Original Article

Stimulation of astrocyte fatty acid oxidation by thyroid hormone is protective against ischemic stroke-induced damage

Journal of Cerebral Blood Flow & Metabolism 0(00) 1–14 ! Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0271678X16629153 jcbfm.sagepub.com

Naomi L Sayre1, Mikaela Sifuentes2, Deborah Holstein2, Sheue-yann Cheng3, Xuguang Zhu3 and James D Lechleiter2,4

Abstract We previously demonstrated that stimulation of astrocyte mitochondrial ATP production via P2Y1 receptor agonists was neuroprotective after cerebral ischemic stroke. Another mechanism that increases ATP production is fatty acid oxidation (FAO). We show that in primary human astrocytes, FAO and ATP production are stimulated by 3,3,5 triiodo-L-thyronine (T3). We tested whether T3-stimulated FAO enhances neuroprotection, and show that T3 increased astrocyte survival after either hydrogen peroxide exposure or oxygen glucose deprivation. T3-mediated ATP production and protection were both eliminated with etomoxir, an inhibitor of FAO. T3-mediated protection in vitro was also dependent on astrocytes expressing HADHA (hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase), which we previously showed was critical for T3-mediated FAO in fibroblasts. Consistent with previous reports, T3-treatment decreased stroke volumes in mice. While T3 decreased stroke volume in etomoxir-treated mice, T3 had no protective effect on stroke volume in HADHA þ/ mice or in mice unable to upregulate astrocyte-specific energy production. In vivo, 95% of HADHA co-localize with glial-fibrillary acidic protein, suggesting the effect of HADHA is astrocyte mediated. These results suggest that astrocyte-FAO modulates lesion size and is required for T3-mediated neuroprotection post-stroke. To our knowledge, this is the first report of a neuroprotective role for FAO in the brain.

Keywords Astrocyte, fatty acid oxidation, neuroprotection, stroke, thyroid hormone Received 28 February 2015; Revised 27 August 2015; Accepted 26 December 2015

Introduction Currently, the only treatment available to victims of ischemic stroke is lysis of blood clots, which can be used in only a small fraction of patients.1 Stroke is the 5th leading cause of death and the primary cause of long-term disability in the USA.2 Healthy outcomes after stroke could be greatly increased with more efficacious treatment options. Research strategies to minimize brain damage and enhance repair after stroke have focused almost exclusively on neurons but have met with little success. An alternative therapeutic target is astrocytes, which are inherently neuroprotective.3 Our laboratory reported that increasing astrocyte mitochondrial metabolism via purinergic signaling was neuroprotective for in vivo mouse models of stroke4 and aging.5

1 Department of Neurosurgery, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA 2 Department of Cellular & Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA 3 Gene Regulation Section, Laboratory of Molecular Biology, National Institutes of Health, Bethesda, MD, 20892 4 Department of Cellular & Structural Biology, South Texas Research Facility Neuroscience Center, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

Corresponding author: James D Lechleiter, Department of Cellular & Structural Biology, South Texas Research Facility Neuroscience Center, University of Texas Health Science Center at San Antonio, 8403 Floyd Curl Drive, San Antonio, TX 78229-3904, USA. Email: [email protected]

2 Signaling pathways that enable mitochondrial ATP production present attractive targets for stroke treatment. In particular, mitochondrial metabolism in astrocytes can play a significant role in neuroprotection, because astrocytes are essential in supporting neuronal function and influencing brain energy homeostasis.3,6,7 In support of this, the presence of astrocytes in neuronal cultures enhances neuronal resistance to oxidative stress.8 Similarly, we reported that increasing astrocyte mitochondrial metabolism protects neurons from oxidative stress in vitro.5 Thus, targeting astrocyte metabolism to increase brain ATP levels is a viable strategy to enhance neuroprotection after injury. Furthermore, by increasing energy production, multiple neuroprotective processes can benefit. Such a therapeutic strategy is more robust than studying a single molecular target that may not be conserved in humans. Here, we investigated the neuroprotective efficacy of a complementary pathway (fatty acid oxidation, FAO) in astrocytes that also stimulates mitochondrial metabolism. It has recently become appreciated that FAO by astrocytes accounts for a portion of their mitochondrial metabolism.9–11 Though most energy in the brain is derived from glucose catabolism, significant energy can also be derived from FAO. In culture, astrocytes utilize fatty acids for energy creation9,10,12 Studies in rat brains revealed that oxidation of the fatty acid octanoate contributes approximately 20% of brain oxidative energy production.11 Similarly, Escartin et al. showed that enhancing astrocyte FAO is protective of neuronal and astrocyte cell death when astrocyte glycolysis is inhibited.13 One way to potentially stimulate FAO involves treatment with thyroid hormones, which is known to increase FAO in the heart,14 muscle,15–17 liver,18 and in skin cells.19 We showed that mitochondrial energy production can be rapidly increased via a mitochondrial targeted thyroid hormone receptor (mTR) after treatment with 3,30 5-triiodo-L-thyronine (T3).20 We also demonstrated that T3 stabilizes subunit assembly of the mitochondrial trifunctional protein FAO complex in an mTR-dependent and genome-independent manner. This, in turn, correlated with increases FAO and energy production.19 Moreover, we showed that a critical component of the T3-mediated energy production involved expression of hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase alpha (HADHA), an essential component of the mitochondrial trifunctional protein complex for long-chain FAO.19 We report that T3 increases astrocyte ATP concentrations in a FAO-dependent manner. T3 treatment was associated with greater astrocyte viability after oxidative stress and oxygen glucose deprivation (OGD).

Journal of Cerebral Blood Flow & Metabolism In vitro, T3-mediated neuroprotection was dependent upon long-chain FAO. Astrocytes that lacked expression of HADHA failed to increase viability in response to T3 after OGD. To measure HADHA localization in vivo, we tested the percentage of HADHA colocalizing with glial-fibrillary acidic protein (GFAP) astrocytes in the cortex and hippocampus, and found that the majority (93–95%) is expressed in astrocytes. We similarly used an in vivo focal ischemia model in mice to determine whether FAO and T3 treatment was protective after stroke. Consistent with previous reports,21,22 thyroid hormone treated mice had significantly smaller lesion size at 24 h. Inhibition of FAO with the pharmacological agent etomoxir significantly increased stroke lesion size. To better test the role for HADHA in neuroprotection, we utilized heterozygous mice (HADHA null mice do not survive past infancy23). T3-treatments were unable to decrease stroke damage in HADHA þ/ mice. Moreover, inhibition of the tricarboxylic acid cycle specifically within astrocytes using fluoroacetate also completely ablated the neuroprotective efficacy of T3 after stroke. Taken together, our data display that inhibition of FAO not only worsens the severity of stroke but that astrocyticspecific expression of HADHA plays an important role in T3-mediated neuroprotection post-stroke.

Materials and methods Cell culture media and reagents Unless otherwise specified, astrocytes were cultured in culture media: DMEM:F12 (Life technologies, Austin TX) containing 10% fetal calf serum (ATCC, Manassas, VA, USA), glutaMAX (2 mM, Life technologies), and Primocin (100 mg/mL, Invivogen, San Diego, CA, USA). 3,3,5 triiodothyronine (EMD Millipore, Billerica, MA, USA) was always prepared just before experiments by diluting from a fresh stock solution in 0.1 N NaOH. Etomoxir (Sigma, St. Louis, MO, USA) was prepared fresh by diluting in deionized H2O before experiments. 20 stock solutions of [9,10-3H(N)]palmitate/BSA (NET043001MC, PerkinElmer, Waltham, MA, USA) were prepared as previously described24: Ethanol stocks of [3H]palmitate were diluted into unlabeled 4 mM palmitic acid (MP Biomedicals, Solon, OH, USA) and the ethanol was evaporated. The palmitate mixture was then resuspended in fatty acid-free Bovine Serum Albumin (BSA, 10 mg/mL; Calbiochem, San Diego, CA, USA) and warmed with occasional sonication until palmitate was fully dissolved. The specific activity of [3H]palmitate/BSA was approximately 8000 DPM/ nmol for each preparation.

Sayre et al.

3

Isolation of human astrocytes

Immunocytochemistry of human astrocytes

In collaboration with the Department of Neurosurgery (University of Texas Health Science Center, San Antonio, TX, USA), fresh tissue was donated from patients undergoing temporal lobe resection for epilepsy. Temporal lobectomies were performed by an American Board-Certified Neurological Surgeon from the Department of Neurosurgery according to the established guidelines from the American Association of Neurological Surgeons and from University Hospital, San Antonio. All tissue-harvesting procedures were performed under approval from the University of Texas Health Science Center Institutional Review Board with patient consent. Investigators had no access to any identifiable information about the patients, other than the age and sex of the patient. Post-surgery, tissue was handled according to protocols approved by the University of Texas Health Science Center, San Antonio Institutional Biosafety Committee. To harvest astrocytes, brain tissue was first manually chopped into small pieces using a single edge carbon razor blade for approximately 1 min. To dissociate cells, tissue was incubated in 0.2% trypsin-EDTA (Life Technologies) for 20 min at 37 C. Trypsin was neutralized with culture media, and the homogenate was drawn through a sterile glass Pasteur pipette 10 times. The homogenate was centrifuged and the supernatant was decanted. The pellet was resuspended in fresh culture media which was drawn again through a sterile glass pipette 10 times. The supernatant was filtered through a 40 mm mesh filter to remove any large debris. The filtered homogenate was cultured at 37 C, 5% CO2 for 24 h, after which the homogenate was removed to a second culture vessel for another 24 h. Cells that had attached to the culture vessel were washed at 24 h post isolation in cold-phosphate buffered saline (PBS; Sigma), and were allowed to grow to confluence in culture media. For experiments, primary human astrocytes were used before they had reached passage 5.

Primary human astrocytes were plated onto glassbottom cell culture dishes (MatTek, Ashland, MA, USA) in culture media and grown to 80% confluence. Washed astrocytes were fixed in 4% paraformaldehyde, and with the exception of astrocytes stained with O4, astrocytes were permeabilized in 0.1% triton X-100 (Sigma) for five minutes, blocked in 5% bovine serum albumin (BSA; Sigma), and then incubated in primary antibody overnight at 4 C. Astrocytes were then washed and then incubated in a 1:100 dilution of fluorophore-conjugated secondary antibody (Life Technologies). All antibodies were prepared in blocking solution, and the antibodies and dilutions used are as follows: a-GFAP (Abcam 7260, 1:1000), a-S100B (Abcam 52642, 1:100), a-MAP2 (Abcam 92434, 1:1000), a-NeuN (Abcam 177487, 1:250), aoligodendrocyte (Millipore MAB1580, 1:1000), a-O4 (Millipore Mab345, 1:100), a-cd11B (Abcam 62817, 1:50), and a-Nestin (Millipore AB5922, 1:200).

Isolation of mouse embryonic astrocytes To isolate embryonic astrocytes, a female mouse that had been bred 14 days prior was humanely sacrificed, and the embryos were individually removed. The brain was removed, titrated with a glass pipette to homogenize, and then grown in culture media containing FibroOut (Chi Scientific, Maynard, MA, USA) until confluent. At this time, media were changed to normal culture media. For experiments, primary astrocytes were used before they had reached passage 5.

Fatty acid oxidation of [3H]palmitate FAO measurements were conducted according to the protocol described by Moon and Rhead, as previously described.24 To obtain the indicated concentration of [3H]palmitate, [3H]palmitate was diluted 20 times using PBS containing Mg2þ and Ca2þ to obtain a working solution in 0.5 mg/mL BSA. Further dilutions were made with BSA/PBS (0.5 mg/mL) to maintain the same BSA concentration across dilutions. On day 0, 1X105 human astrocytes were plated onto six-well culture plates. On day 2, astrocytes were washed 1X with PBS, and the culture media were changed to media containing 10% charcoal-stripped fetal calf serum (ATCC) to lessen the amount of serum hormones fed to the astrocytes. On day 3, astrocytes were subjected to the FAO assay. Astrocytes were washed 2X with PBS, and then incubated in 300 mL of [3H]palmitate/BSA for 3 h. The [3H]palmitate/BSA was harvested into 1.7 mL centrifuge tubes containing 10% trichloroacetic acid (200 mL; Sigma); astrocytes were washed with 100 mL PBS and the wash was pooled with the [3H]palmitate/BSA. After 2 min of incubation, the protein was spun out of solution for 5 min at 8500  g. The supernatant was collected into new tubes containing 6 N NaOH (70 mL) to neutralize acid. The mixture was then poured over mini-columns containing 12, 200–400 mesh Dowex resin (Acros, Geel Belgium) to elute [3H]H2O. [3H]H2O was measured via scintillation to determine the amount of palmitate metabolized. The protein remaining in cell culture dishes was determined via bicinchoninic assay (Thermo-Fisher, Waltham, MA, USA).

4

ATP determination ATP was determined similar to previous experiments.24 On day 0, 1  104 astrocytes were plated onto Costar white-bottom 96-well plates in culture media. On day 1, astrocytes were washed 1XPBS, and replaced with culture media containing charcoal-stripped serum. On day 2, ATP determination was performed using the ATP lite kit (Perkin Elmer) following manufacturer’s instructions.

Oxygen glucose deprivation (OGD) and hydrogen peroxide mediated astrocyte killing On day 0, 1  104 astrocytes were plated onto 12-well glass bottom tissue culture dishes. On day 1, astrocytes were refed with culture media containing 10% charcoal stripped serum. On day 2, astrocytes were subjected to oxygen glucose deprivation or to hydrogen peroxide killing as previously described.25 For OGD, astrocytes were treated with the indicated concentration of T3 in glucose-free Dulbecco’s modified eagle medium (Life Technologies). Astrocytes were then incubated in an anoxic chamber purged with 95% N2 and 5% CO2 for 4 h at 37 C. Immediately after OGD, cell viability was determined. For hydrogen peroxide-mediated killing, astrocytes were treated with H2O2 (100 mM, Sigma cat# 216763) for 4 h, and then viability was determined.

Cell viability determination Viability was determined using calcein-AM (Life technologies C3100MP) according to the manufacturer’s instructions, and as previously performed.25 The number of calcein-retaining viable cells was determined as a percentage of total Hoechst-labeled cells.

Immunohistochemistry of HADHA Wild-type albino C57B6N-Tyrc-Brd/BrdCrCrl mice were humanely sacrificed at three months of age, and the brains were removed directly into 4% paraformaldehyde/PBS. Brains were fixed at 4 C overnight, and then were transferred into 30% sucrose/PBS for three days. Brains were frozen in OCT solution (Sakura, Torrance, CA, USA), and then were cryosectioned onto gelatin-coated superfrost plus slides and stored at 20 C. For IHC, sections were allowed to dry overnight at RT, and then were rehydrated in PBS for 5 min. Heat-mediated antigen retrieval was performed by incubating slides at 95 C in 0.05% citraconic anhydride, pH 7.4 for 20 min, and then allowing the slides to cool in solution to RT. Slides were washed 1  5 minutes in PBS, and then were incubated in 0.1% tritonX100/PBS for 10 min. Slides were washed 2  5 minutes in PBS to remove detergent, and then sections were blocked for 1 h at RT in 5% bovine serum albumin

Journal of Cerebral Blood Flow & Metabolism (BSA)/PBS. Primary antibodies were prepared in 5% BSA/PBS, and were incubated on the sections overnight at 4 C. Anti-HADHA (Abcam ab54477, Cambridge, MA, USA) and anti-GFAP (Millipore ab5541, Billerica, MA, USA) were both diluted at 1:500 concentration. Slides were washed for 3  15 min in PBS, and then were incubated at RT with 1:200 dilution of species-specific secondary antibody in 5% BSA/ PBS: anti-rabbit alexa-488 and anti-chicken alexa-568 (Life Technologies). Slides were washed 1  5 minutes in PBS, and then incubated for 5 min with DAPI/PBS (5 mg/mL). Slides were washed 2  5 min in PBS, and then 1  briefly in distilled water, and allowed to dry. Sections were mounted with Aquapolymount (Polysciences Inc, Warrington, PA, USA) and coverslipped with 1.5 coverslips. For colocalization analysis, imaging was performed in the Core Optical Imaging Facility (which is supported by UTHSCSA, NIH-NCI P30 CA54174 (CTRC at UTHSCSA) and NIH-NIA P01AG19316) on a Prairie Systems Confocal system using a 25X APO WI NA 1.10 objective. High-resolution slices were imaged at 0.5 mm increments through the depth of the tissue section. For image processing, FijiImageJ (NIH) was used. Single-color image stacks were subjected to background removal, and the histogram was normalized throughout the stacks. The HADHA and GFAP channels were merged, and the histograms were again normalized to each other. The total amount of overlapping HADHA/GFAP was determined, as was the total amount of nonoverlapping HADHA. Altogether, three separate image stacks/mouse were obtained in both the cortex and the hippocampus of three separate mice. Overlap analysis was performed on a 5 mm stack of images from each region (10 images/region). All macros used for analysis are detailed in the Supplementary information. For HADHA reorganization analysis, we used procedures similar to those detailed by Ren et al.26 Tile scans over a 4–8 mm2 area around the necrotic lesion were captured using a Zeiss-LSM710 using a 20  Plan APO Chromat NA 0.8 objective. High-resolution slices were image at 0.5 mm increments through the tissue section. For image processing, 5 mm image stacks were projected and tiled into single images. HADHA polarization was determined by measuring the mean pixel intensity in normal brain parenchyma (defined as >500 mm away from penumbra tissue). Next, the total percentage area of pixels with intensity equal to or greater than the mean intensity in normal tissue and in penumbral tissue (defined as within 500 mm of the necrotic lesion) was determined. A lower percentage area of pixels was determined to have greater HADHA polarity. Altogether, 2– 3 separate tiled image stacks/mouse (approximately 32 individual image stacks/mouse) were obtained around

Sayre et al. the photothrombotic lesion of three separate mice per treatment group.

Rose-bengal mediated photothrombosis Procedures for photothrombosis are similar to those previously described.4 Except for thyroid hormone receptor null mice, only males were used for experiments. Mice were given approximately 2 mm craniotomies over the parietal somatosensory cortex, and then were injected intravenously with filtered Rosebengal dye diluted in sterile injectable saline (100 mL of 10 mg/mL). Photothrombosis was achieved by focusing 561 nm laser light (100% power, 10 ms dwell time) over a small field of vessels of approximately 20–30 mm in diameter until clots formed and the vessel was occluded (approximately 5 min).

Transient middle cerebral artery occlusion and western blotting of HADHA Transient ischemia of the left middle cerebral artery was achieved by inserting an intraluminal filament via the carotid artery. A loss of approximately 80% of hemispheric perfusion was confirmed after filament placement using a laser Doppler probe (Perimed) affixed to the skull. The filament remained in place for 1 h, and then was removed to allow reperfusion. Twenty-four hours after ischemia, brains were harvested and the ipsilateral and contralateral cerebral hemispheres were prepared for SDS-PAGE and western blotting. After blocking membranes, HADHA (Abcam ab54477, 1:1000) and actin (Santa Cruz Biotech, 1:10,000) were probed, and relative quantities of HADHA were determined using densitometry after normalization for actin loading.

TTC staining Twenty-four h post photothrombosis, mice were humanely sacrificed and the brain was harvested into coldPBS for 3 min. The brain was sliced into 1 mm sections using a mouse brain mold, and then sections were transferred into a solution of 2% 2, 3, 5-Triphenyl2H-tetrazolium chloride (TTC) in PBS. Color was allowed to develop at 37 C for 5–10 min. Afterwards, brains were fixed overnight in 4% paraformaldehyde. Slices were scanned at high resolution, and lesion size was determined using image J.

Mouse models and animal care All mouse procedures were carried out with approval from the University of Texas Health Science Center Institutional Animal Care and Use Committee, and

5 protocols were utilized which were consistent with the NIH guide on the care and use of animal models. ARRIVE guildelines for how to report animal experiments were followed. Animals for experiments were bred in the vivarium, and used for stroke experiments from 3 to 5 months of age. Wild-type albino C57B6N-Tyrc-Brd/BrdCrCrl were purchased from Charles River laboratories for breeding purposes. Heterozygote HADHA þ/ mice27 were a gift from Arnold Strauss (Nashville, TN, USA).

Statistical analysis Except where indicated, statistical significance was determined using ANOVA followed by Tukey’s HSD text. Statistical significance was defined as p < 0.05.

Results Primary human astrocytes as an in vitro model To improve the translational potential of our research, we chose to utilize primary astrocytes isolated from human tissue for in vitro experiments. Tissue was obtained in collaboration with the Department of Neurosurgery, UTHSCSA. Immunochemical analysis of astrocyte cultures confirmed that the majority of cells (98%) were positively labeled for antibodies to the astrocyte marker glial-fibrillary acidic protein (GFAP). Approximately 95% also labeled for another astrocyte marker, S100B. Less than 2% of cultured cells showed labeling for other brain cells such as neurons (Map2, NeuN), oligodendrocytes (O4), microglia (cd11B) or stem cells (Nestin) (Figure S1, Supplementary material).

Primary human astrocytes increase FAO in response to T3 treatment To test whether human astrocytes utilize fatty acids for energy, we measured the ability of astrocytes to degrade the long-chain fatty acid [9,10-3H(N)]palmitate into 3 H2O. Notably, this assay measures the ability of astrocytes to metabolize palmitate through at least five rounds of beta-oxidation. We found that human astrocytes do indeed oxidize palmitate (Figure 1(a)). To determine whether palmitate oxidation occurs due to mitochondrial FAO, astrocytes were pretreated with the carnitine palmitoyl transferase-1 inhibitor etomoxir. Etomoxir prevents entry of long-chain fatty acids into the mitochondria, and so can help determine the contribution of mitochondria to the observed FAO. Treatment with etomoxir completely ablated the palmitate oxidation observed in human astrocytes (Figure 1(b)), indicating that all observed FAO in human astrocytes occurred in the mitochondria. We previously showed that T3 treatment increased ATP production

6

Journal of Cerebral Blood Flow & Metabolism

Figure 1. T3 stimulates fatty acid oxidation in human primary astrocytes. Primary human astrocytes were tested for the ability to oxidize palmitate to water. (a) Increasing concentrations of specific radioactivity (0–200 mM) of palmitate were given to human astrocytes for 2 h, and the amount of 3H-palmitate metabolized into 3H2O was measured via scintillation. (b) Human astrocytes were treated with a dose curve of the CPT-1 inhibitor etomoxir (0–600 mM) for 1 h prior to measuring 3H-palmitate metabolism. (c) Astrocytes were stimulated with T3 (0–10 nM) and FAO of 3H-palmitate was determined. (d) ATP concentration in human astrocytes was measured 15 min after stimulation with T3 (0–10 nM) in astrocytes treated with vehicle (open circles) or pretreated with etomoxir (closed circles; 100 mM for 1 h). Results shown are mean  SEM of triplicate independent experiments (n ¼ 3–5) with astrocytes from at least two patients. *p < 0.05 using one-way ANOVA followed by Tukey’s HSD.

of fibroblasts in a FAO-dependent manner.24 Treating astrocytes with T3 similarly increased FAO by approximately 150% when used at a 2 nM concentration. Surprisingly, a 10 nM T3 concentration failed to increase FAO, indicating a possible negative feedback mechanism on FAO at high T3 concentrations (Figure 1(c)).

T3-stimulated ATP increases in human astrocytes is dependent on FAO We next measured ATP levels, and found that, similar to the FAO, the amount of ATP in astrocytes increased at T3 concentrations of 2 nM, but did not change at 10 nM. To test whether the ATP increase was dependent on FAO, etomoxir was used to inhibit astrocyte FAO before stimulation with T3. Inhibition of FAO prevented the T3-mediated increase in astrocyte ATP (Figure 1(d)), indicating that the rise in ATP levels was dependent on stimulation of FAO.

T3 improves human astrocyte survival after hydrogen peroxide treatment We previously showed that stimulation of astrocyte energy production via purinergic signaling is protective

during aging,25 stroke,4 and traumatic brain injury.28 To test whether T3-mediated stimulation of ATP production could similarly improve astrocyte survival after stress, human astrocytes were subjected to hydrogen peroxide treatment, which causes cells to lose viability (Figure 2(a) and (b); white arrows). Viability of cells was measured by the ability of astrocytes to retain Calcein inside the intact cell (Figure 2(a) and (b)). Normally, treating human astrocytes with hydrogen peroxide will cause cultures to be only 50% viable (Figure 2(a) and (c)). However, treatment with 1 nM of T3 improved viability to 80% (Figure 2(b) and (c)), a number much closer to the 90% viability of astrocytes that were not given hydrogen peroxide (Figure 2(c)).

T3 does not improve survival in embryonic mouse astrocytes lacking HADHA after oxygen glucose deprivation To further test the contribution of FAO to the ability of astrocytes to survive, we isolated embryonic astrocytes from WT mice, or from mice lacking hydroxyacylCoA dehydrogenase/3-ketoacyl-CoA thiolase/enoylCoA hydratase alpha (HADHA). HADHA is a protein

Sayre et al.

7

Figure 2. T3 prevents loss of human astrocyte viability in response to stress. Primary human astrocytes were pretreated with the indicated concentration of T3 for 15 min, and then were subjected to H2O2 mediated killing to achieve 50% viability (100 mM, 15 min). Viability was assessed by measuring the percentage of cells that retained the green vital dye Calcein compared to total numbers of Hoescht-labeled cells. (a) Representative image of astrocytes treated with vehicle and H2O2. Viable cells are labeled with both Calcein (green) and Hoescht (blue), while dying cells are only labeled with Hoescht (white arrows). (b) Representative image of astrocytes pretreated with T3 (1 nM) 15 min before H2O2 treatment. (c) Quantification from triplicate independent experiments showing mean  SEM (n ¼ 3–5) of the percentage of viable astrocytes after treatment with H2O2. Letters (a, b, c, d) represent statistical groupings, in which groups labeled by unique letters are significantly different (p < 0.05) from other letter-labeled groups after one-way ANOVA followed by Tukey’s HSD.

complex subunit that is responsible for mitochondrial long-chain fatty acid oxidation.29 We previously showed that cultured fibroblasts null for HADHA failed to upregulate FAO or ATP after T3 treatment.24 To test whether the T3-mediated effect on astrocyte survival depended on FAO through HADHA, astrocytes were subjected to oxygen-glucose deprivation (OGD). Similar to human astrocytes treated with hydrogen peroxide, WT mouse astrocytes subjected to OGD lose about 50% viability compared to control non-OGD astrocytes. T3 treatment similarly increased the viability in WT astrocytes after OGD (Figure 3(a), gray bars). In contrast, in HADHA null astrocytes, T3 failed to improve viability compared to vehicle-treated HADHA-null astrocytes (Figure 3(a), black bars). As an additional test for the role of FAO, embryonic astrocytes were pretreated with etomoxir to inhibit transport of fatty acids into mitochondria. Pretreatment with etomoxir eliminated any T3mediated stimulation of viability in WT astrocytes (Figure 3(b), gray bars). Notably, etomoxir pretreatment did not affect the viability of HADHA-null astrocytes (Figure 3(b), black bars). Altogether, these results suggest that T3-mediated improvements in astrocytes viability are dependent on FAO in vitro.

HADHA colocalizes with GFAP in the cortex and hippocampus In vitro results suggest that HADHA plays an important role in astrocyte upregulation of FAO in response to T3. To determine what amount of brain HADHA is expressed specifically by astrocytes, we performed

Figure 3. T3-mediated astrocyte protection is dependent on FAO after oxygen glucose deprivation. (a) Embryonic mouse astrocytes from wild-type (WT, gray bars) or HADHA null mice (black bars) were pretreated with T3 (1 nM) for 15 min, and then were subjected to 2 h of OGD. Viability was determined via Calcein retention, and results are expressed as a percentage of cells that were not subjected to OGD. (b) Embryonic mouse astrocytes were pretreated with etomoxir (100 mM, 1 h) and then subjected to the same experiment as in part (a). Results shown are mean  SEM of triplicate independent experiments (n ¼ 16) with astrocytes from at least two separate embryos. Letters (a, b, c) represent statistical groupings, in which groups labeled by unique letters are significantly different (p < 0.05) from other letter-labeled groups after one-way ANOVA followed by Tukey’s HSD.

colocalization analysis of HADHA in the cortex and hippocampus to glial-fibrillary acidic protein (GFAP), which is expressed predominantly by astrocytes (Figure 4(a) to (f)). We found 95%  2% of HADHA colocalized to GFAP-positive cells in the cortex (Figure 4(g)), and 93%  3% colocalized in the hippocampus (not

8

Journal of Cerebral Blood Flow & Metabolism

Figure 4. The FAO protein complex HADHA is expressed primarily in astrocytes in the cortex and hippocampus. (a) Representative image of HADHA immunohistochemistry (green) in the mouse cortex. (b) Representative image of GFAP immunohistochemistry (red) in mouse cortex. (c) Panels from A and B have been merged, along with a blue DAPI channel to show nuclei. (d–f) Higher resolution images were acquired of the area outlined with a white box in panels (a–c) of HADHA immunolabeling (d, green), GFAP immunolabeling (e, red), and the color merged image (f). (g) Colocalization analysis was performed as detailed in supplementary methods from cortical regions of n ¼ 3 mice, with 3–4 slices per mouse and 10 imaging planes/slice. Shown are the number of images that had the given percentage of overlap between HADHA and GFAP. The red arrow shows the mean 95%  2% of HADHA immunohistochemistry co-localized to GFAP positive cells in the cortex. Not shown is a similar histogram for the hippocampus, in which the co-localization was 93%  3%.

shown). Altogether, these results suggest that any manipulation of HADHA activity in the brain will primarily affect astrocyte response.

In vivo T3 treatment lessens stroke damage after focal ischemia in mice Others have found that T3 decreases the damage after stroke.21,22,30 In concordance with previous reports, we

found that T3 treatment 30 min after focal ischemia (25 mg/kg) decreased the volume of stroke damage in mice by approximately half when measured 24 h poststroke (Figure 5). To test whether the T3-mediated protection from stroke damage was dependent on thyroid hormone receptor and was not a non-specific response, mice lacking both the a isoform and the b isoform of the thyroid hormone receptor were subjected to focal ischemia (Figure S2, Supplementary material).

Sayre et al.

9

Figure 5. T3 reduces stroke lesion size, while inhibition of FAO increases lesion size. Wild-type mice were subjected to photothrombosis using Rose-Bengal dye, and lesion size was quantified 24 h later using TTC staining. Mice treated with etomoxir (30 mg/kg) were given the drug at the same time as onset of photothrombosis. (a) Representative TTC-stained brain sections of mice treated with vehicle or T3 (25 mg/kg) at 30 min post-stroke. (b) Lesion volume was quantified using Image-J in control and etomoxir-treated mice given either vehicle (gray bars) or T3 (black bars) at 30 min after stroke. Results shown are mean  SEM of quantification of lesions from n ¼ 6–8 mice/treatment group. *p < 0.05, **p < 0.01.

Consistent to previous reports,22 we found that T3 treatment failed to decrease stroke volume at 24 h in thyroid hormone receptor null mice, indicating that the effect of T3 was dependent on receptor interaction.

Inhibition of FAO in mice increases stroke damage after focal ischemia

post stroke. Twenty-four hours later, we found that stroke volume was decreased compared to etomoxironly treated mice, but not significantly different than control, untreated stroke lesions (Figure 5). Altogether, these results appear to suggest that in vivo, T3 exerts its neuroprotective effect only partially through FAO.

Mice with a single copy of HADHA fail to decrease lesion size after T3

Because mice lacking HADHA do not survive past birth,27 we used a pharmacological approach to determine whether T3-mediated reduction in stroke volume was dependent on FAO. At the initiation of photothrombosis, mice were treated with etomoxir (30 mg/kg) to inhibit FAO in the body. During stroke, a breakdown of the blood brain barrier allows penetration of etomoxir into the brain, where it inhibits FAO. Mice treated with etomoxir had significantly greater stroke volume compared to control treated mice (Figure 5). Moreover, virtually every mouse subjected to both focal ischemia and to etomoxir treatment revived normally, but by 24 h post-stroke had become quiet and moribund. Importantly, mice treated with etomoxir only did not become moribund after sham surgery (not shown). Altogether, this indicates that etomoxir treatment combined with focal ischemia significantly worsened stroke damage and complications.

To better determine the contribution of the astrocyteexpressed HADHA to T3-mediated stimulation of FAO and neuroprotection after stroke, we utilized mice expressing a single copy of HADHA. HADHA null mice do not survive into adulthood,23 and floxed versions of HADHA do not yet exist in mouse models. We hypothesized that mice expressing only a single copy of HADHA would fail to efficiently upregulate FAO in response to T3, thereby causing T3 to have a reduced ability to protect after stroke. Compared to WT littermates, HADHA þ/ mice exhibited similar stroke volume (Figure 6(a) and (b)). When HADHA þ/ mice were treated with T3, lesion size did not significantly decrease, (Figure 6(a) and (b)). We concluded from this result that T3-neuroprotective efficacy is dependent on the presence of HADHA.

T3-treatment decreased stroke damage even when FAO is inhibited

Inhibition of mitochondrial energy production ablates neuroprotective efficacy of T3

In vitro experiments indicated that T3-mediated astrocyte protection was dependent on FAO. To test whether the same was true in vivo, mice were treated with etomoxir at onset of focal ischemia, and then were given T3 at 30 min

Because FAO can increase mitochondrial ATP production in multiple cell types, we next tested whether astrocyte-specific inhibition of mitochondrial energy production would alter the ability of T3 to protect

10

Journal of Cerebral Blood Flow & Metabolism

Figure 6. Loss of HADHA expression and inhibition of astrocyte mitochondrial energy production ablate T3-mediated neuroprotection after stroke. HADHA þ/ mice and wild-type littermates were subjected to photothrombosis using Rose-Bengal dye, and lesion size was quantified at 24 h. (a) Representative TTC-stained brain sections from mice treated with vehicle or T3 (25 mg/kg) at 30 min post-stroke. (b) Lesion volume was quantified using Image-J in mice treated with either vehicle (gray bars) or T3 (black bars) at 30 min after stroke. (c) Mice were pretreated with fluoroacetate (10 mm, 100 mL) immediately before photothrombosis. Representative TTC-stained brain sections are shown in vehicle or T3-treated mice. (d) Lesion volume was quantified in mice pretreated with fluoroacetate (Fla) and compared to vehicle-treated mice. Results shown are mean  SEM of quantification of lesions from n ¼ 6–8 mice/treatment group. *p < 0.05, **p < 0.01.

after ischemia. To do this, mice were pretreated with fluoroacetate (10 mm, 0.1 mL) prior to induction of photothrombotic stroke. Fluoroacetate is preferentially transported into astrocytes via the action of monocarboxylic acid transporter 1 (MCT1), where it acts as a toxin to the tricarboxylic acid cycle and significantly lowers the ability of astrocytes to make ATP.31 We found that pretreatment with low concentrations of fluoroacetate did not affect thrombotic lesion size compared to vehicle treated mice (Figure 6(c) and (d)). However, in the presence of this toxin, T3 treatment was unable to reduce stroke lesion size (Figure 6(c) and (d)). This result indicates that T3 neuroprotective efficacy is dependent on the ability of astrocyte mitochondria to produce ATP.

HADHA re-organizes toward astrocyte somas in the penumbra after stroke Our evidence indicates that HADHA activity is increased by T3 treatments after stroke to stimulate ATP production. To test whether T3 causes HADHA to increase expression in vivo, brain tissue was harvested

24 h after middle cerebral artery occlusion (MCAO), and total HADHA levels were measured in the cerebral hemispheres. We did not find any post-stroke change in global expression of HADHA in either ipsilateral or contralateral hemispheres in treated mice (Figure S3, Supplementary material). We next tested whether HADHA localization was affected in mice after photothrombosis. We found that in normal tissue (>500 mm from the penumbra), HADHA localization could be observed throughout the soma and within fine astrocytic processes (Figure 7). Interestingly, HADHA localization became significantly more polarized, and concentrated into the cell bodies of astrocytes near the penumbra (within 500 mm of necrotic core) of vehicle-treated mice, (Figure 7(a) to (c) and (g)). In contrast, mice treated with T3 did not exhibit this relocalization of HADHA to astrocyte somas (Figure 7(d) to (f) and (g)).

Discussion Data presented in this paper demonstrate FAO, and not just glucose metabolism, plays an important role

Sayre et al.

11

Figure 7. HADHA reorganizes into astrocyte cell bodies within the stroke penumbra. WT mice were subjected to photothrombosis and were then treated with vehicle or T3 (25 mg/kg) at 30 min post-stroke. Twenty-four hours later, brain was harvested for immunohistochemistry of HADHA and the astrocyte marker GFAP. Tile scans were obtained of a 4–8 mm2 region near the necrotic core to obtain images of the penumbra and of normal brain parenchyma. (a) Representative tiled image of HADHA in vehicle-treated mice. (b) Representative image of GFAP in vehicle-treated mice. (c) Images from (a) and (b) were merged together with a DAPI channel. The dotted line outlines the edge of the necrotic core. Insets show representative astrocytes that have diffuse, non-polarized localization of HADHA (on left) or polarized, cell-body centered localization of HADHA (on right near necrotic core). (d) Representative tile-scanned image of HADHA in T3-treated mice. (e) Representative imaged of GFAP in T3-treated mice. (f) Images from (d) and (e) were merged together with a DAPI channel. The dotted line outlines the edge of the necrotic core. Insets show individual astrocytes in the normal parenchyma (left) and in the penumbra (right). (g) Quantification of HADHA distribution as a measurement of polarity. The mean intensity of HADHA in the normal tissue region was determined for each tile-scanned image stack using ImageJ. The percentage of pixels at or above the mean intensity were then determined in the normal parenchyma (>500 mm away from penumbra) and in the penumbra (>500 mm away from core). Polarity is expressed as a ratio of percentage area in normal tissue: percentage area in penumbra for vehicle (gray) and T3-treated (black) tissue. Results are mean  SEM of 2–3 tile scanned images/mouse (16–32 panels), n ¼ 3 mice/treatment group. *p < 0.05.

12 in mediating neuroprotection after a cerebral ischemic stroke. Moreover, our results support the hypothesis that thyroid hormones, which have previously been reported to be neuroprotective,21,22,30 extend neuroprotection in a manner that has some dependence on FAO. This finding was clear in vitro, where mouse astrocytes lacking HADHA did not exhibit increased cell survival after OGD in the presence of T3. Additionally, mice heterozygous for HADHA were unresponsive to T3-treatments. Finally, pharmacological inhibition of astrocyte-specific mitochondrial metabolism completely blocked T3-mediated neuroprotection. In vivo, we did observe a significant increase in the lesion sizes of mice that were pharmacologically injected with etomoxir. These data also suggested that FAO metabolism played a major protective role after stroke. This is a particularly important finding given that a reduction in blood flow is expected to limit the supply of glucose to the brain. A local supply of fatty acids would be rapidly available to aid mitochondrial metabolism under conditions of reduced circulation. Assuming that increased FAO metabolism increases ATP production, as we previously reported for fibroblasts and demonstrated above for cultured astrocytes, the drop in energy production would be expected to increase lesion sizes after strokes. A complication of treating with etomoxir in vivo is that we were not able to differentiate between the effects of FAO inhibition in the brain versus the liver. It is entirely possible that inhibition of ATP production in the liver accounts for some or much of the effects observed in brain infarct size  the liver produces glucose via gluconeogenesis, and inhibition of FAO would be expected to strain the ability of the liver to produce glucose. A lowered release of glucose from the liver would cause even less glucose to reach the vulnerable portions of the stroke lesions  a region termed the ‘‘penumbra’’ which can expand as a lesion grows.32 Another relevant observation first reported by Bing et al. is that liver damage occurs in a middle cerebral artery occlusion rat model.33 Consequently, etomoxir treatments might be expected to exacerbate any complications with liver function, and further increase the morbidity associated with stroke. Despite the worsened lesion sizes with etomoxir injections, if the effect of T3 on stroke damage was completely dependent on stimulation of FAO, then inhibition of FAO would be expected to ablate any T3-dependent protective effects. T3 treatment in the presence of etomoxir resulted in stroke volumes comparable to non-treated control mice. T3 treatments did reverse the larger lesions observed with etomoxir. However, when we tested mice heterozygous for HADHA, we observed a complete loss of T3-mediated neuroprotection, clearly demonstrating a dependence

Journal of Cerebral Blood Flow & Metabolism on HADHA activity. Because HADHA is primarily expressed in astrocytes, we postulated that astrocytespecific FAO played the key role in limiting stroke lesions size, and was a major part of the mechanism by which T3 exerts neuroprotection. This conclusion is further supported by our finding that pretreatment with low concentrations of toxin fluoroacetate, which specifically targets astrocyte mitochondrial metabolism, ablated the ability of T3 to protect from stroke damage. An additionally interesting observation came from localization analysis of HADHA in tissue near the necrotic lesion. We found that HADHA was re-organized more toward the cell body of astrocytes, whereas in properly perfused normal tissue, HADHA was observed equally in the fine processes around astrocytes and the soma. Notably, this reorganization of HADHA was not observed in mice treated with T3. We cannot distinguish between a mechanism in which T3 targets HADHA organization within astrocytes, or one that is re-organized as a consequence of astrocyte activation, as gliosis appears decreased around the necrotic lesion of T3-treated astrocytes,. As noted above, T3 treatments could also be affecting liver function. In addition, astrocytes are not the only cells that express thyroid hormone receptors in the brain.34,35 Consequently, we cannot rule out additional T3-mediated mechanisms of action important for neuroprotective efficacy, even though astrocytes account for the vast majority of brain FAO. For example, Hiroi et al. reported that Akt and endothelial nitric oxide synthase (eNOS) became activated in endothelial cells after thyroid hormone treatment and also played a role in T3 neuroprotection after stroke.22 Given that activation of Akt also plays a major role in upregulating the import of glucose into cells, these data have notable importance to our findings.36 Together, these functions of thyroid hormone would complement our observed upregulation of FAO and together help neural cells generate ATP. At this time, no ideal conditional model exists to test the exact contribution of astrocyte FAO to stroke damage  mice null for HADHA cannot oxidize long chain fatty acids and fail to survive after birth.27 Nevertheless, our results strongly support a model in which FAO by astrocytes increases the total available energy after stroke. With more carbon sources of energy, astrocytes are better equipped to protect neurons, thereby reducing the amount of damage after stroke. Previous data generated in our lab demonstrated that T3 stimulation of FAO was driven by acute, non-genomic effects.24 It is also well established that whole animal treatments with T3 are expected to exert long-term transcriptional effects, as the hormone remains in circulation for extended periods of time after injection. Consequently, some of the protective effects

Sayre et al. we observed with T3 treatments, e.g. when FAO was pharmacologically inhibited, could also involve genomic effects of T3. In support of this, Genovese and colleagues showed that thyroid hormone treatment lowered reactive gliosis, inflammation, and cell death after middle cerebral artery occlusion, and also helped to restore neurotrophin expression.21 Lourbopoulos et al. showed that thyroid hormone receptor expression is reorganized after cerebral ischemia,37 supporting an expanded role for thyroid hormone in the post-ischemic environment. Whatever the relative role of these underlying processes, such effects are downstream of the initial ischemic event, and it is not clear whether thyroid hormone treatments are helping to drive the recovery and repair process, or preventing the spread of the ischemic core. Also notable is that we did not choose to measure stroke damage beyond 24 hours poststroke. We previously showed that, in our hands, stroke damage leveled in vehicle-treated mice after 24 h.4 As such, we presume that reduced damage at 24 h will similarly result in smaller lesion over the long term. What we cannot account for with this study, however, are the potential long-term effects that FAO and T3 could have in repair and recovery. Our data do not reveal whether the mechanism of FAO stimulation by T3 treatments has long-term protective effects, since we did not measure stroke damage beyond 24 h post-ischemia. We do note that others have reported long-term benefit of T3 treatments after stroke.21,22,30 Overall, our data clearly demonstrate an acute protective role for T3 after stroke. We note that although T3 treatments for hypothyroidism are FDA approved, increased serum thyroid hormone levels are associated with significant potential cardiac side effects.38 Thus, the utility of thyroid hormone by itself as a stroke treatment is not clear. However, we envision that a potential treatment could eventually involve stimulation of the astrocyte’s ability to utilize fatty acids for energy  such a treatment would be particularly attractive for an environment with limited glucose availability such as after stroke. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research in this paper was supported by NIH AG007218 (JDL) and AHA 13Post16820029 (NLS).

Acknowledgements We gratefully acknowledge Dr Ramaswamy Sharma, Dr Linda McManus, and Dr Mark Shapiro for critical insight and discussions related to this project. We thank Mr Shane Sprague of the Department of Neurosurgery in his assistance with the MCAO mouse model.

13 Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions NLS was responsible for writing the manuscript, editing figures, and performing experiments for Figures 1(a) to (c), 2, 4, 5, 7, S2. DH performed experiments for Figures 1(d), 2, and S1 and helped in editing the manuscript. MS performed experiments for Figures 3, 6, 7, and S3 and helped in editing the manuscript. SYC and XZ provided thyroid receptor null mice, which NLS and MS genotyped and used for experiments. SYC also helped in editing the manuscript. JL helped in editing the manuscript.

Supplementary material Supplementary material for this paper can be found at http:// jcbfm.sagepub.com/content/by/supplemental-data

References 1. Marsh JD and Keyrouz SG. Stroke prevention and treatment. J Am Coll Cardiol 2010; 56: 683–691. 2. Mozaffarian D, Benjamin EJ, Go AS, et al. Heart disease and stroke statistics—2015 update: a report from the American Heart Association. Circulation 2014; e30: e29–e322. 3. Zhao Y and Rempe DA. Targeting astrocytes for stroke therapy. Neurotherapeutics 2010; 7: 439–451. 4. Zheng W, Watts LT, Holstein DM, et al. Purinergic receptor stimulation reduces cytotoxic edema and brain infarcts in mouse induced by photothrombosis by energizing glial mitochondria. PLoS One 2010; 5: e14401. 5. Wu J, Holstein JD, Upadhyay G, et al. Purinergic receptor-stimulated IP3-mediated Ca2þ release enhances neuroprotection by increasing astrocyte mitochondrial metabolism during aging. J Neurosci 2007; 27: 6510–6520. 6. Brown AM and Ransom BR. Astrocyte glycogen and brain energy metabolism. Glia 2007; 55: 1263–1271. 7. Watts L and Lechleither J. The impact of astrocyte mitochondrial metabolism on neuroprotection during aging. In: Parpura V and Haydon P (eds) Astrocyte in (patho) physiology of the nervous system. Boston, MA: Springer, 2008, pp.569–590. 8. Mattson MP and Rychlik B. Glia protect hippocampal neurons against excitatory amino acid-induced degeneration: involvement of fibroblast growth factor. Int J Dev Neurosci 1990; 8: 399–415. 9. Edmond J, Robbins RA, Bergstrom JD, et al. Capacity for substrate utilization in oxidative metabolism by neurons, astrocytes, and oligodendrocytes from developing brain in primary culture. J Neurosci Res 1987; 18: 551–561. 10. Edmond J. Energy metabolism in developing brain cells. Can J Physiol Pharmacol 1992; 70: S118–S129. 11. Ebert D, Haller RG and Walton ME. Energy contribution of octanoate to intact rat brain metabolism measured

14

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Journal of Cerebral Blood Flow & Metabolism by 13C nuclear magnetic resonance spectroscopy. J Neurosci 2003; 23: 5928–5935. Waniewski RA and Martin DL. Preferential utilization of acetate by astrocytes is attributable to transport. J Neurosci 1998; 18: 5225–5233. Escartin C, Pierre K, Colin A, et al. Activation of astrocytes by CNTF induces metabolic plasticity and increases resistance to metabolic insults. J Neurosci 2007; 27: 7094–104. Heather LC, Cole MA, Atherton HJ, et al. Adenosine monophosphate-activated protein kinase activation, substrate transporter translocation, and metabolism in the contracting hyperthyroid rat heart. Endocrinology 2010; 151: 422–431. Irrcher I, Walkinshaw DR, Sheehan TE, et al. Thyroid hormone (T3) rapidly activates p38 and AMPK in skeletal muscle in vivo. J Appl Physiol 2008; 104: 178–185. de Lange P, Senese R, Cioffi F, et al. Rapid activation by 3,5,3’-L-triiodothyronine of adenosine 5’-monophosphate-activated protein kinase/acetyl-coenzyme a carboxylase and akt/protein kinase B signaling pathways: relation to changes in fuel metabolism and myosin heavy-chain protein content in rat gastrocnemius muscle in vivo. Endocrinology 2008; 149: 6462–6470. Lombardi A, de Lange P, Silvestri E, et al. 3,5-Diiodo-lthyronine rapidly enhances mitochondrial fatty acid oxidation rate and thermogenesis in rat skeletal muscle: AMP-activated protein kinase involvement. Am Journal of Physiol Endocrinol Metab 2009; 296: E497–E502. Lanni A, Moreno M, Lombardi A, et al. 3,5-Diiodo-Lthyronine powerfully reduces adiposity in rats by increasing the burning of fats. FASEB J 2005; 19: 1552–1554. Chocron ES, Sayre NL, Holstein D, et al. The trifunctional protein mediates thyroid hormone receptordependent stimulation of mitochondria metabolism. Molecular Endocrinol 2012; 26: 1117–1128. Saelim N, John LM, Wu J, et al. Nontranscriptional modulation of intracellular Ca2þ signaling by ligand stimulated thyroid hormone receptor. J Cell Biol 2004; 167: 915–924. Genovese T, Impellizzeri D, Ahmad A, et al. Post-ischaemic thyroid hormone treatment in a rat model of acute stroke. Brain Res 2013; 1513: 92–102. Hiroi Y, Kim HH, Ying H, et al. Rapid nongenomic actions of thyroid hormone. Proc Natl Acad Sci U S A 2006; 103: 14104–14109. Ibdah JA, Paul H, Zhao Y, et al. Lack of mitochondrial trifunctional protein in mice causes neonatal hypoglycemia and sudden death. J Clin Invest 2001; 107: 1403–1409. Chocron ES, Sayre NL, Holstein D, et al. The trifunctional protein mediates thyroid hormone receptordependent stimulation of mitochondria metabolism. Mol Endocrinol 2012; 26: 1117–1128.

25. Lin DT, Wu J, Holstein D, et al. Ca2þ signaling, mitochondria and sensitivity to oxidative stress in aging astrocytes. Neurobiol Aging 2007; 28: 99–111. 26. Ren Z, Iliff JJ, Yang L, et al. ’Hit & Run’ model of closed-skull traumatic brain injury (TBI) reveals complex patterns of post-traumatic AQP4 dysregulation. J Cerebral Blood Flow Metab: Official journal of the International Society of Cerebral Blood Flow and Metabolism 2013; 33: 834–845. 27. Ibdah JA, Paul H, Zhao Y, et al. Lack of mitochondrial trifunctional protein in mice causes neonatal hypoglycemia and sudden death. J Clin Invest 2001; 107: 1403–1409. 28. Watts LT, Sprague S, Zheng W, et al. Purinergic 2Y1 Receptor stimulation decreases cerebral edema and reactive gliosis in a Traumatic Brain Injury Model. J Neurotrauma 2012; 30: 55–66. 29. Angdisen J, Moore VD, Cline JM, et al. Mitochondrial trifunctional protein defects: molecular basis and novel therapeutic approaches. Curr Drug Target Immune Endocr Metabol Disord 2005; 5: 27–40. 30. Mdzinarishvili A, Sutariya V, Talasila P, et al. Engineering triiodothyronine (T3) nanoparticle for use in ischemic brain stroke. Drug Deliv Transl Res 2013; 3: 309–317. 31. Fonnum F, Johnsen A and Hassel B. Use of fluorocitrate and fluoroacetate in the study of brain metabolism. Glia 1997; 21: 106–113. 32. Sayre NL, Chen Y, Sifuentes M, et al. Purinergic receptor stimulation decreases ischemic brain damage by energizing astrocyte mitochondria. Adv Neurobiol 2014; 11: 121–150. 33. Bing Y, Zhu S, Jiang K, et al. Reduction of thyroid hormones triggers down-regulation of hepatic CYP2B through nuclear receptors CAR and TR in a rat model of acute stroke. Biochem Pharmacol 2014; 87: 636–649. 34. Wallis K, Dudazy S, van Hogerlinden M, et al. The thyroid hormone receptor alpha1 protein is expressed in embryonic postmitotic neurons and persists in most adult neurons. Mol Endocrinol 2010; 24: 1904–1916. 35. Bernal J. Thyroid hormone receptors in brain development and function. Nat Clin Pract Endocrinol Metab 2007; 3: 249–259. 36. Schultze SM, Hemmings BA, Niessen M, et al. PI3K/ AKT, MAPK and AMPK signalling: protein kinases in glucose homeostasis. Expert Rev Mol Med 2012; 14: e1. 37. Lourbopoulos A, Mourouzis I, Karapanayiotides T, et al. Changes in thyroid hormone receptors after permanent cerebral ischemia in male rats. J Mol Neurosci 2014; 54: 78–91. 38. Vargas-Uricoechea H, Bonelo-Perdomo A and SierraTorres CH. Effects of thyroid hormones on the heart. Clin Investig Arterioscler 2014; 26: 296–309.

Stimulation of astrocyte fatty acid oxidation by thyroid hormone is protective against ischemic stroke-induced damage.

We previously demonstrated that stimulation of astrocyte mitochondrial ATP production via P2Y1 receptor agonists was neuroprotective after cerebral is...
2MB Sizes 0 Downloads 9 Views