Lipids

, 43:171 | Cite as

High-Throughput Analysis of Plasma Fatty Acid Methyl Esters Employing Robotic Transesterification and Fast Gas Chromatography

Methods

Abstract

Fatty acid analysis is an important research tool, and indices derived from essential fatty acid contents serve as useful biomarkers related to cardiovascular and other chronic disease risk. Both clinical and basic studies of essential fatty acid composition are becoming ever larger in magnitude leading to delays while the rather laborious lipid analyses are performed. A robotic transesterification procedure has been developed for high-throughput analysis of plasma fatty acid methyl esters. In this approach, robots perform most steps including plasma and reagent transfer, transesterification reaction via heating at 80 °C in open tubes with multiple reagent additions, followed by two-phase extraction and transfer of lipid extracts to GC vials. The vials are then placed directly onto a GC autosampler carousel for robotic sample injection. An improved fast GC method is presented in which the peaks of interest are eluted within 6 min. This method is readily scalable to prepare and analyze 200 samples per day (1,000 samples per week) so that large clinical trials can be accommodated.

Keywords

Robotic transesterification Plasma fatty acid methyl esters Fast gas chromatographic High throughput sample preparation High throughput gas chromatography Essential fatty acids 

Introduction

Fatty acids are important components of biological membranes and their measurement is useful from a variety of biological perspectives. The chain length, degree of unsaturation and other structural features of fatty acids are one determinant of the physico-chemical properties of biological membranes [1, 2]. Essential fatty acids are the focus of many compositional [3, 4, 5], physiological [6, 7, 8], behavioral [9, 10, 11, 12, 13, 14] and metabolic experiments [15, 16, 17, 18, 19]. The n-3 polyunsaturates eicosapentaenoic acid (20:5n-3, EPA) and docosahexaenoic acid (22:6n-3, DHA) have been inversely associated with several disease states in epidemiological studies [20, 21, 22, 23]. They antagonize the n-6 polyunsaturated arachidonic acid (20:4n-6, AA) and the eicosanoids that are produced from it [24, 25]. Moreover, the blood stream contents of these n-3 and n-6 polyunsaturates provide an important biomarker of cardiovascular disease risk [21, 22, 24]. EPA and DHA intake and tissue content have been related to sudden cardiac death and arrhythmias [26, 27, 28, 29]. These observations have led to many clinical experiments and more recently large clinical trials (e.g., GISSI study [30], primary prevention CVD [31], AREDS2 [32], DINO infant [33], NINDS Alzheimer’s [34]), where the effects of increased EPA and/or DHA intake are examined in various pathological states.

Thus the analysis of fatty acid profiles in tissues, particularly blood components, has become increasingly important as endpoints in clinical trials, as biomarkers in cross-sectional studies and as measures of compliance. Such trials may generate very large numbers of samples for analysis necessitating quick, efficient and inexpensive analytical methods for fatty acid analysis. The standard methods for such analysis typically involve total lipid extraction followed by a laborious transmethylation procedure, methyl ester extraction, concentration and gas chromatographic analysis. It is typical for this labor-intensive methodology to consume 2 days of a skilled chemist’s efforts for 10–20 samples. As a result, it has not been generally feasible to study thousands of subjects where each subject in a clinical trial can generate many samples. Clearly, a high throughput method for fatty acid analysis is needed.

In previous work from our laboratory, a simplified chemical procedure was developed for lipid transmethylation [35] that was based upon the Lepage and Roy procedure [36]. A method for fast gas chromatography of the typical profile of fatty acid methyl esters found in biological samples was also introduced [35, 37]. This paper presents a complete method for robotic transmethylation and analysis of plasma lipids. Plasma is directly reacted without prior lipid extraction, as in the Lepage and Roy method, and a robot carries out the reaction, methyl ester extraction and sample concentration, leaving the sample in a GC autosampler vial ready for injection. A faster method for GC analysis is also presented here. Together, these methods are efficient and cost effective. They may potentially increase laboratory productivity by an order of magnitude. These procedures may be used as the basis for a high-throughput method suitable for analysis of large populations such as are found in clinical trials.

Materials and Methods

Reagents and Plasma Samples

Acetyl chloride, 2[6]-di-tert-butyl-p-cresol (butylated hydroxytoluene, BHT), analytical grades methanol, hexane and toluene, and K2CO3 were purchased from Sigma-Aldrich Chemical Co. (St Louis, MO, USA). BHT was added to methanol (50 μg BHT/sample) to prevent fatty acid oxidation. The internal fatty acid standards (23:0 methyl ester, and 22:3n-3) and external (GLC-462) were purchased from Nu-Chek Prep (Elysian, MN, USA). Internal standard solutions were prepared by dissolving them, respectively, in the methanol BHT solution at a concentration of 250 μg/ml. Blood was collected by venipuncture into a heparinized tube and was immediately centrifuged for 5 min at 2,000×g. The resulting plasma was then aliquoted in batches of 5 ml, frozen and stored at −80 °C. The 13 mm × 100 mm Pyrex disposable culture tubes for transesterification reactions were purchased from PGC Scientific (Frederick, MD, USA).

For the robotic transesterification procedure, Stock A and Stock B reagents were prepared in bulk. The 2.4 ml of Stock Solution A required initially for the reaction of 100–200 μl aliquots of plasma included 1.9 ml of methanol, 100 μl of acetyl chloride, 0.3 ml of toluene, and 100 μl each of the two internal standard solutions (containing 25 μg of 23:0 and 20 μg of 22:3n-3 as ethyl esters). Each 900 μl volume of Stock B solution was comprised of 740 μl of methanol, 40 μl of acetyl chloride solution, and 120 μl of toluene.

Robotic Apparatus

A 1.5-m long Freedom EVO robot equipped with a liquid handling arm (LH-arm), a multi-channel 1-ml syringe pipetting system with an integrated liquid detection system and liquid detection disposable tips, a robotic manipulator arm, a fast wash pump and a syringe wash station, low disposable tips ejector option, and a nitrogen manipulation station was purchased from Tecan US Inc. (Research Triangle Park, NC, USA). The working space on the Tecan deck also includes other peripherals such as tip rack with a waste tip disposal station, twelve 13-mm test tube carrier racks each with a capacity to hold up to 16 plasma vials along the y-axis, three troughs for holding reagents, a reaction block, and a heated GC vial sample block. The layout of the robotic deck is illustrated in Fig. 1. The Tecan EVOware™ scripting programming language is used to control the robotic arm and other modules. The temperature regulated GC vial sample block and the heating-cooling reaction block were custom made by J-Kem Scientific Inc. (St Louis, MO, USA) and are under computer control. The reaction block is composed of a hot plate connected to a chiller through a cryogenic solenoid valve which regulates the flow of cooling liquid. The GC vial sample plate and reaction block hot plate are an array containing 16 × 12 holes to hold a maximum of 192 GC vials (13-mm diameter) and 192 reaction tubes of 13 × 100 mm dimension, respectively. Both the reaction block and GC vial block are controlled by a digital temperature controller either via a local module or by a program driver that was incorporated into the software of the Tecan programming command set running under Windows. The robot was encased in a hood fabricated by Airline Hydraulics (Bensalem, PA, USA) and is connected at the top to an exhaust outlet and is well ventilated.
Fig. 1

Diagrammatic representation of the layout of the robotic deck

Instrumentation

Fast GC Analyses

Analyses were performed on an Agilent 6890N Network Gas Chromatograph (Agilent Technologies, Palo Alto, CA, USA) equipped with a split/splitless injector, a 7683 automatic liquid sampler and flame ionization detector. The GC was also equipped with a 208-V power supply to enable fast temperature ramping. The column used was a DB-FFAP of 15 m × 0.1 mm i.d. × 0.1 μm film thickness (J&W Scientific from Agilent Technologies, Palo Alto, CA, USA). Instrument control and data collection was performed by a GC Chemstation, Rev. B.01.01 (Agilent Technologies, Palo Alto, CA, USA). Temperature program was as follows: initial conditions, 150 °C with 0.25 min hold; ramp 35 °C/min to 200 °C, 7 °C/min to 225 °C with a 3.2 min hold and then 80 °C/min to 245 °C with 2.75 min hold. Instrumental conditions were as follows: carrier gas was H2 at a flow rate of 56 cm/s and a constant head pressure of 344.7 kPa; FID detector set at 250 °C; air and N2 make-up gas flow rates of 450 and 10 ml/min; respectively, with a split ratio of 100:1; sampling frequency of 50 Hz; autosampler injections of 2 μl volume. Run time for a single sample was 11 min with all the fatty acid peaks of interest eluting within 8 min of the injection time. The sample injection-to-injection time was 16 min including a re-equilibration time of 1.50 min. The liner used was split with a cup containing no glass wool.

For the development of a faster GC method, a constant flow mode of 0.6 ml/min was used using the same column as above. Temperature program was as follows: initial conditions, 160 °C with 0.10 min hold; ramp 60 °C/min to 220 °C with a 0.5 min hold, 80 °C/min to 175 °C, 70 °C/min to 230 °C with 1.82 min hold, 70 °C/min to 220 °C with 0.96 min hold, and final ramping at 60 °C/min to 245 °C with 2.40 min hold. Instrumental conditions were as follows: carrier gas was H2 at a starting flow rate of 58 cm/s and a starting head pressure of 344.7 kPa; FID detector set at 250 °C; air and N2 make-up gas flow rates of 450 and 10 ml/min, respectively; split ratio of 100:1; sampling frequency of 50 Hz; autosampler injections of 2 μl volume. Run time for a single sample was 8.69 min with all the fatty acid peaks of interest eluting within 5.8 min of the injection time. The sample injection-to-injection time was 12 min, including a re-equilibration time of 1.50 min. The data was quantified according to a method previously described [35].

Transesterification Methods

The standard Lepage and Roy transesterification method [36] was used as the reference point for comparisons to samples generated by the robotic method. All manipulations were performed under a nitrogen atmosphere. Briefly, 200 μl of plasma and 100 μl each of the two internal standard solutions (providing 25 μg of 23:0 and 20 μg of 22:3n-3 methyl ester) were added to 13 mm × 100 mm borosilicate or Pyrex screw-capped glass tubes, followed by an addition of 2 ml of a methanol:hexane (4:1 v/v) mixture. Samples were vortexed and the tubes were placed on ice, and then acetyl chloride (200 μl) was added drop-wise while swirling the tubes. The tubes were tightly capped under nitrogen, and heated at 100 °C for 60 min. Afterwards, the samples were placed on ice, uncapped, and neutralized by the addition of 5 ml of a 6% solution of K2CO3. The tubes were recapped and vortexed for 1 min followed by centrifugation for 2 min at 3,000 rpm on a refrigerated tabletop centrifuge at 9 °C to separate the mixture into two phases and the upper (organic) phase was collected. The extraction procedure was repeated on the lower phase by adding 0.5 ml of hexane, vortexing and centrifuging. The organic phases were combined and evaporated under nitrogen to a volume of 75–100 μl. This solution was transferred to a GC vial, and the vial was crimped under nitrogen for FAME analysis by GC. All reactions were performed in sextet in a well ventilated fume hood.

Automated Robotic Transesterification Procedure

For robotic transesterification procedure Stock A and Stock B reagents were freshly prepared in bulk as stated above. The hexane solution and the bulk stock solutions A and B were placed in three different Teflon troughs. Plasma sample vials were placed on the carrier racks and clean uncapped pyrex tubes were labeled and placed in the reaction block (Fig. 1). The robot was initialized and the prompt for the number of samples to be processed is answered. Then the liquid handling arm (LH-arm) picks up disposable tips, aspirates 100–200 μl of plasma per tip, and then dispenses the plasma into the reaction tubes on the reaction block. Tips are ejected into a waste tip dispenser for biohazardous waste (labeled as “plasma disposable tips waste station” in Fig. 1) between plasma samples to avoid cross-contamination. The process is repeated until all plasma samples have been pipetted into reaction tubes. The LH-arm then obtains a fresh set of eight disposable tips and adds 800 μl of stock solution A to each reaction tube three times (2.4 ml total). The tips are then ejected into a general purpose waste tip dispenser. The program then issues a command to the reaction block to heat to 80 °C for 2.5 h. During the course of the 2.5 h reaction time, five additions of 0.9 ml each of stock B solution (every 20 m) are performed by the robot to prevent complete drying of the reaction tubes as well as to compensate for evaporation of the solution and to complete the transesterification reaction. This multiple addition helps to wash down the sides of the tubes and thus yield a complete reaction. Also after each round of addition of the stock B solution, the LH-arm was stationed away from the reaction block in order to minimize exposure of the robot arm pneumatics to corrosive vapors arising from the heated reaction tubes. However, corrosion is still a possibility and so the instrument accuracy is validated every day by weighing a certain volume of liquid that is dispensed by the LH-arm. After the last addition, the reaction block was allowed to cool to 25 °C via a command issued by the program. This sets the solenoid sensor to open to allow chilled fluid (10 °C) to circulate around the reaction block to bring the temperature down to 20 °C within approximately 30 min. The LH-arm picks up a fresh set of disposable tips, aspirates 700 μl of hexane from the third trough and dispenses this into each of the reaction tubes. After this task is performed, the tips are ejected. The LH-arm obtains fresh tips, aspirates 400 μl of the upper hexane phase and dispenses it at the bottom of the reaction tube, forcing the hexane to mix with the aqueous phase as it passes thru. This mixing step is repeated 20 times followed by the ejection of the tips. It was observed that 15–20 hexane aspirations for extraction of FAMEs was sufficient since fewer than 15 aspiration cycles resulted in incomplete extraction and more than 20 aspiration cycles did not produce any benefit. The LH-arm then obtains a fresh set of disposable tips and aspirates from the reaction tubes 250 μl of the upper phase hexane solution. The system is able to detect the position at which it becomes immersed in the hexane solution as it is lowered into the tube through detection of a change in pressure (or by capacitance for detection of the interface between the two liquids). The hexane is transferred into GC vials positioned in the GC vial block. Tips must be changed prior to mixing or sampling a new set of tubes. Since only the tips come in contact with sample tube contents and a new tip is used for each step, there is no possibility of sample cross-contamination. The GC vials can be heated in order to concentrate the sample through hexane evaporation.

Results and Discussion

Validation of Robotic Transmethylation Procedure

In a previous publication from this laboratory, we demonstrated that a simplification of the Lepage and Roy procedure could produce valid results [35]. In that work an “open tube” method was implemented that was amenable to robotic application. This chemical procedure was transferred to a robotic procedure as described above. A bulk plasma sample was transmethylated by the fully automated robotic method and the results compared to the standard Lepage and Roy procedure (n = 6). A second set of samples were similarly treated with the robotic procedure, and the hexane extracts were subjected to heating in the GC vial station in order to concentrate the extract further (Table 1). The mean values for each fatty acid concentration are similar in the three methods. The coefficients of variation are as good or better in the robotic method as the standard method. The fatty acid percentages in the three methods are very similar. The robotic method was in this way validated for plasma analysis. We have run the robotic method with an n of up to 40 samples and obtained similar results. The fatty acids measured here are the common ones reported for many other mammalian tissue analyses as well.
Table 1

Comparison of human plasma fatty acid concentrations in standardized and robotic methods using fast gas chromatography (n = 6)

Fatty acids

Standardized method

Robotic method

Robotic method (concentrated)

Mean (μg/ml)

CV (%)

Percentage of total fatty acids

Mean (μg/ml)

CV (%)

Percentage of total fatty acids

Difference (%)

Mean (μg/ml)

CV (%)

Percentage of total fatty acids

Differencea (%)

14:0

26.4

7.3

1.1

25.7

1.7

1.1

−2.5

27.1

1.9

1.2

2.5

16:0

538.2

6.5

22.5

568.3

2.6

24.1

5.6

538.8

1.6

23.1

0.1

18:0

198.0

6.6

8.3

214.1

3.1

9.1

8.1

206.0

1.1

8.8

4.0

20:0

6.4

4.9

0.3

6.8

1.8

0.3

5.7

7.2

3.4

0.3

12.7

22:0

19.7

2.2

0.8

21.5

5.1

0.9

9.1

21.4

5.0

0.9

8.6

24:0

16.0

1.4

0.7

17.2

4.7

0.7

7.0

17.3

6.3

0.7

8.0

Total saturates

804.8

6.1

33.6

853.6

1.5

36.2

6.1

817.8

1.1

35.0

1.6

16:1n-7

45.7

9.9

1.9

44.5

4.7

1.9

−2.6

42.5

8.7

1.8

−7.0

18:1n-7

43.9

5.7

1.8

42.7

4.4

1.8

−2.8

41.3

4.4

1.8

−5.9

18:1n-9

416.9

8.7

17.4

379.0

4.0

16.1

−9.1

390.4

0.8

16.7

−6.4

20:1n-9

2.7

5.7

0.1

2.8

3.6

0.1

1.8

2.6

7.4

0.1

−5.3

24:1n-9

19.8

2.3

0.8

18.7

3.0

0.8

−5.5

21.6

8.0

0.9

9.5

Monounsaturates

529.0

7.6

22.1

487.6

3.8

20.7

−7.8

498.4

1.1

21.3

−5.8

18:2n-6

663.1

8.1

27.7

636.1

3.6

26.9

−4.1

641.6

0.5

27.5

−3.2

18:3n-6

12.4

8.2

0.5

11.8

2.2

0.5

−4.9

10.6

0.7

0.5

−14.2

20:2n-6

4.9

7.2

0.2

5.0

2.2

0.2

0.7

5.1

3.4

0.2

2.2

20:3n-6

38.0

3.9

1.6

35.8

4.0

1.5

−5.8

36.2

3.3

1.5

−4.8

20:4n-6

192.5

3.0

8.0

188.5

2.6

8.0

−2.1

184.0

1.8

7.9

−4.4

22:4n-6

7.0

8.5

0.3

6.6

7.4

0.3

−4.8

6.6

5.9

0.3

−5.7

22:5n-6

5.4

8.4

0.2

5.0

2.6

0.2

−7.1

5.1

6.8

0.2

−7.0

Total n-6 PUFA

923.3

6.5

38.6

888.8

3.3

37.6

−3.7

889.1

0.7

38.1

−3.7

18:3n-3

15.5

6.5

0.6

14.3

7.6

0.6

−7.9

14.7

6.2

0.6

−5.4

20:5n-3

38.8

2.3

1.6

40.3

3.8

1.7

4.0

41.5

7.7

1.8

7.2

22:5n-3

17.8

2.6

0.7

16.3

6.3

0.7

−8.2

17.0

7.4

0.7

−4.2

22:6n-3

63.4

3.1

2.7

60.0

2.3

2.5

−5.5

57.5

4.0

2.5

−9.3

Total n-3 PUFA

135.5

2.6

5.7

130.8

3.2

5.5

−3.4

130.8

4.0

5.6

−3.5

Total PUFA

1,058.8

6.0

44.3

1,019.7

3.1

43.2

−3.7

1,019.9

1.1

43.7

−3.7

Total fatty acid

2,392.6

5.8

100.0

2,360.9

2.1

100.0

−1.3

2,336.1

0.8

100.0

−2.4

aPercentage difference of the concentration values

Fatty acid extraction using 10% decane in pentane has also been used as an alternative extraction solvent. During heating in the GC vial block, the pentane is easily evaporated leaving mainly the decane volume containing the methyl esters. It is also possible to leave an open GC vial on the GC autosampler and allow it to evaporate prior to analysis. We encountered some difficulties with dripping during the LH-arm transfer of pentane to the GC vials. However, this problem was overcome by first aspirating and dispensing pentane prior to sample transfer.

Another difficulty is the leaching of organic compounds from the disposable tips. The black tips used for capacitance based detection are worse in this respect, in our experience. These impurities elute primarily in the first 1.5 min of the GC run. They create difficulties for automated peak integration of minor peaks of 10–14 carbons with zero or one double bonds. Tips from a variety of vendors have given similar results, however, the clear tips used in pressure-based sensing exhibit a lower level of leaching.

Improvements in the Fast GC Method

We earlier developed a Fast GC method utilizing a DB-FFAP column of dimensions 15 m × 0.1 mm i.d. × 0.1 μm film thickness. A high temperature program was used in which all the fatty acid peaks of interest eluted within 8 min of injection, with a sample to sample injection time of 16 min (Fig. 2a). An attempt was made to further shorten the GC run times using an aggressive, single ramp method of 40 °C or 50 °C/min and starting from 135 to 150 °C up to 240 °C. However, this approach resulted in overlaps of critical peaks found in a 28 component quantitative standard (Nu Chek Prep 462); 18:1n-9, 18:1n-7 and 24:1n-9, 22:6n-3 were the overlapping peak pairs in this case. A method was then developed based on the constant flow mode of operation in which the oven temperature was ramped up and down to effect critical pair separation. With this method, peaks of interest eluted in less than 6 min and sample-to-sample injection time was under 13 min allowing for the elution of cholesterol (Fig. 2b). An initial hold time of 0.10 min and starting temp of 160 °C was used. We observed that even a small decrease in the hold time or a slight increase in the initial oven temperature results in coelution of the 10:0 and 12:0 peaks with the solvent peak. Similarly, small changes (e.g., 0.10 m) in the hold times after the temperature ramps resulted in a loss of peak separation. Use of the constant pressure mode instead of the constant volume mode using an identical temperature program led to an increase of 1 min in analysis time but maintained good separation of peaks. GC Methods have been published where the fatty acid components elute within 2 min [38], however these methods are not feasible for the analysis of FAMEs from plasma samples due to the coelution of several critical pairs (18:1n-9, 18:1n-7 and 24:1n-9, 22:6n-3) which then prevent correct quantitation.
Fig. 2

Improvements in Fast GC methods for separation of a mixed FAME external reference standard

The improved fast GC method that we have developed has been applied to the separation of plasma fatty acid methyl esters. In both the former method and the faster method, all of the peaks of interest are well separated, even with the shorter analysis time with the improved method (Fig. 3). The concentrations of human plasma fatty acids obtained by the robotic method and determined by both fast GC and by the improved GC method are presented in Table 2. Very similar results were obtained for each fatty acid concentration and its percentage of total fatty acids, and the CV values were similar between the two. Thus the faster GC method is suitable for fatty acid analyses of human plasma and can significantly reduce analytical run times.
Fig. 3

Comparison of fast and improved GC method gas chromatograms of human plasma FAME, including 23:0 and 22:3n-3 internal standards. Sample was prepared robotically. BHT butylated hydroxytoluene

Table 2

Comparison of human plasma fatty acid concentrations in the robotic methods using improved methods for fast gas chromatography (n = 6)

Fatty acids

Fast GC method

Improved fast GC method

Differencea (%)

Mean (μg/ml)

CV (%)

Percentage of total fatty acids

Mean (μg/ml)

CV (%)

Percentage of total fatty acids

14:0

25.7

1.7

1.1

27.2

5.5

1.1

5.7

16:0

568.3

2.6

24.1

573.3

4.1

23.8

0.9

18:0

214.1

3.1

9.1

220.2

4.4

9.1

2.9

20:0

6.8

1.8

0.3

6.6

4.6

0.3

−3.1

22:0

21.5

5.1

0.9

21.3

6.1

0.9

−1.0

24:0

17.2

4.7

0.7

17.6

9.8

0.7

2.5

Total saturates

853.6

1.5

36.2

866.2

3.8

35.9

1.5

16:1n-7

44.5

4.7

1.9

41.1

7.6

1.7

−7.6

18:1n-7

42.7

4.4

1.8

41.3

7.9

1.7

−3.2

18:1n-9

379.0

4.0

16.1

392.5

3.1

16.3

3.6

20:1n-9

2.8

3.6

0.1

2.7

7.6

0.1

−2.0

24:1n-9

18.7

3.0

0.8

20.3

5.0

0.8

8.5

Monounsaturates

487.6

3.8

20.7

497.9

2.7

20.6

2.1

18:2n-6

636.1

3.6

26.9

650.2

5.6

27.0

2.2

18:3n-6

11.8

2.2

0.5

11.3

4.6

0.5

−4.3

20:2n-6

5.0

2.2

0.2

5.1

6.3

0.2

2.4

20:3n-6

35.8

4.0

1.5

38.8

8.4

1.6

8.3

20:4n-6

188.5

2.6

8.0

196.4

2.8

8.1

4.2

22:4n-6

6.6

7.4

0.3

6.6

5.8

0.3

−1.2

22:5n-6

5.0

2.6

0.2

5.2

6.7

0.2

3.0

Total n-6 PUFA

888.8

3.3

37.6

913.4

4.1

37.9

2.8

18:3n-3

14.3

7.6

0.6

15.4

5.4

0.6

7.5

20:5n-3

40.3

3.8

1.7

37.7

3.0

1.6

−6.5

22:5n-3

16.3

6.3

0.7

16.9

8.8

0.7

3.4

22:6n-3

60.0

2.3

2.5

63.8

4.8

2.6

6.4

Total n-3 PUFA

130.8

3.2

5.5

133.7

3.2

5.5

2.2

Total PUFA

1,019.7

3.1

43.2

1,047.1

3.4

43.4

2.7

Total fatty acid

2,360.9

2.1

100.0

2,411.2

3.2

100.0

2.1

aPercentage difference of the concentration values

Detecting the liquid interface is one of the most critical steps in developing a high throughput transesterification automation application. The Tecan robot integrated liquid detection system employs two different physical principles: detection of changes in capacitance (c-LLD) or pressure (p-LLD). Both types of detection have been used in our studies. The capacitance-based system is useful for detecting an alkane–aqueous interface. This was used in collecting the upper phase as interface detection is followed by a 1-mm retraction into the organic layer. Pressure detection detects an air–liquid interface. This is useful when aspiration of only the upper organic phase is desired, such as when transferring the methyl ester extracts into GC vials or during the mixing phase when upper phase is squirted thru the lower phase.

The simplified transmethylation procedure used here dispensed with several steps once considered to be necessary. These steps include addition of cold acetyl chloride drop-wise, working under a nitrogen atmosphere, base addition for reaction mixture neutralization, vortexing, centrifugation and multiple extractions. Results in Table 1 and also our previous analysis [35] demonstrate that these steps are unnecessary or can be replaced by simple pipetting procedures that the robotic arm can perform. Such procedures were devised to overcome difficulties in implementing robotic processes that mimic the traditional chemical procedure. For example, vortexing can be performed by a robot using specialized apparatus, but it is a slow process whereby one sample at a time is mixed. Similarly, some robots are capable of centrifuging samples with integrated centrifuges, but this is a laborious, technically challenging and time-consuming process. What is required is for the organic and aqueous phases to be well mixed and then separated to effect methyl ester extraction. This has been achieved here by repeated mixing by aspirating the upper phase, moving the tip to the bottom of the tube and rapidly dispensing the organic solvent thru the aqueous phase. The organic phase can then be transferred into GC vials after the phases have settled and concentrated further for fast GC analysis, if necessary. Use of an internal standard has obviated the need for complete upper phase transfer and multiple extractions. The use of fast GC also makes possible the analysis of a lower concentration of solutes and, if an appropriate amount of plasma is used, no organic phase concentration may be required. Alternatively, if a microanalysis is required where sample is limited, the organic phase can be concentrated by heating the GC vial rack, and multiple extractions may be performed to increase solute concentrations.

The robotic method described here has been applied to the simultaneous analysis of larger numbers of samples. For example, preliminary experiments transmethylated and analyzed the same bulk plasma sample 40 times. When multiple sets of eight samples are to be analyzed, the program may perform the mixing steps, change tips, wait for a prescribed length of time, and then perform the transfer of the upper phase extract into the GC vials. Another alternative that may improve overall efficiency with a larger number of samples is to perform the mixing step on two sets of eight samples, and then return to the first set for upper phase transfer to GC vials. This creates a delay time that will allow adequate phase separation and also maintain a controlled and constant interval for mixing and sampling for each set of tubes to be processed. This method is then amenable to the analysis of 192 samples in one batch. In combination with the rapid GC methodology presented, two GCs can process the 192 samples within 1 day. This method then results in an order of magnitude increase in productivity and makes possible large clinical studies.

This method may then be used for both cross-sectional studies of populations and for monitoring compliance in interventional trials with fatty acid supplements and for the use of essential fatty acid related parameters as biomarkers for disease or disease risk. It should also be noted that where more detailed information is required concerning lipid class and molecular species content, that mass spectrometric techniques may be used for highly efficient analyses [39, 40, 41].

Analysis of about 200 samples per day would generate over 5,000 peaks per day that must be correctly assigned and entered into spreadsheets. This is a non-trivial task that requires automation. It will be the next step in development of a high throughput system for analysis of large sets of clinical samples.

Notes

Acknowledgments

This research was supported by the Intramural Research Program of the National Institutes of Health, NIAAA. The authors wish to thank Dr. William E. M. Lands for his help and encouragement during the course of these studies.

Supplementary material

11745_2007_3130_MOESM1_ESM.doc (2.1 mb)
ESM (DOC 2,111 kb)

References

  1. 1.
    Grossfield A, Feller SE, Pitman MC (2006) Contribution of omega-3 fatty acids to the thermodynamics of membrane protein solvation. J Phys Chem B Condens Matter Mater Surf Interfaces Biophys 110:8907–8909PubMedGoogle Scholar
  2. 2.
    Eldho NV, Feller SE, Nagle ST, Polozov IV, Gawrisch K (2003) Polyunsaturated docosahexaenoic vs docosapentaenoic acid-differences in lipid matrix properties from the loss of one double bond. J Am Chem Soc 125:6409–6421PubMedCrossRefGoogle Scholar
  3. 3.
    Harris JL, Hibbeln JR, Mackey RH, Muldoon MF (2003) Statin treatment alters fatty acid composition in hypercholesterolemic patients. Circulation 107:119Google Scholar
  4. 4.
    Salem N Jr, Hullin F, Yoffe AM, Karanian JW, Kim HY (1989) Fatty acid and phospholipid species composition of rat tissues after a fish oil diet. Adv Prostaglandin Thromboxane Leukot Res 19:618–622PubMedGoogle Scholar
  5. 5.
    Salem N Jr (1989) Omega-3 fatty acids: molecular and biochemical aspects. In: Spiller G, Scala J (eds) New protective roles of selective nutrients in human nutrition. Liss AR, New York, pp 263–317Google Scholar
  6. 6.
    Makrides M, Neumann MA, Simmer K, Pater J, Gibson RA (2000) A critical appraisal of the role of dietary long chain polyunsaturated fatty acids on neural indices of term infants: a randomized, controlled trail. Pediatrics 105:32–38Google Scholar
  7. 7.
    Weisenger HS, Armitage JA, Sinclair AJ, Vingrys AJ, Burns P, Weisinger RS (2001) Perinatal omega-3 fatty acid deficiency affects blood pressure later in life. Nat Med 7:258–259CrossRefGoogle Scholar
  8. 8.
    Neuringer M, Connor WE, Lin DS, Barstad L, Luck S (1986) Biochemical and functional effects of prenatal and postnatal omega 3 fatty acid deficiency on retina and brain in rhesus monkeys. Proc Natl Acad Sci 83:4021–4025PubMedCrossRefGoogle Scholar
  9. 9.
    Bourre JM, Francois M, Youyou A, Dumont O, Piciotti M, Pascal G, Durand G (1989) The effects of dietary alpha-linolenic acid on the composition of nerve membranes, enzymatic activity, amplitude of electrophysiological parameters, resistance to poisons and performance of learning tasks in rats. J Nutr 119:1880–1892PubMedGoogle Scholar
  10. 10.
    Yamamoto N, Saitoh M, Moriuchi A, Nomura M, Okuyama H (1987) Effects of dietary alpha-linolenate/linoleate balance on brain lipid compositions and learning ability of rats. J Lipid Res 28:144–151PubMedGoogle Scholar
  11. 11.
    Moriguchi T, Salem N Jr (2003) Recovery of brain docosahexaenoate leads to recovery of spatial task performance. J Neurochem 87:297–309PubMedCrossRefGoogle Scholar
  12. 12.
    Carlson SE, Werkman SH, Peeples JM, Wilson WM (1994) Growth and development of premature infants in relation to omega 3 and omega 6 fatty acid status. World Rev Nutr Diet 75:63–69PubMedGoogle Scholar
  13. 13.
    Willats P, Forsyth JS, DiModugno MK, Varma S, Colvin M (1998) Effects of long-chain polyunsaturated fatty acids in infant formula on problem solving at 10 months of age. Lancet 352:688–691CrossRefGoogle Scholar
  14. 14.
    Birch EE, Garfield S, Hoffman DR, Uauy R, Birch DG (2000) A randomized controlled trial of early dietary supply of long-chain polyunsaturated fatty acids and mental development of in term infants. Dev Med Child Neurol 42:174–181PubMedCrossRefGoogle Scholar
  15. 15.
    Contreras MA, Greiner RS, Chang MC, Myers CS, Salem N Jr, Rapoport SI (2000) Nutritional deprivation of alpha-linolenic acid decreases but does not abolish turnover and availability of unacylated docosahexaenoic acid and docosahexaenoyl-CoA in rat brain. J Neurochem 75:2392–2400PubMedCrossRefGoogle Scholar
  16. 16.
    Su HM, Bernardo L, Mirmiran M, Hong MX, Thomas CN, Nathanielsz PW, Thomas BR (1999) Bioequivalence of dietary alpha-linolenic and docosahexaenoic acids as sources of docosahexanenoate accretion in brain and associated organs of neonatal baboons. Pediatr Res 45:87–93PubMedCrossRefGoogle Scholar
  17. 17.
    Lefkowitz W, Lim SY, Lin Y, Salem N Jr (2005) Where does the developing brain obtain its docosahexaenoic acid? Relative contribution of dietary [alpha]-linolenic acid, docosahexaenoic acid, and body stores in the developing rat. Pediatr Res 57:157–165PubMedCrossRefGoogle Scholar
  18. 18.
    Pawlosky RJ, Hibbeln JR, Lin Y, Goodson S, Riggs P, Sebring N, Brown GL, Salem N Jr (2003) Effects of beef- and fish-based diets of the kinetics of n-3 fatty acid metabolism in human subjects. Am J Clin Nutr 77:565–572PubMedGoogle Scholar
  19. 19.
    Lin YH, Pawlosky RJ, Salem N Jr (2005) Simultaneous quantitative determination of deuterium- and carbon-13-labeled essential fatty acids in rat plasma. J Lipid Res 46:1974–1982PubMedCrossRefGoogle Scholar
  20. 20.
    Simopoulos AP (2004) The traditional diet of Greece and cancer. Eur J Cancer Prev 13:219–230PubMedCrossRefGoogle Scholar
  21. 21.
    Harris WE (2007) Omega-3 fatty acids and cardiovascular disease: a case for omega-3 index as a new risk factor. Pharmacol Res 55:217–223PubMedCrossRefGoogle Scholar
  22. 22.
    Lands WEM (1995) Long-term fat intake and biomarkers. Am J Clin Nutr 61:721S-725SPubMedGoogle Scholar
  23. 23.
    von Schacky C (2007) Omega-3 fatty acids and cardiovascular disease. Curr Opin Clin Nutr Metab Care 10:129–135Google Scholar
  24. 24.
    Lands WEM, Libelt B, Morris A, Kramer NC, Prewitt TE, Bowen P, Schmeisser D, Davidson M, Burns JH (1992) Maintenance of lower proportions of (n-6) eicosanoid precursors in phospholipids of human plasma in response to added dietary (n-3) fatty acids. Biochim Biophys Acta 1180:147–162PubMedGoogle Scholar
  25. 25.
    Lands WEM (1993) Eicosanoids and health. Ann N Y Acad Sci 676:40–59CrossRefGoogle Scholar
  26. 26.
    Albert CM, Campos H, Stampfer MJ, Ridker PM, Manson JE, Willett WC, Ma J (2002) Blood levels of long-chain n-3 fatty acids and the risk of sudden death. N Engl J Med 346:1113–1118PubMedCrossRefGoogle Scholar
  27. 27.
    Lemaitre RN, King IB, Mozaffarian D, Kuller LH, Tracy RP, Siscovick DS (2003) n-3 Polyunsaturated fatty acids, fatal ischemic heart disease, and nonfatal myocardial infarction in older adults: the cardiovascular health study. Am J Clin Nutr 77:319–325PubMedGoogle Scholar
  28. 28.
    Simon JA, Hodgkins ML, Browner WS, Neuhaus JM, Bernert JT Jr, Hulley SB (1995) Serum fatty acids and the risk of coronary heart disease. Am J Epidemiol 142:469–476PubMedGoogle Scholar
  29. 29.
    Kang JX, Leaf A (1996) The cardiac antiarrhythmic affects of polyunsaturated fatty acids. Lipids 31:S41–S44PubMedCrossRefGoogle Scholar
  30. 30.
    Marchiolli R, Barzi F, Bomba E, Chieffo C, Di Gregorio D, Di Mascio R, Franzosi MG, Geraci E, Levantesi G, Maggioni AP, Mantini L, Marfisi RM, Mastrogiuseppe G, Mininni N, Nicolosi GL, Santini M, Schweiger C, Tavazzi L, Tognoni G, Tucci C, Valagussa F (2002) Early protection against sudden death by n-3 polyunsaturated fatty acids after myocardial infarction: time course analysis of the results of the gruppo italiano per lo studio della sopravvivenza nell’infarto miocardico (GISSI)-Prevenzione. Circulation 105:1897–1903CrossRefGoogle Scholar
  31. 31.
    Mihrshahi S, Peat JK, Webb K, Tovey ER, Marks GB, Mellis CM, Leeder SR (2001) The childhood asthma prevention study (CAPS): design and research protocol of a randomized trial for the primary prevention of asthma. Control Clin Trials 22:333–354PubMedCrossRefGoogle Scholar
  32. 32.
    Coleman H, Chew E, AREDS2 (2007) Nutritional supplementation in age-related macular degeneration. Curr Opin Ophthalmol 3:220–223CrossRefGoogle Scholar
  33. 33.
    Smithers LG, McPhee AJ, Gibson RA, Makrides M (2004) Visual development of preterm infants fed high dose docosahexaenoic acid. Asia Pac J Clin Nutr 13:S50Google Scholar
  34. 34.
    New Alzheimer’s clinical trials to be undertaken by NIA nationwide consortium. http://www.nia.nih.gov/Alzheimers/ResearchInformation/NewsReleases/PR20061017ADC
  35. 35.
    Masood MA, Stark KD, Salem N Jr (2005) A simplified and efficient method for the analysis of fatty acid methyl esters suitable for large clinical studies. J Lipid Res 46:2299–2305PubMedCrossRefGoogle Scholar
  36. 36.
    Lepage G, Roy CC (1986) Direct transesterification of all classes of lipids in a one-step reaction. J Lipid Res 27:114–120PubMedGoogle Scholar
  37. 37.
    Stark KD, Salem N Jr (2005) Fast gas chromatography methods and column selection for the identification of fatty acid methyl esters from mammalian samples. Lipid Technol 17:181–185Google Scholar
  38. 38.
  39. 39.
    Kuksis A (2007) Lipidomics in triacylglycerol and cholesteryl and cholesterol ester oxidation. Front Biosci 12:3203–3246PubMedCrossRefGoogle Scholar
  40. 40.
    Merrill AH Jr, Sullards MC, Allegood JC, Kelly S, Wang E (2005) Sphingolipidomics: high-throughput, structure-specific, and quantitative analysis of sphingolipids by liquid chromatography tandem mass spectrometry. Methods 36:207–224PubMedCrossRefGoogle Scholar
  41. 41.
    Han X, Gross RW (2005) Shotgun lipidomics: electrospray ionization mass spectrometric analysis and quantitation of cellular lipidomes directly from crude extracts of biological samples. Mass Spectrom Rev 24:367–412PubMedCrossRefGoogle Scholar

Copyright information

© AOCS 2007

Authors and Affiliations

  1. 1.Laboratory of Membrane Biochemistry and BiophysicsNational Institutes on Alcohol Abuse and Alcoholism, National Institutes of HealthRockvilleUSA
  2. 2.BethesdaUSA
  3. 3.Laboratory of Proteomics and Analytical Technologies, Advanced Technology Program, SAIC-Frederick, Inc.National Cancer Institute at FrederickFrederickUSA

Personalised recommendations