Contents lists available atScienceDirect
Neuropharmacology
journal homepage:www.elsevier.com/locate/neuropharm
Blood-brain barrier integrity in a mouse model of Alzheimer's disease with or without acute 3D6 immunotherapy
So fia Gustafsson
a, Tobias Gustavsson
b, Sahar Roshanbin
b, Greta Hultqvist
c, Margareta Hammarlund-Udenaes
a, Dag Sehlin
b, Stina Syvänen
b,∗aDepartment of Pharmaceutical Biosciences, Translational PKPD, Box 591, 751 24, Uppsala, Sweden
bDepartment of Public Health and Caring Sciences, Molecular Geriatrics, Dag Hammarskjölds väg 20, 751 85, Uppsala, Sweden
cDepartment of Pharmaceutical Biosciences, Protein Drug Design, Box 591, 751 24, Uppsala, Sweden
H I G H L I G H T S
•
Amyloid-β (Aβ) pathology did not affect brain partitioning of large molecules.•
3D6 bound to Aβ deposited in the vessels rather than to Aβ in the brain parenchyma.•
No effect on the blood-brain barrier integrity despite the 3D6-Aβ interaction.A R T I C L E I N F O
Keywords:
Blood-brain barrier Alzheimer's disease Large molecules
Cerebral amyloid angiopathy Antibody treatment Immunotherapy
A B S T R A C T
The blood-brain barrier (BBB) is suggested to be compromised in Alzheimer's disease (AD). The concomitant presence of vascular amyloid beta (Aβ) pathology, so called cerebral amyloid angiopathy (CAA), also predisposes impairment of vessel integrity. Additionally, immunotherapy against Aβ may lead to further damage of the BBB.
To what extent this affects the BBB passage of molecules is debated. The current study aimed to investigate BBB integrity to large molecules in transgenic mice displaying abundant Aβ pathology and age matched wild type animals, with or without acute anti-Aβ antibody treatment. Animals were administered a single i.v. injection of PBS or 3D6 (10 mg/kg), i.e. the murine version of the clinically investigated Aβ antibody bapineuzumab, sup- plemented with [125I]3D6. Three days post injections, a 4 kDa FITC and a 150 kDa Antonia Red dextran were administered i.v. to all animals. After termination,fluorescent detection in brain and serum was used for the calculation of dextran brain-to-blood concentration ratios. Further characterization of antibody fate and the presence of CAA were investigated using radioactivity measurements and Congo red staining. BBB passage of large molecules was equally low in wild type and transgenic mice, suggesting an intact BBB despite Aβ pa- thology. Neither was the BBB integrity affected by acute 3D6 treatment. However, CAA was confirmed in the transgenes and local antibody accumulations were observed in the brain, indicating CAA-antibody interactions.
The current study shows that independently of Aβ pathology or acute 3D6 treatment, the BBB is intact, without extensive permeability to large molecules, including the 3D6 antibody.
1. Introduction
In recent years, increasing focus has been directed towards the brain microvasculature and its relation to Alzheimer's disease (AD) pathology and progression. Of particular interest are the endothelial cells which constitute the vessel walls of brain capillaries. Through their tight
junctions and specialized properties, the endothelial cells form the blood-brain barrier (BBB). The BBB is a protective interface, both physically and dynamically, assigned to regulate uptake and efflux of endogenous and exogenous molecules from the circulation into and out of the brain. While not a sole actor, the BBB is part of the so called neurovascular unit (NVU) in which the glycocalyx, basement
https://doi.org/10.1016/j.neuropharm.2018.09.001
Received 20 June 2018; Received in revised form 27 August 2018; Accepted 1 September 2018
∗Corresponding author. Molecular Geriatrics, Department of Public Health and Caring Sciences, Uppsala University, Daghammarskölds väg 20, 751 85, Uppsala, Sweden.
E-mail addresses:sofia.gustafsson@farmbio.uu.se(S. Gustafsson),tobias.gustavsson@pubcare.uu.se(T. Gustavsson),
sahar.roshanbin@pubcare.uu.se(S. Roshanbin),greta.hultqvist@farmbio.uu.se(G. Hultqvist),mhu@farmbio.uu.se(M. Hammarlund-Udenaes), dag.sehlin@pubcare.uu.se(D. Sehlin),stina.syvanen@pubcare.uu.se(S. Syvänen).
Available online 07 September 2018
0028-3908/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T
membrane, astrocytes, pericytes, neurons and immune cells act to- gether to control the brain microenvironment (Neuwelt et al., 2011).
The BBB and the functioning of the NVU are proposed to be dis- rupted in AD, with for instance tight junction loss and dysregulated transport across the barrier (Montagne et al., 2017; Nelson et al., 2016).
BBB and NVU alterations have been reported in animal models of AD and in human post-mortem tissue. Imaging studies also imply subtle BBB impairment in patients with mild cognitive impairment and early stages of AD, even before the manifestation of brain atrophy and de- mentia (Montagne et al., 2015; van de Haar et al., 2016a; van de Haar et al., 2016b). Thus, NVU dysfunction might be part of AD etiology or a consequence of the disease. However, in what way and to what extent AD related pathology might impact the passage of drugs across the BBB, which under normal conditions is highly restricted, remains debated (Greenwood et al., 2017). Recent studies in animal models of AD in- dicate that the BBB is able to maintain enough capacity for normal regulation of small molecular drugs, both considering rate and extent of transport, independently of pathology (Bourasset et al., 2009; Cheng et al., 2010; Gustafsson et al., 2018). Still, other studies suggest a de- creased rate of transport of small molecular drugs into the brain of transgenic AD mice, increased BBB penetrability in transgenic mice or an increased uptake in AD patients (Mehta et al., 2013a; Mehta et al., 2013b; van Assema et al., 2012; Deo et al., 2014; Dickstein et al., 2006).
General BBB integrity, with respect to larger molecules such as proteins and antibodies, is not extensively studied under conditions related to AD pathophysiology. However, in a study by Bien-Ly et al. no BBB impairment affecting antibody passage was detected in diverse AD models (Bien-Ly et al., 2015).
The extent of BBB impairment is important to investigate in patients with AD but also in animal models used for translational studies of AD pathophysiology and drug development. In addition, it is important to delineate whether BBB impairment originates from the disease itself or a possible intervention introduced to target and treat the disease. The murine monoclonal antibody (mAb) 3D6 was developed for passive immunotherapy of AD, and later implemented in clinical trials in its humanized form, known as bapineuzumab (AAB-001; Pfizer Inc., New York, NY, and Janssen Pharmaceuticals Inc., Raritan, NJ). The antibody binds the N-terminal neoepitope of amyloid-β (Aβ), resulting from β- secretase cleavage of the Aβ protein precursor, and targets both fibrillar and soluble Aβ, which are known hallmarks of AD neuropathology (Bard et al., 2000). In animal models of AD, 3D6 treatment showed promising outcomes with reduced amyloid plaque burden in brain, possibly through Fc receptor-mediated microglial phagocytosis of Aβ deposits (Bard et al., 2000). While also showing potential beneficial effects in phase II clinical trials, bapineuzumab later failed in phase III due to non-significant changes on primary outcomes and incidence of unfavorable side effects (Rinne et al., 2010; Salloway et al., 2014;
Salloway et al., 2009). Adverse events like brain vasogenic edema and microhemorrhages were documented in Phase I, II and III studies, probably relating to a loss of integrity of cerebral vessels and hence BBB disruption (Rinne et al., 2010; Salloway et al., 2014; Salloway et al., 2009; Black et al., 2010; Sperling et al., 2012).
While immunoglobulins, peptides, and small molecular compounds are essential in studies targeting BBB dysfunction, the passage of such molecules across the BBB is often influenced by multiple factors. It can therefore be difficult to separate proposed BBB leakage, due to for
instance tight junction loss, from changes in transporter expression or function. Different molecular sized dextrans, tagged with fluorescent dyes, are therefore important tools when investigating BBB integrity.
Dextrans are hydrophilic polysaccharides, exhibiting low in vivo toxicity and immunogenicity, and are suggested to not interact with transpor- ters of the BBB, which is advantageous in BBB permeability studies (Hultstrom et al., 1983; Saunders et al., 2015).
By elucidating BBB passage of different molecular sized dextrans in wild type (WT) and AD transgenic mice, as well as animals subjected to 3D6 treatment, the aim of the current study was to investigate BBB integrity with respect to large molecules during AD pathology and after passive immunotherapy with a mAb known to interfere with Aβ in cerebral vessels.
2. Materials and methods 2.1. Dextrans
Afluorescein isothiocyanate (FITC) labelled dextran of 4 kDa and an Antonia Red (AR) labelled dextran of 150 kDa were obtained from TdB Consultancy AB (Uppsala, Sweden).
2.2. Animals
Heterozygous female and male transgenic mice, expressing the human Aβ protein precursor with the Arctic (E693G) and Swedish (KM670/671NL) mutations (tg-ArcSwe) (Lord et al., 2006; Mullan et al., 1992; Nilsberth et al., 2001), were used at an age of 18–19 months together with their WT littermates. The animals were bred in- house and maintained on a C57BL/6 background. Already at a young age, tg-ArcSwe mice display increased levels of soluble Aβ protofibrils while also developing Aβ plaque pathology from about 6 months of age (Lord et al., 2006; Lord et al., 2009; Philipson et al., 2009). The animals were housed in a humidity and temperature controlled environment on a 12 h light/dark cycle, with ad libitum access to food and water. The study was approved by the local Animal Ethics Committee in Uppsala, Sweden (reference numbers C17/14 and 13350/2017).
2.3. Generation of murine monoclonal antibody 3D6
The antibody 3D6 was expressed in a murine IgG2c framework in Expi293f mammalian cells according to a previously published protocol (Fang et al., 2017). Briefly, cells were transiently transfected with a mix of pcDNA3.4 vectors carrying the sequence of either the heavy or light chain of 3D6. Polyethylenimine (PEI) was used as transfection reagent and valproic acid (VPA) as a cell cycle inhibitor. The antibody was purified using an ÄKTA chromatography system with a protein G column (GE Healthcare AB, Uppsala, Sweden).
2.4. Radiochemistry
Direct radioiodination of 3D6 with iodine-125 (125I) was performed by using Chloramine-T, as described elsewhere (Greenwood et al., 1963; Sehlin et al., 2016). In brief, 150-300 pmoles of 3D6 antibody (depending on labelling occasion),125I stock solution (PerkinElmer Inc., Waltham, MA, USA), and 5μg Chloramine-T (Sigma Aldrich, Abbreviations
Aβ Amyloid beta AD Alzheimer's disease AR Antonia Red BBB Blood-brain barrier
CAA Cerebral amyloid angiopathy
Ctx Cortex
FITC Fluorescein isothiocyanate Hpc Hippocampus
mAb Monoclonal antibody NVU Neurovascular unit
WT Wild type
Stockholm, Sweden) were mixed in PBS to afinal volume of 110 μL for the125I labelling. The reaction proceeded for 90 s and was stopped by the addition of double molar excess of sodium metabisulfite (Sigma Aldrich), with further dilution to 500μL in PBS. Radiolabeled 3D6 was purified from free iodine and low-molecular weight components using a disposable NAP-5 size exclusion column, 5 kDa Mw cutoff, according to the manufacturer's instructions (GE Healthcare AB, Uppsala, Sweden), and eluted in 1 mL of PBS. Based on the added radioactivity and the radioactivity in the purified radioligand solution, the yield of the re- action was calculated. Labelling was always performed less than 2 h before the experiment. Affinity of [125I]3D6 to Aβ was tested with ELISA directly after radiolabeling according to a previously described method (Sehlin et al., 2016).
2.5. BBB integrity study
Tg-ArcSwe (n = 22) and WT animals (n = 14) were i.v. injected with either 3D6, supplemented with [125I]3D6, at a dose of 10 mg/kg (5 mL/kg, 6.53 ± 1.94 MBq/mL, n = 12 or 10, respectively per gen- otype) or PBS (5 mL/kg, n = 10 or 4, respectively per genotype) (Fig. 1). After 72 h the animals were anesthetized using isoflurane (Baxter Medical AB, Kista, Sweden) and a cannula for i.v. administra- tion of dextrans was inserted into the tail vein. The 150 kDa AR dextran (150 mg/kg) was administered to all animals, followed 5 min later by an injection of the 4 kDa FITC dextran (600 mg/kg). Ten minutes after thefirst dextran injection and 5 min after the second injection, animals were sacrificed by heart puncture and blood was collected and stored at 4 °C until further processed. The heart puncture was immediately fol- lowed by transcardial perfusion using 0.9% NaCl for 2 min, at a rate of 10.5 mL/min. The brain was extracted and dissected in half where the right hemisphere, including cerebellum but not the olfactory bulb, was instantly frozen on dry ice and kept frozen in−20 °C until use. The left hemisphere was further dissected by removing the olfactory bulb and separating cerebrum from cerebellum. The cerebrum (later referred to as brain) was placed in pre-weighed 2 mL Precellys CK14 tubes, con- taining 1.4 mm ceramic beads (Bertin Technologies, Montigny-le-Bre- tonneux, France). The brain tissue was weighed and PBS containing complete protease inhibitors (Roche, Basel, Switzerland) was added at a 1:3 w:v ratio. All samples were stored at 4 °C until further processing and analysis. Radioactivity was measured in all blood and brain sam- ples, using aγ-counter (2480 Wizard™, PerkinElmer, Waltham, USA).
2.6. Dextran analysis
Serum was prepared by centrifuging the blood samples at 2000 g for 10 min in 4 °C, and collecting the supernatant. Brain samples were homogenized using a Precellys Evolution homogenizer (Bertin Technologies, Montigny-le-Bretonneux, France). The brain samples were further centrifuged at 16 000 g for 1 h at 4 °C, and the supernatant
was collected. Pellets were stored at−20 °C until further use in pa- thological assessment. Brain supernatants were directly loaded onto a black, low-binding 96-well plate (Greiner, Kremsmünster, Austria).
Serum samples were diluted 10 times in PBS before being loaded onto plates. For the brain samples, a standard curve was prepared in blank brain supernatant, with the same 1:3 w:v tissue dilution as the samples.
Blank serum was used for the serum standard curve, with the same serum dilution as that of samples. Standards were prepared in the re- spective matrices to reduce any impact of tissue autofluorescence in the samples. AR and FITCfluorescence were analyzed using a Tecan Infinite 200 pro plate reader (Tecan, Männedorf, Switzerland), with an ex- citation wavelength set to 568 nm, and an emission wavelength set to 602 nm for the AR dextran, and 490 nm and 525 nm, correspondingly, for the FITC dextran. All analyses of dextran levels in the processed samples were achieved within 6 h after euthanization of the animals.
2.7. Aβ analysis
Aβ protofibrils were measured with a homogeneous ELISA, using Aβ N-terminal specific 82E1 (IBL International/Tecan Trading AG, Switzerland) as both capture and detection antibody. A 96-well half- area plate was coated with 12.5 ng per well of 82E1 and blocked with 1% BSA in PBS. PBS brain extracts were diluted 1:25 and incubated overnight at 4 °C, then detected with biotinylated 82E1 (0.25μg/mL), SA-HRP (Mabtech AB, Nacka Strand, Sweden) and K blue aqueous TMB substrate (Neogen Corp., Lexington, KY, USA). For Aβ1-40 and Aβ1-42 analyses, 96-well half-area plates were coated overnight with 50 ng per well of polyclonal rabbit anti-Aβ40 or anti-Aβ42 (Agrisera, Umeå, Sweden), and blocked with 1% BSA in PBS. Formic acid extracts were neutralized with 2M Tris and diluted 100 000 times for Aβ40 analysis or 10 000 times for Aβ42 analysis before being plated and incubated overnight at 4 °C. After incubation with biotinylated 82E1 (0.25μg/mL) signals were developed and read as above. All dilutions were made in ELISA incubation buffer (PBS, 0.1% BSA, 0.05% Tween, 0.15% proclin).
2.8. Autoradiography
Ex vivo autoradiography was conducted to visualize the distribution of [125I]3D6 in the brain. Sagittal sections, 20μm, were prepared from the right hemisphere. The sections were then placed in an X-ray cassette along with 125I standards of known radioactivity. Positron-sensitive phosphor screens (MS, MultiSensitive, PerkinElmer, Downers grove, USA) were placed onto the samples forfive days of exposure and then scanned at a resolution of 600 dots per inch in a Cyclone Plus Imager system (Perkin Elmer). The resulting digital images were converted to a false color scale (Royal) with ImageJ (Schneider et al., 2012) and normalized to the standards.
Fig. 1. Experimental design used to investigate BBB integrity in aged tg-ArcSwe and WT animals, with and without 3D6 treatment.
2.9. Congo red staining
Congo red staining was performed on sagittal sections adjacent to those used for autoradiography according to a previously published protocol (Wilcock et al., 2006). Briefly, 20 μm sections were fixated for 15 min in 4% PFA and rinsed twice in PBS. The sections were subse- quently incubated 20 min in an alkaline solution containing 80%
ethanol saturated with NaCl, followed by 0.2% Congo red in alkaline solution containing 80% ethanol saturated with NaCl for 30 min. Lastly, sections were dipped in an increasing ethanol gradient and mounted with DPX mounting medium. Brightfield images of cortex and hippo- campus were taken using a Nikon ECLIPSE 80i microscope with a Nikon DS-Ri1 camera (Nikon Instruments Inc., Melville NY, USA).
2.10. Nuclear track emulsion (micro autoradiography)
To further investigate the localization of radiolabeled antibody, nuclear track emulsion was performed on Congo red stained sections (20μm) prepared from brains isolated from [125I]3D6 administered mice according to a previously published protocol (Sehlin et al., 2016).
2.11. Statistical analysis
All data were analyzed using GraphPad Prism 6. The results are presented as mean ± S.D. if not otherwise stated. Student's t-tests were used to evaluate 3D6 concentrations in brain, and 3D6 brain-to-blood ratios in transgenes versus WT, as well as Aβ pathology in treated versus non-treated transgenes. A two-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test was used to analyze the dextran brain-to-blood concentration ratios in treated and
Fig. 2. Representative brain sections from transgenes expressing high levels of pathology (Tg 1-2) and moderate levels (Tg 3), as well as WT animals lacking pathology. (a) Autoradiography images of [125I]3D6 visualized on tg-ArcSwe (Tg 1-Tg 3) and WT (WT 1-WT 3) sagittal brain sections, displayed in a caudal to rostral (left to right) direction. Radioactivity originating from [125I]3D6 mainly appeared as high activity deposits in the tg-ArcSwe brain, especially pronounced in the frontal cortex, while the WT brain was largely devoid of [125I]3D6. (b) Congo red staining visualized Aβ pathology in the brain areas where high [125I]3D6 activity was found, and (c) the Congo stained sections further revealed CAA in endothelial vessels, primarily in cortex (Ctx) and hippocampus (Hpc) of tg-ArcSwe animals, while no such staining was found in WT brains, black scale bar 250μm. Adjacent sections from the same animals are shown in (a–c). (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.)
non-treated transgenes and WT. Correlations were investigated by the use of the Pearson correlation coefficient. Statistical significance was set to p < 0.05.
3. Results
3.1. Autoradiography of 3D6 brain distribution and visualization of Aβ pathology
Autoradiography showed rather limited brain distribution of [125I]
3D6 in both tg-ArcSwe and WT mice. However, high signals appeared as local deposits in the brain tissue in tg-ArcSwe mice, particularly in cortex (Fig. 2a). Congo staining of fibrillar Aβ pathology confirmed cortex and hippocampus as regions with abundant Aβ pathology in the tg-ArcSwe animals. In addition, Congo staining also revealed Aβ pa- thology in endothelial vessels (Fig. 2b and c), while no staining was detected in WT mice. Track emulsion was used to further investigate the localization of [125I]3D6 (Fig. 3). [125I]3D6 appeared to be highly confined to vessel-shaped Congo stained structures in the brain of tg- ArcSwe mice, as shown by the black staining (Fig. 3a and b). Brain sections from WT mice were, in line with the lack of Congo staining, almost completely devoid of [125I]3D6 staining in the track emulsion analysis (Fig. 3c).
3.2. 3D6 brain retention
Radioactivity, originating from [125I]3D6, was measured in tg- ArcSwe and WT brain tissue isolated from 3D6 treated mice.
Concentrations of 3D6 were significantly increased in tg-ArcSwe com- pared with WT mice (P = 0.015, Fig. 4a). The brain-to-blood con- centration ratio, which reflects the equilibration across the BBB, was also significantly increased in tg-ArcSwe mice compared to WT (P = 0.0072, Fig. 4b). Both observations are likely to reflect the pre- sence of Aβ, i.e. the target for 3D6, in the tg-ArcSwe brain.
3.3. Aβ pathology in 3D6 treated and non-treated tg-ArcSwe mice Genotype and pathological status of all animals were confirmed by
the manifestation of Aβ species characteristic of the tg-ArcSwe model.
While aged WT animals were devoid of Aβ pathology, aged tg-ArcSwe animals treated with either 3D6 or PBS showed similar Aβ levels, with equal variance, of Aβ protofibrils (P = 0.76), Aβ1-40 (P = 0.82), and Aβ1-42 (P = 0.94) (Fig. 5). Aβ1-40 was the most abundantly expressed Aβ species in the tg-ArcSwe animals, as shown byFig. 5b.
3.4. BBB passage of dextrans in tg-ArcSwe and WT mice with or without 3D6 treatment
BBB integrity in tg-ArcSwe and WT mice, subjected to 3D6 or PBS treatment, was investigated by analyzing the brain-to-blood parti- tioning of different molecular sized dextrans. Two animals were ex- cluded from the 4 kDa dextran analysis since their i.v. injection of this dextran was unsuccessful. One animal was excluded from the 150 kDa dextran analysis due to technical problems during sample analysis.
Fig. 3. Track emulsion (micro autoradiography) of [125I]3D6 localization in cortex. Black grains originate from the125I label attached to the 3D6 antibody. (a–b) High accumulation of [125I]3D6 was found in Congo red vessel-shaped structures in the tg-ArcSwe brain. (c) There were no high intensity accumulation regions, nor Congo positive structures, in the WT brain. Black scale bar 500μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
a. b.
ArcSwe
WT 0.0
0.1 0.2 0.3
[125 I]3D6 %IDpergrambrain
*
ArcSwe
WT 0.00
0.02 0.04 0.06 0.08 0.10
[125 I]3D6 brain-to-bloodratio
**
Fig. 4. Brain retention of 3D6 shown as (a) concentrations of [125I]3D6 in saline perfused brain tissue, expressed as percentage of injected dose per gram brain tissue, at 3 days post administration, or (b) brain-to-blood concentration ratio of [125I]3D6 at 3 days post administration. Circles represent individual animal data points and the bars represent mean ± S.D.*P < 0.05; **P < 0.01.
Regardless of group, the brain-to-blood ratio did not exceed mean values of 0.0015 for the 4 kDa dextran, and 0.00026 for the 150 kDa dextran (Fig. 6). Neither genotype, nor treatment had an effect on the brain-to-blood ratio for either the 4 kDa dextran (P = 0.92 and P = 0.076, respectively) or the 150 kDa dextran (P = 0.57 and P = 0.097, respectively) (Fig. 6). No correlation was observed between any of the Aβ species and the brain partitioning of the 4 kDa dextran (Fig. 7a). However, while a significant correlation was observed be- tween the 150 kDa brain-to-blood ratio and Aβ1-40 (P = 0.031) as well as Aβ1-42 (P = 0.010) (Fig. 7b), the relationship was primarily driven by two animals, and thus, there was no significant correlation (P = 0.66 and P = 0.32, respectively) if these subjects were excluded.
Interestingly, the data further showed a relationship between the brain-to-blood ratio of the injected antibody ([125I]3D6) and Aβ1-40 (P = 0.0021), the most abundantly deposited Aβ species in tg-ArcSwe animals and also a target of the antibody (Fig. 7c). A correlation was also observed between the [125I]3D6 brain-to-blood ratio and the brain- to-blood ratio of the 150 kDa dextran (P = 0.0069), while no significant correlation was detected for the 4 kDa dextran (P = 0.083) (Fig. 7d).
4. Discussion
In the present study we sought to investigate BBB integrity, by as- sessing the passage of different molecular sized dextrans across the BBB, in an animal model of Aβ pathology with or without passive im- munotherapy treatment. BBB integrity was shown to be preserved, with very low BBB passage of a 150 kDa and a 4 kDa dextran, in aged, non- treated tg-ArcSwe mice as well as in tg-ArcSwe mice treated with the murine mAb 3D6. The current results are in line with a previous study investigating the impact of Aβ pathology on the extent of small
molecular drug transport across the BBB in the same AD mouse model (Gustafsson et al., 2018). The former study showed that the BBB, in- dependently of Aβ pathology, was capable of sustaining sufficient function in order to maintain normal drug transport when targeting passive diffusion, as well as influx and efflux transport across the BBB (Gustafsson et al., 2018). The current animal model has also been used in multiple studies investigating BBB transport mechanisms for en- hanced delivery of antibodies and antibody fragments to the brain (Sehlin et al., 2016; Hultqvist et al., 2017; Syvanen et al., 2017). In support of the presentfindings, these previous studies suggest no in- dication of an enhanced delivery across the BBB attributed to a leakage of the molecular complexes across an impaired barrier (Hultqvist et al., 2017; Magnusson et al., 2013). In addition, no increased BBB passage of antibodies, caused by a disrupted barrier, was observed in the study by Bien-Ly et al., using multiple mouse models of AD pathology (Bien-Ly et al., 2015).
Adverse events, involving brain edema and microhemorrhages, as- sociated with 3D6 and bapineuzumab treatment were suggested in animal models of AD to result from antibody interactions with amyloid deposits in the vessel walls, so called cerebral amyloid angiopathy (CAA) (Pfeifer et al., 2002; Wilcock et al., 2004a; Racke et al., 2005).
CAA is a common concomitant feature of AD pathology. Whereas the presence of CAA in itself weakens the vessel wall, it has also been speculated that the anti-Aβ antibody interactions with CAA leads to further activation of an immune response related to vascular Aβ clearance, which in turn could exacerbate the instability of the vessel integrity (Wilcock et al., 2004b). In addition, it has been shown that Aβ accumulation in mouse models of AD leads to a disruption of tight junctions, and hence, a further increased risk of BBB breakdown (Biron et al., 2011). The development of CAA is suggested to be a time
a. b. c.
3D6 PBS
0 200 400 600 800
A1-40 (ng/mgbraintissue)
3D6 PBS
0 20 40 60
Aprotofibrils (pg/mgbraintissue)
3D6 PBS
0 20 40 60
A1-42 (ng/mgbraintissue)
Fig. 5. Brain concentration measurements of (a) Aβ protofibrils, (b) Aβ1-40, and (c) Aβ1-42, in 3D6 and PBS treated tg-ArcSwe mice. Circles represent individual animal data points and the bars represent mean ± S.D.
tg-ArcSwe
WT 0.000
0.001 0.002 0.003
4kDaFITCdextran brain-to-bloodratio
tg-ArcSwe
WT 0.0000
0.0005 0.0010 0.0015
150kDaARdextran brain-to-bloodratio 3D6
PBS
a. b.
Fig. 6. Brain-to-blood concentration ratios of (a) the 4 kDa dextran in tg-ArcSwe and WT mice, treated with 3D6 or PBS, and (b) the 150 kDa dextran in tg-ArcSwe and WT mice, treated with 3D6 or PBS. After analysis by two-way ANOVAs, followed by Tukey's multiple comparisons test, no significant differences were detected in the ratios for any of the dextrans, independent of genotype and treatment. Circles represent individual animal data points and the bars represent mean ± S.D.
dependent process and many preclinical immunotherapy studies tar- geting Aβ involve mice at progressed stages of Aβ pathology where CAA is present to interact with the therapeutic antibody and induce micro- hemorrhages (Racke et al., 2005). The development of CAA pathology is also likely genotype dependent and while the manifestation of CAA was reported in previous characterizations of the tg-ArcSwe model (Philipson et al., 2009; Lillehaug et al., 2014), evaluation of extent or time dependency of CAA development in the tg-ArcSwe model has so far not been conducted. However, the present study showed widespread presence of CAA deposits in 18–19 months old tg-ArcSwe animals, which enables antibody-Aβ interactions in the vessels of the tg-ArcSwe mice. To further support these observations, 3D6 was supplemented with [125I]3D6 before injection in the present study to better determine the fate of the antibody. Interestingly, autoradiography showed local accumulations of 3D6 in the brains of tg-ArcSwe animals but not in WT mice, suggestive of antibody-Aβ interactions. Track emulsion supported these observations by revealing accumulation of [125I]3D6 in what appeared to be Congo stained vessels in tg-ArcSwe animals. The brain- to-blood ratio of [125I]3D6 was also shown to correlate with brain concentrations of Aβ40, which is further indicative of antibody binding to deposits in the vasculature. Aβ40 is the dominating form of Aβ in CAA in humans (Alonzo et al., 1998; Suzuki et al, 1994), and con- sidering the dominant manifestation of Aβ40 in the tg-ArcSwe animals it could be postulated that Aβ40 also dominates the CAA deposits found in these animals.
Despite antibody interaction with CAA in the present study, an in- creased BBB passage of dextrans was not detected and the brain-to- blood ratios of the two dextrans showed generally very low brain par- titioning. The brain-to-blood ratio of the 4 kDa dextran was about 10 times higher than that observed for the 150 kDa dextran, for which the
BBB appeared almost impermeable. The 4 kDa dextran ratio was also more in line with ratios generally observed for antibodies, where a brain concentration below 0.1% of the injected dose is a common value if no enhanced delivery system is applied (Bard et al, 2000; Poduslo et al., 1994). Interestingly, two tg-ArcSwe animals with among the highest Aβ brain levels also displayed the highest dextran ratios. While these two animals were driving the significant correlation found be- tween the 150 kDa dextran brain-to-blood ratio and Aβ brain levels, and to some extent also the relationship between the [125I]3D6 versus the 150 kDa dextran ratio, these animals were kept in the analysis as no technical concerns were posted during the experimental procedure.
Autoradiography showed that the absolute majority of the brain radioactivity, originating from [125I]3D6, was found in local deposits (also in the two animals with high Aβ brain levels and high dextran ratios, i.e. Tg 1 and Tg 2 inFig. 2). This indicated that 3D6, in line with the dextrans, penetrated into the brain parenchyma in a negligible degree, and that 3D6 rather interacted with Aβ deposits in the vessel walls.
The single dose of 3D6 used in the current study was not expected to decrease the Aβ load. This was confirmed by the observation that both 3D6 and PBS treated mice displayed the same Aβ levels three days post treatment. The applied treatment design is considerably shorter com- pared to previous studies of peripherally administered anti-Aβ anti- bodies in AD mouse models where immunotherapy, aimed mainly to investigate Aβ removal, has been ongoing for up to several months. The rationale for using a single dose was to enable investigation of 3D6 and Aβ interaction and its impact on the BBB isolated from long-term treatment. We have previously shown that maximum concentrations of an IgG anti-Aβ antibody occurs around three days after administration (Magnusson et al, 2013). Further, microhemorrhages in animal models Fig. 7. The relationships between brain partitioning of dextrans and 3D6 were investigated, as well as their respective relationship to Aβ species, using linear regression and by the calculation of respective Pearson correlation coefficients (r). The 4 kDa and 150 kDa dextran brain-to-blood ratios, (a) and (b) respectively, were put in relation to Aβ1-40 and Aβ1-42 brain concentrations. Correlations were also analyzed between the [125I]3D6 brain-to-blood ratio and (c) Aβ1-40 and Aβ1- 42, as well as (d) the brain-to-blood ratios of the 4 kDa and 150 kDa dextrans. Correlations involving Aβ1-40 and Aβ1-42 (a–c), only include tg-ArcSwe animals, independent of treatment, as WT mice lack pathology. The [125I]3D6 versus dextran brain-to-blood ratio (d), included corresponding values from both tg-ArcSwe and WT mice treated with 3D6. P < 0.05 was considered statistically significant.
and humans have shown to be a fairly rapid and transient process, occurring early in antibody treatment (Sperling et al, 2012; Blockx et al, 2016; Arrighi et al, 2016; Zago et al, 2013). In line with suchfindings, the current treatment design could give rise to similar outcomes given the presence of CAA in the tg-ArcSwe model. However, an immune reaction may need further time in order to provoke permeability changes in our model or it might be that the leakage, due to antibody induced bleedings, is limited to molecules smaller than 4 kDa. It is also possible that even further advanced pathology in the current AD model would increase passage of molecules across the BBB after 3D6 treat- ment, as indicated by the occurrence of high Aβ levels, high dextran ratios, and high [125I]3D6 ratios in two of the tg-ArcSwe animals in the present study.
Further studies are needed to evaluate the impact of AD pathology on BBB integrity and brain drug delivery. It is also essential to in- vestigate to what extent certain treatments could affect BBB perme- ability to molecules of varying sizes, given the fact that the AD patient population is often subjected to multidrug treatments. The future use and development of translational imaging techniques such as magnetic resonance imaging (MRI) and positron emission tomography (PET) could aid in co-localization of subtle BBB leakage and brain drug pharmacokinetics. Microhemorrhages and vasogenic edema have been reported for several of the Aβ antibodies studied in clinical trials and the risk of these adverse events has been shown to correlate with the administered antibody dose (Salloway et al, 2014; Salloway et al, 2009;
Sperling et al, 2012; Sevigny et al, 2016). Since the effect on Aβ re- moval is also related to the administered dose, the BBB occurring side effects need to be avoided or mitigated while still reaching sufficient effect outcomes (Sevignyet al, 2016). As shown in the present study, the amount of 3D6 in the brain was very low, and even more importantly, most of the antibody was found as local deposits, likely in the vessels. A recent study using RmAb158, another Aβ antibody, showed a similar binding pattern in autoradiography as observed for 3D6 in the present study, i.e. local high intensity deposits (Syvanen et al, 2018). Interest- ingly, a bispecific version of RmAb158, targeting the transferrin re- ceptor to enable receptor mediated transcytosis into the brain showed a completely different spatial distribution in the brain covering all areas affected by pathology and without high intensity “hot spots”. Bispecific antibodies engineered for enhanced delivery s across the BBB are being developed mainly to increase antibody brain concentrations and thus therapeutic effect (Hultqvist et al., 2017; Niewoehner et al, 2014).
However, it can also be speculated that these modified antibodies to a large extent may promote a circumvention of interactions with CAA in the vessels due to higher distribution into the brain parenchyma.
In conclusion, this study showed that the BBB passage of large molecules was intact and independent of the presence of Aβ pathology in 18–19 month aged mice. Additionally, although [125I]3D6 interacted with Aβ, most likely in the vasculature, this did not per se cause ex- tensive loss of BBB integrity with respect to molecules in the size of 4 and 150 kDa. It is possible that a longer antibody treatment involving both Aβ removal and immune responses would lead to a more impaired BBB. Furthermore, this study showed very low 3D6 concentrations in the brain, indicating that the BBB was not only limiting the passage of dextrans across the BBB, but also the passage of the 3D6 antibody.
Conflicts of interest
All authors declare no conflict of interest.
Funding
This work was supported by grants from the Swedish Research Council (2017-02413), Alzheimerfonden, Åhlén-stiftelsen, Stiftelsen Fondkistan, Magnus Bergwalls stiftelse, Stohnes Stiftelse, Stiftelsen för gamla tjänarinnor and Goljes stiftelse.
Acknowledgement
We are grateful to Lars N.G. Nilsson, who developed and char- acterized the transgenic mouse model and to the staff at the SciLifeLab Pilot Facility for Preclinical PET-MRI, a Swedish nationally available imaging platform at Uppsala University, Sweden, where the radi- olabeling of 3D6 was performed.
References
Alonzo, N.C., Hyman, B.T., Rebeck, G.W., Greenberg, S.M., 1998. Progression of cerebral amyloid angiopathy: accumulation of amyloid-beta40 in affected vessels. J.
Neuropathol. Exp. Neurol. 57, 353–359.
Arrighi, H.M., et al., 2016. Amyloid-related imaging abnormalities-haemosiderin (ARIA- H) in patients with Alzheimer's disease treated with bapineuzumab: a historical, prospective secondary analysis. J. Neurol. Neurosurg. Psychiatry 87, 106–112.
Bard, F., et al., 2000. Peripherally administered antibodies against amyloid beta-peptide enter the central nervous system and reduce pathology in a mouse model of Alzheimer disease. Nat. Med. 6, 916–919.
Bien-Ly, N., et al., 2015. Lack of widespread BBB disruption in alzheimer's disease models:
focus on therapeutic antibodies. Neuron 88, 289–297.
Biron, K.E., Dickstein, D.L., Gopaul, R., Jefferies, W.A., 2011. Amyloid triggers extensive cerebral angiogenesis causing blood brain barrier permeability and hypervascularity in Alzheimer's disease. PLoS One 6, e23789.
Black, R.S., et al., 2010. A single ascending dose study of bapineuzumab in patients with Alzheimer disease. Alzheimer Dis. Assoc. Disord. 24, 198–203.
Blockx, I., et al., 2016. Monitoring blood-brain barrier integrity following amyloid-beta immunotherapy using gadolinium-enhanced MRI in a PDAPP mouse model. J.
Alzheimers Dis. 54, 723–735.
Bourasset, F., et al., 2009. Reduction of the cerebrovascular volume in a transgenic mouse model of Alzheimer's disease. Neuropharmacology 56, 808–813.
Cheng, Z., et al., 2010. Central nervous system penetration for small molecule therapeutic agents does not increase in multiple sclerosis- and Alzheimer's disease-related animal models despite reported blood-brain barrier disruption. Drug Metab. Dispos. 38, 1355–1361.
Deo, A.K., et al., 2014. Activity of P-Glycoprotein, a beta-amyloid transporter at the blood-brain barrier, is compromised in patients with mild alzheimer disease. J. Nucl.
Med. 55, 1106–1111.
Dickstein, D.L., et al., 2006. Abeta peptide immunization restores blood-brain barrier integrity in Alzheimer disease. Faseb. J. 20, 426–433.
Fang, X.T., Sehlin, D., Lannfelt, L., Syvanen, S., Hultqvist, G., 2017. Efficient and in- expensive transient expression of multispecific multivalent antibodies in Expi293 cells. Biol. Proced. Online 19, 11.
Greenwood, F.C., Hunter, W.M., Glover, J.S., 1963. The preparation of I-131-Labelled human growth hormone of high specific radioactivity. Biochem. J. 89, 114–123.
Greenwood, J., et al., 2017. Current research into brain barriers and the delivery of therapeutics for neurological diseases: a report on CNS barrier congress London, UK, 2017. Fluids Barriers CNS 14, 31.
Gustafsson, S., Lindstrom, V., Ingelsson, M., Hammarlund-Udenaes, M., Syvanen, S., 2018. Intact blood-brain barrier transport of small molecular drugs in animal models of amyloid beta and alpha-synuclein pathology. Neuropharmacology 128, 482–491.
Hultqvist, G., Syvanen, S., Fang, X.T., Lannfelt, L., Sehlin, D., 2017. Bivalent brain shuttle increases antibody uptake by monovalent binding to the transferrin receptor.
Theranostics 7, 308–318.
Hultstrom, D., Malmgren, L., Gilstring, D., Olsson, Y., 1983. FITC-Dextrans as tracers for macromolecular movements in the nervous system. A freeze-drying method for dextrans of various molecular sizes injected into normal animals. Acta Neuropathol.
59, 53–62.
Lillehaug, S., et al., 2014. Brainwide distribution and variance of amyloid-beta deposits in tg-ArcSwe mice. Neurobiol. Aging 35, 556–564.
Lord, A., et al., 2006. The Arctic Alzheimer mutation facilitates early intraneuronal Abeta aggregation and senile plaque formation in transgenic mice. Neurobiol. Aging 27, 67–77.
Lord, A., et al., 2009. Amyloid-beta protofibril levels correlate with spatial learning in Arctic Alzheimer's disease transgenic mice. FEBS J. 276, 995–1006.
Magnusson, K., et al., 2013. Specific uptake of an amyloid-beta protofibril-binding anti- body-tracer in AbetaPP transgenic mouse brain. J. Alzheimers Dis. 37, 29–40.
Mehta, D.C., Short, J.L., Nicolazzo, J.A., 2013a. Reduced CNS exposure of memantine in a triple transgenic mouse model of Alzheimer's disease assessed using a novel LC-MS technique. J. Pharmaceut. Biomed. Anal. 85, 198–206.
Mehta, D.C., Short, J.L., Nicolazzo, J.A., 2013b. Altered brain uptake of therapeutics in a triple transgenic mouse model of Alzheimer's disease. Pharm. Res. (N. Y.) 30, 2868–2879.
Montagne, A., Zhao, Z., Zlokovic, B.V., 2017. Alzheimer's disease: a matter of blood-brain barrier dysfunction? J. Exp. Med. 214, 3151–3169.
Montagne, A., et al., 2015. Blood-brain barrier breakdown in the aging human hippo- campus. Neuron 85, 296–302.
Mullan, M., et al., 1992. A pathogenic mutation for probable Alzheimer's disease in the APP gene at the N-terminus of beta-amyloid. Nat. Genet. 1, 345–347.
Nelson, A.R., Sweeney, M.D., Sagare, A.P., Zlokovic, B.V., 2016. Neurovascular dys- function and neurodegeneration in dementia and Alzheimer's disease. Biochim.
Biophys. Acta 1862, 887–900.
Neuwelt, E.A., et al., 2011. Engaging neuroscience to advance translational research in
brain barrier biology. Nat. Rev. Neurosci. 12, 169–182.
Niewoehner, J., et al., 2014. Increased brain penetration and potency of a therapeutic antibody using a monovalent molecular shuttle. Neuron 81, 49–60.
Nilsberth, C., et al., 2001. The 'Arctic' APP mutation (E693G) causes Alzheimer's disease by enhanced Abeta protofibril formation. Nat. Neurosci. 4, 887–893.
Pfeifer, M., et al., 2002. Cerebral hemorrhage after passive anti-Abeta immunotherapy.
Science 298, 1379.
Philipson, O., et al., 2009. A highly insoluble state of Abeta similar to that of Alzheimer's disease brain is found in Arctic APP transgenic mice. Neurobiol. Aging 30, 1393–1405.
Poduslo, J.F., Curran, G.L., Berg, C.T., 1994. Macromolecular permeability across the blood-nerve and blood-brain barriers. Proc. Natl. Acad. Sci. U. S. A. 91, 5705–5709.
Racke, M.M., et al., 2005. Exacerbation of cerebral amyloid angiopathy-associated mi- crohemorrhage in amyloid precursor protein transgenic mice by immunotherapy is dependent on antibody recognition of deposited forms of amyloid beta. J. Neurosci.
25, 629–636.
Rinne, J.O., et al., 2010. 11C-PiB PET assessment of change infibrillar amyloid-beta load in patients with Alzheimer's disease treated with bapineuzumab: a phase 2, double- blind, placebo-controlled, ascending-dose study. Lancet Neurol. 9, 363–372.
Salloway, S., et al., 2009. A phase 2 multiple ascending dose trial of bapineuzumab in mild to moderate Alzheimer disease. Neurology 73, 2061–2070.
Salloway, S., et al., 2014. Two phase 3 trials of bapineuzumab in mild-to-moderate Alzheimer's disease. N. Engl. J. Med. 370, 322–333.
Saunders, N.R., Dziegielewska, K.M., Mollgard, K., Habgood, M.D., 2015. Markers for blood-brain barrier integrity: how appropriate is Evans blue in the twenty-first cen- tury and what are the alternatives? Front. Neurosci. 9, 385.
Schneider, C.A., Rasband, W.S., Eliceiri, K.W., 2012. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675.
Sehlin, D., et al., 2016. Antibody-based PET imaging of amyloid beta in mouse models of Alzheimer's disease. Nat. Commun. 7, 10759.
Sevigny, J., et al., 2016. The antibody aducanumab reduces Abeta plaques in Alzheimer's
disease. Nature 537, 50–56.
Sperling, R., et al., 2012. Amyloid-related imaging abnormalities in patients with Alzheimer's disease treated with bapineuzumab: a retrospective analysis. Lancet Neurol. 11, 241–249.
Suzuki, N., et al., 1994. High tissue content of soluble beta 1-40 is linked to cerebral amyloid angiopathy. Am. J. Pathol. 145, 452–460.
Syvanen, S., et al., 2017. A bispecific Tribody PET radioligand for visualization of amy- loid-beta protofibrils - a new concept for neuroimaging. Neuroimage 148, 55–63.
Syvanen, S., et al., 2018. Efficient clearance of Abeta protofibrils in AbetaPP-transgenic mice treated with a brain-penetrating bifunctional antibody. Alzheimer's Res. Ther.
10, 49.
van Assema, D.M., et al., 2012. Blood-brain barrier P-glycoprotein function in Alzheimer's disease. Brain 135, 181–189.
van de Haar, H.J., et al., 2016a. Blood-brain barrier leakage in patients with early alz- heimer disease. Radiology 281, 527–535.
van de Haar, H.J., et al., 2016b. Neurovascular unit impairment in early Alzheimer's disease measured with magnetic resonance imaging. Neurobiol. Aging 45, 190–196.
Wilcock, D.M., Gordon, M.N., Morgan, D., 2006. Quantification of cerebral amyloid an- giopathy and parenchymal amyloid plaques with Congo red histochemical stain. Nat.
Protoc. 1, 1591–1595.
Wilcock, D.M., et al., 2004a. Passive immunotherapy against Abeta in aged APP-trans- genic mice reverses cognitive deficits and depletes parenchymal amyloid deposits in spite of increased vascular amyloid and microhemorrhage. J. Neuroinflammation 1, 24.
Wilcock, D.M., et al., 2004b. Passive amyloid immunotherapy clears amyloid and tran- siently activates microglia in a transgenic mouse model of amyloid deposition. J.
Neurosci. 24, 6144–6151.
Zago, W., et al., 2013. Vascular alterations in PDAPP mice after anti-Abeta im- munotherapy: implications for amyloid-related imaging abnormalities. Alzheimers Dement 9, S105–S115.
