Triglyceride Uptake and Lipoprotein Lipase–Generated Fatty Acid Spillover in the Splanchnic Bed of Dogs Free

Nine adult mongrel dogs (average weight 22.2 kg) were studied after a 42-h fast. The animals’ care was in accordance with National Institutes of Health guidelines for the care and use of laboratory animals. The protocol was approved by the Vanderbilt University Institutional Animal Care and Use Committee.

Each dog underwent a laparotomy under general anesthesia 14–16 days before the experiment. Catheters (0.04-inch internal diameter) for blood sampling were placed into the left hepatic vein, the hepatic portal vein, and left femoral artery as previously described (10), together with transonic flow probes around the portal vein and hepatic artery. All catheters were filled with heparinized saline (200 units/ml; Abbott Laboratories, North Chicago, IL), and their free ends were knotted and then tunneled to a subcutaneous pocket before closure of the skin. The animals were fed standard dog chow and were weight stable before study.

[1-¹⁴C]oleate (GE Healthcare) was prepared for infusion as previously described (11). A commercial lipid emulsion (10% Intralipid) was labeled with [9,10-³H]triolein (Perkin Elmer) and heat sterilized for infusion (12).

On the day of study, the blood sampling catheters were exteriorized and carefully aspirated to clear residual heparinized saline. Infusions of [9,10-³H]triolein (∼1.2 μCi/min) and [1-¹⁴C]oleate (∼0.3 μCi/min) were started in a peripheral vein 90 and 60 min, respectively, before the first blood samples were taken from the artery, the portal vein, and the hepatic vein. These time frames were chosen to establish tracer equilibrium; the tracer infusions were continued until blood sampling had been completed. The radiolabeled lipid emulsion is of sufficiently high specific activity that it does not affect plasma triglyceride concentrations when infused (2). An arterial blood sample was drawn before the tracer infusions to serve as a blank. Subsequently, five matched arterial, portal vein, and hepatic vein blood samples were taken at 7.5-min intervals for a total of 30 min for measurement of plasma triglyceride radioactivity and FFA concentration and specific activity. Plasma flow was calculated from Doppler probe and hematocrit.

Blood samples were immediately transferred to chilled 10-ml EDTA tubes containing 0.5-mg of paraoxon to inhibit LPL (13) and kept on ice until centrifugation at 4°C. The resulting plasma was then frozen at −70⁰ C until analysis. Modification of a previously published procedure (11) allowed measurement of plasma triglyceride radioactivity, FFA concentration, and specific activity on the same aliquot of plasma. Fifty microliters of ∼2.0 mmol/l [²H₃₁]palmitate was added to 1.0 ml plasma (14). A lipid extraction was then performed using the method of Dole and Meinertz (15). The organic layer was then back-extracted twice with 0.02 N NaOH to separate FFA from triglyceride radioactivity. The organic layer was dried down under a stream of nitrogen and resuspended in 10 ml Optiflor for determination of triglyceride radioactivity by liquid scintillation counting. Triglyceride radioactivity was corrected for recovery, which averaged ∼95%, and also for incomplete removal of radiolabeled FFA from the triglyceride fraction, which averaged 5% of FFA radioactivity. The FFA (aqueous) layer was then acidified with 20 μl 3 N HCl, dried under nitrogen, and derivatized as previously described for analysis of FFA using high-performance liquid chromatography (HPLC) (11). The HPLC method was modified for measurement of total FFA using a 15-cm 4μm Waters C18 Nova-Pak column eluted with 78% acetonitrile in water and connected to a Waters 2487 UV detector (260 nm). With this modification, oleate elutes at 14–15 min. The oleate peak was collected for measurement of [³H] and [¹⁴C] radioactivity by liquid scintillation spectrometry. Baseline arterial plasma triglyceride concentrations in each animal were determined using a Sigma TR0100 kit.

All calculations were made using steady-state assumptions. Systemic spillover of triglyceride-derived oleate (estimated from [³H]oleate generation) was calculated from oleate concentrations, [³H] specific activities and [¹⁴C] specific activities in arterial plasma, as previously described (2). The availability of portal venous blood sampling in this study allowed the separate determination of spillover in two components of the splanchnic bed: the liver and nonhepatic tissues. A cartoon depicting the mathematical basis for estimating spillover in tissue beds is shown in the online appendix (available at http://dx.doi.org/10.2337/db06–1657).

Net splanchnic spillover was calculated from the following formulas, which do not take the portal venous circulation into consideration. These formulas use arterial and hepatic venous values, analogous to determination of spillover in the human forearm (2).

Net fractional extraction of oleate for the splanchnic bed was calculated from the formula

where SA is specific activity, C is concentration, and subscripted A and H refer to arterial and hepatic venous plasma, respectively.

The net rate of [³H]oleate release in the splanchnic bed that would be expected if there were no local uptake of triglyceride-derived oleate (i.e., if fractional spillover were 100%) was calculated from the uptake of [³H]triglyceride from arterial plasma according to the formula

where [³H] TG equals the concentration of [³H]triglyceride in dpm · mL⁻¹, PF_P equals portal venous plasma flow, and PF_A equals hepatic artery plasma flow.

The actual net release of triglyceride-derived [³H]oleate in the splanchnic bed was calculated from the formula

where the expression ([³H]SA_A × C_A × [1-FE_S]) equals the predicted concentration of [³H]oleate in hepatic venous plasma if fractional spillover of triglyceride-derived oleate is zero.

Fractional extraction of oleate for nonhepatic splanchnic tissues was calculated from the formula

The rate of [³H]oleate release in nonhepatic splanchnic tissues that would be expected if there were no local uptake of triglyceride-derived oleate (i.e., if fractional spillover were 100%) was calculated from the uptake of [³H]triglyceride from arterial plasma according to the formula

The actual release of triglyceride-derived [³H]oleate in nonhepatic splanchnic tissues was calculated from the formula

Hepatic spillover was calculated from the following formulas, which take into account that organ's dual blood supply.

Fractional extraction of oleate for the liver was calculated from the formula

The rate of [³H]oleate release in the liver that would be expected if there were no local uptake of LPL-generated oleate was calculated as follows:

The actual release of triglyceride-derived [³H]oleate in the liver was calculated from the formula

Fractional spillover of [³H]oleate for the splanchnic bed, nonhepatic tissues, and the liver was calculated by the formula

using the respective R and E values for the three tissue beds.

“True” fractional spillover for the splanchnic bed was calculated from nonhepatic splanchnic and hepatic spillover, the contribution from each tissue bed weighted for the relative distribution of spillover in the two tissue beds:

Nonsplanchnic spillover was calculated from the difference between systemic and true splanchnic [³H]triglyceride uptake and [³H]oleate appearance. The contribution of nonhepatic splanchnic and hepatic tissues to systemic [³H]triglyceride disappearance was calculated by dividing [³H]triglyceride uptake (equal to ER_NH and ER_H, respectively, in Eqs. 5 and 8 above) by the [³H]triolein infusion rate. Because whole plasma triglyceride radioactivity was measured, an estimate of the contribution of hepatic extraction of labeled emulsion remnants to total hepatic extraction of labeled triglyceride was made. Based on the half-life of large emulsion particles and remnants, this contribution was estimated to be 0.8%. A detailed description of this calculation is provided in the online appendix.

Systemic oleate and FFA flux was calculated as previously described (16). Oleate uptake and release were calculated for the liver and for nonhepatic splanchnic tissues as previously described in dogs (17) and extrapolated to total FFA uptake and release using the ratio of total FFA:oleate concentrations (16). The contribution of nonhepatic and hepatic tissues to systemic oleate uptake and release was determined by dividing oleate uptake and release by systemic oleate flux.

Data are expressed as means ± SE. Mean values were calculated by averaging results from the five samples taken from each site. Comparisons between site values in the group of subjects were analyzed by a t test for paired samples for means using the Excel (Microsoft) data analysis package. Correlations were tested for significance using Pearson's correlation coefficient. The relationship between net and true splanchnic spillover was tested for significance with Spearman's rank correlation coefficient, using JMP 6.0.0 by SAS Institute. A P value of <0.05 was required for statistical significance in all cases.

A total of 11 animals were studied, but usable data were available on only 9, because steady-state conditions were not achieved in two of the animals. Baseline arterial plasma triglyceride concentrations were 65 ± 6 mg/dl (data not shown). Plasma [³H]triglyceride concentrations are shown in Fig. 1. Portal venous [³H]triglyceride levels (7,545 ± 882 dpm/ml) were significantly lower than arterial (8,609 ± 1,081 dpm/ml) (P = 0.001). Hepatic venous concentrations (6,493 ± 1,008 dpm/ml) were lower than both arterial and portal venous values (P < 0.01). Fractional extraction of [³H]triglyceride was 26.0 ± 3.5% for the total splanchnic bed, 11.8 ± 1.3% for nonhepatic tissues, and 17.7 ± 4.3% for the liver. Plasma total FFA and oleate concentrations are shown in Table 1 together with plasma oleate specific activities. Portal venous oleate and FFA concentrations were higher than arterial (374 ± 36 and 934 ± 74 vs. 329 ± 30 and 825 ± 68 μmol/l, respectively, both P ≤ 0.02). Hepatic venous oleate and FFA concentrations (305 ± 26 and 755 ± 55 μmol/l, respectively) were lower than both arterial (P < 0.01) and portal (P < 0.001) concentrations. Portal and hepatic [¹⁴C]oleate specific activities were lower than arterial (2.10 ± 0.25 and 2.14 ± 0.23, respectively, vs. 2.47 ± 0.30 dpm/nmol (P ≤ 0.01 for both comparisons). [³H]oleate specific activity was higher in the hepatic vein than either the artery or portal vein (8.57 ± 0.69 vs. 7.03 ± 0.58 and 7.04 ± 0.0.68 dpm/nmol, respectively; P < 0.005 for both).

Plasma flow was 60 ± 6 ml/min in the hepatic artery and 314 ± 29 ml/min in the portal vein (data not shown). Table 2 shows systemic and regional kinetics of FFAs and labeled triglycerides. Splanchnic uptake of [³H]triglyceride was 33 ± 7% of systemic triglyceride disappearance; 45 ± 9% of splanchnic uptake occurred in nonhepatic tissues, and 55 ± 9% in the liver. The splanchnic bed, using net calculations (without portal vein data), accounted for 12 ± 2% of systemic oleate uptake and 6 ± 2% of systemic oleate appearance (not shown). Nonhepatic splanchnic tissues accounted for 19 ± 4% of FFA delivery to the liver. Fractional spillover data are shown in Fig. 2. Systemic, nonsplanchnic, and true splanchnic spillover were 48 ± 5, 62 ± 20, and 41 ± 7%, respectively. Spillover was 50 ± 10 and 45 ± 9% in nonhepatic splanchnic tissues and liver, respectively. Correcting for spillover, nonhepatic tissues and the liver accounted for 6.2 and 9.6%, respectively of the systemic disappearance of [³H]triglyceride fatty acids (data not shown). Net splanchnic spillover (32 ± 5%) was slightly, but not significantly, lower than true splanchnic spillover (P = 0.07). There was a significant correlation between net and true splanchnic spillover (ρ = 0.84, P < 0.005; Fig. 3).

There was a significant correlation between nonhepatic splanchnic spillover and nonhepatic FFA release (ρ = 0.81, P < 0.01; Fig. 4). There was also a significant correlation between nonhepatic spillover and portal venous FFA concentrations (R = 0.80, P < 0.01; data not shown).

The present study demonstrates significant spillover of LPL-generated triglyceride fatty acids in both nonhepatic splanchnic tissues and in the liver of fasting dogs. Systemic spillover in the animals was ∼48%, slightly higher than the 36% spillover previously reported in overnight fasted humans (2). The majority of systemic spillover likely occurs in adipose tissue, where fractional spillover is higher than in skeletal muscle (2,5). The triglyceride uptake and fatty acid spillover in nonsplanchnic tissues was expected and is presumed to occur in visceral fat, which has abundant LPL (18).

There was a strong correlation between fractional spillover and FFA release in nonhepatic splanchnic tissues in this study. Presumably, these two processes reflect LPL and HSL activity, respectively, in visceral adipose tissue. A link between spillover and FFA release has not previously been demonstrated. It is possible that the net flow of fatty acids may be governed by the concentration gradient between adipose tissue and the extracellular fluid. A graphic model of this process is shown in Fig. 5. In the fed state, intracellular lipase activity is low, and the flow gradient is inward, promoting efficient uptake of LPL-generated fatty acids. Presumably, the vast majority of these fatty acids taken up by adipose tissue are directed toward esterification and ultimately triglyceride storage. In the fasting state, intracellular lipase activity is higher, and the gradient of flow tends to be outward, promoting spillover. Such a reciprocal relationship between intracellular lipolysis and the efficiency of LPL-mediated fat storage would not necessarily mean that the direction of fatty acid flow is the only factor that affects spillover.

If this model accurately depicts the relationship between the efficiency of LPL-mediated fat storage and intracellular lipase activity, it does not explain our data on triglyceride uptake by the liver. This finding was unexpected but not unprecedented. Hepatic uptake of triglycerides from triglyceride-rich, chylomicron-like lipid emulsion particles has previously been reported in rats (19) and also in sheep and dogs (20). In rats, hepatic uptake was considerably lower in fed compared with fasting animals (19). However, spillover was not measured. In another study in rats, chylomicrons and a lipid emulsion, each containing labeled triolein, were administered intravenously to rats; with both tracers, fractional hepatic uptake of labeled triglyceride fatty acids was greater on a low fat diet than on a high fat diet (21). Considering that the chylomicron-sized particles from the lipid emulsion in the present study were administered to fasting animals in trace quantities, the observed hepatic uptake of labeled triglyceride is of uncertain relevance to meal fat disposal. Because the liver does not contain significant LPL activity (18), triglyceride uptake from triglyceride-rich lipoproteins in that tissue is due to another lipase, the identity of which is unclear. Hepatic lipase is unlikely to be responsible, because it resides within the Space of Disse, from which chylomicron-sized particles are excluded by the hepatic sieve (22,23). In the present studies, there was significant spillover in the liver of similar magnitude to that observed in nonhepatic tissues, and yet there was essentially no release of unlabeled FFA into the hepatic vein. This suggests that the microanatomic relationship in the liver between the extracellular lipase(s) responsible for triglyceride hydrolysis and hepatocytes is fundamentally different than the relationship between LPL and adipocytes.

Net splanchnic spillover in our study was calculated from arterial and hepatic venous data, without measurements from the portal vein. There was a significant correlation (P < 0.005) between net spillover and true spillover. Also, net spillover was somewhat lower than “true” spillover (calculated as the average spillover in nonhepatic tissues and the liver when portal venous measurements were used), although the difference did not reach statistical significance. This indicates that measurement of net spillover in humans, where access to the portal vein is not possible, could provide significant underestimates of true events.

Nonhepatic spillover correlated significantly with portal venous FFA concentrations. Although it has previously been shown that subcutaneous, not visceral lipolysis, is the chief source of FFA in the portal vein (7), this correlation may be the consequence of a strong linkage between subcutaneous and visceral lipolysis.

In this study, visceral lipolysis (i.e., nonhepatic splanchnic FFA release) accounted for ∼19% of hepatic FFA delivery. This is a somewhat higher proportion than reported in an earlier study (17) of 18-h–fasted dogs (∼11%); the discrepancy between the two studies likely relates to the fact that the animals in the present study had been fasted for 42 h and, as a result, had higher plasma FFA concentrations than 18-h–fasted dogs.

The splanchnic bed accounted for nearly one-third of systemic triglyceride uptake in our study. Corrected for spillover, nonhepatic tissues (presumably visceral fat) were responsible for ∼6% of the systemic disappearance of radiolabeled triglyceride fatty acids. This is very similar to the ∼8% fractional meal fat disposal in visceral fat previously reported in humans (6) and suggests that the radiolabeled lipid emulsion is metabolized by LPL very similarly to native chylomicrons. We previously reported that infusion of a radiolabeled lipid emulsion accurately detects the absorption of dietary fat (12), which is also consistent with this interpretation.

Redgrave and Maranhao (24) have demonstrated that triglyceride clearance from lipid emulsion particles that lack cholesterol is nearly identical to triglyceride clearance from endogenous chylomicrons in rats; on the other hand, they found that cholesterol content in the emulsion particle is a critical determinant of the metabolic behavior of particle remnants. Another study in dogs strongly supports the usefulness of the radiolabeled lipid emulsion as a surrogate tool in the study of chylomicron triglyceride metabolism. Bergman et al. (20) harvested chylomicrons from donor dogs who had ingested [³H]palmitate and administered the radiolabeled chylomicrons simultaneously with [¹⁴C]palmitate intravenously to recipient postabsorptive dogs, sampling arterial, portal venous, and hepatic venous blood. The authors reported “percent chylomicron uptake released as FFA ” analogous to percent spillover. The results of that study using endogenously labeled chylomicrons were very similar to the respective results of the present study in terms of triglyceride fractional extraction in nonhepatic splanchnic tissues (∼9 vs. ∼12%), fractional extraction in the liver (∼22 vs. ∼18%), nonhepatic splanchnic spillover (∼48 vs. ∼50%), and hepatic spillover (∼39 vs. ∼45%).

Potential limitations of these data should be acknowledged. It is theoretically possible that radiolabeled oleate could recycle through synthesis and secretion of VLDL triglyceride. However, previous studies in dogs (25) and recent studies in humans (26) indicate that spillover of fatty acids from LPL hydrolysis of VLDL is very low to negligible. The explanation for lower spillover from VLDL compared with chylomicrons is not certain, but it may be related to kinetic differences, considering that LPL has a much higher affinity for chylomicrons compared with VLDL (2,27–29). There is uncertainty regarding whether data obtained in fasting dogs is relevant to the circumstance of meal absorption in humans. It is reassuring that systemic spillover in postabsorptive dogs (∼49% in the present study) is similar to values of 36–48% previously reported in postabsorptive humans (2,28). A recent study indicates that spillover in human adipose tissue ranges from 13% at 1 h to 52% at 6 h after meal ingestion (29).

In summary, there is significant spillover of fatty acids derived from circulating triglycerides in nonhepatic splanchnic tissues and in the liver in 42-h–fasted dogs. In nonhepatic tissues, which presumably represent visceral fat, spillover correlates with lipolysis or the release of FFA into the portal vein. Thus, efficient storage of dietary fat may require adequate suppression of adipose tissue lipolysis, which typically depends on the antilipolytic effects of the increased insulin concentrations that are observed after meal ingestion. Considering that the amount of dietary fat that traverses the circulation can be similar to systemic FFA flux (2), the contribution of spillover of fatty acids from chylomicrons to total portal FFA is of major potential importance. Abnormally high spillover rates in visceral fat could contribute to both hepatic insulin resistance and increased VLDL triglyceride production by raising postprandial portal venous FFA concentrations (9,30). Ultimately, the relevance of our observations in fasted dogs to meal fat disposal is uncertain and will require further study.

This study has received support from the U.S. Public Health Service (Grants HL67933 and DK18243) and from the Mayo Foundation.

We thank Tiffany Rodewald, Doss Neal, and Danielle Vlazny for technical assistance.