molecular alterations in the hippocampus after ... · molecular alterations in the hippocampus...

ORIGINAL ARTICLE

Molecular alterations in the hippocampus after experimentalsubarachnoid hemorrhageSang Myung Han1,4, Hoyee Wan1,4, Gen Kudo2, Warren D Foltz2,3, Douglass C Vines2,3, David E Green2, Tommaso Zoerle1,Asma Tariq1, Shakira Brathwaite1, Josephine D’Abbondanza1, Jinglu Ai1 and R Loch Macdonald1

Patients with aneurysmal subarachnoid hemorrhage (SAH) frequently have deficits in learning and memory that may or may not beassociated with detectable brain lesions. We examined mediators of long-term potentiation after SAH in rats to determine whatprocesses might be involved. There was a reduction in synapses in the dendritic layer of the CA1 region on transmission electronmicroscopy as well as reduced colocalization of microtubule-associated protein 2 (MAP2) and synaptophysin. Immunohisto-chemistry showed reduced staining for GluR1 and calmodulin kinase 2 and increased staining for GluR2. Myelin basic proteinstaining was decreased as well. There was no detectable neuronal injury by Fluoro-Jade B, TUNEL, or activated caspase-3 staining.Vasospasm of the large arteries of the circle of Willis was mild to moderate in severity. Nitric oxide was increased and superoxideanion radical was decreased in hippocampal tissue. Cerebral blood flow, measured by magnetic resonance imaging, and cerebralglucose metabolism, measured by positron emission tomography, were no different in SAH compared with control groups. Theresults suggest that the etiology of loss of LTP after SAH is not cerebral ischemia but may be mediated by effects of subarachnoidblood such as oxidative stress and inflammation.

Journal of Cerebral Blood Flow & Metabolism advance online publication, 25 September 2013; doi:10.1038/jcbfm.2013.170

Keywords: cognitive dysfunction; LTP; rats; subarachnoid hemorrhage; synaptic plasticity

INTRODUCTIONSurvivors of aneurysmal subarachnoid hemorrhage (SAH) oftensuffer memory and cognitive deficits that affect their quality oflife.1 Brain magnetic resonance imaging (MRI) of patients after SAHshows that there may be atrophy in the medial temporal lobes andthat this may correlate with neurocognitive deficits in the absenceof any focal structural lesions.2,3 The hippocampal formation in themedial temporal lobe has been postulated to be involved incognitive functions such as learning and memory. Long-termpotentiation (LTP), a form of synaptic plasticity, is believed to be acellular substrate of memory and learning.4 We previouslydemonstrated cognitive dysfunction5 and loss of LTP in thehippocampal synapses after SAH in rats.6 We found SAHcaused significant vasospasm of the middle cerebral artery anddecreased postsynaptic responses and loss of LTP at CA3–CA1synapses. However, the cause of the loss of LTP was not workedout.

Here we use a prechiasmatic rat model of SAH5,6 to test thehypothesis that alterations of proteins that participate in thesubcellular maintenance of LTP contribute to loss of LTP afterexperimental SAH. Positron emission tomography (PET), MRI, andimmunohistochemistry were used to investigate neuronal celldeath, cerebral blood flow (CBF), and brain metabolism to

determine if ischemia contributed to the loss of LTP. There areseveral different models of SAH in rats, including the endovascularperforation model7 and the cisterna magna injection models.8–10

The prechiasmatic SAH model was used, as this deposits bloodaround the anterior circle of Willis where most aneurysmal SAHoccurs in humans. The model also has less variability than theendovascular perforation model.11

MATERIALS AND METHODSAnimalsAll procedures performed on animals were approved by the institutionalAnimal Care Committee and complied with regulations of CanadianCouncil on Animal Care. Male Sprague–Dawley rats (350 to 450 g, n¼ 80)were randomly allocated to one of four groups: naı̈ve control, sham-operated control, saline (0.9% NaCl)-injected control, and blood-injectedSAH. Animals were housed in temperature and humidity controlledenvironments with a 12-hour light–dark cycle.

Subarachnoid HemorrhageThe SAH model was created as previously reported.5 We injected 300mLnon-heparinized autologous blood into the prechiasmatic cistern of rats.Rats were anesthetized with isofluorane by inhalation (1% to 2%) and were

1Division of Neurosurgery, St Michael’s Hospital, Labatt Family Centre of Excellence in Brain Injury and Trauma Research, Keenan Research Centre of the Li Ka Shing KnowledgeInstitute of St Michael’s Hospital, Department of Surgery, University of Toronto, Toronto, Ontario, Canada; 2STTARR Innovation Centre, Department of Radiation Oncology,Princess Margaret Hospital, Toronto, Ontario, Canada and 3Department of Radiation Oncology, University of Toronto, Toronto, Ontario, Canada. Correspondence:Dr RL Macdonald, Division of Neurosurgery, St Michael’s Hospital, Labatt Family Centre of Excellence in Brain Injury and Trauma Research, Keenan Research Centre of the Li KaShing Knowledge Institute of St Michael’s Hospital, Department of Surgery, University of Toronto, 30 Bond Street, Toronto, Ontario M5B 1W8, Canada.E-mail: [email protected] study was supported by an operating grant to Dr Jinglu Ai as the North Shore University Hospital, Brain Aneurysm Center Chair of Research from the Brain AneurysmFoundation with Dr RLoch Macdonald as the mentor. RL Macdonald receives grant support from the Physicians Services Incorporated Foundation, Brain Aneurysm Foundation,Canadian Stroke Network and the Heart and Stroke Foundation of Ontario. RL Macdonald is a consultant for Actelion Pharmaceuticals and Chief Scientific Officer of EdgeTherapeutics, Inc.4These authors contribute equally to this work.Received 29 May 2013; revised 6 August 2013; accepted 27 August 2013

Journal of Cerebral Blood Flow & Metabolism (2013), 1–10& 2013 ISCBFM All rights reserved 0271-678X/13 $32.00

www.jcbfm.com

http://dx.doi.org/10.1038/jcbfm.2013.170

mailto:[email protected]

http://www.jcbfm.com

Jinglu Ai

Highlight

positioned in a stereotactic frame (David Kopf Instruments, Tujunga, CA,USA). The scalp was prepared sterilely with povidone iodine. Subarachnoidhemorrhage was produced by injection of 300mL blood withdrawn fromthe tail artery, through a 27-gauge spinal needle with a rounded tip andside hole, which was inserted into the prechiasmatic cistern. The injectionswere made over 20 seconds using a syringe pump (Harvard Apparatus,Holliston, MA, USA). The prechiasmatic injection model of SAH wasmodified from the model initially described by Prunell et al9,11 Theinjection speed was similar to what was used in their original studies, inwhich 200mL of blood was injected manually over 12 seconds. Theinjection speed was initially set aiming to maintain the intracranialpressure at the same level as the mean arterial blood pressure during theinjection.9,11 A syringe pump was used to keep the injection timeconsistent. The needle was angled 401 caudally in the sagittal plane, andplaced through a burr hole placed just off the midline, 7.5 mm anterior tothe bregma. Correct location of the needle in the subarachnoid space wasobtained by advancing the needle until it touched the skull base and thenwithdrawing it 0.5 mm. CBF was monitored with a laser Doppler flow probe(Transonics model BLF21, type N flow probe, Ithaca, NY, USA) that wasplaced over an area of the left side of the skull just posterior to the bregmaline. Care was taken to avoid large vessels. The tail artery was catheterizedwith PE10 tubing that was sutured in place and connected to a pressuretransducer (World Precision Instruments, Sarasota, FL, USA) with rigidtubing filled with 0.9% NaCl for blood pressure measurement. Bodytemperature was maintained and monitored with a heating pad(Homeothermic systems, Harvard Apparatus, Holliston, MA, USA) andrectal temperature probe at 371C. Baseline CBF, blood pressure, and bodytemperature were recorded for 15 minutes. Saline-control animalsunderwent injection of 300mL of 0.9% NaCl. After completing theseprocedures, 5 mL of 0.9% NaCl was injected subcutaneously to preventdehydration and the rats were returned to an incubator maintained at26±11C for 48 hours with access to soft food. Buprenorphine (0.05 mg/kg)was given twice daily for 48 hours after surgery. There was no mortality.Saline-injected animals were injected with 300mL saline. Sham-operatedcontrols were subjected to all procedures without injecting anything. Naivecontrols did not undergo any surgical procedures. Rats were killed undergeneral anesthesia induced by administration of ketamine/xylazine (120/10 mg/kg) on day 0, 2, or 6 after surgery.

For biochemical assays, the brains were isolated immediately afterkilling. For histology, animals were perfused through the left cardiacventricle with 0.9% NaCl for 2 minutes, followed by 250 mL of 4%paraformaldehyde in phosphate-buffered saline (PBS). The brains wereremoved and fixed in 4% paraformaldehyde for another 48 hours. Sectionswere obtained from coronal cuts of the brain centered at 5 mm interaural(B� 4 mm bregma). They were processed and embedded in paraffin.Sections 7 mm thick were cut using a microtome (Leica, Wetzlar, Germany).The coronal hippocampal slides were from the vicinity of bregma, � 4 mmarea on a 2-mm block of the brain.

Transmission Electron MicroscopyQuantification of synapses was done using transmission electron micro-scopy (TEM) as described previously.12 Six days after surgery, SAH, orsaline-injected control rats (n¼ 4 per group) were perfused with saline for2 minutes followed by 2% paraformaldehyde/2.5% glutaraldehyde. Thebrains were removed and sectioned in the coronal plane (1 mm) on avibratome (VT1200, Leica Microsystems, Richmond Hill, Canada). Sampleswere taken from the region of the CA1 layer in the hippocampus on the1 mm brain slices using a biopsy punch (1.5 mm diameter, Acuderm, FortLauderdale, FL, USA). The tissue blocks (1.5 mm circular block of 1 mmthickness) were then processed.12 Tissue blocks were fixed with 2%osmium tetroxide/1.5% potassium ferricyanide. They were embedded inepon and 1-mm sections were cut from a block of embedded tissue fromeach rat using an ultramicrotome. The sections were collected on slotted,copper grids and stained with 2% uranyl acetate followed by lead citrate.The slides were examined by TEM (JEM 1011, JEOL, Munchen, Germany).

Double Immunofluorescence StainingTo verify the data from TEM, we carried out double immunofluorescentstaining using synaptophysin (a presynaptic marker) and MAP2 (neuronaland postsynaptic marker). Active caspase-3/NeuN double staining was usedfor detection of neuronal cell death. Immunofluorescence staining wasperformed as previously described.12 Sections were deparaffinized andrehydrated, followed by antigen retrieval in a 961C water bath for25 minutes using Vector antigen unmasking solution (Vector Laboratories,

Burlingame, CA, USA). Sections were permeabilized with 0.3% Triton X-100for 15 minutes and blocked by incubating with 10% normal goat serum in1% BSA for 60 minutes. Primary antibodies were mouse monoclonal anti-NeuN (1:200 clone A60, Millipore, Billerica, MA, USA), rabbit polyclonal anti-human active caspase-3 (1:400, BD Pharmingen, San Diego, CA, USA),mouse monoclonal anti-rat synaptophysin (1:10, Millipore), and rabbitpolyclonal anti-MAP2 (1:200, Abcam, Cambridge, MA, USA). After incubationovernight at 41C, the slides were washed four times in PBS. They were thenincubated at room temperature for 1 hour with Alexa Fluor 488-conjugatedgoat anti-mouse IgG (1:1000, Life Technologies, Burlington, ON, Canada) tovisualize synaptophysin or NeuN and Alexa Fluor 568-conjugated goat anti-rabbit IgG (1:1000, Life Technologies) to visualize MAP2 or caspase-3.Sections were then rinsed three times in PBS. Slides were rinsed two timesin PBS and cover-slipped with anti-fading mounting medium. The stainedsections were examined with a confocal microscope (Olympus BX50confocal microscope, Olympus, Richmond Hill, ON, Canada).

Immunohistochemistry for GluR1/2, CaMK-II, and Myelin BasicProteinFor GluR2 staining, slides were deparaffinized and rehydrated throughxylene and ethanol solutions and antigen was retrieved in a 961C waterbath for 25 minutes using Vector antigen unmasking solution (VectorLaboratories). Endogenous peroxidase activity was quenched using 0.3%hydrogen peroxide in water for 45 minutes. Sections were blocked with10% horse serum in 1% bovine serum albumin in PBS for 40 minutes,incubated with mouse polyclonal anti-rat GluR2 (1:400 in 1% bovine serumalbumin (BSA)/PBS, Millipore) for 90 minutes. They were then washed withPBS and incubated with biotinylated secondary antibody (horse anti-mouse 1:200 in 1% BSA/PBS, Vector) for 30 minutes. Staining was visualizedwith VIP using the VECTASTAIN ABC Kit (Vector) and counterstained with0.5% methyl green. For immunohistochemical staining of GluR1, CaMK-II,and MBP, the above protocol was repeated where samples were blockedwith 10% goat serum in 1% bovine serum albumin, then incubated withthe primary antibodies rabbit polyclonal anti-rat GluR1 (1:200 in 1% BSA/PBS, Millipore), rabbit polyclonal anti-CaMK-II (1:200 in 1% BSA/PBS, SantaCruz, Dallas, TX, USA), and rabbit polyclonal anti-rat MBP (1:500, BosterBiological Technology, Wuhan, China, in 1% BSA/PBS). For the secondaryantibody goat anti-rabbit biotinylated antibody was used (1:200 in 1% BSA/PBS, Vector).

Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labelingand Fluoro-Jade StainingDeoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining wasdone as reported.12,13 Rats injected with kainic acid (10 mg/kg) were usedas positive controls. Apoptosis was assessed using TUNEL in accordancewith the manufacturer’s protocol (DeadEnd Flurometric kit, Promega,Fitchburg, WI, USA). Slides were counterstained with 40 ,6-diamidino-2-phenylindole, washed, and cover-slipped with a water-based mountingmedium, and sealed with nail polish.

Fluoro-Jade B (Histo-Chem, Jefferson, AR, USA) staining was used as amarker for neuronal injury. After deparaffinization, rehydration, andsubsequent incubation with deionized water, the slides were incubatedin 0.06% potassium permanganate (VWR International, Strasbourg, France)for 15 minutes with gentle shaking. Slides were then rinsed in deionizedwater to remove excess permanganate and immersed in 0.001% Fluoro-Jade B working solution (0.1% acetic acid) for 30 minutes. They werewashed and dried in an incubator (601C) for 15 minutes. Sections werecleared in xylene and cover-slipped with a non-aqueous, low fluorescence,styrene based mounting medium (DPX, Sigma-Aldrich, St Louis, MO, USA).Slides were examined with a fluorescence microscope (Leica). Images weretaken with the same parameters for all sections.

Nitric Oxide and Superoxide Anion Radical DetectionThe biochemical assay was performed as described.13 After killing, thebrain was removed and placed on dry ice. A coronal section of the frontalcortex was cut and placed in an Eppendorf tube and frozen at � 811C.Bilateral hippocampal sections were dissected and transferred to � 811Cas well for homogenization at a later date.

Superoxide anion radical (O2- ) and nitric oxide (NO) were detected in

homogenized fresh or deep frozen brain tissue using spectrophoto-metric methods.13 To detect NO, cell-permeable DAF-2DA (fluoro-phore 4,5-diaminofluorescein-2-diacetate, Alexis Biochemicals, Gruenberg,Germany) was used and a chemiluminescence probe, 2-methyl-6-(p-

Mechnisms of loss of LTP after SAHSM Han et al

2

Journal of Cerebral Blood Flow & Metabolism (2013), 1 – 10 & 2013 ISCBFM

methoxyphenyl)-3,7-dihydroimidazo[1,2-]-pyrazin-3-one (MCLA) was usedfor O2

� detection. Ten microlitersmL of homogenized cortical orhippocampal tissues were incubated with either 10 mmol/L DAF-2DA for30 minutes or 4 mmol/L MCLA for 10 minutes in transparent 96-well plates(Fisher, Ottawa, ON, Canada) at room temperature in the dark. DAF-2DAwas excited at 495 nm and emission read at 515 nm in aspectrofluorometer (SpectraMAX-Gemini, Molecular Devices, Sunnyvale,CA, USA). The luminescence signal of MCLA was read directly at 495 nm.Experiments were repeated three times for each sample. MCLA crosses cellmembranes and therefore the O2

� detected is from both intracellular andextracellular sources. Nitric oxide values are expressed as relativefluorescence units. Superoxide anion radical values are expressed asrelative luminescence unit.

ImagingRats underwent MRI and PET imaging before the killing time. The basis forco-registered MRI/PET was the multi-modal Minerve beds (Bioscan,Washington, DC, USA) so that rats could be transported between modalities.While on the Minerve bed, rats were oriented in the prone position withanesthesia delivered by inhalation (1.5% to 2% isoflurane). A pneumaticpillow was placed under the animal to enable respiratory monitoring whilewithin the MRI suite (SA Instruments, Stony Brook, NY, USA).

Position Emission TomographyOn the day of euthanasia, rats were fasted (water access allowed) for atleast 6 hours. Rats were anesthetized with 2% isoflurane and put in asupine position on the surgical table. The tail artery was exposed andcannulated with a 0.75-inch 24-gauge catheter (AngioCath, BD Biosciences,San Jose, CA, USA) to allow arterial blood sampling during the dynamicscan. A lateral tail vein was cannulated with polyethylene (PE10) tubingfilled with heparinized saline for radiotracer injection. The rats were thensecured on the bite bar of a warmed rat Minerve bed.

Rats were then advanced into a Focus 220 microPET system (SiemensMedical Systems, Erlangen, Germany), which had been cross-calibrated tothe dose calibrator by imaging a known source of 18F and determining aquantification calibration factor. Transmission brain scans were acquiredfor 8 minutes using a 57Co point source to allow attenuation correction tobe performed after the dynamic scan. Rats were then injected with 0.25 mLcontaining 75 MBq of 18F-FDG over 3 to 6 seconds and the 60-minutedynamic scan started at the same time as 18F-FDG injection. Multiplearterial blood samples were obtained to enable determination of a wholeblood time activity curve, which then allowed calculation of cerebralglucose metabolic rates (CMRglc). After injection, the syringes and tubingline with residual radiotracer were measured in the dose calibrator todetermine the net injected amount of radiotracer at injection time.

Magnetic Resonance ImagingAfter completion of the PET acquisition, rats were transported to a 7 TeslaBiospec MRI system (Bruker Corporation, Ettlingen, Germany), fitted withthe B-GA12 gradient coil insert, 7.2 cm inner diameter linearly polarizedvolume resonator for radiofrequency transmission, and dedicated flatsurface coil for radiofrequency reception. The flat coil was fixed superior tothe Minerve bed without compression of the rat head.

The rat brain was first visualized as a stack of transverse slices using aT2-weighted rapid acceleration relaxation enhancement technique(echo time¼ 72 milliseconds; repetition time¼ 4,800 milliseconds, 150�150� 1,000mm voxels; 30� 30 mm field-of-view; 16 slices; 2-minute40-second scan time). These images were used for PET registration. Threetransverse slices separated by 3 mm at the level of the corpus callosumwere then selected for CBF measurements based on a flow-alternating-inversion-recovery technique (two-dimensional fat-suppressed single-shotecho planar imaging; echo time¼ 13.4 milliseconds; repetition time¼ 17seconds; 400� 400� 1,000mm voxels; 30� 30 mm field-of-view; 18 inver-sion times ranging from 25 milliseconds to 6,825 milliseconds separatedby 400 milliseconds; 227 kHz effective readout bandwidth, 10-minute12- second scan time). The flow-alternating-inversion-recovery acquisitionquantifies CBF as the difference in T1 relaxation rate between measure-ments with global and slice-selective inversion pulses, as per equation 1.

P ¼ l�ð1=T1;sliceselective� 1=T1; globalÞðml blood=ð100 g tissue�minuteÞÞð1Þ

where P is perfusion and l is the blood–brain partition coefficient for water(60 mL/(100 g*minute)).

Data AnalysisAll data were from animals killed 6 days after surgery, except PET and MRI,which was from day 0, 2, or 6 as labeled in the corresponding figures andtables. All data was expressed as means±s.d.. Comparison of the groupswas analyzed by Student’s t-test or one-way analysis of variance with posthoc Holm–Sidak test using SigmaPlot 12.0. Pp0.05 was consideredsignificant.

For TEM, four squares of the copper grids in the CA1 dendritic area werechosen for taking images. Five images were acquired under � 15,000magnification. Synapses were counted by an investigator who was maskedto the animal group (SB and JD). Unbiased stereological methods forcounting were used, and the final number used was the mean of 20 images.

Double stained sections (MAP2/synaptophysin) were imaged using alaser confocal microscope (Axiovert 200, Carl Zeiss, Hawthorn, PA, USA).Data were analyzed with imaging software (Fiji: an open-source platformfor biologic-image analysis14) that determines the coefficients of thecolocalization of the double staining, and which represents functionalsynapses.15 The quantification of colocalization was expressed as Manders’coefficient, which is the portion of signals above background thresholdfrom both channels.

For GluR1, GluR2, CamK-II, and MBP quantification, masks of positivestained signals were generated using ImageJ (NIH, Bethesda, MD, USA),and the total area of positive staining was quantified and averaged againsttotal neurons in the field. Raw SAH data were further normalized to thatfrom saline controls. Six images of right and left CA1 subfields were takenunder consistent light and exposure conditions at � 400 magnification byan operator masked to the animal group (HW). Images were color-matchedagainst a randomly selected image to ensure a consistent background, andconverted into an 8-bit image. The region of interest (ROI) and thresholdfor positive signal was predefined, and the percent of positive signal ineach ROI was quantified using ImageJ. For CAMK2, GluR1, and GluR2, theamount of positive signal was normalized to the number of methyl green-positive cells in the ROI.

Magnetic resonance imaging/positron emission tomography co-regis-tration used Inveon Research Workplace (Siemens Medical Systems), and

Figure 1. Quantification of synapses using transmission electronmicroscopy (TEM). (A) Representative images of TEM showingsynapses in CA1 area of the hippocampus (red arrow heads). (B) Bargraph showing significantly decreased number of synapses insubarachnoid hemorrhage (SAH) group as compared with salinecontrols (Pp0.01, t-test). Data are mean±s.d. All animals were killed6 days after surgery.


3

& 2013 ISCBFM Journal of Cerebral Blood Flow & Metabolism (2013), 1 – 10

PET ROI placement was guided by the rat brain atlas of Paxinos, et al16

Image analysis of PET activity used three-dimensional-ROIs drawnmanually over five bilateral regions (frontal cortex, parietal cortex,caudate-putamen, thalamus, and hippocampus) by a single observerwho was masked to the animal group (GK).

The Patlak graphical analysis method was used to determine CMRglc,17

based on equation 2:

rCMRglc ¼ KCp=LC ð2ÞIn this equation, Cp is the arterial plasma glucose concentration, LC is the

lumped constant in the method of Sokoloff, et al,18 and K is the constantslope.19 The LC value used was 0.625 for the normoglycemic range: 6.0 to19.7 mmol/L.19 The interval of 30 to 60 minutes was used to estimate K.

Magnetic resonance imaging cerebral blood flow analysis used a custommMtlab script for flow-alternating-inversion-recovery map generation, andMIPAV (NIH) for manual segmentation and histogram analysis of ROIswithin cortex, sub-cortex, and hippocampus.

RESULTSSubarachnoid hemorrhage Reduces the Hippocampal SynapsesThe quantification with TEM showed that the number of synapsesin the dendritic layer of CA1 was significantly decreased in SAHanimals as compared to saline controls (55±4.7 versus 74±3.1,Po0.01, Figure 1B). Similar to the data from TEM, the portion ofMAP2 and synaptophysin colocalization in the immunochemical

double staining was significantly decreased in the dendritic areaof CA1 pyramidal layer of SAH rats as compared with that of salineand naive controls (Figure 2B). The Manders’ coefficients were0.24±0.05 for SAH, 0.38±0.07 for saline-injected and 0.37±0.1for naı̈ve controls (Pp0.05, Figure 2B).

Subarachnoid Hemorrhage Decreased Postsynaptic GluR1, CaMK-II, and Presynaptic Myelin Basic Protein, but Increased PostsynapticGluR2The loss of synapses prompted us to investigate whether therewas a change in the expression of key proteins that mediatemaintenance of LTP. Immunohistological staining for GluR1, GluR2,and CaMK-II showed that there was a significant decrease in theimmunoreactivity of CaMK-II and a near significant decrease ofGluR1 in the dendritic layer of CA1 pyramidal cells in comparisonwith those of saline-control and naı̈ve rats (Figure 3). Thenormalized CamK-II positive signal in SAH group was significantlydecreased (0.53±0.13 as compared with saline 1±0.13, Pp0.001,Figure 3B). Similarly, the normalized GluR1 signal in the SAH groupwas 0.74±0.23, as compared with saline control (1±0.12,P¼ 0.077, Figure 3B). In contrast, the normalized GluR2 positivestaining in the SAH group was 1.2±0.07, which was significantlyincreased as compared with saline (1±0.17, Pp0.05, Figure 3B).

Figure 2. Manders’ coefficient analysis of active synapses in the hippocampus. (A) Representative images showing the colocalization ofpresynaptic marker synaptophysin and postsynaptic marker microtubule-associated protein 2 (MAP2) (white areas as indicated by whitearrows in c, f, and i). Decreased colocalization (white area in images) is apparent in subarachnoid hemorrhage (SAH) group (i).(B) Subarachnoid hemorrhage induced a significant decrease of Manders’ coefficient on the colocalization of synaptophysin/MAP2 (Pp0.05,analysis of variance). Data are mean±s.d. All animals were killed 6 days after surgery.


4


To determine the involvement of a presynaptic component(axons) in the loss of synapses, we stained for myelin basic protein(MBP). We found that there was a decrease in immunoreactivity toMBP in the CA1 area of hippocampus of the SAH group incomparison with saline-control rats (Figure 4). The normalizedtotal area of MBP positive area in SAH rats was significantly lowerthan in saline controls (1±0.09, n¼ 4 for saline, 0.73±0.06, n¼ 6for SAH, Pp0.05).

Subarachnoid Hemorrhage did not Cause Significant NeuronDeathTo determine whether the decrease in number of synapses andrelated proteins in CA1 area of the hippocampus was associatedwith neuron death, we assessed the brains of naı̈ve, saline-injected,and SAH rats using three different stains. We did not find positivecells in TUNEL, caspase-3/NeuN, or Fluoro-Jade B staining in theCA3–CA1 area of the hippocampus of rats from any of the groups(Figure 5). Qualitative examination of other brain regions likeentorhinal cortex also did not show neuron death.

Subarachnoid Hemorrhage Induces Large Artery VasospasmVasospasm was evaluated in the anterior cerebral artery using aratio of lumen diameter/wall thickness. There was no significantdifference among the seven groups (analysis of variance). Therewas a trend for vasospasm in the SAH group as compared withnaive controls on day 6 (Figure 6, lumen/wall ratio was 18±5.5 fornaive control, 12.5±7.3 for SAH day 6, P¼ 0.065).

Nitric Oxide and Superoxide Anion Radical DetectionWe measured NO and O2

� in the hippocampus and cerebral cortexto assess the involvement of oxidative stress. There were nosignificant changes in NO or O2

� in the cerebral cortex (Figures 7Band 7D). However, there was a significant increase of NO anddecrease of O2

� in the hippocampus after SAH as compared withsaline controls. Nitric oxide is 1252±79 relative fluorescence unitsfor SAH compared with 518±118 relative fluorescence units forsaline controls (Figure 7A, Pp0.001, t-test). O2

� was 377±22relative luminescence unit for SAH and 590±49 RLU for salinecontrol (Figure 7C, Pp0.001, t-test).

Figure 3. Immunohistological staining of GluR1, GluR2, and CaMK-II. Representative images are shown in panel A, decreased signal of GluR1and CaMK-II staining, but increased signal of GluR2 staining was seen in subarachnoid hemorrhage (SAH) group as compared with that innaive or saline controls. Quantified data are shown in panel B. Data are mean±s.d. All animals were killed 6 days after surgery. *Po0.05,#Po0.001.


5


Cerebral Blood Flow and MetabolismUsing laser Doppler flow measurements, we found a dramaticdecrease in CBF ranging from 5% to 30% of baseline immediatelyafter injection of blood or saline (Supplementary Figure 1,Pp0.001 as compared with sham controls). Cerebral blood flowrecovered and stabilized at around 80% of the baseline level bythe time of transfer to the Minerve bed. There was no significantdifference among the groups in acute CBF recovery after theinjection (Supplementary Figure 1). Similarly, we did not find anysignificant difference between any of the groups (SAH and naı̈ve,sham-operated, and saline-injected controls) at any time points(day 0, 2, and 6 after SAH) in MRI measurements of CBF and PETmeasurements of glucose metabolism (Supplementary Figure 2,PX0.05, analysis of variance; Supplementary Tables 1 and 2).

DISCUSSIONNew FindingsThe new findings of this study are that SAH in rats was associatedwith a significant decrease of synapses in neurons in the CA1 area.This was accompanied by a significant reduction in severalproteins involved in LTP, includingCamK II and MBP, and a trend

towards reduction of GluR1. There was an increase in GluR2. Thesechanges may be responsible for the loss of LTP after SAH.Importantly, histologic studies with TUNEL, active caspase-3/NeuNand Fluoro-Jade B staining did not show significant neuron death,suggesting the loss of LTP in the hippocampus, which weobserved in a prior study in the same model,6 could occur inthe absence of cell death. Although there was a significantdecrease of CBF at the time point of saline or blood injection,there was no significant CBF or metabolism change at various timepoints after SAH (1 hour to 6 days), suggesting that the effects ofSAH are not mediated by ischemia. Oxidative stress may have arole in mediating the dysfunction of the hippocampal neurons.

Cognitive Impairment after Subarachnoid HemorrhageCognitive impairment, measured on the Montreal CognitiveAssessment, was detected in 40% of patients followed up afterSAH.20 Few animal studies have examined cognitive function usingneurobehavioral testing or electrophysiology.5,6,21,22 Takata, et al22

created SAH in rats by two injections into the cisterna magna.Subarachnoid hemorrhage was associated with significantlyincreased escape latency and swimming distance compared withsaline-injected or sham-operated animals in the Morris water maze.Similarly, we found that the escape latency and swimming distancewere significantly increased in rats with SAH as compared withcontrols. Subarachnoid hemorrhage rats also tended to do poorlyon accuracy in spatial and working memory tests.5 Similar findingswere reported by two other groups using an endovascularperforation model of SAH.21,23

The discrepancy between our studies may be due to differencesin methods. In the current study, rats were housed in an intensivecare incubator (at 261C, with easily accessible soft food and waterfor 48 hours) before their return to a room temperature cage. Incomparison, Jeon, et al5 returned injured rats directly to theirnormal room temperature cage after recovery from anesthesia.Mortality decreased from 30% to less than 0% between thesestudies, despite the same blood injection volume (300 mL).

It must be noted that the current results reflect early neuronalchanges after experimental SAH. Future studies should examineneuronal changes weeks after SAH, given the fact that cognitive/memory deficits in patients often persist for long period of time.1

In other studies, rats with SAH did have cognitive/memorydysfunction up to 5 weeks after ictus.21–23

Mechanism of Long-Term Potential Loss in Subarachnoid andOther Neurologic DiseasesOur previous study showed that the induction phase of LTP didnot change after SAH but that the maintenance phase of LTP wasdiminished.6 This suggests that the molecular machinery forinduction of LTP such as N-methyl-D-aspartate (NMDA) receptorsis intact, but the molecular components for maintenance of LTPsuch as a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic (AMPA)receptors and their associated trafficking, CamK-II activation, andother downstream processes, are impaired.24 The current resultsare consistent with this hypothesis. It seems that a decrease in keyproteins affects the function of these receptors, resulting in loss ofplasticity of the synapses, and that neuron loss in thehippocampus is not necessary. This is consistent with studies inhumans with SAH that show that cognitive dysfunction may beassociated with temporal lobe volume loss in the absence ofinfarctions.3,20 Animal studies in models of traumatic brain injuryalso found memory impairment in the absence of hippocampalcell death.25

Role of Key Proteins in Long-Term PotentiationEnhanced expression of phosphorylated GluR1 by CamK-II is a keymolecular event for induction and maintenance of LTP.26 In the

Figure 4. Immunohistological staining of myelin basic protein (MBP).(A) Images of MBP-stained slices from saline and subarachnoidhemorrhage (SAH) group. A mask of the positive signals wasgenerated for each image, and the total area of the signal wasquantified in panel B. There was significantly decreased MBP signalin the SAH group (Pp0.05, t-test). Data are mean±s.d. All animalswere killed 6 days after surgery.


6


current study, the decreased expression of CamK-II (and a trendtowards decreased expression of GluR1) in CA1 area of thehippocampus could explain the loss of LTP after SAH.6 Impairedcognitive function and loss of LTP associated with decreasedexpression or dysfunction of GluR1 and CamK-II were found inother diseases such as Fragile X syndrome and Alzheimerdisease27,28 and in GluR1 and CamK-II knockout mice29,30 butnot GluR2 or 3 knockout mice.31 Interestingly, IL-1b, a proinflam-matory cytokine was found to impair both memory and AMPA

receptor expression.32 This is of interest because IL-1b wassignificantly increased in cerebrospinal fluid and blood fromSAH patients and animals.33

Etiology of Hippocampal Changes after SubarachnoidHemorrhageSubarachnoid hemorrhage can cause ischemia due to increasedintracranial pressure or in the form of delayed cerebral ischemia.

Figure 5. Neuronl death stained by deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL), Fluoro-Jade B (B), and caspase-3/NeuN (C).No apparent neuronal cell death in subarachnoid hemorrhage (SAH) animals was detected by any method. All animals were killed 6 days aftersurgery. Kainate-treated animals were used as positive controls in TUNEL staining (A, a to c, white arrows).


7


We measured CBF with MRI and glucose metabolism with PET atdifferent time points after SAH to determine if ischemia mightmediate the hippocampal changes we observed. Experimentalevidence showed that persistent synaptic failure may result from

mild or moderate cerebral ischemia involving both pre- andpostsynaptic components of the hippocampus.34 Although therewas an acute decrease in CBF at the time of blood or salineinjection, we did not find differences between CBF or glucose

Figure 6. Vasospasm of anterior cerebral artery (ACA). (A) Representative images showing ACA from different groups of animals.(B) Quantification of lumen diameter/wall thickness ratio in all animal groups. There is no statistical significant difference among the groups(PX0.05 analysis of variance). However, there was a trend toward vasospasm in subarachnoid hemorrhage (SAH) day 6 animals (P¼ 0.065 ascompared with naive controls). Data are mean±s.d. All animals were killed 6 days after surgery.

Figure 7. Nitric oxide and superoxide anion radical (O2� ) analysis. Significantly increased NO was found in the hippocampus (A), Pp0.001,

Student’s t-test) but not cerebral cortex (B) in subarachnoid hemorrhage (SAH) group as compared with saline controls. In contrast,significantly decreased O2

� was found in the hippocampus (C), Pp0.001, Student’s t-test) but not cerebral cortex (D) in SAH group ascompared with saline controls. Data are mean±s.d. All animals were killed 6 days after surgery.


8


metabolism in SAH groups as compared with saline-injectedcontrols at any time (1 to 2 hours, 2 days, and 6 daysafter injection). Thus, ischemia may not be contributing tothe decreased expression of LTP-related proteins in this SAHmodel. One question is whether microcirculatory constriction orthrombosis could cause the changes in the absence of detectableCBF changes. There are reports that microthrombi mediatingfocal ischemia are likely formed in situ rather than from thebreakdown of distal larger emboli, and that these microthrombiwere accompanied by arteriolar vasoconstriction.12

If the hippocampal changes are not due to ischemia, then theymust be mediated by the subarachnoid blood. In the currentmodel, the blood was injected in the prechiasmatic cistern. If theblood is directly responsible for the hippocampal changes, onepossibility is that the blood flows retrogradely into the ventricles.We did see blood in the ventricles from SAH rats in histology(data not shown) and iron deposition based on Perl staining. Inthe T2-weighted data sets, some rats display abnormal patterns ofintraventricular T2 contrast, and some also present with brightintraventricular signal, which is characteristic of blood. However,the specificity of the T2 contrast is not good enough to distinguishblood retention from cerebrospinal fluid in the ventricles.

Subarachnoid blood has been suggested to cause oxidativestress.35 In this study, we found increased NO and decreased O2

� ,which was opposite to what we observed in our previous studies ina mouse model of SAH.13 This may be due to species differences orto the different times at which these parameters were assessed(two days after SAH in mice, 6 days after SAH in rats). Yatsushige,et al,36 found NO levels increased above basal levels at 24 hoursafter SAH created by endovascular perforation in rats. This wasbelieved to have detrimental effects on the brain becauseincreased NO could act as a free radical and form peroxynitritethat can damage the cell membrane of endothelial and smoothmuscle cells.37 At this point, the etiology of the changes in thehippocampus remains uncertain. Potential etiologies may includecortical spreading depolarization, oxidative stress, andinflammation.35,38 Interestingly, cortical spreading depolarizationhas been reported to enhance LTP,39 suggesting depolarizationwaves invading the hippocampus could be a mechanism ofendogenous protection antagonizing the mechanisms that causethe reduction of LTP after SAH.38,40

CONCLUSIONSThe anterior circulation model of SAH in rat causes loss of LTP,which is independent of global ischemia and neuron loss butwhich is associated with a decreased number of synapses, and adiminished density of AMPA receptors and CamK-II proteins at theSchaffer collateral—CA1 synapses in the hippocampus. The loss ofthese key proteins and possibly their functions in mediating LTPmaintenance may account for, or at least contribute to, thecognitive and neurologic impairments of patients with SAH. Thesekey proteins may be used as potential target to enhance or restorecognitive function in SAH patients.

DISCLOSURE/CONFLICT OF INTERESTThe authors declare no conflict of interest.

REFERENCES1 Al-Khindi T, Macdonald RL, Schweizer TA. Cognitive and functional outcome after

aneurysmal subarachnoid hemorrhage. Stroke 2010; 41: e519–e536.2 Bendel P, Koivisto T, Kononen M, Hanninen T, Hurskainen H, Saari T et al. MR

imaging of the brain 1 year after aneurysmal subarachnoid hemorrhage:randomized study comparing surgical with endovascular treatment. Radiology2008; 246: 543–552.

3 Tam AK, Kapadia A, Ilodigwe D, Li Z, Schweizer TA, Macdonald RL. Impact ofglobal cerebral atrophy on clinical outcome after subarachnoid hemorrhage.J Neurosurg 2013; 119: 198–206.

4 Shapiro ML, Eichenbaum H. Hippocampus as a memory map: synaptic plasticityand memory encoding by hippocampal neurons. Hippocampus 1999; 9: 365–384.

5 Jeon H, Ai J, Sabri M, Tariq A, Macdonald RL. Learning deficits after experimentalsubarachnoid hemorrhage in rats. Neuroscience 2010; 169: 1805–1814.

6 Tariq A, Ai J, Chen G, Sabri M, Jeon H, Shang X et al. Loss of long-term potentiationin the hippocampus after experimental subarachnoid hemorrhage in rats.Neuroscience 2010; 165: 418–426.

7 Bederson JB, Germano IM, Guarino L. Cortical blood flow and cerebral perfusionpressure in a new noncraniotomy model of subarachnoid hemorrhage in the rat.Stroke 1995; 26: 1086–1091.

8 Piepgras A, Thome C, Schmiedek P. Characterization of an anterior circulation ratsubarachnoid hemorrhage model. Stroke 1995; 26: 2347–2352.

9 Prunell GF, Mathiesen T, Svendgaard NA. A new experimental model in rats forstudy of the pathophysiology of subarachnoid hemorrhage. Neuroreport 2002; 13:2553–2556.

10 Solomon RA, Antunes JL, Chen RY, Bland L, Chien S. Decrease in cerebral bloodflow in rats after experimental subarachnoid hemorrhage: a new animal model.Stroke 1985; 16: 58–64.

11 Prunell GF, Mathiesen T, Diemer NH, Svendgaard NA. Experimental subarachnoidhemorrhage: subarachnoid blood volume, mortality rate, neuronal death, cerebralblood flow, and perfusion pressure in three different rat models. Neurosurgery2003; 52: 165–175.

12 Sabri M, Ai J, Lakovic K, D’abbondanza J, Ilodigwe D, Macdonald RL. Mechanismsof microthrombi formation after experimental subarachnoid hemorrhage.Neuroscience 2012; 224: 26–37.

13 Sabri M, Ai J, Knight B, Tariq A, Jeon H, Shang X et al. Uncoupling of endothelialnitric oxide synthase after experimental subarachnoid hemorrhage. J Cereb BloodFlow Metab 2011; 31: 190–199.

14 Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T et al.Fiji: an open-source platform for biological-image analysis. Nat Methods 2012; 9:676–682.

15 Zinchuk V, Zinchuk O, Okada T. Quantitative colocalization analysis of multicolorconfocal immunofluorescence microscopy images: pushing pixels to explorebiological phenomena. Acta Histochem Cytochem 2007; 40: 101–111.

16 Paxinos G, Franklin KBJ. The Rat Brain in Stereotaxic Coordinates. Academic Press:San Diego, 2001.

17 Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-braintransfer constants from multiple-time uptake data. J Cereb Blood Flow Metab 1983;3: 1–7.

18 Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, Pettigrew KD et al.The [14C]deoxyglucose method for the measurement of local cerebral glucoseutilization: theory, procedure, and normal values in the conscious and anesthe-tized albino rat. J Neurochem 1977; 28: 897–916.

19 Toyama H, Ichise M, Liow JS, Modell KJ, Vines DC, Esaki T et al. Absolutequantification of regional cerebral glucose utilization in mice by 18F-FDG small ani-mal PET scanning and 2-14C-DG autoradiography. J Nucl Med 2004; 45: 1398–1405.

20 Schweizer TA, Al-Khindi T, Macdonald RL. Mini-Mental State Examination versusMontreal Cognitive Assessment: rapid assessment tools for cognitive and func-tional outcome after aneurysmal subarachnoid hemorrhage. J Neurol Sci 2012;316: 137–140.

21 Sherchan P, Lekic T, Suzuki H, Hasegawa Y, Rolland W, Duris K et al. Minocyclineimproves functional outcomes, memory deficits, and histopathology after endo-vascular perforation-induced subarachnoid hemorrhage in rats. J Neurotrauma2011; 28: 2503–2512.

22 Takata K, Sheng H, Borel CO, Laskowitz DT, Warner DS, Lombard FW. Long-termcognitive dysfunction following experimental subarachnoid hemorrhage: newperspectives. Exp Neurol 2008; 213: 336–344.

23 Silasi G, Colbourne F. Long-term assessment of motor and cognitive behaviours inthe intraluminal perforation model of subarachnoid hemorrhage in rats. BehavBrain Res 2009; 198: 380–387.

24 Miyamoto E. Molecular mechanism of neuronal plasticity: induction and main-tenance of long-term potentiation in the hippocampus. J Pharmacol Sci 2006;100: 433–442.

25 Lyeth BG, Jenkins LW, Hamm RJ, Dixon CE, Phillips LL, Clifton GL et al. Prolongedmemory impairment in the absence of hippocampal cell death followingtraumatic brain injury in the rat. Brain Res 1990; 526: 249–258.

26 Barria A, Muller D, Derkach V, Griffith LC, Soderling TR. Regulatory phosphoryla-tion of AMPA-type glutamate receptors by CaM-KII during long-term potentiation.Science 1997; 276: 2042–2045.

27 Hsiao YH, Kuo JR, Chen SH, Gean PW. Amelioration of social isolation-triggeredonset of early Alzheimer’s disease-related cognitive deficit by N-acetylcysteine ina transgenic mouse model. Neurobiol Dis 2012; 45: 1111–1120.


9


28 Li J, Pelletier MR, Perez Velazquez JL, Carlen PL. Reduced cortical synaptic plas-ticity and GluR1 expression associated with fragile X mental retardation proteindeficiency. Mol Cell Neurosci 2002; 19: 138–151.

29 Andrasfalvy BK, Smith MA, Borchardt T, Sprengel R, Magee JC. Impaired regulationof synaptic strength in hippocampal neurons from GluR1-deficient mice. J Physiol2003; 552: 35–45.

30 Hinds HL, Tonegawa S, Malinow R. CA1 long-term potentiation is diminished butpresent in hippocampal slices from alpha-CaMKII mutant mice. Learn Mem 1998;5: 344–354.

31 Meng Y, Zhang Y, Jia Z. Synaptic transmission and plasticity in theabsence of AMPA glutamate receptor GluR2 and GluR3. Neuron 2003; 39:163–176.

32 Avital A, Goshen I, Kamsler A, Segal M, Iverfeldt K, Richter-Levin G et al.Impaired interleukin-1 signaling is associated with deficits in hippocampalmemory processes and neural plasticity. Hippocampus 2003; 13: 826–834.

33 Hendryk S, Jarzab B, Josko J. Increase of the IL-1 beta and IL-6 levels in CSF inpatients with vasospasm following aneurysmal SAH. Neuro Endocrinol Lett 2004;25: 141–147.

34 Hofmeijer J, van Putten MJ. Ischemic cerebral damage: an appraisal of synapticfailure. Stroke 2012; 43: 607–615.

35 Pluta RM, Hansen-Schwartz J, Dreier J, Vajkoczy P, Macdonald RL, Nishizawa Set al. Cerebral vasospasm following subarachnoid hemorrhage: time for a newworld of thought. Neurol Res 2009; 31: 151–158.

36 Yatsushige H, Calvert JW, Cahill J, Zhang JH. Limited role of inducible nitric oxidesynthase in blood-brain barrier function after experimental subarachnoidhemorrhage. J Neurotrauma 2006; 23: 1874–1882.

37 Ayer RE, Zhang JH. Oxidative stress in subarachnoid haemorrhage: significance inacute brain injury and vasospasm. Acta Neurochir Suppl 2008; 104: 33–41.

38 Dreier JP, Woitzik J, Fabricius M, Bhatia R, Major S, Drenckhahn C et al. Delayedischaemic neurological deficits after subarachnoid haemorrhage are associatedwith clusters of spreading depolarizations. Brain 2006; 129: 3224–3237.

39 Wernsmann B, Pape HC, Speckmann EJ, Gorji A. Effect of cortical spreadingdepression on synaptic transmission of rat hippocampal tissues. Eur J Neurosci2006; 23: 1103–1110.

40 Dreier JP. The role of spreading depression, spreading depolarization andspreading ischemia in neurological disease. Nat Med 2011; 17: 439–447.

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (http://www.nature.com/jcbfm)


10


http://www.nature.com/jcbfm

http://www.nature.com/jcbfm

molecular alterations in the hippocampus after ... · molecular alterations in the hippocampus...

Documents