Long–echo time MR spectroscopy for skeletal muscle acetylcarnitine detection

Animal models suggest that acetylcarnitine production is essential for maintaining metabolic flexibility and insulin sensitivity. Because current methods to detect acetylcarnitine involve biopsy of the tissue of interest, noninvasive alternatives to measure acetylcarnitine concentrations could facilitate our understanding of its physiological relevance in humans. Here, we investigated the use of long–echo time (TE) proton magnetic resonance spectroscopy (¹H-MRS) to measure skeletal muscle acetylcarnitine concentrations on a clinical 3T scanner. We applied long-TE ¹H-MRS to measure acetylcarnitine in endurance-trained athletes, lean and obese sedentary subjects, and type 2 diabetes mellitus (T2DM) patients to cover a wide spectrum in insulin sensitivity. A long-TE ¹H-MRS protocol was implemented for successful detection of skeletal muscle acetylcarnitine in these individuals. There were pronounced differences in insulin sensitivity, as measured by hyperinsulinemic-euglycemic clamp, and skeletal muscle mitochondrial function, as measured by phosphorus-MRS (³¹P-MRS), across groups. Insulin sensitivity and mitochondrial function were highest in trained athletes and lowest in T2DM patients. Skeletal muscle acetylcarnitine concentration showed a reciprocal distribution, with mean acetylcarnitine concentration correlating with mean insulin sensitivity in each group. These results demonstrate that measuring acetylcarnitine concentrations with ¹H-MRS is feasible on clinical MR scanners and support the hypothesis that T2DM patients are characterized by a decreased formation of acetylcarnitine, possibly underlying decreased insulin sensitivity.

The use of proton magnetic resonance spectroscopy (¹H-MRS) in human skeletal muscle has been instrumental in establishing the importance of ectopic fat accumulation in the development of type 2 diabetes mellitus (T2DM). It was found in a large number of studies that intramyocellular lipid (IMCL) levels are inversely related to insulin sensitivity (1, 2). However, the same methodology also resulted in the identification of the so-called athletes’ paradox: endurance-trained athletes are, despite being very insulin sensitive, also characterized by high IMCL levels (3). These findings have led to the development of new hypotheses to explain the relation between fat accumulation in muscle and insulin sensitivity.

One of the interesting and novel hypotheses suggests a role for carnitine metabolism (4, 5). It has long been known that carnitine permits mitochondrial import of long-chain fatty acids for subsequent β-oxidation (6). However, it has recently been suggested that carnitine may also play a crucial regulatory role in substrate switching and glucose homeostasis (4, 5, 7). Acetylcarnitine is formed in conditions in which acetyl-CoA formation, either as end product of glycolysis or β-oxidation, exceeds its entry into the tricarboxylic (TCA) cycle. Free carnitine can act as a sink for excess acetyl groups in a reversible reaction catalyzed by the enzyme carnitine acetyltransferase (CRAT) (7). Acetylcarnitine, like other acylcarnitine esters, can readily be exported out of the mitochondria (Figure 1).

Figure 1

Formation of acetylcarnitine. When acetyl-CoA formation exceeds use by the TCA cycle, carnitine can function as a sink for excess acetyl groups inside the mitochondria, thereby forming acetylcarnitine. This reversible reaction is catalyzed by the enzyme CRAT and leads to the release of free CoA. The formation of acetylcarnitine can help to keep the mitochondrial acetyl-CoA/free CoA ratio low, which is essential to sustaining TCA cycle flux and PDH activity. Acetylcarnitine can be transformed back to acetyl-CoA or can be exported outside the mitochondria. The protons contributing to the resonance of acetylcarnitine at 2.13 ppm in ¹H-MRS are highlighted in the molecule structure. CPT1, carnitine palmitoyltransferase 1; FFA, free fatty acids; LCFA, long-chain fatty acids.

The formation of acetylcarnitine helps to keep the mitochondrial acetyl-CoA/free CoA ratio low. A low acetyl-CoA/free CoA ratio is needed to maintain pyruvate dehydrogenation (PDH) activity (8), which is known to control the rate of aerobic carbohydrate oxidation. A compromised capacity to generate acetylcarnitine, either due to reduced CRAT activity or low carnitine concentrations, may reduce PDH activity, hence reducing oxidative degradation of glucose, a major concern in insulin-resistant muscle. Compromised mitochondrial entrance of pyruvate is reflected in decreased metabolic flexibility, an early detectable feature of insulin-resistant muscle. Indeed, Crat-knockout mice, which are unable to convert acetyl-CoA to acetylcarnitine, are characterized by decreased glucose tolerance. On the other hand, gain-of-function experiments of CRAT in primary human myotubes showed increased acetylcarnitine efflux and higher PDH activity (5). Similarly, a reduced availability of free carnitine caused by high-fat overfeeding hampered acetylcarnitine formation in rats, and carnitine supplementation was able to reverse diet-induced mitochondrial dysregulation, including restoration of PDH activity. Furthermore, acetylcarnitine formation was increased with increased levels of free carnitine in skeletal muscle of rats and in primary human skeletal myotubes (4). To test the relevance of acetylcarnitine in insulin resistance and T2DM in humans, it is key to measure acetylcarnitine levels in tissue.

Acetylcarnitine has previously been visualized noninvasively using short–echo time (TE) ¹H-MRS by subtracting pre- and postexercise spectra. Because the peak was not visible in the rest spectrum, only an exercise-induced difference signal could be quantified at 2.13 ppm (9). More recently, the observation of an alternative resonance peak of acetylcarnitine at 3.17 ppm has been reported at 7T (10). However, due to the overlap with other trimethyl ammonium (TMA) peaks, this peak also remained only quantifiable after exercise, when it was increased considerably. Furthermore, due to the decreased spectral resolution, this peak was not resolved and therefore remained undetected on clinically available 3T MR systems. We hypothesized though that acquisition strategies using a long TE would yield valuable information about the concentration of acetylcarnitine in vivo. The acetylcarnitine peak at 2.13 ppm is normally covered by broad lipid resonances at short TE. Acquisition strategies using a long TE led to the relative suppression of these broad lipid resonances or short transversal relaxation time (T₂) metabolite signals in general (11, 12). The use of a long TE would enhance the visibility of the acetylcarnitine peak irrespective of the magnetic field strength without being limited to an exercise-induced increase.

We here developed a long-TE acquisition strategy to measure acetylcarnitine levels at rest in a single scan on a clinical 3T MR scanner. Detecting acetylcarnitine levels noninvasively gives the opportunity to study acetylcarnitine levels in relation to pathological conditions such as mitochondrial dysfunction, insulin resistance, and diabetes. To evaluate the physiological relevance of the measurement of skeletal muscle acetylcarnitine levels, we applied the long-TE ¹H-MR methodology in skeletal muscle of endurance-trained athletes, lean sedentary subjects, obese sedentary subjects, and T2DM patients. Our results indicate that higher acetylcarnitine concentration is associated with increased insulin sensitivity and better in vivo mitochondrial function, as measured with phosphorus-MRS (³¹P-MRS), and that no athletes’ paradox is present for acetylcarnitine levels.

Subject characteristics.

Endurance-trained athletes and lean sedentary subjects as well as obese sedentary subjects and T2DM patients were matched for age and BMI. Age and BMI were different between endurance-trained athletes/lean sedentary subjects and obese sedentary subjects/T2DM patients (P < 0.01). Fat percentage also showed a similar distribution, with no significant differences between endurance-trained athletes versus lean sedentary subjects and obese sedentary subjects versus T2DM patients, but with a significantly higher fat percentage in the obese sedentary subject and T2DM patient groups when compared with the lean sedentary subject and endurance-trained athlete groups (P < 0.01). Maximal oxygen uptake (VO₂ max ) was highest in endurance-trained athletes, and this was significantly different from that of the lean sedentary subject, obese sedentary subject, and T2DM patient groups (P < 0.01 for all comparisons). VO₂ max was also significantly higher in the lean sedentary subject group when compared with either the obese sedentary subject or the T2DM patient group (P < 0.01), while there was no significant difference between the latter two (P = 0.87). All subject characteristics are summarized in Table 1.

Table 1

Subject characteristics

Long-TE spectrum.

Enhanced visibility of the acetylcarnitine peak in a long-TE spectrum of the vastus lateralis muscle, as compared with a conventional short-TE spectrum, is apparent from Figure 2. Total scan time per spectrum was 2 minutes.

Figure 2

Comparison of short (TE = 37 ms) and long-TE (TE = 350 ms) spectra from the vastus lateralis muscle. Because of the relatively long T₂ of acetylcarnitine (at 2.13 ppm) when compared with overlapping broad lipid resonances, the visibility of the peak is greatly enhanced in the spectrum with the long TE. Because of the relatively short T₂ of water, no water suppression was applied for the long-TE spectrum. Furthermore, because of the long TE, the t-Cr peak (at 3.03 ppm) shows up as a single and sharp peak.

With long TE, the acetylcarnitine peak at 2.13 ppm could be assigned by simple visual inspection. In contrast, with short TE, the peak was hardly visible due to broad lipid resonances in the region 2.0 to 2.5 ppm obscuring the peak. Also very characteristic for the long-TE spectrum was the sharp, single, and symmetric appearance of the total creatine (t-Cr) peak at 3.03 ppm. This peak was relatively broad and asymmetric in the short-TE spectrum. The residual water signal in the long-TE spectrum was relatively low without the application of sophisticated water suppression techniques.

The T₂ of acetylcarnitine was found to be 265 ± 45 ms (270 ± 29 ms for lean subjects versus 262 ± 37 ms for obese subjects, P = 0.88), which is approximately 10 times longer than the T₂ of water at 3T (see also Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI74830DS1). The T₂ of t-Cr was found to be 162 ± 4 ms (157 ± 3 ms for lean subjects versus 166 ± 6 ms for obese subjects, P = 0.24), which is in correspondence with the long decay component as reported in a previous study (13).

Results from the phantom experiments are depicted in Figure 3. The MR-based quantification, with the long-TE protocol, correlates perfectly with the known acetylcarnitine concentration in the phantom experiments (Pearson’s r = 0.99, R² = 0.99, P < 0.01).

Figure 3

Phantom experiments. Five phantoms were prepared with 0, 0.5, 1.0, 2.0, and 5.0 mM acetylcarnitine. The acetylcarnitine peak at 2.13 ppm for the different concentrations is shown in A. The MR-based concentrations of acetylcarnitine are plotted versus the known acetylcarnitine concentrations in B. The measured concentrations are in concordance with the known phantom concentrations, as indicated by the very strong correlation (Pearson’s r = 0.99, R² = 0.99, P < 0.01).

Internal reference.

As the unsuppressed water signal is used extensively as an internal reference in ¹H-MRS, the use of t-Cr as an internal reference was compared with water as an internal reference. The intensity of the water peak was determined from an additionally acquired short-TE spectrum. The ratio of acetylcarnitine over water and the ratio of acetylcarnitine over t-Cr were calculated and plotted in Figure 4, and a very strong correlation was found between the 2 ratios (Pearson’s r = 0.96, R² = 0.91, P < 0.01, n = 21).

Figure 4

Comparison between the use of t-Cr and water as internal reference. Data are expressed as the T₂ uncorrected ratios of acetylcarnitine over the reference metabolite. The fitted line is forced to go through the origin (dotted line). The correlation between both ratios is strong and highly significant.

Insulin sensitivity.

Insulin sensitivity was significantly different among groups (glucose infusion rate [GIR]: 76.8 ± 5.1 μmol/min/kg in endurance-trained athletes, 61.9 ± 5.4 μmol/min/kg in lean sedentary subjects, 30.3 ± 4.6 μmol/min/kg in obese sedentary subjects, and 23.6 ± 2.7 μmol/min/kg in T2DM patients; P < 0.01, Figure 5A). Post-hoc analysis revealed that GIRs of endurance-trained and lean sedentary subjects were significantly higher when compared with the values obtained in obese sedentary subjects and T2DM patients (P < 0.01).

Figure 5

Peripheral insulin sensitivity and PCr recovery rates per group. (A) Insulin sensitivity is given as GIR (μmol/kg/min). There are significant differences between endurance-trained athletes/lean sedentary subjects and the obese sedentary subjects/T2DM patients. (B) Mitochondrial function is depicted as PCr recovery rate τ (s). The rate was significantly higher in T2DM patients when compared with endurance-trained athletes, lean sedentary subjects, and obese sedentary subjects. T, endurance-trained athletes; S, lean sedentary subjects; O, obese sedentary subjects; T2DM, T2DM patients. *P < 0.05, endurance-trained athletes/lean sedentary subjects versus obese sedentary subjects/T2DM patients; ^#P < 0.05, compared with endurance-trained athletes; ^§P < 0.05, compared with T2DM patients. Data are expressed as mean ± SEM.

In vivo mitochondrial function.

We determined in vivo mitochondrial function by determining the phosphocreatine (PCr) recovery kinetics after exercise. The PCr recovery rate constant was significantly different among groups, with fastest recovery in endurance-trained athletes and slowest recovery in T2DM patients (endurance-trained athletes: 20.7 ± 3.4 s; lean sedentary subjects: 27.9 ± 0.6 s; obese sedentary subjects: 29.9 ± 1.7 s; T2DM patients: 36.5 ± 2.4 s; P < 0.01, Figure 5B). Mitochondrial function was significantly lower in T2DM patients compared with endurance-trained, lean sedentary, and obese sedentary subjects (P < 0.01, P < 0.01, and P = 0.04 respectively), as indicated by the post-hoc analysis.

Acetylcarnitine concentrations (¹H-MRS).

Two representative long-TE spectra, from an endurance-trained athlete and a T2DM patient, respectively, are depicted in Figure 6. Spectra are scaled to the t-Cr peak to ensure that a difference in the acetylcarnitine peak intensity is apparent by visual inspection.

Figure 6

Typical long-TE spectra collected from an endurance-trained athlete and a T2DM patient. Spectra are scaled to the t-Cr peak. Most apparent is the difference between the acetylcarnitine peak intensities in both spectra. EMCL signals are higher in the T2DM patients, as was expected.

Overall, acetylcarnitine concentration was significantly different between groups, with highest values in endurance-trained athletes and lowest values in T2DM patients (endurance-trained athletes: 1.58 ± 0.30 mmol/kg wet weight (kgww); lean sedentary subjects: 1.28 ± 0.22 mmol/kgww; obese sedentary subjects: 0.70 ± 0.22 mmol/kgww; T2DM patients: 0.42 ± 0.19 mmol/kgww; P < 0.01, Figure 7A). Post-hoc analysis revealed that these differences in acetylcarnitine concentrations reached statistical significance between T2DM patients and endurance-trained athletes (P = 0.01).

Figure 7

Skeletal muscle acetylcarnitine concentrations per group and the association between mean acetylcarnitine levels and mean insulin sensitivity. (A) The acetylcarnitine concentration is given in mmol/kgww. Post-hoc analysis showed a significant lower acetylcarnitine concentration in T2DM patients when compared with endurance-trained athletes. (B) Correlation between group means of acetylcarnitine concentration (given in mmol/kgww) and GIR (μmol/kg/min). *P < 0.05, compared with endurance-trained athletes. All results are expressed as mean ± SEM.

Significant correlations of the individual acetylcarnitine concentrations with BMI (Pearson’s r = –0.56, R² = 0.31, P < 0.01), age (Pearson’s r = –0.55, R² = 0.30, P < 0.01), VO₂ max (Pearson’s r = 0.57, R² = 0.32, P < 0.01), and fat percentage (Pearson’s r = –0.45, R² = 0.20, P < 0.01) were found. Acetylcarnitine and PCr recovery rate correlated near significance (Pearson’s r = –0.30, R² = 0.09, P = 0.08).

Carnitine metabolism in human muscle material.

To further examine underlying molecular mechanisms for the group differences in acetylcarnitine levels, we took muscle biopsies after an overnight fast from 11 endurance-trained athletes, 6 lean subjects, 5 obese subjects, and 9 T2DM patients. We first determined liquid chromatography–mass spectrometry–based (LC-MS–based) levels of acetylcarnitine (endurance-trained athletes: 593.4 ± 106.3 pmol/kg tissue; lean sedentary subjects: 585.1 ± 111.4 3 pmol/kg tissue; obese sedentary subjects: 336.0 ± 76.7 pmol/ kg tissue; T2DM patients: 341.0 ± 73.2 pmol/kg tissue; P = 0.129) and total carnitine (endurance-trained athletes: 3758.2 ± 487.8 pmol/kg tissue; lean sedentary subjects: 3311.0 ± 727.8 pmol/kg tissue; obese sedentary subjects: 2584.0 ± 496.0 pmol/kg tissue; T2DM patients: 2587.0 ± 555.3 pmol/kg tissue; P = 0.355), which were not significantly different among groups, although in both cases, a pattern of increased concentrations in endurance-trained athletes was found. Acetylcarnitine as measured with ¹H-MRS did not significantly correlate with LC-MS–based acetylcarnitine concentrations (Pearson’s r = 0.28, R² = 0.08, and P = 0.131) nor with the total carnitine concentrations as measured in the biopsies (Pearson’s r = 0.101, R² = 0.01, and P = 0.587).

However, since it is known that acetylcarnitine levels can fluctuate over the day, we also determined the activity of CRAT, which is the most important enzyme responsible for acetylcarnitine formation. Interestingly, CRAT activity was significantly higher in the endurance-trained athletes when compared with the other groups (endurance-trained athletes: 91.9 ± 7.5 nmol/mg protein/min; lean sedentary subjects: 65.4 ± 3.8 nmol/mg protein/min; obese sedentary subjects: 53.4 ± 3.5 nmol/mg protein/min; T2DM patients: 52.6 ± 4.7 nmol/mg protein/min; P < 0.01). Moreover, MR-derived acetylcarnitine concentration correlated with CRAT activity (Pearson’s r = 0.37, R² = 0.134, and P = 0.04).

Association between acetylcarnitine and insulin sensitivity.

A significant correlation between MR-based acetylcarnitine concentration and insulin sensitivity was found (Pearson’s r = 0.59, R² = 0.35, P < 0.01) based on the individual data. When we performed a per-group correlation analysis (Figure 7B), the correlation between the acetylcarnitine concentration group means and the insulin sensitivity group means was very strong and highly significant (Pearson’s r = 0.99, R² = 0.99, P < 0.01). Furthermore, CRAT activity was positively correlated with insulin sensitivity (Pearson’s r = 0.74, R² = 0.55, P < 0.01 for CRAT).

A stepwise linear regression analysis with BMI, age, fat percentage, and VO₂ max and insulin sensitivity as possible determinants of acetylcarnitine was performed on the individual data and revealed that all variables, other than insulin sensitivity, were excluded from the model, further illustrating the association between acetylcarnitine levels and insulin sensitivity (included variable: insulin sensitivity, P < 0.01; excluded variables: BMI, P = 0.416; age, P = 0.157; fat percentage, P = 0.606; PCr recovery rate, P = 0.803; VO₂ max, P = 0.756).

The potential of in vivo ¹H-MRS in skeletal muscle has not been fully exploited. Due to the small chemical shift range in which the metabolites resonate, many of the metabolite resonances are overlapping, thereby obscuring their presence. Here, we show that the use of a long TE can help to enhance the visibility of metabolites with a relative long T₂, such as acetylcarnitine. In conventional short-TE spectroscopy, the acetylcarnitine peak at 2.13 ppm is covered by broad lipid resonances, but it is apparent from our spectra that acetylcarnitine appears as a single, sharp, and symmetrical peak when a long TE is used at 3T. The long-TE approach that we propose here creates the opportunity to noninvasively measure muscle acetylcarnitine levels on clinically available MR systems, which can boost the research into our understanding of the (patho)physiological relevance of this metabolite. To investigate a potential role in insulin sensitivity, we determined acetylcarnitine concentrations in 4 groups, spanning a wide range of values of insulin sensitivity. We carefully phenotyped these groups in terms of insulin sensitivity and in vivo mitochondrial function and indeed found an increase in insulin sensitivity and in vivo mitochondrial function across the groups from T2DM patients to trained subjects. With our long-TE approach, we were indeed able to robustly measure differences in skeletal muscle acetylcarnitine concentrations among the 4 groups. Insulin sensitivity was strongly associated with the measured skeletal muscle acetylcarnitine concentration, with higher acetylcarnitine concentrations in subjects with higher insulin sensitivity.

Besides the enhanced visibility of the acetylcarnitine peak, the long-TE approach offers additional advantages compared with the conventional short-TE methodology. The methyl t-Cr peak is visible as a single sharp peak in long-TE spectra, making it an excellent internal reference compound. In short-TE spectroscopy, this t-Cr peak at 3.03 ppm is contaminated by extramyocellular lipid (EMCL) signals around 2.95 ppm and is influenced by residual dipolar coupling (14). A longer TE results in a destructive interference of phase differences introduced by residual dipolar coupling in the t-Cr peak (13) and the suppression of the EMCL signal. The use of a long TE also diminishes the need for water suppression. In short-TE measurements, the water signal is approximately 10³ to 10⁵ times as abundant as other metabolite signals and can obscure much smaller resonances, which are present in the millimolar range. Normally, sophisticated water suppression methods are used that unavoidably distort signals near the water resonance. Water intrinsically has a short T₂, leading to a very low signal with longer TE. It should also be mentioned that the use of a long TE leads to spectral simplification and a flat baseline, which make fitting procedures straightforward.

Short TE is normally preferred in ¹H-MRS, as it helps minimize the effect of T₂ relaxation, thereby yielding a higher signal-to-noise ratio (SNR). In our protocol, we used a relatively large voxel size to compensate for the loss in SNR with long TE. Using a larger voxel did not lead to increased overlap of the acetylcarnitine peak with other resonances and inherently led to better tissue averaging. In our experience, this is a robust protocol to obtain spectra with reasonable SNR within a relatively short time.

The choice of the methyl t-Cr peak as the internal reference metabolite in this study, instead of the unsuppressed water signal, was based on 2 considerations. In short-TE spectra, analysis and fitting of the dipolar coupled t-Cr resonance is difficult due to the appearance of satellite peaks due to the residual dipolar coupling effect (13). The phase modulations due to the residual dipolar coupling effect are averaged out when using a long TE, resulting in disappearance of the satellite peaks of the t-Cr triplet. The sharp and symmetric appearance of the t-Cr peak at long TE increases reliable fitting of this metabolite. This enables measurement of both the t-Cr and the acetylcarnitine peaks in a single acquisition. Furthermore, it has been reported that the concentration of the t-Cr pool is rather stable, approximately 30 mmol/kgww in muscular tissue, and shows little interindividual and regional variations, which makes it an excellent reference metabolite (15).

Water has extensively been used as an internal reference in MR spectroscopy of skeletal muscle. In this study, we therefore compared the use of the t-Cr peak to the use of water as internal reference. The water signal was measured in a separately recorded short-TE spectrum. We found a very strong correlation between the ratio of acetylcarnitine/t-Cr and acetylcarnitine/water. This indicates that the t-Cr peak is a reliable internal reference and that differences found between groups cannot be explained by the use of a different internal reference. This makes it possible to measure the acetylcarnitine concentration in a single scan. To calculate the absolute concentrations in skeletal muscle, we used fixed T₂ of both acetylcarnitine and t-Cr in all subjects, as T₂ relaxation measurements were not different between lean and obese subjects. Furthermore, it is suggested that a 2-component modulation pattern exists for the t-Cr peak due to the dipolar coupling, giving rise to short and long decay components (short decay in the range of 0 to 35 ms) (13). When correcting the t-Cr peak intensity based on a single exponential T₂ decay function, this could lead to a systematic overestimation of the acetylcarnitine concentration measured. To avoid this, we used a fixed correction factor for the short decay component, and we report acetylcarnitine concentrations of 1.3 to 1.6 mmol/kgww in lean (sedentary and trained) subjects, which is in accordance with present and earlier results of biopsy studies, reporting comparable concentrations of acetylcarnitine (16–18) when converted to mmol/kgww (ranging from ~0.5 to 1.6 mmol/kgww).

The phantom measurements that we performed in this study confirmed that the long-TE MR measurement can accurately determine acetylcarnitine concentrations, with the use of t-Cr as a reference metabolite. However, when comparing the acetylcarnitine concentrations as derived from MRS and LC-MS, we did not find a significant correlation, although a similar pattern across the groups was visible. However, it is important to note that the muscle biopsies were taken early in the morning, while our MR experiments were performed in the late afternoon to be more closely synchronized with the hyperinsulinemic-euglycemic clamp. This time difference may explain the lack of correlation, since acetylcarnitine concentrations are known to vary significantly during the day (19).

As mentioned above, using our long-TE spectroscopy approach, we found that acetylcarnitine levels are highest in endurance-trained athletes and lowest in insulin-resistant T2DM patients. Furthermore, we showed that insulin sensitivity strongly associates with skeletal muscle acetylcarnitine concentrations. In the first instance, however, our finding of low acetylcarnitine concentrations in T2DM patients may seem to be in contrast with findings of a previous publication in which a high acetylcarnitine plasma concentration was associated with T2DM patients (20). High plasma concentrations of acetylcarnitine may be caused by an increased efflux of acetylcarnitine from muscle tissue; however, acetylcarnitine is also formed by other tissues, such as liver or heart (21, 22). Moreover, high concentrations of acetylcarnitine in plasma could potentially lead to a loss of carnitine from the body when the (acetyl)carnitine is secreted via urine. Without a concomitant increase in the biosynthesis or dietary intake of carnitine, this may lead to a lower carnitine status in skeletal muscle or in the total body. The fact that plasma acetylcarnitine concentrations are not in agreement with skeletal muscle concentrations illustrates that the investigation of plasma concentrations is not sufficient and emphasizes the need for tools to quantify this metabolite directly in tissue.

The transformation of excessive acetyl-CoA into acetylcarnitine has recently been suggested to be important to maintain metabolic flexibility and insulin sensitivity under conditions of excessive lipid supply to the mitochondria (4, 5). It is well known that in T2DM patients, excessive accumulation of fat in skeletal muscle is associated with the development of insulin resistance (1, 2). Low acetylcarnitine levels are consistent with the hypothesis that excessive fat availability in muscle of T2DM patients may lead to the development of insulin resistance and metabolic inflexibility due to a decreased conversion of excessive acetyl-CoA into acetylcarnitine, in turn leading to an inhibition of the PDH complex. It has indeed been reported that patients with T2DM are characterized by lower PDH activity (23). Conversely, dietary l-carnitine supplementation administered to prediabetic subjects increased muscle PDH activity (5). Interestingly, endurance-trained athletes also have high amounts of fat in skeletal muscle, yet they are very insulin sensitive, a phenomenon called the athletes’ paradox (3). Our finding of higher acetylcarnitine concentrations in endurance-trained athletes fits with the concept that the conversion of acetyl-CoA into acetylcarnitine may be important in maintaining insulin sensitivity when muscle fat content is high, further underscored by the significant correlation between acetylcarnitine and insulin sensitivity.

Acetylcarnitine is formed by CRAT, a freely reversible enzyme that resides principally in the mitochondrial matrix and that is undetectable in the cytoplasm. Unlike acetyl-CoA, acetylcarnitine is membrane permeant and readily exported from the mitochondrial lumen to other cellular and extracellular compartments. Thus, episodes of increased acetylcarnitine production (e.g., during a meal) could lead to increased efflux and extramitochondrial accumulation of the metabolite (4, 5). We found that muscle CRAT activity correlated positively with MRS-derived acetylcarnitine concentrations and that CRAT activity was similarly disparate among the 4 groups, with the highest activity measured in exercise-trained subjects and the lowest activity found in those with T2DM. Together, these data suggest that the formation and extramitochondrial storage of acetylcarnitine is influenced by CRAT activity and is different between endurance-trained athletes and T2DM patients, which may in turn be related to insulin sensitivity.

In summary, to date, the determinants of acetylcarnitine concentrations in skeletal muscle have not been investigated in detail and it is yet unclear whether acetylcarnitine levels are reflective of a mitochondrial imbalance and/or are indicative of the development of metabolic disease and decreased mitochondrial flexibility (24). With the successful application of a long-TE strategy, we showed that there is an association among the acetylcarnitine concentration, CRAT activity, and insulin sensitivity. The ¹H-MRS protocol can easily be applied on any 3T MR scanner, and the long-TE strategy provides a unique tool to further unravel the role of acetylcarnitine levels in relation to metabolic health and insulin resistance.

View Supplemental data

The authors would like to thank Maarten Vosselman, Silvie Timmers, Tineke van de Weijer, and Lena Bilet for the recruitment of subjects and for their help in the study. V.B. Schrauwen-Hinderling is supported by a VENI grant (no. 91611136) for innovative research from the Netherlands Organization for Scientific Research (NWO). M.E. Kooi is supported by an Aspasia (grant 015.008.047) from the NWO. R.S. Stevens, T. Koves, and D. Muoio are supported by grants from the United States Public Health Service (R01-DK089312 and 2P01-DK058398 to D. Muoio).

Address correspondence to: Patrick Schrauwen, Department of Human Biology, Maastricht University Medical Center, PO Box 5800, 6202 AZ Maastricht, The Netherlands. Phone: 31.0.43.388.1502; E-mail: P.Schrauwen@maastrichtuniversity.nl.

Note regarding evaluation of this manuscript: Manuscripts authored by scientists associated with Duke University, The University of North Carolina at Chapel Hill, Duke-NUS, and the Sanford-Burnham Medical Research Institute are handled not by members of the editorial board but rather by the science editors, who consult with selected external editors and reviewers.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2014;124(11):4915–4925. doi:10.1172/JCI74830.

1. Krssak M, et al. Intramyocellular lipid concentrations are correlated with insulin sensitivity in humans: a 1H NMR spectroscopy study. Diabetologia. 1999;42(1):113–116.
   View this article via: PubMed CrossRef Google Scholar
2. Boesch C, Slotboom J, Hoppeler H, Kreis R. In vivo determination of intra-myocellular lipids in human muscle by means of localized 1H-MR-spectroscopy. Magn Reson Med. 1997;37(4):484–493.
   View this article via: PubMed CrossRef Google Scholar
3. Goodpaster BH, He J, Watkins S, Kelley DE. Skeletal muscle lipid content and insulin resistance: evidence for a paradox in endurance-trained athletes. J Clin Endocrinol Metab. 2001;86(12):5755–5761.
   View this article via: PubMed CrossRef Google Scholar
4. Noland RC, et al. Carnitine insufficiency caused by aging and overnutrition compromises mitochondrial performance and metabolic control. J Biol Chem. 2009;284(34):22840–22852.
   View this article via: PubMed CrossRef Google Scholar
5. Muoio DM, et al. Muscle-specific deletion of carnitine acetyltransferase compromises glucose tolerance and metabolic flexibility. Cell Metab. 2012;15(5):764–777.
   View this article via: PubMed CrossRef Google Scholar
6. Fritz IB, Yue KT. Long-chain carnitine acyltransferase and the role of acylcarnitine derivatives in the catalytic increase of fatty acid oxidation induced by carnitine. J Lipid Res. 1963;4:279–288.
   View this article via: PubMed Google Scholar
7. Stephens FB, Constantin-Teodosiu D, Greenhaff PL. New insights concerning the role of carnitine in the regulation of fuel metabolism in skeletal muscle. J Physiol. 2007;581(pt 2):431–444.
   View this article via: PubMed Google Scholar
8. Constantin-Teodosiu D, Carlin JI, Cederblad G, Harris RC, Hultman E. Acetyl group accumulation and pyruvate dehydrogenase activity in human muscle during incremental exercise. Acta Physiol Scand. 1991;143(4):367–372.
   View this article via: PubMed CrossRef Google Scholar
9. Kreis R, Jung B, Rotman S, Slotboom J, Boesch C. Non-invasive observation of acetyl-group buffering by 1H-MR spectroscopy in exercising human muscle. NMR Biomed. 1999;12(7):471–476.
   View this article via: PubMed CrossRef Google Scholar
10. Ren J, Lakoski S, Haller RG, Sherry AD, Malloy CR. Dynamic monitoring of carnitine and acetylcarnitine in the trimethylamine signal after exercise in human skeletal muscle by 7T 1H-MRS. Magn Reson Med. 2013;69(1):7–17.
    View this article via: PubMed CrossRef Google Scholar
11. Ren J, Sherry AD, Malloy CR. 1H MRS of intramyocellular lipids in soleus muscle at 7 T: spectral simplification by using long echo times without water suppression. Magn Reson Med. 2010;64(3):662–671.
    View this article via: PubMed CrossRef Google Scholar
12. Skoch A, et al. Intramyocellular lipid quantification from 1H long echo time spectra at 1.5 and 3 T by means of the LCModel technique. J Magn Reson Imaging. 2006;23(5):728–735.
    View this article via: PubMed CrossRef Google Scholar
13. Hanstock CC, Thompson RB, Trump ME, Gheorghiu D, Hochachka PW, Allen PS. Residual dipolar coupling of the Cr/PCr methyl resonance in resting human medial gastrocnemius muscle. Magn Reson Med. 1999;42(3):421–424.
    View this article via: PubMed Google Scholar
14. Kreis R, Boesch C. Spatially localized, one- and two-dimensional NMR spectroscopy and in vivo application to human muscle. J Magn Reson B. 1996;113(2):103–118.
    View this article via: PubMed CrossRef Google Scholar
15. Boesch C, Machann J, Vermathen P, Schick F. Role of proton MR for the study of muscle lipid metabolism. NMR Biomed. 2006;19(7):968–988.
    View this article via: PubMed CrossRef Google Scholar
16. Watt MJ, Heigenhauser GJ, Stellingwerff T, Hargreaves M, Spriet LL. Carbohydrate ingestion reduces skeletal muscle acetylcarnitine availability but has no effect on substrate phosphorylation at the onset of exercise in man. J Physiol. 2002;544(pt 3):949–956.
    View this article via: PubMed Google Scholar
17. Roepstorff C, et al. Malonyl-CoA and carnitine in regulation of fat oxidation in human skeletal muscle during exercise. Am J Physiol Endocrinol Metab. 2005;288(1):E133–E142.
    View this article via: PubMed Google Scholar
18. Wall BT, Stephens FB, Constantin-Teodosiu D, Marimuthu K, Macdonald IA, Greenhaff PL. Chronic oral ingestion of L-carnitine and carbohydrate increases muscle carnitine content and alters muscle fuel metabolism during exercise in humans. J Physiol. 2011;589(pt 4):963–973.
    View this article via: PubMed Google Scholar
19. Kien CL, et al. A lipidomics analysis of the relationship between dietary fatty acid composition and insulin sensitivity in young adults. Diabetes. 2013;62(4):1054–1063.
    View this article via: PubMed CrossRef Google Scholar
20. Adams SH, et al. Plasma acylcarnitine profiles suggest incomplete long-chain fatty acid β-oxidation and altered tricarboxylic acid cycle activity in type 2 diabetic African-American women. J Nutr. 2009;139(6):1073–1081.
    View this article via: PubMed CrossRef Google Scholar
21. Agius L, Wright PD, Alberti KG. Carnitine acyltransferases and acyl-CoA hydrolases in human and rat liver. Clin Sci (Lond). 1987;73(1):3–10.
    View this article via: PubMed Google Scholar
22. Minkler PE, Brass EP, Hiatt WR, Ingalls ST, Hoppel CL. Quantification of carnitine, acetylcarnitine, and total carnitine in tissues by high-performance liquid chromatography: the effect of exercise on carnitine homeostasis in man. Anal Biochem. 1995;231(2):315–322.
    View this article via: PubMed CrossRef Google Scholar
23. Kelley DE, Mokan M, Mandarino LJ. Intracellular defects in glucose metabolism in obese patients with NIDDM. Diabetes. 1992;41(6):698–706.
    View this article via: PubMed Google Scholar
24. Schooneman MG, Vaz FM, Houten SM, Soeters MR. Acylcarnitines: reflecting or inflicting insulin resistance? Diabetes. 2013;62(1):1–8.
    View this article via: PubMed CrossRef Google Scholar
25. Kemp GJ, Radda GK. Quantitative interpretation of bioenergetic data from 31P and 1H magnetic resonance spectroscopic studies of skeletal muscle: an analytical review. Magn Reson Q. 1994;10(1):43–63.
    View this article via: PubMed Google Scholar
26. Kuipers H, Verstappen FT, Keizer HA, Geurten P, van Kranenburg G. Variability of aerobic performance in the laboratory and its physiologic correlates. Int J Sports Med. 1985;6(4):197–201.
    View this article via: PubMed CrossRef Google Scholar
27. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237(3):E214–E223.
    View this article via: PubMed Google Scholar
28. Phielix E, et al. Lower intrinsic ADP-stimulated mitochondrial respiration underlies in vivo mitochondrial dysfunction in muscle of male type 2 diabetic patients. Diabetes. 2008;57(11):2943–2949.
    View this article via: PubMed CrossRef Google Scholar
29. An J, et al. Hepatic expression of malonyl-CoA decarboxylase reverses muscle, liver and whole-animal insulin resistance. Nat Med. 2004;10(3):268–274.
    View this article via: PubMed CrossRef Google Scholar
30. Haqq AM, et al. The Study of the Effects of Diet on Metabolism and Nutrition (STEDMAN) weight loss project: Rationale and design. Contemp Clin Trials. 2005;26(6):616–625.
    View this article via: PubMed CrossRef Google Scholar
31. Schrauwen-Hinderling VB, et al. Impaired in vivo mitochondrial function but similar intramyocellular lipid content in patients with type 2 diabetes mellitus and BMI-matched control subjects. Diabetologia. 2007;50(1):113–120.
    View this article via: PubMed CrossRef Google Scholar
32. Gruetter R. Automatic, localized in vivo adjustment of all first- and second-order shim coils. Magn Reson Med. 1993;29(6):804–811.
    View this article via: PubMed CrossRef Google Scholar
33. Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann N Y Acad Sci. 1987;508:333–348.
    View this article via: PubMed CrossRef Google Scholar
34. Vanhamme L, van den Boogaart A, Van Huffel S. Improved method for accurate and efficient quantification of MRS data with use of prior knowledge. J Magn Reson. 1997;129(1):35–43.
    View this article via: PubMed CrossRef Google Scholar
35. Krssak M, Mlynarik V, Meyerspeer M, Moser E, Roden M. 1H NMR relaxation times of skeletal muscle metabolites at 3 T. MAGMA. 2004;16(4):155–159.
    View this article via: PubMed CrossRef Google Scholar