Free fatty acid flux measured using [1-¹¹C]palmitate positron emission tomography and [U-¹³C]palmitate in humans

Abstract

PET radiopharmaceuticals can noninvasively measure free fatty acid (FFA) tissue uptake. Investigators often use PET scan-derived data to calculate FFA flux. We tested whether the [1-¹¹C]palmitate PET measures of palmitate flux provide results equivalent to a continuous infusion of [U-¹³C]palmitate. Nine volunteers participated in study 1 to evaluate whether a rapidly (10–20 s) given bolus of [1-¹¹C]palmitate affects calculated flux results. Thirty volunteers participated in study 2, which was identical to study 1 except that the [1-¹¹C]palmitate bolus was given over 1 min. Volunteers in both studies also received a continuous intravenous infusion of [U-¹³C]palmitate. Plasma palmitate concentrations and enrichment were measured by liquid chromatography-mass spectrometry. The PET/CT images were analyzed on a workstation running PMOD. Palmitate flux was estimated using PET time-activity curve (TAC) data from regions of interest in the left ventricle (LV) and aorta both with and without hybrid TACs that employed the ¹¹CO₂-corrected data for the first 5 min and the ¹¹CO₂-corrected blood radioactivity for the remainder of the PET scan. Palmitate flux in study 1 measured with PET [1-¹¹C]palmitate and [U-¹³C]palmitate were not correlated, and the PET [1-¹¹C]palmitate flux was significantly less than the [U-¹³C]palmitate measured flux. In study 2, the palmitate flux using PET [1-¹¹C]palmitate hybrid LV models provided closer mean estimates of [U-¹³C]palmitate measured flux. The best PET calculation approaches predicted 64% of the interindividual variance in [U-¹³C]palmitate measured flux. Palmitate kinetics measured using [1-¹¹C]palmitate/PET do not provide the same palmitate kinetic results as the continuous infusion [U-¹³C]palmitate approach.

INTRODUCTION

Elevated circulating free fatty acids (FFA) are implicated in the metabolic complications of insulin resistance, obesity, and type 2 diabetes (3). The release of FFA from adipose tissue is increased in some obesity phenotypes (1), resulting in higher plasma FFA concentrations and disordered lipid metabolism. Understanding what tissues are most affected by FFA requires measurements of tissue uptake, which until recently required arteriovenous balance studies with FFA tracers (7, 9, 19, 25) or tissue biopsies (13, 24). The availability of positron emission tomography (PET) imaging now provides investigators with a noninvasive method to measure tissue uptake of FFA (8, 16, 21). [1-¹¹C]palmitate has been used to assess myocardial FFA metabolism (2, 17). Similar to naturally occurring FFA, [1-¹¹C]palmitate that enters the mitochondria for β-oxidation is oxidized to CO₂ and is then released into the bloodstream.

An important aspect of measuring FFA metabolism using [1-¹¹C]palmitate PET is the simultaneous assessment of systemic palmitate flux, which allows tissue uptake data to be presented in the context of overall FFA availability/flux. Palmitate flux during PET studies is calculated by dividing the administered dose of [1-¹¹C]palmitate by the area under the curve (AUC) of blood [1-¹¹C]palmitate content. To our knowledge, this measure of palmitate flux has not been compared with an independent measurement of FFA flux known to be quantitatively accurate (6, 18, 20).

Therefore, our study aimed to 1) compare the measurement of palmitate flux by PET and stable isotope [U-¹³C]palmitate in humans to understand whether we can forgo the stable isotope infusion method to measure FFA flux during [1-¹¹C]palmitate PET studies; 2) test whether the rapidity of the PET tracer bolus affects palmitate kinetic results; and 3) assess whether regions of interest (ROI) drawn in the left ventricle (LV) vs. aorta of PET images provide a better time-activity curve (TAC) in relation to measuring palmitate kinetics.

MATERIALS AND METHODS

Subjects.

Nine nonobese (BMI < 30 kg/m²) adults participated in study 1 to evaluate whether timing of the [1-¹¹C]palmitate bolus affects calculated flux results. Thirty volunteers ranging from normal weight to obese (BMI 19.7–36.9 kg/m²) participated in study 2, which was identical to the first study with the exception that the [1-¹¹C]palmitate bolus was given over 1 min. All volunteers in both studies also received a continuous intravenous infusion of [U-¹³C]palmitate. The protocol was approved by the Institutional Review Board and conducted at the Center for Translational Science Activities’ (CTSA) Clinical Research Unit and the Nuclear Medicine Department at Mayo Clinic, Rochester, MN. Written informed consent was obtained from all volunteers.

Materials.

[U-¹³C]palmitate was purchased from Cambridge Isotope Laboratories, Andover, MA. [1-¹¹C]palmitate was synthesized by the Nuclear Medicine PET Radiochemistry Laboratory of the Mayo Clinic using previously described techniques (15, 26).

Study protocol.

The study participants came to the Mayo Clinic CTSA Clinical Research Unit, where a screening blood sample was collected and body fat was measured using dual energy X-ray absorptiometry (DEXA; GE Lunar; GE Healthcare, Madison, WI). On a separate day, the volunteers reported to the Division of Nuclear Medicine after a 12-h overnight fast for the combined [U-¹³C]palmitate infusion/[1-¹¹C]palmitate PET/CT study. Two catheters were inserted, one into the antecubital vein for infusion of [U-¹³C]palmitate and [1-¹¹C]palmitate and the other into a hand vein in the opposite arm for blood sampling. The hand with the intravenous catheter used for blood sampling was placed in a heated box to facilitate sampling of arterialized blood (12). The [U-¹³C]palmitate continuous infusion (2 nmol·kg⁻¹·min⁻¹) was started ~30 min before the anticipated PET scan to ensure that isotopic steady state was achieved. Each patient underwent a PET/CT study on a Discovery RX system (GE Healthcare, Waukesha, WI) (14). A scout CT scan was acquired in order to select the scan range for the subsequent CT and PET acquisitions. A low-dose CT scan, used for PET attenuation correction and localization, was acquired next. Upon completion of the CT scan, an intravenous bolus of [1-¹¹C]palmitate was manually administered over 10–20 s (study 1 subjects) or 1 min (study 2 subjects). A dynamic 2D PET scan was acquired over the lower chest/upper abdomen, commencing with the injection of ~853 ± 93 MBq (mean ± SD) (~23 mCi) of [1-¹¹C]palmitate using the following scan sequence: 20 frames at 3 s, 12 frames at 10 s, 4 frames at 20 s, and 4 frames at 300 s. Hand vein blood samples were obtained before and at frequent intervals after the [1-¹¹C]palmitate bolus. These data were used to determine [U-¹³C]palmitate enrichment as well as to measure the proportion of blood ¹¹C radioactivity in CO₂ vs. palmitate. After completion of the study, the intravenous catheters were removed and the volunteers were allowed to leave.

Sample analysis.

Plasma palmitate concentrations and enrichment were measured by ultra-performance liquid chromatography-mass spectrometry (18, 20), as previously described. The PET image arterial TACs were corrected for ¹¹CO₂ as follows. Two 1-ml aliquots of blood samples were placed into tubes containing 3 ml of isopropyl alcohol and 1.0 ml of 0.9 M sodium bicarbonate. One milliliter of 6 N HCl was added to one aliquot and 1 ml of 0.1 NaOH was added to the other. These samples were infused with N₂ for 10 min. Radioactivity in the samples was assayed with a γ-well counter, and the data were corrected for isotopic decay. Total ¹¹C radioactivity was measured in the alkalinized aliquot and non-¹¹CO₂ radioactivity (i.e., palmitate) was measured in the acidified aliquot. The ¹¹CO₂ in each sample was calculated from the difference.

PET/CT image analysis.

All PET images were reconstructed with an iterative 2D OSEM reconstruction algorithm and a Gaussian post-filter of 7 mm. The PET data were corrected for attenuation (using the CT images), scatter, and random coincidences, and all frames were decay-corrected to the start of the acquisition. Cross-calibration of image data and blood samples were performed on the day of each study by imaging a known activity concentration and counting aliquots of this solution in the γ-well counter.

The PET images were transferred to a workstation running PMOD (PMOD Technologies, version 2.8; Zurich, Switzerland). The dynamic frames were summed to create a high-count, low-noise image that was used, along with the coregistered CT images, as a guide for ROI placement. ROIs were drawn in the LV cavity and abdominal aorta for the generation of TACs. We used the average (AVG) and maximum (MAX) values for the LV and aorta statistical comparisons. All aorta TACs were corrected for the limited spatial resolution of the PET system by using the methodology of Croteau et al. (5), where the counts in the TAC were divided by a recovery coefficient whose value depended on the diameter of the aorta as measured on the CT.

TAC corrections.

All data are provided as mean ± SD unless otherwise indicated. Because palmitate flux in all participants met the criteria for steady state, palmitate flux was calculated using the measured infusion rate of [U-¹³C]palmitate divided by the plasma palmitate enrichment, as previously described (20). The contribution of ¹¹CO₂ to whole blood ¹¹C radioactivity over time was determined by regressing the fraction of blood ¹¹C radioactivity that remained in the acidified blood relative to alkalinized blood vs. the time of collection to develop the slope and intercept needed to correct the PET ROIs in LV and aorta for non-[¹¹C]palmitate radioactivity. Because we had blood samples for measurement of total and non-CO₂ ¹¹C radioactivity, we also used these data to create hybrid TACs that employed the ¹¹CO₂-corrected LV and aorta data for the first 5 min and the ¹¹CO₂-corrected blood radioactivity for the remainder of the PET scan. Thus, palmitate flux estimated by PET was calculated using the following parameters for the TAC: LV_AVG, LV_MAX, hybrid LV-blood TAC, aorta_AVG, aorta_MAX, hybrid aorta_AVG-blood TAC, and hybrid aorta_MAX-blood TAC. The palmitate rate of appearance using the [1-¹¹C]palmitate was calculated as previously described (22). Essentially, the [1-¹¹C]palmitate dose (kBq) was divided by the blood TAC. Because the rate of appearance values are in milliliters per minute, the palmitate rate of appearance is converted to systemic flux by dividing the rate of appearance in milliliters per minute by palmitate concentration (µmol/ml).

Statistical analysis.

We used paired t-tests to determine whether the [U-¹³C]palmitate flux results were systematically different from the [1-¹¹C]palmitate flux measurement-derived PET using LV_AVG, LV_MAX, hybrid LV-blood TAC, aorta_AVG, aorta_MAX, hybrid aorta_AVG-blood TAC, and hybrid aorta_MAX-blood TAC. Pearson’s linear regression analysis was used to determine the degree of agreement between the stable isotope approach and the various [1-¹¹C]palmitate PET approaches.

RESULTS

Subject characteristics.

Sex, age, BMI, body composition, and plasma and whole blood palmitate concentration data for the volunteers in both studies are provided in Table 1. Study 2 included a wider range of participants’ BMIs than study 1, with the goal of achieving a wide range of palmitate flux.

Table 1.

Subjects' characteristics

	Study 1	Study 2
n	9	30
Sex (male/female)	5/4	6/24
Age, yr	35 ± 10	36 ± 8
BMI, kg/m2	25.5 ± 1.7	27.9 ± 5.2
Body fat, %	31 ± 7	36 ± 10
Fat-free mass, kg	52.8 ± 12.4	47.8 ± 10.1
Plasma palmitate, µmol/l	113 ± 28	169 ± 43
Whole blood palmitate, µmol/l	73 ± 22	104 ± 27

Values represent means ± SD. BMI, body mass index.

[U-¹³C]palmitate flux vs. PET [1-¹¹C]palmitate flux.

We frequently found discrepancies between the ¹¹C content of the LV and aorta ROIs by the 3- to 5-min time points; beginning at 6–8 min the corrected blood ¹¹C content was much closer to the aorta than the LV (Fig. 1). On review of the PET images it was clear that ¹¹C present in myocardium at end systole was often present in the LV ROIs, causing our LV ROI ¹¹C to be a combination of arterial blood and myocardium. In contrast, there was little discrepancy between the aorta ROI ¹¹C and the ¹¹C content of blood during these time intervals.

Fig. 1.

¹¹C activity concentration curve of measurement from blood samples, left ventricle (LV), and aorta for PET palmitate flux. Beginning at 3–5 min, a discrepancy was shown between LV and aorta regions of interest (ROIs); between 6 and 8 min, corrected blood ¹¹C content was much closer to aorta than the LV. Conc, concentration.

The study 1 (10- to 20-s bolus of [¹¹C]palmitate) palmitate flux values measured using the PET approaches and the [U-¹³C]palmitate approach are provided in Table 2. All of the average palmitate flux values from PET were less than the [U-¹³C]palmitate-derived values; most were highly significantly different. Although the differences between palmitate flux using [U-¹³C]palmitate flux and PET [1-¹¹C]palmitate (aorta_AVG and hybrid aorta_AVG-blood TAC) were not statistically significant (both ~P = 0.097), even these approaches were ~22% less than [U-¹³C]palmitate flux. None of the PET [1-¹¹C]palmitate flux approaches provided values that correlated significantly with [U-¹³C]palmitate measured flux. These data indicate that the rapid bolus [1-¹¹C]palmitate method results in a systematic underestimate and poor agreement with [U-¹³C]palmitate measured flux.

Table 2.

[U-¹³C]palmitate flux compared with PET [1-¹¹C]palmitate flux calculations

			PET [1-11C]Palmitate Flux
		[U-13C]Palmitate Flux	LV AVG	LV MAX	Hybrid LV-Blood TAC	Aorta AVG	Aorta MAX	Hybrid Aorta AVG-Blood TAC	Hybrid Aorta MAX-Blood TAC
Study 1	Mean ± SD	154 ± 28	89 ± 25	63 ± 37	99 ± 24	120 ± 59	70 ± 28	120 ± 59	70 ± 28
	t-Test P value		0.0001	0.0001	0.0001	0.0977	<0.0001	0.0965	<0.0001
Study 2	Mean ± SD	146 ± 53	137 ± 52	80 ± 29	142 ± 39	191 ± 74	124 ± 50	177 ± 62	122 ± 54
	t-Test P value		0.3127	<0.0001	0.6129	0.0003	0.0012	0.0014	0.0071
	R2		0.38	0.39	0.32	0.34	0.64	0.43	0.44
	P value		0.0003	0.0002	0.0012	0.0007	<0.0001	<0.0001	<0.0001

Means ± SD are in µmol/min. LV, left ventricle; AVG, average; MAX, maximum; TAC, time-activity curve. t-Test P value, compared with [U-¹³C]palmitate flux; R² and P value, correlated with [U-¹³C]palmitate flux using Pearson’s linear regression analysis.

The study 2 (60 s bolus of [¹¹C]palmitate) palmitate flux values measured using the PET and [U-¹³C]palmitate are also provided in Table 2. Palmitate flux using PET [1-¹¹C]palmitate LV_AVG and hybrid LV-blood TAC were not significantly different than [U-¹³C]palmitate flux (P > 0.05) (Table 2). We found significant correlations between each PET [1-¹¹C]palmitate flux calculation and [U-¹³C]palmitate flux (Table 2). The aorta_MAX approach had the strongest correlation with [U-¹³C]palmitate flux (R² = 0.64, P < 0.0001; Fig. 2A), but these values were significantly less than the [U-¹³C]palmitate flux values. The average palmitate flux values from the hybrid LV-blood TAC approach were closest to the [U-¹³C]palmitate approach, but the predictive value of PET [1-¹¹C]palmitate flux to measure [U-¹³C]palmitate flux was modest (R² = 0.32, P = 0.001; Fig. 2B). These data indicate that the 1-min bolus [1-¹¹C]palmitate method can provide better estimates of average palmitate flux, as well as better agreement with an independent method.

Fig. 2.

Correlation of PET [1-¹¹C]palmitate flux using aorta maximum (MAX; A) and hybrid LV-blood TAC (B) with [U-¹³C]palmitate flux. TAC, time-activity curve.

DISCUSSION

PET radiopharmaceuticals are valuable tools to measure tissue uptake of FFA and palmitate flux in humans (10, 23). We wished to determine whether we could substitute the [1-¹¹C]palmitate PET measurements of palmitate flux for our usual stable isotope approaches. We found that the rapid bolus (10- to 20-s injection) resulted in unacceptable underestimations of palmitate flux relative to [U-¹³C]palmitate. By extending the [1-¹¹C]palmitate injection time to 60 s, we were able to achieve closer mean estimates of PET-measured flux to [U-¹³C]palmitate-measured flux by using the hybrid LV-blood TAC approach. Unfortunately, the PET approaches to measuring palmitate flux did not predict [U-¹³C]palmitate-measured flux with sufficient accuracy to meet our needs. Although our analysis included a recovery coefficient correction, it did not include a correction spill in from adjacent structures. The best of the PET calculation approaches could predict only 64% of the interindividual variance in [U-¹³C]palmitate-measured flux. Despite our trying a number of PET approaches to predict palmitate flux, none of the methods agreed sufficiently well with [U-¹³C]palmitate-measured flux to convince us to abandon including [U-¹³C]palmitate with future PET studies.

We have strong evidence that the constant FFA tracer infusion approach provides quantitatively accurate measurements of FFA kinetics. We (11) have shown that a constant infusion of a radiolabeled (³H or ¹⁴C) FFA tracer accurately predicts the inflow of a known amount of FFA into the systemic circulation under both steady-state (18) and non-steady-state conditions. We (6) also found that stable isotope approaches using [U-¹³C]palmitate provide equivalent flux measurements relative to radioisotope approaches. Thus, the lack of good agreement between [1-¹¹C]palmitate PET and [U-¹³C]palmitate measurements of FFA kinetics points toward problems with the former rather than with the latter technique. That said, it is likely that FFA flux measurements using any bolus tracer technique are less accurate and precise than a continuous tracer infusion approach. Because PET allows virtuously continuous monitoring of blood tracer content from images, whereas other approaches require frequent blood sampling, PET might be better than other tracer bolus approaches.

The gross underestimate of PET-measured palmitate flux that was seen using a 10- to 20-s bolus injection of [1-¹¹C]palmitate was the result of a greater AUC of blood ¹¹C activity (corrected for ¹¹CO₂) than during an equivalent dose with a 60-s bolus time. This may relate to the circulation time of palmitate (~1 min), the incomplete distribution of tracer in the circulation, and the modest fractional extraction of palmitate by tissues. We reviewed a coronal plane (including the heart) time course of the tracer in the circulation. There were at least one, and sometimes two, obvious recirculations of the tracer bolus through the LV after the rapid injection, which are evident in the TAC (Fig. 1) even with the 60-s injection. It is possible that the incomplete mixing of tracer with the circulating palmitate pool with the rapid bolus was responsible for the substantially greater blood ¹¹C radioactivity AUC and thus the underestimation of palmitate flux.

On the basis of these findings, we recommend that FFA PET studies include a continuous stable isotope tracer infusion such as [U-¹³C]palmitate, if one of the goals is to measure FFA flux. This should not be an excessive burden. Some investigators are already performing these types of studies (4), and it is necessary to have intravenous infusions and blood sampling during an FFA PET study in any case. The tracers are relatively inexpensive and the analysis straightforward with modern equipment (20).

In summary, we report that palmitate kinetics measured using [1-¹¹C]palmitate/PET do not accurately reflect palmitate kinetics as measured by an independent [U-¹³C]palmitate approach. If palmitate kinetics is an important aspect of FFA PET studies, we recommend employing a continuous stable isotope tracer to provide more accurate data.

GRANTS

This work was supported by NIH Grants DK-40484, DK-50456, and R00585 from the US Public Health Service, and by the Mayo Foundation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Q.H., N.W.G., and B.J.K. performed experiments; Q.H., Y.C., N.W.G., B.J.K., and M.D.J. analyzed data; Q.H., Y.C., B.J.K., and M.D.J. interpreted results of experiments; Q.H., Y.C., N.W.G., B.J.K., and M.D.J. approved final version of manuscript; Y.C. and M.D.J. prepared figures; Y.C. drafted manuscript; B.J.K. and M.D.J. edited and revised manuscript; M.D.J. conceived and designed research.