A Dual Tracer PET-MRI Protocol for the Quantitative Measure of Regional Brain Energy Substrates Uptake in the Rat

Podsumowanie

Small-animal positron emission tomography enables the assessment of the brain's two main energy substrates: glucose and ketones. In the present method, ¹¹C-acetoacetate and ¹⁸F-fluorodeoxyglucose are injected sequentially in each animal, and their uptake is measured quantitatively in specific brain regions determined from the magnetic resonance images.

Streszczenie

We present a method for comparing the uptake of the brain's two key energy substrates: glucose and ketones (acetoacetate [AcAc] in this case) in the rat. The developed method is a small-animal positron emission tomography (PET) protocol, in which ¹¹C-AcAc and ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) are injected sequentially in each animal. This dual tracer PET acquisition is possible because of the short half-life of ¹¹C (20.4 min). The rats also undergo a magnetic resonance imaging (MRI) acquisition seven days before the PET protocol. Prior to image analysis, PET and MRI images are coregistered to allow the measurement of regional cerebral uptake (cortex, hippocampus, striatum, and cerebellum). A quantitative measure of ¹¹C-AcAc and ¹⁸F-FDG brain uptake (cerebral metabolic rate; μmol/100 g/min) is determined by kinetic modeling using the image-derived input function (IDIF) method. Our new dual tracer PET protocol is robust and flexible; the two tracers used can be replaced by different radiotracers to evaluate other processes in the brain. Moreover, our protocol is applicable to the study of brain fuel supply in multiple conditions such as normal aging and neurodegenerative pathologies such as Alzheimer's and Parkinson's diseases.

Wprowadzenie

Context and Rationale

Positron emission tomography (PET) enables the minimally-invasive study of functional processes in the brain. Glucose is the brain's main energy substrate, but in conditions of glucose deficiency, ketones (acetoacetate [AcAc] and β-hydroxybutyrate) are the main alternative energy substrates. Brain energy metabolism has been widely studied by PET using the most common PET tracer, ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG), a glucose analog. Our group recently developed a novel radiotracer -¹¹C-AcAc - to measure brain ketone metabolism¹. Magnetic resonance imaging (MRI) is a much higher resolution technique (0.1 mm × 0.1 mm in-plane resolution) than PET, and is needed to clearly localize anatomical brain regions required for the regional PET analysis of brain energy metabolism.

PET data are commonly expressed as standardized uptake values (SUV)^2-6. SUV are the tissue activity concentration normalized by the fraction of the injected dose/unit weight, as initially proposed over 70 years ago⁷. These units are still widely used because they require simpler PET acquisition and image analysis methodologies. However, an important limitation is that SUV are relative not absolute units, making it difficult to compare results across different studies. This difficulty of comparison may contribute to contradictory findings in the literature on brain glucose uptake in the elderly⁸.  Therefore, the quantitative cerebral metabolic rate (CMR; μmol/100 g/min) has particular advantages^9-11. Generating CMR values requires a dynamic PET acquisition and the plasma radioactivity counts as a function of time, i.e. the plasma time-activity curve (TAC) or input function. The input function can be obtained by multiple blood samplings throughout the PET acquisition^9,12 or by the image-derived input function (IDIF) method, in which a region of interest is drawn on a blood pool (heart's left ventricle or major artery)^11,13-16.

Goal

The aim of our method was to quantitatively compare for the first time the uptake of the brain's two key energy substrates, glucose and ketones, using PET and MRI in rodents. ¹¹C-AcAc and ¹⁸F-FDG were used sequentially in the same animal. The protocol was designed to measure regional uptakes in different relevant brain structures (cortex, hippocampus, striatum, and cerebellum) clearly visible in the MR images. The protocol was also specifically intended to permit quantitative analysis of tracer brain uptake, i.e. CMR of both ¹⁸F-FDG and ¹¹C-AcAc. Although this protocol was developed to study brain energy substrates, the radiotracers we used could be replaced by others, and the same methodology can be used to study different brain functional processes.

Advantages over existing methods

PET and MRI do not require the animal to be sacrificed after the acquisition. Therefore, follow-up studies of treatments are possible. Thus, baseline data followed by an experimental condition can be measured within the same animal, thereby reducing both biological variability and the number of animals required. A key advantage of our dual tracer PET protocol is to compare both tracers uptake in the same animal under the same physiological conditions within the same imaging session, thereby reducing even more biological variability and systematic discrepancies. This dual tracer PET protocol is feasible primarily because of the short physical half-life of ¹¹C (20.4 min) and the fast biological washout of ¹¹C-AcAc, which leave minimal residual ¹¹C radioactivity during the second acquisition with ¹⁸F-FDG. The MRI scan is an important feature of this protocol as it enables the tracer uptake to be studied in specific brain areas. In addition, this method enables an absolute quantitative measure of brain tracer uptake in contrast to relative units obtained by the SUV method. Finally, ¹¹C-AcAc images have a low signal-to-noise ratio because of relatively low brain AcAc uptake under physiological conditions, which makes automatic ¹¹C-AcAc and MR images registration challenging. Hence, because ¹¹C-AcAc and ¹⁸F-FDG acquisitions are sequential (no motion of the animal), the ¹⁸F-FDG to MRI alignment can be applied to ¹¹C-AcAc images.

Key papers where the protocol has been used

We have used the dual tracer PET protocol in a study involving the comparison of the fasted state and the ketogenic diet (KD) in young rats². We showed that both fasting and the KD increase significantly both ¹¹C-AcAc and ¹⁸F-FDG brain uptake. However, these were not quantitative results as we did not use the dynamic PET imaging and tracer kinetic modeling methodology at that time. Thereafter, we undertook a regional and quantitative study of brain metabolism in aged rats, where the effect of aging and a KD were evaluated on brain ¹¹C-AcAc and ¹⁸F-FDG uptake¹¹. We also showed that the percentage of distribution across brain regions was different between ¹¹C-AcAc and ¹⁸F-FDG. Furthermore, not only the CMR of ¹¹C-AcAc but also that of ¹⁸F-FDG was increased in the whole brain as well as in the striatum of aged rats on the KD.

Access restricted. Please log in or start a trial to view this content.

Protokół

All experiments were completed in accordance with the Animal Care and Use Committee at the Université de Sherbrooke and with the Canadian Council on Animal Care. The experimental protocol was approved by the Institutional Animal Research Ethics Review Board (protocol #011-09).

1. Brain Anatomy with MRI

1. Let rats acclimatize in the animal facility for a minimum of 7 days prior to the protocol. Perform brain MRI scans 1-2 weeks prior to the dual tracer PET protocol to allow full recovery from the anesthesia.
2. Anesthetize the rat in an induction chamber. Use 2% isoflurane and 1.5 L/min oxygen throughout the protocol for anesthesia. Pinch the hind leg to ensure the animal is fully anesthetized. CAUTION: inhalation of isoflurane can cause headache, dizziness, or unconsciousness in some cases; it should always be used in the presence of an air exchange system in a suitably ventilated room.
3. Position the rat on the MRI examination table in the head-first prone position with a nose cone for isoflurane. Position the respiration and rectal probes. Monitor the respiration rate during the experiment and maintain body temperature at 37 °C with an automated air warming system.
4. Acquire T₂-weighted MR images using a fast spin-echo pulse sequence with acquisition parameters as detailed previously¹¹. Place the rat on a heated mat during recovery from anesthesia.

2. Dual Tracer PET Acquisitions

1. Use a small-animal PET scanner equipped with avalanche photodiode detectors, an axial field of view of ~7.5 cm and an isotropic spatial resolution of 1.2 mm¹⁷. Due to its short physical half-life, prepare ¹¹C by proton bombardment of natural nitrogen (through the ¹⁴N(ρ,α)¹¹C nuclear reaction) in a cyclotron on-site before each experiment.
2. Fast the rat for 18 hr prior to PET scanning. Fasting allows higher brain uptake of ¹¹C-AcAc and ¹⁸F-FDG, thereby improving signal-to-noise ratio.
3. Anesthetize the rat in an induction chamber. Transfer the rat onto a heating mat with a nose cone for isoflurane. Start ¹¹C-AcAc synthesis at this time. The synthesis takes 18 min from the end of bombardment, as previously described¹.
4. Position the rat on its side. Prepare a PE50 polyethylene catheter filled with heparinized 0.9% sodium chloride solution (saline). Install the catheter in the tail vein for tracer injection. Install a second catheter in the mid-ventral tail artery for blood sampling throughout the acquisitions.
5. Rapidly transfer the rat to the scanner table in the head-first prone position with a nose cone for isoflurane. Position the respiration and rectal probes. Monitor the respiration rate during the experiment and maintain body temperature at 37 °C with an automated warming air system. Move the scanner table forward into the scanner to ensure the appropriate imaging of the brain and the heart simultaneously. As an anatomical landmark, position the edge of the field of view on the rat's eyes (with the help of laser lines).
6. Using a concentrated ¹¹C-AcAc solution (~1 GBq/ml), prepare a syringe of ~50 MBq of radioactivity (a range of 45-55 MBq is acceptable). Adjust the volume to 300 μl with saline. CAUTION: Radioactivity can be harmful for the health. It should always be handled behind a lead shield. The experimenter should be wearing a body and ring dosimeter. A Geiger counter should also be on during the experiment.
7. Install the syringe on an injection pump. Start the bolus injection of ¹¹C-AcAc at a rate of 1 ml/min (injection duration: ~19 sec). On a second pump, immediately after terminating the ¹¹C-AcAc injection, start the injection of 300 μl of saline at a rate of 1 ml/min, which is important for an optimal injection of tracer into the blood circulation.
8. Start a dynamic PET data acquisition 30 sec before starting the bolus injection for a total duration of 20.5 min. The 30 sec data acquisition before the tracer injection provides a measure of the ambient background to be subsequently subtracted from the PET data. Set the regular sampling mode and the energy window at 250-650 keV.
9. Collect two 200 μl blood samples at ~15 and 18 min after starting the ¹¹C-AcAc injection. Inject heparinized saline in the catheter after each sampling to avoid blood clotting. Take note of the time at the beginning and the end of blood sampling and use the mean time. Centrifuge at 6,000 RPM for 5 min and collect plasma. Measure radioactivity counts in plasma using a gamma counter calibrated with the PET scanner.
10. After the ¹¹C-AcAc scan, allow a waiting period of 20 min to ensure that most of radioactivity has decayed. The period can vary from 15-30 min. Do not move the scanner bed and leave the rat under isoflurane throughout this period.
11. During this time, using a concentrated ¹⁸F-FDG solution (~5.5 GBq/μl), prepare a syringe of ~50 MBq. Proceed exactly as mentioned in steps 2.6 and 2.7. Start a dynamic acquisition of a total duration of 40.5 min, including the 30 sec before the injection. Collect two 200 μl blood samples at ~30 and 35 min after starting the ¹⁸F-FDG injection.
12. At the end of the ¹⁸F-FDG acquisition, take one final 200 μl blood sample. Centrifuge and collect plasma. Keep plasma samples at -80 °C for glucose and AcAc analysis.
13. Finally, place the rat on a heating mat during the waking period. Keep the animal for a longitudinal follow-up, or alternatively, euthanize the rat for brain sampling and further biochemical studies.
14. Reconstruct the PET images according to the following time frame sequences: 1 × 30 sec; 12 × 5 sec; 8 × 30 sec; and n × 300 sec, where n = 3 for ¹¹C-AcAc acquisition and n = 7 for ¹⁸F-FDG acquisition.

3. Plasma Glucose and AcAc Analysis

1. Measure plasma glucose and AcAc with a clinical chemistry analyzer. Perform assays within 48 hr of the blood sampling to minimize AcAc decarboxylation into acetone. Measure glucose concentration with DF40 kit as previously described¹⁸ and AcAc with an open channel². This type of analyzer is more accurate than strips (0.1 μM) and has a broad detection window.

4. Quantitative PET Analysis

1. Use PMOD software or an equivalent system for small-animal PET image analysis.
2. Plasma time-activity curve (TAC)
   1. Load ¹⁸F-FDG data. Sum image frames of the first 60 sec following injection (when tracer is mainly in blood).
   2. Using a manual drawing tool, draw a volume of interest (VOI) on the left ventricular cavity blood pool (a large pool centered in the bottom part of images). Draw the VOI  ~1 mm inside of the edge of the blood pool (red voxels) to ensure no inclusion of tissue and avoid tissue radioactivity spill into the blood pool (Figure 1A). Copy the VOI to the entire dynamic image series and generate a curve of radioactivity as a function of time (TAC).
   3. In a spreadsheet, use radioactivity counts of the two plasma samples taken during the acquisition to correct plasma TAC (Figure 1B). Use the following equation:
      
      Some smoothing of the plasma TAC can be performed.
   4. Verify if residual radioactivity from ¹¹C-AcAc is present in the first 30 sec of ¹⁸F-FDG scan, prior to injection. If so, subtract the following factor from all the subsequent time frames of the TAC:
      ,
      where R is the residual radioactivity, t_1/2 is the half-life of ¹¹C (20.4 min) and T is the time frame^2,13. Residual radioactivity from ¹¹C-AcAc after a waiting period of 20 min can be up to 1.5% of maximal ¹⁸F-FDG counts.
3. ¹⁸F-FDG/MRI coregistration
   1. Apply coregistration obtained with ¹⁸F-FDG images to the¹¹C-AcAc images. This is necessary because ¹¹C-AcAc images have a low signal-to-noise ratio and automatic coregistration is difficult.
   2. Load the corresponding MRI data. Load all 28 ¹⁸F-FDG time frames in the same orientation as the MR images (on three axes).
   3. Compute the ¹⁸F-FDG summed image of the 28 time frames. This generates an image with higher counts, so it is easier to work with for coregistration with MRI than individual image frames.
   4. Perform the automatic coregistration of the ¹⁸F-FDG summed image on MR images (Figure 2). Apply the transformation to all individual ¹⁸F-FDG image frames. Save the transformation which will later be applied to ¹¹C-AcAc images.
4. Volumes of interest (VOIs)
   1. Using segmentation software such as PMOD, select MRI data and choose preferred plane for manual drawing. Choose the manual or the semi-automatic drawing tool, depending on brain region size and image contrast. Locate brain structures according to a standard rat brain atlas¹⁹.
   2. Draw the VOIs. As shown in Figure 3, segment the whole brain, cortex, hippocampus, striatum and cerebellum, but other regions could be chosen. Save the VOIs, which will be applied to the coregistered ¹⁸F-FDG and ¹¹C-AcAc images.
   3. Apply VOIs to ¹⁸F-FDG coregistered images. Select the VOI statistics to visualize brain TAC. If needed, correct TAC by subtracting the decay-corrected mean radioactivity in the first 30 sec from all the subsequent time frames.
5. Cerebral metabolic rate calculation
   1. Load brain and the corresponding plasma TACs.
   2. Select Patlak plot kinetic model^20,21. Set the lumped constant to 0.48²² and the corresponding plasma glucose concentration (Figure 4B). Set the maximal relative deviation from the Patlak plot to 5%. Fit the data to the model.
6. ¹¹C-AcAc images
   1. Compute the ¹⁸F-FDG summed image of the 24 first time frames to have the same number of frames as the ¹¹C-AcAc scan. Otherwise, ¹¹C-AcAc/MRI coregistration fails. Perform the automatic coregistration of ¹⁸F-FDG on MRI. Save the transformation process which will be used for ¹¹C-AcAc image coregistration.
   2. Repeat steps 4.2 and 4.3 for ¹¹C-AcAc images. Use ¹⁸F-FDG transformation in step 4.6.1 and apply it to all individual ¹¹C-AcAc image frames. Repeat steps 4.4 and 4.5 and use ¹⁸F-FDG saved VOIs. Set the lumped constant to 1.

Access restricted. Please log in or start a trial to view this content.

Wyniki

As seen in Figure 2,¹¹C-AcAc uptake is low within the brain itself. As mentioned earlier, ketones consumption by the brain is very low on a short-term fasting. ¹¹C-AcAc uptake is higher in the tongue and cheek muscles. Indeed, ketones are rapidly taken up by rat skeletal muscles²³. In contrast, ¹⁸F-FDG uptake is mostly in the brain and the cheek muscles. Figure 2 shows that during the coregistration process, MR images are fixed and PET ima...

Access restricted. Please log in or start a trial to view this content.

Dyskusje

Critical steps

A critical step in this dual tracer PET protocol is to be able to simultaneously scan the heart's left ventricle and the brain at the same time. This requires a PET scanner with a sufficient axial length, i.e. a minimum of 7.5 cm. A few test scans are needed to determine the exact position of the scanner table (x, y, and z values), where the brain and the heart are scanned correctly.

Tracer injection is also a crucial point for a succes...

Access restricted. Please log in or start a trial to view this content.

Materiały

Name	Company	Catalog Number	Comments
MRI scanner	Varian		7 Tesla
Small-animal PET scanner	Gamma Medica		Lab-PET/-Triumph
Heat mat	Sunbeam	PN 143937
Heater system	SA Instruments	761 100 Rev B
Respiratory gating	SA Instruments	SAII's P-resp
Clinical chemistry analyzer	Siemens Healthcare Diagnosis	765000.931	Dimension Xpand Plus
Polyethylene tubing 50	Becton Dickinson	427411
Injection pump	KD Scientific	Model 210
Gamma-counter	GMI	Packard Cobra II
Centrifuge	Thermo Scientific	75002416	Heraeus Pico 21
PMOD software	PMOD Technologies		PMOD 3.2 version
Geiger counter	Fluke Biomedical	ASM-990	Advanced Survey Meter
Reagent
Isoflurane	Abbott Laboratories, Ltd	B506
0.9% NaCl solution	Hospira	4888010
Heparin	Sandoz	1004336
Isopropenyl acetate	Aldrich	11778	99%
Methyllithium	Aldrich	197343	1.6 M
THF	Aldrich	87371
Flex reagent cartridge glucose	Siemens Healthcare Diagnosis	DF40
Trizma base	Sigma	T6066-500G	prepare tris buffer 100 mM pH 7.0
Sodium oxamate	Sigma	O2751	20 mM
NADH	Roche	10128015001	0.15 mM
b-Hydroxybutyrate dehydrogenase	Toyobo	HBD-301	1 U/ml