Lipopolysaccharide Induced Opening of the Blood Brain Barrier on Aging 5XFAD Mouse Model

Abstract

The development of neurotherapeutics for many neurodegenerative diseases has largely been hindered by limited pharmacologic penetration across the blood-brain barrier (BBB). Previous attempts to target and clear amyloid-β (Aβ) plaques, a key mediator of neurodegenerative changes in Alzheimer’s disease (AD), have had limited clinical success due to low bioavailability in the brain because of the BBB. Here we test the effects of inducing an inflammatory response to disrupt the BBB in the 5XFAD transgenic mouse model of AD. Lipopolysaccharide (LPS), a bacterial endotoxin recognized by the innate immune system, was injected at varying doses. 24 hours later, mice were injected with either thioflavin S, a fluorescent Aβ-binding small molecule or 30 nm superparamagnetic iron oxide (SPIO) nanoparticles, both of which are unable to penetrate the BBB under normal physiologic conditions. Our results showed that when pretreated with 3.0 mg/kg LPS, thioflavin S can be found in the brain bound to Aβ plaques in aged 5XFAD transgenic mice. Following the same LPS pretreatment, SPIO nanoparticles could also be found in the brain. However, when done on wild type or young 5XFAD mice, limited SPIO was detected. Our results suggest that the BBB in aged 5XFAD mouse model is susceptible to increased permeability mediated by LPS, allowing for improved delivery of the small molecule thioflavin S to target Aβ plaques and SPIO nanoparticles, which are significantly larger than antibodies used in clinical trials for immunotherapy of AD. Although this approach demonstrated efficacy for improved delivery to the brain, LPS treatment resulted in significant weight loss even at low doses, resulting from the induced inflammatory response. These findings suggest inducing inflammation can improve delivery of small and large materials to the brain for improved therapeutic or diagnostic efficacy. However, this approach must be balanced with the risks of systemic inflammation.

INTRODUCTION

Alzheimer’s disease (AD) is among the very few illnesses that has been known for over a century but continues to impact millions of people world-wide in an unabated manner, particularly on elderly people [1]. The epidemic threat of AD is particularly worrisome in an aging population because no disease-modifying therapies are currently available. The mechanism that regulates neuronal degeneration in AD remains unknown; however, the cytopathological hallmarks of the disease appear to be the formation of extracellular amyloid-β (Aβ) plaques between the neurons, which leads ultimately to profound neuron toxicity and atrophy [2]. Since the formation of Aβ plaques is one of the underlying mechanisms implicated in AD, detection and eventual disruption of their formation, particularly at the onset of the disease, would be an ultimate goal in treating AD. Toward that end, experimental therapeutics, including small molecules [3, 4] and antibodies [5] that target Aβ were developed. Although this approach demonstrated significant promise and widespread appeal, several limitations exist in the development of drugs intended for the brain. For instance, a variety of criteria must be considered in all related initiatives, such as the drug’s ability to cross the blood-brain barrier (BBB) and bind to Aβ. As basic neurological research continues to reveal novel targets for AD therapy, the necessity to deliver therapeutic agents across the BBB becomes of paramount importance. The improved bioavailability of drugs or imaging probes in the brain for the detection/treatment of AD produces a prolonged therapeutic effect and, as a consequence, reduces toxicity, cost, and importantly is expected to eradicate the disease [6–9]. Unfortunately, drug delivery across the BBB is still a significant challenge and lags behind other areas in molecular neuroscience, because of the difficulties posed by this barrier [10]. The brain contains blood capillaries that are different from the blood capillaries in other tissues. Peripheral capillaries have open interendothelial junction spaces and active pinocytosis, which forms a paracellular route and a transcellular route, respectively, for the free diffusion of molecules from the blood to the organ interstitium. However, brain capillaries are lined with a unique layer of endothelial cells that lack fenestrations and are sealed with intercellular tight junctions that effectively prevent diffusion through paracellular pathways. Furthermore, these cells have minimal pinocytosis that eliminates the nonspecific transcellular route of molecular transport from blood to brain [11]. This is exacerbated by the fact that brain capillaries are also reinforced to eliminate nonspecific molecular transport into the brain with a matrix of astrocyte foot processes and P-glycoprotein (Pgp)-active drug efflux transporter proteins in the luminal membrane of the cerebral capillary endothelium [12, 13]. Due to these unique properties of vertebrate brain capillaries, the distribution of molecules from blood to the brain interstitial space is facilitated mainly via two mechanisms, namely, lipid-mediated free diffusion of small molecules and catalyzed transport of small or large molecules [10]. In fact, it is known that water, CO₂, glucose and oxygen penetrate the brain with ease owing to their small molecular size, as do small lipid soluble entities. Certain large molecules such as peptides can be transported across the BBB via receptor-mediated transport (RMT) systems [14]. To facilitate the delivery of drugs, which do not belong to the aforementioned categories or in special circumstances, a number of active techniques have been developed, such as intracranial injection [15], electroporation or incorporation of a drug into cationic liposomes [16, 17], dendrimers [18] or bacterial toxins [19, 20], focused ultrasound [21, 22], pulsed electric field [23], and intracerebroventricular injection methods using convection-enhanced delivery [24] and intraventricular catheters to bypass the BBB [25]. It has been known that the BBB also can be compromised during disease conditions, such as AD [26–29], meningitis, encephalitis, sepsis, and local and systemic infections [30]. In fact, inflammation has also long been known for disrupting the BBB [31–33]. For example, lipopolysaccharide (LPS), a powerful cytokines releaser can affect the permeability of the BBB in animal models [34]. LPS is a molecular motif structurally similar amongst gram-negative bacteria that is recognized by the innate immune system and results in pro-inflammatory cytokine release mediated by toll-like receptor 4 (TLR–4) [35]. Vascular changes occur as a result of this inflammatory state, including increased expression of leukocyte adhesion molecules, cytokine release and increased permeability, which are all seen in brain endothelium as well [36–38]. Previous studies have shown that LPS can alter BBB permeability [39, 40], increase absorptive endocytosis and immune cell adhesion/trafficking [37, 41, 42], and direct application of LPS to brain endothelial cells increases intracellular leakage [43]. However, innate immune response to LPS mediated by leukocyte recruitment and activation [44–50] can become maladaptive in severe localized and systemic infections, leading to septic shock due to rapid cytokine release [51].

Given the complexity of the biological and immunological events that emerge after treating animals with LPS, questions arise over the relative benefits and pernicious effects of LPS. In this work, we want to examine the toxicity threshold, kinetics of BBB opening and the benefit of LPS-induced opening of the BBB on a transgenic mouse model of AD. The data suggest that the BBB in 5XFAD mice remains intact even when the animals are old; however, it is prone to be compromised by LPS compared to the wild type counterpart. We conclude that LPS has gravitas as a BBB opener, albeit careful dosing is necessary to avoid potential toxicity.

MATERIALS AND METHODS

Materials

Thioflavin S and LPS were obtained from Sigma Aldrich (St. Louis, MO) and iron oxide nanoparticles (SPIO) were developed in-house as described in the past [52]. All reagents and solvents were of analytical grade and used as received from the commercial source without further purifications.

Animals

5XFAD and control C57BL/6J mice were maintained at Vanderbilt University under standard conditions, in a 12-h light/dark cycle and with free access to food and water. The 5XFAD mice over express both mutant human amyloid precursor protein (APP) and presenilin 1 (PS1), correlating with high burden and accelerated accumulation of the Aβ. A colony of 5XFAD transgenic mice obtained from Jackson Laboratories was maintained by crossing 5XFAD mice with a wild-type (wt) C57BL/6J strain. The mice were genotyped by a standard polymerase chain reaction using DNA isolated from tail tips with the following primers: PSEN1 forward, 5′–TCATGACTATCCTCCTGGTGG–3′ and reverse, 5′–CGTTA TAG GTTTTAAACACTTCCCC–3′. For APP, forward, 5′–AGGACTGACCACTCGACCAG–3′ and reverse, 5′–CGGGGGTCTAGTTCTGCAT–3′. We also genotyped mice for the presence of retinal degeneration Pde6brd1 mutation using forward, 5′–AAGCTAGCTGCAGTAACGCCATTT–3′ and reverse, 5′–ACCTGCATGTGAACCCAGTATTCTATC–3′. After polymerase chain reaction amplification, the DNA product of each reaction was analyzed by size fractionation through a 1% agarose gel; with Pde6b mutant = 560 bp, APP transgene = 377 bp, and PSEN1 transgene = 608 bp. The 5XFAD mice were maintained as heterozygous. Animal experiments were conducted per the guidelines established by Vanderbilt University’s Institutional Animal Care and Use Committee. At the end of the study, animals were euthanized by cervical dislocation after sedation with isoflurane. Clinical signs were used to verify euthanasia, including heartbeats and reflection to toe-pinching. Further, if animals showed signs of illness (weight loss, food withdrawal, or infection) they were sacrificed before the endpoints. All experimental procedures in this study were approved by the Vanderbilt University IACUC panel.

LPS dose thresholds

In a series of pilot studies, the maximal dose with which LPS, derived from Escherichia coli O111:B4 (Sigma Aldrich, St Louis, MO), could be administered was determined to be 3 mg/kg. Additional pilot studies supported previously published results [36] that LPS-mediated BBB permeation can be observed 24 h after administration. To illustrate the dosage-dependent effect of LPS on BBB opening, LPS was administered via tail vein injection at varying doses (0.01 mg/kg, 0.1 mg/kg, 1 mg/kg, 3 mg/kg) on sedated mice, which received approximately 2.5% isoflurane mixed with 2 liter/minute of oxygen. Approximately 24 h later, LPS-treated mice (3–5–month-old or 13–15-month-old) were treated with either i.v. thioflavin S (5XFAD n = 3) or SPIO nanoparticles [5XFAD n = 6 (3 young, 3 old mice), wt n = 10 (3 young, 7 old mice) age-matched]. Brain specimens were harvested 8 h following this final injection using the cardiac perfusion protocol (described below).

Thioflavin S/SPIO formulation and administration

Thioflavin S was prepared in saline and filtered using 45 μm Luer-Lok syringe filters. A 2 mg/kg dose of thioflavin S was administered to each animal. Approximately, 10 mg/kg SPIO nanoparticles were given to respective mouse cohorts. Intravenous injection of both thioflavin S and SPIO solutions was accomplished through the caudal tail vein in mice anesthetized with isoflurane.

Cardiac perfusion procedure

Following exposure or treatment, deeply anesthetized animals were laid to ice after which the thoracic cavity was accessed via a sharp transverse incision into the abdomen. This was followed by a series of longitudinal cuts with a scalpel to open the thoracic cavity, which then was stabilized with a retractor. Perfusion was performed with a 25-gauge syringe containing ice-cold PBS (30 mL, pH 7.4) inserted through the left ventricle and injected slowly into the ascending aorta. Upon initiation of perfusion, the right atrium was snipped to facilitate drainage of the systemic venous return. Immediately following PBS perfusion, 30 mL of 4% paraformaldehyde (PFA) (pH 7.4) was perfused. With perfusion completed, the animals were decapitated and their brains quickly harvested and fixed in 4% PFA overnight at 4C. Brains to be immuno-stained were cryoprotected for two days in 10% sucrose at 4°C. Cryoprotected brains were then embedded in Tissue-Tek optimum cutting temperature (OCT) compound for cryosectioning and stored at −80°C. All other fixed tissue was embedded in paraffin for sectioning.

Immunohistochemistry

Brains embedded in OCT were cut into sagittal sections (10 μm) using a Tissue-Tek cryostat and mounted onto charged glass slides. Prior to staining, slides were washed with PBS before being subjected to a citrate buffer antigen retrieval protocol. Briefly, slide mounted sections were incubated for 30 min in a 100°C bath of sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6.0) and allowed to cool for 20 min before PBS washing and blocking. Treated sections were then incubated overnight at 4°C with primary antibodies: Beta Amyloid 17–24 (4G8) mouse monoclonal antibody (1:200 dilution, Biolegend, San Diego, CA, USA). Following PBS washes, the sections were subsequently incubated with secondary antibodies goat anti-mouse Alexa 647-conjugated IgG (1:200, ThermoFisher Scientific, Pittsburgh, PA, USA) for 1 h at room temperature. The sections were then washed with PBS and cover-slipped with an antifade mounting medium (Vector Laboratories, Burlingame, CA).

Perls’ staining for SPIO nanoparticle detection

Brains were manually cut into 2-mm coronal sections using a mouse brain matrix and all sections were embedded into a single paraffin block. Approximately, 5-μm coronal sections were cut on a microtome and mounted onto charged glass slides. Paraffin sections were deparaffinized in xylene and rehydrated in a series of ethanol baths. Following rehydration, slides were incubated in freshly prepared Perls’ Prussian Blue solution (1:1 potassium ferrocyanide and hydrochloric acid) for detection of ferric iron, including SPIO nanoparticles found in tissue parenchyma. After washing in water, slides were counterstained in nuclear fast red solution for 5 min. Finally, sections were washed, dehydrated, cleared in xylene, and mounted using Organo/Limonene mount (Sigma Aldrich, St. Louis, MO).

Assessment of the distribution of the probes via stereological analysis

High-resolution digital images of whole slides in brightfield at 20x magnification to a resolution of 0.5 μm/pixel were produced using the Leica SCN400 Slide Scanner. Images were then analyzed in a web-based digital slide-viewing environment called the Digital Imaging Hub (Leica Biosystems) using Tissue Image Analysis (TIA) tools. A color definition file was made to identify positive Perls’ Prussian Blue stain in mouse spleen tissue, which contains a high level of iron and serves as a positive control tissue. Regions of interests (ROIs) drawn for individual brain sections were then analyzed for total tissue area and positive stained area as determined by the color definition profile. These values were used to calculate percent positive tissue area, which represented SPIO nanoparticle distribution in the brain.

Confocal microscopy

Whole section fluorescence imaging was performed on mounted slides from LPS-treated and untreated mouse cohorts with i.v. injection of thioflavin S to assess the BBB permeability for small molecules. Imaging was performed on spinning disk confocal system equipped with a Nikon Eclipse Ti microscope, a Yokogawa CSU-X1 spinning disk heard, and Andor DU–897 EMCCD camera. The sections were visualized by montage scan with 10x Plan Apo (air) 0.45 NA WD 4.0 mm objective, with or without 1.5x intermediate magnification, with 488 nm and 647 nm laser lines to excite Thioflavin S and Alexa Fluor 647 respectively, and appropriate emission filters (525 nm (±18 nm), and 641 nm (±75 nm)). All slides were imaged with identical settings utilizing NIS-Elements imaging software for acquisition and analysis.

Statistical analysis

All statistical analysis was performed using Graph-pad Prism 7.0 (Graphpad Software, Inc.). Error bars represent the standard error of the mean (SEM). To assess differences in survival at 24 h post LPS treatment between 5XFAD and wt mice receiving the same treatment, a chi-square test was used. Comparison of means of more than two groups was done using one-way analysis of variance (ANOVA) followed by the Newman-Keuls post-test.

RESULTS

The maximal tolerable dose of LPS in wt mice is 3 mg/kg, but not repeated doses

To assess the extent to which LPS could induce effective BBB opening without causing severe side effects or lethality, animals were treated with a series of concentrations of LPS via i.v. injection. Previous studies reported that i.p. injection of 3 mg/kg LPS is safe and significantly increased BBB permeability 24 h post injection in a mouse model [36, 53]. In this work, we want to compare the same dose using wt mice albeit with i.v. injection. In the first approach to assess the dose-toxicity relationship, mice (n = 2) were treated with 3 doses of 3 mg/kg LPS, with each injection 5 h apart. Immediately after injection, all animals were recovered from anesthesia and appeared healthy, nevertheless, the slow but probably strong inflammation [54] caused by repeated doses resulted in 100% fatality (Table 1). When the injection frequency was reduced to two doses (n = 3) of 3 mg/kg LPS, 5 h apart, treatment was still highly lethal at 24 h post LPS injection. Similar to the previous report for i.p. injections, we found that i.v. injection of a single dose of 3 mg/kg of LPS was well tolerated with approximately 90% survival in wt mice (n = 18), albeit 5XFAD mouse (n = 21) survival using the same dose was much lower, at 57% (p < 0.05). There is a small chance of death in mice if combined with large injection volume. However, if the injection volume was reduced from 120 μL to less than 100 μL, the animals were able to tolerate the treatment. At low LPS doses ranging from 0.01–1 mg/kg, animals (n = 4, each) there was 100% survival in wt mice.

Table 1.

Effect of LPS dose on lethality of wt and 5XFAD mice

LPS Dose (mg/kg)	# Animals treated	# Died	% Lethality
0.01	4	0	0
0.1	4	0	0
1.0	4	0	0
3.0	18 (21)	2 (9)	11 (43)
3.0 (x2)	3	2	67
3.0 (x3)	2	2	100

From 0.01–3 mg/kg of LPS was injected i.v. via the tail vein of mice in a volume <100 μL. Double (x2) or triple doses (x3) were injected 5 h apart. Data for 5XFAD mice is shown only for 3 mg/kg dose, in parentheses.

LPS-induced BBB opening on 5XFAD mice

Next, we assessed BBB permeability on 5XFAD mice using the highest tolerable dose of 3 mg/kg LPS as described in the toxicity study above. To facilitate the detection and visualize the distribution of the delivered molecules upon BBB compromise, we used thioflavin S, since not only does it have high affinity for Aβ but it cannot cross the BBB under normal physiological conditions [55]. Further, thioflavin S is a fluorophore that can readily be detected in tissue sections. As shown in Fig. 1A, ex vivo treatment of a 5XFAD mouse brain section with thioflavin S resulted in robust staining of Aβ in the hippocampus, cortex, and subiculum. However, due to its inability to penetrate the BBB, i.v. injection of thioflavin S to 5XFAD does not show signal in the brain (Fig. 1B).

Fig. 1.

Assessment of BBB penetration capability and Aβ specificity of thioflavin S on 5XFAD mice. A) ex vivo thioflavin S staining of the brain slide depicting the hippocampal region. B) i.v. injection of thioflavin S resulted in no fluorescence in the brain (CTX, cortex; SUB, subiculum; DG, dentate gyrus).

In the next experiment, as depicted in the timeline of Fig. 2, aged 5XFAD mice (13–15–month-old) were injected i.v. with 3 mg/kg LPS, then 24 h later the same LPS-treated animals were injected with i.v. thioflavin S (2 mg/kg). This optimized BBB opening timeframe was reported in the past [36]. Eight hours later, animals were cardiac perfused, sacrificed, and brain sections were prepared for imaging. As shown in Fig. 3A, if 5XFAD mice were treated with thioflavin S alone, this resulted in no fluorescence signal in the brain, albeit remarkable Aβ amyloid expression was detected via immunohistochemistry (Fig. 3B, C). In contrast, when 5XFAD mice were i.v. injected with the same thioflavin S dose after BBB compromise by LPS, the compound now can reach the brain and bind to Aβ plaques at regions of high plaque burden, such as the cortex (Fig. 3D) (thioflavin S distribution to the other areas in the brain, including subiculum, CA1, and CA2 regions can be found in Supplementary Figure 1) and the data correlate with Aβ immunostaining using Abs on consecutive slides (Fig. 3E, F).

Fig. 2.

Timeline and procedures to test LPS-mediated opening the BBB in transgenic and wt mice using small and large materials.

Fig. 3.

Effect of LPS on BBB opening using age-matched 5XFAD mice. A) representative brain sample of an aged 5XFAD mice (14 months old) treated with 2 mg/kg thioflavin S alone via i.v. injection. Data were imaged under thioflavin S channel (green). B) The same brain section (same slide, double stained) was imaged using red-shift dyes labeled on secondary antibodies after the brain specimen was treated with anti-Aβ Abs, and (C) merged data of (A) and (B). D) In a different cohort of animals, age-matched 5XFAD mice were treated with LPS (3 mg/kg) 24 h prior to thioflavin S i.v. injection. The presence of thioflavin S in the brain was detected in thioflavin S channel (green). E) Aβ was detected (same slide) via red-shift dyes labeled on secondary antibodies after the brain specimen was treated with anti-Aβ antibodies. F) Merged data of (D) and (E).

LPS-induced BBB opening for large materials

Next, we sought to determine whether LPS-pretreatment makes the BBB permeable to large materials for future delivery applications of not only small molecules but also large vehicles, such as Abs or nanoparticles. In this work, 30-nm super paramagnetic iron oxide (SPIO) nanoparticles (10 mg/kg) were injected i.v. into wt and 5XFAD mice 24 h post LPS treatment (3 mg/kg) also via i.v. injection (Fig. 2). Brains were collected 8 h later following transcardial perfusion, and prepared sections were stained for iron using our previously described protocols [52, 56]. When young (3–5-month-old) wt and 5XFAD mice were pretreated with LPS, they showed similar results of minimal positive Perl’s stain in the brain, at the detection threshold, suggesting limited nanoparticle delivery to the brain (Fig. 4G). However, when these procedures were repeated on aged 5XFAD mice (13–15-month-old) treated with LPS before SPIO injection, significant SPIO positive stain was detected in the brain parenchyma (Fig. 4C–F). Aged 5XFAD mice injected with SPIO but not LPS showed lit tle stain, suggesting that with advanced age and disease progression, the BBB is intact enough to limit SPIO delivery to the brain in the absence of LPS pretreatment (Fig. 4B). Overall, the data suggest that even at advanced stages of disease in our model of AD, the BBB is still somewhat intact but more susceptible to increased BBB permeability mediated by LPS to allow delivery of nanoparticles as large as 30 nm to the brain.

Fig. 4.

LPS-induced BBB opening allows delivery of large molecules. A) Timeline of the study with a figure showing caudal tail vein injection. B) Mouse brain section receiving SPIO injection but no LPS pretreatment. C, D) Detection of SPIO nanoparticles (arrows) in the brain parenchyma of 5XFAD mice after i.v. injection. E. F) Default color of (C & D) for analysis. G) Quantitative analysis of SPIO distribution in the brain of 5XFAD versus wt mice of different ages after the animals were pretreated with or without LPS, aged mice were 13–15-month-old while young mice were 3–5-month-old; *p < 0.05.

Next, we sought to determine whether LPS-mediated BBB opening is due to AD pathology or through normal aging processes. Aged wt animals (13–15-month-old, n = 7) were treated with LPS and SPIO in identical dose and process as mentioned with other cohorts. As shown in Fig. 4G, there is no significant difference in the SPIO permeability in the aged wt mice versus the young, 3-month-old counterparts. Taken together, the data suggest that LPS-induced opening of the BBB is due to AD pathology and not aging alone.

DISCUSSION

The natural paradox of the BBB is that is serves as a formidable barrier to protect the brain from being harmed by exogenous and foreign materials. On the other hand, it also excludes many drugs intended for therapy [57]. The endothelial cells of the BBB provide an important interface between the brain and other tissue environments and possess tight junctions of severely limited permeability [58]. Apparently, actively induced opening of the BBB faces the reality of the impact of the material on the body as a whole, as the effect may not be targeted to only brain endothelium. Although it has been demonstrated that bacterial products such as LPS could be used to induce BBB opening via robust inflammation [59]; the toxicity caused by the agent is conspicuous because of LPS production of different inflammatory mediators including IL-1β, IL–6, TNF-#, nitric oxide, MMP–2, and prostaglandins [60–62]. Specifically, IL-1β is implicated in the failure of BBB [63, 64]. Our data suggest that the maximal tolerable i.v. dose of LPS for wt mice is 3 mg/kg without co-administration of anti-inflammatory drugs. When this threshold dose is combined with large injection volume or with repeated doses, even at an extended interval between injections, the combination can cause fatality. However, the dosing in the 5XFAD mice was approximately 30% more lethal (p < 0.05), suggesting that they are more susceptible to LPS-mediated toxicity.

To test the BBB of aged 5XFAD mice, a transgenic model of AD, one of the candidate compounds we studied is thioflavin S. It is a small molecule but its possession of a charge moiety prevents it from penetrating the BBB [55]. Another advantage for this particular study is that thioflavin S is a fluorescent dye, suitable for characterization of BBB compromise via direct visualization using microscopic imaging and histology. Further, thioflavin S is an Aβ-binding molecule and we demonstrated that if the BBB is compromised by LPS, thioflavin S not only will cross the barrier but it will report the expression of Aβ in the areas of interest, such as the subiculum, CA1, CA3, dentate gyrus (Supplementary Figure 1), and cortex (Fig. 3D, F), all known to have high plaque burden in 5XFAD mice. The data also showed that the BBB in aged 5XFAD mice (13–15-month-old) is somewhat intact as thioflavin S fluorescence is not detected in the brain without LPS treatment (Fig. 3A, C), suggesting that even at advanced ages, AD-related pathology has not disrupted the BBB significantly enough to allow thioflavin S to enter the brain parenchyma and target Aβ plaques. Further, the results suggested that aside from small molecules, the LPS-compromised BBB in aged 5XFAD mice is also permeable to large structures such as SPIO nanoparticles. These data provide useful information for future applications of antibodies and nanoparticles as therapeutics for AD. The SPIO nanoparticles used in this work are 30 nm in size, much larger than any tested anti-Aβ antibodies in clinical trials, and could cross the BBB to enter the brain parenchyma in aged 5XFAD mice with LPS pretreatment. Interestingly, the degree of SPIO detected in 5XFAD mice was 4–fold higher than that found in wt mice (both 3–5 and 13–15-month-old cohorts) receiving the same LPS pretreatment (Fig. 4G), alluding that AD pathology and not aging-related changes play a role in BBB compromise after LPS treatment. A caveat worth mentioning is that regardless of the route of distribution, our i.v. injection data corroborates earlier studies using i.p. injections [36] that all tested doses of LPS, even a very small amount, causes remarkable body weight loss (Fig. 5). Animals treated with LPS experienced significant weight loss within 24 h after i.v. injection of any LPS dose described above. All doses of LPS produced a significant decrease in body weight compared to untreated controls, even at a very low dose, such as 0.01 mg/kg LPS, where 8% weight loss occurred. If comparing the effect across the tested doses, there is an increase in weight loss of animals received a dose of 3 mg/kg LPS compared to those received 0.01 mg/kg.

Fig. 5.

Effect of LPS on weight loss 24 h after wt mice were injected i.v. with different doses of LPS ranging from 0.01–3 mg/kg; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

In conclusion, we have demonstrated that the BBB of aging 5XFAD mice is intact, albeit more vulnerable than wt counterparts 24 h after i.v. injection of LPS. Theoretically, LPS-induced opening of the BBB can potentially be used to deliver antibodies or large materials for therapy or diagnostic application, but for future clinical translation, the use of LPS should be administered with great circumspection. As demonstrated in this work, probably, the dose should be reduced to prevent toxicity. Another caveat worth mentioning is that even at low dose of LPS, weight loss of treated subject is inevitable. Therefore, clinical implementation of LPS-induced BBB opening, particularly for AD patients would require ways to overcome the weight loss and increased lethality seen in our preclinical study either through the use of a different pro-inflammatory compound, different dosing, or even pairing with an anti-inflammatory agent that reduced unintended effects while not inhibiting BBB opening.