Probing Cerebral Metabolism with Hyperpolarized ¹³C Imaging after Opening the Blood-Brain Barrier with Focused Ultrasound

Abstract

Transient disruption of the blood-brain barrier (BBB) with focused ultrasound (FUS) is an emerging clinical method to facilitate targeted drug delivery to the brain. The focal noninvasive disruption of the BBB can be applied to promote local delivery of hyperpolarized substrates. In this study, we investigated effects of FUS on imaging brain metabolism using two hyperpolarized ¹³C-labeled substrates in rodents: [1-¹³C]pyruvate and [1-¹³C]glycerate. The BBB is a rate-limiting factor for pyruvate delivery to the brain and glycerate minimally passes through the BBB. First, cerebral imaging with hyperpolarized [1-¹³C]pyruvate resulted in an increase in total ¹³C signals (p = 0.05) after disrupting the BBB with FUS. Significantly higher levels of both [1-¹³C]lactate (lactate/total ¹³C signals, p = 0.01) and [¹³C]bicarbonate (p = 0.008) were detected in the FUS-applied brain region as compared to the contralateral FUS-unaffected normal-appearing brain region. Application of FUS without opening the BBB in a separate group of rodents resulted in comparable lactate and bicarbonate productions between the FUS-applied and the contralateral brain regions. Second, ¹³C imaging with hyperpolarized [1-¹³C]glycerate after opening the BBB showed increased [1-¹³C]glycerate delivery to the FUS-applied region (p = 0.04) relative to the contralateral side, and [1-¹³C]lactate production was consistently detected from the FUS-applied region. Our findings suggest that FUS accelerates the delivery of hyperpolarized molecules across the BBB and provides enhanced sensitivity to detect metabolic products in the brain; therefore, hyperpolarized ¹³C imaging with FUS may provide new opportunities to study cerebral metabolic pathways as well as various neurological pathologies.

Introduction

The blood-brain barrier (BBB) is a highly selective and relatively impermeable barrier that limits molecules in the blood from entering into the brain interstitial space. Typically, only lipophilic molecules less than 400 Da can traverse an intact BBB¹. These severe limitations can be overcome by reversible disruption of the BBB using focused ultrasound (FUS). By combining with intravenous microbubbles, FUS can exert distinct biologic effects, including BBB opening, ablation, and neuromodulation, depending on the frequency and duty cycle². The ability to open the BBB has been used for a wide range of neurological therapies in preclinical and clinical studies including targeted antibody transport to reduce plaque load in Alzheimer’s disease and introduction of chemotherapy agents for poorly vascularized cerebral tumors^3,4.

Unlike conventional image-guided treatment modalities, therapeutic ultrasound at non-ablative doses has the benefit of treating disease without causing overt tissue destruction. Recent publications have demonstrated the safety and reproducibility of FUS to disrupt the BBB in human patients with glioblastoma, multiple sclerosis, and Alzheimer’s disease^5–7. Currently, there is an NIH-sponsored clinical trial underway to evaluate the safety of opening the BBB in patients with Alzheimer’s disease using a commercially available 220-kHz FUS system (NCT02986932)⁸. After intravenous injection of microbubbles, targeted FUS results in microbubble oscillation, which induces a downregulation of tight junctions associated proteins, and establishes a temporary opening in the BBB^9,10. The BBB disruption can permit the passage of molecules as large as 200 kDa and can last for several hours⁹. While there is evidence that targeted FUS can induce a focal sterile cerebral inflammatory response, there is no evidence that it causes irreversible neuronal injury or cell death,^11–13 indicating that FUS is able to safely disrupt the BBB.

Carbon-13 (¹³C) magnetic resonance spectroscopy (MRS) is a well-established tool for studying in vivo energy metabolism. The low MR sensitivity and natural abundance of ¹³C nuclei can be overcome by administering solutions that contain hyperpolarized (HP) ¹³C-enriched substrates. This dissolution dynamic nuclear polarization (dDNP) technique¹⁴ dramatically increases MR signals of ¹³C-labeled substrates and metabolic products, which in turn allows for real-time spatial mapping of cellular metabolism in vivo. However, this methodology requires rapid cellular transport of the HP substrates since the dDNP-MRS technique maintains HP signals for only a short time after dissolution (1–3 min)¹⁵. For studies of cerebral metabolism, HP substrates are further limited to low molecular weight, non-polar molecules that can easily pass the BBB unless under a defective condition (e.g., brain tumor)¹⁶. Pyruvate is a small molecule that can rapidly cross the BBB and has been employed in preliminary studies of cerebral metabolism in healthy states and in diseases such as brain tumors, neuroinflammation, and traumatic brain injury^17–21. During the recent years, the efficacy of utilizing HP pyruvate for imaging human brain metabolism has been demonstrated in healthy subjects^22,23 and patients with brain tumors^24–26 or traumatic brain injury²⁷. Nonetheless, the transport across the BBB is a rate-limiting step for HP pyruvate^17,28.

Opening the BBB offers opportunities not only to facilitate the transport of pyruvate but also to study a wide range of metabolic pathways in the brain with a various selection of ¹³C-probes. For instance, ¹³C-glutamate and ¹³C-glycerate are hyperpolarized probes that maintain excellent polarization levels, adequate longitudinal relaxation time (T₁), and an importance in cerebral metabolism. However, these agents have proven unsuccessful for investigations of neurotransmission and cerebral glycolysis, respectively, due to their inability to cross the BBB^29,30.

Previous strategies to facilitate the BBB transport of HP agents include administration of hyperosmolar solutions through the internal carotid artery and chemically-induced BBB disruption. Administration of intra-arterial hyperosmolar solutions results in simultaneous vasodilation and shrinkage of endothelial cells, which disrupts the BBB by widening the tight junctions³¹. However, hyperosmolar solution-mediated opening of the BBB requires internal carotid artery administration, which induces variable effects on differing regions of the brain. In addition, this response is relatively short-lived (~10 min) and only widens the inter-endothelial tight junctions to a radius of 200 Å. These limitations preclude targeting of specific regions of the brain³¹. A study by Mazuel et al. showed the delivery of HP ¹³C-glutamate to the brain as well as production of ¹³C-glutamine after BBB disruption with mannitol²⁹. Recently, FUS was applied to open the BBB in rodents, demonstrating increased HP [1-¹³C]pyruvate and [1-¹³C]lactate delivery in the FUS-applied regions^32,33. However, precise metabolic analyses remained challenging in these studies as both pyruvate and lactate were primarily found in the vasculature and extracellular space, so FUS only induced small increase in the levels of HP pyruvate and lactate. As an alternative strategy, utilizing a lipophilic analog of a targeted substrate for HP was suggested to improve BBB permeability. As such, a previous study demonstrated that [1-¹³C]ethyl-pyruvate is transported across the BBB faster than [1-¹³C]pyruvate³⁴. However, this approach is limited to select substrates and often requires substantial investigation in chemical synthesis and optimization.

The goals of this study are two-fold: (1) investigate the impact of FUS on pyruvate metabolism by comparing [1-¹³C]lactate and [¹³C]bicarbonate production from HP [1-¹³C]pyruvate in FUS-applied brain region with the products in FUS-unaffected contralateral brain region in BBB-opening and BBB-intact groups, and (2) demonstrate metabolic imaging of a hydrophilic HP substrate after temporary BBB disruption with FUS.

Results and Discussion

Effects of BBB-Opening FUS on Metabolic Imaging with HP [1-¹³C]Pyruvate

The Fischer rats underwent FUS and then were directly transported to the MR suite for in vivo imaging, followed by immediate ¹H MRI for localization and HP ¹³C MR spectroscopic imaging (MRSI). The overall FUS-HP protocol is illustrated in Figure 1. Figure 2A shows prescription of a ¹³C imaging slice relative to the FUS-applied region of a representative BBB-opening rat. T₂-weighted image did not show noticeable contrast between the FUS-applied region and the contralateral region, indicating that FUS did not induce edema formation (Figure 2B). Within 10 min from the injection of HP solution, contrast-enhanced (CE) ¹H MRI was acquired. The BBB disruption of the BBB-opened group was confirmed by increased contrast in the CE T₁-weighted ¹H MRI (Figure 2C). In particular, the region with increased [¹³C]bicarbonate production was aligned with the hyperintense region in the ¹H MRI (Figure 2D). The BBB-intact group did not show any contrast change in the CE ¹H MRI.

Figure 1. FUS-HP protocol.

After intravenous infusion of nanobubbles, FUS was used to target the striatum for BBB disruption. After FUS, rats were immediately positioned in a ¹H MRI scanner and imaged with a bolus injection of HP [1-¹³C]pyruvate or [1-¹³C]glycerate (~30–45 mins post BBB opening). After ¹³C images were obtained, CE T₁-weighted ¹H images were obtained to confirm BBB opening. FUS: focused ultrasound; HP: hyperpolarized; BBB: blood-brain barrier; CE: contrast-enhanced.

Figure 2. BBB opening after FUS.

(A) ¹³C slice was prescribed to include FUS-applied region. BBB opening by localized FUS was not visible in (B) T₂-weighted ¹H MRI but (C) was confirmed by hyperintense signals in CE T₁-weighted ¹H MRI. (D) The brain region with increased ¹³C-bicarbonate production was aligned with the enhancing region in the ¹H MRI. FUS: focused ultrasound; BBB: blood-brain barrier; CE: contrast-enhanced.

Opening BBB facilitated cerebral transport of HP pyruvate and lactate, which can be further metabolized to carbon dioxide and bicarbonate (Figure 3A). Figure 3B shows metabolite maps of HP [1-¹³C]pyruvate, HP [1-¹³C]lactate and HP [¹³C]bicarbonate from a representative rat from the BBB-opening FUS group. The metabolite maps are applied with a brain mask and overlaid over the matching ¹H slice. The unmasked ¹³C metabolite maps are available in the supporting information (Figure S1). Normalized lactate and bicarbonate maps by total HP signal were also higher in the BBB-opened region than the FUS-unaffected contralateral region (Figure 3C). As summarized in Figure 3D, the BBB-opened group with nanobubbles showed 7.91 ± 3.95 % larger total ¹³C (tC) signal in the FUS-applied region as compared to the contralateral region (p = 0.05). When normalized by total ¹³C signals, [1-¹³C]lactate (lactate/tC = 0.204 ± 0.121, n = 4) and [¹³C]bicarbonate (bicarbonate/tC = 0.039 ± 0.014, n = 3) production in the FUS-applied region were significantly increased as compared to the metabolic signals (lactate/tC = 0.175 ± 0.109, bicarbonate/tC = 0.028 ± 0.014) in the FUS-unaffected region (p = 0.01 for lactate/tC, p = 0.008 for bicarbonate/tC). The lactate/tC and bicarbonate/tC signals were 15.2 ± 4.5 % and 44.4 ± 26.1 % higher in the FUS-applied region than in the contralateral region (Table S1). In contrast, the BBB-intact FUS group without nanobubbles did not display significant differences in signals between the FUS and FUS-unaffected regions. These regions had comparable lactate (lactate/tC = 0.170 ± 0.052 for FUS region, 0.166 ± 0.058 for FUS-unaffected region, p = 0.27) and bicarbonate (bicarbonate/tC = 0.042 ± 0.015 for FUS region, 0.040 ± 0.015 for FUS-unaffected region, p = 0.26) production as well as the total ¹³C signal (p = 0.38).

Figure 3. Effects of FUS on HP [1-¹³C]pyruvate metabolism.

(A) Metabolic pathway of HP [1-¹³C]pyruvate. (B) Brain-masked metabolite maps of HP [1-¹³C]pyruvate, [1-¹³C]lactate, and [¹³C]bicarbonate. The ¹³C maps are overlaid on the corresponding ¹H image. Both lactate and bicarbonate production were larger in the FUS-applied brain region than the contralateral brain region. The numbers for the color bar normalized by the largest [1-¹³C]pyruvate signal. (C) Normalized lactate and bicarbonate maps by the total ¹³C signal (tC) showed increased lactate and bicarbonate signals in the FUS-applied brain region. (D) The contrast between the FUS-applied region and the contralateral FUS-unaffected region was significant in the normalized lactate and bicarbonate maps of the BBB-opening FUS group (* indicates p < 0.05) but the difference was insignificant in the BBB-intact FUS group. FUS: focused ultrasound; HP: hyperpolarized; tC: total ¹³C; BBB: blood-brain barrier; LDH: lactate dehydrogenase; PDH: pyruvate dehydrogenase; TCA: tricarboxylic acid.

The enhanced [1-¹³C]pyruvate and [1-¹³C]lactate signals are consistent with the previous observations³². Furthermore, we could detect increased [¹³C]bicarbonate production after BBB opening. Although [1-¹³C]lactate can be converted from HP [1-¹³C]pyruvate in other organs or in the vasculature and then delivered to the brain, [¹³C]bicarbonate is produced within the targeted brain region, and thus is a direct indicator of local mitochondrial metabolism. It should be noted that the rats were anesthetized with isoflurane during the assessment of brain metabolism, which dampened [¹³C]bicarbonate production^35,36. The increased [1-¹³C]lactate/tC and [¹³C]bicarbonate/tC in the BBB-opened region indicate that the lactate and bicarbonate result from an improved delivery via diffusion of [1-¹³C]pyruvate through the disrupted BBB into the brain parenchyma. This data suggests that the transport of HP [1-¹³C]pyruvate through the intact BBB, mediated by pyruvate transporters, may be rate-limiting and can be enhanced by diffusion following disruption of the BBB. This interpretation is supported by our immunofluorescence data, showing no evidence of neuronal loss or astrogliosis (Figure 4). Immunohistochemistry of neuron-specific protein, NeuN, and astrocyte marker, glial fibrillary acidic protein (GFAP), showed comparable staining between FUS-applied group with nanobubbles and control group that did not receive FUS. Since neurons and astrocytes are the most abundant cell types in the forebrain and are the most metabolically active, the presence of viable neurons and astrocytes supports that the impact of FUS induced disruption of the BBB on oxidative metabolism is transient. Previous studies have also reported that the permeability changes after BBB opening by FUS largely return to normal after 24 hours^37,38 and no histopathological or structural changes by MRI changes following FUS¹².

Figure 4. Immunofluorescent staining of cortical neurons and astrocytes after FUS.

No significant change in NeuN (cortical neurons) and GFAP (astrocytes) staining in (A) a normal rat brain (somatosensory cortex, M2) exposed to FUS and compared to (B) an animal that did not receive FUS.

BBB-Opening FUS Enables Imaging Cerebral Metabolism with HP [1-¹³C]Glycerate

The BBB-opening FUS was applied to a separate group of rats to investigate the feasibility of imaging cerebral metabolism using HP [1-¹³C]glycerate. Metabolic pathways of delivered [1-¹³C]glycerate into the brain after BBB opening is illustrated in Figure 5A. Significantly increased (36.6 ± 13.8 %, n = 3, p = 0.04) glycerate signals (178.9 parts per million [ppm]) were observed in the FUS-applied brain region as compared to the contralateral hemisphere (Figure 5B). A discernable [1-¹³C]lactate peak was consistently detected at 185.0 ppm only from the FUS-applied region (Figure 5C). However, other metabolite signals such as pyruvate or bicarbonate were not initially detected. To confirm these observations, an additional injection of HP [1-¹³C]glycerate was performed, followed by a dynamic slice-selective MRS (Figure 5D–E). In addition to ¹³C-lactate, ¹³C-pyruvate was detected at 172.5 ppm, and ¹³C-bicarbonate peak at ~161 ppm was observed at noise level.

Figure 5. Effects of FUS on HP [1-¹³C]glycerate imaging.

(A) Metabolic pathway of HP [1-¹³C]glycerate after BBB opening. (B) Axial ¹³C image of rat brain, acquired immediately (25 s) after a bolus injection of HP [1-¹³C]glycerate. (C) FUS-applied left brain hemisphere had larger uptake of glycerate than the contralateral unaffected hemisphere. In addition to glycerate, [1-¹³C]lactate peak was detected in the FUS-applied brain region. (D) Time-averaged spectrum (0–120 s) and (E) time-curves of ¹³C-metabolites were acquired using slice-selective dynamic ¹³C MRS, and demonstrate that increased amount of glycerate was delivered to the FUS-applied brain region and was metabolized to lactate and pyruvate. HP: hyperpolarized; FUS: focused ultrasound; ROI: region of interest.

The results show that BBB opening allows hydrophilic molecules, which do not cross the BBB, to successfully reach the brain and subsequently undergo cerebral metabolism. HP [1-¹³C]glycerate is a promising agent to assess glycolysis,³⁰ but this agent was previously unsuccessful for imaging brain metabolism due to its highly polar molecular structure. After opening the BBB with FUS, we could detect the presence of glycerate in the brain as well as downstream metabolic products including lactate, pyruvate, and bicarbonate. Together with the HP pyruvate results, these experiments demonstrate that BBB opening facilitates delivery of exogenous HP molecules to the brain.

Perspectives and Future Studies

While unfocused ultrasound beams are appealing for covering large brain regions, they are difficult to control through the skull and typically result in very inhomogeneous pressure field in the brain. In contrast, focused beams cover much smaller regions, but there is more control over the pressure distribution. In rodents, the use of focused transducers results in the ultrasound energy passing through a small skull area, so the variation in thickness and angle across this area is minimal. In humans, where ultrasound is typically passed through most of the skull using a hemispherical array, there is the ability to do phase and amplitude corrections in order to adjust for beam distortions through different skull regions. Larger regions of brain can be covered by steering the ultrasound beam electronically with the phased array. Thus, the use of a focused beam, and electronic scanning is the most practical means for covering larger brain regions with ultrasound for delivery of metabolites.

In this study, we investigated brain metabolism with an injection of HP pyruvate and HP glycerate after opening the BBB using FUS in rodents. For both HP agents, we observed an increase in HP ¹³C signals of the substrates and the metabolic products in the BBB-opened region. In contrast to previous studies³², we were able to acquire extended metabolic profiles reliably from HP [1-¹³C]pyruvate by using rats (vs. mice) with improved polarization condition (5 T, 0.8 K vs. 3.38 T, 1.25 K). Moreover, our method allowed for assessment of both [¹³C]bicarbonate and [1-¹³C]lactate. It should be noted that HP images in this study were acquired using a conventional pulse sequence, resulting in suitable SNRs and spatial resolution at a single timepoint. Further optimizations of various parameters to improve image acquisition, SNR, and spatial resolution are possible and may allow for dynamic observation of HP kinetics. Our future studies will therefore include a formal analysis of transfer kinetics to delineate the temporal course of transport of HP substances across the opened BBB. In addition, HP probes that target unique metabolic pathways can be explored to evaluate any potential metabolic impacts of FUS.

Maintaining structural and metabolic integrity of the brain is essential as researchers advance the FUS BBB opening technique towards integrated therapies. In the group of animals that received FUS without nanobubbles, the BBB remained closed. In addition, lactate and bicarbonate production was unaltered, indicating that the altered HP imaging results are predominantly due to the BBB opening rather than FUS. In particular, production of [¹³C]bicarbonate from HP [1-¹³C]pyruvate implies that the activity of pyruvate dehydrogenase, an entry enzyme complex that catalyzes pyruvate oxidation in mitochondria, remains intact in the brain with FUS exposure. These findings support the premise that FUS-mediated BBB opening is safe, reversible, and preserves microstructural integrity of the brain. The unique ability of FUS to preserve tissue integrity and function makes the technique particularly attractive for neurological studies. Moreover, FUS-mediated delivery of HP substances across the BBB allows for metabolic imaging with MRSI and can enable studies of brain metabolism. This technical advancement further provides unique opportunities to assess a myriad of metabolic pathways that have previously been difficult to evaluate in vivo. For instance, we demonstrated that after opening the BBB, HP [1-¹³C]glycerate enters the brain and subsequently produces glycolytic intermediates within the brain. An immediate application of HP [1-¹³C]glycerate is in the evaluation of Huntington’s disease, D-glyceric acidemia, and Alzheimer’s disease, which are all associated with altered glycolytic pathways in relation to glycerate^39–41.

Conclusion

In conclusion, we investigated the effects of FUS-mediated opening of the BBB on imaging cerebral metabolism using HP [1-¹³C]pyruvate. In addition, we successfully advanced this technology through the application of FUS with nanobubbles in HP brain imaging using a ¹³C-labeled hydrophilic molecule, [1-¹³C]glycerate. Our study demonstrates that the transient disruption of BBB by FUS provides opportunities to study cerebral metabolism including in vivo analyses of biochemical pathways that were previously inaccessible or lacked suitable methods to obtain precise metabolic measurements. Lastly, our report further emphasizes the need for careful calibration of how the kinetics of HP substrates may be impacted following disruption of the BBB by FUS.

Methods

Animal Preparation

Healthy male Fischer rats (n = 10, 210–270 g, 10–15 weeks old) were purchased from Charles River Laboratories (Wilmington, MA, USA) for the study. The rats were divided into two groups: a BBB-opening FUS group with nanobubbles (n = 7) and a BBB-intact FUS group without nanobubbles (n = 3). Four of the BBB-opening FUS group rats were imaged using HP [1-¹³C]pyruvate and the other three were imaged using HP [1-¹³C]glycerate. The substrate, BBB status, and acquired data of each animal are summarized in Table 1. All animal procedures followed the Guide for Care and Use of Laboratory Animals of US National Research Council and were approved by the University of Texas Southwestern Medical Center Institutional Animal Care and Use Committee. All animal experiments are in accordance with the Animal Research: Reporting In Vivo Experiments (ARRIVE) guidelines.

Table 1.

HP substrates and acquisition conditions.

Rat ID	Weight [g]	HP Substrate	BBB Status	13C Data
1	226	[1-13C]Pyruvate	Open	CSI
2	210	[1-13C]Pyruvate	Open	CSI
3	245	[1-13C]Pyruvate	Open	CSI
4	268	[1-13C]Pyruvate	Open	CSI
5	257	[1-13C]Pyruvate	Intact	CSI
6	234	[1-13C]Pyruvate	Intact	CSI
7	270	[1-13C]Pyruvate	Intact	CSI
8	262	[1-13C]Glycerate	Open	CSI
9	242	[1-13C]Glycerate	Open	CSI
10	213	[1-13C]Glycerate	Open	CSI, MRS

Blood-Brain Barrier (BBB) Opening by Focused Ultrasound (FUS)

Before applying FUS, rats were anesthetized with 1–3 % isoflurane mixed with oxygen and were fixed to a stereotactic frame (Robot Stereotactic, Neurostar, Tubingen, Germany). After registration of the skull with a rat brain atlas, a transducer was affixed to the system, and a series of targets in the left hemisphere were identified using the rat brain atlas. Nanobubbles (synthesized in-house according to the protocol described previously⁴²) were infused into the rat through the tail vein at a rate of 0.3 mL/min, and pulsed ultrasound exposures were delivered at a transmission frequency (f₀) of 0.5 MHz with a pulse repetition frequency of 1 Hz and a pulse length of 10 msec using custom-developed software (Labview, National Instruments, Austin, TX, USA). At each target location, stimulated acoustic emissions were recorded using a 0.75-MHz hydrophone connected to a digital oscilloscope (ATS460, Alazar Technologies, Inc, Pointe-Claire, Quebec, Canada) and the frequency spectrum was analyzed to calculate the area under the curve (AUC) between 0.70 and 0.80 MHz. The software adjusted the output pressure from the transducer after each burst to maintain the AUC at a target threshold selected in the software, which typically corresponded to a pressure level between 0.3 and 0.4 MPa. The algorithm used to control the transducer output has been described in more detail previously⁴³. After each 50-sec exposure, the system translated the transducer to the next target brain location and the process was repeated until the entire target region of brain was exposed to open the BBB in that volume. In this study, a total of 8 target locations in a 2×4 grid manner were selected, covering a rectangular spot approximately 4 mm by 8 mm. The rats in the BBB-intact FUS group underwent the same procedure without nanobubbles to serve as a control group without BBB opening.

Sample Preparation and Dynamic Nuclear Polarization (DNP)

Two substrates were used for hyperpolarization: [1-¹³C]pyruvate and [1-¹³C]glycerate. For pyruvate, a 35-μL sample of 14-M [1-¹³C]-labeled pyruvic acid (MilliporeSigma, Miamisburg OH, USA), mixed with 15-mM trityl radical OX063 (Oxford Instruments America Inc., Concord MA, USA), was prepared for each dissolution and polarized using a SPINlab^™ clinical DNP polarizer (GE Healthcare, Waukesha WI, USA). Since three hours of polarization using the SPINlab typically achieves 30–40 % of liquid-state polarization at the time of dissolution for [1-¹³C]pyruvic acid⁴⁴, we polarized the samples for 3–4 hours to ensure at least 30 % of polarization. The pyruvate samples were dissolved by 130 °C solvent (saline with 0.1 g/L EDTA-Na₂) and mixed with pH-neutralization media (NaOH). Final solution contained 70-mM [1-¹³C]pyruvate with pH of ~7.5. For glycerate, the labeled compound was synthesized as previously described³⁰. The glycerate samples were prepared in 3.0-M glycerol/water solution containing 15-mM OX063. After 3–5 hours of polarization, each sample was dissolved in a solution of hot solvent (saline with 0.1 g/L EDTA-Na₂), leading to a 60-mM solution of HP glycerate with a pH of ~7.5. For both substrates, HP solution was delivered to animals intravenously with an injection rate of 0.25 mL/s (0.875 mmol/kg body weight for pyruvate, 0.75 mmol/kg for glycerate).

MR Protocol

MR imaging data were collected using a 3T clinical MR scanner (GE Healthcare, 750w Discovery) and a ¹³C/¹H dual-tuned birdcage radiofrequency (RF) coil (GE Healthcare, inner diameter = 50 mm). After applying FUS, each animal was positioned inside the MR scanner in prone position to have the brain at the center of the RF coil. First, two-dimensional (2D) T₂-weighted fast spin echo (FSE) images were acquired (echo time [TE] = 11.3 msec, repetition time [TR] = 5 sec, slice thickness = 2 mm, matrix size = 256 × 256, field of view [FOV] = 9.6 cm × 9.6 cm) to acquire anatomical reference and to identify the imaging slice for ¹³C imaging. Then ¹³C data was acquired ~30 mins after FUS exposure using a single time-point 2D free-induction decay chemical shift imaging (FID CSI) sequence (spectral width = 5,000 Hz, spectral points = 256, slice thickness = 7.7 mm, matrix size = 16 × 16, FOV = 48 mm × 48 mm) 25 secs after the start of the injection of HP solution. In one animal (rat ID#: 10), approximately 10 mins after the MRSI acquisition, a slice-selective ¹³C MRS was acquired (slice thickness = 8 mm, flip angle = 10°, TR = 3 sec, spectral width = 5,000 Hz, spectral points = 2,048, #timepoints = 80) with additional injection of HP [1-¹³C]glycerate. Within 10 min after HP scans, the BBB opening was confirmed by contrast-enhanced (CE) T₁-weighted images (TE = 12 ms, TR = 700 ms, slice thickness = 2 mm, matrix size = 256 × 256, FOV = 9.6 cm × 9.6 cm) that was acquired after injecting 0.9 mL solution of a 1:2 mixture of Gadolinium (MAGNEVIST^®, Bayer AG, Germany) and saline through the tail vein. The rats were euthanized immediately after the imaging session.

¹³C Data Reconstruction and Analysis

All ¹³C data sets were processed using MATLAB (Mathworks, Natick MA, USA) from the raw data (P-file in GE format). The k-space raw data of ¹³C CSI were apodized by a 10-Hz Gaussian filter and zero-filled by a factor of 4 both in the spectral and spatial dimensions. After a fast Fourier transform (FFT) in the time domain, an inverse 2-dimensional FFT in the spatial frequency domains was performed. Metabolic maps were calculated by integrating the corresponding metabolite peaks in absorption mode. For analysis, the FUS-applied region of interest (ROI_FUS) was selected over voxels that only include brain tissues in the FUS-applied (left) brain hemisphere, based on the matching proton images. Similarly, the FUS-unaffected ROI (ROI_noFUS) was chosen in the contralateral side of the brain. The sizes of the ROIs were similar to each other in each animal. For comparing HP ¹³C signals between ROI_FUS and ROI_noFUS, individual metabolite peaks were integrated from ¹³C spectra that were averaged over each ROI. For HP [1-¹³C]pyruvate images, metabolite data for each ROI were evaluated as ratios to total carbon (tC), which was calculated as the sum of the detected ¹³C signals. In addition, image contrast between the ROIs was evaluated by taking the ratio of signal difference between the ROIs to the signal from the FUS-unaffected ROI ([S_FUS-S_noFUS]/S_noFUS).

Histology

Histological evaluations of the rat brain following FUS exposure were performed in separate Fischer rats (FUS-BBB opening: n = 2; normal control n = 2) 72 hours post-FUS. The rats were euthanized via trans-cardiac perfusion with PBS followed by 4% (w/v) paraformaldehyde (PFA) in PBS. The harvested brains were fixed in cold PFA (4 °C) for 12 hours, washed thoroughly in PBS and dehydrated sequentially in 15% and 30% sucrose in PBS until they were completely submerged. Fixed and dehydrated brains were embedded O.C.T for cryosectioning, 20 μm (CM3050S, Leica). Sections were permeabilized with 0.25% Triton X-100 and blocked with 5% BSA and 3% normal goat serum with 0.25% Triton X-100 for 2 hours. Primary antibodies against GFAP (1:300 rabbit, G9269, Sigma-Aldrich, Inc. St. Louis MO, USA), NeuN (1:300, mouse, MAB377, MilliporeSigma), were incubated with brain sections for 24 hours at 4°C. After three washes in PBS, sections were incubated with secondary antibodies conjugated with Alexa 488, Alexa 647, (1:500, Life Technologies, Carlsbad CA, USA) for 2 hours at room temperature. For nuclear labeling, sections were washed in PBS and incubated with DAPI (Life Technologies) for 10 mins. Sections were coverslipped with anti-fade mounting medium Fluoro-Gel EMS for confocal microscopy (LSM510).

Statistical Analysis

Values are reported as mean ± standard deviation. Differences of the lactate/tC and bicarbonate/tC from the HP [1-¹³C]pyruvate study and tC from the HP [1-¹³C]glycerate study between FUS-applied and FUS-unaffected contralateral brain regions were assessed for statistical significance using a paired Student’s t-test (two-tailed analysis, α = 0.05).

Supplementary Material

Supporting Information

Figure S1. Metabolite maps acquired after BBB opening by FUS. Unmasked metabolite maps of HP [1-¹³C]pyruvate, [1-¹³C]lactate, [¹³C]bicarbonate, [1-¹³C]alanine, and [1-¹³C]pyruvate-hydrate.

Table S1. Contrast between FUS applied region and contralateral brain region. Percent differences of spatially averaged HP lactate/tC, bicarbonate/tC, and tC between FUS applied region and contralateral brain region.