- Email: [email protected]
The effect of physiological levels of non-esterified fatty acids on the radioimmunoassay of prostaglandins
PROSTAGLANDINS
THE EFFECT OF PHYSIOLOGICALLEVELS OF NON-ESTERIFIEDFATTY ACIDS ON THE RADIOIMMUNOASSAYOF PROSTAGLANDINS Edward W. Gold and Paul R. Edgar Orthopedic Research Laboratory Ohio State University College of Medicine Columbus Ohio 43210
Abstract Non-esterifiedfatty acids (NEFA) can significantlyinterferewith the radioimnunoassayof PGE and PGF2olusingcommerciallyavailable anti-sera. PGBI antigen-antibodybinding is 50% inhibited by 110 pg of PGB 48 ng of PGE , 3.5.w of PGF2a(,or 9.Oug linoleic, 14wg arachidonic,21'ug '6-1' ino1eic, 4Oug palmitoleicor 45 wg oleic acids. PGF2q antigenantibody binding is 50% inhibited by 270 pg of PGF2*, 70 ng of PGEI, or 4.2&g arachidonic,14 ug ‘6-linolenic,22 ug linoleic, 70&g palmitoleic or 11Oug oleic acids. Physiologicallevels of NEFA, such as the quantities found in small volumes of plasma, are sufficient to prohibit accurate prostaglandinmeasurements. Chromatographyon small columns of silicic acid proved to be an effective technique for separation of NEFA and prostaglandin from lipid extracts, however, the results of this study suggest that the interferenceproduced by the presence of NEFA in the measurement of prostaglandin from certain physiologicalfluids may be avoided if the prostaglandins are not extracted prior to radioimmunoassay. Introduction Radioimnunoassayis the most sensitive technique for measuring prostaglandins. The procedure typically involves extracting the lipids from the tissue being examined using appropriate solvents, followed perhaps by a wash of the extract with petroleum ether to remove neutral lipids. The high degree of specificityof the antibodies used could allow radioimmunoassayto be carried out on this fraction without further treatment. The measurement of prostaglandinin certain tissues, however, may present additional difficulties because of the presence of relatively large amounts of non-esterifiedfatty acids (NEFA). Studies on the specificityof the prostaglandinanti-sera have focused on the cross reactivitiesof other prostaglandinsin the antigen-antibody reactions. The specificityhas clearly been shown to be sufficient to allow measurement of a particular prostaglandinin the presence of others unless the quantity of the competing prostaglandinis quite large (1,2). The effects of fatty acids on the antigen-antibodyreactions might be regarded as negligible since the anti-sera are many orders of magnitude more sensitive to a
DECEMBER 1978 VOL. 16 NO. 6
945
PROSTAGLANDINS
specific prostaglandin than to any non-prostaglandin fatty acid. I t should be noted, however, that physiological levels of NEFAare many orders of magnitude higher than corresponding levels of any prostaglandin. In this report fatty acids were examinedfor their a b i l i t y to interact with commerciMly available anti-sera to PGB1 (used in the radioimmunoassay of PGE1) and PBF20~. The aim of this investigation was to determine i f physiological )evels of NEFA (e.g. levels in quantities of plasma suitable for prostaglan~in measurementsby radioimmunoassay)were sufficient to interfere withaccurate PGE and PGF2=
946
DECEMBER 1978 VOL. 16 NO. 6
PROSTAGLANDINS
I00
Z
" ~lliOQlt~lUi6~lll1'OR
Fig. 1 Inhibition of 3H-PGB1- anti-PGB1 binding by prostaglandins and fatty acids. Commercially available anti-serum to PGB1 was tested in the presence of the indicated amounts of PGBI ( A ) , PGE1 ( I I ) , PGF2~ ( 0 ) , and linoleic (Z~), arachidonic ( n ) , ~r-linolenic ( 0 ) , palmitoleic ( X ) , oleic (V) and palmitic ()K) acids. Radioimmunoassays were carried out according to standard procedures. Results The ability of various fatty acids to inhibit the reaction between PGB1 and antibody to PGB~as a function of fatty acid concentration is shown in Fig. 1. Linoleic acia was the best inhibitor tested followed in decreasing order by arachidonic, ~ - l i n o l e n i c , palmitoleic and oleic acids. The saturated fatty acids palmitic and stearic, produced negligible inhibition. The PGBI anti-serum exhibited substantial specificity. I t was nearly 82,000 times more sensitive to PGB1 than to linoleic acid, the most effective inhibitor of the fatty acids tested. Fig. 1 also shows the high degree of specificity of the PGB] anti-serum toward other prostaglandins. More than 400 times greater PGEI levels and almost 32,000 times greater PGF2~ levels were required to achleve the equivalent inhibition produced by PGBI. Similar results were obtained when the fatty acids were examined for their ability to inhibit the reaction between PGF~ and its anti-serum. As shown in Fig. 2, the best inhibitor was arachidonlc acid followed by -linolenic, linoleic, palmitoleic and oleic acids. Palmitic and stearic acids again produced negligible inhibition at these concentration ranges. The PGF2~, anti-serum did not exhibit the high degree of specificity toward fatty acids that was observed with the PGBI anti-serum. However, i t was found to be more than 15,500 times more sensitive to PGF2~ than to arachidonic acid, the most effective fatty acid inhibitor. The PGF2~ anti-serum was also not quite as specific toward other prostaglandins, as shown by the requirement for only 260 times more PGE1 to achieve the equivalent inhibition produced by PGF2~k.
DECEMBER 1978 VOL. 16 NO. 6
947
PROSTAGLANDINS
I00
Ol 0.01
o.,
,.o
.
,~'
~,~
,o"
~'
NANOGRADm INHlalTOR
Fig. 2 Inhibition of 3H-PGFp~ - anti-PGFp~ binding by prostaglandins and fatty acids. CommerciallyavaTlable anti-serum to PGF~x was tested in the presence of the indicated amounts of PGFp~ I O ) , PGEI (m), and arachidonic (17), ~-linolenic ( 0 ) , liBoleic ( A ) , palmitoleic ( x ) , oleic ( V ) and palmitic (~C) acids. Radioimmunoassays were carried out according to standard procedures. The amount of individual fatty acid or prostaglandin required to produce 50% inhibition of the PGBI and the PGF2~ antigen-antibody reactions is shown in Table I. I t is evident that for both antibodies the most potent inhibitors were those that most resembled the homologous prostaglandin. Sufficient cross reaction was being produced by fatty acids to warrant further investigation of the effect of physiological levels of NEFA (e.g. normal plasma levels) on the radioimmunoassay of prostaglandins. The total NEFA level in human plasma is 0.45 - 0.90 meq/L (4). Levels in other animals are probably comparable, although rabbit plasma contains slightly higher amounts (5). The composition of the human NEFA pool has been determined by several investigators (6,7) and is similar, with a few exceptions, to the NEFA composition of the plasma of other animals (8). In a l l cases, the f a t t y acid precursors of prostaglandin constitute only a small fraction of the total NEFA pool. I t is therefore possible to calculate the levels of the principal NEFA in normal human plasma (Table I I ) . I t was apparent from Figs. I and 2 and Table I I that the inhibition expected from the NEFA of plasma, even in relatively small volumes, could significantly affect the radioimmunoassay of both PGEI and PGF2~ . The actual inhibition produced by the fatty acids expected in 0,I ml of plasma, i n d i v i dually and collectively, is shown in Table I I I . Oleic and linoleic are the fatty acids which account for most of the interference. In order to obtain accurate measurements of prostaglandin in samples of plasma or in tissues containing comparable levels of NEFA, removal of NEFA contaminants from tissue extracts prior to radioimmunoassay would be necessary. The results of using s i l i c i c acid ~olumn chromatography for such a separation are shown in Table IV. ~ith both ~H-labeled and unlabeled f a t t y acids, i t was observed that most of the applied oleic, linoleic and arachidonic acids were
948
DECEMBER 1978 VOL. 16 NO. 6
PROSTAGLANDINS
TABLE I Inhibition of 3H-PGBI and 3H-PGF2w antigen-antibody binding by prostaglandins and fatty acids. The quantities (picograms) required to produce 50% inhibition were obtained from the curves in Fig. 1 and Fig. I.
~
Anti-PGB 1
Anti-PGF~ec
"110
- - -
4.8 X 104
7.0 X 104
3.5 x 106
270
1.4 x 107
4.2 x 106
"6-Linolenic Acid (18:3)
2,2 x 107
1.4 x 107
Linoleic Acid
9.0 x 106
2.2 x 107
Oleic Acid
4.5 x 107
1,1 x 108
4.0 x 107
7,0 x 107
PGB1
~
PGEI
o
WO"
PGF2ec
,
~
,
~
J
~
,
v Arachidonic Acid (20:4)
e
~
yvv
~~__~__~,/coo~
(1B:2)
(18:1)
Palmitoleic Acid
(16:1)
~ e e o ~
eluted from s i l i c i c acid columns in the ether wash. The non-physiological PGB1was also found in this fraction. PGE1 and PGF2~ were effectively recovered in the ethyl acetate/methanol washes. I t was possi61e to obtain PGEI and PGF2ot fractions with relatively l i t t l e contamination by interfering free fatty acids, but the separation of PGE1and PGF2~ from each other was limited.
DECEMBER 1978 VOL. 16 NO. 6
949
PROSTAGLANDINS
TABLE I I The distribution of non-esterified fatty acids in normal humanplasma (see text). Proportion of total plasma NEFApool (%)
Fatty acid 18:101eic 16:0 Palmitic 18:2 Linoleic 18:0 Stearic 16:1Palmitoleic 20:3 Prostaglandin 20:4 precursors
C o n c e n t r a t i orange n ~g/ml
37 25 15 11 5
47 29 19 14 6
1
-
94 58 13 28 11
1.4 - 2.7
TABLE I l l Inhibition of prostaglandin antigen-antibody binding by the fatty acids contained in 0.1 ml of normal humanplasma, The low end of the concentration ranges of each fatty acid in Table I I was used to determine the amount of fatty acid to be tested. These quantities, in ~g, appear in parenthesis. Eachfatty acid was tested in duplicate. The expected levels of inhibitions were calculated from the data in Fig. 1 and Fig. 2. The fatty acids were also tested in combination. % Inhibition PGB1 Fatt~ acid
PGF2o~
observed
expected
observed
expected
11.2 2.7 12.6 0 2.4 1.2 17.5
10 0 12 0 0 0 22
11.5 2.2 7.0 0 0 0 17.2
12 0 11 0 0 0 23
Oleic (5.0) Palmitic (3.0) Linoleic (2.0) Stearic (1.5) Palmitoleic (02)(016) Arachidonic Combined
TABLE IV S i l i c i c acid chromatographyof prostaglandins (PGB1, PGEI and PGF2~) and fatty acids. The chromatographywas carried out as described in Methods. Fraction
PGBI%
Ether Ethyl acetate Ethyl acetate: Methanol (49:1) Ethyl acetate: Methanol (42:8)
96.9 3.1 0
0.9 8.1 86.1
1.3 2.9 38.7
74.5 12.6 9.0
74.8 3.1 12,0
77.9 8.5 5.6
0
1.9
52,1
4,0
10.1
8.0
950
PGEl PG~ counts applieo
Oleic Linoleic Arachidonic "--'-~'-fattx-~recovered
DECEMBER 1978 VOL. 16 NO. 6
PROSTAGLANDINS
Discussion The quantities of PBGI, PGE1 and PGF2~ required to produce 50% inhibition of the H-PGB1 antigen-antiBody reaction is in accord with the values previously reported by Levine et al. (1). In the present study the amount of PGBI which would be required-fo-r-5O% inhibition was 110 pg. This quantity of PGEI (converted to PGBI before measurement) will be contained in volumes of plasma much greater than 0.1 ml according to currently accepted plasma PGEI levels. Yet such a quantity of plasma contains sufficient NEFA, principally oleic and linoleic acids, to seriously interfere with accurate PGE% measurements (Table I I I ) . The antibody to PGF2~ was found to be slightly more sensitive to prostaglandins than had be~n reported (1). The quantity of PGF2~ required for 50% inhibition of the H-PGF2~ antigen-antibody reaction was 270 pg. Again, a volume of 0.1 ml of plasma would contain substantially less than that amount of PGF2~ according to published accounts. Therefore, considerably more plasma would be needed to provide enough PGF2o( for radioimmunoassay measurements, but the NEFA in even 0.1 ml of plasma is sufficient (Table I I I ) to produce significant inhibition of PGF2~ determinations. Consequently, plasma NEFAwill affect the radioimmunoassay of PGF2~ even more adversely than they will affect the estimations of PGEI, The relevance of the observed results to measurements of plasma prostaglandin levels are limited, in that primary prostaglandins in plasma are not regarded as important as their 13,14-dihydro-15-keto metabolites. However, i t was the aim of this investigation to establish the possible interference of NEFA in the radioimmunoassay of PGE1 and PGF~ in any tissue which contains relatively high levels of NEFA. Plasma was useB as an example. If one were to attempt to measure the PGE1 or PGF2~ in a small volume of plasma which contained the proper amount of prostaglandin to fall in the range of sensitivity of radioimunoassay, enough NEFAwould be present to prohibit accurate measurements. Therefore, a step must be included in such protocols to remove NEFA contaminants from extracts prior to measurement. This is especially important in subjects with a lipid imbalance, in which circulating fatty acid levels are elevated, such as following prolonged corticosteroid administration (5,9) or in various forms of hyperlipidemia. I t should also be noted that in cell culture systems, the addition of exogenous prostaglandin precursors must be carefully controlled in order to avoid errors in prostaglandin determinations, unless a separation is incorporated into the procedure before measurements are made. Unfortunately there are no simple techniques available for separation of prostaglandins from other fatty acids. Theuse of s i l i c i c acid columns, as described in this report, is a satisfactory method. However, this system may require modification since the incomplete separation of PGEI and PGF.2~ could s t i l l lead to problems in measuring one in the presence of very hlgh concentrations of the other. In any case, i f a large number of samples are to be processed, this method could become cumbersome. Despite these d i f f i c u l t i e s , however, such efforts appear to be unavoidable.
DECEMBER 1978 VOL. 16 NO. 6
951
PROSTAGLANDINS
I t was previously reported (5) that 14-20 per cent of the material estimated as PGE in l i p i d extracts of plasma could be washed off s i l i c i c acid columns with ether. The remainder could be eluted with 2% methanol in ethyl acetate. This distribution is approximately what would be expected according to the results of chromatography of fatty acids (ether fraction) and prostaglandins (ethyl acetate/methanol fraction) as described in this report. The results of this investigation support the suggestion that the extraction of prostaglandin from tissues such as plasma may not be the best procedure to follow in their measurement. In unextracted plasma, albumin w i l l bind both NEFA and prostaglandins almost completely. In the presence of an antibody to a prostaglandin, the prostaglandin w i l l be.9omp]etely dissociated f~om ~he weakly binding albumin (Ka approximately I0~-i0 IU M"1 vs Ka less than 10b M-I). However, the prostaglandin antibody w i l l not compete e f f ~ t i v e l y with albumin for the NEFA. Therefore, the presence of albumin in the unextracted samples may provide an efficient protection from the influence of even high levels of NEFA. In extracted samples, where albumin has been removed, the influence of NEFAw i l l be much greater as described in this report, Extraction may therefore not only be unnecessary but may actually be detrimental for prostaglandin estimations in certain physiological fluids. Several different lots of commercially available anti-sera to both PGE1 and PGF2~ were examined in this study. For each group of antibodies the response to fatty acids was consistent. Preliminary investigations using anti-serum specific for 13,14 dihydro-15-keto PGF2o( (Clinical Assays Inc.) have also indicated a sensitivity to NEFA similar to that found for PGB] and PGF2~. I t is anticipated, therefore, that anti-sera from other sources and for other prostaglandins w i l l also be affected by physiological levels of NEFA. References I.
Levine, L., Gutierrez-Cernosek, R.M., and Van Vunakis, H., Specificities of Prostaglandins BI, FI~ , and F2K Antigen-Antibody Reactions. J. Biol. Chem. 246___:6782,1971. 2. Levine, L., Antibodies to Pharmacologically Active Molecules:Specificities and SomeApplications of Antiprostaglandins. Pharmacol. Rev. 25:293, 1973: 3. Smith, S.W., A New Salting-Out Technique for Colorimetric Fre~Fatty Acid Assays. Anal. Biochem. 67:531, 1975. 4 . Henry, R.J., Cannon, D.C., and Winkelman, J.W., Clinical Chemistry, Harper and Row, Hagerstown, Maryland, 1974. 5. Gold, E.W., Fox, O.D., and Edgar, P.R., The Effect of Long Term Corticosteroid Administration on Lipid and Prostaglandin Levels. J. Steroid Biochem. 4:303, 1978. 6. Hagenfeldt, L. and Webster, P.O., Plasma Levels of Individual Free Fatty Acids in Patients with Acute Myocardial Infarction. Acta Med. Scand. 194:357, 1973. 7. Rogiers, V., Gas-Liquid Chromatography Method for the Determina~n of Non-Esterified Fatty Acids in Blood Plasma in Children. Clin. Chim. Acta 78:227, 1977. 8. Leat, W.M.F., Fatty Acid Compositions of the Plasma Lipids of Newborn and Maternal Ruminants. Biochem. J. 98:598, 1966. 9. Gold, E.W., Fox, O.D., Weissfeld,~. and Curtiss, P.H., CorticosteroidInduced Avascular Necrosis: An Experimental Study in Rabbits. Clin. Orthop. 135:295, 1978. Received
952
8/7/78
- Approved
9/27/78
DECEMBER 1978 VOL. 16 NO. 6
THE EFFECT OF PHYSIOLOGICALLEVELS OF NON-ESTERIFIEDFATTY ACIDS ON THE RADIOIMMUNOASSAYOF PROSTAGLANDINS Edward W. Gold and Paul R. Edgar Orthopedic Research Laboratory Ohio State University College of Medicine Columbus Ohio 43210
Abstract Non-esterifiedfatty acids (NEFA) can significantlyinterferewith the radioimnunoassayof PGE and PGF2olusingcommerciallyavailable anti-sera. PGBI antigen-antibodybinding is 50% inhibited by 110 pg of PGB 48 ng of PGE , 3.5.w of PGF2a(,or 9.Oug linoleic, 14wg arachidonic,21'ug '6-1' ino1eic, 4Oug palmitoleicor 45 wg oleic acids. PGF2q antigenantibody binding is 50% inhibited by 270 pg of PGF2*, 70 ng of PGEI, or 4.2&g arachidonic,14 ug ‘6-linolenic,22 ug linoleic, 70&g palmitoleic or 11Oug oleic acids. Physiologicallevels of NEFA, such as the quantities found in small volumes of plasma, are sufficient to prohibit accurate prostaglandinmeasurements. Chromatographyon small columns of silicic acid proved to be an effective technique for separation of NEFA and prostaglandin from lipid extracts, however, the results of this study suggest that the interferenceproduced by the presence of NEFA in the measurement of prostaglandin from certain physiologicalfluids may be avoided if the prostaglandins are not extracted prior to radioimmunoassay. Introduction Radioimnunoassayis the most sensitive technique for measuring prostaglandins. The procedure typically involves extracting the lipids from the tissue being examined using appropriate solvents, followed perhaps by a wash of the extract with petroleum ether to remove neutral lipids. The high degree of specificityof the antibodies used could allow radioimmunoassayto be carried out on this fraction without further treatment. The measurement of prostaglandinin certain tissues, however, may present additional difficulties because of the presence of relatively large amounts of non-esterifiedfatty acids (NEFA). Studies on the specificityof the prostaglandinanti-sera have focused on the cross reactivitiesof other prostaglandinsin the antigen-antibody reactions. The specificityhas clearly been shown to be sufficient to allow measurement of a particular prostaglandinin the presence of others unless the quantity of the competing prostaglandinis quite large (1,2). The effects of fatty acids on the antigen-antibodyreactions might be regarded as negligible since the anti-sera are many orders of magnitude more sensitive to a
DECEMBER 1978 VOL. 16 NO. 6
945
PROSTAGLANDINS
specific prostaglandin than to any non-prostaglandin fatty acid. I t should be noted, however, that physiological levels of NEFAare many orders of magnitude higher than corresponding levels of any prostaglandin. In this report fatty acids were examinedfor their a b i l i t y to interact with commerciMly available anti-sera to PGB1 (used in the radioimmunoassay of PGE1) and PBF20~. The aim of this investigation was to determine i f physiological )evels of NEFA (e.g. levels in quantities of plasma suitable for prostaglan~in measurementsby radioimmunoassay)were sufficient to interfere withaccurate PGE and PGF2=
946
DECEMBER 1978 VOL. 16 NO. 6
PROSTAGLANDINS
I00
Z
" ~lliOQlt~lUi6~lll1'OR
Fig. 1 Inhibition of 3H-PGB1- anti-PGB1 binding by prostaglandins and fatty acids. Commercially available anti-serum to PGB1 was tested in the presence of the indicated amounts of PGBI ( A ) , PGE1 ( I I ) , PGF2~ ( 0 ) , and linoleic (Z~), arachidonic ( n ) , ~r-linolenic ( 0 ) , palmitoleic ( X ) , oleic (V) and palmitic ()K) acids. Radioimmunoassays were carried out according to standard procedures. Results The ability of various fatty acids to inhibit the reaction between PGB1 and antibody to PGB~as a function of fatty acid concentration is shown in Fig. 1. Linoleic acia was the best inhibitor tested followed in decreasing order by arachidonic, ~ - l i n o l e n i c , palmitoleic and oleic acids. The saturated fatty acids palmitic and stearic, produced negligible inhibition. The PGBI anti-serum exhibited substantial specificity. I t was nearly 82,000 times more sensitive to PGB1 than to linoleic acid, the most effective inhibitor of the fatty acids tested. Fig. 1 also shows the high degree of specificity of the PGB] anti-serum toward other prostaglandins. More than 400 times greater PGEI levels and almost 32,000 times greater PGF2~ levels were required to achleve the equivalent inhibition produced by PGBI. Similar results were obtained when the fatty acids were examined for their ability to inhibit the reaction between PGF~ and its anti-serum. As shown in Fig. 2, the best inhibitor was arachidonlc acid followed by -linolenic, linoleic, palmitoleic and oleic acids. Palmitic and stearic acids again produced negligible inhibition at these concentration ranges. The PGF2~, anti-serum did not exhibit the high degree of specificity toward fatty acids that was observed with the PGBI anti-serum. However, i t was found to be more than 15,500 times more sensitive to PGF2~ than to arachidonic acid, the most effective fatty acid inhibitor. The PGF2~ anti-serum was also not quite as specific toward other prostaglandins, as shown by the requirement for only 260 times more PGE1 to achieve the equivalent inhibition produced by PGF2~k.
DECEMBER 1978 VOL. 16 NO. 6
947
PROSTAGLANDINS
I00
Ol 0.01
o.,
,.o
.
,~'
~,~
,o"
~'
NANOGRADm INHlalTOR
Fig. 2 Inhibition of 3H-PGFp~ - anti-PGFp~ binding by prostaglandins and fatty acids. CommerciallyavaTlable anti-serum to PGF~x was tested in the presence of the indicated amounts of PGFp~ I O ) , PGEI (m), and arachidonic (17), ~-linolenic ( 0 ) , liBoleic ( A ) , palmitoleic ( x ) , oleic ( V ) and palmitic (~C) acids. Radioimmunoassays were carried out according to standard procedures. The amount of individual fatty acid or prostaglandin required to produce 50% inhibition of the PGBI and the PGF2~ antigen-antibody reactions is shown in Table I. I t is evident that for both antibodies the most potent inhibitors were those that most resembled the homologous prostaglandin. Sufficient cross reaction was being produced by fatty acids to warrant further investigation of the effect of physiological levels of NEFA (e.g. normal plasma levels) on the radioimmunoassay of prostaglandins. The total NEFA level in human plasma is 0.45 - 0.90 meq/L (4). Levels in other animals are probably comparable, although rabbit plasma contains slightly higher amounts (5). The composition of the human NEFA pool has been determined by several investigators (6,7) and is similar, with a few exceptions, to the NEFA composition of the plasma of other animals (8). In a l l cases, the f a t t y acid precursors of prostaglandin constitute only a small fraction of the total NEFA pool. I t is therefore possible to calculate the levels of the principal NEFA in normal human plasma (Table I I ) . I t was apparent from Figs. I and 2 and Table I I that the inhibition expected from the NEFA of plasma, even in relatively small volumes, could significantly affect the radioimmunoassay of both PGEI and PGF2~ . The actual inhibition produced by the fatty acids expected in 0,I ml of plasma, i n d i v i dually and collectively, is shown in Table I I I . Oleic and linoleic are the fatty acids which account for most of the interference. In order to obtain accurate measurements of prostaglandin in samples of plasma or in tissues containing comparable levels of NEFA, removal of NEFA contaminants from tissue extracts prior to radioimmunoassay would be necessary. The results of using s i l i c i c acid ~olumn chromatography for such a separation are shown in Table IV. ~ith both ~H-labeled and unlabeled f a t t y acids, i t was observed that most of the applied oleic, linoleic and arachidonic acids were
948
DECEMBER 1978 VOL. 16 NO. 6
PROSTAGLANDINS
TABLE I Inhibition of 3H-PGBI and 3H-PGF2w antigen-antibody binding by prostaglandins and fatty acids. The quantities (picograms) required to produce 50% inhibition were obtained from the curves in Fig. 1 and Fig. I.
~
Anti-PGB 1
Anti-PGF~ec
"110
- - -
4.8 X 104
7.0 X 104
3.5 x 106
270
1.4 x 107
4.2 x 106
"6-Linolenic Acid (18:3)
2,2 x 107
1.4 x 107
Linoleic Acid
9.0 x 106
2.2 x 107
Oleic Acid
4.5 x 107
1,1 x 108
4.0 x 107
7,0 x 107
PGB1
~
PGEI
o
WO"
PGF2ec
,
~
,
~
J
~
,
v Arachidonic Acid (20:4)
e
~
yvv
~~__~__~,/coo~
(1B:2)
(18:1)
Palmitoleic Acid
(16:1)
~ e e o ~
eluted from s i l i c i c acid columns in the ether wash. The non-physiological PGB1was also found in this fraction. PGE1 and PGF2~ were effectively recovered in the ethyl acetate/methanol washes. I t was possi61e to obtain PGEI and PGF2ot fractions with relatively l i t t l e contamination by interfering free fatty acids, but the separation of PGE1and PGF2~ from each other was limited.
DECEMBER 1978 VOL. 16 NO. 6
949
PROSTAGLANDINS
TABLE I I The distribution of non-esterified fatty acids in normal humanplasma (see text). Proportion of total plasma NEFApool (%)
Fatty acid 18:101eic 16:0 Palmitic 18:2 Linoleic 18:0 Stearic 16:1Palmitoleic 20:3 Prostaglandin 20:4 precursors
C o n c e n t r a t i orange n ~g/ml
37 25 15 11 5
47 29 19 14 6
1
-
94 58 13 28 11
1.4 - 2.7
TABLE I l l Inhibition of prostaglandin antigen-antibody binding by the fatty acids contained in 0.1 ml of normal humanplasma, The low end of the concentration ranges of each fatty acid in Table I I was used to determine the amount of fatty acid to be tested. These quantities, in ~g, appear in parenthesis. Eachfatty acid was tested in duplicate. The expected levels of inhibitions were calculated from the data in Fig. 1 and Fig. 2. The fatty acids were also tested in combination. % Inhibition PGB1 Fatt~ acid
PGF2o~
observed
expected
observed
expected
11.2 2.7 12.6 0 2.4 1.2 17.5
10 0 12 0 0 0 22
11.5 2.2 7.0 0 0 0 17.2
12 0 11 0 0 0 23
Oleic (5.0) Palmitic (3.0) Linoleic (2.0) Stearic (1.5) Palmitoleic (02)(016) Arachidonic Combined
TABLE IV S i l i c i c acid chromatographyof prostaglandins (PGB1, PGEI and PGF2~) and fatty acids. The chromatographywas carried out as described in Methods. Fraction
PGBI%
Ether Ethyl acetate Ethyl acetate: Methanol (49:1) Ethyl acetate: Methanol (42:8)
96.9 3.1 0
0.9 8.1 86.1
1.3 2.9 38.7
74.5 12.6 9.0
74.8 3.1 12,0
77.9 8.5 5.6
0
1.9
52,1
4,0
10.1
8.0
950
PGEl PG~ counts applieo
Oleic Linoleic Arachidonic "--'-~'-fattx-~recovered
DECEMBER 1978 VOL. 16 NO. 6
PROSTAGLANDINS
Discussion The quantities of PBGI, PGE1 and PGF2~ required to produce 50% inhibition of the H-PGB1 antigen-antiBody reaction is in accord with the values previously reported by Levine et al. (1). In the present study the amount of PGBI which would be required-fo-r-5O% inhibition was 110 pg. This quantity of PGEI (converted to PGBI before measurement) will be contained in volumes of plasma much greater than 0.1 ml according to currently accepted plasma PGEI levels. Yet such a quantity of plasma contains sufficient NEFA, principally oleic and linoleic acids, to seriously interfere with accurate PGE% measurements (Table I I I ) . The antibody to PGF2~ was found to be slightly more sensitive to prostaglandins than had be~n reported (1). The quantity of PGF2~ required for 50% inhibition of the H-PGF2~ antigen-antibody reaction was 270 pg. Again, a volume of 0.1 ml of plasma would contain substantially less than that amount of PGF2~ according to published accounts. Therefore, considerably more plasma would be needed to provide enough PGF2o( for radioimmunoassay measurements, but the NEFA in even 0.1 ml of plasma is sufficient (Table I I I ) to produce significant inhibition of PGF2~ determinations. Consequently, plasma NEFAwill affect the radioimmunoassay of PGF2~ even more adversely than they will affect the estimations of PGEI, The relevance of the observed results to measurements of plasma prostaglandin levels are limited, in that primary prostaglandins in plasma are not regarded as important as their 13,14-dihydro-15-keto metabolites. However, i t was the aim of this investigation to establish the possible interference of NEFA in the radioimmunoassay of PGE1 and PGF~ in any tissue which contains relatively high levels of NEFA. Plasma was useB as an example. If one were to attempt to measure the PGE1 or PGF2~ in a small volume of plasma which contained the proper amount of prostaglandin to fall in the range of sensitivity of radioimunoassay, enough NEFAwould be present to prohibit accurate measurements. Therefore, a step must be included in such protocols to remove NEFA contaminants from extracts prior to measurement. This is especially important in subjects with a lipid imbalance, in which circulating fatty acid levels are elevated, such as following prolonged corticosteroid administration (5,9) or in various forms of hyperlipidemia. I t should also be noted that in cell culture systems, the addition of exogenous prostaglandin precursors must be carefully controlled in order to avoid errors in prostaglandin determinations, unless a separation is incorporated into the procedure before measurements are made. Unfortunately there are no simple techniques available for separation of prostaglandins from other fatty acids. Theuse of s i l i c i c acid columns, as described in this report, is a satisfactory method. However, this system may require modification since the incomplete separation of PGEI and PGF.2~ could s t i l l lead to problems in measuring one in the presence of very hlgh concentrations of the other. In any case, i f a large number of samples are to be processed, this method could become cumbersome. Despite these d i f f i c u l t i e s , however, such efforts appear to be unavoidable.
DECEMBER 1978 VOL. 16 NO. 6
951
PROSTAGLANDINS
I t was previously reported (5) that 14-20 per cent of the material estimated as PGE in l i p i d extracts of plasma could be washed off s i l i c i c acid columns with ether. The remainder could be eluted with 2% methanol in ethyl acetate. This distribution is approximately what would be expected according to the results of chromatography of fatty acids (ether fraction) and prostaglandins (ethyl acetate/methanol fraction) as described in this report. The results of this investigation support the suggestion that the extraction of prostaglandin from tissues such as plasma may not be the best procedure to follow in their measurement. In unextracted plasma, albumin w i l l bind both NEFA and prostaglandins almost completely. In the presence of an antibody to a prostaglandin, the prostaglandin w i l l be.9omp]etely dissociated f~om ~he weakly binding albumin (Ka approximately I0~-i0 IU M"1 vs Ka less than 10b M-I). However, the prostaglandin antibody w i l l not compete e f f ~ t i v e l y with albumin for the NEFA. Therefore, the presence of albumin in the unextracted samples may provide an efficient protection from the influence of even high levels of NEFA. In extracted samples, where albumin has been removed, the influence of NEFAw i l l be much greater as described in this report, Extraction may therefore not only be unnecessary but may actually be detrimental for prostaglandin estimations in certain physiological fluids. Several different lots of commercially available anti-sera to both PGE1 and PGF2~ were examined in this study. For each group of antibodies the response to fatty acids was consistent. Preliminary investigations using anti-serum specific for 13,14 dihydro-15-keto PGF2o( (Clinical Assays Inc.) have also indicated a sensitivity to NEFA similar to that found for PGB] and PGF2~. I t is anticipated, therefore, that anti-sera from other sources and for other prostaglandins w i l l also be affected by physiological levels of NEFA. References I.
Levine, L., Gutierrez-Cernosek, R.M., and Van Vunakis, H., Specificities of Prostaglandins BI, FI~ , and F2K Antigen-Antibody Reactions. J. Biol. Chem. 246___:6782,1971. 2. Levine, L., Antibodies to Pharmacologically Active Molecules:Specificities and SomeApplications of Antiprostaglandins. Pharmacol. Rev. 25:293, 1973: 3. Smith, S.W., A New Salting-Out Technique for Colorimetric Fre~Fatty Acid Assays. Anal. Biochem. 67:531, 1975. 4 . Henry, R.J., Cannon, D.C., and Winkelman, J.W., Clinical Chemistry, Harper and Row, Hagerstown, Maryland, 1974. 5. Gold, E.W., Fox, O.D., and Edgar, P.R., The Effect of Long Term Corticosteroid Administration on Lipid and Prostaglandin Levels. J. Steroid Biochem. 4:303, 1978. 6. Hagenfeldt, L. and Webster, P.O., Plasma Levels of Individual Free Fatty Acids in Patients with Acute Myocardial Infarction. Acta Med. Scand. 194:357, 1973. 7. Rogiers, V., Gas-Liquid Chromatography Method for the Determina~n of Non-Esterified Fatty Acids in Blood Plasma in Children. Clin. Chim. Acta 78:227, 1977. 8. Leat, W.M.F., Fatty Acid Compositions of the Plasma Lipids of Newborn and Maternal Ruminants. Biochem. J. 98:598, 1966. 9. Gold, E.W., Fox, O.D., Weissfeld,~. and Curtiss, P.H., CorticosteroidInduced Avascular Necrosis: An Experimental Study in Rabbits. Clin. Orthop. 135:295, 1978. Received
952
8/7/78
- Approved
9/27/78
DECEMBER 1978 VOL. 16 NO. 6