Revealing the Penumbra through Imaging ... - ACS Publications

Subscriber access provided by MT ROYAL COLLEGE

Letter

Revealing the penumbra through imaging elemental markers of cellular metabolism in an ischemic stroke model M. Jake Pushie, Andrew Crawford, Nicole J Sylvain, Huishu Hou, Mark J. Hackett, Graham N. George, and Michael E Kelly ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.7b00382 • Publication Date (Web): 25 Jan 2018 Downloaded from http://pubs.acs.org on January 26, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Chemical Neuroscience is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Neuroscience

Revealing the penumbra through imaging elemental markers of cellular metabolism in an ischemic stroke model

M. Jake Pushie (PhD),1 Andrew M. Crawford (PhD),2 Nicole J. Sylvain (MSc),1 Huishu Hou (MSc),1 Mark J. Hackett(PhD),3,4 Graham N. George (DPhil),2 Michael E. Kelly (MD, PhD)1*

1. Department of Surgery, Division of Neurosurgery, College of Medicine, University of Saskatchewan, 107 Wiggins Road, Saskatoon, Saskatchewan, S7N 5E5, Canada 2. Geological Sciences, College of Arts & Science, University of Saskatchewan, 114 Science Place, Saskatoon, Saskatchewan, S7N 5E2, Canada 3. Curtin Institute for Functional Molecules and Interfaces, Department of Chemistry, Faculty of Science & Engineering, Curtin University, Kent Street, Bentley, Perth, Western Australia 6102 4. Curtin Health Innovation Research Institute, Curtin University, Bentley, Western Australia 6102, Australia

*Corresponding Author: E-mail: [email protected]; Ph: 1-306-844-1104

Cover Title: Imaging the penumbra with synchrotron X-rays Display Items: Figures: 5 Key Words: ischemic stroke, X-ray fluorescence imaging, penumbra, metabolites, ions Subject terms: Ischemia; Metabolism; Imaging; Ischemic Stroke;

0 ACS Paragon Plus Environment

ACS Chemical Neuroscience 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Stroke exacts a heavy financial and economic burden, is a leading cause of death, and the leading cause of long term disability in those who survive. The penumbra surrounds the ischemic core of the stroke lesion and is comprised of cells that are stressed and vulnerable to death, which is due to an altered metabolic, oxidative and ionic environment within the penumbra. Without therapeutic intervention many cells within the penumbra will die and become part of the growing infarct, however, there is hope that appropriate therapies may allow potential recovery of cells within this tissue region, or at least slow the rate of cell death, therefore, slowing the spread of the ischemic infarct and minimising the extent of tissue damage. As such, preserving the penumbra to promote functional brain recovery is a central goal in stroke research. While identification of the ischemic infarct, and the infarct/penumbra boundary is relatively trivial using classical histology and microscopy techniques, accurately assessing the penetration of the penumbra zone into undamaged brain tissue, and evaluating the magnitude of chemical alterations in the penumbra, has long been a major challenge to the stroke research field. In this study we have used synchrotron-based X-ray fluorescence imaging to visualise the elemental changes in undamaged, penumbra, and infarct brain tissue, following ischemic stroke. We have employed a Gaussian mixture model to cluster tissue areas based on their elemental characteristics. The method separates the core of the infarct from healthy tissue, and also demarcates discrete regions encircling the infarct. These regions of interest can be combined with elemental and metabolic data, as well as with conventional histology. The cell populations defined by clustering provide a reproducible means of visualizing the size and extent of the penumbra at the level of the single cell and provide a critically-needed tool to track changes in elemental status and penumbra size.


Page 2 of 31

Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60




A new stroke occurs every 10 minutes in Canada, making it the third leading cause of death in the country.1 Nearly 30% of these strokes are fatal, while the remaining cases make stroke a leading cause of disability. The risk factors for stroke are well-known, such as age, smoking habits, dyslipidemia, hypertension, and family history. If not resolved quickly, acute cerebrovascular occlusion rapidly leads to irreversible cerebral ischemia due to lack of oxygen and glucose. According to the World Health Organization, while stroke prevalence is overshadowed by lower respiratory infections and HIV/AIDS in low-income countries, stroke is the leading cause of death in developed countries.2 Once a stroke has occurred the most severely impacted tissue comprises the core of the infarct (Fig. 1). The surrounding peri-infarct tissue, partially supplied by continued peripheral circulation, is regarded as containing two conceptually distinct regions – immediately adjacent to the infarct is the penumbra, comprised of oxidatively and metabolically stressed cells, which may become part of the growing infarct if left untreated, or with an appropriate therapy can potentially recover (Fig. 1). The second peri-infarct region, isa more distant region of benign oligemia which is perfused at a near-normal level (Fig. 1). While many risk factors for stroke can be managed, a stroke can occur regardless the level of risk mitigation, and it is impossible to predict who will have a stroke, or when one will occur. Although stroke mitigating therapies, such as rtPA,3-6 or endovascular therapy for large vessel occlusions,7-12 can be highly beneficial, only a small number of patients are candidates for these treatments. Developing new therapies that reduce the size of the stroke lesion once it has already occurred therefore remains an outstanding challenge for the stroke field. In animal stroke models the ischemic infarct size (sometimes referred to as lesion size) is typically determined histologically.13 In haematoxylin and eosin (H&E) stained ischemic stroke tissue, the infarct is identifiable as a region of pan-necrosis with reduced eosin staining 3 ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


surrounded by normal-appearing tissue. At early post-stroke time points, the infarct border demonstrates evidence of selective neuronal necrosis. On-going oxidative and metabolic stresses, including generation of reactive oxygen and nitrogen species, and the lack of ATP to maintain normal operation of ion-channel pumps, within the affected tissue, is thought to contribute to continued cell death post-stoke, which peaks approximately 24h post-stroke.14 Ourselves and others have previously demonstrated that alterations in elemental content may serve as useful markers of tissue damage driven by oxidative and or metabolic stress.15-17 It is also well established that metabolic stress and the associated failure of ion pumps results in large intracellular increases in Cl- and Ca2+ and decreases in K+.18,19 Cell death from within the penumbra contributes to further functional decline in the immediate period post-stroke. Due to the above, a central goal of stroke treatment, therefore, is to rescue the penumbra to promote functional recovery. While the border of reduced eosin staining can be readily identified histologically, reliable differentiation of the extent of potentially viable penumbra tissue surrounding the infarct is otherwise difficult to discern with conventional methods. Cells within the penumbra suffer a compromised metabolic and oxidizing environment,althoughpotentially recoverable, some cells may show signs of selective necrosis, identifiable as pyknotic cells with shrunken nuclei surrounded by red cytoplasm in H&E stained tissue sections, but this is inadequate for identifying the volume of tissue within the penumbra or the extent of penumbra preservation following treatments. The ability to accurately and reliably define the size of the penumbra is anticipated to aid characterization and analysis by improving statistical analysis of therapeutic treatments in animal models. Further the ability to characterise the extent and magnitude of elemental alterations in the penumbra, in particular, changes in the ocncnetraitons of ions such as Cl-, K+, and Ca2+, which are heavily influence by cell metabolic “health”18,19 may



also serve to improve our understanding of the timeline and location of altered metabolic state of the cells in this critical region. Synchrotron sources provide a tunable source of X-rays which can be used to perform Xray fluorescence imaging (XFI), an element-specific technique that can map the elemental distributions in a wide range of sample types.20 The incident X-rays excite elements in the tissue, inducing elemental fluorescence. This fluorescence can be recorded, allowing a broad range of fluorescence energies from multiple elements to be recorded simultaneously from the area of sample illuminated by the beam. Applied to biological specimens XFI can be used to map the metabolic status of tissue, with a pixel resolution of 2-50 µm routinely available,15,20,21 and submicron imaging accessible with specialized setups.22,23 While particularly bright X-ray sources can be damaging to samples over long durations, the type of imaging performed herein is non-descructive and allows follow-on experiments to be performed on the same tissue after XFI data collection, such as histological staining. Employing XFI to map elemental distributions in coronal sections of stroke models, followed by analysis of multiple elemental maps which groups pixels into statistically similar clusters, to define regions of interest, we have identified concentric areas of tissue surrounding the infarct which are metabolically distinct from both the core of the infarct and unaffected tissue and relate them to the concept of the ischemic penumbra.24 Herein, we describe, for the first time, the application of XFI characterization of the peri-infarct zone in the photothrombotic mouse model of ischemic stroke.

Elemental distributions in the PT stroke model 5 ACS Paragon Plus Environment

Page 6 of 31

Page 7 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


XFI reveals the distribution of a wide range of elements at their physiological concentrations (Fig. 2), except for Na and Mg because their X-ray fluorescence energies are too low to observe in our experiments. Resolving elements with overlapping fluorescence emissions has been problematic when integrating photon counts over discrete energy ranges. In the case of K and Ca, their Kα fluorescence peaks overlap appreciably, leading to X-ray fluorescence photons from areas with high K concentration erroneously contributing to a falsely high measurement for Ca (vice versa within the infarct).Error! Bookmark not defined.,25 By deconvoluting the X-ray fluorescence spectrum elements with overlapping emission peaks can be reliably separated with improved detection limits. Cellular metabolic failure following ischemic stroke results in loss-of-function of ion channel pumps, both in the infarct as well as in the metabolically stressed cells within the surrounding penumbra. Disruption of ion levels within and surrounding the stroke lesion are an indirect marker of unhealthy tissue and a reflection of metabolic failure. The deconvoluted XFI data (Fig. 2) reveals the elemental levels in normal tissue as well as the significant change in elemental content within the stroke lesion. Reductions in total P, S, K, Mn, Fe, Cu, and Zn are observed in the damaged tissue, whereas Cl and Ca are greatly elevated compared to normal tissue. Increased Ca, Cl, and Na are associated with ischemic infarction and are a hallmark of cellular dyshomeostasis and necrotic changes.19,26 Neuroanatomical variability can also be discerned in most elemental maps, with tissue types demonstrating markedly different levels for all elements shown in Fig. 2 (with the exception of Mn, which is at trace levels). We have observed that the apparent distribution of Ca in the XFI maps is an artifact of tissue drying prior to data collection, an artifact not apparent for other elements. As the tissue sections dry we observe microcrystals form on the surface of the lesion. We have confirmed that these 6 ACS Paragon Plus Environment


microcrystals are the source of the high concentration Ca spots (as demonstrated in the Ca map in Fig. 2). We hypothesize that these microcrystals form as a result of high extracellular Ca2+, which complexes with anionic ligands also abundant in the infarct (such as glutamate and aspartate).27 Upon drying these complexes self-organize into crystalline arrangements. While the local distribution of Ca is an artifact the total Ca content within the tissue remains unchanged. There also appears to be at least two distinct distributions of Fe – the first is widely distributed through all tissues in the brain, providing a low Fe background in the tissue, while the intensely concentrated areas of Fe are associated with blood vessels (Fig. 2, confirmed histologically). Cu is also widely distributed throughout all tissues in the brain at a low concentration, however, elevated levels are found in the periventricular zones (Fig. 2).28,29

Clustering reveals distinct peri-infarct zones The elemental content, particularly the ions integral to maintaining membrane potential in neuronal cells, serve as an indicator of the metabolic state of cells within tissue. We hypothesized that by using data from multiple elements from stroke model imaging we could generate regions of interest for metabolically-related areas of tissue. To this end we have carried out Gaussian clustering, using expectation maximization with XFI maps of P, S, Cl, K, and Zn, where each pixel is represented by a concentration for each of these 5 elements. Figure 3 shows the distributions of the 5 selected elements and the resulting regions of interest (in no particular order). The regions can be related to the core of the ischemic infarct, unaffected tissue, and the peri-infarct zones (PIZ-1 and PIZ-2 in Fig. 3). Clustering was optimal when a 5th cluster was included, which contained pixels along the edge of the tissue (not shown). These distinct ROIs


Page 8 of 31

Page 9 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


can be used to determine the size (area) of a particular region, or mask pixels and obtain regionspecific elemental quantification data. By registering and overlaying the regions of interest from clustering and the XFI data with the corresponding histological image of the tissue the clusters can be used to demarcate metabolically distinct tissue regions (Fig. 4). In the photothrombotic model depletion of P, S, K, Fe, Cu, and Zn, and elevation of Cl and Ca are associated with the infarct and PIZ-1, with a trend toward lower levels of these elements in PIZ-2. Intermediate levels of S, K, and Zn are found within PIZ-1. The levels of Cl and Ca, however, are significantly elevated above concentrations that can be explained by infiltration of extracellular ions present in the brain parenchyma. Although the photothrombotic stroke model is understood to generate a permanent occlusion,39 without subsequent reperfusion, the blood-brain-barrier within and surrounding the ischemic stroke lesion becomes permeable.30 Disruption of the blood-brain-barrier allows low molecular weight species, particularly water and ions, to re-equilibrate between the blood and brain parenchyma. We hypothesize that the large increase in Cl and Ca (and likely Na, which cannot be measured at biologically-relevant concentrations with current XFI methods) is likely to be largely derived from blood, which has a significantly higher concentration of these ions than the brain.31 Similarly, chemical species that are higher concentration in the brain parenchyma relative to blood, will re-equilibrate and demonstrate a net loss within the stroke lesion. This hypothesis is supported by observations of increased Zn and Glu levels following hypoxic challenge.32,33 Nitric oxide production is also increased in ischemic tissue34 and has been shown to contribute to loss of Fe from within ferritin stores.35 Although there is a trend toward lower Fe within the peri-infarct zones relative to



unaffected tissue (Fig. 5) the average concentration of Fe within all regions of interest (except the infarct) is complicated by the presence of blood vessels. The regions of interest from Gaussian clustering (Fig. 3) can be used to selectively analyze XFI data and compare elemental differences within these distinct regions (Fig. 5). As noted earlier, aside from gross general distributions, Ca localization is complicated by postsectioning drying artifacts confounding interpretation. The average concentration between PIZ-1 and PIZ-2, however, do show significant differences for this element, despite its exclusion from elements used to define the clusters. While the tissue along the borders of the regions of interest obtained from clustering do not demonstrate readily identifiable histological differences (e.g. Fig. 4), the underlying differences in elemental concentrations revealed by XFI demonstrate the altered metabolic states of the tissue. The photothrombotic infarct contains widespread pan-necrosis, where all cells within the tissue are necrotic, whereas within PIZ-1 the tissue begins to transition from pannecrosis toward selective neuronal necrosis with Cl and Ca infiltration reaching similar levels to those found in the necrotic core. Both of these regions contain eosinophilic neurons, identifiable with shrunken cell bodies, red eosinophilic cytoplasm, and darkly stained pyknotic nuclei. A larger density of cells is observed within the PIZ-1 region, particularly along the PIZ-2 border, although this region still contains primarily eosinophilic neurons. The PIZ-1/PIZ-2 border shown in Figure 4 closely follows the histologic border used to differentiate the infarct at 1-day post-stroke.13,36 It is important to note that the region that clusters as unaffected tissue looks indistinguishable from tissue within PIZ-2 at 24h, however, it is clear from the elemental


Page 10 of 31

Page 11 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


maps that the tissue within PIZ-2 is metabolically altered from otherwise normal tissue for most elements (Fig. 5). We have applied the clustering algorithm to a multi-dimensional dataset, combining elemental maps of P, S, Cl, K, and Zn distribution from an ischemic stroke model to generate regions of interest that differentiate metabolically distinct regions of tissue based on the elemental concentrations. Clustering of XFI data is a robust and non-subjective method for revealing metabolically distinct regions of tissue. In the case of the unaffected tissue and PIZ-2 in Figure 4, these tissue regions can appear indistinguishable histologically. The PT stroke model employed herein induces a permanent stroke and the model therefore the ischemic penumbra lacks post-stroke damage due to subsequent reperfusion of the tissue. The data clustering method presented effectively differentiates metabolic states of the tissue surrounding the infarct. While regions defined through clustering do not inherently define the ischemic penumbra, the regions do represent statistical differences in key elemental concentrations and are therefore relatable to the concept of the ischemic penumbra. Nevertheless, these types of experimental and data analysis methods are likely to prove useful in the evaluation of tissues surrounding the infarct during post-stroke time course studies, as well as provide a non-biased method for evaluating treatment efficacy for shrinking infarct area and penumbra size. The number of clusters obtained can be reduced, for example, forcing the algorithm to group data into fewer regions, in which case the infarct and unaffected tissue cluster separately, while PIZ-1 and PIZ-2 (shown in Fig. 4) cluster as a single region (albeit with some slight variability along the borders). Although the maximum likelihood expectation maximization



algorithm is nondeterministic, we implement a K-means clustering algorithm first, which is used as the initial seed for subsequent clustering, thereby generating reproducible regions of interest. The changes in elemental levels that we see with the photothrombotic model agree with the central dogma underlying the metabolic description of membrane potential dyshomeostasis in the pathophysiology of ischemic stroke. An increase in Ca and Cl are in agreement with this mechanism, as are the significant loss of most other elements, including P, S, K, Mn, Fe, Cu, and Zn. While histology is typically employed to characterize the boundary between the ischemic infarct and the penumbra our XFI data and cluster analysis reveals the full extent of the penumbra, including the boundary between the penumbra and healthy tissue, which is not otherwise evident histologically. It is known that cells within the peri-infarct zone will degenerate and that the cells themselves, as well as the extracellular space, have altered ion levels resulting from metabolic failure due to disrupted energy metabolism during the ischemic event; however, there have not been any previous methods which allow this important physiological aspect of ischemic injury to be imaged. This study has identified a new method of analysis for characterizing the extent of tissue injury after an ischemic stroke using XFI. We anticipate that this method will be beneficial in future stroke models for monitoring the efficacy of stroke therapies through quantification of the size and extent of the ischemic infarct as well as the penumbra, along with the ability to quantitatively image ion dyshomeostasis following a stroke. The photothrombotic stroke model generates a highly focal stroke, mimicking permanent vessel occlusion, without subsequent reperfusion. Future XFI experiments and clustering analysis using alternative stroke models, such as the middle cerebral artery occlusion or a hypoxia/ischemia models37,38 will be necessary to identify if the trends observed here are common across other ischemic stroke models. 11 ACS Paragon Plus Environment

Page 12 of 31

Page 13 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Methods Animal Model and Tissue Preparation Five male BALB/c mice were housed in the University Saskatchewan’s Large Animal Services Unit with a 12h light/dark cycle, with ad libitum access to chow and water. Preparation of mice in the photothrombotic (PT) stroke treatment group followed previously published methods.Error! Bookmark not defined.,39 Briefly, eleven week old mice were anesthetized using isoflurane (2% induction, 1.5-2% maintenance in 40% O2 and 60% N2O, Baxter Corp., Toronto ON, CA. The fur above the skull was removed and the head secured in a stereotactic frame. Body temperature was maintained with a RightTemp monitor and homeothermic controller (Kent Scientific Corp. USA), and breathing rate, pulse and O2 SAT was monitored with a MouseOX Plus (Starr Life Sciences Corp. USA). The skull was exposed with an incision in the midline of the scalp. The area of skull overlying the primary somatosensory cortex was identified for alignment of the laser (S1FL, -1 to -2.5 mm midline to lateral; +1 to -0.5 mm anterior to posterior of bregma, right hemisphere). An intraperitoneal injection of Rose Bengal (100mg/kg, Sigma, USA) was administered and the dye was allowed to circulate for 5 minutes following injection. A sterile mask was applied surrounding the region of interest to minimize scatter of the laser light, and the S1FL was illuminated for 20 minutes with a green laser (532 nm, positioned 3cm above the skull) to photoactive the Rose Bengal. Time zero coincides with the time the laser was shut off. Sham mice were omitted from exposure to the laser. Marcaine (1.67 mg/mL, 18 µL/15g) was applied to the open wound before suturing. Animals recovered in single animal cages post-surgery.



Animals were sacrificed by first anaesthetizing with isoflurane, followed immediately by removing the heads with a decapitator. To preserve biochemical speciation and distribution of labile ions the heads were then submerged in liquid N2 in order to rapidly freeze the brain.40,41 Heads were maintained at ca. −20°C while the fur, skin, muscle and finally bone were chiseled away from the frozen brain. The brains were sectioned using a cryomicrotome, at an operating temperature of −18°C, starting +0.25 mm anterior to the bregma. The water content of biological samples can lead to X-ray beam-induced photo-damage42,43 and therefore the 30 µmthick sections were collected on metal-free Thermanox coverslips and allowed to air dry. Unlike many glass or plastic alternatives, Thermanox coverslips are largely free of metals and other elements of interest (some trace Co is detectable with XFI), making them an advantageous choice for imaging trace levels of elements in biological tissue sections. This sample thickness was chosen because the neuroanatomical features in this brain region do not change significantly over the 30 µm distance, and the incident X-ray beam is sufficiently penetrating that trace levels of Mn, Fe, Cu, and Zn can be readily detected in the ~35 µm x 35 µm x 30 µm volume of tissue contributing fluorescence signal for each pixel. Prior to XFI analysis tissue sections were allowed to air-dry at ambient temperature. Following XFI data collection the sections were postfixed by immersion in 10% buffered formalin for histological staining. All animal work was conducted with approval from the University of Saskatchewan’s Animal research Ethics Board and carried out in accordance with the Canadian Council on Animal Care guidelines for humane animal use. Synchrotron X-ray Fluorescence Imaging


Page 14 of 31

Page 15 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


XFI was carried out at the Stanford Synchrotron Radiation Lightsource (SSRL) on beamline 10-2, with the SPEAR3 storage ring operating in top-up mode at 3GeV and 500 mA. Beamline 10-2 is comprised of a 33-pole 1.27-Tesla wiggler, using a Si(111) double-crystal monochromator (phi = 90o) with a Rh-coated mirror for focusing. The incident energy was set below the Br K-edge to 13,450 eV. A microfocussed beam approximately 35 µm x 35 µm was achieved using an aperture upstream of the I0 ion chamber which removes the less intense portions of the beam in the horizontal and vertical directions. Samples were mounted at 45° to the incident beam and raster scanned using Newport (Irvine CA, USA) IMS Series stages in 30 µm steps, giving a 30 µm pixel resolution with a 0.5 µm position precision over its 600 mm of travel, and a dwell time of 200 ms per point. A silicon-drift Vortex detector was positioned 45° to the sample normal and 90° to the incident beam, and fitted with a collimator to reduce total counts from background. The full multi-channel array spectrum for each point in the XFI map was saved for off-line processing.

Data Processing and Analysis Multi-channel array spectra were deconvoluted for each pixel in the XFI data using the program MBlank, which fits the total X-ray fluorescence emission using a series of equations that model the emission lines for each element, and also includes contributions from background and scatter.44 This method of deconvolution allows contributions from overlapping fluorescence lines (i.e. K Kα and Kβ with Ca Kα) to be separated, and their individual contributions to the spectrum quantified. This approach to fitting the MCA spectrum yields more accurate elemental quantification, improves signal-to-noise, and affords somewhat improved detection limits.



Fluorescence intensities were converted to aerial concentrations (µg/cm2) using reference standards deposited on 6.3 µm-thick mylar film: K and Cl (KCl, 98.8 µg/cm2), Ca (CaF2, 56.8 µg/cm2), P (47.2 µg/cm2), Fe (Fe 56.0 µg/cm2), Cu (CuSx 95.9 µg/cm2), and Zn (ZnTe, 45.8 µg/cm2) (Micromatter, Vancouver, CA).

Data Clustering XFI data from P, S, Cl, K, and Zn for each pixel were collectively analyzed using expectation maximization (EM), a Gaussian mixture-based soft clustering method similar to that described previously for similar synchrotron imaging data.45 Each region of interest was randomly seeded and optimized using K-means clustering during the initial analysis pass. The refined K-means clusters were then used to seed subsequent clustering using EM. The log likelihood was calculated for each EM run and used as an objective measure to identify completion. The clustering was repeated 3 times for each image to verify reproducibility. Clustering was performed in the XFI data analysis package MBlank.44


Page 16 of 31

Page 17 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Author Information MEK and MJH conceived the experiments; NJS and HH performed animal surgeries and prepared tissue sections for analysis; MJH, NJS, HH, and MJP collected X-ray fluorescence data, AMC and NJS processed data; AMC suggested the clustering approach for data analysis and developed software for this; MJP performed clustering and analyzed all data; MJP prepared all figures; MJP wrote the manuscript; MJP, NJS, AMC, MJH, GNG, and MEK edited the manuscript.

Funding Sources MEK is the Saskatchewan Research Chair in Clinical Stroke Research, and this project was supported by the Heart & Stroke Foundation, the Saskatchewan Health Research Foundation, and the University of Saskatchewan College of Medicine. GNG is a Tier I Canada Research Chair in X-ray Absorption Spectroscopy. MJH was supported by a Canadian Institute of Health Research Fellowship during a portion of this research. Additionally, this work was supported by: a SHRF Establishment Grant awarded to MEK; a joint Canadian Institutes of Health Research (CIHR)/Heart and Stroke Foundation (HSFC) Synchrotron Medical Imaging Team Grant #CIF 99472 awarded to Dr. Helen Nichol, MEK, GNG, and others; and a SHRF Gene Expression Mapping using Synchrotron Light grant awarded to Dr. Helen Nichol, MEK, GNG, and others.

Acknowledgements



NJS, MJP, MJH and AMC are Associate members of the Canadian Institutes of Health Research-funded program Training in Health Research Using Synchrotron Techniques (THRUST). The authors thank Drs. Roland Auer and Phyllis Paterson for useful discussions, the use of Paterson lab surgical equipment, Sharleen Weese Maley for administrative support, and Dr. Sally Caine for surgical assistance. XFI experiments were carried out at the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, which is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515.

Conflict of Interest Disclosures: The authors declare no competing financial interest.


Page 18 of 31

Page 19 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References 1

Statistics Canada, G.o.C., Leading causes of death, by sex. 2012 (CANSIM table 102-0561).

2

WHO Fact sheet N°310, http://www.who.int/mediacentre/factsheets/fs310/en/, accessed Dec 7, 2016

3

The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group (1995) Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 333:1581-1587.

4

Boysen G, ECASS Study Group (1995) European Cooperative Acute Stroke Study (ECASS): (rt-PA-Thrombolysis in acute stroke) study design and progress report. Eur J Neurol. 1:213219.

5

Ford G, Freemantle N (1999) ECASS-II: intravenous alteplase in acute ischaemic stroke. European Co-operative Acute Stroke Study-II. Lancet. 353:65.

6

Albers GW, Clark WM, Madden KP, Hamilton SA. (2002) ATLANTIS Trial: Results for Patients Treated Within 3 Hours of Stroke Onset. Stroke. 33:493-496.

7

Campbell BC, Mitchell PJ, Yan B, Parsons MW, Christensen S, Churilov L, Dowling RJ, Dewey H, Brooks M, Miteff F, Levi C, Krause M, Harrington TJ, Faulder KC, Steinfort BS, Kleinig T, Scroop R, Chryssidis S, Barber A, Hope A, Moriarty M, McGuinness B, Wong AA, Coulthard A, Wijeratne T, Lee A, Jannes J, Leyden J, Phan TG, Chong W, Holt ME, Chandra RV, Bladin CF, Badve M, Rice H, de Villiers L, Ma H, Desmond PM, Donnan GA, Davis SM; EXTEND-IA investigators. (2014) A multicenter, randomized, controlled study to investigate EXtending the time for Thrombolysis in Emergency Neurological Deficits with Intra-Arterial therapy (EXTEND-IA). Int J Stroke. 9:126-132.

8

Fransen PSS, Beumer D, Berkhemer OA, van den Berg LA, Lingsma H, van der Lugt A, van Zwam WH, van Oostenbrugge RJ, Roos YBWEM, Majoie CB, Dippel DWJ; MR CLEAN 18 ACS Paragon Plus Environment


Trial Investigators. (2014) MR CLEAN, a multicenter randomized clinical trial of endovascular treatment for acute ischemic stroke in the Netherlands: study protocol for a randomized controlled trial. Trials 15:343. 9

Goyal M, Demchuk AM, Menon BK, Eesa M, Rempel JL, Thornton J, Roy D, Jovin TG, Willinsky RA, Sapkota BL, Dowlatshahi D, Frei DF, Kamal NR, Montanera WJ, Poppe AY, Ryckborst KJ, Silver FL, Shuaib A, Tampieri D, Williams D, Bang OY, Baxter BW, Burns PA, Choe H, Heo JH, Holmstedt CA, Jankowitz B, Kelly M, Linares G, Mandzia JL, Shankar J, Sohn SI, Swartz RH, Barber PA, Coutts SB, Smith EE, Morrish WF, Weill A, Subramaniam S, Mitha AP, Wong JH, Lowerison MW, Sajobi TT, Hill MD; ESCAPE Trial Investigators. (2015) Randomized assessment of rapid endovascular treatment of ischemic stroke. N Engl J Med. 372:1019-1030.

10

Molina CA, Chamorro A, Rovira À, de Miquel A, Serena J, Roman LS, Jovin TG, Davalos A, Cobo E. (2015) REVASCAT: a randomized trial of revascularization with SOLITAIRE FR device vs. best medical therapy in the treatment of acute stroke due to anterior circulation large vessel occlusion presenting within eight-hours of symptom onset. Int J Stroke. 10:619626.

11

Saver JL, Goyal M, Bonafe A, Diener HC, Levy EI, Pereira VM, Albers GW, Cognard C, Cohen DJ, Hacke W, Jansen O, Jovin TG, Mattle HP, Nogueira RG, Siddiqui AH, Yavagal DR, Devlin TG, Lopes DK, Reddy V, du Mesnil de Rochemont R, Jahan R; SWIFT PRIME Investigators. (2015) Solitaire™ with the Intention for Thrombectomy as Primary Endovascular Treatment for Acute Ischemic Stroke (SWIFT PRIME) trial: protocol for a randomized, controlled, multicenter study comparing the Solitaire revascularization device with IV tPA with IV tPA alone in acute ischemic stroke. Int J Stroke. 10:439-448. 19 ACS Paragon Plus Environment

Page 20 of 31

Page 21 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


12

Goyal M, Menon BK, van Zwam WH, Dippel DW, Mitchell PJ, Demchuk AM, Dávalos A, Majoie CB, van der Lugt A, de Miquel MA, Donnan GA, Roos YB, Bonafe A, Jahan R, Diener HC, van den Berg LA, Levy EI, Berkhemer OA, Pereira VM, Rempel J, Millán M, Davis SM, Roy D, Thornton J, Román LS, Ribó M, Beumer D, Stouch B, Brown S, Campbell BC, van Oostenbrugge RJ, Saver JL, Hill MD, Jovin TG2; HERMES collaborators. (2016) Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet. 387:1723-1731.

13

Zille M, Farr TD, Przesdzing I, Müller J, Sommer C, Dirnagl U, Andreas Wunder A (2012) Visualizing cell death in experimental focal cerebral ischemia: promises, problems, and perspectives. J Cereb Blood Flow Metab. 32:213–231.

14

Asahi M, Hoshimaru M, Uemura Y, Tokime T, Kojima M, Ohtsuka T, Matsuura N, Aoki Y, Shibahara K, Kikuchi H. (1997)Expression of interleukin-1 beta converting enzyme gene family and bcl-2 gene family in the rat brain following permanent occlusion of the middle cerebral artery, J. Cereb. Blood Flow Metab. 17:11–18.

15

Caine S, Hackett MJ; Kumar S; Maley J; Ivanishvili Z, Suen B, Szmigielski A, Jiang Z; Sylvain, NJ; Nichol H, Kelly ME. (2016) A Novel Multi-Modal Platform To Image Molecular And Elemental Alterations In Ischemic Stroke. Neurobiology of Disease. 91: 132142.

16

Hackett MJ, DeSouza M, Caine S, Bewer B, Nichol H, Paterson PG, Colbourne F. (2015) A new method to image heme-Fe, total Fe, and aggregated protein levels after intracerebral hemorrhage. ACS Chem Neurosci 6:761-770.

17

Chwiej J, Patulska A, Skoczen A, Matusiak K, Janeczko K, Ciarach M, Simon R, Setkowicz Z. (2017) Various ketogenic diets can differently support brain resistance against 20 ACS Paragon Plus Environment


experimentally evoked seizures and seizure-induced elemental anomalies of hippocampal formation. J Trace Elem Med Biol. 42:50-58. 18

Hansen AJ, Zeuthen T. (1981) Extracellular ion concentrations during spreading depression and ischemia in the rat brain cortex. Acta Physiol Scand. 113:437-445.

19

Hansen AJ, Nedergaard M. (1988) Brain ion homeostasis in cerebral ischemia. Neurochem Pathol. 9:195-209.

20

Pushie MJ, Pickering IJ, Korbas M, Hackett MJ, George GN. (2014) Elemental and chemically specific X-ray fluorescence imaging of biological systems. Chem Rev. 114:8499541.

21

Lins BR, Pushie MJ, Jones M, Howard DL, Howland JG, Hackett MJ. (2016) Mapping Alterations to the Endogenous Elemental Distribution within the Lateral Ventricles and Choroid Plexus in Brain Disorders Using X-Ray Fluorescence Imaging. PLoS One. 11:e0158152.

22

Hackett MJ, George GN, Pickering IJ, Eames BF. (2016) Chemical Biology in the Embryo: In Situ Imaging of Sulfur Biochemistry in Normal and Proteoglycan-Deficient Cartilage Matrix. Biochemistry. 55:2441-2451.

23

Summers KL, Fimognari N, Hollings A, Kiernan M, Lam V, Tidy RJ, Paterson D, Tobin MJ, Takechi R, George GN, Pickering IJ, Mamo JC, Harris HH, Hackett MJ. (2017) A Multimodal Spectroscopic Imaging Method To Characterize the Metal and Macromolecular Content of Proteinaceous Aggregates ("Amyloid Plaques"). Biochemistry. 56:4107-4116.

24

J Astrup, B K Siesjö, L Symon. (1981) Thresholds in cerebral ischemia - the ischemic penumbra. Stroke. 12:723-725.


Page 22 of 31

Page 23 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


25

Williamson MR, Dietrich K, Hackett MJ, Caine S, Nadeau CA, Aziz JR, Nichol H, Paterson PG, Colbourne F. (2017) Rehabilitation Augments Hematoma Clearance and Attenuates Oxidative Injury and Ion Dyshomeostasis After Brain Hemorrhage. Stroke. 48:195-203.

26

Hansen AJ, Zeuthen T. (1981) Extracellular ion concentrations during spreading depression and ischemia in the rat brain cortex. Acta Physiol Scand. 113:437-445.

27

Hurn PD, Koehler RC, Norris SE, Blizzard KK, Traystman RJ (1991) Dependence of cerebral energy phosphate and evoked potential recovery on end-ischemic pH. Am J Physiol . 260:H532-H541.

28

Pushie MJ, Pickering IJ, Martin GR, Tsutsui S, Jirik FR, George GN. (2011) Prion protein expression level alters regional copper, iron and zinc content in the mouse brain. Metallomics. 3:206-214.

29

Pushkar Y, Robison G, Sullivan B, Fu SX, Kohne M, Jiang W, Rohr S, Lai B, Marcus MA, Zakharova T, Zheng W. (2013) Aging results in copper accumulations in glial fibrillary acidic protein-positive cells in the subventricular zone. Aging Cell. 12:823-832.

30

Nahirney PC, Reeson P, Brown CE. (2016) Ultrastructural analysis of blood-brain barrier breakdown in the peri-infarct zone in young adult and aged mice. J Cereb Blood Flow Metab. 36:413-425.

31

Squire LR (Ed) (2008) Fundamental Neuroscience, Third Edition, Burlington, MA. Elsevier Inc.

32

Kitamura Y1, Iida Y, Abe J, Mifune M, Kasuya F, Ohta M, Igarashi K, Saito Y, Saji H. (2006) In vivo measurement of presynaptic Zn2+ release during forebrain ischemia in rats. Biol. Pharm. Bull. 29:821–823.



33

Kitamura Y, Iida Y, Abe J, Mifune M, Kasuya F, Ohta M, Igarashi K, Saito Y, Saji H. (2006) Release of vesicular Zn2+ in a rat transient middle cerebral artery occlusion model. Brain Res. Bull. 69:622–625.

34

Eliasson MJ, Huang Z, Ferrante RJ, Sasamata M, Molliver ME, Snyder SH, Moskowitz MA. (1999) Neuronal nitric oxide synthase activation and peroxynitrite formation in ischemic stroke linked to neural damage. J Neurosci. 19:5910-5918.

35

Reif DW & Simmons RD (1990) Nitric oxide mediates iron release from ferritin. Arch Biochem Biophys. 283:537-541

36

Hossmann KA (2008) Cerebral ischemia: models, methods and outcomes. Neuropharmacology. 55:257-270.

37

Koizumi J, Yoshida Y, Nakazawa T, Ooneda G. (1986) Experimental studies of ischemic brain edema. I. A new experimental model of cerebral embolism in which recirculation can introduced into the ischemic area. Jpn J Stroke 8:108.

38

Ashwal S, Tone B, Tian HR, Chong S, Obenaus A. (2007) Comparison of two neonatal ischemic injury models using magnetic resonance imaging. Pediatr Res. 61:9–14.

39

Carmichael ST (2005) Rodent models of focal stroke: size, mechanism, and purpose. NeuroRx. 2:396-409.

40

Hackett MJ, McQuillan JA, El-Assaad F, Aitken JB, Levina A, Cohen DD, Siegele R, Carter EA, Grau GE, Hunt NH, Lay PA. (2011) Chemical alterations to murine brain tissue induced by formalin fixation: implications for biospectroscopic imaging and mapping studies of disease pathogenesis. Analyst. 136:2941-2952.

41

M. J. Hackett, C. J. Britz, H. Nichol, P. G. Paterson, I. J. Pickering and G. N. George, ACS Chemical Neuroscience, 2015, 6, 226-238 23 ACS Paragon Plus Environment

Page 24 of 31

Page 25 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


42

Jones MWM, Hare DJ, James SA, de Jonge MD, McColl G. (2017) Radiation Dose Limits for Bioanalytical X-ray Fluorescence Microscopy. Anal Chem. 89:12168-12175.

43

George GN, Pickering IJ, Pushie MJ, Nienaber K, Hackett MJ, Ascone I, Hedman B, Hodgson KO, Aitken JB, Levina A, Glover C, Lay PA. (2012) X-ray-induced photo-chemistry and Xray absorption spectroscopy of biological samples. J Synchrotron Radiat. 19:875-886.

44

Crawford AM. Methodologies in XRF Cytometry (Thesis). University of Michigan, 2015.

45

Ward J, Marvin R, O’Halloran T, Jacobsen C, Vogt S. (2013) Rapid and Accurate Analysis of an X-Ray Fluorescence Microscopy Data Set through Gaussian Mixture-Based Soft Clustering Methods. Microscopy and Microanalysis, 19:1281-1289.



Figure Legends Figure 1. Scheme of ischemic penumbra depicting the conceptually distinct peri-infarct zones.

Figure 2. Coronal images of the post-stroke cortex from the midbrain region, 24-hours post-PT stroke. Neuroanatomical schematic reference and H&E section were generated from the same tissue section used in XFI analysis. cc – corpus callosum, CPu – caudate putamen, CTX – cortex, LF – lateral fissure, LV – lateral ventricles, V3 – third ventricle. Pixel resolution = 30 µm.

Figure 3. Regions of interest from Gaussian clustering with n=5 clusters, based on P, S, Cl, K, and Zn (the 5th region of interest, corresponding to tissue edges, is not shown). Histological specimen and elemental maps are shown for reference.

Figure 4. Regions of interest identified from XFI clustering overlaid on the H&E-stained tissue section and elemental maps at 24h post-stroke.

Figure 5. Average aerial concentrations for regions of interest from Gaussian clustering of 24h post-stroke specimens (n=5).

Graphical Abstract.


Page 26 of 31

Page 27 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figures Figure 1.

Figure 2.



Figure 3. H&E

P

S

Cl

K

Zn


Page 28 of 31

Page 29 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4.



Figure 5.


Page 30 of 31

Page 31 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Graphical Table of Contents Image


Revealing the Penumbra through Imaging ... - ACS Publications

Recommend Documents