Isoflurane titration improves detection of hippocampal lactate by ¹H-MRS

Highly elevated cerebral lactate levels are typically associated with ischemic events (Henriksen et al., 1992; Higuchi et al., 1996; Hyacinthe et al., 2020; Woo et al., 2010). However, numerous studies have demonstrated that under non-ischemic conditions, more subtle changes in lactate play multiple key roles in brain physiology, including synaptic plasticity, long-term potentiation, and memory formation (Bingul et al., 2020; Dembitskaya et al., 2022; Suzuki et al., 2011; Yang et al., 2014). In addition, lactate can function as a signaling molecule within the brain by binding to the G-protein coupled receptor-81 (GPR81), also known as hydroxycarboxylic acid receptor 1 (HCAR1), thereby decreasing neuronal excitability (Ahmed et al., 2010; Barros, 2013; Briquet et al., 2022; Cai et al., 2008; Liu et al., 2009). Altered lactate metabolism has been implicated in various neurological disorders, including Schizophrenia, Attention Deficit Disorder, Alzheimer’s, and Parkinson’s disease (Bonomi et al., 2021; Liguori et al., 2016; Medin et al., 2019; Rowland et al., 2016). Thus, there is great interest in detecting both major alterations in cerebral lactate levels under pathological conditions and subtle alterations in cerebral lactate levels under normal circumstances.

Proton magnetic resonance spectroscopy (¹H-MRS) is a non-invasive technique for measuring brain metabolites in vivo (van der Graaf, 2010). Detection of lactate by ¹H-MRS using a single echo time, which generally appears as a doublet at a frequency of approximately 1.3 ppm, can be difficult due to its low abundance and spectral overlap with lipids, particularly at short echo times. However, increased detection of lactate and suppression of water signal can be achieved using an array of echo times and 2-dimensional J-resolved MRS to detect the J-coupling behavior of lactate (Weaver et al., 2015) or by adjusting the echo time such that the lactate peaks are inverted relative to lipids. MRS studies typically require that animals are anesthetized to prevent any movement during the scanning process. Assessment of lactate by ¹H-MRS can be confounded by elevation of this metabolite following exposure to halogenated volatile anesthetics such as isoflurane and sevoflurane (Boretius et al., 2013; Horn & Klein, 2010; Makaryus et al., 2011; L. Zhang et al., 2022). Moreover, ¹H-MRS measurement revealed that lactate levels in the brain are specifically raised while other macromolecules remain relatively stable when isoflurane dose was increased from 1% to 2%, despite alterations to diffusion of intracellular metabolites (Valette et al., 2007). Several lines of evidence suggest that isoflurane interferes with mitochondrial function (Fedorov et al., 2023; Kishikawa et al., 2018; Y. Zhang et al., 2012), possibly by inhibiting complex I (Zimin et al., 2018), thereby promoting NADH accumulation which, in turn, can lead to increased lactate production to maintain ATP levels. Therefore, it is critical to carefully control for the concentration and duration of exposure of isoflurane when assessing cerebral lactate levels by ¹H-MRS. Moreover, this is especially important when attempting to measure non-pathological changes in cerebral lactate which may be of a small magnitude.

The astrocyte-to-neuron lactate shuttle hypothesis posits that astrocytes produce and release lactate, which is then transported and oxidized by adjacent neurons to fuel synaptic transmission, a highly energy demanding process (Barros & Deitmer, 2010; Pellerin & Magistretti, 1994). Many studies corroborate the seminal 2011 study (Suzuki et al., 2011) showing the astrocyte-neuron lactate shuttle fuels cognition. However, elevated local lactate associated with neuronal activity (Hu & Wilson, 1997; Prichard et al., 1991; Sappey-Marinier et al., 1992) may also be explained by neuronal glycolysis and lactate release (Díaz-García et al., 2017; Ivanov et al., 2014). To assess the role of neuronal-produced lactate on hippocampal-dependent cognitive function, we generated a transgenic mouse with inducible expression of lactate dehydrogenase A (Ldha) within cortical and hippocampal neurons. Because LDHA is predominantly biased towards the conversion of pyruvate to lactate (Kaplan et al., 1968; Vesell, 1965), we anticipated that neuronal Ldha induction mice would promote an increase in hippocampal lactate levels. Surprisingly, hippocampal lactate levels did not differ in neuronal Ldha induction mice compared with control mice when measured using ¹H-MRS analysis under standard isoflurane anesthetic conditions. Because standard anesthetic conditions employ high level isoflurane exposure, which can artificially elevate cerebral lactate levels (Boretius et al., 2013; Horn & Klein, 2010; Makaryus et al., 2011; L. Zhang et al., 2022), we hypothesized that progressively increasing isoflurane concentrations over a prolonged period of time would allow the brain to adapt and lessen the artificial elevation of lactate levels typically caused by high isoflurane levels initially used for induction of anesthesia. Using a modified anesthetic protocol in which mice were progressively exposed to increasing concentrations of isoflurane over a longer period of time revealed a significant increase in hippocampal lactate levels in neuronal Ldha induction compared to control mice. Our findings indicate that conventional ¹H-MRS protocols using isoflurane at concentrations that are not systematically controlled may be inadequate for detection of cerebral lactate under non-pathological conditions. Use of a protocol whereby isoflurane concentration and duration of exposure is precisely monitored, such as the isoflurane titration method utilized in this study, is recommended to improve detection of subtle increases of hippocampal lactate by ¹H-MRS.

All animal procedures were conducted in accordance with Canadian Council of Animal Care guidelines and protocols (2011-079 and 2020-112) approved by the animal care committee (ACC) of Western University. Male mice were housed in shoebox cages under standard conditions, including 22–25°C temperature control, 12 h light-dark cycle, and ad libitum access to chow diet (Teklad, 2018, Envigo) and water. Expression of hemagglutanin (HA) tagged Ldha in neurons was achieved by crossing neuronal tetracycline transactivator driver mice (calcium-calmodulin-dependent kinase II (CaMKIIα) promoter; CaMKII-tTA; 007004) (Mayford et al., 1996) obtained from The Jackson Laboratory and mice containing a transgene with a tetracycline response element (TRE) promoter driving HA-tagged Ldha expression (TRE-LDHA) produced in house (RRID:MGI:7645158), as previously described (Frame et al., 2024). The TetOff system (Gossen & Bujard, 1992; Schönig & Bujard, 2013) utilized in these transgenic mice was restricted to adulthood by withdrawal of doxycycline (S3888, Bio-Serv) from the diet during the post-weaning period (Fig. 1A). Male mice were scanned using ¹H-MRS at 7–9 months of age and euthanized at 9.5 months of age with carbon dioxide for brains to be harvested. Prior to brains being removed, mice were cardiac perfused with phosphate buffered saline (PBS) containing phenylmethylsulfonyl fluoride (PMSF) protease inhibitor (P7626, Millipore Sigma) and sodium orthovanadate phosphatase inhibitor (S6508, Millipore Sigma).

During scans using a standard isoflurane protocol, isoflurane concentration was started at 4% and then maintained between 1.5% and 2.5% (Fig. 2A) with an oxygen flow rate of 1−1.5 L/min through a custom-built nose cone. For scans where the isoflurane was titrated, isoflurane concentration was maintained for 40 min prior to each scan and changed sequentially from 1.6% to 2% to 3% (Fig. 3A). Animal temperature was monitored with a rectal temperature probe, and respiration was monitored with a pneumatic pillow connected to a pressure transducer that was placed on the thoracic region. Body temperature was maintained at 36.9−37.1°C throughout imaging by blowing warm air over the animal using a model 1,025 small-animal monitoring and gating system (SA Instruments Inc., Stony Brook, NY, USA). Magnetic resonance imaging (MRI) experiments were performed on a 9.4-T/31-cm small animal MRI scanner (Agilent, Palo Alto, CA, USA) interfaced to a Bruker Avance III HD console (Bruker BioSpin Corp, Billerica, MA) and equipped with a 6-cm gradient coil of 1,000 mT/m strength, running Paravision-6 software at the Centre for Functional and Metabolic Mapping located within the Robarts Research Institute at the University of Western Ontario. A Varian 30-mm millipede volume radiofrequency coil was used for data collection. At the beginning of each scan, coronal T2-weighted anatomical images were acquired using a TurboRARE2D pulse sequence (16 averages, 31 slices with slice thickness of 0.5 mm, FOV = 19.2 × 19.2 mm², matrix size = 128 × 128, in-plane resolution = 0.15 × 0.15 mm², TE = 40 ms, TR = 5.0 s, echo spacing = 10 ms, and rare factor = 8). For planning the MRS voxel, a 2 × 6 × 3 mm³ voxel was positioned over both hippocampi (Supplementary Fig. S1) for MRS data acquisition using the semi-LASER (Semi-Localization by Adiabatic Selective Refocusing) sequence that provides localized spectra from rectangular voxels selected with a sequence of a 90° excitation pulse and two pairs of 180° adiabatic refocusing pulses (TR/TE = 5,000/136 ms) (Garwood & DelaBarre, 2001). A water spectrum was acquired first (8 acquisitions), followed by a metabolite spectrum with water suppression (128 acquisitions). Metabolite spectra (including N-acetylaspartate, lactate, alanine, glutamate, glutamine, creatine (total creatine, including creatine and phosphocreatine), taurine, choline, glycerophosphocholine, and myoinositol) were post-processed using eddy current correction (ECC) and fitted using the fitMAN software (Bartha et al., 1999) incorporated into a graphical user interface written in the IDL (Interactive Data Language) programming language (Kassem & Bartha, 2003) to determine the amplitude of lactate and creatine resonances. The lactate over creatine ratio was calculated by dividing the sum of each amplitude resonance after correction for the T2 relaxation time-constant of lactate (161) and creatine (104) using the formula: as previously described (Wong et al., 2018).

The left hemisphere of each brain was post-fixed in 4% paraformaldehyde (15713, Electron Microscopy Sciences) in phosphate buffer for 3 days, stored in ethanol, embedded in paraffin wax, and sectioned with a rotary microtome (RM2055, Leica Biosystems) at a 5 µm thickness sagittally until the hippocampal dentate gyrus was visible. Sections were baked onto glass slides (12-550-15, ThermoFisher Scientific), deparaffinized, and subjected to heat-induced epitope retrieval with 10 mM sodium citrate buffer pH 6.0 (C7129, Millipore Sigma). Background autofluorescence was quenched with UV exposure, 1 mg/ml sodium borohydride, and TrueBlack (23007, Biotium). Sections were blocked with goat anti-mouse affinity-purified fab fragment antibodies (115-007-003, Jackson ImmunoResearch Inc.; 1:40) and Background Sniper (BS966, Biocare Medical). Primary antibodies were applied at 4°C overnight, including mouse anti-HA.11 epitope tag (901513, BioLegend; 1:500), rabbit anti-MAP2 (ab32454, Abcam; 1:1,000), and secondary antibodies were applied at room temperature for 40 min, including goat anti-mouse Alexa Fluor 568 (A11031, Thermofisher Scientific; 1:500) and goat anti-rabbit Alexa Fluor 647 (A21244, ThermoFisher Scientific; 1:500). 4’,6-diamidino-2-phenylindole (DAPI; D1306, ThermoFisher Scientific; 1:300) was used to counterstain for nuclei. Fluorescence microscopy images were taken of each section using a Nikon ECLIPSE Ti2-E microscope system with plan apochromatic lambda dry 20 x objective (MRD00205, Nikon Instruments) and a monochrome backside illuminated scientific complementary metal–oxide–semiconductor image sensor (pco.edge 4.2 bi, Excelitas Technologies). The excitation and emission wavelengths (nm) used for detection of fluorescence from Alexa Fluor 568 were 554 and 595, and from Alexa Fluor 647 were 635 and 681. NIS-Elements AR (RRID:SCR_014329, Nikon Instruments) with the 2D deconvolution module was used to automatically remove out-of-focus light.

Protein was extracted from the right hemisphere of each brain using a lysis buffer containing 250 mM sucrose (S0389, Millipore Sigma), 50 mM tris (BP152, ThermoFisher Scientific), 25 mM potassium chloride (P4504, Millipore Sigma), 1% triton X (T9284, Millipore Sigma), 0.5 mM PMSF (P7626, Millipore Sigma), 1X Halt™ protease inhibitor cocktail (87786, ThermoFisher Scientific), and 0.1 mM sodium orthovanadate (S6508, Millipore Sigma) and quantified using a detergent compatible assay (5000111, Bio-Rad). Protein extracts, alongside a protein standard (1610373, Bio-Rad), were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene fluoride (PVDF) membrane blot. Blots were blocked with 1% w/v bovine serum albumin (BSA) and 3% w/v milk, probed with primary antibodies at 4°C overnight, including rabbit anti-LDHA (2012, Cell Signaling; 1:1,000) and mouse anti-actin (sc-47778, Santa Cruz Biotechnology; 1:2,000), and secondary antibodies at room temperature for 1 h, including goat anti-mouse (AP130P, Millipore Sigma; 1:10,000) and goat anti-rabbit (AP132P, Millipore Sigma; 1:10,000) conjugated with horse radish peroxidase (HRP). Chemiluminescent signal was detected using Forte Western HRP substrate (WBLUF, Millipore Sigma), imaged using a ChemiDoc XRS imaging system (170-8070, Bio-Rad), and quantified using Image Lab software (RRID:SCR_014210, Bio-Rad). Western blot HA band intensity was standardized to actin band intensity for quantification.

Data were analyzed statistically and visualized using GraphPad Prism version 10.0.0 (153) (RRID:SCR_002798). Data were presented as mean ± SEM with statistical comparisons described in each figure legend. Power analysis was conducted using the pwr package (Champely, 2018) in R version 4.3.1 (2023-06-16 ucrt, RRID:SCR_001905) using RStudio (RRID:SCR_000432) (R Core Team, 2015; RStudio Team, 2020).

To assess the effects of altered lactate levels in the brain by ¹H-MRS, we generated a transgenic mouse line with a hemagglutinin (HA)-tagged murine Ldha cDNA under the control of the tetracycline response element promoter (TRE-LDHA). This line was subsequently crossed to transgenic mice expressing the tetracycline transactivator (tTA) under regulatory control of the neuron-specific calcium/calmodulin-dependent kinase IIα promoter (CaMKIIα-tTA), a driver highly expressed in the cortex and hippocampus (Fig. 1A). In dual transgenic mice, induced expression of Ldha in CNS neurons was achieved following the removal of doxycycline, a tetracycline analogue, from the diet of weaned mice. Induction of LDHA was confirmed by both immunofluorescence microscopy (Fig. 1B) and western blotting (Fig. 1C).

To determine if neuronal induction of Ldha expression promoted increased lactate production in the hippocampus, a key brain structure required for learning and memory, we performed ¹H-MRS analysis under standard anesthesia conditions using isoflurane at a concentration of 4% for induction followed by 1.6–2% for maintenance (Fig. 2A). Surprisingly, no significant elevation of hippocampal lactate/creatine levels was detected in neuronal induction Ldha mice compared to control mice (Fig. 2B). Typically, the time between isoflurane induction followed by a maintenance level of isoflurane is generally not systematically controlled in most studies. Therefore, the absence of altered hippocampal lactate levels in Ldha induction mice relative to controls may have potentially been attributed to confounding effects related to the concentration and duration of isoflurane exposure used for induction and maintenance of anesthesia.

A previous study by Horn and Klein (2010) demonstrated that continuous isoflurane administration promotes a progressive increase in cerebral lactate levels that eventually reaches a plateau. In addition, lactate levels increase in a manner proportional to isoflurane concentration (Boretius et al., 2013; Horn & Klein, 2010). Mice exposed to 1.75% isoflurane followed by a period of no anesthetic show a pronounced drop in cerebral lactate levels (Boretius et al., 2013). Interestingly, following re-exposure to isoflurane, lactate levels once again increased but not to the same level as during the initial exposure (Boretius et al., 2013). In addition, numerous studies have shown that isoflurane preconditioning leads to adaptive responses in the brain (McMurtrey & Zuo, 2010). Thus, we hypothesized that progressively increasing isoflurane concentrations over a prolonged period of time would allow the brain to adapt and lessen the artificial elevation of lactate levels typically caused by high isoflurane levels initially used for induction of anesthesia.

To test this hypothesis, we measured hippocampal lactate/creatine in neuronal Ldha induction mice with an altered ¹H-MRS protocol by systematically titrating the isoflurane concentration from 1.6%, 2%, and 3% (Fig. 3A). This protocol specifies that scanning occurs 40 min after each isoflurane concentration increase to allow for lactate levels to stabilize (Fig. 3A). As expected, we found that increasing the isoflurane concentration resulted in an increase in hippocampal lactate/creatine levels in all mice (Fig. 3B). However, by employing a sequential elevation in the concentration of isoflurane used for anesthesia revealed a significant increase in hippocampal lactate/creatine in neuronal Ldha induction mice compared with control mice (Fig. 3B). Importantly, the absence of an isoflurane by genotype effect indicates that these two genotypes do not differ in their susceptibility to isoflurane induction of lactate (Fig. 3B).

Normalization of lactate to creatine eliminates two problems associated with measurement of absolute values, including variance introduced by measurement of tissues within the voxel that are required to distinguish cerebrospinal fluid and tissue fractions, and error introduced by scaling to water in the voxel. Nonetheless, normalizing to creatine could potentially be problematic if this metabolite is also affected by isoflurane. Thus, we quantified absolute levels of both lactate and creatine under standard and isoflurane titration concentrations. Similar to lactate/creatine measurements, standard isoflurane conditions revealed no differences in absolute lactate levels between control and Ldha induction mice (Supplementary Fig. S2A). As observed in lactate/creatine measurements, the progressive increase in isoflurane concentration resulted in elevated absolute levels of lactate in all mice (Supplementary Fig. S2B). However, Ldha induction mice showed a trend of increased absolute lactate levels similar to the lactate/creatine measurements (Supplementary Fig. S2B). In contrast, absolute levels of creatine did not change under standard isoflurane conditions (Supplementary Fig. S2C) or with the isoflurane titration protocol (Supplementary Fig. S2D) for either control or Ldha induction mice. Thus, the genotype specific elevation of lactate/creatine levels in Ldha induction mice is attributed to changes in lactate but not creatine.

These data provide strong evidence that isoflurane titration is a valuable method for measuring increases in hippocampal lactate by ¹H-MRS in genetically manipulated mice.

In the present study, we demonstrate the utility of systematically controlling isoflurane concentration used for anesthesia during measurement of ¹H-MRS-based hippocampal lactate levels in mice. We show here that mice genetically manipulated to promote increased capacity for lactate production within CNS neurons exhibit increased hippocampal lactate/creatine levels detectable by ¹H-MRS but only when isoflurane concentration is titrated. The utility of this method may be expanded to studies seeking to detect cerebral lactate under circumstances where small changes are expected. For example, animal models of neurodegenerative disease which may have progressive accumulation of cerebral lactate (Hagihara et al., 2024) may employ isoflurane titration during ¹H-MRS to detect small increases in lactate that develop earlier. Furthermore, the increased power afforded by our isoflurane titration method also reduced the number of animals required to detect changes in hippocampal lactate. If we intended on designing a study using standard anesthetic conditions to achieve an 80% chance of detecting an increase in hippocampal lactate in our neuronal Ldha induction mice with the effect size we observed in this study (Fig. 2B; Cohen’s d = 0.63), then a power analysis (α = 0.05, β = 0.2) indicates that we would have required over 63 animals per condition. In contrast, we were able to detect an increase in hippocampal lactate in our neuronal Ldha induction mice applying our isoflurane titration method using less than 5 animals per condition. Therefore, isoflurane titration during cerebral lactate measurement with ¹H-MRS may facilitate more humane studies, according to the reduction principle of human experimental technique (Russell & Burch, 1959) while increasing the probability of detecting changes that would have otherwise been undiscovered under standard anesthesia conditions.

In this study, lactate levels were standardized to a reference peak representing creatine and phosphocreatine combined as a measure of total creatine. In the past, studies have shown that total creatine in the brain is largely unaffected by isoflurane level used for anesthesia (Boretius et al., 2013; Söbbeler et al., 2018). A relatively low concentration of isoflurane (1.75%) compared to no isoflurane caused lactate/creatine levels to increase greatly (533%) (Boretius et al., 2013). Here, we showed that absolute levels of creatine do not change with increasing concentrations of isoflurane. This suggests that elevated lactate/creatine levels with increasing isoflurane exposure are primarily due to changes in lactate and not creatine.

One could seek alternative methods of assessing lactate in the brain in vivo while avoiding potential confounds introduced by anesthesia. For example, a neuromuscular blocking agent could be administered to mice without anesthesia such that MRS can be performed while mice are incapable of movement. However, the use of a neuromuscular blocking agent could also cause confounding effects. For example, does the chosen neuromuscular blocking agent impact CNS lactate? Does a lack of anesthesia allow the animal to perceive the MRS procedure as stressful and leave the animal susceptible to stress-induced CNS lactate alterations? Moreover, local animal care committees, following institutional and governmental guidelines that differ around the world, may not consider it ethical to administer a potentially stressful procedure without anesthetic. For example, the Canadian Council on Animal Care (CCAC) suggests that anesthesia should be used during potentially stressful procedures (CCAC, 2019). Therefore, while alternatives to anesthesia may avoid confounding anesthesia related effects or can provide a non-anesthetized baseline in vivo measurement, ethical concerns should be also considered. In light of the ubiquity in which volatile anesthetics are used during MRS procedures, ideally customizing a protocol that retains anesthesia while permitting the ability to discriminate between the effects of anesthesia versus physiological or pathological changes would provide the most utility.

In recent years, there have been newly developed adaptations to MRS methods for detection of lactate, such as deuterium metabolic imaging (DMI) (De Feyter et al., 2018; Flatt et al., 2021), quantitative exchanged-label turn-over MRS (qMRS) (Rich et al., 2020), diffusion-weighted magnetic resonance spectroscopy (DW-MRS) (Ligneul et al., 2019), chemical shift imaging (CSI) (Brender et al., 2019), double selective multiple quantum filter technique (SelMQC) (Pickup et al., 2008), and lactate chemical exchange saturation transfer LATEST (DeBrosse et al., 2016). To our knowledge, systematic titration of anesthetic concentration has not been adopted as a method for lactate detection using any of the aforementioned newer methods. Moreover, these studies are similarly susceptible to isoflurane-induced changes in cerebral lactate. Therefore, adding isoflurane titration to any MRS protocol offers a safe and effective strategy for reducing isoflurane-induced lactate changes regardless of the detection method. Overall, future use of the simple modification to MRS protocols we describe here has the potential to greatly increase the power and detection capabilities when measuring cerebral lactate levels in animal models.

This study provides a relatively simple protocol that can enhance the use of isoflurane anesthesia during MRS assessment of brain lactate levels in mice. However, alternative methods of conducting MRS without the use of anesthesia could eliminate or reduce artificial anesthesia-induced effects on brain metabolites, particularly lactate (Ferris, 2022; Madularu et al., 2017). MRS without anesthesia is limited by potential confounding effects that come with increased stress and required acclimation. Thus, until MRS analysis of awake mice has been perfected, and as long as isoflurane remains the most common method of anesthesia for MRS, the protocol described here should still retain utility.

All raw data and statistical analyses are available on the Borealis data repository (https://doi.org/10.5683/SP3/ROJAYY).

Conceptualization, A.K.F. and R.C.C.; Methodology, A.K.F., M.B., R.B., and R.C.C.; Software, A.K.F., S.K.W., and R.B.; Formal Analysis, A.K.F.; Investigation, A.K.F., R.K., M.C., M.B., and A.X.L.; Resources, R.B. and R.C.C.; Data Curation, A.K.F. and S.K.W.; Writing—Original Draft, A.K.F. and R.C.C.; Writing—Review & Editing, A.K.F., M.B., A.X.L., R.B., and R.C.C.; Visualization, A.K.F.; Supervision, R.B. and R.C.C.; Project Administration, A.K.F. and R.C.C.; and Funding Acquisition, R.B. and R.C.C.

A.K.F. received an Ontario Graduate Scholarship and Postgraduate Scholarship — Doctoral from the Natural Sciences and Engineering Research Council of Canada. R.K. received a Canada Graduate Scholarships — Doctoral from the Natural Sciences and Engineering Research Council of Canada. M.C. received an Ontario Graduate Scholarship. Funding support for R.C.C. was provided by the Natural Sciences and Engineering Research Council of Canada, grant number RGPIN-2019-355803. Funding support for R.B. was provided by the Natural Sciences and Engineering Research Council of Canada, grant number RGPIN-2020-05290, and the Canadian Consortium on Neurodegeneration in Aging, grant # 137794 (Phase I) and 163902 (Phase 2) which receives funding from the Canadian Institutes of Health Research and several other partner organizations.

All animal procedures were approved by the Western University animal care committee.

The authors declare no competing interests.

We thank Dr. Richard Harris for helping create the TRE-LDHA transgenic mouse line and the Biotron integrated microscopy facility at Western university for equipment/software use & support.

Supplementary material for this article is available with the online version here: https://doi.org/10.1162/imag_a_00305