Systemic and Forearm Triglyceride Metabolism: Fate of Lipoprotein Lipase-Generated Glycerol and Free Fatty Acids

Eight healthy male subjects, aged 25.6 ± 2.2 years and weighing 73.7 ± 3.1 kg (BMI 24.0 ± 0.6 kg/m²), participated in the study. Individuals on chronic medications or with fasting plasma triglycerides >150 mg/dl were excluded. Subjects refrained from alcohol consumption and vigorous exercise for 48 h before the study. The protocol was approved by the Saint Luke’s Hospital Institutional Review Board, and informed, written consent was obtained from the subjects after the nature of the study was explained.

Subjects were admitted to the Clinical Study Unit the afternoon before the experiment and at 1900 h. A mixed meal (50% carbohydrate, 30% fat, and 20% protein) was given containing calories equal to one-third of energy requirements for weight maintenance, calculated as 135% of estimated basal energy expenditure (Harris-Benedict equation). The next morning, after an overnight (12–14 h) fast, a 20-gauge catheter was placed in a radial artery under local anesthesia and an additional sampling catheter was placed in an antecubital vein of the ipsilateral forearm. Oxygen saturation was measured in blood from this site to confirm that it was <60%, ensuring that the selected vein was draining primarily deep forearm tissues. An infusion catheter was inserted into a contralateral forearm vein. All three catheters were kept patent by controlled (15 ml/h each) infusions of 0.9% NaCl. After obtaining blood samples for assay blanks at 0800 h, continuous infusions of [1-¹⁴C]oleate (∼0.3 μCi/min), [1,3-¹⁴C]glycerol (∼0.45 μCi/min), [oleyl-9,10-³H]triolein (∼0.53 μCi/min), and [glycerol-1,2,3-³H]triolein (∼1.2 μCi/min) were started and continued to the end of the study. Forearm blood flow was measured at 0900 h by mercury strain-gauge plethysmography, which produces data expressed in milliliter per unit mass per minute (8). After allowing 2 h to establish isotope equilibrium, seven paired blood samples were taken at 10-min intervals for determination of steady-state concentrations and specific activities of plasma glycerol, free oleate, and triglycerides.

All radioactive materials were purchased from American Radiolabeled Chemicals (St. Louis, MO). [1,3-¹⁴C]glycerol and [1-¹⁴C]oleate were prepared for infusion as previously described (8). The [9,10-³H]triolein and the [glycerol-1,2,3-³H]triolein were purified by thin-layer chromatography and separately added to aliquots of 10% Liposyn (Abbott Laboratory, Chicago, IL), and labeled triglyceride-rich particles were then isolated by size-exclusion high-performance liquid chromatography (HPLC) (9,10). The material was collected in a sterile vial containing unlabeled lipid emulsion in order to restore the phospholipid excess present in commercial lipid emulsions, thus maintaining particle stability, and autoclaved (10,11). The labeled particles in the emulsion have previously been shown to be a suitable surrogate for the study of chylomicron triglyceride metabolism (10).

Blood samples were taken, immediately transferred to precooled EDTA tubes containing 0.5 mg of paraoxon (an inhibitor of LPL), and kept on ice until centrifugation at 4°C. Plasma triglyceride concentrations were determined by replicate analysis (n = 15 on each specimen) using a Cobas Fara analyzer and a lipase/glycerol kinase-based kit (Bayer TO1-1868-02); for each sample, the three highest and three lowest values were dropped, and the remaining nine values were averaged. This approach produces excellent precision, since the intra-assay coefficient of variation on the centrifugal analyzer is already excellent, usually <2%.

The concentration and specific activity of plasma oleate was measured by HPLC (12), using [²H₃₁]palmitate as an internal standard (13). The concentration and specific activity of plasma glycerol was measured by HPLC (14), using 1,2,6-hexanetriol as an internal standard. In both assays, two standard curves are used: one utilizing the internal standard to determine concentration, and a second “absolute” standard curve containing ¹⁴C-labeled oleate or glycerol to determine the mass of material in the unknown sample for the calculation of specific activity. A modified Folch extraction of whole plasma, followed by removal of labeled FFAs with 0.01N Na₂CO₃, was performed to determine whole plasma ³H triglyceride concentrations.

Steady-state equations were used for all calculations (15).

1) The systemic rates of appearance (R_a) of glycerol and oleate were calculated from the formula:

where SA_A is the specific activity in arterial plasma.

2) Systemic metabolic clearance rates (MCRs) of glycerol and oleate were calculated from the formula:

3) Fractional extraction (FE) of glycerol and oleate were calculated from the formula:

where SA_V is the specific activity in venous plasma, and C_V and C_A are the concentrations in venous and arterial plasma, respectively.

4) Uptake and release of glycerol and oleate in the forearm tissue beds were calculated from the formulae:

where plasma flow (ml · 100 g⁻¹ · min⁻¹) equals blood flow × (1 − hematocrit). Calculations are made on the assumption that fatty acids are only carried in plasma, whereas glycerol is distributed in whole blood (16).

5) The systemic rates of appearance of [³H] were calculated from the formula:

6) Systemic fractional release (%) was calculated from the formula:

assuming that the [³H] R_a (dpm/min) would equal the [³H]triolein infusion rate if there were no local uptake of [³H]glycerol or [³H]oleate

7) The rate of [³H]oleate or glycerol release in a tissue bed that would be expected if there were no local uptake of LPL-generated oleate or glycerol was calculated from the uptake of ³H-triglyceride from arterial plasma according to the formula:

Expected [³H] release (dpm · 100 g tissue⁻¹ · min⁻¹) = ([³H TG]_A − [³H TG]_V) · blood/plasma flow, where [³H TG] equals concentration of ³Htriglyceride in dpm/ml.

8) The actual release of triglyceride-derived [³H] in a tissue was calculated from the formula:

where the expression ([³H] SA_A · C_A) · FE equals the predicted concentration of ³H-oleate or ³H-glycerol in venous plasma if LPL activity in the tissue bed is zero.

9) Forearm fractional release (%) of [³H]oleate or glycerol was calculated by the formula:

The conceptual basis for formulae 7–9 is illustrated schematically in Fig. 1. After tracer equilibration is achieved, a constant amount of both [³H] and [¹⁴C] is delivered to the tissue bed via arterial blood. Since the [¹⁴C]-labeled material can only disappear across the tissue bed, the decrement in [¹⁴C] concentration represents only uptake. The same is true for the [³H] triglyceride tracers. It is assumed that the arteriovenous difference in [³H] triglyceride concentration (see formula 7) is due to complete hydrolysis of labeled triglyceride (17). Only [³H]oleate and [³H]glycerol can both be produced and taken up by forearm tissue. If one assumes that LPL activity in the tissue bed is zero, then one can predict (for example) venous [³H]oleate concentrations from the extraction fraction of [¹⁴C]oleate and the arterial concentration of [³H]oleate (formula 8). If there is LPL activity in the tissue bed, then the release, or escape, of [³H]oleate due to LPL activity can be calculated from the difference between measured and predicted venous [³H]oleate concentration. Fractional release, then, represents the relationship between actual release of [³H]oleate (formula 8) and the disappearance of [³H]triglyceride (formula 7).

Means ± SE for arterial and venous plasma (n = 7 each) were calculated for each subject, and a P value of <0.05 was required for statistical significance. The concentrations, radioactivities, and specific activities of oleate, glycerol, and triglyceride, as well as fractional release of oleate and glycerol between venous and arterial plasma, were compared using Student’s paired t test, as was the systemic versus forearm fractional release of oleate and glycerol.

Arterial and venous plasma triglyceride concentrations and radioactivity are shown in Fig. 2. Venous plasma triglyceride concentrations were significantly lower than arterial values (0.811 ± 0.116 vs. 0.830 ± 0.116 mmol/l, P < 0.005). Venous plasma triglyceride radioactivity was also lower than arterial (2,721 ± 961 vs. 3,856 ± 904 dpm/ml, P < 0.005). Forearm FE of whole plasma triglyceride and ³H-triglyceride was 2.6 ± 0.6 and 31 ± 4%, respectively (Table 1). Forearm total triglyceride uptake was 34 ± 10 nmol · 100 g⁻¹ · min⁻¹ (Table 1).

Arterial and venous plasma glycerol concentrations were similar, averaging 40 ± 45 μmol/l (Fig. 3). In contrast, venous oleate concentrations were lower than arterial (126 ± 19 vs. 151 ± 22 μmol/l, P < 0.05, Fig. 3). Plasma oleate specific activities are shown in Table 2. Venous ¹⁴C oleate specific activity was lower than arterial (4.2 × 0.5 vs. 6.6 ± 0.9 dpm/nmol, P < 0.001). Venous ³H oleate specific activity was slightly, but not significantly, lower than arterial (4.3 ± 1.0 vs. 4.9 ± 1.1 dpm/nmol, NS). Plasma glycerol specific activities are depicted in Table 2. Venous ¹⁴C glycerol specific activity was lower than arterial (5.0 ± 0.9 vs. 11.4 ± 1.6 dpm/nmol, P < 0.001), whereas there was no difference between venous and arterial ³H glycerol specific activity (7.4 ± 2.9 vs. 7.4 ± 2.2 dpm/nmol, NS).

Forearm blood flow was 2.77 ± 0.27 ml · 100 g⁻¹ · min^−1, and forearm plasma flow was 1.63 ± 0.14 ml · 100 g⁻¹ · min⁻¹. Systemic R_a and MCR for oleate and glycerol are shown in Table 1. Glycerol MCR was higher than oleate MCR (P < 0.005). Forearm uptake, release, and FE of glycerol and oleate are also shown in Table 1. It can be estimated from the specific activity of the infused lipid emulsion (409 ± 55 dpm/nmol, data not shown), the arterial and venous ³H-triglyceride concentrations (3,856 ± 904 and 2,721 ± 962 dpm/ml, respectively) and the forearm triglyceride uptake (18.8 ± 4.5 μmol/l) that the contribution of emulsion particles to forearm triglyceride uptake was ∼0.0028 mmol/l, or ∼10% of total uptake. Forearm FE was higher for glycerol than for oleate (P < 0.02). If one assumes that 75% of forearm tissue is skeletal muscle and that muscle is 25% of body weight, one can extrapolate from forearm glycerol uptake to a value for whole-body skeletal muscle glycerol uptake [(66 nmol · 100 g muscle⁻¹ · min⁻¹/0.75) × 0.25 × 73,700 g body wt = ∼16 μmol/min] that is ∼14% of whole-body glycerol turnover.

Fractional release of LPL-generated (i.e., ³H) oleate and glycerol, shown in Fig. 4, were calculated by dividing actual systemic or forearm release by expected systemic or forearm release. Expected release represents the rate of release that would be observed if there were no local sequestration of LPL-generated [³H]oleate or glycerol and is equal to the amount of radiolabeled triglyceride that is taken up by the forearm (see formulae 5–7 above). Systemic and forearm fractional release of ³H-glycerol were similar (51 ± 4 vs. 59 ± 1%, NS). For oleate, however, forearm fractional release was lower than systemic (14 ± 2 vs. 36 ± 4%, P < 0.0001).

The present study indicates that a radiolabeled lipid emulsion can be used effectively to investigate the systemic and regional trafficking of LPL-generated glycerol and fatty acids in humans. We have previously demonstrated that the triglyceride-rich particles in our labeled lipid emulsion are metabolized similarly to native chylomicrons (10). A strength of this method is that the labeled lipid emulsions serve as near-massless tracers, which, when used in combination with independent FFA and glycerol tracers, can be used to partition the lipolytic products of LPL action from the lipolytic products of HSL action. Our current results indicate that chylomicrons are a potentially significant source of plasma FFAs in healthy volunteers. We also found that there is significant uptake of triglyceride-derived glycerol in tissues, without prior release into the systemic circulation. The results further suggest that there may be tissue-specific differences in the efficiency of LPL-mediated fatty acid storage.

The chief role of endothelium-bound LPL is the mediation of fat storage (2). In our study, we calculated the fractional release of fatty acids from a radiolabeled triglyceride substrate as a reflection of the relative efficiency of this process. Systemically, ∼36% of LPL-generated fatty acids were released into the circulation, suggesting that fat storage is not a completely efficient process. This apparent inefficiency may be important in redistributing FFAs to tissues that lack LPL, such as liver. It also may play a role in maintaining a small supply of FFAs as fuel for tissues that depend on FFAs for energy (liver, myocardium) during the postprandial period, when insulin suppresses FFA availability (18) via its action on HSL in adipose tissue (19). We found markedly lower fractional FFA release by forearm LPL (∼14%), indicating that the relative efficiency of LPL-mediated fatty acid storage in forearm tissue is greater than in the whole body. Since adipose tissue is underrepresented (and skeletal muscle overrepresented) in the forearm compared with whole body fat and muscle mass, and since the majority of systemic LPL activity resides in adipose tissue (2), these results suggest that the uptake of LPL-generated fatty acids is less efficient in adipose tissue than in muscle. It should be emphasized that these observations were made under postabsorptive conditions. LPL in adipose tissue and skeletal muscle are regulated reciprocally; specifically, LPL activity increases in adipose tissue and decreases in skeletal muscle after meals (2). These changes are mediated in part by increases in plasma insulin concentrations, and it is possible that insulin could influence the efficiency of fat storage. Thus, since our studies were not done in the postprandial state, they may not be entirely relevant to the storage of fat that occurs after meal ingestion.

The finding that ∼36% of LPL-generated fatty acids are released from tissues without initial storage are similar to previous data in humans. Roust and Jensen (6) gave a meal containing ¹⁴C-triolein to humans and found that ∼33% of the ¹⁴C-labeled oleate appeared in the systemic circulation. This suggests that our intravenous radiolabeled lipid emulsion is metabolized by LPL in a fashion similar to native chylomicrons, an interpretation consistent with a previous report from our laboratory (10). Evans et al. (7) recently reported results that were qualitatively similar to our findings, with some quantitative differences. Healthy subjects received a mixed meal containing ¹³C-labeled oleic acid and palmitic acid. Blood samples were taken from an arterialized hand vein and from veins draining abdominal adipose tissue and the forearm. The fractional release of LPL-generated fatty acids from adipose tissue varied from ∼20% in the early postprandial period to ∼80% at 6 h after the meal. The fractional release from the forearm was essentially zero, although it was not possible to partition uptake of arterial FFAs and release of LPL-generated FFAs because an independent FFA tracer was not administered (7). A strength of the Evans study was the fact that the study was conducted during meal absorption. The general observation that LPL-mediated fatty acid uptake is more efficient in the forearm than in adipose tissue is concordant with our results in postabsorptive subjects showing that the spillover of LPL-generated fatty acids is greater systemically than in the forearm.

These data in humans contrast with results of studies in rodents, however. Olivecrona injected chylomicrons labeled with ³H in the fatty acid moiety into rats and found that only ∼10% of the triglyceride-fatty acid isotope appeared in plasma FFAs, but a considerable amount of ³H was recovered in adipose tissue and myocardium (4). Teusink et al. (5) recently reported studies in which both lipid emulsion and chylomicrons labeled with ³H-triolein were given intravenously to anesthetized mice, and the appearance of ³H-oleate was traced with ¹⁴C-oleate. The results of experiments using the lipid emulsion and chylomicrons were concordant, suggesting that 80–90% of triglyceride fatty acids traverse the plasma FFA pool (5). The explanation for the discrepancy between these results and other studies is not apparent. Olivecrona (4) found an ∼10% escape of chylomicron fatty acids into plasma in rats, whereas our results and those of others (6,7) suggest that a higher figure is operative in humans. It is possible that the study of anesthetized animals influenced the results, since general anesthesia has a marked effect on FFA metabolism (20). However, the inconsistencies may relate more to species differences. Human chylomicrons are reported to have particle diameters of 316–352 nm (10,21,22); this correlates well with commercial lipid emulsions, which have particle sizes in the 286- to 340-nm range (10,23). In contrast, rat chylomicrons have been found to have mean diameters of 130–160 nm (24–26). Whether differences in chylomicron particle size could account for differences in the trafficking of LPL-generated fatty acids is not clear.

Although there appears to be general agreement that plasma triglycerides, specifically chylomicrons, contribute to circulating FFAs, recent studies in mice have suggested that the distinction between triglyceride-derived fatty acids and plasma FFAs cannot be made (5). Our data show substantial systemic sequestration of triglyceride fatty acids without appearance in plasma FFAs, indicating that the distinction is both important and measurable in humans. We have previously reported that postabsorptive FFA uptake is substantial in forearm tissue but negligible in subcutaneous adipose tissue in postabsorptive subjects (8). Taken together, these observations suggest that FFAs and triglyceride fatty acids are partitioned in tissues. Such partitioning could be due to differential transport of FFAs and triglyceride fatty acids through the interstitial space; labeled FFAs are present in fairly high concentrations in lymph during constant tracer infusion in animals (27).

Available evidence suggests that VLDL triglyceride, unlike chylomicron triglyceride, is not a significant source of plasma FFAs. Wolfe et al. (3) administered biosynthetic doubly-labeled VLDLs to dogs (harvested from donor dogs who had been given [³H]glycerol and [¹⁴C]palmitate); they found that >95% of VLDL-triglyceride fatty acids were taken up directly into tissues without traversing the plasma FFA pool. In recent unpublished studies, VLDL labeled with triolein, using both in vivo and ex vivo labeling techniques, was given to normal volunteers and no detectable labeled FFAs were found in plasma (M.D. Jensen, personal communication). Even if 5% of LPL-generated fatty acids from VLDL escaped into the systemic circulation (assuming complete triglyceride hydrolysis by LPL), it would account for only ∼0.5% of total FFA appearance, based on VLDL triglyceride turnover rates of ∼0.2 μmol · kg⁻¹ · min⁻¹ (28) and FFA flux rates of ∼6 μmol · kg⁻¹ · min⁻¹ (8) in normal subjects. The relative contribution of VLDL to total FFA appearance would be expected to increase during the postprandial period because of suppression of HSL (6) if VLDL triglyceride hydrolysis continued at the same rate. However, a 75–90% decrease in VLDL triglyceride hydrolysis occurs during mixed-meal absorption, due to competition for LPL by chylomicrons (29). Although VLDL secretion rates are increased to ∼0.4 μmol · kg⁻¹ · min⁻¹ in obese nondiabetic (30) and type 2 diabetic (31,32) individuals, FFA appearance is also increased in the postabsorptive (18,33,34) and postprandial (6,18) states in these individuals. Thus, the relative contribution of VLDL would remain low in normal and insulin-resistant subjects under both postabsorptive and postprandial conditions.

We found that the fractional release of LPL-generated glycerol was 50–60%, with no difference between the systemic and forearm values. The substantial (∼40%) uptake of triglyceride-derived glycerol in the forearm is consistent with our previous work (8) and indicates the presence of glycerol kinase in skeletal muscle. However, the observation that glycerol generated by the action of LPL on triglyceride-rich lipoproteins is substantially sequestered in tissues without systemic release contrasts with previous studies in animals showing near-quantitative mixing with the plasma glycerol pool, both in the case of VLDL and with chylomicrons (3,4). Species differences may explain these discrepant findings.

Our results demonstrate significant uptake of triglyceride by the forearm, which in normotriglyceridemic postabsorptive subjects is due almost entirely to LPL’s interaction with VLDL (35) The contribution of emulsion particles to forearm triglyceride uptake was ∼10% of total uptake. The fractional extraction of radiolabeled triglyceride in the forearm was 31%, which underestimates the avidity of chylomicron triglyceride extraction because the calculation was made on whole plasma triglyceride radioactivity, which includes chylomicron remnants that do not interact with LPL. Nonetheless, it amply demonstrates the preference LPL has for chylomicrons compared with VLDL, as shown previously (16).

The implications of our results concerning the contribution of dietary fat to FFA turnover are considerable and can be estimated. Assuming for a 70 kg adult 1) postabsorptive FFA flux of ∼6 μmol · kg⁻¹ · min⁻¹ (8); 2) postprandial FFA flux averaging about half of postabsorptive values (18), prevailing for at least half of the 24-h period; 3) dietary fat intake equaling 35% of a daily energy intake of ∼2,000 kcal; and 4) fractional systemic release of chylomicron derived fatty acids of 36%, as much as 20–25% of total FFAs may come from LPL. Thus, it is possible that dietary fat content could have a nearly around-the-clock impact on ambient FFA levels. The proportion of FFAs derived directly from triglyceride-rich lipoproteins could be even higher if the relative efficiency of LPL-generated fatty acid storage were impaired in certain disease states. An increased contribution of LPL to the FFA pool could result in FFA concentrations even higher than would occur solely with the increased HSL-generated FFA release that occurs in insulin-resistant states, such as obesity and type 2 diabetes. This could have important pathophysiological consequences, considering that elevated FFAs contribute to systemic insulin resistance (36), impair endothelial function (37), stimulate the sympathetic nervous system (38), and acutely raise blood pressure (39).

In summary, the results of the present study indicate that there is escape, or spillover, of the lipolytic products of LPL action on triglyceride-rich lipoproteins in humans. This suggests that LPL-mediated fat storage is somewhat inefficient. Additional studies will be needed to determine whether the efficiency of fat storage is altered in conditions such as obesity and diabetes.

This study was resented in part at the annual meeting of the American Diabetes Association, San Francisco, California, June 2002, and was supported by a grant from the Saint Luke’s Hospital Foundation. Y.S.P. was supported by an American Heart Association postdoctoral fellowship.

We thank Dr. D. Paramore for advice on blood flow measurements; Drs. R. Hudson and R. Nelson for assistance with arterial cannulation; and T. al-Maliky, H Bunyard, and D. Klatt for technical assistance.