Development of Muscle Insulin Resistance After Liver Insulin Resistance in High-Fat-Fed Rats EDWARD W. KRAEGEN, PETER W. CLARK, ARTHUR B. JENKINS, EUGENE A. DALEY, DONALD J. CHISHOLM, AND LEONARD H. STORLIEN
Muscle and hepatic insulin resistance are two major defects of non-insulin-dependent diabetes mellitus. Dietary factors may be important in the etiology of insulin resistance. We studied progressive changes in the development of high-fat-diet-induced insulin resistance in tissues of the adult male Wistar rat. In vivo insulin action was compared 3 days and 3 wk after isocaloric synthetic high-fat or high-starch feeding (59 and 10% cal as fat, respectively). Basal and insulin-stimulated glucose metabolism were assessed in the conscious 5- to 7-h fasted state with the euglycemic clamp (600 pM insulin) with a [3-3H]glucose infusion. Fat feeding significantly reduced suppressibility of hepatic glucose output by insulin after both 3 days and 3 wk of diet (P < 0.01). However, a significant impairment of insulin-mediated peripheral glucose disposal was only present after 3 wk of diet. Further in vivo [3H]-2-deoxyglucose uptake studies supported this finding and demonstrated adipose but not muscle insulin resistance after 3 days of high-fat feeding. Muscle triglyceride accumulation due to fat feeding was not significant at 3 days but had doubled by 3 wk in red muscle (P < 0.001) compared with starch-fed controls. By 3 wk, high-fat-fed animals had developed significant glucose intolerance. We conclude that fat feeding induces insulin resistance in liver and adipose tissue before skeletal muscle with early metabolic changes favoring an oversupply of energy substrate to skeletal muscle relative to metabolic needs. This may generate later muscle insulin resistance. Diabetes 40:1397-1403, 1991
R
educed potency of insulin action (insulin resistance) is a characteristic of non-insulin-dependent diabetes mellitus (NIDDM) and several other possibly related metabolic disorders such as obesity and hypertension. Although the contribution of insulin resistance to the NIDDM state is widely recognized, its precise etiology is not understood. However, it is likely that both genetic and life-style factors, i.e., diet and activity, are in-
DIABETES, VOL. 40, NOVEMBER 1991
volved (1). Insulin resistance occurs early in the development of NIDDM (2). There are conflicting reports favoring the primacy of hepatic (3) or peripheral (4) insulin resistance, but the general inference is that it might occur at different times in specific insulin target tissues. Our study is concerned with the development of diet-induced insulin resistance and glucose intolerance. Dietary factors such as highfat feeding have been implicated in the development of hepatic and muscle insulin resistance in rats (5-7), although we are unaware of any studies investigating the relative rate of development of diet-induced insulin resistance in specific tissues. Rats fed for 3 wk with a high-fat diet (either saturated or co-6 polyunsaturated fat) develop widespread peripheral and hepatic insulin resistance characterized by reduced uptake of glucose into skeletal muscle and adipose tissue and impaired suppressibility of hepatic glucose output by insulin (6-8). However, the relative time of onset of these metabolic responses and the mechanisms responsible are unknown. One contributing factor may be an increase in muscle triglyceride content (8), although the time relationship between an increase in muscle triglyceride and development of insulin resistance has not been established. Therefore, we compared in vivo insulin action in 3-day and 3-wk high-fat-fed rats to better understand the onset of insulin resistance in this model.
RESEARCH DESIGN AND METHODS Adult male Wistar rats (300-380 g) with free access to food and water were used in this study. Rats were housed in individual cages in a temperature-controlled room (22 ± 1°C) with a 12-h light-dark cycle (lights on 0600). All stud-
From the Garvan Institute of Medical Research, St. Vincent's Hospital, Sydney, New South'Wales, Australia. Address correspondence and reprint requests to Dr. E.W. Kraegen, Garvan Institute of Medical Research, St. Vincent's Hospital, Sydney, New South Wales 2010, Australia. Received for publication 4 September 1990 and accepted in revised form 28 May 1991.
1397
DEVELOPMENT OF DIET-INDUCED INSULIN RESISTANCE
ies were approved by the Garvan Institute Animal Experimental Ethics Committee and complied with the National Health and Medical Research of Australia guidelines for animal experimentation. Cannulations were performed under pentobarbital sodium anesthesia (60 mg/kg Nembutal, Abbott, Sydney). Cannulas were implanted in the left carotid artery and right jugular vein and exteriorized through the back of the neck. These lines were kept patent by filling with a 60% solution of polyvinylpyrrolidone (40,000 Mr Fluka Bioscientific, Sydney) made up in heparin sodium (1000 U/ml, Fisons, Sydney). Rats underwent a 5-day postoperative recovery period, during which there was close monitoring of dietary intake. Rats were fed isocalorically matched high-fat or highstarch diets (Table 1) for either a long-term (3-wk) or shortterm (3-day) period. Long-term rats were fed for 16 days precannulation on their diet with a precise record of dietary intake and spillage. All spillage was added to the next day's amount, but we found that little or no spillage occurred after the 1st day on the diet. These rats were cannulated as described above and given their specific diet for the 5-day postoperative recovery period. Short-term diet-treated rats were fed normal laboratory chow up to the morning of cannulation. After cannulation, all rats were fed the high-starch diet because it was of precisely known composition. After 2 days on the high-starch diet, rats were switched to the highfat diet or remained on the high-starch diet (controls) for the 3 days before acute study. On the morning of the 6th day, acute glucose turnover studies were performed 5-7 h after the removal of food. Sequential basal and euglycemic hyperinsulinemic glucose turnover studies as described elsewhere (9) were performed in conscious rats on each of the four groups (3-day or 3-wk high-fat- or high-starch-fed rats, n = 5-6 rats/ group). Briefly, infusion and sampling lines were connected to the cannulas, and rats were returned to individual cages and allowed a 30- to 40-min settling period before comm-
TABLE 1 Composition of high-fat and high-starch diets used in diet treatment Ingredient
High-fat diet (g/kg)
High-starch diet (g/kg)
Safflower oil Comstarch Casein Methionine Gelatin Wheat bran Vitamin mix* Mineral mix|
339 254 254 3 19 51 13 67
41 658 188 2 14 38 9 50
By calories, the fat diet contained 59% fat, 2 1 % protein, and 20% carbohydrate, and the starch diet contained 10% fat, 21% protein, and 69% carbohydrate. 'Supplied 3 g thiamine mononitrate, 3 g riboflavin; 3.5 g pyridoxine HCI; 15 g nicotinamide; 8 g d-calcium pantothenate; 1 g folic acid; 0.1 g d-biotin; 0.005 g cyanocobalamin; 0.013 g cholecalciferol; 0.025 g acetomenaphthone; 0.6 g vitamin A acetate; 25 g d-RRRa-tocopherol acetate and 10 g choline chloride per kg mix. tSupplied 30.5 g MgSO4 • 7H2O; 65.2 g NaCI; 105.7 g KCI; 200.2 g KH2PO4; 38.8 g MgCO3 • Mg(OH)2 • 3H2O; 40.0 g FeC6H5O7-5H2O; 512.4 g CaCO3; 0.8 g Kl; 0.9 g NaF; 1.4 g CuSO4 • 5H2O; 0.4 g MnSO4; and 0.05 g CoNO3 per kg mix.
1398
encing the study. [3-3H]glucose (Amersham, Sydney) was infused into the jugular vein as a priming dose (1.25 |xCi) followed by a constant infusion (0.05 |xCi/min). Steady-state glucose specific activity was measured 50, 60, and 70 min after the start of the infusion. The 60-min sample was also assayed for the basal plasma insulin, corticosterone, and free fatty acids. After the 70-min sample, rats were infused additionally with pork insulin (25 pmol • kg" 1 • min" 1 ; Actrapid, Novo, Bagsvaerd, Denmark), and euglycemic clamps were performed (10). Having reached a steady state, plasma glucose and glucose specific activity were determined 60, 70, and 80 min after the insulin infusion started. The final sample was assayed for the plasma parameters as above. Peripheral glucose disposal (R6) and hepatic glucose output (HGO) were calculated as Ra =
and
HGO = Ra - GIR,
where R*a is the tracer infusion rate (dpm/min SAg is the steady-state value of glucose specific activity, and GIR is the glucose infusion rate (0 in the basal state) taken over a 30-min period from 50 to 80 min after the start of the insulin infusion. Further studies were conducted on 3-day high-fat (n = 6) and 3-day high-starch (n = 6) -fed rats to assess relative tissue insulin sensitivity in a number of hindlimb muscles and fat depots. On the morning of the study, rats were anesthetized and cannulated (as described above), and euglycemic hyperinsulinemic clamps were combined with bolus [3H]-2deoxyglucose administration (11). Because these studies were essentially to confirm main data, they were performed in anesthetized rats for simplicity, although hindlimb blood flow and absolute glucose uptake are reduced under these circumstances (12). To investigate the accumulation of energy storage products in hindlimb muscle, a further series of studies was performed. Rats were fed either the 3-day or 3-wk high-fat or high-starch diet as described above, but only a single jugular cannula was implanted. On the morning of the study, a blood sample was taken to measure basal plasma parameters (i.e., glucose, insulin, nonesterified fatty acids (NEFA), triglyceride, and corticosterone), then the animal was killed with an overdose of pentobarbital sodium (60 mg), and soleus, plantaris, and red gastrocnemius muscle samples were rapidly removed and immediately freeze-clamped. Tissues were kept at -70°C until assayed for glycogen (plantaris; 13) and triglyceride. Triglycerides (soleus and red gastrocnemius) were extracted by a modification of the method of Denton and Randle (14). Intravenous glucose tolerance tests (IVGTT; 500 mg/kg glucose, 50% wt/vol solution, David Bull, Sydney) were performed on 5-h fasted chronically cannulated conscious rats maintained on 3-day high-fat (n = 6) or high-starch {n = 5) diets. Blood samples (300 jxL) were taken 0, 2, 5, 10, 15, 20, and 50 min after the intravenous glucose bolus for estimation of plasma glucose and insulin. Erythrocytes were resuspended and returned to the animal after the 0- and 20min sample. Glucose disappearance was approximated by first-order kinetics up to 15 min, and a K value was estimated with loglinear regression over this period. For comparison,
DIABETES, VOL. 40, NOVEMBER 1991
E.W. KRAEGEN AND ASSOCIATES
TABLE 2 Postcannulation dietary intake 3 wk diet
3-day diet Days postcannulation 1 2 3 4 5
High fat 63 79 97 100 100
±6 ± 8 ± 1 ±0 ± 0
High starch 59 78 88 99 99
± ± ± ± ±
High fat
8 6 7 1 1
74 93 98 98 97
± ± ± ± ±
7 3 2 2 2
High starch 88 94 97 99 99
± ± ± ± ±
3 3 2 1 1
Results are expressed as a percentage of normal precannulation dietary intake, n, 9-13 animals; diets are high fat or high starch as indicated.
K values for 3-wk high-fat and high-starch -fed animals (both n = 7) were calculated from previously obtained data (15). Blood and plasma glucose concentrations were determined with a glucose analyzer (model 23AM, Yellow Springs, OH). Plasma samples for determination of insulin, corticosterone, and NEFA were kept at - 20°C until assayed. Insulin (16) and corticosterone (17) were assayed by radioimmunoassay. Enzymatic colorimetric kits were used to determine plasma concentrations of free fatty acids (NEFA-C code no. 279-75409, Wako, Osaka, Japan). Plasma samples for determination of tracer concentration were deproteinized immediately in 2.8% ZnSO4 and saturated Ba(OH)2. An aliquot of supernatant was dried down to remove tritiated water and redissolved in an aqueous scintillant (Picofluor 30, Packard, Rockville, MD). Radioactivity counting was performed in a liquid-scintillation spectrophotometer (Beckman, Fullerton, CA). All experimental data were expressed as means ± SE. Statistical analyses were performed with paired or unpaired two-tailed Student's t tests or a two-way analysis of variance (ANOVA) as appropriate. P < 0.05 was considered significant. RESULTS
The postcannulation dietary intake of both the 3-day and 3wk high-fat- and high-starch-fed rats expressed as a per-
centage of normal precannulation dietary values was similar to that reported previously for starch-fed rats (9; Table 2). Dietary intake increased progressively in each group over the first 2 days of the postcannulation recovery period. On the 1st day, dietary intake was reduced to - 6 0 % of normal in rats not previously on the diet (3-day rats) and - 8 0 % in rats used to the diet (3-wk rats). Dietary intake increased over day 2 and was 90-100% of normal for the next 3 days before the study (i.e., normal over the full test period of shortterm feeding). There was no significant difference in energy intake among the four groups over the 3-day period before acute study. Plasma concentrations of glucose, insulin, NEFA, and triglyceride during the basal and clamp phases of the study are displayed in Table 3. There was no significant effect of fat feeding (either 3 day or 3 wk) on the absolute basal or clamp plasma concentrations of any of these parameters. Although basal NEFA levels were similar, there was a tendency toward reduced suppression of NEFA by insulin with fat feeding (significant suppression in both starch groups, P < 0.05, but not significant in either fat-fed group). Basal corticosterone levels were similar in all four groups (mean 40-70 nM, data not shown). Basal and clamp glucose turnover (HGO and Ra) and clamp glucose infusion rates are displayed in Table 4, and HGO and P>d are plotted against prevailing insulin levels in Fig. 1. Fat feeding reduced basal glucose turnover (HGO and P>d) by - 2 0 % at 3 days (P < 0.05) and at 3 wk (P < 0.09). In the insulin-stimulated state (-100 mU/L), fat feeding induced whole-body insulin resistance (as indexed by the clamp GIR both at 3 days (reduced by 46%, P < 0.001) and 3 wk (reduced by 59%, P < 0.001). This was associated with significantly less suppression of HGO during the clamp both at 3 days (fat vs. starch, 26 ± 9 vs. 106 ± 4% suppression of basal HGO, P < 0.001) and 3 wk (fat vs. starch, 43 ± 10 vs. 102 ± 10% suppression of basal HGO, P < 0.01). However, there was no significant difference in insulin-stimulated P.d between high-fat- and high-starch-fed rats after 3 days on the diet (Fig. 1; Table 4). After 3 wk of fat feeding, peripheral insulin resistance developed. Insulin-stimulated P.d
TABLE 3 Basal and euglycemic hyperinsulinemic clamp plasma parameters in rats maintained on high-fat or high-starch diets for 3 days or 3 wk 3 day High fat Basal Glucose (mM) Blood Plasma Insulin (pM) NEFA (mM) Triglyceride (mM) Clamp Glucose (mM) Blood Plasma Insulin (pM) NEFA (mM)
3 wk High starch
High fat
High starch
4.6 7.2 240 0.73 0.92
± ± ± ± ±
0.1 0.2 30 0.13 0.06
4.5 7.0 276 0.83 0.88
± ± ± ± ±
0.2 0.2 42 0.13 0.11
4.9 7.3 186 0.94 0.79
± ± ± ± ±
0.1 0.3 24 0.08 0.03
4.5 6.9 240 1.28 0.92
± ± ± ± ±
0.1 0.2 60 0.14 0.08
4.5 7.2 700 0.73
± ± ± ±
0.1 0.3 30 0.08
4.3 6.8 680 0.49
± ± ± ±
0.2 0.2 20 0.10*
4.7 7.0 590 0.66
± ± ± ±
0.1 0.2 70 0.17
4.6 6.9 580 0.74
± ± ± ±
0.1 0.2 70 0.11*
Results are expressed as means ± SE. n = 6 for all groups except 3-wk fat (n = 4). NEFA, nonesterified fatty acids. *P < 0.05 vs. basal value in same group.
DIABETES, VOL. 40, NOVEMBER 1991
1399
DEVELOPMENT OF DIET-INDUCED INSULIN RESISTANCE
TABLE 4 Glucose turnover parameters in rats maintained on high-fat or high-starch diets for 3 days or 3 wk 3 wk
3 day Parameter (mg • kg" 1 • mirr 1 )
High fat
Basal HGO and Rd Clamp GIR R« HGO
High starch
High fat
High starch
11.0 ± 0.5*
13.7 ± 0 . 9
11.0 ± 0.5
14.0 ± 1.2
12.8±0.6§ 21.5 ± 1.3
23.7 ± 1.5 23.2 ± 1.5
10.7 ± 1.0§ 17.1 ± 0.9§
26.1 ± 1.2 26.1 ± 0.4
8.7 ± 1.4§
- 0 . 5 ± 0.4
6.4 ± 1.2f
0.0 ± 1.2
Results are expressed as means ± SE. n = 6 for all groups except 3-wk fat (n = 4). In the basal state, glucose infusion rate (GIR) is 0, and therefore hepatic glucose output (HGO) equals peripheral glucose disposal (f?d). *P < 0.05, -fP < 0.01, §P < 0.001, vs. corresponding high-starch group.
was reduced by 35% vs. starch control (P < 0.001), and the increment in insulin-stimulated P.d above respective basal levels was also significantly reduced. Therefore, after 3 days of fat feeding, rats developed hepatic but not peripheral insulin resistance. However, by extending the diet treatment to 3 wk, both hepatic and peripheral insulin resistance were present. To confirm the implication that muscle insulin resistance was absent after 3 days of high-fat feeding, additional studies were performed to examine tissue-specific peripheral glucose metabolic index (R'g) as derived from [3H]-2-deoxyglucose uptake during the clamp (Fig. 2). Confirmatory to the previous results, there was no significant difference in R'g in any of the hindlimb muscles after 3 days of fat feeding, indicating no insulin resistance in these muscles (Student's t test). In fact, with ANOVA, high-fat-fed rats had a significantly higher insulin-stimulated Rg over the muscles sampled than the starch controls (P < 0.005). There was no significant
t
difference in insulin-stimulated Rg with 3 days high-fat feeding in either the diaphragm (fat vs. starch, 23.8 ± 2.3 vs. 27.8 ± 4.9 ixmol • 100 g~1 • mirr 1 ; NS) or the heart (fat vs. starch, 82.4 ± 10.5 vs. 89.3 ± 19.1 fjimol • 100 g~1 • mirr 1 ; NS). In contrast to the muscle results, after 3 days of fat feeding, Pig in white and brown adipose tissue was significantly reduced (Fig. 3). There were no significant differences in basal muscle glycogen levels after either 3 days or 3 wk on a high-fat diet (Table 5). There was no significant difference in muscle triglyceride content after 3 days on a high-fat diet. However, after 3 wk, fat feeding led to a twofold increase in mean muscle triglyceride content (P < 0.001 vs. 3-wk starch control). Furthermore, there was also a significant increase in muscle triglycerides from 3 days to 3 wk of high-fat feeding (P < 0.05). Additional studies examined intravenous glucose tolerance of diet-treated animals. Intravenous glucose (500
B
lo:
14106-
o o
2
^
2-
ffl -2 120 25-,
240 360
480
600
720
c
•2 120
30i
240
360
480
600
720
D
25
20-
20i 15 10 120
>¥* 240 360
480
600
720
10 120
240 360
Plasma Insulin (pM)
1400
480
600
720
FIG. 1. Hepatic glucose output (HGO) and peripheral glucose disposal (fid) in 3-day (A and C) and 3-wk (B and O) high-fat-fed ( • ) and highstarch-fed (D) rats in basal and euglycemic hyperinsulinemic clamp states, n = 4 - 6 rats in each group.
DIABETES, VOL. 40, NOVEMBER 1991
E.W. KRAEGEN AND ASSOCIATES
TABLE 5 Basal muscle glycogen and triglyceride in rats maintained on high-fat or high-starch diets for 3 days or 3 wk
20E o> o o "o E
3 wk
3 day High fat Muscle glycogen (% of 3-day highstarch group)* Muscle triglyceride (% of 3-day highstarch group)t
10-
O)
SOL
FG
EDL
Hlndlimb
PLA
V\Q
W3
Muscle
FIG. 2. Insulin-stimulated muscle glucose metabolic index (fig) in hindlimb muscles of 3-day fat-fed (n = 6, solid bars) and starch-fed (n = 6, open bars) rats. RQ, red quadriceps; SOL, soleus; RG, red gastrocnemius; EDL, extensor digitorum longus; PLA, plantaris; WQ, white quadriceps; WG, white gastrocnemius.
mg/kg) elevated plasma glucose to ~20 mM at 2-min postinjection in all groups. In animals after 3 days on diet, there was a tendency toward glucose intolerance in the high-fatfed group, although differences in K values were not quite significant (starch vs. fat, 0.089 ± 0.003 vs. 0.075 ± 0.006 min"1, P < 0.20). However, data supported the implication from the clamp studies that there may have been an impairment in suppression of HGO in the 3-day high-fat-fed group; significantly higher plasma glucose levels were observed in 15- and 20-min postinjection (e.g., 15-min starch vs. fat, 6.4 ± 0.4 vs. 8.6 ± 0.8 mM, P < 0.05), and the nadir in plasma glucose was lower in the high-starch-fed group than in the high-fat-fed group, either expressed as absolute levels {P < 0.01, data not shown) or as a decrease below respective basal levels (starch vs. fat, - 1 . 0 ± 0 . 1 vs. - 0 . 3 ± 0 . 1 mM, P
Muscle and hepatic insulin resistance are two major defects of non-insulin-dependent diabetes mellitus. Dietary factors may be important in the etiology of insulin resistance. We studied progressive changes in the development of high-fat-diet-induced insulin resistance in tissues of the adult male Wistar rat. In vivo insulin action was compared 3 days and 3 wk after isocaloric synthetic high-fat or high-starch feeding (59 and 10% cal as fat, respectively). Basal and insulin-stimulated glucose metabolism were assessed in the conscious 5- to 7-h fasted state with the euglycemic clamp (600 pM insulin) with a [3-3H]glucose infusion. Fat feeding significantly reduced suppressibility of hepatic glucose output by insulin after both 3 days and 3 wk of diet (P < 0.01). However, a significant impairment of insulin-mediated peripheral glucose disposal was only present after 3 wk of diet. Further in vivo [3H]-2-deoxyglucose uptake studies supported this finding and demonstrated adipose but not muscle insulin resistance after 3 days of high-fat feeding. Muscle triglyceride accumulation due to fat feeding was not significant at 3 days but had doubled by 3 wk in red muscle (P < 0.001) compared with starch-fed controls. By 3 wk, high-fat-fed animals had developed significant glucose intolerance. We conclude that fat feeding induces insulin resistance in liver and adipose tissue before skeletal muscle with early metabolic changes favoring an oversupply of energy substrate to skeletal muscle relative to metabolic needs. This may generate later muscle insulin resistance. Diabetes 40:1397-1403, 1991
R
educed potency of insulin action (insulin resistance) is a characteristic of non-insulin-dependent diabetes mellitus (NIDDM) and several other possibly related metabolic disorders such as obesity and hypertension. Although the contribution of insulin resistance to the NIDDM state is widely recognized, its precise etiology is not understood. However, it is likely that both genetic and life-style factors, i.e., diet and activity, are in-
DIABETES, VOL. 40, NOVEMBER 1991
volved (1). Insulin resistance occurs early in the development of NIDDM (2). There are conflicting reports favoring the primacy of hepatic (3) or peripheral (4) insulin resistance, but the general inference is that it might occur at different times in specific insulin target tissues. Our study is concerned with the development of diet-induced insulin resistance and glucose intolerance. Dietary factors such as highfat feeding have been implicated in the development of hepatic and muscle insulin resistance in rats (5-7), although we are unaware of any studies investigating the relative rate of development of diet-induced insulin resistance in specific tissues. Rats fed for 3 wk with a high-fat diet (either saturated or co-6 polyunsaturated fat) develop widespread peripheral and hepatic insulin resistance characterized by reduced uptake of glucose into skeletal muscle and adipose tissue and impaired suppressibility of hepatic glucose output by insulin (6-8). However, the relative time of onset of these metabolic responses and the mechanisms responsible are unknown. One contributing factor may be an increase in muscle triglyceride content (8), although the time relationship between an increase in muscle triglyceride and development of insulin resistance has not been established. Therefore, we compared in vivo insulin action in 3-day and 3-wk high-fat-fed rats to better understand the onset of insulin resistance in this model.
RESEARCH DESIGN AND METHODS Adult male Wistar rats (300-380 g) with free access to food and water were used in this study. Rats were housed in individual cages in a temperature-controlled room (22 ± 1°C) with a 12-h light-dark cycle (lights on 0600). All stud-
From the Garvan Institute of Medical Research, St. Vincent's Hospital, Sydney, New South'Wales, Australia. Address correspondence and reprint requests to Dr. E.W. Kraegen, Garvan Institute of Medical Research, St. Vincent's Hospital, Sydney, New South Wales 2010, Australia. Received for publication 4 September 1990 and accepted in revised form 28 May 1991.
1397
DEVELOPMENT OF DIET-INDUCED INSULIN RESISTANCE
ies were approved by the Garvan Institute Animal Experimental Ethics Committee and complied with the National Health and Medical Research of Australia guidelines for animal experimentation. Cannulations were performed under pentobarbital sodium anesthesia (60 mg/kg Nembutal, Abbott, Sydney). Cannulas were implanted in the left carotid artery and right jugular vein and exteriorized through the back of the neck. These lines were kept patent by filling with a 60% solution of polyvinylpyrrolidone (40,000 Mr Fluka Bioscientific, Sydney) made up in heparin sodium (1000 U/ml, Fisons, Sydney). Rats underwent a 5-day postoperative recovery period, during which there was close monitoring of dietary intake. Rats were fed isocalorically matched high-fat or highstarch diets (Table 1) for either a long-term (3-wk) or shortterm (3-day) period. Long-term rats were fed for 16 days precannulation on their diet with a precise record of dietary intake and spillage. All spillage was added to the next day's amount, but we found that little or no spillage occurred after the 1st day on the diet. These rats were cannulated as described above and given their specific diet for the 5-day postoperative recovery period. Short-term diet-treated rats were fed normal laboratory chow up to the morning of cannulation. After cannulation, all rats were fed the high-starch diet because it was of precisely known composition. After 2 days on the high-starch diet, rats were switched to the highfat diet or remained on the high-starch diet (controls) for the 3 days before acute study. On the morning of the 6th day, acute glucose turnover studies were performed 5-7 h after the removal of food. Sequential basal and euglycemic hyperinsulinemic glucose turnover studies as described elsewhere (9) were performed in conscious rats on each of the four groups (3-day or 3-wk high-fat- or high-starch-fed rats, n = 5-6 rats/ group). Briefly, infusion and sampling lines were connected to the cannulas, and rats were returned to individual cages and allowed a 30- to 40-min settling period before comm-
TABLE 1 Composition of high-fat and high-starch diets used in diet treatment Ingredient
High-fat diet (g/kg)
High-starch diet (g/kg)
Safflower oil Comstarch Casein Methionine Gelatin Wheat bran Vitamin mix* Mineral mix|
339 254 254 3 19 51 13 67
41 658 188 2 14 38 9 50
By calories, the fat diet contained 59% fat, 2 1 % protein, and 20% carbohydrate, and the starch diet contained 10% fat, 21% protein, and 69% carbohydrate. 'Supplied 3 g thiamine mononitrate, 3 g riboflavin; 3.5 g pyridoxine HCI; 15 g nicotinamide; 8 g d-calcium pantothenate; 1 g folic acid; 0.1 g d-biotin; 0.005 g cyanocobalamin; 0.013 g cholecalciferol; 0.025 g acetomenaphthone; 0.6 g vitamin A acetate; 25 g d-RRRa-tocopherol acetate and 10 g choline chloride per kg mix. tSupplied 30.5 g MgSO4 • 7H2O; 65.2 g NaCI; 105.7 g KCI; 200.2 g KH2PO4; 38.8 g MgCO3 • Mg(OH)2 • 3H2O; 40.0 g FeC6H5O7-5H2O; 512.4 g CaCO3; 0.8 g Kl; 0.9 g NaF; 1.4 g CuSO4 • 5H2O; 0.4 g MnSO4; and 0.05 g CoNO3 per kg mix.
1398
encing the study. [3-3H]glucose (Amersham, Sydney) was infused into the jugular vein as a priming dose (1.25 |xCi) followed by a constant infusion (0.05 |xCi/min). Steady-state glucose specific activity was measured 50, 60, and 70 min after the start of the infusion. The 60-min sample was also assayed for the basal plasma insulin, corticosterone, and free fatty acids. After the 70-min sample, rats were infused additionally with pork insulin (25 pmol • kg" 1 • min" 1 ; Actrapid, Novo, Bagsvaerd, Denmark), and euglycemic clamps were performed (10). Having reached a steady state, plasma glucose and glucose specific activity were determined 60, 70, and 80 min after the insulin infusion started. The final sample was assayed for the plasma parameters as above. Peripheral glucose disposal (R6) and hepatic glucose output (HGO) were calculated as Ra =
and
HGO = Ra - GIR,
where R*a is the tracer infusion rate (dpm/min SAg is the steady-state value of glucose specific activity, and GIR is the glucose infusion rate (0 in the basal state) taken over a 30-min period from 50 to 80 min after the start of the insulin infusion. Further studies were conducted on 3-day high-fat (n = 6) and 3-day high-starch (n = 6) -fed rats to assess relative tissue insulin sensitivity in a number of hindlimb muscles and fat depots. On the morning of the study, rats were anesthetized and cannulated (as described above), and euglycemic hyperinsulinemic clamps were combined with bolus [3H]-2deoxyglucose administration (11). Because these studies were essentially to confirm main data, they were performed in anesthetized rats for simplicity, although hindlimb blood flow and absolute glucose uptake are reduced under these circumstances (12). To investigate the accumulation of energy storage products in hindlimb muscle, a further series of studies was performed. Rats were fed either the 3-day or 3-wk high-fat or high-starch diet as described above, but only a single jugular cannula was implanted. On the morning of the study, a blood sample was taken to measure basal plasma parameters (i.e., glucose, insulin, nonesterified fatty acids (NEFA), triglyceride, and corticosterone), then the animal was killed with an overdose of pentobarbital sodium (60 mg), and soleus, plantaris, and red gastrocnemius muscle samples were rapidly removed and immediately freeze-clamped. Tissues were kept at -70°C until assayed for glycogen (plantaris; 13) and triglyceride. Triglycerides (soleus and red gastrocnemius) were extracted by a modification of the method of Denton and Randle (14). Intravenous glucose tolerance tests (IVGTT; 500 mg/kg glucose, 50% wt/vol solution, David Bull, Sydney) were performed on 5-h fasted chronically cannulated conscious rats maintained on 3-day high-fat (n = 6) or high-starch {n = 5) diets. Blood samples (300 jxL) were taken 0, 2, 5, 10, 15, 20, and 50 min after the intravenous glucose bolus for estimation of plasma glucose and insulin. Erythrocytes were resuspended and returned to the animal after the 0- and 20min sample. Glucose disappearance was approximated by first-order kinetics up to 15 min, and a K value was estimated with loglinear regression over this period. For comparison,
DIABETES, VOL. 40, NOVEMBER 1991
E.W. KRAEGEN AND ASSOCIATES
TABLE 2 Postcannulation dietary intake 3 wk diet
3-day diet Days postcannulation 1 2 3 4 5
High fat 63 79 97 100 100
±6 ± 8 ± 1 ±0 ± 0
High starch 59 78 88 99 99
± ± ± ± ±
High fat
8 6 7 1 1
74 93 98 98 97
± ± ± ± ±
7 3 2 2 2
High starch 88 94 97 99 99
± ± ± ± ±
3 3 2 1 1
Results are expressed as a percentage of normal precannulation dietary intake, n, 9-13 animals; diets are high fat or high starch as indicated.
K values for 3-wk high-fat and high-starch -fed animals (both n = 7) were calculated from previously obtained data (15). Blood and plasma glucose concentrations were determined with a glucose analyzer (model 23AM, Yellow Springs, OH). Plasma samples for determination of insulin, corticosterone, and NEFA were kept at - 20°C until assayed. Insulin (16) and corticosterone (17) were assayed by radioimmunoassay. Enzymatic colorimetric kits were used to determine plasma concentrations of free fatty acids (NEFA-C code no. 279-75409, Wako, Osaka, Japan). Plasma samples for determination of tracer concentration were deproteinized immediately in 2.8% ZnSO4 and saturated Ba(OH)2. An aliquot of supernatant was dried down to remove tritiated water and redissolved in an aqueous scintillant (Picofluor 30, Packard, Rockville, MD). Radioactivity counting was performed in a liquid-scintillation spectrophotometer (Beckman, Fullerton, CA). All experimental data were expressed as means ± SE. Statistical analyses were performed with paired or unpaired two-tailed Student's t tests or a two-way analysis of variance (ANOVA) as appropriate. P < 0.05 was considered significant. RESULTS
The postcannulation dietary intake of both the 3-day and 3wk high-fat- and high-starch-fed rats expressed as a per-
centage of normal precannulation dietary values was similar to that reported previously for starch-fed rats (9; Table 2). Dietary intake increased progressively in each group over the first 2 days of the postcannulation recovery period. On the 1st day, dietary intake was reduced to - 6 0 % of normal in rats not previously on the diet (3-day rats) and - 8 0 % in rats used to the diet (3-wk rats). Dietary intake increased over day 2 and was 90-100% of normal for the next 3 days before the study (i.e., normal over the full test period of shortterm feeding). There was no significant difference in energy intake among the four groups over the 3-day period before acute study. Plasma concentrations of glucose, insulin, NEFA, and triglyceride during the basal and clamp phases of the study are displayed in Table 3. There was no significant effect of fat feeding (either 3 day or 3 wk) on the absolute basal or clamp plasma concentrations of any of these parameters. Although basal NEFA levels were similar, there was a tendency toward reduced suppression of NEFA by insulin with fat feeding (significant suppression in both starch groups, P < 0.05, but not significant in either fat-fed group). Basal corticosterone levels were similar in all four groups (mean 40-70 nM, data not shown). Basal and clamp glucose turnover (HGO and Ra) and clamp glucose infusion rates are displayed in Table 4, and HGO and P>d are plotted against prevailing insulin levels in Fig. 1. Fat feeding reduced basal glucose turnover (HGO and P>d) by - 2 0 % at 3 days (P < 0.05) and at 3 wk (P < 0.09). In the insulin-stimulated state (-100 mU/L), fat feeding induced whole-body insulin resistance (as indexed by the clamp GIR both at 3 days (reduced by 46%, P < 0.001) and 3 wk (reduced by 59%, P < 0.001). This was associated with significantly less suppression of HGO during the clamp both at 3 days (fat vs. starch, 26 ± 9 vs. 106 ± 4% suppression of basal HGO, P < 0.001) and 3 wk (fat vs. starch, 43 ± 10 vs. 102 ± 10% suppression of basal HGO, P < 0.01). However, there was no significant difference in insulin-stimulated P.d between high-fat- and high-starch-fed rats after 3 days on the diet (Fig. 1; Table 4). After 3 wk of fat feeding, peripheral insulin resistance developed. Insulin-stimulated P.d
TABLE 3 Basal and euglycemic hyperinsulinemic clamp plasma parameters in rats maintained on high-fat or high-starch diets for 3 days or 3 wk 3 day High fat Basal Glucose (mM) Blood Plasma Insulin (pM) NEFA (mM) Triglyceride (mM) Clamp Glucose (mM) Blood Plasma Insulin (pM) NEFA (mM)
3 wk High starch
High fat
High starch
4.6 7.2 240 0.73 0.92
± ± ± ± ±
0.1 0.2 30 0.13 0.06
4.5 7.0 276 0.83 0.88
± ± ± ± ±
0.2 0.2 42 0.13 0.11
4.9 7.3 186 0.94 0.79
± ± ± ± ±
0.1 0.3 24 0.08 0.03
4.5 6.9 240 1.28 0.92
± ± ± ± ±
0.1 0.2 60 0.14 0.08
4.5 7.2 700 0.73
± ± ± ±
0.1 0.3 30 0.08
4.3 6.8 680 0.49
± ± ± ±
0.2 0.2 20 0.10*
4.7 7.0 590 0.66
± ± ± ±
0.1 0.2 70 0.17
4.6 6.9 580 0.74
± ± ± ±
0.1 0.2 70 0.11*
Results are expressed as means ± SE. n = 6 for all groups except 3-wk fat (n = 4). NEFA, nonesterified fatty acids. *P < 0.05 vs. basal value in same group.
DIABETES, VOL. 40, NOVEMBER 1991
1399
DEVELOPMENT OF DIET-INDUCED INSULIN RESISTANCE
TABLE 4 Glucose turnover parameters in rats maintained on high-fat or high-starch diets for 3 days or 3 wk 3 wk
3 day Parameter (mg • kg" 1 • mirr 1 )
High fat
Basal HGO and Rd Clamp GIR R« HGO
High starch
High fat
High starch
11.0 ± 0.5*
13.7 ± 0 . 9
11.0 ± 0.5
14.0 ± 1.2
12.8±0.6§ 21.5 ± 1.3
23.7 ± 1.5 23.2 ± 1.5
10.7 ± 1.0§ 17.1 ± 0.9§
26.1 ± 1.2 26.1 ± 0.4
8.7 ± 1.4§
- 0 . 5 ± 0.4
6.4 ± 1.2f
0.0 ± 1.2
Results are expressed as means ± SE. n = 6 for all groups except 3-wk fat (n = 4). In the basal state, glucose infusion rate (GIR) is 0, and therefore hepatic glucose output (HGO) equals peripheral glucose disposal (f?d). *P < 0.05, -fP < 0.01, §P < 0.001, vs. corresponding high-starch group.
was reduced by 35% vs. starch control (P < 0.001), and the increment in insulin-stimulated P.d above respective basal levels was also significantly reduced. Therefore, after 3 days of fat feeding, rats developed hepatic but not peripheral insulin resistance. However, by extending the diet treatment to 3 wk, both hepatic and peripheral insulin resistance were present. To confirm the implication that muscle insulin resistance was absent after 3 days of high-fat feeding, additional studies were performed to examine tissue-specific peripheral glucose metabolic index (R'g) as derived from [3H]-2-deoxyglucose uptake during the clamp (Fig. 2). Confirmatory to the previous results, there was no significant difference in R'g in any of the hindlimb muscles after 3 days of fat feeding, indicating no insulin resistance in these muscles (Student's t test). In fact, with ANOVA, high-fat-fed rats had a significantly higher insulin-stimulated Rg over the muscles sampled than the starch controls (P < 0.005). There was no significant
t
difference in insulin-stimulated Rg with 3 days high-fat feeding in either the diaphragm (fat vs. starch, 23.8 ± 2.3 vs. 27.8 ± 4.9 ixmol • 100 g~1 • mirr 1 ; NS) or the heart (fat vs. starch, 82.4 ± 10.5 vs. 89.3 ± 19.1 fjimol • 100 g~1 • mirr 1 ; NS). In contrast to the muscle results, after 3 days of fat feeding, Pig in white and brown adipose tissue was significantly reduced (Fig. 3). There were no significant differences in basal muscle glycogen levels after either 3 days or 3 wk on a high-fat diet (Table 5). There was no significant difference in muscle triglyceride content after 3 days on a high-fat diet. However, after 3 wk, fat feeding led to a twofold increase in mean muscle triglyceride content (P < 0.001 vs. 3-wk starch control). Furthermore, there was also a significant increase in muscle triglycerides from 3 days to 3 wk of high-fat feeding (P < 0.05). Additional studies examined intravenous glucose tolerance of diet-treated animals. Intravenous glucose (500
B
lo:
14106-
o o
2
^
2-
ffl -2 120 25-,
240 360
480
600
720
c
•2 120
30i
240
360
480
600
720
D
25
20-
20i 15 10 120
>¥* 240 360
480
600
720
10 120
240 360
Plasma Insulin (pM)
1400
480
600
720
FIG. 1. Hepatic glucose output (HGO) and peripheral glucose disposal (fid) in 3-day (A and C) and 3-wk (B and O) high-fat-fed ( • ) and highstarch-fed (D) rats in basal and euglycemic hyperinsulinemic clamp states, n = 4 - 6 rats in each group.
DIABETES, VOL. 40, NOVEMBER 1991
E.W. KRAEGEN AND ASSOCIATES
TABLE 5 Basal muscle glycogen and triglyceride in rats maintained on high-fat or high-starch diets for 3 days or 3 wk
20E o> o o "o E
3 wk
3 day High fat Muscle glycogen (% of 3-day highstarch group)* Muscle triglyceride (% of 3-day highstarch group)t
10-
O)
SOL
FG
EDL
Hlndlimb
PLA
V\Q
W3
Muscle
FIG. 2. Insulin-stimulated muscle glucose metabolic index (fig) in hindlimb muscles of 3-day fat-fed (n = 6, solid bars) and starch-fed (n = 6, open bars) rats. RQ, red quadriceps; SOL, soleus; RG, red gastrocnemius; EDL, extensor digitorum longus; PLA, plantaris; WQ, white quadriceps; WG, white gastrocnemius.
mg/kg) elevated plasma glucose to ~20 mM at 2-min postinjection in all groups. In animals after 3 days on diet, there was a tendency toward glucose intolerance in the high-fatfed group, although differences in K values were not quite significant (starch vs. fat, 0.089 ± 0.003 vs. 0.075 ± 0.006 min"1, P < 0.20). However, data supported the implication from the clamp studies that there may have been an impairment in suppression of HGO in the 3-day high-fat-fed group; significantly higher plasma glucose levels were observed in 15- and 20-min postinjection (e.g., 15-min starch vs. fat, 6.4 ± 0.4 vs. 8.6 ± 0.8 mM, P < 0.05), and the nadir in plasma glucose was lower in the high-starch-fed group than in the high-fat-fed group, either expressed as absolute levels {P < 0.01, data not shown) or as a decrease below respective basal levels (starch vs. fat, - 1 . 0 ± 0 . 1 vs. - 0 . 3 ± 0 . 1 mM, P
