Muscle-Type Specific Intramyocellular and Hepatic Lipid Metabolism During Starvation in Wistar Rats

Male Wistar rats (HsdCpd:WU) were obtained from Charles River (Sulzfeld, Germany) and were used at the age of 16 weeks for the study. Animals were housed in pairs at 20°C and on a 12-h light-dark cycle with ad libitum access to water.

All experimental procedures were conducted according to the German Animal Protection Law. The study protocol ran for 5 days and consisted of the following study points: normal fed; 12, 24, 48, and 72 h of starvation; and refed. For each study point, body weight, metabolic serum, and tissue (muscle and liver) parameters, as well as IMCLs and HepCLs, were measured.

Eight rats were used for normal fed; 24, 48, and 72 h starvation; and refed conditions to obtain intraindividually IMCL values by in vivo ¹H-MRS of the soleus (oxidative), tibialis anterior (glycolytic), and extensor digitorum longus (EDL; intermediary) muscles. Eight additional rats were used for the parallel intraindividual monitoring of HepCL by in vivo ¹H-MRS for the 5-day protocol. In both groups, metabolic serum parameters were obtained daily. Because rats are night-active and consume food predominantly during the dark phase, eight additional rats were kept on a reverse light-dark cycle for 1 week before the MRS studies and were used for the measurements of their IMCL and HepCL levels at the end of the light phase after a 12-h starvation period.

In parallel groups (n = 8 for each study point of the study protocol), blood samples for determination of metabolic serum parameters, as well as defined tissue samples, were collected during terminal isoflurane anesthesia. Soleus, tibialis anterior, and EDL of both hind legs, as well as an ∼1-g tissue sample from the longissimus dorsi muscle and the liver, were isolated and immediately frozen in liquid nitrogen for determination of malonyl-CoA (soleus, tibialis anterior, EDL, liver), LCACoAs (longissimus dorsi, liver) and total creatine (tCr) (soleus, tibialis anterior, EDL). Forty-eight additional rats were used for determination of liver weight and the content (liver) of triglycerides, glycogen, and water.

In vivo MRS studies in anesthetized rats were performed on a 7-Tesla Biospec system (Bruker BioSpin, Ettlingen, Germany) as described previously (28). Briefly, IMCLs and HepCLs were observed using a single-voxel spectroscopy sequence. IMCLs were determined in the soleus, tibialis anterior, and EDL muscles and expressed as IMCL-to-tCr ratio. HepCLs were measured without water suppression and were expressed as the ratio of the fat-to-water signal in percent.

To investigate the influence of starvation on insulin sensitivity, euglycemic-hyperinsulinemic glucose clamp studies were performed on three groups of rats (n = 7): 1) 12 h (reverse light-dark cycle), 2) 24 h, and 3) 72 h of starvation. The study was performed in anesthetized animals as described previously (28). Briefly, the study design consisted of a 120-min basal period followed by a 2-h euglycemic-hyperinsulinemic clamp. Additionally, the rats were given a constant infusion of [U-¹³C]glucose (1 mg · kg⁻¹ · min⁻¹) to estimate rates of glucose production and utilization (29). Every 15 min, blood samples were obtained via tail-tip bleeds for determination of glucose enrichment. Enrichments were calculated from the ratio of [U-¹³C]glucose/[¹²C]glucose during the last 30 min of the basal period and during the last 60 min of the clamp (i.e., steady-state conditions). This ratio was determined by gas chromatography-mass spectrometry (GC-MS) analyses of derivatized glucose from blood samples following literature protocols (29,30).

Blood samples were taken at 30 min for determination of baseline insulin levels and at 180, 210, and 240 min for determination of insulin levels under hyperinsulinemia. At 120 and 240 min, additional blood samples for determination of FFAs were taken. Animals were killed by pentobarbital overdose.

During steady state, the total glucose appearance in the circulation (R_a) equals the rate of disappearance of glucose (R_d), and was calculated by dividing the [U-¹³C]glucose infusion rate by the steady-state value of glucose enrichment. EGP was calculated as follows: EGP equals R_d minus the glucose infusion rate. For each animal, two values of EGP were obtained: one during basal conditions and one during the euglycemic-hyperinsulinemic clamp (31).

Blood samples were taken for determination of blood glucose and lactate from the tip of the tail. Samples for monitoring of FFAs, HO-butyrate, acetoacetate, triglycerides, cholesterol, and insulin were obtained by puncture of the sternal venous plexus during short-term isoflurane anesthesia. These blood samples were placed in serum gel tubes and were centrifuged at 6°C at 5,000 rpm.

Standard procedures were used to determine blood glucose, lactate, FFAs, HO-butyrate, acetoacetate, triglycerides, and cholesterol, as well as hepatic contents of glycogen, triglycerides, and water (32). Plasma insulin concentrations were assayed with an antibody radioimmunoassay kit obtained from Linco.

High-performance liquid chromatography (HPLC) analysis were performed using a Waters Alliance 2690 system (2487 detector; Millennium 2010 chromatographic manager).

Measurement of tCr in soleus, tibialis anterior, and EDL was performed as described previously (33). Linear detection was achieved between 0 and 50 μmol/ml for both analytes, with r² values of >0.999 for Cr and Cr-P. The retention times for Cr and Cr-P were 1.24 min (relative standard deviation [RSD] 0.78%) and 2.4 min (RSD 0.98%), respectively. Total Cr results represent the sum of detectable Cr plus Cr-P relative to the amount of dry muscle mass used for homogenization.

Analyses of LCACoAs in muscle (longissimus dorsi) and liver were performed in freeze-dried tissue samples. The analysis was performed using a modified method described by Deutsch et al. (34). To ∼100 mg (50-mg liver samples) of dried tissue, 2 ml of 100 mmol/l KH₂PO₄ (pH 2) and 20 μl of a 100-μmol/l aqueous solution of the internal standard (C17:0-CoA) were added. Subsequently the tissue was homogenized and 200 μl of 4.5% KOH, 2 ml of 2-propanol, 0.25 ml of saturated (NH₄)₂SO₄, and 4 ml of acetonitrile were added for extraction. After 15 min of extraction, the LCACoAs were separated by centrifugation at 5,000 rpm (5 min). The supernatant was diluted 1:2 with 100 mmol/l KH₂PO₄ (pH 4.9) and applied to solid-phase extraction on a Bond Elut PPL-Column, 100 mg of 3 ml (Varian, Darmstadt, Germany), conditioned with 5 ml of acetonitrile and 2 ml of 25 mmol/l KH₂PO₄ (pH 4.9). Elution of the bound LCACoAs was performed after washing the cartridge (4 ml of 25 mmol/l KH₂PO₄, pH 4.9) by passing 1 ml of 40 mmol/l KH₂PO₄ (pH 4.9) in 60% acetonitrile without applying pressure. A Synergi 4-μm hydro (3-mm) column (Phenomenex, Aschaffenburg, Germany) was used for chromatography (at 40°C). Detection of absorbance occurred at 260 nm. Separation of acyl-CoAs was performed in 30 μl of the extract making use of a gradient system including two mobile phases for the muscle samples: 25 mmol/l KH₂PO₄ (pH 4.9) (buffer A); 100% acetonitrile, plus a third phase for the liver samples (buffer B); and 100% methanol (buffer C). For the muscle samples, a gradient consisting of buffer A (70–30%) and buffer B (30–70%) for 30 min was used. The flow rate was set at 0.4 ml/min. For the liver samples, a gradient consisting of buffer A (40–20%) and buffer B (30–50%) plus 30% of buffer C for 30 min was used; the flow was set at 0.5 ml/min. Calibration was performed using selected pure LCACoAs as their respective Li-salts. Linear detection was achieved between 0.05 and 1 μmol/l for all analytes, with r² values of >0.998 in buffer and >0.98 when calibrators had been spiked in muscle extracts. The sums of the major species (C16:0, C16:1, C18:0 [liver], C18:1, C18:2) were referred to as total LCACoAs. In muscle samples, C18:2 and C20:4 CoA esters could not be separated using the described method and, thus, were eluted in one peak.

Analysis of malonyl-CoA in tissues of muscle (soleus, tibialis anterior, EDL) and liver was performed after homogenization of ∼100 mg of the frozen tissues in 0.5 ml perchloric acid (0.6 mol/l) and centrifugation at 5,000 rpm (10 min). The following solutions were added to 0.5 ml of the resulting perchloric acid supernatant: 5 μl of 2 mmol/l diothiothreitol, 10 μl of the internal standard (isobutyryl-CoA), and 100 μl of 4 mol/l sodium acetate. After mixing and short centrifugation, 600 μl of the prepared sample was applied to solid-phase extraction. The column (Bond Elut C18 3CC/200 mg) had been preconditioned with 3 ml methanol and washed with 3 ml water/HCl (pH 3). After the sample was run through the column, the matrix was again washed with 3 ml water/HCl and then dried by applying vacuum. Elution of malonyl-CoA was performed with 0.1 mol/l NH₄-acetate in 65% ethanol. After drying the eluate at 40°C under a nitrogen stream, the material was reconstituted with 400 μl water. HPLC separation was performed using a modified method described by Saddik et al. (35). An aliquot (100 μl) of each sample was passed through a Phenomenex Aqua C18 (150 × 4.6, 5-μm particle size) column. Detection of absorbance occurred at 254 nm; the flow rate was set at 1 ml/min. A gradient was initiated using two buffers: buffer A consisted of 25 mmol/l NaH₂PO₄ (pH 5), and buffer B was a mixture of 80% 20 mmol/l NaH₂PO₄ (pH 5) and 20% acetonitrile. Initially, 97% buffer A was held constant for 2.5 min. Subsequently, buffer B increased to 25% within 5.5 min, thereafter to 30% within 7 min, and finally to 100% within 2 min. This condition was held constant for 2.5 min before reequilibration of the column. Calibration was performed using malonyl-CoA at preset concentrations between 0.1 and 10 nmol/ml. Linear detection was achieved with r² values of >0.98 in buffer and >0.96 when spiked in muscle homogenate. Malonyl-CoA elutes after 9.7 min of retention.

Data are presented as means ± SE. Statistical differences were determined using a one-way ANOVA (for repeated measures, if applicable) followed by a post hoc analysis with Bonferroni’s correction when testing for differences between three or more experimental groups (SigmaStat; Jandel Scientific, Erkrath, Germany). A P < 0.05 was considered to be statistically significant.

Changes in body weight and metabolic serum parameters during the 5 days of the study period were similar for the MRS (IMCL, HepCL) and satellite groups (representatively shown for the IMCL group in Fig. 1). During transition from the fed to the fasted state, a decrease in blood glucose, insulin, and triglycerides was observed, whereas FFAs and ketone bodies increased, which is consistent with metabolism during starvation. Refeeding after 3 days of starvation caused an immediate reversal of all parameters (Fig. 1). Malonyl-CoA decreased during starvation in soleus, tibialis anterior, and EDL, and tended to increase after refeeding in tibialis anterior and EDL (Fig. 2A). The values for tCr in soleus, tibialis anterior, and EDL were constant during the study period (Table 1). Thus, tCr could be used as a stable reference marker to express IMCL levels as the IMCL-to-tCr ratio in the ¹H-MR-spectra.

Despite significant elevations of FFAs during starvation, IMCLs in soleus remained relatively constant (Fig. 3A). However, in tibialis anterior and EDL, IMCLs increased significantly by 170 and 450% after 72 h of starvation, respectively. Refeeding caused an immediate drop of the elevated IMCLs in both muscles. The intramyocellular level of total LCACoAs in longissimus dorsi, as well as the relative contribution of the respective lipid types, remained constant during the whole study period, as indicated in Table 2.

Hepatic malonyl-CoA did not vary during starvation, but increased dramatically after refeeding (Fig. 2B). Despite the obvious dynamics in plasma lipids (FFAs, triglycerides) during the study period, HepCL levels remained constant (Fig. 3B). As HepCL levels were quantified as fraction of the water signal but hepatic glycogen (and, therefore, the content of glycogen-bound water) changed dramatically during starvation, we determined biochemically hepatic triglyceride levels (Table 3). The water content (mg/g wet wt) remained constant in spite of dramatic reductions of total organ weight and hepatic glycogen levels. Relative (mg/g wet wt) as well as total hepatic triglycerides (mg/liver) did not change during starvation, confirming the HepCL data. Although during early starvation lipid content in the liver tended to increase slightly as seen both MR-spectroscopically and biochemically, the changes failed to be significant.

In contrast to the muscle, hepatic LCACoAs demonstrated significant changes during starvation (Table 2), with a peak value of total LCACoAs after 24 h of starvation. Oleoyl-CoA (C18:1) remained constant during the study, as did palmitoyl-CoA (C16:0), which only showed an increase after refeeding. All other determined types of LCACoAs, except for palmitoleoyl-CoA (C16:1), showed an increase during starvation and a reversal after refeeding. In contrast, palmitoleoyl-CoA (C16:1) decreased during starvation and increased after refeeding significantly (Table 2).

Prolonged starvation reduced basal rates of glucose turnover (R_d), which equals basal EGP, but did not affect peripheral insulin sensitivity in Wistar rats during euglycemic-hyperinsulinemic clamp conditions. Rats starved for 12, 24, and 72 h demonstrated similar values for peripheral glucose disposal (R_d) and antilipolysis (FFA suppression) during the clamp, whereas EGP was lower after prolonged starvation (Table 4).

The present study was performed to investigate the physiological dynamics and interplay between serum triglycerides and FFAs, HepCLs, and IMCLs in muscle of different fiber types. The newly developed ability to determine noninvasively HepCL and IMCL contents by in vivo ¹H-MRS allows investigations for a better understanding of lipid metabolic regulation during periods of starvation and refeeding in rats. Our study resulted in several unexpected findings: 1) muscles with different fiber types coped with starvation-induced elevated FFA levels differently; 2) increased lipolysis in adipocytes during starvation occurred in parallel to increased FFA reesterification in glycolytic (tibialis anterior) and intermediary (EDL) muscles but not in oxidative muscle (soleus); 3) IMCL levels changed rapidly in tibialis anterior and EDL within hours; 4) HepCL levels remained constant despite increased levels of total LCACoAs; and 5) hepatic palmitoleoyl-CoA (C16:1) showed an inverse dynamic to all other investigated types of LCACoAs.

The changes of metabolic serum parameters during starvation have been known for decades (rev.in 36). During starvation, blood glucose and, subsequently, serum insulin levels decrease, while an increased lipolysis in adipocytes results in high FFA and subsequently elevated ketone body levels produced by the liver. In the presence of low insulin during starvation, lipoprotein lipase in the muscle is more active than in the adipose tissue, and fatty acids from triglyceride-rich VLDLs are shunted in addition to the readily available FFAs into skeletal muscle cells in order to produce energy by oxidation. At low glucose levels, malonyl-CoA, the endogenous inhibitor of carnitine palmitoyl transferase-1 (CPT-1) and the rate-limiting enzyme in fatty acid oxidation in liver (2,37) and muscle (38,39), is decreased and thereby directs LCACoAs into mitochondria for oxidation.

The changes in malonyl-CoA in this study, especially for the oxidative muscle soleus, were not as impressive as could be expected from their regulatory fuel-sensing role. Recently, a malonyl-CoA-insensitive pathway of lipid oxidation in oxidative muscles was reported (40). Nevertheless, in our study malonyl-CoA levels in muscle demonstrated a dynamic profile during the study period, which might be consistent with its role as an endogenous regulator of fatty acid oxidation predominantly for the nonoxidative and intermediary type muscles tibialis anterior and EDL.

IMCL levels were relatively high in soleus as compared with tibialis anterior and EDL in the fed state and did not change during starvation, implying the capacity of soleus to cope with high FFA values. The same observation was made with HepCLs, which also remained constant during prolonged starvation. However, IMCL values in tibialis anterior and EDL rose significantly and rapidly during starvation, reflecting an increased FFA reesterification in these muscles in parallel to an increased lipolysis in adipocytes. Recently, it was concluded from systemic and regional glycerol and FFA kinetic studies that FFAs must be reesterified in tissues other than liver after a 60-h fast in men, and muscle tissue was suggested as an important site of FFA reesterification (41). In fact, our study in rats also demonstrates increasing IMCL levels, which would support the concept of FFA reesterification in muscle during fasting, but only in muscles with nonoxidative fiber composition (tibialis anterior, EDL). During refeeding, after 72 h of starvation, the elevated IMCLs in tibialis anterior and EDL returned within 24 h to the very low values observed on day 1 before fasting, thus indicating the rapid physiological dynamics of this lipid pool.

Assuming that longissimus dorsi reflects a mixed fiber-type muscle and therefore mirrors all three investigated muscles measured by ¹H-MRS in this study, the total amount as well as the individual types of LCACoAs in the muscles remained constant in spite of significant changes in plasma FFA and IMCL levels in tibialis anterior and EDL during starvation. Therefore, this increase of IMCL levels in the nonoxidative muscle fibers might be an adaptive or protective cellular mechanism to avoid accumulation of LCACoAs in the cytoplasm, because LCACoAs have been identified as major candidates affecting insulin signaling (42,43) and as a major link to insulin resistance in obesity and diabetes (19). In contrast, in the liver there was a significant increase in total LCACoAs based predominantly on the increases of linoleoyl-CoA (C18:2) and stearoyl-CoA (C18:0). Only palmitoleoyl-CoA (C16:1) showed an inverse dynamic with significant decreases during starvation and a significant increase on refeeding that exceeded the level of the normal fed state. This decrease might be caused by a preferable oxidation of palmitoleoyl-CoA, as well as by a reduced expression of stearoyl-CoA desaturase (SCD) due to low insulin levels during starvation (44). Interestingly, this regulation appeared only in the liver and not in the muscle. Recently it has been described that SCD1(−/−) mice are more insulin sensitive than wild-type mice (45). Thus it can be speculated that the decrease in palmitoleoyl-CoA is an adaptive mechanism to maintain insulin sensitivity during starvation. However, the underlying mechanism is far from being understood, as monounsaturated fatty acids are generally believed to be metabolically more beneficial than saturated ones (46).

In contrast to the prevailing view that starvation induces insulin resistance (47), but in line with constant values for FFAs, HepCLs, IMCLs, and muscle LCACoAs from 12 to 72 h of starvation in this study, there were no significant changes detectable in R_d and antilipolysis during the euglycemic-hyperinsulinemic glucose clamp at 12, 24, and 72 h of starvation. During the basal period of the clamp studies, the values for FFAs were lower and those for insulin were higher than those in the respective MRS studies, which might be related to the pentobarbital anesthesia during the clamp experiment. Although there was a tendency toward reduced values for R_d during the hyperinsulinemic condition of the clamp with prolonged duration of starvation, the value after 72 h failed to be significant (P = 0.064 vs. 12 h). The increased values for EGP after 12 h of starvation during basal and clamp conditions as compared with the values after 24 and 72h starvation reflected more of a residual enteral glucose absorption after this short time of starvation rather than exclusive EGP by the liver or the kidneys. The elevated plasma insulin level after 12 h of starvation, predominantly during the basal period, also supports this explanation. Furthermore, no decrease in insulin sensitivity due to starvation was reported in rats (24,25), although their cellular glucose metabolism had changed from glycolysis to glycogen synthesis during maximal insulin infusion after prolonged starvation. In this context, it is important to stress that normal insulin-sensitive Wistar rats were used in this study. Although IMCLs demonstrated strong dynamics in this study, all absolute IMCL-to-tCr values were still below 1 and, therefore, lower than in fed insulin-resistant obese Zucker diabetic fatty rats (28).

It is concluded that peripheral lipolysis during starvation is adjusted to the energy demands of oxidative muscles (soleus) and that nonoxidative muscles (tibialis anterior, EDL) reesterify excess FFAs to maintain constant levels of LCACoAs in order to prevent impairment of intracellular signaling. Whether there is a threshold level of IMCLs associated to insulin resistance needs further investigation in animal disease models of insulin resistance.

We thank M. Roden, Vienna, for helpful discussions during the preparation of the manuscript.