraction and probably chromatography will be required to avoid the significant problems of specificity induced by the interference caused by free fatty acids and albumin in this and perhaps other RIA's. Our data also suggest that other radioimmunoassays for native prostaglandins or PG metabolites which have not used extraction and chromatography and documented exclusion of free fatty acids by these procedures will require reevaluation in view of a possible un- recognized lack of specificity.

856 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

ACKNOWLEDGEMENTS

The authors are grateful to Dr. Lawrence Levine for his generous supply of antiserum, and for his help and suggestions in establishing this radioinununoassay in our laboratory.

We also gratefully acknowledge the excellent technical assistance of Ms. Kate Pfeifer, Ms. Barbara Williams, Ms. Susan Coates, Ms. Connie Holmes, the secretarial assistance of Ms. Terri Stevens and Ms. Anne Bartlett, and the nursing assistance of Ms. Martha Pleasant.

JUNE 1979 VOL. 17 NO. 6 857

PROSTAGLANDINS

REFERENCES

1. Robertson, R.P., D.J. Baylink, S.A. Metz, and K.B. Cummings. Plasma prostaglandin E in patients with cancer with and without

hypercalcemia. J. Clin. Endocrinol. Metab. 43:1330, 1976. -

2. Seyberth, H.W., G.V. Segre, J.L. Morgan, B.J. Sweetman, J.T. Potts, Jr. and J.A. Oates. Prostaglandins as mediators of hypercalcemia associated with certain types of cancer. N. Eng. J. Med.

293:1278, 1975. --

3. Voelkel, E.F., A.H. Tashjian, Jr., R. Franklin, E. Wasserman and L. Levine. Hypercalcemia and tumor-prostaglandins: the VK2 carcinoma model in the rabbit. Metabolism 24:973, 1975. -

4. Robertson, R.P. and M. Chen. A role for prostaglandin E in defective insulin secretion and carbohydrate intolerance in diabetes mellitus. J. Clin. Invest. 60~747, 1977. -

5. Fichman, M.P., N. Tefler, P. Zia, P. Speckart, M. Golub, and R. Rude. Role of prostaglandins in the pathogenesis of Bartter's syndrome. Am J. Med. 60:785, 1976. --

6. Gill, J.R., J.C. Frdlich, R.E. Bowden, A.A. Taylor, H.R. Keiser, H.W. Seyberth, J.A. Oates and F.C. Bartter. Bartter's syndrome: a disorder characterized by high urinary prostaglandins and a dependence of hyperreninemia on prostaglandin synthesis. Am. J. Med. 61:43, 1976. -

7. Jaffe, B.M. and S. Condon. Prostaglandin E and F in endocrine diarrheagenic syndromes. Ann. Surg. 184:516, 1976.

8. Silver, M.J., J.B. Smith, C.C. Ingerman, and J.J. Kocsis. Prostaglandins in blood: measurement, sources, and effects. In: Progress in Hematology (E.B. Brown, ed.), Grune and Stratton, New Yok, 1973, p. 235.

9. Granstrom, E. Radioimmunoassay of prostaglandins. Prosta- glandins 15:3, 1978. -

10. Waitzman, M.B. and M.L. Law. Changes in prostaglandin concentration in blood subjected to repetitive freezing and thawing. Prosta- glandins 10: 949, 1975. -

11. Jubiz, W., and J. Frailey. Prostaglandin E generation during storage of plasma samples. Prostaglandins 1:339, 1974.

12. Challis, J.R.G. and D. Tulchinsky. A comparison between the concentration of prostaglandin F in human plasma and serum. Prosta- glandins 5~27, 1974.

858 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

13. Silver, M.J., J.B. Smith, C. Ingerman, and J.J. Kocsis. Human blood prostaglandins: formation during clotting. Prostaglandins &:429, 1972.

14. Levine, L. Levels of 13,14-dihydro-15-keto-PGE2 in some biological fluids as measured by radioimmunoassay. Prostaglandins 14:1125, - 1977.

15. Hsmberg, M. and B. Samuelsson. On the metabolism of prostaglandin El and E2 in man. J. Biol. Chem. 246:6713, 1971.

16. Powles, T.J., R.C. Coombes, A.M. Neville, H.T. Ford, J.C. Gazet and L. Levine. 15-keto-13,14-dihydroprostaglandin E concentrations in serum of patients with breast cancer. Lancet 2~1 8, 5 1977.

17. Tashjian, A.H., Jr., E.F. Voelkel and L. Levine. concentrations of 13,14-dihydro-15-keto-prostaglandin bearing the VX2 carcinoma: effects of hydrocortisone Prostaglandins 14:309, 1977. -

18. Hubbard, W.C. and J.T. Watson. Determination of dihydro-metabolites of PGE, and PGF?, in plasma using

Plasma E2 in rabbits and indomethacin.

15-keto-13,14- high performance

liquid chromatography and gas chromafography-mass spectrometry. Prostaglandins 1_2:21, 1976.

19. Samuelsson, B. and K. G&en. Endogenous levels of 15-keto- dihydro-prostaglandins in human plasma. Parameters for monitoring prostaglandin synthesis. Biochem. Med. 11:298, 1974. -

20. Seyberth, H.W., W.C. Hubbard, 0. Oelz, B.J. Sweetman, J.T. Watson and J.A. Oates. Prostaglandin-mediated hypercalcemia in the VX2 carcinoma-bearing rabbit. Prostaglandins 14:319, 1977. -

21. Robertson, R.P. Differential in vivo pulmonary degradation of prostaglandin El, Bl, and Al. Am. J. Physiol. 228:68, 1975.

22. Dole, V.P. A relation between non-esterified fatty acids in plasma and the metabolism of glucose. J. Clin. Invest. 35:150, 1956. -

23. Carlson, L.A. and J. Sstman. Inhibition of the mobilization of free fatty acids from adipose tissue in diabetes. II. Effect of nicotinic acid and acetylsalicylate on blood glucose in human diabetics. Acta Med. Stand. 178~71, 1965.

24. Vik-Mo, H., and O.D. Mj&. Myocardial metabolism and performance during sodium salicylate infusion in dogs. Stand. J. clin. Lab. Invest. 36:763, 1976. -

JUNE 1979 VOL. 17 NO. 6 859

PROSTAGLANDINS

25. Robertson, R.P. and M. Chen. Modulation of insulin secretion

in normal and diabetic humans by prostaglandin E and sodium sali-

cylate. Trans. Assoc. Am. Phys. 90:353, 1977.

26. Kaplan, L., S-C. Lee, and L. Levine. Pattial purification and some properties of human erythrocyte prostaglandin 9-ketoreductase and 15-hyuroxy-prostaglandin dehydrogenase. Arch. Biochem. Biophys. 167~287, 1975.

27. Lee, S.-C. and L. Levine. Prostaglandin metabolism. II. Identification of two 15-hydroxyprostaglandin dehydrogenase types. J. Biol. Chem. 250:548, 1975.

28. Knapp, H.R., 0. Oelz, B.J. Sweetman, and J.A. Oates. Synthesis and metabolism of prostaglandin E2, F intestinal tract.

and D2 by the rat gastro- Stimulation by a h&ertonic environment in vitro.

Prostaglandins 15:751, 1978. -

29. Forbes, A.L. and J.A. Camlin. Effects of storage on serum non- esterified fatty acid concentrations. Proc. Sot. Exp. Biol. Med. 102:709, 1959.

30. Shaw, J.E. and P.W. Ramwell. Release of prostaglandin from rat epididymal fat pad on nervous and hormonal stimulation. J. Biol. Chem. 243:1498, 1968.

31. Hirsch, J. Fatty acid patterns in human adipose tissue. In: Handbook of Physiology, Vol. 1, Am. Physiol. Sot., Washington, D.C., 1972, D.F. Steiner and N. Freinkel, eds., p. 181.

32. Spitzer, J.J., H. Nakamura, M. Cold; H. Altschuler and M. Lieberson. Correlation between release of individual free fatty acids and fatty acid composition of adipose tissue. Proc. Sot. Exp. Biol. Med. 122:1276, 1966. -_

33. Hagenfeldt, L. and J. Wahren. Turnover of plasma-free arachidonic and oleic acids in resting and exercising human subjects. Metabolism 24:799, 1975. -

34. Hagenfeldt, L. The concentrations of individual free fatty acids in human plasma and their interrelationships. Arkiv For Kemi 29:57, 1968. -

35. Meinertz, H., and V.P. Dole. Changes in the composition of plasma non-esterified fatty acids after insulin, glucose and norepinephrine. Fed. Proc. 19:148, 1960. -

36. Hagenfeldt, L., J. Wahren, B. Pernow and L. R'af. Uptake of individual free fatty acids by skeletal muscle and liver in man. J. Clin. Invest. 51:2324, 1972. -

860 JUNE 1979 VOL. 17 NO. 6

PROSTAGLANDINS

37. Christ, E.J. and D. H. Nugteren. The biosynthesis and possible function of prostaglandins in adipose tissue. Biochim. Biophys. Acta 218:296, 1970.

38. Steinberg, D. Catecholamine stimulation of fat mobilization and its metabolic consequences. Pharmacol. Rev. 18:217, 1966. -

39. Samuelsson, B. Quantitative aspects on prostaglandin synthesis in man. Adv. Biosci. 9~7, 1973.

40. Bizzi, A., A.M. Codegoni, and S. Garattini. Salicylate, a powerful inhibitor of free fatty acid release. Nature 204:1205, 1964.

41. Paoletti, R., and L. Puglisi. Lipid metabolism. In: The Prostaglandins, Vol. 1 (P.W. Ramwell, ed.), Plenum Press, New York, 1973, p. 317.

42. Granstrom, E., and H. Kindahl. Radioimmunoassays for prostaglandin metabolites. In: Advances in Prostaglandin and Thromboxane Research, Vol. 1 (B. Samuelsson and R. Paoletti, eds.), Raven Press, New York, 1976, p. 81.

43. Levine, L., and R.M. Gutierrez-Cernosek. Levels of 13,14- dihydro-15-keto-PGF20 in biological fluids as measured by radioimmunoassay. Prostaglandins 2~785, 1973.

JUNE 1979 VOL. 17 NO. 6 861

applications and limitations of measurement of 15-keto,13,14-dihydro prostaglandin e2 in human blood...

Documents