Fast HPLC Determination of Serum Free Fatty Acids in the Picomole Range. Michael. P#{252}ttmann,1'3 Harald. Krug,2. Elke von Ochsenstein,' and Reinhard.
CLIN. CHEM. 39/5, 825-832
Fast HPLC Michael
(1993)
Determination
P#{252}ttmann,1’3
Harald
of Serum
Krug,2
Elke
von
Free Fatty Acids in the Picomole Ochsenstein,’
We developed a method for determining individual free fatty acids in serum by using a modified one-step Dole extraction, derivatization, and a new high-performance
liquid chromatographic (HPLC) separation. Sample handling is minimized to a single transfer of the fatty acids (upper layer of the Dole extract), which are readily derivatized at 85 #{176}C with p-bromophenacyl bromide without significant hydrolysis of esterifled fatty acids. The derivatization mixture is directly injected into the HPLC apparatus. The new method, which uses C6 (3-zm particle) column material and an isocratic acetonitrile-water eluent, separates nearly to baseline 12 of the physiologically most abundant long-chain fatty acids (C12-C) in 99%” (Sigma, Deisenhofen, acetonitrile for HPLC (Fisons, Loughborough, UK), glass Reacti-vials (1.5 and 10 mL) with screwcaps and Teflon-lined septa, and magnetic bars for the derivatization of fatty acids, Chromabond NH2#{174} and fiberglass discs for lipid separations (Macherey und Nagel, D#{252}ren, Germany) were purchased as indicated. Other solvents, inorganic acids, and laboratory chemicals were “pro analysi” quality and Extrelut#{174} bulk material from Merck (Darmstadt, Germany). All glassware was rinsed with chloroform/methanol (1/1 by vol). For solvents used during Dole’s extraction, plastic pipette tips (Ratiolab, Dreieich, Germany) could be used after having being rinsed several times with solvent; we exactly weighed the pipetted volume. In every series we included two blank runs (water instead of serum) to check for contaminations. Blood from 38 apparently healthy, fasting volunteers (ages 20-55 years) was drawn into serum Monovettes#{174} (Sarstedt, N#{252}mbrecht, Germany), which were shown to be free of FFA contamination. The sera used for FFA determinations were stored at -18 #{176}C, usually for sev-
Germany),
eral weeks, the analysis.
and
centrifuged
after
thawing
just
before
partides, 25 x 4 mm, -140 000 plates/m (Phase Separation, Deeside, UK), used with a CN-5pm guard column (4 x 4 mm) at a column temperature of 30#{176}C.The mobile phase was acetonitrile/water (77/23 by vol), the flow rate 1.3 mJ.Imin (back pressure -220 kPa). Absorbance of the eluate was monitored at 254 nm and the signal processed by the integrator. Peaks were identified via retention times in relation to the internal standard heptadecanoic acid. We calculated micromolar concentrations of fatty acids by the internal standard method, using peak areas corrected for different detector responses by calibration runs with a mixture of fatty acid standards (Figure la). Extractions. We performed the extraction method of Bligh and Dyer essentially as described (20), extracting 100 p1 of serum with 375 iL of methanol/chloroform (2/1 by vol, with 2 g/L added formic acid), 125 p1 of water, and 125 p1 of chloroform. The resulting aqueous phase was shaken twice with 1 mL of chloroform, all organic layers were combined, and finally the solvent was removed under a stream of nitrogen. Extrelut extractions were performed as indicated by Ikeda et al. (21): 100 p1 of serum mixed with 900 p1 of a)
Final Method for Serum FFA Dole’s extraction (modified). Mix 100 L of serum with 20 pL of heptadecanoic (margaric) acid in isopropanol (5 mxnol/L) as internal standard (included also in the added fatty acid mixture in recovery experiments) in a 1.5-mL Reacti-vial. Add 500 pL of the modified Dole’s mixture-isopropanolln-heptane/phosphoric acid (2 mol/ L), 40/10/1 (by vol)-and mix again. After 5-10 miii of incubation at room temperature add 200 L of n-heptane and 300 j.L of water, thoroughly vortex-mix, and centrifuge for 5 miii at 1000 x g. Transfer 200 pL of the upper organic layer (i.e., 62.5% of the total 320 L) to another Reacti-vial and remove the solvent under a stream of nitrogen. Derivatization of FFA Reconstitute the lipid extract with 6 L of chromophore-catalyst reagent (50 mmol/L p-bromophenacylbromide and 5 mmol/L 18-crown-6 in acetonitrile, stored protected from light), 500 j.tL of acetonitrile, and -1 mg of KHCO3. Insert a magnetic bar, top the liquid with a few milliliters of nitrogen gas, close the vial tightly (use the Teflon-lined septa), and heat to 85#{176}C for 45 mm with vigorous stirring. Weigh the vials before and after heating to check for leakage. After cooling the samples, remove the magnetic bars, centrifuge the reaction mixture, and inject 10-25 p1 of the clear supernate into the HPLC apparatus. The amount of reagent used is sufficient for FFA concentrations up to 3.5 mmol/L; the given volumes should be increased slightly (60
>60
96-100
>90a
40-80
No
Contaminations,
imoI/L
100 p1 of sample is extracted initially and the derivatized FFA are concentrated before injection.
in the ranges
ence
With Comparison
with
Other
Methods
There is a good correlation between the TLC/GC method (13,15,27) and our HPLC method for determining total serum FFA concentrations, as shown by the analysis of sera from 10 volunteers (Figure 2a). This is especially true for the individual saturated and oligounsaturated fatty acids (summarized in Figure 2, b and c) and even for ilnolenic acid (linear regression y = 1.014x + 0.98, r 0.968, n = 10; data not separately =
shown).
The
about
4-8%
lower
values
for
total
FFA
measured with the caused by the much
GC method are at least partially lower findings for the polyunsatuarachidonic acid and docosahexaenoic
rated fatty acids, acid (no significant correlation, Figure 2d). This discrepancy was already observed by Cordis et al. (27) and is probably due to the isolation of FFA during TLC, leading to losses of polyunsaturated fatty acids by oxidation. Unfortunately, the “one-step” GC method by Lepage and Roy (15), which circumvents the TLC step, was not useful in our hands because it yielded two- to threefold higher total serum FFA contents than the three methods above. Figure 2e shows the correlation between the enzymatic and the HPLC method for total serum FFA concentrations measured in the sera of 38 volunteers. The slight difference of 2-5% lower values for the normal range by the HPLC method might be caused in part by the still incomplete extraction of linoleic and other polyunsaturated fatty acids by the modified Dole method (Table 2). The percentage ranges for the individual fatty acids determined with our method fit well
Table 5. Percentage
of Total Serum Human Subjects Mien
HPLC (n = 38)
FFA
12:0
1.5±
14:0
3.1
14:1
0.3
16:0 16:1 18:0
26.7 3.8± 10.2±
18:1 c/s 18:1 trans
±
±
FFA
SD, % TLC/GC (n = 10)
0.8
1.8±
1.2
3.7
±
0.8
0.4
±
0.3
± 0.2
In Heafthy
0.9
%
‘13’
n.r. 25-32 4-6
(Table
5 and
refer-
new
our
method,
HPLC
only
100 p1
of serum
and basic HPLC equipment are necessary for a fast and reliable determination of total and individual serum FFA. The method correlates well with two other generally accepted methods and is suitable for routine analysis in series of 10-20 samples, e.g., to determine precursor fatty acids, arachidonic and docosahexaenoic
It is also transferable to matrices other than e.g., for the investigation of acyl chain selectivity of lipolytic enzymes in serum (28) or thrombocytes acid.
serum,
(Preuss, P#{252}ttmann, Patscheke, unpublished). the FFA are not destroyed, fraction collection for further investigation, e.g., by GC, mass try, or scintillation counting. We thank
Birgit
Waidele
for excellent
technical
Because
is possible spectrome-
assistance.
References 1. Mead JF,
AIfm-Slater RB, Howton DR., Popjak G. Lipidschemistry, biochemistry and nutrition, let ed. New York: Plenum Press, 1986:273-83. 2. Thiele OW. Lipide, Isoprenoide mit Steroiden, 2nd ed. Stuttgart: Georg Thieme Verlag, 1982:369-70. 3. Spector AA. Fatty acid binding to plasma albumin. J Lipid Roe 1975;16:165-79. 4. Peters T. Serum albumin [Review]. Mv Protein Chem 1985;37: 161-245. 5. Moulder C, Shouten JA, Popp-Snijders C. Determination of free fatty acids: a comparative study of the enzymatic versus the gas chromatographic and colorimetric method. J Clin Chem Clin Biochem 1983;21:823-7. 6. Christie WW. Lipid analysis-isolation, separation, identification and structural analysis of lipids, 2nd ed. Oxford: Pergamon Press, 1982:99. 7. Itaya K. A more sensitive and stable colorimetric determination of free fatty acids in the blood. J Lipid Res 1977;18:663-5. 8. Dole VP, Meinertz H. Microdetermination of long-chain fatty acids in plasma and tissues. J Biol Chem 1960;235:2595-9. 9. Shimizu 5, Tani Y, Yamada H, Tabata M, Murachi T. Enzymatic determination of serum-free fatty acids: a colorimetric method. Anal Biochem 1980;107:193-8. 10. Degen MM, Van der Vies J. Enzymatic microdetermination of free fatty acids in plasma of animgla using paraoxon to prevent lipolysis. Scand J Clin Lab Invest 1985;45:283-5. 11. Chilliard Y, Bauchart D, Barnouin J. Determination of plasma non-esterifled fatty acids in herbivores and man: a comparison of values obtained by manual or automatic chromato-
graphic,
n.r.
by others
titrimetric,
and enzymatic
colorimetric
Nutr Devel 1984;24:469-82. 12. Shukla VKS. Recent chromatography of lipids
advances [Review].
13. Christie
analysis-isolation,
WW. Lipid
methods.
Reprod
in the high performance liquid Prog Lipid Res 1988;27:5-38.
3.5 1.8 2.9
29.4± 3.8± 10.6±
31.6±
4.8
1.9±
1.3
30.8 ± 1.6 1.3± 0.5
18:2
15.4±
4.2
12.9±
18:3
1.4± 1.9±
1.1 0.5
1.1 ± 0.8
1-3
J Chromatogr
0.9
0.4
1
0.6
0.2
±
0.2
n.r. n.r.
15. Lepage G, Roy C. Specific methylation of plasma nonesterifled fatty acids in a one-step reaction. J Lipid Res 1988;29:227-35. 16. Hoffmann GE, Neumann U. Modified photometric method for determination of phospholipase A activities. Klin Wochenschr
289
552
±
188
n.r.
20:4 20:5
±
n.f.
22:6 Total,
imol/L
0.8
±
made
512
±
3.0 ±
n.f.
up from five different references therein). Extreme values listed In parentheses. n.r., not reported; n.j., not found. Range
1.5 1.6 2.8
Rang.,
determined
14).
(2)
8-13
cation Press,
and structural 1982:51,
analysis
separation,
of lipids,
2nd ed. Oxford:
identifiPergamon
63.
14. Van derhigh Horstperformance FA, Post Automated
n.r. 10-19
(ref. 14 and references
tion
of plasma
MI!, Holthuis JJ, Brinkman liquid chromatographic determina- UA. acids using on-line derivatization with based on micellar phase transfer catalysis.
free fatty
9-bromomethylacridin
1990;500:443-52.
1989;67:106-9. cited
17. Imamura enzymatic
in serum 18.
S, Hirayama T, Arai method using 1,2-diglyceride
[Abstract].
Schalkwijk
CG,
T, Takoa K, Misaki H. An for pancreatic lipase test
Cliii Chem 1989;35:1126. MSrki
CUNICAL
F, Van
den
Bosch
H. Studies
CHEMISTRY, Vol. 39, No.
5, 1993
on the
831
acyl-chain
selectivity of cellular 1990;1044:139-46. 19. Jensen RG. Detection and erol hydrolase) activity from
phospholipases
A2. Biochim
Bio-
phys Acta
determination various
sources.
of lipase Lipids
(acylglycl983;18:
650.-7. 20. Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol 1959;37:911-7. 21. Ikeda M, Shimada K, Sakaguchi T. High performance liquid chromatographic determination of free fatty acids with l-naphthylamine. J Chromatogr 1983,272:251-9. 22. Kaluzny MA, Duncan LA, Merrit MV, Eppa DE. Rapid separation of lipid classes in high yield and purity using bonded phase columns. J Lipid Roe 1985;26:135-40. 23. Korte K, Chien KR., Casey ML. Separation and quantitation of fatty acids by high-performance liquid chromatography. J Chromatogr 1986;375:225-3l. 24. Thirst liD, Milano M, Kikta EJ, Connelly SA, Grushka E. Phena#{231}ylesters of fatty acids via ether catalysts for enhanced ultraviolet detection in liquid chromatography. Anal Chem 1975;
Measurement of polyunsaturated by high performance liquid
fatty acids chromatography.
of myocsrdial lipids Chromatographia
1987;24:687-90. 28.
Puttmann
M, Aufenanger
van Ackern K, Harenberg pase A activities in sera specificity but no acyl-chain
J, von Ochsenstein E, Durholt 5, J, Hoffmann GE. Increased phospholiof intensive-care patients show sn-2 selectivity. Clin Chem 1993;39:782-8.
25.
29. Jork H, Funk W, Fischer W, Winuner H. DSnnschichtchromatographie; Reagenzien und Nachweismethoden, Vol. la. Physikaliache mid chemiache Nachweismethoden, let ed. Weinheim: Verlag Chemie, 1989:342-6. 30. L.epage G, Roy CC. Direct transesterification of all classes of lipids in a one-step reaction. J Lipid Roe 1986;27:114-20. 31. Van der Vusse GJ, Roemen THM, Reneman RS. The content of non-esterifled fatty acids in rat myocardial tissue. A comparison between the Dole and the Foich extraction procedures. J Mol Cell Cardiol 1985;17:527-31. 32. Duncombe WG, Rising TJ. Quantitative extraction and determination of noneaterifled fatty acids. J Lipid Roe 1973;14:258-.61. 33. Iannou PV, Golding BT. Cardiolipins: their chemistry and
phase
biochemistry
47:l797-.801. Pei PTS, Kossa WC, Ramachandran S. High pressure reverse liquid chromatography of fatty acidp-bromophenacyl esters. Lipids 1976;ll:814-6. 26. Wood R, Lee T. High performance liquid chromatography of fatty acids: quantitative analysis of saturated, monoenoic polyenoic and geometrical isomers. J Chromatogr 1983;254:237-46. 27. Cordis GA, Prasad MR., Otani H, Engelman RM, Kas DK
832
CUNICAL CHEMISTRY, Vol. 39, No.5,1993
[Review].
Prog Lipid Roe 1979;17:279-318.
Kukais A, Myher JJ. Gas chromatographic analysis of plasma lipids [Review]. Mv Chromatogr 1989;28:267-332. 35. Roemen THM, Keizer H, Van der Vusse GJ. Quantitative assessment of y-linolemc acid in human blood and plasma with capillary gas chromatography. J Chroinatogr 1990;528:447-62. 34.
