[18F]CB251 PET/MR imaging probe targeting translocator protein (TSPO) independent of its polymorphism in a neuroinflammation model

Kim et al.

LAB/HAB = 37.28

N N

O

O CH3

OCD218F

[18F]fmPBR28-d2

Ki=3.6 nM

0.1 2.1

HAB

LAB

HAB

LAB

3.30.2

HAB LAB HAB LAB

LAB/HAB = 1.14

NN

Cl

ClO

N

O

18F

[18F]CB251

Ki=0.27 nM

1.0 30.1

HAB

LABHAB

LAB

3.30.2

HAB LAB HAB LAB

[18F]CB251 PET/MR BLI of immune cells

[18F]CB251 PET/MR imaging probe targeting translocator 1

protein (TSPO) independent of its polymorphism in a 2

neuroinflammation model 3

4 Kyungmin Kim1,2,3, Ha Kim1,3, Sung-Hwan Bae1,3, Seok-Yong Lee1,2,3, Young-Hwa Kim1,3,5, 5

Juri Na1,2,3, Chul-hee Lee1,2,3, Min Sun Lee1, Guen Bae Ko1,8, Kyeong Yun Kim1,8, Sang-Hee 6

Lee4, In Ho Song4,5, Gi Jeong Cheon1,3, Keon Wook Kang1,2,3, Sang Eun Kim4,5,6, June-Key 7

Chung1,9, Euishin Edmund Kim1, Sun-Ha Paek2,7, Jae Sung Lee1,2,8, Byung Chul Lee4,6*, 8

Hyewon Youn1,3* 9 10 1Department of Nuclear Medicine, Seoul National University Hospital, 2Biomedical Sciences, 11 3Laboratory of Molecular Imaging and Therapy, Cancer Research Institute, Seoul National 12

University College of Medicine, 4Department of Nuclear Medicine, Seoul National University 13

Bundang Hospital, 5Department of Transdisciplinary Studies, Graduate School of 14

Convergence Science and Technology, Seoul National University, 6Center for Nanomolecular 15

Imaging and Innovative Drug Development, Advanced Institutes of Convergence Technology, 16

Seoul National University, 7Department of Neurosurgery, Seoul National University Hospital, 17 8Brightonix Imaging Inc., 9Department of Nuclear Medicine, National Cancer Institute, 18

Republic of Korea 19

20

21

*For correspondence 22

Hyewon Youn, Ph.D. 23

Department of Nuclear Medicine, Seoul National University Hospital, 24

#207-4 Samsung Cancer Research Bldg. 25

103 Daehak-No, Chongno-Gu, Seoul, 03080, Korea. 26

Tel: +822-3668-7026; Fax: +822-745-7690; E-mail: hwyoun@snu.ac.kr 27

28

Byung Chul Lee, Ph.D. 29

Department of Nuclear Medicine, Seoul National University Bundang Hospital 30

Goomie-Ro 173, Boondang-Gu, Sungnam, Kyunggi-Do, 13620, Korea 31

Tel: +8231-787-2956; Fax: +8231-787-2957; E-mail: leebc@snu.ac.kr 32

Abstract 33

The 18 kDa translocator protein (TSPO) has been proposed as a biomarker for the 34

detection of neuroinflammation. Although various PET probes targeting TSPO have been 35

developed, a highly selective probe for detecting TSPO is still needed because single 36

nucleotide polymorphisms in the human TSPO gene greatly affect the binding affinity of 37

TSPO ligands. Here, we describe the visualization of neuroinflammation with a 38

multimodality imaging system using our recently developed TSPO-targeting radionuclide 39

PET probe [18F]CB251, which is less affected by TSPO polymorphisms. 40

41

Methods: To test the selectivity of [18F]CB251 for TSPO polymorphisms, 293FT cells 42

expressing polymorphic TSPO were generated by introducing the coding sequences of wild-43

type (WT) and mutant (Alanine Threonine at 147th Amino Acid; A147T) forms. 44

Competitive inhibition assay was conducted with [3H]PK11195 and various TSPO ligands 45

using membrane proteins isolated from 293FT cells expressing TSPO WT or mutant-A147T, 46

representing high-affinity binder (HAB) or low-affinity binder (LAB), respectively. IC50 47

values of each ligand to [3H]PK11195 in HAB or LAB were measured and the ratio of IC50 48

values of each ligand to [3H]PK11195 in LAB to HAB was calculated, indicating the 49

sensitivity of TSPO polymorphism. Cellular uptake of [18F]CB251 was measured with 50

different TSPO polymorphisms, and phantom studies of [18F]CB251-PET using 293FT cells 51

were performed. To test TSPO-specific cellular uptake of [18F]CB251, TSPO expression was 52

regulated with pCMV-TSPO (or shTSPO)/eGFP vector. Intracranial lipopolysaccharide (LPS) 53

treatment was used to induce regional inflammation in the mouse brain. Gadolinium (Gd)-54

DOTA MRI was used to monitor the disruption of the blood-brain barrier (BBB) and 55

infiltration by immune cells. Infiltration of peripheral immune cells across the BBB, which 56

exacerbates neuroinflammation to produce higher levels of neurotoxicity, was also monitored 57

with bioluminescence imaging (BLI). Peripheral immune cells isolated from luciferase-58

expressing transgenic mice were transferred to syngeneic inflamed mice. Neuroinflammation 59

was monitored with [18F]CB251-PET/MR and BLI. To evaluate the effects of anti-60

inflammatory agents on intracranial inflammation, an inflammatory cytokine inhibitor, 2-61

cyano-3, 12-dioxooleana-1, 9-dien-28-oic acid methyl ester (CDDO-Me) was administered in 62

intracranial LPS challenged mice. 63

64

Results: The ratio of IC50 values of [18F]CB251 in HAB to LAB indicated similar binding 65

affinity to WT and mutant TSPO and was less affected by TSPO polymorphisms. [18F]CB251 66

was specific for TSPO, and its cellular uptake reflected the amount of TSPO. Higher 67

[18F]CB251 uptake was also observed in activated immune cells. Simultaneous [18F]CB251-68

PET/MRI showed that [18F]CB251 radioactivity was co-registered with the MR signals in the 69

same region of the brain of LPS-injected mice. Luciferase-expressing peripheral immune 70

cells were located at the site of LPS-injected right striatum. Quantitative evaluation of the 71

anti-inflammatory effect of CDDO-Me on neuroinflammation was successfully monitored 72

with TSPO-targeting [18F]CB251-PET/MR and BLI. 73

74

Conclusion: Our results indicate that [18F]CB251-PET has great potential for detecting 75

neuroinflammation with higher TSPO selectivity regardless of polymorphisms. Our 76

multimodal imaging system, [18F]CB251-PET/MRI, tested for evaluating the efficacy of anti-77

inflammatory agents in preclinical studies, might be an effective method to assess the severity 78

and therapeutic response of neuroinflammation. 79

80

Keywords: neuroinflammation, TSPO, polymorphism, PET/MRI, multimodal imaging 81

Introduction 82

Neuroinflammation is the primary defense mechanism against viral and bacterial 83

infections, toxic metabolites, and injury in the central nervous system (CNS). Several studies 84

have shown that prolonged neuroinflammation is linked to the onset and progression of 85

neurodegenerative conditions such as Alzheimer’s and Parkinson’s diseases, and multiple 86

sclerosis [1-3]. Bidirectional interactions between the CNS and peripheral immune systems 87

adversely affect the CNS [4-6] and are considered the main cause of prolonged 88

neuroinflammation and neurodegenerative diseases [5, 6]. These interactions induce not only 89

the activation of microglia and resident immune cells in the brain but also the disruption of 90

the blood-brain barrier (BBB), which results in the migration of peripheral immune cells into 91

the region of CNS neuroinflammation [6-8]. Therefore, imaging the region of 92

neuroinflammation and investigating the interactions between CNS immune cells and 93

infiltrating peripheral immune cells are important for understanding neurodegenerative 94

diseases. 95

Several studies of various imaging techniques have been conducted for identifying 96

the region of neuroinflammation [9-13]. Single-photon emission computed tomography 97

(SPECT) using 99mTc-hexamethylpropyleneamineoxime (HMPAO) or 99mTc-ethyl cysteinate 98

dimer (ECD) has been frequently used to visualize the increased blood flow that occurs in 99

neuroinflammation [14, 15]. Disruption of the BBB and changes in vascular permeability can 100

also be visualized by 99mTc-DTPA SPECT and gadolinium (Gd)-DTPA magnetic resonance 101

imaging (MRI) [16, 17]. Also, [11C]arachidonic acid and [18F]FDG positron emission 102

tomography (PET) have been used for imaging neuroinflammation reflected by changes in 103

incorporation of arachidonic acid and glucose metabolism, respectively, in neurons and 104

inflammatory cells [18]. However, [18F]FDG-PET cannot be used for early diagnosis of 105

neuronal diseases because the basal level of [18F]FDG uptake is too high in the brain to 106

distinguish neuroinflammation from normal brain regions. Therefore, recent studies have 107

focused on identifying better imaging biomarkers to overcome the limitation of [18F] FDG-108

PET for imaging neuroinflammation. 109

The translocator protein 18 kDa (TSPO) has been proposed as an attractive target 110

biomarker of neuroinflammation [19-31]. TSPO, located in the mitochondrial outer 111

membrane, acts as a peripheral benzodiazepine receptor and a cholesterol transporter. 112

Because activated immune cells have been reported to show increased TSPO expression, 113

TSPO-targeting probes such as [11C]PK11195 and [11C]PBR28 have been suggested for PET 114

imaging of inflammation. However, these PET probes have limitations, such as nonspecific 115

binding, a low target-to-background ratio, and different binding sensitivities to TSPO 116

polymorphisms [28-31]. Although radiochemists have been working on developing improved 117

imaging ligands for targeting TSPO, a highly selectable probe for detecting TSPO is still 118

needed. 119

Our group recently developed a new imidazopyridine TSPO ligand, [18F]CB251, 120

which showed promise as a TSPO radiotracer [26]. We introduced fluorine-18 because 18F-121

labelled compounds with a longer half-life (109.8 min for fluorine-18 and 20.3 min for 122

carbon-11) are preferable to 11C-labelled compounds for clinical applications. We reported on 123

the synthesis of [18F]CB251, which showed a 5.1-times higher TSPO binding affinity than 124

[11C]PK11195 [26, 31]. 125

In this study, we evaluated the selectivity of the [18F]CB251 probe for human TSPO 126

harboring genetic variants, and its ability to reveal activated immune cells in regions of 127

neuroinflammation. We established a mouse model of intracranial lipopolysaccharide (LPS)-128

induced regional inflammation to visualize neuroinflammation using multimodality imaging. 129

Both CNS and peripheral immune cells infiltrating across the BBB play an important role in 130

neuroinflammation. We isolated peripheral immune cells from spleens of luciferase-131

expressing transgenic mice and systemically injected them into mice with intracranial LPS-132

induced regional neuroinflammation, and visualized with bioluminescence imaging (BLI). By 133

comparing TSPO-targeting [18F]CB251 PET/MR and peripheral immune cells-targeting BLI, 134

we attempted to determine whether the [18F]CB251 probe can quantitatively assess the 135

severity of neuroinflammation and therapeutic response to anti-inflammatory agents. 136

Results 137

In vitro evaluation of [18F] CB251 as a probe for targeting TSPO to image activated 138

immune cells 139

As we reported previously [26], an 18F-labeled alpidem analog, [18F]CB251, was 140

synthesized from 2-phenyl-imidazo [1,2-a] pyridine, and used as a specific probe for TSPO. 141

We also reported that the binding affinity of CB251 was higher than that of the widely used 142

TSPO ligands, PK11195 [26, 32] , PBR28 [26, 32] , fluoremethyl-PBR28-d2 (fmPBR28-d2), 143

and GE-180 [33] (Figure 1A). 144

For TSPO ligand development, TSPO selectability regardless of TSPO 145

polymorphism (rs6971; A147T) is important because TSPO polymorphism is known to affect 146

both in vitro and in vivo radio-ligand bindings. TSPO ligands show substantial differences in 147

affinity between subjects classified as a high-affinity binder (HAB; TSPO wild-type (WT)) 148

and low-affinity binder (LAB; TSPO A147T mutant (Mut). TSPO genotypes and 149

classification is showed in Figure S1A. PBR28 was reported to have a different binding 150

affinity to TSPO polymorphism compared to PK11195 [28-31]. 151

To test the ligand-binding affinity and cellular uptake, we generated 293FT cells 152

expressing polymorphic TSPOs, including the most abundant type (WT) and mutant TSPO 153

(C to T point mutation generating Alanine to Threonine at 147th amino acid of TSPO; Mut-154

A147T). Figure S1B shows the mutagenesis site and the map of expression vectors. 155

To compare ligand binding affinity to the TSPO polymorphism, we conducted 156

competitive inhibition assay with [3H]PK11195 and various TSPO ligands such as PK11195, 157

fluoromethyl-PBR28-d2 (fmPBR28-d2) and CB251 using membrane proteins isolated from 158

293FT cells expressing TSPO polymorphism (TSPO WT and TSPO Mut-A147T) (Figure 159

1B). The ratio of IC50 values (IC50 of LAB/ IC50 of HAB) for PK11195 was 0.83 and similar 160

to earlier reports [34, 35]. Unlike fmPBR28-d2 (IC50 of LAB/ IC50 of HAB=37.28), CB251 161

showed a similar ratio (IC50 of LAB/ IC50 of HAB =1.14), which was close to 1. These results 162

indicated that fmPBR28-d2 ligand had a difference in binding affinity according to the TSPO 163

polymorphism, whereas CB251 showed ratios similar to PK11195. 164

The higher affinity of ligand binding to the target protein is an important factor. 165

However, clinical applications also require consideration of biochemical, cellular, and 166

physiological factors, such as lipophilicity of the ligand and its non-specific insertion in the 167

membrane, cellular metabolism of the ligand, endogenous cellular TSPO levels, and in vivo 168

kinetics. Also, the higher affinity of a ligand to the target proteins does not automatically 169

reflect target selectivity. Therefore, we performed cellular uptake of various radiolabeled 170

ligands targeting TSPO because it reflects the entire cellular process from infiltration through 171

the cell membrane to its binding to TSPO on the mitochondrial outer membrane. As expected, 172

our results demonstrated that [18F]fmPBR28-d2 was sensitive to the A147T mutation. GE-173

180, another well-known TSPO ligand, showed less sensitivity to A147T mutation. However, 174

[11C]PK1195 and [18F]CB251 were much less sensitive to TSPO polymorphism, and 175

[18F]CB251 showed similar cellular uptake levels between the polymorphic TSPOs (Figure 176

1C). 177

We also observed differences in PET signals from the cells with different 178

polymorphic TSPOs using phantoms. [18F]fmPBR28-d2 was used as a sensitive probe for 179

TSPO polymorphism. Similar to the cellular uptake test results, [18F]CB251 was less 180

sensitive to TSPO polymorphism (Figure 1D, 1E), indicating that it could be a good PET 181

probe with better selectivity for targeting TSPO. 182

In a previous study, the specificity of [18F CB251 ligand for TSPO was not directly 183

addressed in a biological system [26]; therefore, in this study, the specificity of [18F]CB251 184

was tested by regulating TSPO expression in cell culture. [18F]CB251 uptake was measured 185

in the cells with different levels of TSPO expression, which was regulated by transfection 186

with pCMV-TSPO/eGFP for overexpression and pCMV-shTSPO/eGFP for reduced 187

expression. 293FT cells expressing low levels of TSPO were used for TSPO overexpression 188

(Figure 2A), and MDA-MD-231 cells expressing high levels of TSPO were used for reduced 189

expression (Figure 2C). Fluorescence microscopic images showed increased eGFP signals, 190

reflecting the successful transfection of each vector in the target cells (Figure 2A-C). 191

Increased TSPO expression was observed by both immunoblotting and fluorescence 192

microscopy of pCMV-TSPO/eGFP transfected cells (Figure 2A), and [18F]CB251 uptake was 193

increased in TSPO-overexpressing cells (Figure 2B). In contrast, decreased TSPO expression 194

was observed in pCMV-shTSPO/eGFP-transfected cells (Figure 2C), and [18F]CB251 uptake 195

was decreased in cells with reduced TSPO expression (Figure 2D). These data demonstrated 196

that [18F]CB251 uptake was directly reflected by changes in TSPO expression, indicating that 197

[18F]CB251 is specific for TSPO at the cellular level. 198

Next, we monitored TSPO expression as a marker for activated immune cells and 199

tested whether [18F]CB251 could be used as a specific imaging probe for activated immune 200

cells. The mouse macrophage cell line, RAW264.7, and primary white blood cells obtained 201

from mouse spleens (splenocytes) were treated with LPS and IFN-γ for 24 h to activate 202

immune cells. Following activation, immunoblotting in RAW264.7 cells and mouse 203

splenocytes showed increased expression of TSPO as well as iNOS, a marker of activated 204

macrophages (Figure 2E). Increased [18F]CB251 uptake was also observed in activated 205

immune cells with increased TSPO expression (Figure 2F). These data indicated that 206

[18F]CB251 directly reflected the activation of immune cells by targeting TSPO. 207

208

BLI of peripheral immune cells in mice receiving an intracranial injection of LPS 209

In addition to activation of CNS immune cells, infiltration of peripheral immune cells 210

through the BBB into the inflammatory area is considered to be critical for the 211

neuroinflammatory process [34-36]. Therefore, imaging depicting the infiltration of 212

peripheral immune cells across the BBB provides useful information for the evaluation of the 213

inflammatory process in the brain. Since BLI has been frequently used to visualize the 214

distribution and proliferation of immune cells in vivo [37, 38] , we employed this technique 215

to observe peripheral immune cell infiltration into the neuroinflammatory region after 216

systemic administration of luciferase-expressing peripheral immune cells. Splenocytes from 217

luciferase-expressing transgenic mice (C57BL/6.LucTg) were used as a source of peripheral 218

immune cells [38]. We previously reported that the composition and function of immune cells 219

in the C57BL/6.LucTg mice were similar to those of the normal C57BL/6 mice [38]. 220

Flow cytometry analysis showed that splenocytes from reporter transgenic mice 221

(C57BL/6.LucTg) used in this study had immune cell populations similar to those of normal 222

(C57BL/6) mice (Figure 3A). The BLI signals of splenocytes isolated from luciferase-223

expressing transgenic mice increased in a cell-number-dependent manner with a high 224

correlation value (r2 = 0.9991), which could quantitatively assess the number of splenocytes 225

that exhibited BLI signals (Figure 3B). To assess peripheral immune cell biodistribution by 226

BLI, we isolated splenocytes from reporter mice and grafted them into inflamed mice (Figure 227

3C). 228

BLI signals produced by grafted luciferase-expressing splenocytes were observed in 229

peripheral immune cell-targeting organs such as the lung, heart, liver, spleen, and intestine. 230

However, BLI signals from grafted splenocytes were observed only in the brain of inflamed 231

mice 4 days after intracranial LPS injection (Figure 3D). Most of the immune cell-homing 232

organs showed bioluminescent signals in both saline-injected control mice and LPS-injected 233

inflamed mice. However, the brain showed higher signals in mice receiving intracranial LPS 234

injection than the saline injection. Since the mouse brain is not a homing organ of peripheral 235

immune cells, our results indicated that grafted peripheral immune cells infiltrated the 236

inflamed brain of mice. 237

238

MRI using Gd-DOTA to monitor BBB disruption, immune cell infiltration, and 239

activation 240

Our observations on peripheral immune cell infiltration of the LPS-injected mouse 241

brains led us to monitor BBB disruption induced by mechanical stress. Our mouse model was 242

created by injecting saline (control agent) or LPS (inflammation-inducing agent) into the 243

mouse brain using a robotic stereotaxic device with a Hamilton syringe. We, therefore, had to 244

investigate the possibility of peripheral immune cell infiltration through the injection hole 245

rather than BBB. For this, closure of the injection hole and no BBB disruption in control mice 246

had to be verified. Previously, Gd-DOTA (Dotarem), a gadolinium-based MRI contrast agent 247

with a macrocyclic structure, has been used as an efficient paramagnetic contrast agent to 248

increase MR sensitivity and specificity for the identification of BBB defects [39-41]. 249

Generally, T1-weighted MRI has been used to visualize the presence of gadolinium 250

penetration through disrupted BBB. On the other hand, T2-weighted MRI usually provides 251

information about water content or vasogenic edema of brain lesions induced by LPS 252

injection. Therefore, we performed MRI using Gd-DOTA as a contrast agent to assess the 253

level of BBB disruption and vasogenic edema in mouse models. 254

We injected saline or LPS into the right striatum of the mouse brain to create 255

intracranial inflammation (Figire 4A). The MR signals in the right striatum were higher in the 256

LPS-injected mice than in the saline-injected mice in both the T1- (Figure 4B) and T2- 257

weighted images (Figure 4C). Representative MR images demonstrated that BBB disruption 258

caused by mechanical stress (injection hole) was recovered on day 1, as shown in the saline-259

injected group, while in the LPS-induced group, signal peaks indicating BBB disruption were 260

still visible on day 3 (Figure 4B; N=3). As shown in Figure 4C, the T2 signal representing 261

vasogenic edema due to LPS injection was observed. Peak T2 signal in LPS-treated mice, 262

representing tissue damage resulting from inflammation, was observed on day 3. On day 4, 263

T2 signals were reduced in the LPS-treated mice. Since T2-weighted signal enhancement was 264

reported to be produced by gadolinium-based chelates trapped in viable immune cells [41], 265

high T2 signals are thought to indicate the presence of viable immune cells in the region. Our 266

results showed that tracer uptake on day 4 in T1- and T2-weighted MRIs with Gd-DOTA was 267

due to the upregulation of neuroinflammation induced by LPS and not due to BBB disruption 268

induced by mechanical stress. Based on these results, we used only T2-weighted MRI with 269

Gd-DOTA to monitor the severity of neuroinflammation for further investigation. 270

271

Simultaneous [18F]CB251 PET/MRI and BLI for visualization of neuroinflammation 272

It has been reported that simultaneous PET/MRI provides information on anatomical 273

and functional changes with precise co-registration of signals and accurate correction of 274

attenuation [42, 43]. Herein, we performed simultaneous [18F]CB251 PET/MRI and BLI to 275

visualize neuroinflammation indicative of immune cell activation and peripheral immune cell 276

infiltration in brain regions in the intracranial LPS mouse model (Figure 5A). Although T1- 277

and T2-weighted MR signals peaked on day 3 after LPS injection, neuroinflammation was 278

still observed on day 4. Imaging studies at this time point could provide additional useful 279

information about peripheral immune cell infiltration without interference due to severe 280

tissue damage. Therefore, we performed [18F]CB251 PET/MRI and BLIs on day 4. 281

Radiation signals from [18F]CB251-targeting TSPO proteins reflected activation of 282

immune cells, such as resident microglia and infiltrated peripheral immune cells. T2-283

weighted MR signals reflected localization of all viable immune cells, and the BLI signals 284

reflected only luciferase-expressing peripheral immune cells. On day 4, LPS-injected mice 285

showed higher signals from each imaging modality in the inflamed region of the right 286

striatum than in the saline-injected mice (Figure 5B-C). T2-weighted MR signals from LPS-287

injected mice appeared to emanate from a slightly wider area of immune cells (Figure 5B 288

left) than the [18F]CB251 PET signals (Figure 5B middle), but the critical region of T2-289

weighted MR signal was co-localized with the region of increased [18F]CB251 PET signals. 290

In the BLI, luciferase-expressing splenocytes representing peripheral immune cells were 291

visualized in the right striatum of the LPS-injected mice (Figure 5B right). The degree of 292

neuroinflammation could be observed using T2-weighted MR reflecting localization of all 293

viable immune cells, including residential and peripheral immune cells. However, we used 294

MR images of simultaneous PET/MR to obtain anatomical information for locating 295

inflammation and BBB disruption, since the observation of inflammatory cell activation 296

using quantitative PET is more sensitive. 297

Next, we compared the differences between the signal intensity ratio of the right 298

(saline or LPS injection) to the left (no injection) striatum in the saline- (N=6) and LPS-299

injected (N=7) mice (Figure 5C). [18F] CB251 PET signal was expressed as standard uptake 300

value (SUVmax), and BLI optical signal was expressed as total flux (photon/sec/cm2/sr). In 301

LPS-injected mice, [18F]CB251 radioactivity and BLI optical signals were significantly 302

increased compared to saline-injected mice (p=0.023 for [18F]CB251; p=0.011 for BLI ). 303

When we compared the difference between the non-injected left striatum and the saline-304

injected right striatum, [18F]CB251 PET signal increased slightly in the saline-treated right 305

striatum (p=0.012). However, the BLI signal did not change significantly in the saline-treated 306

right striatum. Since [18F]CB251 PET reflects residential microglial activity as well as 307

peripheral immune cell activity, these results suggested that [18F]CB251 PET was more 308

sensitive to mechanical stress due to intracranial injection than BLI, which only reflected 309

peripheral immune cells. Considering the change in the strength of BLI signals after LPS 310

injection, BLI appeared to be more sensitive to the development of inflammation than PET 311

MRI using the TSPO-targeting probe [18F]CB251. However, the BLI signals were not 312

detectable through the black skin of mice, which had to be removed to detect and assess the 313

BLI signals. Taken together, the simultaneous analysis by [18F]CB251 PET targeting TSPO 314

and BLI targeting grafted peripheral immune cells revealed resident and peripheral immune 315

cell activation due to inflammatory processes, as well as the location of infiltrating peripheral 316

immune cells. 317

318

Identification of immune cells in inflamed regions of the mouse brain 319

We performed immunohistochemical (IHC) staining to identify the presence of 320

different types of immune cells in the brain region of neuroinflammation (Figure 5D). 321

Antibodies were used for staining TSPO-expressing cells (anti-TSPO) and immune cells in 322

the CNS parenchyma including perivascular macrophages, monocytes, microglia (anti-CD68, 323

anti-Iba-1), antigen-presenting cells (anti-CD86), and T cells (anti CD4, CD8). Cells from 324

monocyte lineages such as circulating macrophages and resident microglia expressing CD68 325

were clearly increased and were located with TSPO expressing cells at the site of 326

inflammation (right striatum). Also, the antigen-presenting cells expressing CD89, which are 327

expressed only in peripheral immune cells and critical for T cell activation, increased 328

significantly in the right striatum of LPS-injected mice. Since macrophages and dendritic 329

cells from bone marrow generally exhibit higher antigen-presenting ability than microglial 330

cells, we confirmed that infiltrated peripheral immune cells were localized at the site of 331

inflammation. However, helper T cells (anti-CD4) and cytotoxic T cells (anti-CD8) appeared 332

widely distributed on the cross-sections of mouse brains. 333

Animals were sacrificed 4 days after LPS injections, and only the initial response 334

consisting of activated resident immune cells and recruited antigen-presenting cells could be 335

observed at the sites of inflammation. Our multimodality imaging data and IHC data 336

colocalizing TSPO-, CD68- and CD86-expressing cells in the inflamed region of the brain, 337

clearly demonstrated that immune cells in the LPS-injected region were activated, and 338

recruitment of peripheral immune cells occurred after intracranial LPS challenges. 339

340

Comparative analysis of neuroinflammation by [18F]CB251 PET and BLI after anti-341

inflammatory treatment 342

We compared the usefulness of [18F]CB251 PET with BLI to evaluate the therapeutic 343

efficacy of an anti-inflammatory agent in neuroinflammation. We used 2-cyano-3, 12-344

dioxooleana-1, 9-dien-28-oic acid methyl ester (CDDO-Me), which is known to inhibit 345

inflammatory cytokines such as IL-6, IL-10, and IL-12 [44-48]. We administered CDDO-Me 346

daily for 3 days after intracranial LPS-injection in the right striatum (Figure 6A). As shown in 347

the previous experiments, the [18F]CB251 uptake ratio (SUVmax, right/left striatum) of the 348

inflamed region in LPS-injected mice was significantly increased compared to saline-injected 349

control mice (p=0.002).On the other hand, the [18F]CB251 uptake ratio of CDDO-Me-treated 350

mice after LPS injection significantly decreased compared to LPS-injected mice (p=0.0323, 351

Figure 6B). A similar pattern was also confirmed by the ratios of SUVmean values (data not 352

shown). 353

In LPS-injected mice, the BLI signal ratio from the brain (expressed as the ratio of 354

the total photon flux of the right and the left striatum) was 3.93-fold higher than that from 355

saline-injected control mice (p=0.0009). The BLI signal ratio from the brain of mice treated 356

with anti-inflammatory CDDO-Me after LPS injection decreased dramatically compared to 357

LPS-injected mice (p=0.0025, Figure 6C). These data demonstrated that CDDO-Me 358

successfully inhibited peripheral immune cell infiltration and reduced their recruitment 359

leading to neuronal damage. Our results clearly showed that both modalities, TSPO-targeting 360

[18F]CB251 PET and peripheral immune cell-targeting BLI, revealed the degree of neuronal 361

inflammation and reduced activation of immune cells. 362

363

Discussion 364

In a previous report [26], we described synthesis, binding, and competition studies 365

with [18F]CB251 and its future applications. In this study, we attempted to test [18F]CB251, 366

targeting TSPO as a clinically applicable PET probe for imaging neuroinflammation. Our 367

results indicated that [18F]CB251 targeted TSPO with higher selectivity in human TSPO 368

polymorphisms (Figure 1), was specific for TSPO protein, and taken up by activated immune 369

cells with high TSPO expression (Figure 2), These observations suggested that [18F]CB251 is 370

a suitable candidate for imaging of TSPO-related neuroinflammation. We also established an 371

intracranial LPS-induced neuroinflammation mouse model, and designed experiments with a 372

multimodal imaging approach using [18F]CB251 PET/MR and BLI technologies to 373

understand the interaction between CNS and the peripheral immune system during 374

neuroinflammation. Furthermore, we tested whether [18F]CB251 PET could be used to 375

evaluate the therapeutic effect of anti-inflammatory drugs in this model. 376

The higher affinity of a ligand for the target protein does not automatically reflect the 377

target selectivity. Especially, in vitro and in vivo discrepancies of targeting ability of TSPO 378

ligands have been reported in many human brain studies [20-25, 27, 28, 33, 49, 50]. While 379

developing TSPO targeting ligands, we noticed that human rs6971 polymorphism was 380

important, and TSPO ligand binding was influenced by A147T mutation (Figure S1A). 381

Generally, TSPO with Ala at 147th amino acid is the abundant phenotype and high-affinity 382

ligand binder (HAB). Single point mutation of C to T in the TSPO gene induces amino acid 383

change (Ala to Thr at 147th amino acid). This phenotype is a low-affinity ligand binder 384

(LAB). PBR28 is a well-known TSPO ligand that shows a different binding ability to the 385

TSPO rs6971 polymorphism. 386

Although testing the sensitivity of the rs6971 polymorphism (A147T) for developing 387

clinically applicable TSPO targeting ligands is important, we could not find enough donors 388

with the rs6971 polymorphism to test the affinity for [18F]CB251 in Korea. About 30% of 389

Caucasian and 3% of Chinese/Japanese are reported as LABs (“TSPO polymorphism by 390

ethnic differences” in the Hapmap database http://hapmap.ncbi.nih.gov or 391

http://asia.ensembl.org/Homo_sapiens/Variation/Explore?r=22:43162420-392

43163420;v=rs6971;vdb=variation;vf=210028659). The LAB percentage of Koreans has not 393

been reported, but less than 3% of Koreans are expected to have LAB considering the genetic 394

similarity between the Chinese and Japanese. 395

Thus, we established cells that ectopically express polymorphic TSPOs (WT for 396

HAB or Mut-A147T for LAB). In vitro site mutagenesis was introduced in the coding 397

sequence of 147th Ala in TSPO (Figure S1B). To test the sensitivity of the rs6971 398

polymorphism, competitive inhibition assays of CB251 and other ligands with [3H]PK11195 399

were performed (Figure 1B). We calculated the IC50 value of each ligand to [3H]PK11195 400

from % inhibition in HAB or LAB. The ratios of IC50 LAB/HAB for each ligand were also 401

calculated. The ratio of IC50 values of PK11195 was consistent with previous reports (0.83 to 402

0.85 for PK11195 [34, 35]). Similar to PBR28, the ratios of IC50 LAB/HAB for fm-PBR28-d2 403

was 37.28 (55 for PBR28 [34, 35]), indicating lower binding to LAB. The ratios of IC50 404

LAB/HAB were 1.14 for CB251 and 3.96 for GE-180 [33]. After comparing the Ki 405

(generated from recombinant TSPO protein) and the ratio of IC50 LAB/HAB (generated from 406

membrane TSPO from cells), we concluded that CB251 had a higher binding affinity 407

(Ki=0.27 nM) and selectivity (ratio of IC50 LAB/HAB=1.14) than other TSPO ligands 408

regardless of TSPO polymorphism. 409

To test in vitro and in vivo selectivity for rs6971 polymorphism of [18F]CB251 as a 410

PET probe, [18F]CB251 cell uptake PET scans were performed. Our results showed that 411

[18F]CB251 had a higher binding affinity (Figure 1A) with favorable selectivity (Figure 1B- 412

E) in vitro and in vivo. However, there were limitations because cellular uptake of TSPO 413

ligands in humans can be influenced by factors such as cellular metabolism, endogenous 414

TSPO level, and in vivo kinetics. [11C]PK11195 and [11C]ER176 were both reported to be 415

significantly insensitive to TSPO polymorphism in vitro, however, in human studies, 416

[11C]PK11195 showed the different uptake of lung and heart in binders and non-binders [49]. 417

In addition, [11C]ER176 represented different radioactivity according to each genotype in the 418

human brain [50]. Therefore, it is difficult to conclude that in vitro uptake completely reflects 419

human kinetics. Unfortunately, we could not perform human study because it is difficult to 420

find Korean volunteers to express rs6971 A147T polymorphism. Evaluation of the ligand in 421

humans should be performed as a further study. 422

In this study, we also tested the specificity of [18F]CB251 as a TSPO-targeting ligand 423

for imaging activated immune cells (Figure 2). By modulating the TSPO expression levels in 424

the cells and immune cell activation, we found that [18F]CB251 was specific for TSPO 425

protein and its uptake was specifically observed in activated immune cells with high TSPO 426

expression. 427

The role of resident immune cells in the brain and peripheral immune cells that 428

infiltrate across the BBB in neuroinflammation has been studied and emphasized [34-36] as 429

the main target of neuroinflammation imaging. The resident CNS immune system usually 430

consists of innate immune cells and operates under homeostatic conditions. However, when 431

peripheral innate and adaptive immune cells enter the CNS, they execute distinct cell-432

mediated effects. Infiltration of peripheral immune cells across the BBB exacerbates 433

neuroinflammation resulting in higher levels of neurotoxicity [4, 8]. Therefore, it is important 434

to visualize peripheral immune cell infiltration to evaluate neuroinflammatory damages. 435

Herein, we tested whether [18F]CB251 PET signal reflects neuroinflammation compared to 436

other conventional methods such as MRI and BLI. 437

We adoptively transferred splenocytes from luciferase-expressing transgenic mice 438

into an intracranial LPS-injected mouse model to visualize the distribution of infiltrating 439

peripheral immune cells in the brain by BLI. The total population of splenocytes was used 440

because immune responses are initiated by the co-stimulation of different populations of 441

immune cells. Despite the limitations of optical imaging, the grafted splenocytes were found 442

in the inflamed region of the mouse brain (Figure 3). 443

We also visualized BBB disruption and immune cell localization by MRI with the 444

contrast agent, Gd-DOTA. When Gd-DOTA crosses the damaged BBB into the extracellular 445

space of brain parenchyma, its paramagnetic property is changed [41, 51]. T1-weighted MRI 446

contrast enhancement shows the rapid diffusion of gadolinium-based chelates from non-447

viable cells, while T2-weighted MRI contrast enhancement shows slow diffusion of 448

gadolinium-based chelates from viable immune cells [39]. We used T1-weighted MRI to 449

visualize gadolinium penetration through disrupted BBB. T2-weighted MRI was used to 450

monitor the vasogenic edema and localization of viable immune cells, including residential 451

microglia and peripheral immune cells due to LPS induced-neuroinflammation (Figure 4). In 452

our mouse model of neuroinflammation induced by intracranial injection of LPS, we opted to 453

use T2-weighted imaging. However, since the sensitivity to detect BBB disruption is limited, 454

there is a limitation that only MRI signal is not sufficient to evaluate BBB disruption. 455

Therefore, it is difficult to exclude the possibility of micro-disruptions of BBB. 456

To validate the usefulness of [18F]CB251 as a PET probe for imaging 457

neuroinflammation, we performed simultaneous PET/MRI and BLI (Figure 5), which 458

provided information on anatomical and functional changes with precise co-registration of 459

signals and accurate correction of attenuation [42, 43]. IHC staining showed increased TSPO 460

expression at the site of neuroinflammation. It also demonstrated that inflammatory 461

monocytes, macrophages, and antigen-presenting cells were present with resident immune 462

cells (i.e., microglia) at the site of inflammation, suggesting cooperation between the 463

peripheral and CNS immune system in the region of neuroinflammation [34-36]. BLI 464

revealed peripheral immune cell infiltration only in the region of the inflamed brain, but the 465

radioactive signal from [18F]CB251 showed the effects of both infiltrating peripheral immune 466

cells and resident microglia. 467

Furthermore, simultaneous [18F]CB251 PET/MR and comparative analysis of 468

[18F]CB251 PET and BLI were performed to evaluate the efficacy of an anti-inflammatory 469

agent and verify the usefulness of [18F]CB251 PET. CDDO-Me was used as an anti-470

inflammatory drug, which was reported to induce antioxidant proteins that prevent oxidative 471

damage from inflammation [44-48]. In our LPS-induced neuroinflammation model, treatment 472

with CDDO-Me reduced [18F]CB251 PET and BLI signals, indicating a decrease in 473

inflammatory response and recruitment of peripheral immune cells (Figure 6). Despite the 474

higher sensitivity of BLI, its signals were obscured by the skin and black fur of the mice in 475

our animal model. Because of the limitations of optical imaging, BLI can be used only for 476

preclinical studies. However, [18F]CB251 PET is a noninvasive imaging procedure that is 477

directly applicable to clinical investigations that provide diagnostic and therapeutic 478

information on neuroinflammation. 479

We demonstrated that [18F]CB251 is specific to TSPO-expressing inflammatory cells 480

with better selectivity for TSPO polymorphism. TSPO-targeting [18F]CB251 PET identified 481

the regions of neuroinflammation and also allowed us to evaluate the therapeutic effect of 482

anti-inflammatory treatments. In preclinical studies, our multimodal imaging approach using 483

PET/MR and BLI is expected to enable further understanding of the interactions between 484

CNS and the peripheral immune system during neuropathogenesis. Furthermore, [18F]CB251 485

PET/MRI, which provides information on anatomical and functional changes with precise co-486

registration of signals, can be an efficient tool for testing the efficacy of various anti-487

inflammatory drugs and treatment planning for both preclinical and clinical application. 488

Materials & Methods 489

Cell culture and reagents 490

RAW264.7, a mouse macrophage cell line and mouse splenocytes were used to evaluate 491

TSPO expression. LPS (50 ng/mL, Sigma-Aldrich, St.Louis, MO, USA) and IFN-γ (20 ng/mL, 492

Sigma-Aldrich) were used to activate immune cells. The 293FT (Invitrogen, Carlsbad, CA, 493

USA) and MDA-MB-231 (ATCC, Manassas, VA, USA) cell lines were used for various TSPO 494

expression levels. Cells were maintained in complete DMEM (for Raw264.7 and MDA-MB-495

231) or MEM (for 293FT) with 10% FBS and 1% antibiotics (penicillin-streptomycin). All 496

media and reagents for cell culture, including FBS and antibiotics were purchased from Gibco 497

(Grand Island, NY, USA). 498

499

TSPO expression in Cells 500

To express the genetic variants of TSPO proteins (Wild-type, WT; Mutant with 501

alanine-threonine at 147 amino acid, A147T), we generated different TSPO expression 502

vectors from the PBR (TSPO) (NM_000714) human tagged ORF clone RG220107 (Origene, 503

Rockville, MD, USA) by site mutagenesis (Figure S1B). We changed the nucleotide 504

sequences encoding alanine at amino acid 147 to threonine (A147T). We then transfected 505

these vectors into 293FT cells and performed an in vitro uptake test of these cells at 24 hours 506

after transfection. 507

For overexpression of the TSPO protein, the 293FT cell line was transiently 508

transfected with vectors (pCMV-eGFP/TSPO) (AddGene, Cambridge, MA, USA) harboring 509

the eGFP and TSPO coding sequences. MDA-MB-231, which highly expresses TSPO, was 510

used for the reduction of TSPO expression by shRNA (pCMV-eGFP/shTSPO; AddGene). 511

Each vector used for various levels of TSPO expression was transfected by Lipofectamine 512

2000 (Invitrogen, Waltham, MA, USA) at a total concentration of 2.5 µg/mL. Fluorescent 513

images were acquired by a confocal laser-scanning microscope (Leica TCS SP8; Wetzlar, 514

Hesse, Germany), and Western blotting was performed for verification. 515

516

Western blotting 517

Cell lysates were extracted by RIPA lysis buffer (Sigma-Aldrich), with a protease 518

inhibitor (Roche, Branchburg, NJ, USA) cocktail. Protein concentrations were measured by 519

the BCA protein assay kit (Pierce Endogen, Rockford, IL, USA), and samples containing 20 520

µg of proteins were separated on 10% polyacrylamide gels. Proteins were then transferred to 521

PVDF membranes (Millipore, Watford, Herts, UK), which were blocked by 5% skim milk in 522

Tris-buffered saline (20 mM Tris, 138 mM NaCl, and pH 7.4) with Tween-20 (0.1%) for 1 h. 523

Membranes were then incubated with primary antibodies followed by incubation with 524

secondary antibodies. The following antibodies were used: anti-TSPO (Abcam, Cambridge, 525

Cambs, UK), anti-iNOS (Santa-Cruz Biotechnology, Santa-Cruz, CA, USA), anti-β-actin 526

(Sigma-Aldrich) were anti-rabbit IgG (Cell Signaling Technology, Danvers, MA, USA), and 527

anti-mouse IgG (Cell Signaling Technology) for β-actin. 528

529

Radiochemical synthesis of radioactive TSPO-targeting ligands 530

The [11C]PK11195 [21], deuterium-substituted fluoromethyl-PBR28 ([18F]fmPBR28-531

d2 [52]), and [18F]GE-180 [33] were synthesized according to previous studies. The 532

[18F]CB251 probe was also synthesized following a protocol described in a previous study 533

[26]. 534

535

Membrane protein extraction 536

Membrane protein extraction was conducted by protocols from the membrane protein 537

extraction kit (Mem-PER plus kit, Thermo Fisher, Waltham, MA, USA). 293FT cells (5 x 538

106/100 Ø) were seeded for transient transfection and transfected with TSPO WT or A147T 539

vector for 24 hours. Cells were harvested using a scraper and washed twice with cell wash 540

solution. After centrifugation (300 x g, 5 min), the cell pellet was permeabilized for 10 min at 541

4°C and centrifuged (16000 x g, 15 min, 4°C). The supernatant containing cytosolic proteins 542

was removed, and the pellet was solubilized with the buffer from the Mem-PER plus kit for 543

30 min at 4°C. The solubilized pellet was centrifuged again (16000 x g, 15 min, 4°C), and the 544

supernatant containing membrane-associated proteins was used for competitive binding 545

assay. 546

547

Competitive inhibition assay 548

To verify the binding affinity of TSPO-targeting ligand for the TSPO polymorphism 549

rs6971, we performed competitive inhibition assay [34, 35] with [3H]PK11195 550

(NET885250UC, 9.25 MBq, Perkin Elmer, Boston, MA, USA) and 3 different TSPO-551

targeting ligands (PK11195, fmPBR28-d2, and CB251). Membrane proteins (100 µg/100 µL) 552

isolated from 293FT cells expressing TSPO WT (HABs) or A147T (LABs) were incubated 553

with [3H]PK11195 (0.019 MBq/100 µL) at different concentrations of each ligand (from 0.1 554

nM to 1000 µM) in assay buffer (50 mM Tris base, 140 mM NaCl, 1.5 mM MgCl2, 5 mM 555

KCl, 1.5 mM CaCl2, pH 7.4) for 1 hour at 37°C. Subsequently, the mixtures were filtrated 556

through Whatman filter paper under vacuum and washed three times with the wash buffer (50 557

mM Tris base, 1 mM MgCl2, pH7.4). Filters were harvested, and the radioactivity of each 558

filter in scintillation buffer (Perkin Elmer) was counted using β-counter (Liquid scintillation 559

analyzer Tri-Carb3100 TR, Perkin Elmer, Waltham, MA, USA). IC50 values of each ligand to 560

[3H]PK11195 in HAB or LAB were measured. The ratio of IC50 values of each ligand to 561

[3H]PK11195 in the LAB to HAB was also calculated, representing the sensitivity of TSPO 562

polymorphism. A value close to 1 indicates lower sensitivity to TSPO polymorphism. 563

564

Measurement of cell uptake of radioactive ligands 565

Cells were seeded in 6-well plates (Sigma-Aldrich) until they reached approximately 566

80% confluency. Before the uptake experiments, the cells were preincubated with glucose-567

free RPMI-1640 medium (Gibco) for 4 h. The cells were then trypsinized (Gibco), counted, 568

and 1×105 cells transferred into a 5 mL test tube containing warmed Hank’s balanced salt 569

solution (HBSS, Sigma-Aldrich) with 0.5% (w/v) bovine serum albumin (Sigma-Aldrich). 570

Subsequently, approximately 0.185 MBq of radioactive ligands were added to the tubes, and 571

the cells were incubated in a humidified incubator with 5% CO2 for 1 h at 37°C. The cells 572

were then washed 3 times with cold HBSS and lysed for 5 min in 200 μL of 1% sodium 573

dodecyl sulfate (SDS, Sigma-Aldrich). Cell lysates were collected, and radioactivity was 574

measured by a Cobra II gamma counter (Canberra Packard; Vaughan, Ontario, Canada). 575

Radioactivity was normalized according to the amount of total protein at the time of assay. 576

Total protein levels were quantified by the BCA protein assay kit (Pierce, Rockford, IL, 577

USA). All experiments were performed in triplicate. 578

579

Preparation of splenocytes from B6. LucTg mice 580

Luciferase-expressing cells were isolated from the dissected spleens of transgenic 581

C57BL/6lucTg mice and centrifuged at 1300 rpm for 5 min at 4 °C). The red blood cell lysis 582

buffer (1 mL/spleen, Sigma-Aldrich) was added to the cell pellet, and the mixture was 583

incubated at room temperature for 3 min to remove red cells. After a double volume of RPMI 584

medium was added to the cells, 2-3 volumes of red blood cell lysis buffer were added until 585

complete lysis of red blood cells. Subsequently, the cells were re-suspended, ensuring a 586

suspension of single cells, and filtered through a cell strainer (40 µm nylon, Falcon, Oneonta, 587

NY). Cells (3.175 × 105 to 5 × 106) were then plated in 24-well plates, and bioluminescent 588

images were acquired after the addition of D-luciferin (300 µg/mL, Caliper Life Science, 589

Hopkinton, MA) to measure luciferase activity. Bioluminescence intensity was measured by 590

an IVIS 100 system (Perkin Elmer, Waltham, MA). Cells (2 × 107) were adoptively 591

transferred by intravenous injection into recipient C57BL/6 mice to monitor the in vivo 592

distribution of luciferase-expressing immune cells. All experimental designs and procedures 593

for animal experiments were approved by the Institutional Animal Care and Use Committee 594

of Seoul National University and in accordance with the ethical standards of Seoul National 595

University Hospital guidelines of Seoul National University Hospital. 596

597

Intracranial inflammation mouse model by LPS administration 598

C57BL/6 mice (male, 6 weeks old) were used for generating the intracranial 599

inflammation model. LPS (Sigma-Aldrich, 5 µg/2 µL dissolved in saline) was used to induce 600

local inflammation in the right striatum of the mouse brain (anteroposterior, +1.8; 601

mediolateral, +2.0; dorsoventral, -3.0) via intracranial injection by a robotic stereotaxic 602

device (Stoelting, Wood Dale, IL) equipped with a Hamilton syringe at a rate of 0.25 µL/min 603

[52-54]. Saline was injected as a control. The mouse was placed in the isoflurane induction 604

box (5% isoflurane, 95% air, 250 cc/min) until the animal was fully anesthetized, and the 605

isoflurane percentage was adjusted to 2 % at a flow rate of 250 cc/min during the process. 606

607

[18F]CB251 PET/MR imaging 608

To visualize neuroinflammation, PET/MR scans were acquired by a simultaneous 609

small animal PET insert (SimPET, Brightonic imaging, Seoul, Korea) combined with an M7 610

1.0 T MRI system (Aspect Imaging, Jerusalem, Israel). [18F]CB251 (9.25 to 11.1 MBq per 611

mouse) was injected intravenously, and images were acquired 4 days after intracranial 612

injection of LPS. Mice were anesthetized with 1.5% isoflurane during imaging. 613

To evaluate breach of the BBB, gadolinium (Gd-DOTA, 0.5 mmol/kg) was injected 614

intravenously, and T1- and T2-weighted MR scans were acquired [26]. The T1-weighted fast-615

spin echo sequence consisted of a repetition time of 200 msec and an echo time of 10 msec. 616

The T2-weighted fast spin echo sequence consisted of a repetition time of 3000 msec and an 617

echo time of 63.84 msec. 618

A PET/MR scan was performed simultaneously for 40 min with a 30 x 30 mm field-619

of-view (FOV) at 20 min post-[18F]CB251 injection. PET images were reconstructed using 620

3D OSEM algorithm (12 subsets and 3 iterations). An ellipsoidal region-of-interest was 621

drawn around the peak PET activity on the right striatum in a coronal PET/MRI slice, and the 622

region-of-interest was copied to the left striatum using AMIDE program (ver 0.9.0, 623

http://amide.sourceforge.net). PET and MR images were reconstructed and analyzed by the 624

AMIDE program (ver 0.9.0, http://amide.sourceforge.net). The ratio of maximum PET SUV 625

in the right and left striatum was then calculated. 626

627

Bioluminescence imaging 628

An in vivo IVIS 100 bioluminescence/optical imaging system (Xenogen, Alameda, 629

CA, USA) was used to follow luciferase-expressing peripheral immune cells. For the ex vivo 630

imaging of luciferase-expressing peripheral immune cells in the brain, D-luciferin (3 631

mg/mouse) was intravenously injected into the tail vein at 10 min before sacrificing the 632

animal in a CO2 chamber. The bioluminescent signal was acquired and analyzed by Living 633

Image ver.2.50.2 software (Xenogen). 634

635

Immunohistochemistry 636

Brain tissues from mice were fixed in 4% paraformaldehyde, embedded in paraffin, 637

and sectioned into 4 µm-thick segments. After heat-induced antigen epitope retrieval, slides 638

were blocked with PBS containing 3% normal serum and incubated overnight at 4 °C with 639

primary antibodies as follows: anti-CD68, anti-CD86 (Santa Cruz Biotechnology, Santa Cruz, 640

CA, USA), anti-CD4, anti-CD8, anti-Iba-1, and anti-TSPO (Abcam, Cambridge, UK). The 641

slides were then washed with PBS and incubated with biotinylated secondary antibodies 642

(anti-goat for CD68; anti-rabbit for CD4, CD8, and Iba-1; all from Santa Cruz 643

Biotechnology). The slides were then treated with avidin-biotin solution (Vector Laboratories, 644

Burlingame, CA, USA) for 1 h. The antigenic signal was developed using DAB (3, 3 –645

diaminobenzidine) substrate (Vector Laboratories, Burlingame, CA, USA), according to the 646

manufacturer’s instructions. Finally, the slides were counterstained with hematoxylin and 647

mounted with the Permount Mounting Medium (Thermo Fisher Scientific, Fair Lawn, NJ, 648

USA). 649

650

CDDO-methyl ester (CDDO-Me) treatment 651

2-Cyano-3,12-dioxo-oleana-1,9(11)-dien-28-oic acid methyl ester (CDDO-methyl 652

ester or CDDO-Me, Sigma-Aldrich) was dissolved in DMSO to produce a 10 mM stock 653

solution. The stock solution was further diluted with 7.5% PBST (1.5 ml Tween-20 dissolved 654

in 20 ml PBS) to obtain a 200 nM CDDO-ME solution. Mice received 100 µL of CDDO-ME 655

solution via intraperitoneal injection administered once daily for 3 days beginning 1 day after 656

intracranial administration of LPS. 657

658

Statistical analysis 659

All results were calculated as means ± standard deviation (SD). Statistically 660

significant differences were analyzed by a paired 2-sample Student t-test. Statistical 661

significance was considered to be p<0.05. GraphPad Prism 5 software (San Diego, CA) was 662

used for statistical analysis.663

Acknowledgments 664

This study was supported by grants the from the National Research Foundation (NRF), 665

funded by the Ministry of Science and ICT (MSIT) of Korea (No. 20110030001, 666

2020R1A2C2011695 to H.Y.), grants from the Korea Health Industry Development Institute 667

(KHIDI), and by the Ministry of Health & Welfare of Korea (HI16C0947 to S.E.K. and 668

B.C.L.). 669

670

Author Contributions 671

K. K., Y-H. K., J. N., and C-H. L. performed transfections, Western blots, 672

immunohistochemistry, confocal microscopy, FACS analysis, animal imaging, and 673

interpretation of data. S-Y L. generated the cells that expressed TSPO polymorphisms 674

(genetic variants) and performed the evaluation of the in vitro cell uptake of radio-labeled 675

compounds. H.K. and S-H. B. maintained transgenic animals and isolated immune cells from 676

transgenic mice. M.S.L., G.B.K., and J.S.L. analyzed and interpreted simultaneous PET/MR 677

imaging. S.H.L., K.Y.K, J.H.J., I.H.S, and B.C.L. synthesized the described compound, 678

performed radiosynthesis of the described radiotracers, and participated in scientific 679

discussions. J.C.P., G.J.C., K.W.K., S.E.K., J-K C., E.E.K., and S-H P. also participated in 680

scientific discussions. K.K. interpreted the data and wrote the initial draft of the manuscript. 681

H.Y. supervised projects, contributed to the conception of the study, interpretation of data, 682

and wrote the final manuscript. This manuscript has been seen and approved by all authors, 683

who contributed to preparing the manuscript, and no person other than the authors listed has 684

contributed significantly to its preparation. 685

686

Competing Interests 687

Guen Bae Ko, Kyeong Yun Kim, and Jae Sung Lee declare that they are employees of 688

Brightonix Imaging Incorporation (for PET/MR). Other authors have nothing to declare. 689

Figure Legends 690

Figure 1. [18F]CB251 as a TSPO-targeting radioactive ligand. (A) Various TSPO ligands 691

and the structure of TSPO-targeting radioactive ligands. CB251 is shown compared with 692

known TSPO ligands, such as PK11195, PBR28, fm-PBR-d2, and GE-180. [18F]fmPBR28-693

d2 is fluoromethyl-PBR28-d2 (ref. 24, 30). Ki of ligands for TSPO from refs. 47a, 50b, 31c, 694

and 24d. (B) Competitive inhibition assay of PK11195, fmPBR28-d2, and CB251 with 695

[3H]PK11195 on membrane proteins . HAB indicates high affinity ligand binding phenotype 696

(wild type TSPO); LAB indicates low affinity ligand binding phenotype (A147T mutant) 697

isolated from 293FT cells expressing polymorphic TSPOs. CPM indicates count per minute. 698

The ratio of IC50 LAB/HAB represents the sensitivity of TSPO polymorphism; a value close 699

to 1 indicates less sensitive to TSPO polymorphism. (C) In vitro cellular uptake of CB251 in 700

293FT cells expressing polymorphic TSPO was compared with TSPO ligands. (D) PET scan 701

images of 293FT cells expressing polymorphic TSPOs (E) Representative PET/MR images 702

of mice bearing 293FT cells with polymorphic TSPOs (N=2 for each treatment) 703

Figure 2. [18F]CB251 as a TSPO-targeting ligand for imaging activated immune cells. 704

(A) TSPO expression in 293FT cells was increased by transfection with pCMV-TSPO/eGFP. 705

(B) [18F]CB251 uptake was increased in TSPO-overexpressing cells. (C) TSPO expression in 706

MDA-MB-231 cells was reduced by transfection with pCMV-shTSPO/eGFP. (D) 707

[18F]CB251 uptake was decreased in cells with reduced TSPO expression. (E) Western blots 708

of TSPO and iNOS (activated immune cell marker) in a mouse macrophage cell line 709

(RAW264.7) and mouse splenocytes. (F) [18F]CB251 uptake was increased in activated cells 710

(both in Raw264.7 and mouse splenocytes). All experiments were replicated at least three 711

times. 712

Figure 3. Bioluminescence imaging of peripheral immune cells in a mouse 713

neuroinflammation model using intracranial LPS injection. (A) Flow cytometry analysis 714

of immune cells from splenocytes of C57BL/6 (wild type) and C57BL/6.LucTg (reporter 715

transgenic) mice. (B) Semiquantitative analysis of luciferase-expressing mouse splenocytes 716

from C57BL/6.LucTg. The intensity of luciferase signal and the number of cells were highly 717

correlated (R2=0.9991) (C) Schematic representation of experimental design using luciferase-718

expressing splenocytes to monitor peripheral immune cell infiltration in the brain of a mouse 719

with neuroinflammation induced by intracranial LPS injection. (D) Representative 720

biodistribution of luciferase-expressing splenocytes in mice 4 days after intracranial injection 721

of saline or LPS (N=2 for each treatment). 722

Figure 4. Representative MR scans using gadolinium-DOTA to monitor the Blood-Brain 723

Barrier (BBB) disruption and immune cell activation. (A) Intracranial LPS injection 724

model (saline as a control). (B) T1-weighted MR scans and (C) T2-weighted MR scans on 725

day 1, 3, and 4 after intracranial injection (N=3). T1-weighted MRI was used to visualize the 726

presence of gadolinium penetration through disrupted BBB, and T2-weighted MRI was used 727

to monitor the vasogenic edema and localization of viable immune cells due to LPS induced-728

neuroinflammation. 729

Figure 5. Visualization of the intracranial LPS-induced neuro-inflammatory region 730

using multi-modal imaging. (A) Experimental scheme of simultaneous PET/MR using 731

[18F]CB251 and BLI. (B) Representative images of [18F]CB251 PET/MR and 732

bioluminescence (BLI). The red circle indicates a region of interest (ROI). (C) Signal 733

intensity ratio of the right striatum (injected with saline or LPS) to the left striatum (no 734

injection) in the saline-injected (N=6) and LPS-injected (N=7) mice. PET signal was 735

expressed as standard uptake value (SUVmax) and the BLI signal was expressed as total flux 736

(photon/sec/cm2/sr). (D) IHC staining of intracranial LPS injected mouse brain tissues. Iba-1 737

(macrophages/microglia marker), CD68 (macrophages/monocytes marker), CD86 (antigen-738

presenting cells), CD4 (helper T cells), and CD8 (cytotoxic T cells) were used as markers for 739

each cell type. 740

Figure 6. Evaluation of the therapeutic effect of the anti-inflammatory drug (CDDO-741

Me) using [18F]CB251 PET and BLI. (A) Experimental design of treatments and imaging 742

schedules, including simultaneous [18F]CB251 PET/MRs and BLIs. (B) Representative 743

[18F]CB251 PET/MR scans and SUVmax ratios were plotted. (C) Representative BLI image 744

and total flux ratio were plotted. Signal ratios from each modality (SUVmax ratio for PET; 745

Total flux ratio for BLI) were calculated from the region of interest (ROI) of right 746

(intracranially injected with saline, LPS, or LPS plus anti-inflammatory agent CDDO-Me) 747

and left striatum (non-injected) in the mouse brain. The values are mean ± SD. (N=8 for 748

saline-treated; N=12 for LPS-treated; N=5 for LPS plus CDDO-Me). 749

Figure S1. TSPO genotypes and classfication. (A) TSPO ligand binding classification. Single 750

nucleotide polymorphism rs6971 TSPO (CT point mutation; rs6971 A147T; Ala to Thr at 751

amino acid 147 of TSPO) influences the binding affinity of TSPO ligands. TSPO with Thr/Thr 752

at amino acid 147 shows a lower binding affinity to its ligand. TSPO ligand binding phenotypes 753

are HAB or LAB, which indicate high- or low-affinity binding of ligands, respectively 754

*Adapted from ref 32, 33. (B) Expression vector design for TSPO mutagenesis 755

References 756

1. Chen WW, Zhang X, Huang WJ. Role of neuroinflammation in neurodegenerative 757

diseases (Review). Mol Med Rep. 2016; 13: 3391-6. 758

2. Barrientos RM, Kitt MM, Watins, LR, Maier, SF. Neuroinflammation in the normal 759

aging hippocampus. Neuroscience. 2015; 309: 84-99. 760

3. Amor S, Peferoen LA, Vogel DY, Breur M, van der Valk P, Baker D, et al. Inflammation 761

in neurodegenerative diseases-an update. Immunology. 2014; 142: 151-66. 762

4. Louveau A, Smirnov I, Keyes TJ, Eccles JD, Rouhani SJ, Peske JD, et al. Structural 763

and functional features of central nervous system lymphatic vessels. Nature. 2015; 523: 764

337-41. 765

5. Russo MV, McGavern DB. Immune surveillance of the CNS following infection and 766

injury. Trends Immunol. 2015; 36: 637-650. 767

6. Pavlov VA, Tracey KJ. Neural regulation of immunity: molecular mechanisms and 768

clinical translation. Nat Neurosci. 2017. 20: 156-166. 769

7. Banks WA, Gray AM, Erickson MA, Salameh TS, Damodarasamy M, Sheibani N, et 770

al. Lipopolysaccharide-induced blood-brain barrier disruption: roles of cyclooxygenase, 771

oxidative stress, neuroinflammation, and elements of the neurovascular unit. J 772

Neuroinflammation. 2015; 12: 223. 773

8. Rezai-Zadeh K, Gate D, Town T. CNS infiltration of peripheral immune cells: D-day 774

for neurodegenerative disease? J Neuroimmune Pharmacol. 2009; 4: 462-75. 775

9. Withana NP, Saito T, Ma X, Garland M, Liu C, Kosuge H, et al. Dual-modality activity-776

based probes as molecular imaging agents for vascular inflammation. J Nucl Med. 2016; 777

57: 1583-1590. 778

10. Airas L, Rissanen E, Rinne J. Imaging of microglial activation in MS using PET: 779

Research use and potential future clinical application. Mult Scler. 2017; 23: 496-504. 780

11. Cagnin A, Gerhard A, Banati RB. In vivo imaging of neuroinflammation. Eur 781

Neuropsychopharmacol. 2002; 12: 581-6. 782

12. Zinnhardt B, Wiesmann M, Honold L, Barca C, Schäfers M, Kiliaan AJ, et al. In vivo 783

Imaging Biomarkers of Neuroinflammation in the Development and Assessment of 784

Stroke Therapies - Towards Clinical Translation. Theranostics. 2018; 8: 2603-2620. 785

13. Gauberti M, Fournier AP, Docagne F, Vivien D, Martinez de Lizarrondo S. Molecular 786

Magnetic Resonance Imaging of Endothelial Activation in the Central Nervous System. 787

Theranostics. 2018; 8: 1195-1212. 788

14. Winkeler A, Boisgard R, Martin A, Tavitial B. Radioisotopic imaging of 789

neuroinflammation. J Nucl Med. 2010; 51: 1-4. 790

15. Nakamura K, Tukatani Y, Kubo A, Hashimoto S, Terayama Y, Amano T, et al. The 791

behavior of 99mTc-hexamethylpropyleneamineoxime (99mTc-HMPAO) in blood and 792

brain. Eur J Nucl Med. 1989; 15: 100-7. 793

16. Jin AY, Tuor UI, Rushforth D, Filfil R, Kaur J, Ni F, et al. Magnetic resonance molecular 794

imaging of post-stroke neuroinflammation with a P-selectin targeted iron oxide 795

nanoparticle. Contrast Media Mol Imaging. 2009; 4: 305-11. 796

17. Jacobs AH, Tavitian B, consortium INMiND. Noninvasive molecular imaging of 797

neuroinflammation. J Cereb Blood Flow Metab. 2012; 32: 1393-415. 798

18. Esposito G, Giovacchini G, Liow JS, Bhattacharjee AK, Greenstein D, Schapiro M, et 799

al. Imaging neuroinflammation in Alzheimer's disease with radiolabeled arachidonic 800

acid and PET. J Nucl Med. 2008; 49: 1414-21. 801

19. Hatori A, Yui J, Yamasaki T, Xie L, Kumata K, Fujinaga M, et al. PET imaging of lung 802

inflammation with [18F]FEDAC, a radioligand for translocator protein (18 kDa). PLoS 803

One. 2012; 7: e45065. 804

20. Varrone A, Mattsson P, Forsberg A, Takano A, Nag S, Gulyas B, et al. In vivo imaging 805

of the 18-kDa translocator protein (TSPO) with [18F]FEDAA1106 and PET does not 806

show increased binding in Alzheimer's disease patients. Eur J Nucl Med Mol Imaging. 807

2013; 40: 921-31. 808

21. Kumar A, Muzik O, Shandal V, Chugani D, Chakraborty P, Chugani HT. Evaluation of 809

age-related changes in translocator protein (TSPO) in human brain using (11)C-[R]-810

PK11195 PET. J Neuroinflammation. 2012; 9: 232. 811

22. Eberl S, Katsifis A, Peyronneau MA, Wen L, Henderson D, Loc'h C, et al. Preclinical 812

in vivo and in vitro comparison of the translocator protein PET ligands [18F]PBR102 813

and [18F]PBR111. Eur J Nucl Med Mol Imaging. 2017; 44: 296-307. 814

23. Zheng J, Boisgard R, Siquier-Pernet K, Decaudin D, Dolle F, Tavitian B. Differential 815

expression of the 18 kDa translocator protein (TSPO) by neoplastic and inflammatory 816

cells in mouse tumors of breast cancer. Mol Pharm. 2011; 8: 823-32. 817

24. Kim T, Pae AN. Translocator protein (TSPO) ligands for the diagnosis or treatment of 818

neurodegenerative diseases: a patent review (2010-2015; part 1). Expert Opin Ther Pat. 819

2016; 26: 1325-1351. 820

25. Kim T, Pae AN. Translocator protein (TSPO) ligands for the diagnosis or treatment of 821

neurodegenerative diseases: a patent review (2010 - 2015; part 2). Expert Opin Ther 822

Pat. 2016; 26: 1353-1366. 823

26. Perrone M, Moon BS, Park HS, Laquintana V, Jung JH, Cutrignelli A, et al. A novel 824

PET imaging probe for the detection and monitoring of translocator protein 18 kDa 825

expression in pathological disorders. Sci Rep. 2016; 6: 20422. 826

27. Wang Y, Coughlin JM, Ma S, Endres CJ, Kassiou M, Sawa A, et al. Neuroimaging of 827

translocator protein in patients with systemic lupus erythematosus: a pilot study using 828

[11C]DPA-713 positron emission tomography. Lupus. 2017; 26: 170-178. 829

28. Kuntner C, Kesner AL, Bauer M, Kremslehner R, Wanek T, Mandler M, et al. 830

Limitations of small animal PET imaging with [18F]FDDNP and FDG for quantitative 831

studies in a transgenic mouse model of Alzheimer's disease. Mol Imaging Biol. 2009; 832

11: 236-40. 833

29. Buscombe J. PET imaging of inflammation. Q J Nucl Med Mol Imaging. 2014; [Epub 834

ahead of print]. 835

30. Wu C, Li F, Niu G, Chen X. PET imaging of inflammation biomarkers. Theranostics. 836

2013; 3: 448-66. 837

31. Le Fur G, Guilloux F, Rufat P, Benavides J, Uzan A, Renault C, et al. Peripheral 838

benzodiazepine binding sites: effect of PK 11195, 1-(2-chlorophenyl)-N-methyl-(1-839

methylpropyl)-3 isoquinolinecarboxamide. II. In vivo studies. Life Sci. 1983; 32: 1849-840

56. 841

32. Denora N, Laquintana V, Pisu MG, Dore R, Murru L, Latrofa A, et al. 2-Phenyl-842

imidazo[1,2-a]pyridine compounds containing hydrophilic groups as potent and 843

selective ligands for peripheral benzodiazepine receptors: synthesis, binding affinity 844

and electrophysiological studies. J Med Chem. 2008; 51: 6876-88. 845

33. Wadsworth H, Jones PA, Chau WF, Durrant C, Fouladi N, Passmore J, et al. [18F]GE-846

180: a novel fluorine-18 labelled PET tracer for imaging Translocator protein 18 kDa 847

(TSPO). Bioorg Med Chem Lett. 2012; 22: 1308-13. 848

34. Owen DR, Howell OW, Tang SP, Wells LA, Benacef I, Bergstrom M, et al. Two binding 849

sites for [3H]PBR28 in human brain: implications for TSPO PET imaging of 850

neuroinflammation. J Cereb Blood Flow Metab. 2010; 30: 1608-18. 851

35. Owen DR, Gunn RN, Rabiner EA, Bennacef I, Fujita M, Kreisl WC, et al. Mixed-852

affinity binding in humans with 18-kDa translocator protein ligands. J Nucl Med. 2011; 853

52: 24-32. 854

36. Unger MS, Schernthaner P, Marschallinger J, Mrowetz H, Aigner L. Microglia prevent 855

peripheral immune cell invasion and promote an anti-inflammatory environment in the 856

brain of APP-PS1 transgenic mice. J Neuroinflammation. 2018; 15: 274. 857

37. Negrin RS, Contag CH. In vivo imaging using bioluminescence: a tool for probing 858

graft-versus-host disease. Nat Rev Immunol. 2006; 6: 484-90. 859

38. Song MG, Kang B, Jeon JY, Chang J, Lee S, Min CK, et al. In vivo imaging of 860

differences in early donor cell proliferation in graft-versus-host disease hosts with 861

different pre-conditioning doses. Mol Cells. 2012; 33: 79-86. 862

39. Nwe K, Bernardo M, Regino CA, Williams M, Brechbiel MW. Comparison of MRI 863

properties between derivatized DTPA and DOTA gadolinium-dendrimer conjugates. 864

Bioorg Med Chem. 2010; 18: 5925-31. 865

40. Ngen EJ, Wang L, Kato Y, Krishnamachary B, Zhu W, Gandhi, N, et al. Imaging 866

transplanted stem cells in real time using an MRI dual-contrast method. Sci Rep. 2015; 867

5: 13628. 868

41. Lemasson B, Serduc R, Maisin C, Bouchet A, Coquery N, Robert P, et al. Monitoring 869

blood-brain barrier status in a rat model of glioma receiving therapy: dual injection of 870

low-molecular-weight and macromolecular MR contrast media. Radiology. 2010; 257: 871

342-52. 872

42. Heiss WD. The potential of PET/MR for brain imaging. Eur J Nucl Med Mol Imaging, 873

2009. 36 (Suppl 1): S105-12. 874

43. Zaidi H, Mawlawi O, Orton CG. Point/counterpoint. Simultaneous PET/MR will 875

replace PET/CT as the molecular multimodality imaging platform of choice. Med Phys. 876

2007; 34: 1525-8. 877

44. Ahmad R, Raina D, Meyer C, Kufe D. Triterpenoid CDDO-methyl ester inhibits the 878

Janus-activated kinase-1 (JAK1) ->signal transducer and activator of transcription-3 879

(STAT3) pathway by direct inhibition of JAK1 and STAT3. Cancer Res. 2008; 68: 2920-880

6. 881

45. Thimmulappa RK, Fuchs RJ, Malhotra D, Scollick C, Traore K, Bream JH, et al. 882

Preclinical evaluation of targeting the Nrf2 pathway by triterpenoids (CDDO-Im and 883

CDDO-Me) for protection from LPS-induced inflammatory response and reactive 884

oxygen species in human peripheral blood mononuclear cells and neutrophils. Antioxid 885

Redox Signal. 2007; 9: 1963-70. 886

46. Vannini N, Lorusso G, Cammarota R, Barberis M, Noonan DM, Sporn MD, et al. The 887

synthetic oleanane triterpenoid, CDDO-methyl ester, is a potent antiangiogenic agent. 888

Mol Cancer Ther. 2007; 6: 3139-46. 889

47. Buendia I, Michalska P, Navarro E, Gameiro I, Egea J, Leon R. Nrf2-ARE pathway: 890

An emerging target against oxidative stress and neuroinflammation in 891

neurodegenerative diseases. Pharmacol Ther. 2016; 157: 84-104. 892

48. Wang YY, Yang YX, Zhe H, He ZX, Zhou SF. Bardoxolone methyl (CDDO-Me) as a 893

therapeutic agent: an update on its pharmacokinetic and pharmacodynamic properties. 894

Drug Des Devel Ther. 2014; 8: 2075-88. 895

49. Kreisl WC, Fujita M, Fujimura Y, Kimura N, Jenko KJ, Kannan P, et al. Comparison of 896

[11C]-(R)-PK 11195 and [11C]PBR28, two radioligands for translocator protein (18 897

kDa) in human and monkey: Implications for positron emission tomographic imaging 898

of this inflammation biomarker. Neuroimage. 2010; 49: 2924-32. 899

50. Ikawa M, Lohith TG, Shrestha S, Telu S, Zoghbi SS, Castellano S, et al. 11C-ER176, a 900

radioligand for 18-kDa translocator protein, has adequate sensitivity to robustly image 901

all three affinity genotypes in human brain. J Nucl Med. 2017; 58: 320-325. 902

51. Herborn CU, Honold E, Wolf M, Kemper J, Kinner S, Adam G, et al. Clinical safety 903

and diagnostic value of the gadolinium chelate gadoterate meglumine (Gd-DOTA). 904

Invest Radiol. 2007; 42: 58-62. 905

52. Moon BS, Jung JH, Park HS, Contino M, Denora N, Lee BC, et al. Preclinical 906

comparison study between [18F]fluoromethyl-PBR28 and its deuterated analog in a rat 907

model of neuroinflammation. Bioorg Med Chem Lett. 2018; 28: 2925-2929. 908

53. Herber DL, Mercer M, Roth LM, Symmonds K, Maloney J, Wilson K, et al. Microglial 909

activation is required for Abeta clearance after intracranial injection of 910

lipopolysaccharide in APP transgenic mice. J Neuroimmune Pharmacol. 2007; 2: 222-911

31. 912

54. Catorce MN, Gevorkian G. LPS-induced murine neuroinflammation model: Main 913

features and suitability for pre-clinical assessment of nutraceuticals. Curr 914

Neuropharmacol. 2016; 14: 155-64. 915

916

D E

NN

O

11CH3

Cl

[11C]PK11195

N

NOO

18F

[18F]GE180

NN

Cl

ClO

N

O

18F

[18F]CB251

N N

O

O CH3

OCD218F

[18F]fmPBR28-d2

Compound Ki (nM)PK11195 1.4a

PBR28 6.1a

fmPBR28-d2 3.6b

GE-180 0.87c

CB251 0.27d

ATSPO ligands

3.30.2

coronal

sagittal

[18F]CB251[18F]fmPBR28-d2

WT Mut

0.1 2.1

WT

Mut

WT

Mut

[18F]CB251[18F]fmPBR28-d2

1.0 30.1

WT

Mut

WT

Mut

FIGURE 1. [18F]CB251 as a TSPO-targeting radioactive ligand (A) Various TSPO ligands and the structure of TSPO-targeting radioactive ligands. CB251 is shown compared with known TSPO ligands, such as PK11195, PBR28, fm-PBR-d2, and GE-180. [18F]fmPBR28-d2 is fluoromethyl-PBR28-d2 (ref. 24, 30). Ki of ligands for TSPO from refs. 47a, 50b, 31c, and 24d. (B) Competitive inhibition assay of PK11195, fmPBR28-d2, and CB251 with [3H]PK11195 on membrane proteins . HAB indicates high affinity ligand binding phenotype (wild type TSPO); LAB indicates low affinity ligand binding phenotype (A147T mutant) isolated from 293FT cells expressing polymorphic TSPOs. CPM indicates count per minute. The ratio of IC50 LAB/HAB represents the sensitivity of TSPO polymorphism; a value close to 1 indicates less sensitive to TSPO polymorphism. (C) In vitro cellular uptake of CB251 in 293FT cells expressing polymorphic TSPO was compared with TSPO ligands. (D) PET scan images of 293FT cells expressing polymorphic TSPOs (E) Representative PET/MR images of mice bearing 293FT cells with polymorphic TSPOs (N=2 for each treatment)

C100806040200

% u

ptak

e

[11C]PK11195 [18F]fmPBR28-d2 [18F]GE-180 [18F]CB251

Ctr WT Mut(A147T)

Ctr WT Mut(A147T)

Ctr WT Mut(A147T)

Ctr WT Mut(A147T)

B

-11 -10 -9 -8 -7 -6 -5 -40

50

100

150

Log10 (M)%

CP

M-11 -10 -9 -8 -7 -6 -5 -4

0

50

100

150

Log10 (M)

% C

PM

LAB (Mut) IC50=7.61 µMHAB (WT) IC50=0.204 µM

LAB (Mut) IC50=0.305 µMHAB (WT) IC50=0.267 µM

-11 -10 -9 -8 -7 -6 -5 -40

50

100

150

200

Log10 (M)

% C

PM

LAB (Mut) IC50=5.01 µMHAB (WT) IC50=6.05 µM

[11C]PK11195 [18F]fmPBR28-d2 [18F]CB251

LAB/HAB of IC50= 0.83 LAB/HAB of IC50= 37.28 LAB/HAB of IC50= 1.14

%C

PM

%C

PM

200

150

100

50

0%

CPM

150

100

50

0

150

100

50

0-11 -10 -9 -8 -7 -6 -5 -4 Log10(M)-11 -10 -9 -8 -7 -6 -5 -4 Log10(M)-11 -10 -9 -8 -7 -6 -5 -4 Log10(M)

20

15

10

5

0

[18F]

CB

251

upta

ke

( 106

cpm

/mg)

ctr TSPO/eGFP(+)

P <0.01

E F

C D

A B

FIGURE 2. [18F]CB251 as a TSPO-targeting ligand for imaging activated immune cells. (A) TSPO expression in 293FT cells was increased by transfection with pCMV-TSPO/eGFP. (B) [18F]CB251 uptake was increased in TSPO-overexpressing cells. (C) TSPO expression in MDA-MB-231 cells was reduced by transfection with pCMV-shTSPO/eGFP. (D) [18F]CB251 uptake was decreased in cells with reduced TSPO expression. (E) Western blots of TSPO and iNOS (activated immune cell marker) in a mouse macrophage cell line (RAW264.7) and mouse splenocytes. (F) [18F]CB251 uptake was increased in activated cells (both in Raw264.7 and mouse splenocytes). All experiments were replicated at least three times.

ctr TSPO/eGFP(+)

β-actin

TSPO

ctr TSPO/eGFP (+)

ctr shTSPO/eGFP(-)

ctr shTSPO/eGFP(-)

β-actin

TSPO

[18F]

CB

251

upta

ke

( 106

cpm

/mg)

2

1.5

1

0.5

0

ctr shTSPO/eGFP(-)

P <0.05

RAW264.7 Splenocytesβ-actin

TSPO

iNOS

ctractivatedLPS+INFγctr

activatedLPS+INFγ

[18F]

CB

251

upta

ke

( 106

cpm

/mg)

5

4

3

2

1

0 ctr activated ctr activatedRAW264.7 Splenocytes

P <0.005P <0.01

Figure 3. Bioluminescence imaging of peripheral immune cells in a mouse neuroinflammation model using intracranial LPS injection. (A) Flow cytometry analysis of immune cells from splenocytes of C57BL/6 (wild type) and C57BL/6.LucTg (reporter transgenic) mice. (B) Semiquantitative analysis of luciferase-expressing mouse splenocytes from C57BL/6.LucTg. The intensity of luciferase signal and the number of cells were highly correlated (R2=0.9991) (C) Schematic representation of experimental design using luciferase-expressing splenocytes to monitor peripheral immune cell infiltration in the brain of a mouse with neuroinflammation induced by intracranial LPS injection. (D) Representative biodistribution of luciferase-expressing splenocytes in mice 4 days after intracranial injection of saline or LPS (N=2 for each treatment).

A B

C D

C57BL/6.LucTg C57BL/6

Saline or

LPS

Luciferase-expressingsplenocytes

BLI

Splenocytes

CD8 CD4 Mac-1

CD

4

C

D4

CD8 CD4 Mac-1

B22

0

B22

0

Gr-

1

G

r-1

B cells

B cellsT cells

T cells

Macrophages

Macrophages

Granulocytes

Granulocytes

WT

LucTg

R2=0.9991

3

2.5

2

1.5

1

0.5

0

Tota

l Flu

x (x

106

phot

on/s

ec/c

m2 /s

r)

0 2 4 6 (106 cells)

0.3175 0.625 1.25 2.5 5 (106 cells/well)

Saline LPS

Brain Brain

Heart

Lung

Kidney

Liver

StomachSpleen

Intestine Intestine

SpleenStomach

Liver

Kidney

Lung

Heart

A

B

C

Figure 4. Representative MR scans using gadolinium-DOTA to monitor the Blood-Brain Barrier (BBB) disruption and immune cell activation. (A) Intracranial LPS injection model (saline as a control). (B) T1-weighted MR scans and (C) T2-weighted MR scans on day 1, 3, and 4 after intracranial injection (N=3). T1-weighted MRI was used to visualize the presence of gadolinium penetration through disrupted BBB, and T2-weighted MRI was used to monitor the vasogenic edema and localization of viable immune cells due to LPS induced-neuroinflammation.

Saline / LPSSaline / LPS

T2- weighted MRDay 1 Day 3 Day 4

T2- weighted MRDay 1 Day 3 Day 4

Saline

LPS

Saline

LPS

Day 1 Day 3 Day 4 T1-weighted MR T1- weighted MR

Day 1 Day 3 Day 4

Figure 5. Visualization of the intracranial LPS-induced neuro-inflammatory region using multi-modal imaging. (A) Experimental scheme of simultaneous PET/MR using [18F]CB251 and BLI. (B) Representative images of [18F]CB251 PET/MR and bioluminescence (BLI). The red circle indicates a region of interest (ROI). (C) Signal intensity ratio of the right striatum (injected with saline or LPS) to the left striatum (no injection) in the saline-injected (N=6) and LPS-injected (N=7) mice. PET signal was expressed as standard uptake value (SUVmax) and the BLI signal was expressed as total flux (photon/sec/cm2/sr). (D) IHC staining of intracranial LPS injected mouse brain tissues. Iba-1 (macrophages/microglia marker), CD68 (macrophages/monocytes marker), CD86 (antigen-presenting cells), CD4 (helper T cells), and CD8 (cytotoxic T cells) were used as markers for each cell type.

A

B

C

D

Day 0 Day 1 Day 4

Intracranial injection with LPS

i.v. injectionLuciferse-expressing splenocytes

[18F]CB251 PET/MRBLI

Saline

LPS

MRI (T2) [18F]CB251 PET/MR BLI

CD4 CD8

CD68

CD86

TSPO Iba-1

Left Right Rightno injection Saline LPS

PET (SUVmax) p=0.023

p=0.012

Left Right Rightno injection Saline LPS

p=0.011

n.s.

BLI

SUVm

axra

tio(R

ight

/Lef

t stri

atum

)

Tota

l Flu

x ra

tio(R

ight

/Lef

t stri

atum

)1.8

1.6

1.4

1.2

1.0

0

6

4

2

0

[18F]CB251 PET/MR

Saline LPS LPS + CDDO-Me

A

B

CBLI

Saline LPS LPS + CDDO-Me

Day 0 Day 1 Day 2 Day3 Day 4

Intracranial injection with LPS

i.v. injectionLuciferse-expressing splenocytes [18F]CB251 PET

T2-MRIBLI i.p. injection 200 nM CDDO-Me

0.7 2.0

PET

***p=0.002 **

p= 0.0323

n.s.

Saline LPS LPSCDDO-Me

SUVm

axra

tio(R

ight

/Lef

t stri

atum

)

1.8

1.6

1.4

1.2

1.0

0

BLI

*** p=0.0009

**p= 0.0025

*p= 0.042

Saline LPS LPSCDDO-Me

Tota

l Flu

x ra

tio(R

ight

/Lef

t stri

atum

)

8

6

4

2

0

Figure 6. Evaluation of the therapeutic effect of the anti-inflammatory drug (CDDO-Me) using [18F]CB251 PET and BLI. (A) Experimental design of treatments and imaging schedules, including simultaneous [18F]CB251 PET/MRs and BLIs. (B) Representative [18F]CB251 PET/MR scans and SUVmax ratios were plotted. (C) Representative BLI image and total flux ratio were plotted. Signal ratios from each modality (SUVmax ratio for PET; Total flux ratio for BLI) were calculated from the region of interest (ROI) of right (intracraniallyinjected with saline, LPS, or LPS plus anti-inflammatory agent CDDO-Me) and left striatum (non-injected) in the mouse brain. The values are mean ± SD. (N=8 for saline-treated; N=12 for LPS-treated; N=5 for LPS plus CDDO-Me).

Vectors for TSPO Mutagenesis

pTSPO-WT or –Mut(7.1kb)

TSPO genotype LigandBinding

PhenotypeDNA polymorphism rs6971

Protein position 147

Wild type C/C Ala/Ala HAB*

Mutant T/T Thr/Thr LAB*

TSPO ligand binding classification

HAB High Affinity Binder ; LAB Low Affinity Binder

B

A

FIGURE S1. TSPO genotypes and classfication. (A) TSPO ligand binding classification. Single nucleotide polymorphism rs6971 TSPO (CT point mutation; rs6971 A147T; Ala to Thr at amino acid 147 of TSPO) influences the binding affinity of TSPO ligands. TSPO with Thr/Thr at amino acid 147 shows a lower binding affinity to its ligand. TSPO ligand binding phenotypes are HAB or LAB, which indicate high- or low-affinity binding of ligands, respectively *Adapted from ref 32, 33. (B) Expression vector design for TSPO mutagenesis.