G Model
ARTICLE IN PRESS
YSCDB 1697 1–10
Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Seminars in Cell & Developmental Biology journal homepage: www.elsevier.com/locate/semcdb
Review
1
Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice
2
3
4 5 6 7 8 9
Q1
Yoash Chassidim a,∗,1 , Udi Vazana a,1 , Ofer Prager a , Ronel Veksler a , Guy Bar-Klein a , Karl Schoknecht b , Michael Fassler a , Svetlana Lublinsky a , Ilan Shelef c
a Departments of Physiology & Cell Biology, Cognitive and Brain Sciences, The Zlotowski Center for Neuroscience, Ben-Gurion University of the Negev, Beer-Sheva, Israel Q2 b Department of Neurophysiology, Charite University of Medicine, Berlin, Germany c Medical Imaging Institute, Soroka Medical Center, Beer-Sheva, Israel
10
11 22
a r t i c l e
i n f o
a b s t r a c t
12 13 14
Article history: Available online xxx
15
21
Keywords: Blood–brain barrier Biomarkers Invasive/non-invasive pre-clinical imaging modalities Clinical imaging modalities
23
Contents
16 17 18 19 20
24 25 26 27 28 29 30
1. 2. 3.
A dysfunctional BBB is a common feature in a variety of brain disorders, a fact stressing the need for diagnostic tools designed to assess brain vessels’ permeability in space and time. Biological research has benefited over the years various means to analyze BBB integrity. The use of biomarkers for improper BBB functionality is abundant. Systemic administration of BBB impermeable tracers can both visualize brain regions characterized by BBB impairment, as well as lead to its quantification. Additionally, locating molecular, physiological content in regions from which it is restricted under normal BBB functionality undoubtedly indicates brain pathology-related BBB disruption. However, in-depth research into the BBB’s phenotype demands higher analytical complexity than functional vs. pathological BBB; criteria which biomarker based BBB permeability analyses do not meet. The involvement of accurate and engineering sciences in recent brain research, has led to improvements in the field, in the form of more accurate, sensitive imaging-based methods. Improvements in the spatiotemporal resolution of many imaging modalities and in image processing techniques, make up for the inadequacies of biomarker based analyses. In preclinical research, imaging approaches involving invasive procedures, enable microscopic evaluation of BBB integrity, and benefit high levels of sensitivity and accuracy. However, invasive techniques may alter normal physiological function, thus generating a modality-based impact on vessel’s permeability, which needs to be corrected for. Non-invasive approaches do not affect proper functionality of the inspected system, but lack in spatiotemporal resolution. Nevertheless, the benefit of medical imaging, even in preclinical phases, outweighs its disadvantages. The innovations in pre-clinical imaging and the development of novel processing techniques, have led to their implementation in clinical use as well. Specialized analyses of vessels’ permeability add valuable information to standard anatomical inspections which do not take the latter into consideration. © 2014 Published by Elsevier Ltd.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In vivo biomarkers for enhanced permeability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In vivo imaging and assessment of brain vessels’ permeability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Pre-clinical imaging modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. Direct optical imaging-BBB permeability assessment using fluorescence imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Direct optical imaging – laser-speckle imaging for BBB permeability assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Non-invasive techniques in pre-clinical imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00 00 00 00
Abbreviations: BBB, blood–brain barrier; EB, Evans-blue; PET, positron emission tomography; SPECT, single photon emission computerized tomography; MRI, magnetic resonance imaging; DCE, dynamic contrast enhanced; SNR, signal to noise ratio; Gd, gadolinium. ∗ Corresponding author. 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.semcdb.2014.11.007 1084-9521/© 2014 Published by Elsevier Ltd.
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
2
3.2.
4.
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 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
Clinical imaging modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. BBB imaging examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction The central nervous system (CNS) is highly affected by changes in its environment. To insure normal function of the neural system, transport of ions and molecules between the CNS and its supporting vascular system must be tightly regulated; keeping homeostasis within the neuropil. The concept of an anatomical separation between the blood and the brain arose in the late 19th century. A German bacteriologist named Paul Ehrlich observed that the intravenous administration of aniline dyes to small animals stained all organs but the brain. Ehrlich’s student, Goldman, continued these experiments and injected Trypan blue to the cerebrospinal fluid (CSF) of rabbits and dogs and demonstrated staining of the entire brain without a trace of the dye in the blood stream [1]. These experiments brought about the recognition of a tight barrier between blood and brain environments, known as the blood–brain barrier (BBB). The BBB is a complex structure and functional mechanism underlying the specialized isolation of the CNS from its supporting vascular system. It is formed at the level of endothelial cells comprising the lumen of blood vessels in the CNS [2,3]. Endothelial cells of brain blood vessels are connected by tight junction protein complexes and junction adhesion molecules; these protein structures restrict para-cellular passage of molecules and force most molecular traffic to take place in a trans-cellular manner. Lipophilic or small gaseous molecules can diffuse freely through the cellular membrane; otherwise molecular passage requires specific transport mechanisms within the membrane. Hence the term: “selective isolation”. Transport mechanisms include solute carriers for specific molecules, ATP binding cassette transporters, receptor-mediated and adsorptive-mediated transcytosis and cellular migration mechanisms [2]. All of these processes are tightly regulated by endothelial intra-cellular processes (e.g. gene translation). Additionally, regulation of the BBB phenotype extends beyond the endothelial cell. Brain blood vessels are innervated by neurons, astrocytes, pericytes/smooth muscle cells and microglia. These cell–cell interactions form the neurovascular unit and induce cellular processes that determine specific features of the BBB phenotype. A dysfunctional BBB is a feature of a variety of neurological disorders Such as traumatic brain injury, stroke, cancer, epilepsy and neurodegenerative diseases [4–7]. In stroke for instance, the damage caused to endothelial cells due to ischemia is suggested to result in formation of reactive oxygen species (ROS) leading to an abnormal ion flux, extravasation of proteins and subsequent brain edema [8]. Studies have also shown disruption of tight junction complexes in human gliomas and metastatic adenocarcinomas [5]. Additionally BBB breakdown has been shown to be associated with epilepsy either as a cause or as a consequence [6,9–11]. Seizures are observed in cases of brain insult such as traumatic brain injury and central nervous system infections, conditions known to result in compromised BBB [6,12,13]. Additionally, it has been shown that seizure activity results in BBB impairment [9]. The major role of BBB dysfunction in brain disease has raised the need for accurate and sensitive diagnostic tools that would assess the level of BBB permeability and provide information regarding degradation of brain tissue in pathology. This review aims to survey past and present diagnostic modalities, emphasizing present day imaging techniques employed in
00 00 00 00 00
pre-clinical research and in clinical use as well. We present here methods in optical, magnetic resonance and nuclear imaging, previously published and validated, both in-house as well as by others. These imaging platforms provide highly sensitive and reliable tools for BBB permeability assessment. 2. In vivo biomarkers for enhanced permeability A simple approach for detecting BBB disruptions in vivo is by post mortem visualization of BBB impermeable tracers within brain tissue. The tracers may be substances systemically administered to the anesthetized animal or normal plasma/brain content that is restricted from brain parenchyma/plasma respectively, when the BBB is intact. Methods, in clinical and pre-clinical use, range between those enabling qualitative assessment alone and those allowing quantitative analysis as well. Systemic administrations of non-BBB permeable tracers, possessing physical properties such as fluorescence, radioactivity, etc., that enable their detection, are commonly used in pre-clinical studies to asses BBB leakage through accumulation of tracer residues in extra vascular tissue. A qualitative macroscopic analysis exemplifying this approach can be found in Fig. 1A/B; following intravenous injection of Evans-blue (EB, 2% in 0.9% NaCl, 2.4 ml/kg), the animal is sacrificed in an open-heart surgical procedure, in which paraformaldehyde (PFA, 4% in phosphate buffered saline) is administered to the cardiovascular system. PFA fixates the brain, which is then extracted for further analysis. EB binds to the serum protein albumin and therefore is BBB impermeable [14]. Thus, blue stains in fixated brains and brain slices indicate local disruption to the BBB (Fig. 1A/B). For more sensitive evaluations, quantitative analysis is required. Given that EB-albumin is fluorescent, the use of spectrophotometers can be applied in order to measure fluorescence intensity and subsequent tracer concentration within brain parenchyma [15]; as was performed by Asahi et al. in attempt to measure BBB permeability in mice following induction of cerebral ischemia [16]. Briefly, extracted brain samples are frozen, homogenized in buffer solution and finally centrifuged. The supernatant is then excited at the appropriate wavelength and the emission is read. The ratio of emission to excitation light intensities can be correlated to substance concentration [17] and therefore the level of extra-vascular EB, reflecting BBB permeability level, can be evaluated. However, this simple calculation is not informative when the dynamic features of molecular passage through the BBB are to be assessed. For that end, there are approaches employing multicompartmental mathematical models in which unidirectional or bidirectional passage between compartments (representing vascular and extra-vascular regions) is applied under restricting factors [18]. Intra and extra-vascular tracer concentrations are evaluated using physical techniques as the one previously mentioned, and are placed in the model as input. Resolving the model benefits numerical constants reflective of tracer passage. Physiological markers for BBB damage are abundant. The focus of this approach has been primarily on proteins given the variety of mechanisms by which they penetrate the BBB. Detection of plasma proteins, normally restricted from the brain by the BBB, in brain parenchyma is a valuable tool for BBB dysfunction diagnosis. An example of such an approach is immunostaining against serum proteins done post mortem. Van Vliet et al. applied
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
88 89 90 91 92
93
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
3
Fig. 1. In vivo biomarkers for blood–brain barrier disruption. (A) The impact of thrombolytic stroke on BBB integrity is exemplified by Evans-blue extravasation (right) in extracted, fixated rat brains. Healthy brains do not exhibit these phenomena (left). (B) Evans-blue extravasation is observed in the course of status epilepticus in the rat cortex. The treated hemisphere (right) alone is affected. (C/D) Immunostained brain slices from the neocortex (top) and the hippocampus (bottom), of untreated control (left) and status epilepticus induced rats (right). BBB dysfunction is evident by albumin extravasation into brain parenchyma, noted as black dots (arrow).
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
this for detection of albumin in human and animal brains for the study of temporal lobe epilepsy progression due to BBB leakage [7]. Following its fixation and embedment in paraffin, brain tissue is sectioned to micron size slices. Slices are incubated with primary and secondary antibodies subsequently enabling microscopic examination and detection of proteins in question (Fig. 2C/D). A dominant disadvantage of this technique is the time point of its implementation-post mortem. Detecting BBB disruption in real time is impossible making this approach clinically irrelevant. An alternative approach aims to detect brain tissue proteins within plasma content, one of which is S100 [19,20]. This astrocytic protein’s levels increase in serum upon BBB damage, regardless of neuronal damage [19]; a fact making it preferable for analysis in comparison with other markers of the sort. S100 levels in serum are usually evaluated with ELIZA [21]. Although S100 detection in serum is applicable as a clinical diagnostic tool, and validated as an indicator for BBB impairment, there are some drawbacks in its regard. It is essentially an indirect method for analyzing the BBB given that it measures the consequences of BBB dysfunction in the periphery and can therefore be affected by peripheral processes. For instance, S100 is expressed in peripheral systems such as the
heart and kidney [22]. Thus it is plausible that S100 levels may increase in serum due to factors unrelated to BBB integrity. Although adding a quantitative analysis dimension, and enabling in vivo examination, non-imaging, biomarker detectionbased techniques do not meet high-end requirements which medical imaging, in all its forms, can satisfy. Online, high resolution, image acquisition sets a platform on which the immediate environment of brain vasculature can be examined. Offline, various image processing techniques provide a highly accurate quantitative approach for monitoring BBB integrity with time and between various conditions. 3. In vivo imaging and assessment of brain vessels’ permeability 3.1. Pre-clinical imaging modalities 3.1.1. Direct optical imaging-BBB permeability assessment using fluorescence imaging This direct vessel imaging approach [23] was developed inhouse and is considered our primary approach for detecting BBB
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
165 166 167 168 169 170 171 172 173 174 175
176 177
178
179 180 181 182
G Model YSCDB 1697 1–10 4
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
Fig. 2. Quantitative assessment of BBB permeability is validated by a well-established model for BBB dysfunction. (A) Fluorescence imaging of cerebral blood flow in an exposed rat cortical section under normal physiologic conditions. Sodium fluorescein is injected intravenously, thus achieving contrast enhancement of brain vasculature. (B) Classification of the fluorescence image into vascular (blue) and extra-vascular (red) regions. (C) Fluorescence intensity–time curves representing the primary artery (red) and extra-vascular space (black). The green arrow marks the time span taken for computation of BBB permeability. (D) Fluorescence imaging of cerebral blood flow in an exposed rat cortical section following exposure to deoxycholic acid (DOC) so to induce BBB disruption. Evident enhancement of the extra-vascular space reflects the latter. (E) BBB permeability enhancement is reflected as in intensification of the extra-vascular signal over the arterial one (DOC) in comparison to the state preceding DOC exposure (control). The permeability indicator is therefore increased between the two conditions. (F) Using each extra-vascular pixel’s intensity–time curve in permeability indicator calculations results in its value being assigned to each pixel and the subsequent formation of a permeability map. The enhancement in BBB permeability is reflected in the maps as color intensification in response to DOC exposure (right) compared to pre-exposure status (left). (G) The DOC effect reflected using direct vessel imaging is statistically significant (p < 0.05) in comparison to both pre-exposure status (n = 6, p = 0.03, Wilcoxon) and to unexposed animals (DOC vs. control, n = 6 vs n = 16, p < 0.001, Mann–Whitney).
183 184 185 186 187 188 189
190 191 192 193 194 195 196
disruption events in vivo. Its primary advantage is the ability to semi-quantify BBB permeability in anesthetized animals. Its validation was achieved in 2010 when its application enabled the detection and quantification of major BBB dysfunction following the use of deoxycholic acid (DOC, [24]) or arterial stroke applied with photothrombosis [60]. The procedure consists of 3 stages: (1) Craniotomy: Animals are deeply anesthetized by intraperitoneal (IP) injection of ketamine (100 mg/ml, 0.08 ml/100 g) and xylasine (20 mg/ml, 0.06 mg/100 g). The tail vein is catheterized and the animal is placed in a stereotactic frame under a fluorescence microscope (Zeiss Lumar). A mid-sagittal incision is made and the top of the cranium is exposed following removal of soft tissue. A 2 mm × 4 mm cranial section is removed by drilling
into the bone. Removal of the 3 meningeal layers results in exposure of a neocortical section with its supporting vascular system. The exposed section is then sealed by a ring-like structure, constructed from bone-wax (Ethicon Ltd.) and dental cement (Unifast Ltd.), and is continuously perfused with artificial cerebrospinal fluid (aCSF) containing in mM: 129 NaCl, 21 NaHCO3 , 1.25 NaH2 PO4 , 1.8 MgSO4, 1.6 CaCl2 , 3 KCl and 10 glucose (pH 7.4). (2) Dynamic contrast enhanced fluorescence imaging: The basis of quantifying BBB permeability is tracking the location of a systemically administered tracer, within the exposed cortical section. The tracer possesses chemical features that render it unable to penetrate the BBB. Locating and measuring tracer amounts in an identified extra-vascular region, indicates BBB disruption. For the purposes of this review we demonstrate the
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
use of sodium-fluorescein (NaFlu, 0.2 ml/injection) as a tracer. NaFlu is fluorescent by its fluorescein component. Fluorescein emits radiation at ∼540 nm when stimulated at ∼470 nm. BBB restriction results from the binding to Na ions, making the endproduct highly soluble in aqueous mediums. NaFlu is injected to the tail vein, while the exposed cortex is constantly excited with 470 nm light. Using an EMCCD camera (Andor Technology, DL-658 M-TIL) we acquire images at 5 fr/s, and obtain a dynamic footage displaying regional cerebral blood flow (Fig. 2A). (3) Computational analysis: Initially, the dynamic footage is registered in order to overcome artifacts caused by small movements [25]. The registration procedure is applied between each given frame and a moving average of its preceding frames in order to diminish the effects of noise in the time dimension. Next, the footage is classified to sections: vascular and extra vascular regions, by using a process of threshold segmentation [26]. Briefly, a reference image of maximal signal/noise ratio is created by averaging several frames in the short time span of tracer entry into the exposed cortical section. A spatial moving average is then performed with a 3 × 3 kernel. The post-processing image and the original image are then compared. Pixels, whose values are elevated above a predetermined threshold as a result of averaging, are considered extravascular. The end result is a binary image according to the vascular/extra vascular criterion (Fig. 2B). This classified image is used as a mask, enabling the evaluation of fluorescence intensities at specific regions in the exposed cortex. According to known anatomical features (width, length) the primary artery is selected from the footage as a region of interest (ROI). Fluorescent intensities in this region alone are then averaged at each time point, yielding an intensity–time curve (IT, Fig. 2C) representative of the primary artery, also termed the “arterial input function” (AIF). This function indicates tracer input levels into the inspected cortical section. An identical process is performed regarding the entire extra-vascular space, yielding an IT curve indicative of tracer residues left in the inspected section. Comparing the two functions yields a parameter reflective of BBB permeability – a permeability indicator (PI): The ratio function between normalized fluorescence intensities of the extra-vascular and arterial IT curves is calculated: (IEV /IAIF )(t). The ratio is averaged in the time span ranging from the critical time point-tcr the start point of the second decline phase(arrow, Fig. 2C), till end of measurement cycle point: t (1/T ) t end (IEV /IAIF )(t)dt T = tend − tcr . cr
This particular time span is chosen for analysis since it represents the tracer’s clearance phase – a period in which intra-vascular tracer is washed out of the brain and tracer residues become apparent. BBB disruptions are visually detected in this time period. By performing a similar analysis with each extra-vascular pixel’s IT curve instead of the one representing the entire extra-vascular space, yields a PI values for each extra-vascular pixel and the formation of a permeability map (Fig. 2F). Such maps represent BBB integrity level in each location of the inspected cortical section, based on color scale. An example for this approach’s ability for detecting and quantitatively assessing alterations in BBB phenotype can be found in Fig. 2D–G DOC (2 mM, diluted in aCSF). Perfusions of the exposed cortex result in leakage of NaFlu into extra-vascular space (Fig. 2D). This leakage is quantified as intensification of the extra-vascular fluorescence signal over the arterial one (Fig. 2E), which results in significantly higher values of PI (Fig. 2G). DOC-induced BBB dysfunction occurs in the entire exposed cortical section, as evident by global color intensification in permeability mapping (Fig. 2F).
5
3.1.2. Direct optical imaging – laser-speckle imaging for BBB permeability assessment In recent time, laser speckle contrast imaging (LSCI) has been widely used to measure blood flow in a variety of tissues. The physical concept of this approach lies in the fact that when coherent laser illuminates an object the beam is scattered, and adds constructively as well as destructively. A random interference pattern is therefore formed and can be visualized with appropriate optical receptors [27]. When the laser illuminates moving objects (e.g. blood cells) the pattern fluctuates with time; its integration results in reduced spatial contrast in areas of increased motion. Thus, motion of moving particles is encoded in the blurring of the time integrated speckle pattern. Quantification of motion is achieved by mathematically analyzing the spatial blurring [27,28]. A recent study [29] employed laser speckle contrast imaging to assess BBB permeability, constituting the first demonstration of LSCI in that aspect. The basis for this modality is that BBB disruption can alter cerebral blood flow [23]. LSCI was performed on an exposed rat cortical section together with fluorescence tracking of a systemically administered non-BBB permeable dye. Chemically induced BBB disruption, noted and localized by fluorescence imaging, resulted in arterial vasodilation accompanied by a net reduction in venous flow velocity. Thus, suggesting a different impact of BBB disruption on flow speeds of arteries and veins. The ratio of output/input flow velocity profiles in the inspected cortical section was eventually chosen as an indicative parameter for BBB disruption, whose drop in values following BBB disruption was found to be significant. 3.1.3. Non-invasive techniques in pre-clinical imaging The direct vessel imaging approach, in what is perhaps its greatest advantage, allows for high spatial resolution (down to 10 m, depending on the imaging equipment). The immediate environment of brain vasculature is examined. However, this is achieved by way of craniotomy; a cortical section is exposed thus compromising the integrity of its supporting vasculature due to a decline in intra-cranial pressure [30]. Therefore, when examining the effects of various treatment approaches on BBB functionality, additional assessments must be performed in order to rule out surgically induced damage to vessel integrity and subsequent harm to BBB functionality. An alternative approach involves the use of non-invasive techniques in which larger scale brain regions are imaged. The modality has no bearing on vessel integrity/permeability; however this is achieved at the price of low spatial and temporal resolution (∼1 mm and ∼1 min respectively). Non-invasive imaging based BBB permeability assessment is commonly done with near infrared fluorescence (NIRF), magnetic resonance imaging (MRI), positron emission tomography (PET), single photon emission tomography (SPECT) and X-ray computerized tomography (CT). A non-BBB permeable reactive tracer is systemically administered, generating image enhancement. Various offline image processing techniques have been developed to assess tracer kinetics over time and space, and translate this enhancement into numerical parameters representing vascular features; among which is BBB permeability. In vivo, non-invasive fluorescence imaging is quite limited given the complexity of photon penetration through tissues. Tissues are heterogeneous surfaces with different refractive indices and a photon traveling to reach a fluorescent contrast agent may be reflected back to its source. The same is true for photons traveling from a fluorophore in attempt to reach a receiver unit [31]. Additionally, tissues may emit light upon excitation even without the presence of a contrast agent; a phenomenon known as “auto-fluorescence”, that may harm the signal to background ratio [31]. However, excitation at NIR wavelengths (>700 nm) can reduce these shortcomings. Water and lipids tend to absorb NIR photons [32] and are nonfluorescent at this excitation [31]; allowing photon propagation
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
G Model YSCDB 1697 1–10 6 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
and high signal to background ratio. The use of NIRF has been implemented in BBB permeability analysis as well. Klohs et al. describe their attempt to detect fluorescently labeled albumin in mice brains [33]. The study has demonstrated extravasation of albumin into brain parenchyma induced by stroke applied with the middle cerebral artery occlusion model [34]. Higher fluorescent intensities were measured in the treated hemisphere in comparison to the non-treated one. These results were correlated to EB extravasation evaluated ex vivo. Nevertheless, NIRF imaging has some limitations. Its sensitivity is limited by contrast agent’s depth and background fluorescence [33]. In the described study the differences in fluorescence intensities between hemispheres were stronger in ex vivo examinations than in the parallel non-invasive procedure. This was due to high fluorescence of the skin over the non-treated hemisphere due to injury. Contrast enhanced nuclear imaging (CENI) can quantitatively assess cellular processes. For instance, PET with the administration of 18 F-2-fluoro-2-deoxy-d-gluocose (18 F-FDG) is widely used to quantify changes in glucose metabolism [35,36]. Glucose enters brain parenchyma through glucose transporter 1 (GLUT1) receptors expressed in brain vascular endothelial cells [3]. Thus altered glucose uptake could serve as a marker indicating modulations of the BBB’s molecular structure. Yang et al. [37] demonstrated the effects of focused ultrasound on glucose metabolism in anesthetized rats using PET combined with 18 F-FDG administration following sonication. Disruption of the BBB, induced by sonication was observed by spectrophotometer readings of EB staining in brain slices, and was significantly reduced following 24 h. A 2-compartmental model was used for quantification of 18 F-FDG uptake into the brain. Interestingly, the study presented that glucose uptake was reduced immediately following sonication in comparison to levels found in untreated controls; and was returned to control level post 24 h. These findings were corroborated by reduced expression of GLUT1 found in treated brains. This study remarkably indicates that while the use of a biomarker in non-imaging based analysis techniques indicated BBB disruption, medical imaging indicated phenotype modulations which are in contradiction to the latter. This raises important questions regarding BBB selectivity even in a state of improper functionality. It exemplifies how the use of medical imaging can improve the resolution in which we explore the BBB. Additionally, CENI is used to investigate efflux mechanisms in the BBB as well as CNS inflammation. Efflux pumps and drug transporters belonging to the ATP-binding cassette family of transporters have become another target to study selective transport mechanisms at the BBB. P-glycoprotein (p-gp), a transporter comprising the BBB’s phenotype, mediates the efflux of xenobiotics [2]. PET Molecular imaging, non-invasively detecting p-gp function, can identify p-gp overexpression at the BBB which may underlie pharmaco-resistance. Bartmann et al. exemplify such an approach [38] by administering the p-gp radiolabeled substrate 2 -methoxyphenyl-(N-2 -pyridinyl)p-fluoro-benzamidoethyipiperazine ([18 F]MPPF) to anesthetized, epileptic rats. The study exhibited a correlation between ([18 F]MPPF) kinetics and responsiveness to anti-epileptic treatment. CNS inflammation is featured by increased vessels’ permeability, leading to infiltration of peripheral immune cells (lymphocytes, granulocytes, neutrophils, monocytes/macrophages), as well as activation of microglia. By radiolabeling autologous leukocytes, this process can be monitored. Tracers in this approach include 111 In-oxine and 111 In-troponolate [35]. MRI in combination of MR-visible contrast agents containing gadolinium [39–41] is considered to be the gold standard in assessment of BBB impairment using non-invasive imaging modalities. Advantages of MRI in terms of safety and spatial resolution, compared to other imaging modalities such as PET [39], contribute to the latter. MRI generates high-resolution information of brain
anatomy and cerebral blood flow; thus setting a platform on which BBB integrity could be assessed (Fig. 2A/B). For instance, T2 and diffusion-weighted scans may identify ischemic brain regions [42], which may also be characterized by enhanced vascular permeability [8,43]. Protocols for BBB permeability quantification in MRI include static contrast enhanced (SCE) imaging, in which statistical differences between scans are calculated; and dynamic contrast enhanced imaging (DCE) where alterations in tracer concentration are monitored in repeated scans. A practical approach for SCE-imaging based BBB integrity evaluation was developed in-house, and is termed the “pre-post” comparison [41]. Briefly, T1-weighted scans are performed prior to and 5 min following systemic gadopentetic acid (Gd-DTPA) administration; making the second scan contrast-enhanced in regions where tracer is accumulated due to BBB dysfunction (Fig. 3C). For quantification purposes, a comparison is done between the scans in several steps: First, a neighborhood of 3 × 3 is selected around each pixel (9 pixels) in both pre and post images. A slice-wise unpaired ttest is then performed between each two parallel neighborhoods in the pre and post images. Pixels indicating accumulation of Gd-DTPA are identified as those whose neighborhood comparison benefited statistically significant results. Each pixel is therefore assigned a P-value, resulting in a binary significance image in which all pixels with scores under a predetermined threshold are assigned the value “1” and the rest, the value “0”. The second step involves the calculations of enhancement differences between pre and post images. Enhancement distribution histograms are calculated for three representing regions: muscle, eyeball and blood vessel in order to determine the BBB breakdown enhancement range for each subject. The muscle represents a “permeable tissue” with no BBB, while the eyeball represents a region with intact BBB (that isolates the retina from its supporting vasculature). The blood vessel region is used to exclude vascular tissue. Fitting a Gaussian to the muscle histogram and estimating the average and standard deviation enables the estimation of an enhancement range indicative of BBB breakdown. Pixels indicating a leaky BBB are therefore those whose statistical analysis indicates a statistically significant intensity increase, and whose post injection intensities are within the enhancement range. The third step is clustering of potentially permeable pixels. Neighboring pixels are aggregated into anatomically connected regions, so to establish a more robust and clinically relevant assessment (Fig. 3D). Advances in non-invasive imaging techniques in general, and in MRI in particular, in terms of safety and temporal/spatial resolution; have led to the development of novel protocols for BBB integrity assessment, and their enrollment in clinical research as well as in every-day use. 3.2. Clinical imaging modalities The advances made in pre-clinical BBB imaging modalities have raised the question of their implementation in clinical practice as well. Imaging of the BBB in human patients has been extensively studied over the last two decades. Although SPECT and CT can be used to evaluate BBB integrity, MRI is the modality-of-choice in most cases. Similarly to the animal model case, contrast-enhanced MRI (CE-MRI) sequences are used for BBB imaging in human, where Gd-based contrast agents (CA) that do not cross the intact BBB are used. 3.2.1. Analytical methods Assessment of BBB integrity can be carried out by means of qualitative, semi-quantitative or quantitative methods. Qualitative methods are very simple in terms of scanning parameters and computational costs. However, they can come short in terms of sensitivity and do not provide information regarding the permeability
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
454
455 456 457 458 459 460 461 462 463
464 465 466 467 468 469
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
7
Fig. 3. MRI analysis for assessment of BBB permeability. (A) In T2-weighted MRI, brain areas suspected of BBB dysfunction-related pathology (e.g. oedeme) appear to have high signal intensity (overlaid in red) in relation to healthy brain tissue. These regions are defined as “lesions”. (B) Analysis of variability in T2-weighted image intensity benefits lesion identification. A mixture of 3 Gaussian probability density functions is proposed to model variability in brain tissue intensities. The cut off threshold, indicating a lesion, is defined as an intersection between the 2d and 3d Gaussians (marked by an arrow). In order to remove small disconnected noisy speckles, a clustering procedure is applied. To this end, neighboring pixels’ connectivity (4-connected) is checked. Clusters of pixels, smaller in size than a minimal area are defined as noisy regions and are excluded from the lesion object. (C) T1 weighted scans are used to quantify BBB dysfunction (overlaid in blue). The quantification approach is based on a comparative analysis between contrast-enhanced MR scans before and 20 min after contrast agent injection. (D) Percentage value of enhancement differences between the pre and post contrast images. Significantly different pixels within an enhancement range of 30–100% are considered to represent brain tissue with potential leaky BBB. The described above “clustering procedure” was applied in order to remove noisy speckles.
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507
values in the affected areas. These methods include comparison of pre- and post-CA injection images. Another example is the delayed contrast extravasation subtraction [44], which utilizes 3 scans with a long delay between the first and third scan (75 min). This method has the advantage of enabling scans with high spatial resolution due to the fact that the acquisition speed is not a factor. It has been claimed that this method can be used to differentiate tumor progression from tissue damage induced by treatment. Despite their shortcomings, qualitative imaging methods may be clinically useful. Semi-quantitative and quantitative methods rely on dynamic contrast-enhanced (DCE-MRI) acquisition protocols, where images are acquired before, during and after the CA injection. The acquisition parameters depend on the analysis method, and are under debate, since a natural tradeoff exists between spatial/temporal resolution/volume coverage. Semi-quantitative approaches make use of measurements of MRI signal over time, and are calculated either from image intensity’s dynamic values or from concentration curves. The measurements include area under curve (AUC), initial area under curve (iAUC), time to peak (TTP), wash-in and wash-out slopes. These metrics arise from a combination of physiological parameters such as blood flow, blood volume, and permeability rather than imply a single parameter. Nevertheless, they can be very useful in clinical settings as they are less demanding in terms of acquisition parameters, especially regarding temporal resolution. Additionally, slower scans result in higher signal to noise ratio (SNR) and allow higher spatial resolution to be obtained. Quantitative methods use pharmacokinetic models with variable number of parameters to infer on physiological properties of different tissues. These models require the calculation of CA concentration–time curves (C(t)) in each voxel. To allow the conversion from intensity values to concentration, additional sequences are required (e.g. [45]) Full description of the common models that are used to fit the curves and the hierarchy between them have been recently reviewed by Sourbron and Buckly [46]. Imaging acquisition protocols for quantitative method are a matter of debate, and although no standard procedure has been established, the
requirements from suitable protocol include: volume coverage, spatial resolution, temporal resolution, length of scan and artifact minimization. Covered volume should usually include the whole brain, unless only a certain region (e.g. tumor region) is the area of interest. Temporal resolution has a direct impact on the accuracy of parameter estimation and the models differ in their sensitivity to the sampling interval [47–52]. Although several works have recommended a maximal sampling interval of
ARTICLE IN PRESS
YSCDB 1697 1–10
Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Seminars in Cell & Developmental Biology journal homepage: www.elsevier.com/locate/semcdb
Review
1
Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice
2
3
4 5 6 7 8 9
Q1
Yoash Chassidim a,∗,1 , Udi Vazana a,1 , Ofer Prager a , Ronel Veksler a , Guy Bar-Klein a , Karl Schoknecht b , Michael Fassler a , Svetlana Lublinsky a , Ilan Shelef c
a Departments of Physiology & Cell Biology, Cognitive and Brain Sciences, The Zlotowski Center for Neuroscience, Ben-Gurion University of the Negev, Beer-Sheva, Israel Q2 b Department of Neurophysiology, Charite University of Medicine, Berlin, Germany c Medical Imaging Institute, Soroka Medical Center, Beer-Sheva, Israel
10
11 22
a r t i c l e
i n f o
a b s t r a c t
12 13 14
Article history: Available online xxx
15
21
Keywords: Blood–brain barrier Biomarkers Invasive/non-invasive pre-clinical imaging modalities Clinical imaging modalities
23
Contents
16 17 18 19 20
24 25 26 27 28 29 30
1. 2. 3.
A dysfunctional BBB is a common feature in a variety of brain disorders, a fact stressing the need for diagnostic tools designed to assess brain vessels’ permeability in space and time. Biological research has benefited over the years various means to analyze BBB integrity. The use of biomarkers for improper BBB functionality is abundant. Systemic administration of BBB impermeable tracers can both visualize brain regions characterized by BBB impairment, as well as lead to its quantification. Additionally, locating molecular, physiological content in regions from which it is restricted under normal BBB functionality undoubtedly indicates brain pathology-related BBB disruption. However, in-depth research into the BBB’s phenotype demands higher analytical complexity than functional vs. pathological BBB; criteria which biomarker based BBB permeability analyses do not meet. The involvement of accurate and engineering sciences in recent brain research, has led to improvements in the field, in the form of more accurate, sensitive imaging-based methods. Improvements in the spatiotemporal resolution of many imaging modalities and in image processing techniques, make up for the inadequacies of biomarker based analyses. In preclinical research, imaging approaches involving invasive procedures, enable microscopic evaluation of BBB integrity, and benefit high levels of sensitivity and accuracy. However, invasive techniques may alter normal physiological function, thus generating a modality-based impact on vessel’s permeability, which needs to be corrected for. Non-invasive approaches do not affect proper functionality of the inspected system, but lack in spatiotemporal resolution. Nevertheless, the benefit of medical imaging, even in preclinical phases, outweighs its disadvantages. The innovations in pre-clinical imaging and the development of novel processing techniques, have led to their implementation in clinical use as well. Specialized analyses of vessels’ permeability add valuable information to standard anatomical inspections which do not take the latter into consideration. © 2014 Published by Elsevier Ltd.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In vivo biomarkers for enhanced permeability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In vivo imaging and assessment of brain vessels’ permeability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Pre-clinical imaging modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. Direct optical imaging-BBB permeability assessment using fluorescence imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Direct optical imaging – laser-speckle imaging for BBB permeability assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Non-invasive techniques in pre-clinical imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00 00 00 00
Abbreviations: BBB, blood–brain barrier; EB, Evans-blue; PET, positron emission tomography; SPECT, single photon emission computerized tomography; MRI, magnetic resonance imaging; DCE, dynamic contrast enhanced; SNR, signal to noise ratio; Gd, gadolinium. ∗ Corresponding author. 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.semcdb.2014.11.007 1084-9521/© 2014 Published by Elsevier Ltd.
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
2
3.2.
4.
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 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
Clinical imaging modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. BBB imaging examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction The central nervous system (CNS) is highly affected by changes in its environment. To insure normal function of the neural system, transport of ions and molecules between the CNS and its supporting vascular system must be tightly regulated; keeping homeostasis within the neuropil. The concept of an anatomical separation between the blood and the brain arose in the late 19th century. A German bacteriologist named Paul Ehrlich observed that the intravenous administration of aniline dyes to small animals stained all organs but the brain. Ehrlich’s student, Goldman, continued these experiments and injected Trypan blue to the cerebrospinal fluid (CSF) of rabbits and dogs and demonstrated staining of the entire brain without a trace of the dye in the blood stream [1]. These experiments brought about the recognition of a tight barrier between blood and brain environments, known as the blood–brain barrier (BBB). The BBB is a complex structure and functional mechanism underlying the specialized isolation of the CNS from its supporting vascular system. It is formed at the level of endothelial cells comprising the lumen of blood vessels in the CNS [2,3]. Endothelial cells of brain blood vessels are connected by tight junction protein complexes and junction adhesion molecules; these protein structures restrict para-cellular passage of molecules and force most molecular traffic to take place in a trans-cellular manner. Lipophilic or small gaseous molecules can diffuse freely through the cellular membrane; otherwise molecular passage requires specific transport mechanisms within the membrane. Hence the term: “selective isolation”. Transport mechanisms include solute carriers for specific molecules, ATP binding cassette transporters, receptor-mediated and adsorptive-mediated transcytosis and cellular migration mechanisms [2]. All of these processes are tightly regulated by endothelial intra-cellular processes (e.g. gene translation). Additionally, regulation of the BBB phenotype extends beyond the endothelial cell. Brain blood vessels are innervated by neurons, astrocytes, pericytes/smooth muscle cells and microglia. These cell–cell interactions form the neurovascular unit and induce cellular processes that determine specific features of the BBB phenotype. A dysfunctional BBB is a feature of a variety of neurological disorders Such as traumatic brain injury, stroke, cancer, epilepsy and neurodegenerative diseases [4–7]. In stroke for instance, the damage caused to endothelial cells due to ischemia is suggested to result in formation of reactive oxygen species (ROS) leading to an abnormal ion flux, extravasation of proteins and subsequent brain edema [8]. Studies have also shown disruption of tight junction complexes in human gliomas and metastatic adenocarcinomas [5]. Additionally BBB breakdown has been shown to be associated with epilepsy either as a cause or as a consequence [6,9–11]. Seizures are observed in cases of brain insult such as traumatic brain injury and central nervous system infections, conditions known to result in compromised BBB [6,12,13]. Additionally, it has been shown that seizure activity results in BBB impairment [9]. The major role of BBB dysfunction in brain disease has raised the need for accurate and sensitive diagnostic tools that would assess the level of BBB permeability and provide information regarding degradation of brain tissue in pathology. This review aims to survey past and present diagnostic modalities, emphasizing present day imaging techniques employed in
00 00 00 00 00
pre-clinical research and in clinical use as well. We present here methods in optical, magnetic resonance and nuclear imaging, previously published and validated, both in-house as well as by others. These imaging platforms provide highly sensitive and reliable tools for BBB permeability assessment. 2. In vivo biomarkers for enhanced permeability A simple approach for detecting BBB disruptions in vivo is by post mortem visualization of BBB impermeable tracers within brain tissue. The tracers may be substances systemically administered to the anesthetized animal or normal plasma/brain content that is restricted from brain parenchyma/plasma respectively, when the BBB is intact. Methods, in clinical and pre-clinical use, range between those enabling qualitative assessment alone and those allowing quantitative analysis as well. Systemic administrations of non-BBB permeable tracers, possessing physical properties such as fluorescence, radioactivity, etc., that enable their detection, are commonly used in pre-clinical studies to asses BBB leakage through accumulation of tracer residues in extra vascular tissue. A qualitative macroscopic analysis exemplifying this approach can be found in Fig. 1A/B; following intravenous injection of Evans-blue (EB, 2% in 0.9% NaCl, 2.4 ml/kg), the animal is sacrificed in an open-heart surgical procedure, in which paraformaldehyde (PFA, 4% in phosphate buffered saline) is administered to the cardiovascular system. PFA fixates the brain, which is then extracted for further analysis. EB binds to the serum protein albumin and therefore is BBB impermeable [14]. Thus, blue stains in fixated brains and brain slices indicate local disruption to the BBB (Fig. 1A/B). For more sensitive evaluations, quantitative analysis is required. Given that EB-albumin is fluorescent, the use of spectrophotometers can be applied in order to measure fluorescence intensity and subsequent tracer concentration within brain parenchyma [15]; as was performed by Asahi et al. in attempt to measure BBB permeability in mice following induction of cerebral ischemia [16]. Briefly, extracted brain samples are frozen, homogenized in buffer solution and finally centrifuged. The supernatant is then excited at the appropriate wavelength and the emission is read. The ratio of emission to excitation light intensities can be correlated to substance concentration [17] and therefore the level of extra-vascular EB, reflecting BBB permeability level, can be evaluated. However, this simple calculation is not informative when the dynamic features of molecular passage through the BBB are to be assessed. For that end, there are approaches employing multicompartmental mathematical models in which unidirectional or bidirectional passage between compartments (representing vascular and extra-vascular regions) is applied under restricting factors [18]. Intra and extra-vascular tracer concentrations are evaluated using physical techniques as the one previously mentioned, and are placed in the model as input. Resolving the model benefits numerical constants reflective of tracer passage. Physiological markers for BBB damage are abundant. The focus of this approach has been primarily on proteins given the variety of mechanisms by which they penetrate the BBB. Detection of plasma proteins, normally restricted from the brain by the BBB, in brain parenchyma is a valuable tool for BBB dysfunction diagnosis. An example of such an approach is immunostaining against serum proteins done post mortem. Van Vliet et al. applied
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
88 89 90 91 92
93
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
3
Fig. 1. In vivo biomarkers for blood–brain barrier disruption. (A) The impact of thrombolytic stroke on BBB integrity is exemplified by Evans-blue extravasation (right) in extracted, fixated rat brains. Healthy brains do not exhibit these phenomena (left). (B) Evans-blue extravasation is observed in the course of status epilepticus in the rat cortex. The treated hemisphere (right) alone is affected. (C/D) Immunostained brain slices from the neocortex (top) and the hippocampus (bottom), of untreated control (left) and status epilepticus induced rats (right). BBB dysfunction is evident by albumin extravasation into brain parenchyma, noted as black dots (arrow).
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
this for detection of albumin in human and animal brains for the study of temporal lobe epilepsy progression due to BBB leakage [7]. Following its fixation and embedment in paraffin, brain tissue is sectioned to micron size slices. Slices are incubated with primary and secondary antibodies subsequently enabling microscopic examination and detection of proteins in question (Fig. 2C/D). A dominant disadvantage of this technique is the time point of its implementation-post mortem. Detecting BBB disruption in real time is impossible making this approach clinically irrelevant. An alternative approach aims to detect brain tissue proteins within plasma content, one of which is S100 [19,20]. This astrocytic protein’s levels increase in serum upon BBB damage, regardless of neuronal damage [19]; a fact making it preferable for analysis in comparison with other markers of the sort. S100 levels in serum are usually evaluated with ELIZA [21]. Although S100 detection in serum is applicable as a clinical diagnostic tool, and validated as an indicator for BBB impairment, there are some drawbacks in its regard. It is essentially an indirect method for analyzing the BBB given that it measures the consequences of BBB dysfunction in the periphery and can therefore be affected by peripheral processes. For instance, S100 is expressed in peripheral systems such as the
heart and kidney [22]. Thus it is plausible that S100 levels may increase in serum due to factors unrelated to BBB integrity. Although adding a quantitative analysis dimension, and enabling in vivo examination, non-imaging, biomarker detectionbased techniques do not meet high-end requirements which medical imaging, in all its forms, can satisfy. Online, high resolution, image acquisition sets a platform on which the immediate environment of brain vasculature can be examined. Offline, various image processing techniques provide a highly accurate quantitative approach for monitoring BBB integrity with time and between various conditions. 3. In vivo imaging and assessment of brain vessels’ permeability 3.1. Pre-clinical imaging modalities 3.1.1. Direct optical imaging-BBB permeability assessment using fluorescence imaging This direct vessel imaging approach [23] was developed inhouse and is considered our primary approach for detecting BBB
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
165 166 167 168 169 170 171 172 173 174 175
176 177
178
179 180 181 182
G Model YSCDB 1697 1–10 4
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
Fig. 2. Quantitative assessment of BBB permeability is validated by a well-established model for BBB dysfunction. (A) Fluorescence imaging of cerebral blood flow in an exposed rat cortical section under normal physiologic conditions. Sodium fluorescein is injected intravenously, thus achieving contrast enhancement of brain vasculature. (B) Classification of the fluorescence image into vascular (blue) and extra-vascular (red) regions. (C) Fluorescence intensity–time curves representing the primary artery (red) and extra-vascular space (black). The green arrow marks the time span taken for computation of BBB permeability. (D) Fluorescence imaging of cerebral blood flow in an exposed rat cortical section following exposure to deoxycholic acid (DOC) so to induce BBB disruption. Evident enhancement of the extra-vascular space reflects the latter. (E) BBB permeability enhancement is reflected as in intensification of the extra-vascular signal over the arterial one (DOC) in comparison to the state preceding DOC exposure (control). The permeability indicator is therefore increased between the two conditions. (F) Using each extra-vascular pixel’s intensity–time curve in permeability indicator calculations results in its value being assigned to each pixel and the subsequent formation of a permeability map. The enhancement in BBB permeability is reflected in the maps as color intensification in response to DOC exposure (right) compared to pre-exposure status (left). (G) The DOC effect reflected using direct vessel imaging is statistically significant (p < 0.05) in comparison to both pre-exposure status (n = 6, p = 0.03, Wilcoxon) and to unexposed animals (DOC vs. control, n = 6 vs n = 16, p < 0.001, Mann–Whitney).
183 184 185 186 187 188 189
190 191 192 193 194 195 196
disruption events in vivo. Its primary advantage is the ability to semi-quantify BBB permeability in anesthetized animals. Its validation was achieved in 2010 when its application enabled the detection and quantification of major BBB dysfunction following the use of deoxycholic acid (DOC, [24]) or arterial stroke applied with photothrombosis [60]. The procedure consists of 3 stages: (1) Craniotomy: Animals are deeply anesthetized by intraperitoneal (IP) injection of ketamine (100 mg/ml, 0.08 ml/100 g) and xylasine (20 mg/ml, 0.06 mg/100 g). The tail vein is catheterized and the animal is placed in a stereotactic frame under a fluorescence microscope (Zeiss Lumar). A mid-sagittal incision is made and the top of the cranium is exposed following removal of soft tissue. A 2 mm × 4 mm cranial section is removed by drilling
into the bone. Removal of the 3 meningeal layers results in exposure of a neocortical section with its supporting vascular system. The exposed section is then sealed by a ring-like structure, constructed from bone-wax (Ethicon Ltd.) and dental cement (Unifast Ltd.), and is continuously perfused with artificial cerebrospinal fluid (aCSF) containing in mM: 129 NaCl, 21 NaHCO3 , 1.25 NaH2 PO4 , 1.8 MgSO4, 1.6 CaCl2 , 3 KCl and 10 glucose (pH 7.4). (2) Dynamic contrast enhanced fluorescence imaging: The basis of quantifying BBB permeability is tracking the location of a systemically administered tracer, within the exposed cortical section. The tracer possesses chemical features that render it unable to penetrate the BBB. Locating and measuring tracer amounts in an identified extra-vascular region, indicates BBB disruption. For the purposes of this review we demonstrate the
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
use of sodium-fluorescein (NaFlu, 0.2 ml/injection) as a tracer. NaFlu is fluorescent by its fluorescein component. Fluorescein emits radiation at ∼540 nm when stimulated at ∼470 nm. BBB restriction results from the binding to Na ions, making the endproduct highly soluble in aqueous mediums. NaFlu is injected to the tail vein, while the exposed cortex is constantly excited with 470 nm light. Using an EMCCD camera (Andor Technology, DL-658 M-TIL) we acquire images at 5 fr/s, and obtain a dynamic footage displaying regional cerebral blood flow (Fig. 2A). (3) Computational analysis: Initially, the dynamic footage is registered in order to overcome artifacts caused by small movements [25]. The registration procedure is applied between each given frame and a moving average of its preceding frames in order to diminish the effects of noise in the time dimension. Next, the footage is classified to sections: vascular and extra vascular regions, by using a process of threshold segmentation [26]. Briefly, a reference image of maximal signal/noise ratio is created by averaging several frames in the short time span of tracer entry into the exposed cortical section. A spatial moving average is then performed with a 3 × 3 kernel. The post-processing image and the original image are then compared. Pixels, whose values are elevated above a predetermined threshold as a result of averaging, are considered extravascular. The end result is a binary image according to the vascular/extra vascular criterion (Fig. 2B). This classified image is used as a mask, enabling the evaluation of fluorescence intensities at specific regions in the exposed cortex. According to known anatomical features (width, length) the primary artery is selected from the footage as a region of interest (ROI). Fluorescent intensities in this region alone are then averaged at each time point, yielding an intensity–time curve (IT, Fig. 2C) representative of the primary artery, also termed the “arterial input function” (AIF). This function indicates tracer input levels into the inspected cortical section. An identical process is performed regarding the entire extra-vascular space, yielding an IT curve indicative of tracer residues left in the inspected section. Comparing the two functions yields a parameter reflective of BBB permeability – a permeability indicator (PI): The ratio function between normalized fluorescence intensities of the extra-vascular and arterial IT curves is calculated: (IEV /IAIF )(t). The ratio is averaged in the time span ranging from the critical time point-tcr the start point of the second decline phase(arrow, Fig. 2C), till end of measurement cycle point: t (1/T ) t end (IEV /IAIF )(t)dt T = tend − tcr . cr
This particular time span is chosen for analysis since it represents the tracer’s clearance phase – a period in which intra-vascular tracer is washed out of the brain and tracer residues become apparent. BBB disruptions are visually detected in this time period. By performing a similar analysis with each extra-vascular pixel’s IT curve instead of the one representing the entire extra-vascular space, yields a PI values for each extra-vascular pixel and the formation of a permeability map (Fig. 2F). Such maps represent BBB integrity level in each location of the inspected cortical section, based on color scale. An example for this approach’s ability for detecting and quantitatively assessing alterations in BBB phenotype can be found in Fig. 2D–G DOC (2 mM, diluted in aCSF). Perfusions of the exposed cortex result in leakage of NaFlu into extra-vascular space (Fig. 2D). This leakage is quantified as intensification of the extra-vascular fluorescence signal over the arterial one (Fig. 2E), which results in significantly higher values of PI (Fig. 2G). DOC-induced BBB dysfunction occurs in the entire exposed cortical section, as evident by global color intensification in permeability mapping (Fig. 2F).
5
3.1.2. Direct optical imaging – laser-speckle imaging for BBB permeability assessment In recent time, laser speckle contrast imaging (LSCI) has been widely used to measure blood flow in a variety of tissues. The physical concept of this approach lies in the fact that when coherent laser illuminates an object the beam is scattered, and adds constructively as well as destructively. A random interference pattern is therefore formed and can be visualized with appropriate optical receptors [27]. When the laser illuminates moving objects (e.g. blood cells) the pattern fluctuates with time; its integration results in reduced spatial contrast in areas of increased motion. Thus, motion of moving particles is encoded in the blurring of the time integrated speckle pattern. Quantification of motion is achieved by mathematically analyzing the spatial blurring [27,28]. A recent study [29] employed laser speckle contrast imaging to assess BBB permeability, constituting the first demonstration of LSCI in that aspect. The basis for this modality is that BBB disruption can alter cerebral blood flow [23]. LSCI was performed on an exposed rat cortical section together with fluorescence tracking of a systemically administered non-BBB permeable dye. Chemically induced BBB disruption, noted and localized by fluorescence imaging, resulted in arterial vasodilation accompanied by a net reduction in venous flow velocity. Thus, suggesting a different impact of BBB disruption on flow speeds of arteries and veins. The ratio of output/input flow velocity profiles in the inspected cortical section was eventually chosen as an indicative parameter for BBB disruption, whose drop in values following BBB disruption was found to be significant. 3.1.3. Non-invasive techniques in pre-clinical imaging The direct vessel imaging approach, in what is perhaps its greatest advantage, allows for high spatial resolution (down to 10 m, depending on the imaging equipment). The immediate environment of brain vasculature is examined. However, this is achieved by way of craniotomy; a cortical section is exposed thus compromising the integrity of its supporting vasculature due to a decline in intra-cranial pressure [30]. Therefore, when examining the effects of various treatment approaches on BBB functionality, additional assessments must be performed in order to rule out surgically induced damage to vessel integrity and subsequent harm to BBB functionality. An alternative approach involves the use of non-invasive techniques in which larger scale brain regions are imaged. The modality has no bearing on vessel integrity/permeability; however this is achieved at the price of low spatial and temporal resolution (∼1 mm and ∼1 min respectively). Non-invasive imaging based BBB permeability assessment is commonly done with near infrared fluorescence (NIRF), magnetic resonance imaging (MRI), positron emission tomography (PET), single photon emission tomography (SPECT) and X-ray computerized tomography (CT). A non-BBB permeable reactive tracer is systemically administered, generating image enhancement. Various offline image processing techniques have been developed to assess tracer kinetics over time and space, and translate this enhancement into numerical parameters representing vascular features; among which is BBB permeability. In vivo, non-invasive fluorescence imaging is quite limited given the complexity of photon penetration through tissues. Tissues are heterogeneous surfaces with different refractive indices and a photon traveling to reach a fluorescent contrast agent may be reflected back to its source. The same is true for photons traveling from a fluorophore in attempt to reach a receiver unit [31]. Additionally, tissues may emit light upon excitation even without the presence of a contrast agent; a phenomenon known as “auto-fluorescence”, that may harm the signal to background ratio [31]. However, excitation at NIR wavelengths (>700 nm) can reduce these shortcomings. Water and lipids tend to absorb NIR photons [32] and are nonfluorescent at this excitation [31]; allowing photon propagation
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
G Model YSCDB 1697 1–10 6 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
and high signal to background ratio. The use of NIRF has been implemented in BBB permeability analysis as well. Klohs et al. describe their attempt to detect fluorescently labeled albumin in mice brains [33]. The study has demonstrated extravasation of albumin into brain parenchyma induced by stroke applied with the middle cerebral artery occlusion model [34]. Higher fluorescent intensities were measured in the treated hemisphere in comparison to the non-treated one. These results were correlated to EB extravasation evaluated ex vivo. Nevertheless, NIRF imaging has some limitations. Its sensitivity is limited by contrast agent’s depth and background fluorescence [33]. In the described study the differences in fluorescence intensities between hemispheres were stronger in ex vivo examinations than in the parallel non-invasive procedure. This was due to high fluorescence of the skin over the non-treated hemisphere due to injury. Contrast enhanced nuclear imaging (CENI) can quantitatively assess cellular processes. For instance, PET with the administration of 18 F-2-fluoro-2-deoxy-d-gluocose (18 F-FDG) is widely used to quantify changes in glucose metabolism [35,36]. Glucose enters brain parenchyma through glucose transporter 1 (GLUT1) receptors expressed in brain vascular endothelial cells [3]. Thus altered glucose uptake could serve as a marker indicating modulations of the BBB’s molecular structure. Yang et al. [37] demonstrated the effects of focused ultrasound on glucose metabolism in anesthetized rats using PET combined with 18 F-FDG administration following sonication. Disruption of the BBB, induced by sonication was observed by spectrophotometer readings of EB staining in brain slices, and was significantly reduced following 24 h. A 2-compartmental model was used for quantification of 18 F-FDG uptake into the brain. Interestingly, the study presented that glucose uptake was reduced immediately following sonication in comparison to levels found in untreated controls; and was returned to control level post 24 h. These findings were corroborated by reduced expression of GLUT1 found in treated brains. This study remarkably indicates that while the use of a biomarker in non-imaging based analysis techniques indicated BBB disruption, medical imaging indicated phenotype modulations which are in contradiction to the latter. This raises important questions regarding BBB selectivity even in a state of improper functionality. It exemplifies how the use of medical imaging can improve the resolution in which we explore the BBB. Additionally, CENI is used to investigate efflux mechanisms in the BBB as well as CNS inflammation. Efflux pumps and drug transporters belonging to the ATP-binding cassette family of transporters have become another target to study selective transport mechanisms at the BBB. P-glycoprotein (p-gp), a transporter comprising the BBB’s phenotype, mediates the efflux of xenobiotics [2]. PET Molecular imaging, non-invasively detecting p-gp function, can identify p-gp overexpression at the BBB which may underlie pharmaco-resistance. Bartmann et al. exemplify such an approach [38] by administering the p-gp radiolabeled substrate 2 -methoxyphenyl-(N-2 -pyridinyl)p-fluoro-benzamidoethyipiperazine ([18 F]MPPF) to anesthetized, epileptic rats. The study exhibited a correlation between ([18 F]MPPF) kinetics and responsiveness to anti-epileptic treatment. CNS inflammation is featured by increased vessels’ permeability, leading to infiltration of peripheral immune cells (lymphocytes, granulocytes, neutrophils, monocytes/macrophages), as well as activation of microglia. By radiolabeling autologous leukocytes, this process can be monitored. Tracers in this approach include 111 In-oxine and 111 In-troponolate [35]. MRI in combination of MR-visible contrast agents containing gadolinium [39–41] is considered to be the gold standard in assessment of BBB impairment using non-invasive imaging modalities. Advantages of MRI in terms of safety and spatial resolution, compared to other imaging modalities such as PET [39], contribute to the latter. MRI generates high-resolution information of brain
anatomy and cerebral blood flow; thus setting a platform on which BBB integrity could be assessed (Fig. 2A/B). For instance, T2 and diffusion-weighted scans may identify ischemic brain regions [42], which may also be characterized by enhanced vascular permeability [8,43]. Protocols for BBB permeability quantification in MRI include static contrast enhanced (SCE) imaging, in which statistical differences between scans are calculated; and dynamic contrast enhanced imaging (DCE) where alterations in tracer concentration are monitored in repeated scans. A practical approach for SCE-imaging based BBB integrity evaluation was developed in-house, and is termed the “pre-post” comparison [41]. Briefly, T1-weighted scans are performed prior to and 5 min following systemic gadopentetic acid (Gd-DTPA) administration; making the second scan contrast-enhanced in regions where tracer is accumulated due to BBB dysfunction (Fig. 3C). For quantification purposes, a comparison is done between the scans in several steps: First, a neighborhood of 3 × 3 is selected around each pixel (9 pixels) in both pre and post images. A slice-wise unpaired ttest is then performed between each two parallel neighborhoods in the pre and post images. Pixels indicating accumulation of Gd-DTPA are identified as those whose neighborhood comparison benefited statistically significant results. Each pixel is therefore assigned a P-value, resulting in a binary significance image in which all pixels with scores under a predetermined threshold are assigned the value “1” and the rest, the value “0”. The second step involves the calculations of enhancement differences between pre and post images. Enhancement distribution histograms are calculated for three representing regions: muscle, eyeball and blood vessel in order to determine the BBB breakdown enhancement range for each subject. The muscle represents a “permeable tissue” with no BBB, while the eyeball represents a region with intact BBB (that isolates the retina from its supporting vasculature). The blood vessel region is used to exclude vascular tissue. Fitting a Gaussian to the muscle histogram and estimating the average and standard deviation enables the estimation of an enhancement range indicative of BBB breakdown. Pixels indicating a leaky BBB are therefore those whose statistical analysis indicates a statistically significant intensity increase, and whose post injection intensities are within the enhancement range. The third step is clustering of potentially permeable pixels. Neighboring pixels are aggregated into anatomically connected regions, so to establish a more robust and clinically relevant assessment (Fig. 3D). Advances in non-invasive imaging techniques in general, and in MRI in particular, in terms of safety and temporal/spatial resolution; have led to the development of novel protocols for BBB integrity assessment, and their enrollment in clinical research as well as in every-day use. 3.2. Clinical imaging modalities The advances made in pre-clinical BBB imaging modalities have raised the question of their implementation in clinical practice as well. Imaging of the BBB in human patients has been extensively studied over the last two decades. Although SPECT and CT can be used to evaluate BBB integrity, MRI is the modality-of-choice in most cases. Similarly to the animal model case, contrast-enhanced MRI (CE-MRI) sequences are used for BBB imaging in human, where Gd-based contrast agents (CA) that do not cross the intact BBB are used. 3.2.1. Analytical methods Assessment of BBB integrity can be carried out by means of qualitative, semi-quantitative or quantitative methods. Qualitative methods are very simple in terms of scanning parameters and computational costs. However, they can come short in terms of sensitivity and do not provide information regarding the permeability
Please cite this article in press as: Chassidim Y, et al. Analyzing the blood–brain barrier: The benefits of medical imaging in research and clinical practice. Semin Cell Dev Biol (2014), http://dx.doi.org/10.1016/j.semcdb.2014.11.007
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
454
455 456 457 458 459 460 461 462 463
464 465 466 467 468 469
G Model YSCDB 1697 1–10
ARTICLE IN PRESS Y. Chassidim et al. / Seminars in Cell & Developmental Biology xxx (2014) xxx–xxx
7
Fig. 3. MRI analysis for assessment of BBB permeability. (A) In T2-weighted MRI, brain areas suspected of BBB dysfunction-related pathology (e.g. oedeme) appear to have high signal intensity (overlaid in red) in relation to healthy brain tissue. These regions are defined as “lesions”. (B) Analysis of variability in T2-weighted image intensity benefits lesion identification. A mixture of 3 Gaussian probability density functions is proposed to model variability in brain tissue intensities. The cut off threshold, indicating a lesion, is defined as an intersection between the 2d and 3d Gaussians (marked by an arrow). In order to remove small disconnected noisy speckles, a clustering procedure is applied. To this end, neighboring pixels’ connectivity (4-connected) is checked. Clusters of pixels, smaller in size than a minimal area are defined as noisy regions and are excluded from the lesion object. (C) T1 weighted scans are used to quantify BBB dysfunction (overlaid in blue). The quantification approach is based on a comparative analysis between contrast-enhanced MR scans before and 20 min after contrast agent injection. (D) Percentage value of enhancement differences between the pre and post contrast images. Significantly different pixels within an enhancement range of 30–100% are considered to represent brain tissue with potential leaky BBB. The described above “clustering procedure” was applied in order to remove noisy speckles.
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507
values in the affected areas. These methods include comparison of pre- and post-CA injection images. Another example is the delayed contrast extravasation subtraction [44], which utilizes 3 scans with a long delay between the first and third scan (75 min). This method has the advantage of enabling scans with high spatial resolution due to the fact that the acquisition speed is not a factor. It has been claimed that this method can be used to differentiate tumor progression from tissue damage induced by treatment. Despite their shortcomings, qualitative imaging methods may be clinically useful. Semi-quantitative and quantitative methods rely on dynamic contrast-enhanced (DCE-MRI) acquisition protocols, where images are acquired before, during and after the CA injection. The acquisition parameters depend on the analysis method, and are under debate, since a natural tradeoff exists between spatial/temporal resolution/volume coverage. Semi-quantitative approaches make use of measurements of MRI signal over time, and are calculated either from image intensity’s dynamic values or from concentration curves. The measurements include area under curve (AUC), initial area under curve (iAUC), time to peak (TTP), wash-in and wash-out slopes. These metrics arise from a combination of physiological parameters such as blood flow, blood volume, and permeability rather than imply a single parameter. Nevertheless, they can be very useful in clinical settings as they are less demanding in terms of acquisition parameters, especially regarding temporal resolution. Additionally, slower scans result in higher signal to noise ratio (SNR) and allow higher spatial resolution to be obtained. Quantitative methods use pharmacokinetic models with variable number of parameters to infer on physiological properties of different tissues. These models require the calculation of CA concentration–time curves (C(t)) in each voxel. To allow the conversion from intensity values to concentration, additional sequences are required (e.g. [45]) Full description of the common models that are used to fit the curves and the hierarchy between them have been recently reviewed by Sourbron and Buckly [46]. Imaging acquisition protocols for quantitative method are a matter of debate, and although no standard procedure has been established, the
requirements from suitable protocol include: volume coverage, spatial resolution, temporal resolution, length of scan and artifact minimization. Covered volume should usually include the whole brain, unless only a certain region (e.g. tumor region) is the area of interest. Temporal resolution has a direct impact on the accuracy of parameter estimation and the models differ in their sensitivity to the sampling interval [47–52]. Although several works have recommended a maximal sampling interval of
