Stereospecific analysis of triacylglycerols rich in long-chain polyunsaturated fatty acids.

Six oils of marine, algal, and microbial origin were analyzed for stereospecific distribution of component fatty acids. The general procedure involved preparation of sn-1,2-(2,3)-diacylglycerols by partial deacylation with ethylmagnesium bromide or pancreatic lipase, separation of X-1,3- and sn-1,2(2,3)-diacylglycerols by borate thin-layer chromatography, resolution of the sn-1,2- and sn-2,3-enantiomers by chiral phase high-performance liquid chromatography following preparation of dinitrophenylurethane derivatives, and determination of the fatty acid composition by gas chromatography. Unexpected complications arose during a stereospecific analysis of triacylglycerols containing over 33% of either 20:4 or 22:6 fatty acids. The sn-1,2(2,3)-diacylglycerols made up of two long-chain polyunsaturated acids migrated with the X-1,3-diacylglycerols and required separate chiral phase resolution. Furthermore, the enzymatic method yielded sn-1,2(2,3)-diacylglycerols, overrepresenting the polyenoic species due to their relative resistance to lipolysis, but prolonged digestion yielded correct composition for the 2-monoacylglycerols. The final positional distribution of the fatty acids was established by pooling and normalizing the data from subfractions obtained by normal- and chiral-phase separation of diacylglycerols. The molecular species of X-1,3-, sn-1,2- and sn-2,3-diacylglycerol dinitrophenylurethanes were identified by chiral-phase liquid chromatography/mass spectrometry with electrospray ionization, which demonstrated a preferential association of the paired long-chain acids with the sn-1,2- and sn-2,3-diacylglycerol isomers.

Quantitation of plasma lipids by gas-liquid chromatography on high temperature polarizable capillary columns.

We have previously demonstrated the potential usefulness of capillary columns coated with a high temperature polarizable phenylmethylsilicone liquid phase in plasma lipid profiling (Kuksis, A., J.J. Myher, and P. Sandra. 1990. J. Chromatogr. 500: 427-441). The present study reports improved operating conditions along with a practical application to the analysis of a series of human plasma samples in comparison to capillary gas-liquid chromatography on nonpolar columns. For this purpose the plasma lipids were dephosphorylated with phospholipase C and converted to the trimethylsilyl ethers. The molecular species of the plasma lipids were identified by comparing the relative retention times to reference standards. The species were quantitated using tridecanoylglycerol as internal standard. The recoveries of the lipid classes were determined by summing the molecular species within each carbon number and comparing the proportion of the carbon numbers obtained on polar and nonpolar columns. The relative recoveries varied with the lipid class and sample load and averaged as follows: FA C16, 78%; FA C18, 78%; FC, 99%; tridecanoylglycerol (TD), 100% (by definition); CER 34:1, 92%; DG C34, 103%; DG C36, 98%; CE C16, 73%; CE C18, 61%; TG C50, 93%; TG C52, 60%; TGC54, 32%. We conclude that high temperature polarizable capillary columns are suitable for qualitative and quantitative assessment of plasma lipids and provide more information per man-hour, instrument time, and unit cost than any other analytical method known.

Gas chromatographic profiles of plasma total lipids as indicators of dietary history. Correlation with carbohydrate and alcohol intake based on 24-h dietary recall.

Quantitative gas chromatographic estimates of the major lipid classes and molecular species in fasting plasma were correlated with total carbohydrate, starch, fibre, sucrose and alcohol intake based on 24-h dietary recall. Spearman coefficients (rs) and tests of significance (P) were obtained for groups of 775 males and 471 females aged 20-59 years from a Toronto-McMaster Lipid Research Clinics Population Study. The most significant correlations varying from rs 0.1 to 0.2 and P 0.001 to 0.0005 (n = 400-773) were between increased intake of alcohol and increased ratios of C50/C54 triacylglycerols, C34/C36 phosphatidylcholines and phosphatidylcholine/free cholesterol (PC/FC) of plasma. Increase in total dietary carbohydrate, starch and fibre correlated with decreasing C50/C54 triacylglycerol, C34/C36 phosphatidylcholine and PC/FC ratios (rs = -0.1-0.2; P less than 0.002-0.04; n = 400-773). In contrast, consumption of high levels of alcohol was associated with increasing C50/C54 triacylglycerol, C34/C36 phosphatidylcholine and PC/FC ratios. A high intake of alcohol (50-150 ml per day) distinguished itself from other simple carbohydrate-induced lipid profiles by its marked effect on increased C50/C52 triacylglycerol and PC/FC ratio.

Gas chromatographic profiles of plasma total lipids as indicators of dietary history. Correlation with fat intake based on 24-h dietary recall.

Fasting plasma total lipid profiles were determined by high-temperature gas chromatography on a total of 1246 free living urban subjects, ages 20-59 years, from the Toronto-McMaster Lipid Research Clinic Population Study. Quantitative estimates of the major molecular species, lipid classes and lipid class ratios were correlated with a total of twelve dietary lipid components, including total saturated and unsaturated fats. oleic and linoleic acids, and cholesterol, to give appropriate Spearman coefficients (rS) and tests of significance (P) for groups of 775 males and 471 females. The intake of the various nutrients was derived from a 24-h dietary recall. The most significant correlations varying from rs +/- 0.1-0.4 and P less than 0.0001-0.0005 were between the intake of total fat, individual saturated and unsaturated fats, and the ratios of C50/C54 triacylglycerols and the C34/C36 phosphatidylcholines, which reflected the nature and quantity of the dietary fat consumed. Increases in dietary cholesterol and saturated fat produced small increases in plasma cholesterol and saturated triacylglycerols, while unsaturated dietary fat produced small decreases in saturated and increases in unsaturated plasma triacylglycerols. These changes in the plasma lipid parameters are consistent with those observed previously in much more limited dietary experiments with accurately known composition of ingested fats. It is, therefore, concluded that direct gas chromatographic profiling of plasma total lipids provides a simple and rapid method of verifying the overall correctness of the dietary recall.

Plasma phosphatidylcholine/free cholesterol ratio as an indicator for atherosclerosis.

Plasma lipid profiles were determined in two inbred strains of mice, C57BR/cdJ and CBA/J, fed either a normal chow or an atherogenic diet for a 15-week period, starting at 10 weeks of age. On the chow diet, the C57BR/cdJ had significantly higher mean free cholesterol, esterified cholesterol, and total lipid values, and a significantly lower mean phosphatidylcholine/free cholesterol ratio than the CBA/J mice. On the atherogenic diet, the C57BR/cdJ had significantly higher mean levels for all lipids classes, except triacylglycerols, than the CBA/J mice. The mean plasma free cholesterol and esterified cholesterol levels of the C57BR/cdJ were four times greater than those of the CBA/J strain on the atherogenic diet. The mean plasma phosphatidylcholine/free cholesterol ratio of the C57BR/cdJ mice on the high cholesterol diet was 0.87 compared to 1.91 for CBA/J mice. These plasma lipid changes were associated with a marked development of atheromatous deposits in the wall of the aortic sinus of the C57BR/cdJ compared to the CBA/J animals. The phosphatidylcholine/free cholesterol ratios of the liver lipids of both strains decreased from 2.5-2.7 on the chow diet to 1.0-1.1 on the high cholesterol diet. It is suggested that a plasma phosphatidylcholine/free cholesterol ratio less than 1 represents a supersaturation of the vascular system and the vessel wall with cholesterol, which leads to a destabilization of the plasma membranes of the endothelial and smooth muscle cells, and an infiltration of the vessel wall by the plasma lipids.

Decreased plasma phosphatidylcholine/free cholesterol ratio as an indicator of risk for ischemic vascular disease.

As part of a population survey and a follow-up study of plasma lipid profiles by high temperature gas-liquid chromatography, we have determined the quantities and relative proportions of all major chemical classes and molecular species of lipids of plasma from 1200 subjects at Visit 2 of the Toronto-McMaster Lipid Research Clinic Prevalence Study. We have compared these values between our 24 subjects with ischemic vascular disease and 73 control subjects matched for age, sex, and plasma total cholesterol and triacylglycerols. The phosphatidylcholine/free cholesterol ratio showed the highest association with ischemic vascular disease of any of over 10 other lipid parameters and all the common risk indicators except high density lipoprotein cholesterol. The phosphatidylcholine/free cholesterol ratio had a relative risk ratio of 20/4 (95% confidence limits, 15/9, 23/1) and high density lipoprotein cholesterol, a risk ratio of 23/1 (95% confidence limits, 24/0, 19/5) for ischemic vascular disease. The average ratio of phosphatidylcholine/free cholesterol for the ischemic vascular disease group was 1.36 and for the controls 1.51, the population average being 1.50. Plasma high density lipoprotein cholesterol had a significant correlation (R = 0.15) with the phosphatidylcholine/free cholesterol ratio in the total population sample. The increased risk for ischemic vascular disease from a lower phosphatidylcholine/free cholesterol ratio may possibly be explained on the basis of decreased fluidity and stability of the lipoproteins due to a relative oversaturation with free cholesterol.

Lipid class and molecular species interrelationships among plasma lipoproteins of type III and type IV hyperlipemic subjects.

As a further appraisal of lipoprotein interconversion and equilibration of lipid components a detailed examination was made of the chemical class and molecular species interrelationships among the major fasting plasma lipoprotein fractions within each of six male Type IIi and Type IV hyperlipemic subjects subsisting on free choice diets. The lipoprotein fractions were prepared by conventional ultracentrifugation and the lipid class and molecular species composition of the corresponding lipoprotein fractions were determined by gas chromatography of the intact glycerol esters and ceramides. In general, each lipoprotein fraction possessed a well defined lipid class composition, which was characterized by a dramatically decreasing triacylglycerol and increasing phospholipid and cholesteryl ester content, when progressing from the very low (VLDL) to the low (LDL) and high (HDL) density lipoproteins, as already established for normolipemic subjects. Likewise, the LDL1, and LDL2 of the hyperlipemic subjects contained about two times higher proportion of total phospholipid as sphingomyelin than VLDL and HDL. Furthermore, the sphingomyelins of the HDL fraction contained about 30% more of the higher and 30% of the lower molecular weight species than the sphingomyelins of the VLDL. Smaller differences were seen in the molecular species composition of the phosphatidylcholines, cholesteryl esters and triacylglycerols among the corresponding lipoproteins. In comparison to normolipemic subjects analyzed previously, the hyperlipemic subjects showed greater individual variability. Despite this variability the lipid class and molecular species composition in the hyperlipemic subjects was again incompatible with the hypothesis which postulates direct VLDL conversion into LDL nd HDL under the influence of lipoprotein lipase and lecithin: cholesterol acyltransferase. The main differences between normolipemic and hyperlipemic plasma were found to reside in the number of the VLDL and LDL, lipoprotein particles and not in their chemical composition or physical structure, or in the apparent mechanism of their metabolic interconversion.

Lipid class and molecular species interrelationships among plasma lipoproteins of normolipemic subjects.

As evidence of lipoprotein interconversion and/or equilibration, a gas-liquid chromatographic (GLC) examination was made of the lipid class and molecular species interrelationships among the major fasting plasma lipoprotein fractions within each of seven male and four female normolipemic subjects subsisting on free choice diets. The lipoprotein fractions were prepared by conventional ultracentrifugation and the lipid class and molecular species composition of the corresponding lipoprotein fractions were determined by GLC of the intact cholesterol and glycerol esters and of ceramides. In general, each lipoprotein fraction possessed a well defined lipid class composition, which was characterized by a dramatically decreasing triacylglycerol and an increasing phospholipid and cholesteryl ester content when progressing from the very low (VLDL), to the low (LDL2) and high (HDL) density lipoproteins, as already established by conventional analyses. However, the LDL2 contained about a two times higher proportion of total phospholipids as sphingomyelin than VLDL and HDL. Furthermore, the sphingomyelins of the HDL fraction contained about 30% more of the higher molecular weight species than the sphingomyelins of either VLDL or LDL. Smaller differences were seen in the molecular species composition of the phosphatidylcholines, cholesteryl esters and triacylglycerols among the corresponding fractions of lipoproteins. In general, the lipid class and molecular species distribution is incompatible with the hypothesis which postulates VLDL conversion into LDL and HDL under the influence of lipoprotein lipase and lecithin:cholesterol acyltransferase. The significant differences noted in the lipid class and molecular species distribution suggest that the true transformations of the lipoproteins are much more complex and may also involve cholesteryl ester-triacylglycerol, triacylglycerol and phosphatidylcholine exchanges via appropriate carrier plasma proteins, as well as possible phase separation of lipids during the removal of the excess surface material from the VLDL remnants, as already demonstrated in in vitro experiments. It is concluded that a direct GLC analysis of the neutral and polar lipid components of plasma lipoprotein classes provides important evidence of lipoprotein interrelationships which may be utilized to test existing and new hypotheses of lipoprotein interconversion.

Comparative determination of plasma phospholipids by automated gas--liquid chromatographic and manual colorimetric phosphorus methods.

Plasma samples obtained during a prevalence study of hyperlipemia in a free-living urban population were analyzed for phosphatidylcholine, sphingomyelin and lysophosphatidylcholine content by automated high-temperature gas--liquid chromatographic (GLC) and manual colorimetric phosphorus (thin-layer chromatographic, TLC) methods. The GLC estimates were obtained from a quantitative analysis of the diacylglycerol, ceramide and monoacylglycerol moieties released from the parent phospholipids by digestion with phospholipase C, while the TLC estimates were derived by manual colorimetric phosphorus analyses of the individual phospholipid classes resolved by TLC. On samples analyzed over a two-year period the methods gave excellent correlation for the total phospholipids (r = 0.98), phosphatidylcholine (r = 0.98) and sphingomyelin (r = 0.90), but resulted in a poor agreement for lysophosphatidylcholine (r = 0.69). Comparable results were obtained for estimates of these phospholipids in plasma very low density, low density and high density lipoproteins. The between-method coefficient of variation ranged from 3 to 5% for phosphatidylcholine and from 5 to 10% for sphingomyelin. The relative error for the estimates of lysophosphatidylcholine ranged from 10 to 25%, and was due to the inclusion in the GLC estimates of a variable proportion of plasma free monoacylglycerols. Other differences between the two methods are due to various analytical errors and biases inherent in the two techniques. The within-day, within GLC, relative error averaged 1% for phosphatidylcholine, 3% for sphingomyelin and 5% for lysophosphatidylcholine. The apparent high precision and accuracy of the GLC method recommend it as an alternative to conventional direct methods of phospholipid analyses based on TLC isolation of lipid classes and colorimetric measurements of their phosphorus content. The GLC analyses of the plasma phospholipids are particularly convenient in conjunction with GLC measurements of plasma cholesterol and triacylglycerols, where a smaller throughput of samples is not a limitation and where both total amount and relative proportion of the lipids are of interest.

Comparative determination of plasma cholesterol and triacylglycerol levels by automated gas--liquid chromatographic and autoanalyzer methods.

Plasma samples obtained during a prevalence study of hyperlipemia in a free living urban population were analyzed for total cholesterol and triacylglycerol content by automated high-temperature gas--liquid chromatographic (GLC) and automated colorimetric (Auto-Analyzer, AA11) methods. The analyses were done over a three-year period. The methods gave excellent overall correlation for both total cholesterol (r = 0.9811) and total triacylglycerols (r = 0.9739). Detailed comparisons of the results obtained by the two methods with natural samples over the entire concentration range, indicated that the GLC method gave cholesterol values 5--10 mg% lower and triacylglycerol values 10--20 mg% lower than the corresponding AA11 determinations. The differences between the two methods are attributed to an overestimation of the cholesterol and triacylglycerol levels by the AA11 method due to presence of variable amounts of interfering chromogens in the plasma extracts. The between-method relative error ranged from 3 to 5% for cholesterol and from 5 to 10% for triacylglycerols. The within-day standard deviation of GLC averaged 2.3 mg% for cholesterol and 3.5 mg% for triacylglycerols. The between-day standard deviation of the GLC method averaged about 6 mg% for both cholesterol and triacylglycerols. The within-day, within GLC, relative error averaged 1.12% for cholesterol and 2.66% for triacylglycerols. The apparent high precision and high accuracy of the GLC method recommend it as an alternative to the indirect methods of plasma cholesterol and triacylglycerol analysis, especially where a smaller throughput of samples is not a limitation and where both total amount and composition of the lipids is of interest.

Oral contraceptive estrogen and plasma lipid levels.

Blood samples were obtained from 20 women, aged 18 to 30 years, before and during the use of selected oral contraceptives. The contraceptive preparation containing 100 microgram mestranol induced increases in triglycerides, esterified cholesterol, free cholesterol, and phospholipids. Ethinyl estradiol with norgestrol, on the other hand, tended to decrease the lipid levels, while preparations containing only 50 microgram mestranol produced no significant change in the levels of the four substances. The levels produced by each of these three preparations did not exceed the normal limits. A cyclic fluctuation was observed in the monthly cycles.

Identification of plant sterols in plasma and red blood cells of man and experimental animals.

Direct gas liquid chromatography (GLC) of total plasma lipids showed small peaks (0.5-1.5% of total free sterol area) corresponding to free C28 and C29 sterols in ca. 50% of some 3,000 normal subjects and patients with hyperlipemia. Comparable proportions of similar peaks were present in the sterol fraction isolated from the red blood cells of many of these subjects. The maximum levels of these components in the plasma and red blood cells of domestic and laboratory animals were up to 10 times higher than those seen in man. Detailed gas chromatography/mass spectrometry analyses of the plasma lipids from a much more limited number of subjects and animals showed that the GLC peaks corresponding to the free C28 and C29 sterols were largely due to the plant sterols campesterol, stigmasterol, and beta-sitosterol. In all instances, variable amounts (0.05-0.2% of the total free sterol area) of 7-dehydrocholesterol, desmosterol, lanosterol, and cholesterol alpha-oxide were also detected. While the total content and composition of the plasma plant sterols appeared to vary greatly among the subjects, it never exceeded 2% of total sterol in the normal subjects and patients examined. There was no evidence for a significant increase in the plant sterol content of the plasma of patients with hypercholesterolemia or hypertriglyceridemia.

Determination of plasma lipid profiles by automated gas chromatography and computerized data analysis.

Plasma or serum [ 0.1-1.0 ml] was digested with phospholipase C and total lipid extracts were prepared and silylated in the presence of tridecanoylglycerol as internal standard. The neutral lipid and free fatty acid profiles were determined by means of an automated GLC system equipped with an unheated on-column inlet, time actuated liquid injector, programmed heating, cooling and equilibration cycles, and an electronic peak area integrator. The separations were accomplished on a 50 cm x 2 mm i.d. steel column packed with 3% OV-1 on100-120 mesh Gas Chrom Q using nitrogen as a carrier gas in the temperature range 175-350 degrees C. The tube number, peak retention time and peak area were recorded on a punched paper tape, which was subsequently read into a computer via a time-share terminal. The composition of the sample was calculated in relation to the internal standard using a modification of a commercially available computer program and the results were expressed as mg or mole % and characteristic molar ratios of lipid classes. In addition to estimates for total cholesterol and triglyceride, the method provides a detailed account of individual or small groups of molecular species of various lipid classes, which is a major advantage over other automated methods of plasma lipid analyses.