Insulin-mediated modifications of myocardial lipoprotein lipase and lipoprotein metabolism.

Recirculating organ perfusion in vitro was conducted with hearts from control rats, animals given a single dose of streptozotocin (65 mg/kg) 48 h earlier, and streptozotocin-treated rats administered insulin (5 units), 2 h prior to organ perfusion. During 45-min perfusions, the lipolysis of very low density lipoprotein (VLDL) triglyceride was significantly less in hearts from diabetics than in controls (41.9 +/- 7.3% of control). This was associated with significant reductions in heparin-releasable (functional) lipoprotein lipase and tissue lipoprotein lipase of perfused hearts. The decreases in VLDL triglyceride metabolism and the levels of myocardial lipoprotein lipase were completely reversed by treatment of diabetic rats with insulin 2 h prior to study. Similar improvement of VLDL triglyceride metabolism and increases in myocardial lipoprotein lipase activity were observed in hearts from diabetic rats by direct addition of 100 milliunits/ml of insulin to the recirculating perfusion media. Under these conditions, the increase in both fractions of lipoprotein lipase in response to insulin was completely inhibited, and utilization of VLDL triglyceride was partially inhibited by pre-perfusion with cycloheximide for 10 min. The data derived from either VLDL triglyceride lipolysis in organ perfusion or direct measurement of myocardial lipoprotein lipase demonstrate a direct effect of insulin on myocardial lipoprotein lipase activity, and suggest that the response to insulin may be due in part to effects on protein synthesis.


Recirculating organ perfusion
in vitro was conducted with hearts from control rats, animals given a single dose of streptozotocin (65 mg/kg) 48 h earlier, and streptozotocin-treated rats administered insulin (5 units), 2 h prior to organ perfusion. During 45-min perfusions, the lipolysis of very low density lipoprotein (VLDL) triglyceride was significantly less in hearts from diabetics than in controls (41.9 f 7.3% of control). This was associated with significant reductions in heparin-releasable (functional) lipoprotein lipase and tissue lipoprotein lipase of perfused hearts. The decreases in VLDL triglyceride metabolism and the levels of myocardial lipoprotein lipase were completely reversed by treatment of diabetic rats with insulin 2 h prior to study. Similar improvement of VLDL triglyceride metabolism and increases in myocardial lipoprotein lipase activity were observed in hearts from diabetic rats by direct addition of 100 milliunits/ml of insulin to the recirculating perfusion media. Under these conditions, the increase in both fractions of lipoprotein lipase in response to insulin was completely inhibited, and utilization of VLDL triglyceride was partially inhibited by pre-perfusion with cycloheximide for 10 min. The data derived from either VLDL triglyceride lipolysis in organ perfusion or direct measurement of myocardial lipoprotein lipase demonstrate a direct effect of insulin on myocardial lipoprotein lipase activity, and suggest that the response to insulin may be due in part to effects on protein synthesis.
It is generally considered that a primary cause of the hypertriglyceridemia associated with insulin deficiency is defective clearance of chylomicrons and/or very low density lipoproteins from the circulation (1)(2)(3)(4). The peripheral metabolism of the triglycerides of these circulating lipoproteins is dependent on the activity of the enzyme, lipoprotein lipase, which appears to be functional at the endothelial surface of capillary beds associated with most tissues (5)(6)(7).
The lipoproteins clearance defect associated with insulin deficiency has been demonstrated in intact animals (3, 8) and in perfusion of intact myocardial tissue (e.g. Ref. 9). However, the demonstrations of a direct relationship of this defect with the activity of lipoprotein lipase have been more elusive. Since the activity of lipoprotein lipase is high in adipose tissue and heart (6,7), these tissues have been extensively exploited for * This work was supported by Grant HL-27848 from the United States Public Health Service. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$To whom communications and reprint requests should be addressed. studies of hormonal influences on lipoprotein lipase synthesis and activity.
Measurement of lipoprotein lipase activity in adipose tissue of diabetic rats indicates a reduction associated with insulin deficiency (10,ll) and a restoration by treatment with insulin (10). The increased lipoprotein lipase activity induced by insulin appears to be related to increased protein synthesis (12,13) and enzyme secretion from the adipocyte to the endothelial site of activity (12,14,15).
Similar studies with heart lipoprotein lipase in response to insulin deficiency have not been consistent. It has been reported that during experimental diabetes, heart lipoprotein lipase activity remains unchanged (8, l l ) , is decreased (9, 16,17), or is increased (10,18) with various experimental approaches. A lack of change or an increase in myocardial lipoprotein lipase during experimental diabetes is inconsistent with the observed defect in lipoprotein clearance by this organ (10,17), if, in fact, lipoprotein lipase activity represents the "rate-limiting" step in lipoprotein triglyceride metabolism.
In the present study, the isolated rat heart in recirculating perfusion has been employed as a model for assessing the overall effects of insulin deficiency and subsequent insulin therapy on (a) VLDL' triglyceride metabolism and, (b) "functional" and residual lipoprotein lipase activity (6,7,19). The data clearly demonstrate a direct relationship between the two parameters, and further suggest the stimulation of myocardial lipoprotein lipase activity by insulin is, at least in part, due to increased protein synthesis. Preliminary reports of these studies have been presented earlier (17,20).

Insulin and
Myocardial Triglyceride Utilizution 12995 to the procedure of Brenneman and Spector (26). After incubation of the lipids on Celite with VLDL for 2 h a t room temperature, the mixture was passed through a 1.2-pm Millipore filter. The VLDL were washed with 1 mM albumin, dialyzed overnight against 0.1% EDTA in 0.15 M NaCI, and reisolated by centrifugation a t d 1.006 g/ ml. Under these conditions, 5-10% of the radioactivity was incorporated into the VLDL, as had been previously reported (26). Generally, of the "C incorporated, 91% was recovered as triglyceride and 3% was as unesterified fatty acid the remainder co-chromatographed with partial glycerides and phospholipids. Of the 'H incorporated, approximately 87% was as triglyceride and the remainder was cochromatographed with phospholipids and partial glycerides. The apolipoprotein profile of VLDL was determined on delipidized samples by sodium dodecyl sulfate 10-15% gradient polyacrylamide slab electrophoresis (23). The major apolipoproteins (Fig. 1) were apo-B, apo-E, and apo-AI, with detectable bands corresponding to apo-AIV and apo-C apoproteins. These were not measurably altered by the labeling procedure.
Animal and Perfusion Procedures-Adult male rats (200-300 g) were allowed laboratory chow ad libitum until use. Streptozotocin was prepared in 0.1 M citric acid, p H 4.5, immediately before use. Animals were given a single intravenous injection of streptozotocin (65 mg/ kg) and were maintained in metabolic cages with free access to chow. They were also provided with 5% glucose in their drinking water due to the severe hypoglycemia occurring 7-12 h after administration of streptozotocin (27). Urine was monitored for glucose and ketone bodies, and urinary glucose levels in excess of 2 g/100 ml were considered as confirmation of severe hyperglycemia. Plasma glucose levels were determined with a Beckman glucose auto analyzer at the time of sacrifice. These are summarized in Table I. For studies on VLDL triglyceride utilization, hearts were removed from animals anesthetized with sodium pentobarbital, the aorta was secured to the perfusion catheter, and blood was cleared by perfusion with 10 ml of modified Krebs' buffer, pH 7.4 (28). The perfusion mixture consisted of 20 ml of Krebs' buffer containing 5 mg/ml of albumin and the VLDL preparation (0.32 mM total fatty acid equivalents). This was circulated in the recirculating perfusion apparatus (28) for 5 min prior to insertion of the heart, and subsequent heart perfusions were conducted for 45 min a t 37 "C. Coronary flow rates of greater than 5 ml/min and beats greater than 150/min were considered satisfactory. Where indicated, 100 milliunits/ml of insulin or 20 YM cvcloheximide were also included in the Derfusion mixture. as described below.
At the initiation of heart perfusion and a t 15-min intervals there-   97.6 f 4.0 Diabetic (12) 320 f 8.0 Diabetic injected with 5 units insulin (19) 37.2 f 4.0 after, 2 ml of the perfusate were removed and extracted in chloroformmethanol (2:1, v/v) according to Folch et al. (29). Additional 0.5-ml aliquots were used for assay of "CO, (30). Following perfusion, hearts were cleared of perfusate with 5 ml of Krebs' buffer. These were blotted, weighed, and homogenized in 20 volumes of chloroformmethanol for subsequent lipid analysis. Lipoprotein Lipase Activity-Analyses of heparin-releasable and tissue lipoprotein lipase activity were conducted on separate hearts from the individual experimental groups. Hearts from control, diabetic, and insulin-treated diabetic rats were perfused for 10 min with 10 ml of Krebs' buffer containing rat serum (10%) and 5 units of heparin/ml. The hearts were homogenized in 3 ml of 50 mM Tris-HC1, pH 8.0, containing 1 M ethylene glycol (31).

MOLECULAR
Studies on the effects of insulin and cycloheximide added to the heart perfusate were modified as follows. Hearts from control and diabetic rat were perfused for 10 min in the absence or presence of 20 p~ cycloheximide, prior to addition of insulin (100 milliunits/ml) and an additional 30 min of perfusion. Heparin was added during the final 10 min of perfusion. (Total perfusion time, 40 min).
Lipoprotein lipase activity of the heart perfusates and heart homogenates was assayed using a stable emulsion of glyceryl [1-"C] trioleate according to the procedure of Nilsson-Ehle and Schotz (31). Assays were conducted for 30 min a t 37 "C in a total volume of 200 pl containing 100 pl of diluted substrate preparation (31) and 100 pl of heart perfusate or heart homogenate. The reaction was stopped by addition of 3.25 ml of methano1:chloroform:heptane (141:125:100, v/ v) and 1.05 ml of 100 mM potassium carbonate-borate buffer, pH 10.5.
Labeled unesterified fatty acids were partitioned into the methanbl:water phase and aliquots were assayed for radioactivity by liquid scintillation spectrometry. Under these conditions, 75.6 f 3.0% of the labeled unesterified fatty acid was recovered in the methanokwater phase, confirming earlier studies on this liquid partition system (31).
The limitations of the assay (31) were confirmed on both perfusate and tissue lipoprotein lipase prior to study. Under the conditions of the present studies, the assay was linear for 30 min a t perfusate additions of 10-200 pl. The substrate concentration (6.68 pmollml) was not rate-limiting and optimal albumin (1%) and serum (7.5-15.0%) concentrations were as described earlier (31). During recirculating perfusion of heparin, there was a rapid release of lipoprotein lipase activity during the first 2 min and a slow release of lipoprotein lipase for at least 60 min thereafter. This latter slow release phase was consistent in all preparations, and therefore comparisons of lipoprotein lipase activities were derived from the initial 10 min rather than 2 min of perfusion.
Control assays were conducted on perfusate obtained in the absence of heparin; heart homogenates of unperfused hearts, heart homogenates to which heparin was added. Corrections were made for the recovery of labeled fatty acid in the partitioning system, and for labeled fatty acid release in the absence of heparin. Enzyme activity is expressed as milliunits where 1 milliunit is 1 nmol of oleic acid released/min a t 37 "C.
Anulytical Procedures-Lipids in the perfusate and heart homogenates were extracted (29) and fractionated into major lipid classes by thin layer silicic acid chromatography in a solvent system of hexane:ethyl ether:acetic acid (8016:2, v/v). Areas corresponding to authentic standards of unesterified fatty acid, triglyceride, diglyceride + monoglyceride, and cholesterol ester were individually scraped and eluted from silicic acid with methanokether (25:75, v/v) as described earlier (29). The origin containing phospholipid was also scraped but was subsequently handled without solvent elution (24). These were subjected to transmethylation with boron trifluoride-methanol and extracted prior to analysis of derivitized fatty acids by quantitative gas-liquid chromatography. The conditions for extraction, derivitization and chromatography, and corrections for recovery have been described previously (24). Unesterified and esterified cholesterol were determined on the silicic acid extracts by gas-liquid chromatography (24). Radioactivity in all silicic acid fractions was determined by liquid scintillation spectrometry (Beckman LS-250). Where indicated, total protein of the lipoproteins and heart was analyzed according to Markwell et al. (32).
Data on individual lipids concentrations of the lipoproteins and heart tissue are expressed as micromoles and are derived from the total fatty acid concentrations of each lipid class. Since the absolute level of triglycerides in VLDL varied between preparations (6.4 ? 0.75 pmol), all data were normalized to 6.4 pmol of VLDL triglyceride/ perfusion. The data are expressed as means/g of wet weight 1?: S.E., and differences between values were analyzed for significance by the Student's t test.

Correlation of VLDL Triglyceride Mass and Radioactivity during Perfusion-The use of VLDL labeled in vitro with [2-
'Hlglyceryl tri[l-14C]oleate for measurement of VLDL triglyceride clearance and metabolism of triglyceride fatty acids was assessed by comparison of the disappearance of triglyceride mass and radioactivity during heart perfusion. As shown in Fig. 2, the disappearance of triglyceride fatty acids during recirculating perfusion of control hearts with VLDL was essentially linear over the 45-min perfusion period. As shown in Table 11, these triglycerides are composed largely (90%) of linoleic (18:2 n-6), oleic (18:l n-9), palmitic (16:0), and stearic (18:O) acids. During the disappearance of approximately 2.6 wmol of total triglyceride fatty acids, the proportions of the individual major fatty acids remained unchanged (Table II), suggesting the lack of fatty acid specificity during lipolysis of the VLDL core triglycerides by myocardial lipoprotein lipase. During this same period, the disappearance of VLDL glyceryl tri[ l-'*C]oleate was proportional to disappearance of triglyceride mass (Table 111) With hearts from diabetic rats under identical perfusion conditions, disappearance of triglyceride fatty acids was also linear but occurred a t a rate of 24.2 k 4.2 nmol/min, or only 42% of that by control hearts. Again, this was not associated with significant accumulations of labeled unesterified fatty acids or partial glycerides in the medium. Accumulation of 13H]glycerol from lipolysis of [2-'H]glyceryl trioleate was linear, representing a release of 6.0 k 2.0 nmol/min. Although the ratio of fatty acid to glycerol release with diabetic hearts    2.60 f 0.13 24 Table  IV. With control hearts, almost 25% of the labeled fatty acid released was recovered as I4CO2 by 45 min of perfusion. Another 52% was recovered as tissue non-lipid (or water-soluble) radioactivity suggesting catabolism of the fatty acid to water-soluble intermediates or metabolites. Only 6.3% was recovered as tissue lipid, and of this, 71.1 f 2.3% was as tissue triglyceride. Thus, 93% of the fatty acid label was utilized by myocardial tissue and only 7% remained in the perfusate. Of the labeled glycerol released during lipolysis, 90% was recovered in the perfusate, and only 5.3% was utilized for lipid synthesis. Of this, 69.5 f 3.7% was associated with tissue triglyceride.
With hearts from diabetic rats, a greater percentage (60.7 -t 11.2%) of the released fatty acid radioactivity was recovered in tissue lipids (0.66 -t 0.12 pmol compared to 0.16 f 0.03 pmol in controls), and of this, 61.2 f 6.5% was as tissue triglyceride. Although the complete oxidation of fatty acid to 14C02 was statistically comparable to that in controls (0.42 f 0.16 pmol uersus 0.64 f 0.13 pmol), there was no radioactivity in the non-lipid fraction of heart extracts. Of the labeled glycerol released, 76% remained in the perfusate, and a greater percentage was recovered as tissue lipid, of which 62.6 f 2.2% was as triglyceride.
Effect of Insulin on Myocardial Triglyceride Utilization-Diabetic rats were given a single intraperitoneal dose of insulin (5 units) 2 h prior to removal of hearts for perfusion. As shown in Table IV, hearts from these animals were normalized with respect to lipolysis of VLDL [2-3H]glyceryl tri[l-14C]oleate. The release of either 14C-fatty acid or [3H]glycerol was comparable to that in hearts from control rats and significantly greater than that observed with diabetic tissue. Despite the normalization of lipolysis of the triglyceride, the subsequent metabolism of the released fatty acid and glycerol was different from either control or diabetic tissue. Only 41% (0.96 pmol) of the released fatty acid was accounted for by complete or partial catabolism of the fatty acid, and a greater percentage (21.9% or 0.52 pmol) was recovered as tissue lipid. Thus, catabolism of fatty acid was improved over that by diabetic tissue but still only 50% of that by control tissue. Accumulation of radioactivity as tissue lipid (primarily triglyceride), in contrast, was 3.2 times that in controls, and similar to that in diabetic tissue (0.52 pmol and 0.66 pmol, respectively). This increased formation of esterified lipid was also observed by analysis of the metabolic fate of the released glycerol. In this case 19.4% (0.21 pmol) was incorporated into tissue lipid compared to 5.3% (0.05 pmol) by hearts from control rats.
The direct effect of insulin on myocardial lipolysis of VLDL triglyceride was demonstrable by including insulin (100 milliunits/ml) in the perfusates of diabetic hearts (Table IV). At this high level of insulin over the 45-min recirculating perfusion period, almost 59% of the available [14C]triglyceride oleic acid was released, representing 1.4 times that by control hearts, and 1.5 times that by hearts from diabetic animals treated with insulin. In this case, however, total catabolism of fatty acid was comparable to that in controls, while accumulation of label in tissue lipid (27.5% or 1.0 pmol) was still elevated (6.2-fold over controls), and even greater than that in hearts from insulin-treated diabetic rats (by 1.9-fold). This increased incorporation of label into tissue lipid was also evident by analysis of the fate of the released glycerol.
Myocardial Lipoprotein Lipase Actiuity-Heparin-releasable (functional) lipoprotein lipase and tissue lipoprotein lipase were determined on the perfusates and hearts following a 10min recirculating perfusion with 5 units of heparin/ml perfusate. It was determined in preliminary studies that during perfusion of VLDL, or perfusions in the absence of heparin, release of lipoprotein lipase activity into the perfusate was only 7.5 milliunits (0.75 milliunits/ml for 10 ml of perfusate). When hearts from controls rats were perfused with heparin, about 71% of the total releasable and non-releasable lipoprotein lipase was found in the perfusate (Table V).
Total lipoprotein lipase activity of diabetic hearts was 74 f 3% of that in control hearts. This decrease was a reflection of both the heparin-releasable fraction (78.6 f 2.7% of control) and the tissue residual activity (63.2 k 4.0% of control). With hearts from diabetic rats treated with 5 units of insulin 2 h before study, total lipoprotein lipase activity (1220 f 74 milliunits), and the percentage of the total lipoprotein lipase released by heparin perfusion (70.7 f 4.4% of total lipoprotein lipase) was the same as in controls. As a control, normal rats were also given a single dose of 5 units of insulin, 2 h prior to study. In this group, levels of total lipoprotein lipase, and the proportions of heparin-releasable and tissue residual lipoprotein lipase activities were indistinguishable from those of controls or of diabetic animals treated with insulin.
The direct effect of insulin on myocardial lipoprotein lipase activity was demonstrated by perfusion of either control or  diabetic hearts with insulin (100 milliunits/ml) prior to analysis of heparin-releaseable and residual lipoprotein lipase activities. Perfusion of hearts from diabetic rats with insulin resulted in an increase ( p < 0.01) in total lipoprotein lipase activities to levels which were comparable to those in controls (1168 f 104 uersus 1135 f 82 milliunits, respectively). This was a result of a proportional increase in both heparinreleasable and residual fraction of the heart (73.7 f 8.1% of the total activity was released by heparin). Insulin had only a small effect on the lipoprotein lipase activity of hearts obtained from control rats (1366 f 22 and 1135 f 82 milliunits, respectively), and again, this was a result of proportional increases in the two tissue lipoprotein lipase fractions.
Effect of Cycloheximide on Myocardial Clearance of VLDL Triglyceride and Lipoprotein Lipase Actiuities-Since the in uiuo effect of insulin on myocardial VLDL triglyceride utilization and lipoprotein lipase activities was largely duplicated by perfusion of hearts with insulin in uitro, perfusion studies were designed to assess the effect of cycloheximide on the insulin-dependent improvement of VLDL triglyceride clearance and on myocardial lipoprotein lipase activity.
For studies on VLDL triglyceride clearance, hearts from diabetic rats were perfused for 10 min with buffer, insulin, or insulin and 20 PM cycloheximide. This was followed by addition of the VLDL preparation labeled with glyceryl tri[l-14C] oleate, and perfusions were continued for 45 min. With these modified perfusion conditions, the effect of insulin on the improvement of the lipolysis of VLDL triglyceride by diabetic hearts (Fig. 3) was even more pronounced than without the 10-min pre-perfusion with insulin (Table IV). This effect of insulin was partially, but not completely, prevented when cycloheximide was present during the preperfusion period (Fig. 3).
Comparable studies were designed to determine the effect of cycloheximide on functional and residual lipoprotein lipase activities (See "Experimental Procedures"). As shown in Fig.  4, these perfusion conditions gave control lipoprotein lipase activity levels comparable to those in Table V (921 f 53 units and 303 f 23 units of functional and residual LPL activity, respectively). Perfusates from diabetic hearts contained only 47% of the lipoprotein lipase activity of that from control hearts ( p < 0.051, and tissue lipoprotein lipase activity was also proportionately less (51% of control) than in controls. The effect of diabetes in this study was more dramatic than in the earlier study (Table V) total of only 10 min. Perfusion of hearts from diabetic rats with 20 pM cycloheximide (40 min) had no effect on the already reduced levels or distribution of functional and residual lipoprotein lipase activities. Perfusion of diabetic hearts with buffer for 10 min, and 100 milliunits of insulin/ml for 30 min, completely reversed the effect of diabetes and resulted in functional and residual lipoprotein lipase activities comparable to those in control hearts. However, when diabetic hearts were perfused with cycloheximide for 10 min prior to addition of insulin, and continued perfusion for 30 min, the increase in lipoprotein lipase activity during perfusion with insulin alone was completely blocked. Under these conditions, both functional and residual lipoprotein lipase activities remained low, with no suggestion of an accumulation of tissueassociated lipoprotein lipase activity.
Since the effects of perfusions with insulin and cycloheximide on myocardial lipoprotein lipase were observed when lipoprotein lipase levels were low, an additional study was conducted to determine the effects of these agents on lipoprotein lipase levels in control hearts. During these 40-min perfusions, insulin had no significant effect on the already elevated levels of functional lipoprotein lipase (906 -+ 76 uersus 761 f 43 units of insulin-perfused and control hearts, respectively). However, in the presence of cycloheximide added 10 min prior to addition of insulin, functional lipoprotein lipase levels were significantly less (662 f 77 units; p < 0.05) than after perfusions with insulin alone.

DISCUSSION
Diabetes with uncontrolled hyperglycemia is generally associated with elevated VLDL triglycerides (e.g. Ref. 34). Nikkila et al. (34) have suggested that these individuals have VLDL secretion rates comparable to non-diabetics, and that the associated hypertriglyceridemia is due to defective clearance rates. Chen et al. (35) have demonstrated that VLDL secretion rates are not elevated in streptozotocin-induced insulin deficiency in rats, and the clearance defect has been clearly demonstrated in a similar animal model by infusion of labeled VLDL triglycerides (3, 36).
The enzyme responsible for peripheral clearance of VLDL triglycerides is lipoprotein lipase. In intact organs such as adipose tissue and heart, lipoprotein lipase exists in both functional and nonfunctional forms (5,14). The functional form appears to be largely associated with the endothelial surface of tissue capillaries and is readily released by heparin (5). This lipoprotein lipase activity is presumably responsible for the degradation of lipoprotein triglyceride during tissue assimilation of the triglyceride fatty acids. The nonfunctional form of lipoprotein lipase which is not readily releasable by heparin, appears to be associated with the tissue parenchymal cells (5). The activity has also been termed "residual" and may represent the ultimate source of the vascular endothelial activity (e.g. Refs. 37 and 38).
Data attempting to relate the defective clearance of lipoprotein triglycerides during human or experimental diabetes, and the activity of either functional or total tissue lipoprotein lipase have not been convincing. The original observations by Bierman et al. (39) suggested a decrease in postheparin lipolytic activity in ketosis-prone diabetic patients. These data, however, have been questioned (e.g. Ref. 35) in part, on the basis that postheparin lipolytic activity represents multiple enzyme activities derived from various tissues (34,40,41). More direct measurements, however, suggested a decrease in adipose tissue lipoprotein lipase in untreated diabetics (42,43). Others, however, have been unable to demonstrate changes in either adipose (36) lipoprotein lipase, or in plasma postheparin lipolytic activity (11) in insulin-deficient rats. In fact, the original studies of Kessler (10) suggested that insulin deficiency in rats, induced by alloxan treatment, resulted in increased heart muscle lipoprotein lipase and decreased adipose lipoprotein lipase and that both changes were reversible by insulin. These data on heart lipoprotein lipase, however, are not compatable with the observation (9, 17) that perfused hearts from diabetic rats demonstrated defective clearance of chylomicron triglycerides. It seems likely, that differences in results on lipoprotein lipase activities under various experimental conditions are due in part to methods of tissue preparation and to the assay procedures. Lipoprotein lipase has been assayed on fresh tissue homogenates (e.g. Refs. 5 and 44) in the absence or presence of heparin (16), on acetone powders (e.g. Refs. 44 and 45) and on organ perfusate (45, 46). Activity measurements have been made on a variety of emulsified and natural triglyceride substrates. These multiple variations have not allowed consistent findings or effective comparison of results. The differences in measurable functional lipoprotein lipase activity are also evident in the present study. Lipoprotein lipase activity, calculated from the extent of VLDL triglyceride lipolysis during controlled perfusion conditions, is likely to represent a more physiological assay. Under these conditions, calculated lipoprotein lipase activity was about 58 milliunits, compared to 800-900 milliunits (Table V, and Fig. 3) of functional lipoprotein lipase activity assayed by using the artificial substrate preparation (31). Nevertheless, as discussed below, the results were internally consistent, and comparable changes during each experimental condition were observed using either approach.
In the present study, the intact heart in recirculating perfusion has been employed as an organ model to assess the role of insulin on both the clearance of VLDL triglycerides, and on the activity of the endothelial, membrane-associated lipoprotein lipase. Studies on the recirculating perfusion of VLDL clearly demonstrated a defective triglyceride clearance associated with insulin deficiency within 48 h of a single dose of 65 mg/kg of streptozotocin. The defect was observed by measurement of either disappearance of triglyceride mass, or of [2-3H]glyceryl tri[l-14C]oleate, or by the accumulation of [2-3H]glycerol in the perfusion media. These data (58% reduction in clearance of 0.32 mM triglyceride) are entirely consistent with those of Kreisberg (9), who initially demonstrated a 70% reduction in chylomicron triglyceride (0.3-0.5 mM) clearance by perfused hearts from rats rendered insulindeficient by a single injection of 45 mg/kg of alloxan 3-5 days earlier.
Although the present studies and those of Kreisberg (9) employed lymph lipoprotein particles for assessment of tissue clearance, the observed data on defective clearance of circulating triglyceride by perfused hearts of diabetic rats were duplicated using isolated plasma VLDL.' Under identical perfusion conditions and triglyceride concentrations, lipolysis of plasma VLDL triglyceride during perfusion of controls hearts was 4.7 ~mo1/45 min while with lymph VLDL, 2.6 pmol were cleared. Lipolysis of plasma VLDL by hearts from diabetic rats was 54% of control while with lymph VLDL, lipolysis was 42% of control. Thus, based on all of the measured criteria, it seems clear that a major defect in the hearts from diabetic rats is a reduced ability to degrade circulating lipoprotein triglycerides.
There were also characteristic metabolic differences in diabetic tissue, with respect to the fate of the released fatty acids, and these were also compatable with previous observations. The accumulation of myocardial triglycerides during diabetes is well recognized, and can occur within 48 h of streptozotocin administration to rats (47). We have determined that myocardial triglycerides are increased by 71.4 +. 9.8% within 48 h of administration of the drug to rats? In the present study, diabetes was associated with an increased percentage of the extracted fatty acid recovered as tissue lipid, and of this, the majority was as triglyceride. Conversely, as had been reported earlier (9), catabolism of the extracted fatty acid to CO, and water-soluble catabolites was markedly reduced in diabetic hearts. Thus, acute diabetes, as defined by marked hyperglycemia in these animals, was associated with decreased myocardial clearance of VLDL triglycerides, and with increased esterification and reduced catabolism of the extracted triglyceride fatty acids.
It seems possible that abnormal tissue metabolism of fatty acids may subsequently influence VLDL triglyceride lipolysis without modifying the activity of functional lipoprotein lipase. However, using direct measurements of heparin-releasable lipoprotein lipase, it was possible to show a significant reduction of functional lipoprotein lipase activity in hearts from diabetic rats, suggesting that the clearance defect was at least in part due to reduced lipolytic capacity of this tissue. The reduced clearance of VLDL triglycerides by perfused hearts from diabetic rats, and the measurable decrease in functional lipoprotein lipase from these tissues are difficult to reconcile with earlier reports of unchanged or even increased levels of myocardial lipoprotein lipase associated with insulin deficiency (8,10,11,18).
Administration of a pharmacological dose of insulin (5 units) to diabetic rats resulted in severe hypoglycemia within 2 h. This treatment also resulted in a normalization of VLDL triglyceride lipolysis by perfused hearts, but did not completely reverse the metabolic defects observed with hearts from diabetic rats. Thus, a significant amount of the tissue fatty acid was retained as lipid (0.52 pmol compared to 0.66 pmol in diabetic tissue and 0.16 pmol in control hearts), and this was largely (75%) as triglyceride. Although complete oxidation of fatty acid was not improved by treatment of diabetic rats with insulin, overall catabolism (to water-soluble products and CO,) was improved (0.95 pmol compared to 0.43 pmol in diabetic hearts).
These data suggested that the improvement in VLDL triglyceride clearance by hearts from insulin-treated diabetic rats is not simply a result of altered metabolic characteristics of the tissue per se, but might reflect a direct or indirect effect of insulin on lipolytic activity of the coronary vasculature. This was again substantiated by the observation that functional lipoprotein lipase activity of hearts from diabetic rats treated with insulin was significantly increased over that in animals not given insulin, and was comparable to that in control hearts.
Studies on the addition of insulin to the recirculating perfusion media were designed to assess the direct effects of insulin on myocardial VLDL triglyceride clearance. With the high levels of insulin (100 milliunits/ml) present during the entire perfusion period, the lipolysis of VLDL triglyceride by diabetic hearts was markedly improved (by 3.3-fold), and was even greater than that observed by perfusion of control hearts or hearts from diabetics rats treated with insulin. Under these conditions, there was also improvement of the catabolism of tissue fatty acids to levels comparable to that in control hearts. However, tissue levels of lipid (primarily as triglyceride) were still elevated compared to controls. These studies suggest a direct effect of insulin on the ability of perfused hearts to metabolize VLDL triglyceride, and this is further supported by the observation that, under these experimental conditions, the activities of both functional and tissue lipoprotein lipase are significantly increased by insulin, and to levels comparable to those in control hearts.
Evidence for the involvement of protein synthesis during insulin-induced increases in myocardial triglyceride utilization was derived from both assay systems. Thus, the increased lipoprotein lipase activities in response to insulin, assayed in vitro under optimized conditions, were completely blocked by pretreatment of hearts with cycloheximide. Using the more physiological measurement of VLDL triglyceride lipolysis during organ perfusion, however, cycloheximide only partially prevented the insulin-stimulated improvement in lipolysis. Nevertheless, these studies collectively imply that protein synthesis is at least in part involved in the effect of insulin on lipoprotein lipase activity.
It has been estimated that the turnover time of the functional enzyme at the capillary endothelium of rat heart is about 2 h (19). Morgan et al. (48) have clearly demonstrated that insulin stimulates and cycloheximide inhibits protein synthesis during short-term perfusions of rat hearts. From analyses of polysome and ribosomal subunit profiles, it has been suggested (48), that insulin accelerates steps involved in initiation of peptide chain synthesis in this tissue.
Furthermore, a reduction of total protein synthesis has been demonstrated in perfused hearts from rats following alloxan-induced diabetes (49). This reduction can be largely restored to normal levels by the perfusion of insulin for 1 h. Considering these results and the data in the present study, it seems likely that the reduced lipoprotein lipase activity during diabetes is due in part to lowered lipoprotein lipase synthesis or depressed synthesis of an activator or processing enzyme. Therefore, a general stimulation of protein synthesis by insulin could restore lipoprotein lipase activity to normal levels in either a specific or nonspecific manner.
The present studies on myocardial tissue are entirely consistent with those employing adipose tissue. Insulin can, under specific experimental conditions, stimulate adipose lipoprotein lipase synthesis (6, 7 ) but may not be directly involved in post-translational modifications (50). Activation, presumably via glucose concentration (13), has also been considered important in regulation of functional lipoprotein lipase activity. Enzyme release from the adipocyte, however, appears to be independent of protein synthesis (52).
Thus, despite reported differences in the K,,, for the enzymes from adipose tissue and heart (6), and the possible differences in effects of circulating metabolite levels, such as VLDL (53), it appears that the enzyme from both tissues may be directly regulated by insulin. The mechanism(s) and site of this regulation in myocardial tissue require additional attention.

Insulin and
Myocardial Triglyceride Utilization 13001