Lipid composition of HDL subfractions in dog plasma and lymph.

We report the lipid composition of dog plasma and peripheral lymph lipoproteins as separated into pre-beta, alpha, and pre-alpha fractions by agarose gel electrophoresis. Plasma lipoproteins with alpha mobility have a composition different from that of plasma lipoproteins with pre-alpha mobility, having 9% versus 11% free cholesterol, 21% versus 17% cholesterol ester, 1% versus 16% triacylglycerol, and 69% versus 56% phospholipid. On the other hand, lymph alpha and pre-alpha lipoproteins have compositions that are quite similar (9% versus 7% free cholesterol, 17% versus 17% cholesterol ester, 2% versus 4% triacylglycerol, and 71% versus 71% phospholipid). The lipid compositions of plasma and lymph alpha lipoproteins are quite similar (9% versus 9% free cholesterol, 21% versus 17% cholesterol ester, 1% versus 2% triacylglycerol, and 70% versus 72% phospholipid). The lipid compositions of plasma and lymph pre-alpha lipoproteins are different (11% versus 7% free cholesterol, 17% versus 17% cholesterol ester, 16% versus 4% triacylglycerol, and 56% versus 71% phospholipid). Peripheral lymph lipoproteins with pre-beta mobility contained 15% cholesterol, 13% cholesterol ester, 10% triacylglycerol, and 61% phospholipid. Compared with plasma, peripheral lymph lipoproteins are free cholesterol-enriched in all fractions. Calculated stoichiometric ratios of lipid to apoA-I per particle, alpha lipoproteins have two molecules of apoA-I per particle, and pre-alpha lipoproteins have four molecules of apoA-I per particle.

Chronic peripheral lymphatic cannulation in the dog.

Peripheral lymph collected acutely has been commonly sampled as representative of non-visceral interstitial fluid. By developing a prenodal lymphatic-lymphatic (L-L) shunt, we were able to collect peripheral lymph for 3-5 days in unanesthetized dogs. The L-L shunt was constructed entirely of medical grade silicone rubber tubing designed with a slip of coupling which allowed the shunt to be disconnected for lymph collection and reconnected at night. Average peripheral lymph flow (4.9 ml/hr leg) in unanesthetized dogs was almost twice the flow rate previously observed in anesthetized dogs. The average lymph/plasma total protein concentration ratio (0.16), however, was similar to that previously found in anesthetized dogs. Lymph protein concentration fell with the collection during the day and became more concentrated at night. Lymph flow did not change greatly during daytime collection. Average peripheral lymph collection volume was greater than 200 ml/dog. The L-L shunt allows collection of prenodal-lymph in experiments where unanesthetized dogs are required (e.g., feeding studies). They also are useful when multiple protocols are conducted on the same dog or when large volumes of peripheral lymph are required.

Two-dimensional electrophoresis of plasma lipoproteins: recognition of new apo A-I-containing subpopulations.

Two-dimensional electrophoresis has been used to resolve 12 distinct apo A-I-containing high-density lipoprotein (HDL) subpopulations in human plasma. The subpopulations were quantitated by 125I-labeled, monospecific antibody and phosphor-imaging. Modification and standardization of the agarose electrophoresis (first dimension) enabled us to recognize new HDL subpopulations. Lipoprotein mobilities in agarose were expressed relative to the mobility of the sample's endogenous albumin. We demonstrated the presence of lipoproteins with mobilities faster than and similar to albumin, as well as subpopulations with mobilities slower than albumin. We refer to these as pre alpha, alpha and pre beta, respectively. Lipoprotein molecular sizes were determined with a non-denaturing polyacrylamide gradient gel electrophoresis (PAGE) (2% to 36%) in the second dimension. Internal standard of 125I-labeled proteins of known molecular size was run simultaneously in each gel permitting accurate size determination. We have demonstrated that ultracentrifugally-isolated lipoproteins are different from the native apo A-I-containing subpopulations. The major difference observed was the loss of pre beta 1 and pre beta 2 particles from the d < 1.21 g/ml fractions to the d > 1.21 g/ml fractions. Possible physiologic and pathologic implications of these findings are also discussed.

Comparison of apo A-I-containing subpopulations of dog plasma and prenodal peripheral lymph: evidence for alteration in subpopulations in the interstitial space.

To study in vivo reverse cholesterol transport, dog plasma and lymph apo A-I-containing subpopulations were compared by two-dimensional electrophoresis. Charge and size of subpopulations were similar in plasma and lymph, but the distribution of subpopulations varied considerably. An increase in pre-beta and pre-alpha particles in lymph suggests these changes are a reflection of in vivo reverse cholesterol transport.

Comparison of the lipid and apolipoprotein composition of skeletal muscle and peripheral lymph in control dogs and in dogs fed a high fat, high cholesterol, hypothyroid-inducing diet.

Most studies of peripheral interstitial fluid lipoprotein composition have been made on interstitial fluid-derived from skin and connective tissue. We developed techniques which allowed simultaneous comparison of lymph (a model of interstitial fluid) from skeletal muscle and skin in control (C) and cholesterol-fed (CF) dogs. Lipoprotein fractions were separated by ultracentrifugation. Skeletal muscle interstitial fluid HDL concentrations were approximately twice those of skin. However, the concentration of VLDL-LDL particles was similar in both interstitial spaces. HDL particles from both microvascular beds showed evidence of extensive remodelling when compared to plasma HDL from the same animal. Relative to apo A-I, skeletal muscle HDL was enriched in free cholesterol and apo E (C and CF dogs) and apo A-IV (CF dogs). Skin-derived HDL was consistently enriched in free cholesterol, apo E and A-IV in both C and CF dogs. These studies indicate that similar remodeling of plasma HDL occurs in widely different tissues which together constitute approximately 70% of the total interstitial space. The relatively high concentration of plasma-derived and remodeled HDL within the interstitial space of skeletal muscle is consistent with that tissue's importance in reverse cholesterol transport.

Lipoprotein lipase and hepatic triacylglycerol lipase activities in peripheral and skeletal muscle lymph.

We studied the interstitial fluid concentration of two lipid-metabolizing enzymes (lipoprotein lipase and hepatic triacylglycerol lipase) to determine their importance in interstitial modification of filtered lipoproteins. Despite the use of a very sensitive lipase assay (1 nmol of fatty acid release/ml/hr), lipase activities in plasma and in peripheral and skeletal muscle lymph from control dogs were below the sensitivity of our assay. After heparin injection, hepatic triacylglycerol lipase and lipoprotein lipase activities in plasma were similar. However, the postheparin hepatic triacylglycerol lipase activities in peripheral and skeletal muscle lymph were only 1.4% and 1.1%, respectively, those of plasma. This concentration is considerably less than the lymph concentration of albumin, which has a similar size to the lipases but has a lymph concentration of 30% to 40% of plasma. Lipoprotein lipase activity in peripheral lymph and skeletal muscle lymph was 2.7% and 4.8%, respectively, of plasma activity. Since lipoprotein lipase has a similar size as hepatic triacylglycerol lipase, the disproportionate amount of lipoprotein lipase in lymph as compared to hepatic triacylglycerol lipase could be due to heparin crossing the capillary endothelium and displacing lipoprotein lipase from peripheral cells. Injection of radioactive heparin confirmed that it does cross into the interstitial space in sufficient concentrations to displace lipase from peripheral cells. We conclude that most of the lipase found in lymph after heparin injection is derived from peripheral cells and not from plasma. Furthermore, hepatic triacylglycerol lipase does not play a role in high density lipoprotein remodeling in interstitial fluid. Therefore, it seems likely that the considerable remodeling of high density lipoprotein that we found previously results from its interaction with peripheral cells.

A sensitive, inexpensive method for determining minute quantities of lipase activity.

We describe a sensitive and reproducible lipase assay based on the binding of 63Ni to fatty acid. This method can detect down to 1 nmol of fatty acid per milliliter of solution. It has been adapted for measuring low concentrations of lipoprotein lipase and hepatic triacylglycerol lipase. Furthermore, in the presence of tritiated triolein, the method is insensitive to radiolabel interference, even when the fatty acid is labeled.

Characterization of dog prenodal peripheral lymph lipoproteins. Evidence for the peripheral formation of lipoprotein-unassociated apoA-I with slow pre-beta electrophoretic mobility.

Dog plasma and prenodal peripheral lymph apoA-I distribution was examined by nondenaturing gradient gel electrophoresis-immunoblot analysis. In control dogs, plasma apoA-I could be localized to two distinct populations of particles with modal diameters of 8.4 nm and 10.4 nm. The smaller sized population accounted for over 50% of plasma apoA-I. Peripheral lymph apoA-I distribution was significantly different. The percentage of apoA-I localized to the 10.4 nm population was reduced by 40% and the modal diameter of the smaller HDL apoA-I population was significantly decreased by 0.1 nm. Additionally, peripheral lymph apoA-I could be localized to particles smaller than albumin (lipoprotein-unassociated apoA-I). The presence of lipoprotein-unassociated apoA-I particles was confirmed by gel filtration chromatography. Immunoblots of column fractions subjected to agarose electrophoresis revealed that these particles had slow pre-beta electrophoretic mobility. In dogs fed an atherogenic diet, lipoprotein-unassociated apoA-I particles with slow pre-beta electrophoretic mobility could be found in both plasma and peripheral lymph. With increasing degree of hypercholesterolemia, the relative amount of plasma lipoprotein-unassociated apoA-I tended to increase. In peripheral lymph, an increasing degree of hypercholesterolemia was associated with a decrease in the relative amount of lipoprotein-unassociated apoA-I. Instead, a population of large apoA-I particles (11-25 nm) became increasingly prominent.

Interstitial fluid lipoproteins.

While a wide variety of techniques has been used to collect samples of interstitial fluid, most of our detailed knowledge about the composition of interstitial fluid lipoproteins has come from lymph collection studies. The considerable variability of lymph data probably reflects the effect of variable metabolic modification and different capillary permeabilities on the lipoprotein composition of interstitial fluid. All density classes of plasma lipoproteins are present in lymph. In peripheral lymph, the lymph/plasma concentration ratios of lipoproteins vary from 0.03 for VLDL-sized particles to 0.2 for HDL. Lymph from more permeable vascular beds, such as lung and myocardium, contains proportionately more lipoproteins. Their lymph/plasma concentration ratios vary from 0.1 to 0.6. In general, lymph lipoproteins are more heterogeneous in size than their plasma counterparts. Lymph HDL and LDL contain larger and smaller particles than their plasma equivalents. Lymph lipoproteins have unusual shapes (square packing and discoidal), chemical compositions, and molecular charge, which suggest de novo formation and/or extensive peripheral modification. Lymph HDL and LDL are enriched in free cholesterol. Lymph HDL also has increased cholesterol/protein and phospholipid/protein (especially sphingomyelin) ratios (Sloop, C.H., L. Dory, and P.S. Roheim, unpublished observations). Lymph HDL apoprotein composition differs from that of plasma, with an increase in apoE and apoA-IV content relative to apoA-I. These discoidal HDL particles may be products of an initial stage of reverse cholesterol transport. We believe further study of their metabolic fate would give important information concerning the later stages of reverse cholesterol transport.

Peripheral synthesis and isoform distribution of dog apoprotein E. An in vivo approach.

Purified dog plasma apo-E is composed of four major isoforms with the following pI values: 5.40, 5.31, 5.26, and 5.22. Treatment with neuraminidase suggests that the multiple forms are due to progressive sialation. The acidic isoforms (pI = 5.22 and less), rarely detectable in plasma lipoprotein samples (except in the d less than 1.05 g/ml fraction of cholesterol-fed dogs), are present in high concentrations in the interstitial fluid high density lipoprotein fraction I (HDLI) of cholesterol-fed dogs, a lipoprotein recently described (Dory, L., Boquet, L.M., Hamilton, R.L., Sloop, C.H., and Roheim, P.S. (1985) J. Lipid Res. 26, 519-527). Apo-Es0 (the most basic isoform) is a major constituent of the plasma d less than 1.05 g/ml fraction, but it is usually only a minor component of other plasma or interstitial fluid lipoproteins. This is likely a reflection of the prolonged residence time of a specific lipoprotein species in this density range, resulting in more extensive desialation. Peripheral apo-E synthesis has been measured under in vivo conditions using a hindlimb cannulation of the lymphatics and collection of prenodal peripheral lymph. Following injection of [3H]leucine or [35S]methionine into the dorsal skin of the toes, the specific activity of the interstitial fluid apo-E far exceeded that of plasma apo-E at all time points examined. Incorporation was measurable 30 min after the isotope injection and peaked at 150 min in control dogs and between 120-150 min in cholesterol-fed dogs. The rate of peripheral synthesis of apo-E in cholesterol-fed dogs appeared to be twice that of control dogs. The newly synthesized and secreted apo-E preferentially associated with the disc-shaped HDLI of the interstitial fluid; less than 15% of the apo-E-associated radioactivity was recovered in the d less than 1.05 g/ml fraction, despite the fact that this fraction contains well over 50% of the total interstitial fluid apo-E mass. The newly secreted, HDLI-associated apo-E can be converted by neuraminidase into apo-Es0.

Heterogeneity of dog interstitial fluid (peripheral lymph) high density lipoproteins: implications for a role in reverse cholesterol transport.

The heterogeneity of dog interstitial fluid (peripheral lymph) high density lipoprotein (HDL) was investigated and compared to plasma HDL. Interstitial fluid and plasma HDL of normal and cholesterol-fed dogs was subfractionated by ultracentrifugation and affinity and molecular weight sieving chromatography. Both plasma (P) and interstitial fluid (L) HDL can be subfractionated into a larger fraction (P-I and L-I) and a smaller one (P-II and L-II). Cholesterol feeding induces a large increase in the P-I and L-I component of HDL, but the increase in L-I is far greater in proportion than that of P-I. Furthermore, L-I of cholesterol-fed dogs appears to be almost exclusively discoid in shape, while only approximately 15% of particles in P-I are discoidal. The discoid HDL of L-I is reflected in its chemical composition: 28% unesterified cholesterol, 6% cholesteryl ester, 45% phospholipid, and 21% protein. It contains large amounts of apoE in addition to apoA-I and apoA-IV. We found that the association of apoE with discoid particles is frequent, but not necessary. Calculations based on known protein mass and quantitation of discoid particles on electron micrographs suggest that the concentration of discoid particles in the peripheral lymph of cholesterol-fed dogs is about fourfold that of the plasma of the same animal. These findings provide strong circumstantial evidence for the peripheral formation of discoid HDL, perhaps as an early event in reverse cholesterol transport.

Characterization of dog peripheral lymph lipoproteins: the presence of a disc-shaped "nascent" high density lipoprotein.

The distribution, chemical, and apoprotein composition of plasma and peripheral lymph lipoproteins were compared in control and cholesterol-fed dogs. In both groups of animals, the agarose electrophoretic patterns of plasma and lymph lipoproteins were similar. In hypercholesterolemic dogs, beta-very low density lipoprotein, beta-migrating intermediate density lipoprotein, and HDLc were major components both in plasma and lymph, providing evidence for a potential interaction of these atherogenic particles with macrophages and other peripheral cells. The chemical composition and physical appearance of peripheral lymph HDL was markedly different from that of plasma HDL (high density lipoprotein), especially in the cholesterol-fed animals. Lymph HDL had a higher cholesterol to protein ratio and a markedly increased free cholesterol content (free cholesterol to cholesteryl ester ratio of 1.7 as opposed to 0.2 in plasma HDL in cholesterol-fed animals). The phospholipid content of lymph HDL was higher than that of plasma HDL, while the protein content was lower. A significant proportion of lymph HDL obtained from cholesterol-fed dogs was in the form of disc-shaped particles stacked in rouleau structures. Changes in plasma apolipoprotein concentrations due to cholesterol feeding were reflected in peripheral lymph to different degrees, depending largely on the relative size of the lipoproteins containing the individual lipoproteins. A considerable enrichment of lymph HDL with apoE and apoA-IV was observed by both immunochemical and electrophoretic methods. In lymph HDL from control and cholesterol-fed dogs, the apoE/apoA-I and apoA-IV/apoA-I ratios were several-fold elevated, compared to those of plasma HDL. It is concluded, therefore, that during cholesterol feeding a substantial portion of interstitial HDL is assembled de novo in the periphery as a crucial stage of reverse cholesterol transport to the liver. It is likely that further modification occurs upon entry to plasma and exposure to lecithin:cholesterol acyltransferase, possibly leading to generation of HDLc. Alternatively, these particles may be directly and rapidly removed by the liver.

Lipoproteins and apolipoproteins in peripheral lymph of normal and cholesterol-fed dogs.

Prenodal peripheral lymph was used as a model of interstitial fluid to obtain information on the composition of lipoproteins and apolipoproteins which are in direct contact with peripheral cells. Lipoproteins resembling plasma lipoproteins in size and electrophoretic mobility were present in the prenodal peripheral lymph of control as well as cholesterol-fed dogs. Most of the lipoproteins in control dogs were high density lipoproteins, both in the plasma and in the lymph. Cholesterol feeding resulted in an increased concentration of lipoprotein particles with decreased electrophoretic mobility (beta-VLDL and, in plasma, HDLc) and decreased concentration of HDL, both in plasma and in lymph. Size distribution of lipoproteins was also markedly altered by cholesterol feeding; most of the lipoproteins were present as IDL and VLDL both in plasma and in lymph. Judged by agarose gel chromatography, the size of the lymph HDL as consistently larger than plasma HDL in both groups of dogs. Furthermore, it appears that cholesterol feeding increased the size of an HDL subfraction, partially resolved by agarose chromatography, both in lymph and plasma. All apolipoproteins present in plasma were also present in lymph. Cholesterol feeding resulted in 3-10-fold increases in plasma apo B, E, and A-IV while apo A-I was drastically decreased. These changes were reflected in lymph to different degrees depending on the size of the lipoprotein fraction containing the individual apolipoproteins. Our findings provide direct evidence that the large, cholesterol-rich, 'atherogenic' lipoproteins found in the plasma of cholesterol-fed dogs (beta-VLDL) are also present in the interstitial fluid and presumably interact with peripheral cells. Our studies furthermore suggest modification of plasma HDL by peripheral cells and/or de novo assembly of an HDL subfraction. The utilization of this animal model may thus provide a direct approach to the study of the interaction of lipoproteins with peripheral cells.

Lecithin:cholesterol acyltransferase-mediated modification of discoidal peripheral lymph high density lipoproteins: possible mechanism of formation of cholesterol-induced high density lipoproteins (HDLc) in cholesterol-fed dogs.

Peripheral lymph high density lipoproteins (HDL) of the cholesterol-fed dog differ in a number of characteristics from plasma HDL of the same animal. Their high content of free cholesterol, phospholipid, apoprotein E, and apoprotein A-IV, their greater heterogeneity in size, and the presence of many discoidal particles suggest that a portion of lymph HDL is assembled within the interstitial fluid. The present experiments demonstrate that the endogenous lecithin:cholesterol acyltransferase (LCAT) activity of whole peripheral lymph of the cholesterol-fed dog is far less (less than 1%) than that found in the plasma of the same animal (0.3 nmol/hr per ml versus 40.6 nmol/hr per ml). Addition of partially purified LCAT to whole lymph induced many changes in the chemical composition of peripheral lymph lipoproteins. After incubation with LCAT, the free cholesterol and phospholipid contents of lymph HDL decreased, from 17% to 12% and from 46% to 33%, respectively, whereas cholesteryl ester content increased from 7% to 13%. These changes were accompanied by a mass transfer of apoprotein E and cholesterol to the p less than 1.05 g/ml fraction, the complete disappearance of the discoidal particles, and a decrease in size heterogeneity of lymph HDL. These results suggest that, in the cholesterol-fed dog, cholesterol efflux into the interstitial spaces may occur in the absence of significant LCAT activity. Furthermore, our studies suggest that the subsequent reaction between lymph HDL and LCAT within the vascular compartment leads to the generation of apoprotein E and cholesteryl ester-rich cholesterol-induced HDL.

Action of angiotensin II on renal blood flow and function during hemorrhagic shock.

The purpose of this study was to evaluate during hemorrhagic hypotension and shock the effect of angiotensin II on renal blood flow, glomerular filtration rate, and sodium and potassium excretions, and to determine its role in the development of irreversible hemorrhagic shock. Anesthetized dogs were subjected to a hemorrhagic shock protocol. Angiotensin II was infused at 100 ng/kg/min i.v. from 50 mm Hg initial hemorrhage until the experiment was terminated. The survival time from 50 mm Hg initial hemorrhage to reinfusion was increased significantly from 2.7 +/- 0.5 h to 4.7 +/- 0.8 h by exogenous angiotensin II. However, once shock had developed, the survival time from reinfusion to 50 mm Hg normovolemic hemorrhagic shock was not affected by exogenous angiotensin II (4.4 +/- 1.4 to 3.6 +/- 0.7 h.) During hemorrhagic shock, exogenous angiotensin II significantly increased sodium excretion and total renal blood flow. Glomerular filtration rate, potassium excretion, and arterial sodium and potassium concentrations were not affected. These data indicate that angiotensin II prolonged the development of irreversible hemorrhagic shock and selectively increased sodium excretion and total renal blood flow.

Lipid absorption in unanesthetized, unrestrained rats. Effects of 4-aminopyrazolopyrimidine and ethinyl estradiol.

Emulsified triacylglycerol containing [14C]palmitate was infused intraduodenally in unanesthetized, unrestrained animals treated with 4-aminopyrazolopyrimidine or pharmacologic doses of ethinyl estradiol. 4-Aminopyrazolo-pyrimidine practically eliminated the appearance of radioactivity in plasma but in ethinyl estradiol-treated animals the peak of radioactivity and shape of the plasma curve were similar to control, although lower in amplitude. A delayed appearance of radioactivity was also observed in 48-h- compared to 15-h-fasted controls, suggesting a requirement for induction of lipoprotein production prior to fat absorption.

Mesenteric lymph apolipoproteins in control and ethinyl estradiol-treated rats: a model for studying apolipoproteins of intestinal origin.

Rat mesenteric lymph contains all serum apolipoproteins. However, it is uncertain whether some of these apolipoproteins are derived from intestinal synthesis or are transferred from plasma. We compared lymph apolipoprotein composition, concentrations, and transport rates in normal rats and in rats treated with pharmacologic doses of ethinyl estradiol which have negligible concentrations of serum lipids and apolipoproteins. Lymph apolipoproteins were examined before and after duodenal lipid infusion. Lymph d less than 1.006 and 1.006-1.21 g/ml lipoproteins were isolated and SDS-electrophoresis was performed using 10 and 3.5% polyacrylamide. During lipid absorption, lymph flow increased in control but not in treated rats. Control lymph contained all major apolipoproteins, but lymph from ethinyl estradiol-treated rats contained only apoB, A-I, and A-IV. Two apoB bands were noted on 3.5% gels in control lymph, but only the lower molecular weight protein was found in lymph from ethinyl estradiol-treated rats. In control rats, transport rates for apoA-I, A-IV, E, and C proteins increased during lipid absorption, but only in the case of A-IV was this a reflection of increased apolipoprotein concentration and not the enhanced lymph flow. In ethinyl estradiol-treated rats only the A-IV transport rate increased due to lipid infusion. It is concluded that in the ethinyl estradiol-treated rat 1) the intestine does not synthesize apoE, C, or the high molecular weight apoB; 2) lymphatic output of A-IV is predominantly increased during lipid absorption; and 3) since plasma apolipoprotein concentrations are negligible, lymph lipoproteins from ethinyl estradiol-treated rats may represent a close approximation to nascent particles of intestinal origin.