Selective Uptake of Macromolecules to the Brain in Microfluidics and Animal Models using the HAVN1 Peptide as a Blood-Brain Barrier Modulator

Abstract

Monoclonal antibodies (mAbs) possess favorable pharmacokinetic properties, high binding specificity and affinity, and minimal off-target effects, making them promising therapeutic agents for central nervous system (CNS) disorders. However, their development as effective therapeutic and diagnostic agents for brain disorders is hindered by their limited ability to efficiently penetrate the blood-brain barrier (BBB). Therefore, it is crucial to develop efficient delivery methods that enhance the penetration of antibodies into the brain. Previous studies have demonstrated the potential of cadherin-derived peptides (i.e., ADTC5, HAVN1 peptides) as BBB modulators (BBBMs) to increase paracellular porosities for penetration of molecules across the BBB. Here, we test the effectiveness of the leading BBBM peptide, HAVN1 (Cyclo(1,6)SHAVSS), in enhancing the permeation of various monoclonal antibodies through the BBB using both in vitro and in vivo systems. In vitro, HAVN1 has been shown to increase the permeability of fluorescently labeled macromolecules such is a 70 kDa dextran, 50 kDa Fab1, and 150 kDa mAb1, by 4 to 9-fold in a 3D-BBB microfluidics model using a human BBB endothelial cell line (i.e., hCMEC/D3). HAVN1 was selective in modulating the BBB endothelial cell compared to the pulmonary vascular endothelial (PVE) cell barrier. Co-administration of HAVN1 significantly improved brain depositions of mAb1, mAb2, and Fab1 in C57BL/6 mice after 15-min in the systemic circulation. Furthermore, HAVN1 still significantly enhanced brain deposition of mAb2 when it was administered 24 h after the administration of the mAb. Lastly, we observed that multiple doses of HAVN1 may have a cumulative effect on the brain deposition of mAb2 within a 24-h period. These findings offer promising insights into optimizing HAVN1 and mAb dosing regimens to control or modulate mAb brain deposition for achieving desired mAb dose in the brain to provide its therapeutic effects.

Graphical Abstract

INTRODUCTION

Monoclonal antibodies (mAbs) have been successfully used as therapeutic and diagnostic agents for patients with cancers,^1–3 autoimmune diseases (e.g., rheumatoid arthritis, multiple sclerosis, psoriasis),^4–6 and Alzheimer’s disease (AD).^7–9 They are the fastest growing class of drugs and represent attractive therapeutic modalities for the diagnosis and treatment of central nervous system (CNS) disorders. Monoclonal antibodies can be designed with known targets and mechanisms of action with high specificity. Another advantage of mAbs is that they have a long residence or circulating half-life (~ 14 days in humans), due in part to the recognition of their Fc region by the Fc neonatal receptor (FcRn) that participates in cellular recycling mechanism, and protects mAbs from degradation and plasma clearance. Despite these advantages, mAbs have had limited success in treating brain diseases partly because of their inability to cross the BBB.^{10, 11} Several clinical trials of mAbs for treatment of brain diseases were not approved by health authorities due to the lack of efficacy.¹² The recent approval of Leqembi (lecanemab-irmb), an anti-Aβ mAb, for clearing brain amyloid plaques in early-stage Alzheimer’s diseases (AD) represents a promising development in the utilization of mAbs as potential treatments for brain diseases.^{13, 14}

Due to the great potential of mAbs as therapeutic and diagnostic agents in peripheral diseases, extensive efforts have been directed toward the use of mAbs for treatment of brain diseases. Unfortunately, the development of mAbs for brain disease treatment has met with various challenges. One of the challenges is that mAb has difficulty to diffuse passively across the BBB unlike small molecule drugs; this is due to mAb’s physicochemical properties (e.g., large size, high hydrogen bonding potential). The physicochemical properties of mAbs are unfavorable for transcellular passive diffusion across the BBB as they are unable to partition into the cell membranes of the BBB endothelial cells. Antibodies also are unable to utilize paracellular diffusion routes across the BBB due to the presence of tight junction and junctional proteins between endothelial cells that restrict passage of macromolecules to the brain tissue. Several strategies have been exploited to increase the CNS exposure of mAbs. One approach is through chemical modification of the mAbs, such as engineering mAbs to bind transferrin or insulin receptors for receptor mediated transcytosis across the BBB.^{15, 16} Other approaches have focused on the brain microvasculature to increase the BBB permeability of mAbs. Bevacizumab (2.5 mg/kg) was delivered to the brain using osmotic BBB disruption (BBBD) to successfully treat pediatric patients with brain tumor to regain muscle strength.¹⁷ Unfortunately, a parallel study demonstrated that osmotic BBBD could cause astrogliosis, seizures and cerebral edema as side effects in some treated patients.¹⁸ More recently, a focused ultrasound (FUS) method in combination with microbubbles (MB) has effectively delivered drugs to the brain.^19–25 FUS-MB has been shown to open the BBB safely, temporarily, and repeatedly in amyloid-rich brain regions of patients with AD.²⁶ Therefore, any method that can safely deliver mAbs to the brain could benefit patients with brain diseases such as Alzheimer’s disease (AD), Parkinson’s disease, multiple sclerosis, and brain tumors.

The osmotic BBBD method has been shown to increase permeation of various molecules through the BBB paracellular pathways. This method utilizes a hypertonic mannitol solution that is perfused through the BBB to induce shrinkage of BBB endothelial cells to disrupt the paracellular pathways. As a result, it increases the porosity of the paracellular pathway and enhances the passage of molecules from the blood into the brain. An alternative method to increasing BBB paracellular permeability is through the inhibition of protein-protein interactions in the adherens junction of the BBB. Specifically, our laboratory has demonstrated that cadherin peptides (i.e., HAV and ADT peptides) that inhibit the protein-protein interactions of cadherin can be used as transient and reversible blood-brain barrier modulators (BBBMs).^27–29 Both the duration and magnitude of BBB permeability response can be determined by the cadherin peptide selected. Thus, the use of these peptides for enhancing the brain delivery of a variety of agents including small molecule anti-cancer drugs (such as daunomycin, camptothecin, adenanthin);^30–32 paracellular-marker molecules (¹⁴C-mannitol, Gd-DTPA, and 25 kDa IRdye800CW-PEG);^{29, 30, 33, 34} drug efflux-pump substrates (R800);³³ as well as various biological agents including peptides (cIBR7 and cLABL);²⁸ and proteins (13 kDa BDNF, 65 kDa albumin, and 150 kDa mAb) have been reported.^{27, 28, 35–37}

One way to deliver mAbs to the brain is by enhancing their permeation through the paracellular pathways of the BBB. In this study, we assessed the activity of HAVN1 peptide (Cyclo(1,6)SHAVSS) as a BBBM to enhance the delivery of mAbs and Fab fragments to the brain in in vitro and in vivo systems.²⁷ The effects of HAVN1 peptide on the distribution of mAbs and antibody Fab fragments in the brain and peripheral organs (i.e., heart, lung, liver, kidney, and spleen) were determined to evaluate the selectivity of HAVN1 in modulating the BBB over vascular endothelial cells of the peripheral organs. The effect of delaying delivery of HAVN1 after mAb administration was determined to confirm the ability of HAVN1 to increase the porosity of the BBB and improve the uptake of circulating mAbs. Finally, we investigated the effects of multiple administrations of HAVN1 to enhance mAb brain depositions.

MATERIALS AND METHODS

Antibody conjugation with IRDye800CW NHS ester

Both mAb1 (IgG1) and mAb2 (IgG4) are humanized antibodies with different CDR sequences. To label the monoclonal antibodies, IRDye^® 800CW NHS Ester (Licor, Lincoln, NE) was reacted to the amino groups of the N-terminus and Lysine residues of the antibody to produce IRDye^® 800CW-labeled mAbs. The unreacted IRDye was removed using a Zeba Spin Desalting Column with a 7 kDa molecular weight cutoff (Fisher Scientific Inc., Hampton, NH). Purity the IRDye-labled mAbs was determined by using SDS-PAGE (Supplemental Figure S1). An Odyssey CLx NIR scanner was used to scan the SDS-PAGE gel at 800 nm to confirm the removal of all excess dye.

The antibody concentrations and degree of labeling (dye-to-protein molar ratio) were determined following purification using UV-Vis spectroscopy with formulas that incorporated a correction factor of 0.03 for IRDye800CW. This correction factor indicates that the absorbance at 280 nm corresponds to 3% of the absorbance at the ${\text{A}}_{\text{max}}$ (780 nm). The extinction coefficients of the antibody $\left(\text{ε}\right)$ and fluorescent dye $\left(\text{ε'}\right)$ were also considered in the calculations.

$$Antibody Concentration \left(M\right)=\frac{A_{280}-\left(A_{max}\right)\times CF}{\epsilon }\times dilution factor$$

$$Degree of labeling= \frac{A_{max} of labeled antibody}{{\epsilon }^{\prime }\times protein concentration \left(M\right)}\times dilution factor$$

Microfluidics 3D-BBB and pulmonary vascular endothelial (PVE) cell culture

The microfluidic BBB culture model was established using three-lane OrganoPlates (Mimetas BV, Netherlands). The OrganoPlate is configured on a 384-well plate format and has 40 individual microfluidic units (Supplemental Figure S2). Each microfluidic unit consists of three channels, a top channel (capillary perfusion lane), middle channel (extracellular matrix gel), and a bottom channel (brain extracellular space). Each channel has both inlet and outlet ports for sampling both the apical (blood) and basolateral (brain) compartments, as well as an observation window for examining barrier formation and integrity using either phase-contrast brightfield or fluorescence microscopy methods (Fig. 1A). The top and bottom channels have dimensions of 300 × 220 µm (WxH) and the gel channel has dimension of 350 × 220 µm. A phase guide separates the top and bottom channels from the gel channel in the microfluidic chamber and is used to pattern the extracellular matrix (ECM) gel into the middle compartment (100 × 55 µm WxH).

Figure 1.

(A) A diagram of Mimetas^™ organoplate plate that is used as an in vitro 3D-BBB model to evaluate the effects of HAVN1 and serum on the BBB permeation of FDX-70 kDa, Fab1 (50 kDa), and mAb1 (150 kDa) after 30 min in circulation. (B) The effects of HAVN1 concentration at 0 mM (control), 0.01 mM, 0.1 mM, and 0.5 mM in modulating the intercellular junction of the in vitro 3D-BBB model in enhancing the transport of FDX-70 kDa across the BBB in 30 min perfusion time. As the dose of HAVN1 increased, there was an increased in the permeation of FDX-70 kDa into the brain side as detected by NIRF imaging. (C) P_app permeabilities (in cm/s) of FDX-70 kDa, Fab1, and mAb1 across the BBB upon coadministration with 0.5 mM HAVN1. Low permeabilities were observed for FDX-70 kDa, Fab1, and mAb1 across the BBB when they were delivered alone. HAVN1 significantly enhanced permeability of all macromolecules in cell culture conditions with and without serum. At concentrations of mAb and Fab used (5 mg/ml), there was greater apical (A)-to-basolateral (B) than B-to-A permeabilities. Serum influenced permeability of all macromolecules examined – presence of serum decreases permeability of molecules in the absence of HAVN1. Values represent mean ± SEM of 3 microfluidic cells. *p<0.05; **p<0.01; ****p< 0.001.

Prior to seeding the hCMEC/d3 or pulmonary vascular endothelial (PVE) cells into the microfluidic units, a 2 µL solution of ECM was dispensed into the middle channel. The ECM consisted of Cultrex 3D rat tail collagen (R&D systems) mixed in a 8:1:1 ratio with 100 mM HEPES, and 37 mg/mL NaHCO₃. The plate was placed in the incubator for 15 min at 37 ºC and 5% CO₂, after which, 30 µL of HBSS was added to gel inlet to prevent the gel from drying out. The plate was then placed back into the incubator for an additional 1.5 h to establish the ECM gel in the middle channel. Human BBB endothelial cell line (hCMEC/d3) or primary PVE cells (ScienCell; Carlsbad, CA) were seeded into the top channel of the microfluidic unit by dispensing 2 µL of a cell suspension (10,000 cells/µL) into the inlet port. Following cell seeding, 50 µL of complete EBM-2 media was added to inlet port and the plate was placed on its side in the incubator to allow the hCMEC/d3 or PVE cells to pattern and attach against the ECM gel. Following a 3-h side incubation, 50 µL of complete EBM-2 media was added to the outlet port and the plate was then placed back in the incubator on a Mimetas^™ interval rocker which switched between +7 and −7 inclinations every 8 min, to create bidirectional fluid flow. Cells were grown to confluency in 37 ^oC and 5% CO₂ incubator with media changes every other day.

In vitro permeability of macromolecules through 3D-BBB and 3D-PVE microfluidics system

At endothelial cell confluency, to determine permeability from the apical to basolateral compartments (A-to-B), 20 µL of EBM-2 media was added to gel and bottom channel inlet and outlet ports. After a 5-min equilibration period, FDX-70 kDa , mAb1or Fab1 in EBM-2 media (final concentration of 1 µM) with and without HAVN1 peptide (final concentration of 0.01, 0.1, or 0.5 mM) were added to the top channel of the microfluidic unit by placing 40 µL and 30 µL into the inlet and outlet ports, respectively. A sample (10 µL) was immediately removed from the inlet port to determine initial apical concentrations. This was diluted 5-fold with EBM-2 media in a black 96-well plate. To determine permeability of the various macromolecules, 20 µL samples were collected from the middle and bottom chamber inlet and outlet ports at 30 and 60 min to assess permeability. To determine permeability from the basolateral (brain)-to-apical (blood) (B-to-A) direction, 20 µL of solutions containing the same fluorescent permeability markers and concentrations as mentioned above were added to the gel and bottom channel inlet and outlet ports while complete EBM-2 media was added to the inlet and outlet ports of the top chamber, 40 and 30 µL respectively. A sample (10 µL) was taken immediately and replaced from the basolateral side to determine initial starting concentrations. At 30 and 60 min, 40 and 30 µL samples were removed from the inlet and outlet ports, respectively, of the top channel to determine permeability. Permeability was also assessed under serum-free conditions using EBM-2 media that contained all the necessary supplements except FBS. Qualitative determination of the permeability responses to the cadherin peptide was determined visually using an Evos cell imaging system with fluorescent filter for measuring fluorescein labeled dextran. Quantitative determination of permeability in the microfluidic units were determined using a Varioskan Lux plate reader, FDX-70 kDa (ex/em-480/520), Fab1 (ex/em −685/712), and mAb1 (ex/em - 785/812) with concentrations of macromolecules from the donor and receiver compartments determined by use of standard curves.

Apparent permeability was determined by Fick’s law of diffusion:

Apparent permeability equation 1: $Papp\left(\frac{cm}{s}\right)=\left(\frac{\Delta Cr}{\Delta time}\right)\ast \left(\frac{Volume}{Area\ast Cd}\right)$ where $\text{Cr}$ is concentration in receiver compartment and $\text{Cd}$ is concentration in the donor compartment.

Values represent the mean ± SEM of three microfluidic units per treatment group.

Animal studies

C57BL/6 mice obtained from Jackson Laboratories or breeding colonies at The University of Kansas were used in this study and with 6 mice per each group (n = 6). A random selection of male and female mice with similar body weights were used for each group. The animal studies were performed according to an animal protocol approved by the Institutional Animal Care and Use Committee (IACUC) at The University of Kansas. All animals were cared for by personnel of the Animal Care Unit (ACU) at The University of Kansas under the supervision of veterinarians.

In vivo antibody delivery in mice

IRDye-labeled mAbs were administered into healthy C57BL/6 mice via intravenous (IV) delivery at a dose of 50 nmol/kg (7.5 mg/kg) of mAb or 160 nmol/kg (8 mg/kg) of Fab with and without the addition of 18 μmol/kg HAVN1 cadherin peptide.²⁷ Blood samples (~100 μLs) from the submandibular vein were collected immediately prior to euthanasia into chilled plasma collection tubes containing lithium heparin as anticoagulant. Samples were kept on ice until centrifugation for 3 min at 10000×g and the plasmas were stored at –80 ºC until analysis. Mice were sacrificed after various time points via transcardial perfusion of an ice-cold HBSS solution while the mice were deeply anesthetized with 4–5% isoflurane.²⁷ The brain, heart, lungs, liver, spleen, and kidney were collected and kept on ice until analysis. The isolated brains were scanned for imaging using NIRF Licor Odyssey CLX by taking eight optical sections (slices) of the brain at 0.5 mm increments. The scan was started at 4 mm depth of the brain to collect 8 image brain slices and the summed of fluorescence intensities of all optical sections (slices) were summed to yield a total fluorescence intensity value per each brain.

Antibody whole tissue and plasma quantitation

All organs were isolated and washed with PBS and mAb depositions were determined in the organ homogenates in 2.0 mL PBS. Homogenized organ (200 μL) was aliquoted to a 96-well plate followed by quantification of IRdye-labeled mAb or Fab intensity using NIRF imaging by an Licor Odyssey CLX scanner. A calibration curve was used to determine the concentrations of mAb or Fab in the homogenate. Organ calibration curves were generated by spiking various standard concentrations of IRDye-labeled antibody into homogenized organ solutions followed by determination of the sample fluorescence intensity.

Effects of HAVN1 dosing schedule on antibody brain deposition

The effects of delayed administration of HAVN1 after injection of IRdye-labeled mAb2 was assessed in C57BL/6 mice (n = 6). For each study, the animal was injected through IV via lateral tail vein with 50 nmol/kg mAb2. After antibody administration, the mice received doses of HAVN1 peptide (18 μmol/kg) or vehicle at 24-h or 72-h time point. The mice were sacrificed 15 min following peptide or vehicle administration. Blood, brain, heart, lung, liver, kidney, and spleen samples were taken from each animal and the antibody deposition in each was quantitated using NIRF imaging by Odyssey CLx scanner.²⁷

To test the effect of multiple doses of HAVN1 on mAb brain deposition, C57BL/6 mice were separated into 2 groups (n = 6) and administered an initial dose of mAb2 (50 nmol/kg) with HAVN1 peptide (18 μmol/kg). After 24 hours, each group was administered either second dose of HAVN1 peptide (18 μmol/kg) or vehicle. Then after 15 min, a blood sample was collected, and each animal was euthanized via transcardiac perfusion. Brain, heart, lung, liver, spleen, and kidney samples were collected from each and analyzed for mAb content.

Statistical analysis

To determine statistical significance for improved delivery of mAb into the brain by HAVN1 peptide, ANOVA with Student–Newman–Keuls post hoc tests were utilized to compare brain depositions of mAb and Fab with and without HAVN1. A criterion for a significant difference of acquired data is a p-value of less than 0.05. The sample size was determined based on the number needed to achieve a power of 90% at a significance level of 0.05 in a study which aims to show a difference in means two animal groups.^{38, 39} Mean ± standard error of the mean (SEM) was used to report the experimental values.

RESULTS

Synthesis and purification of IRDye800CW-labeled antibodies

To fluorescently label all antibodies, IRDye800CW was conjugated to free amino groups of each antibody to form stable conjugates. Excess dye was removed from the reaction mixture by using a Pierce Zeba desalting spin column with a molecular weight cutoff of 7 kDa. To confirm purity, the antibody conjugates were evaluated with SDS-PAGE and scanned with an Odyssey CLx NIR imager. All antibodies showed a single band at the appropriated molecular weight (Supplementary Figure S1). The final degree of labeling was 2.0, 1.0, 0.32, and 0.47 for mAb1, mAb2, Fab1, and Fab2, respectively.

HAVN1 improved permeability of mAbs across in vitro microfluidics 3D-BBB model

A BBB microfluidics in vitro model cultured (Fig. 1A) with human brain capillary endothelial cells was used to examine the permeability of fluorescently labeled large molecules with and without HAVN1 peptide treatment. Initial experiments examined the permeability enhancement of a 70 kDa dextran (FDX-70 kDa) molecule incubated with various concentrations of HAVN1 in the microfluidic BBB culture model. Treatment with HAVN1 produced a concentration-dependent increase in the permeability of FDX-70 kDa across the BBB monolayer (Fig. 1B). Next, the ability of HAVN1 to increase macromolecule permeability was evaluated using FDX-70 kDa, a 50 kDa Fab1, and a 150 kDa mAb1. Treatment with HAVN1 significantly enhanced the permeability of all macromolecules tested compared to when molecules were incubated without HAVN1 peptide (p<0.01, Fig. 1C). Specifically, HAVN1 increased the apical to basolateral (A->B) delivery of macromolecules in the microfluidic BBB culture model by 4.7-fold, 5.6-fold, and 8.9-fold for FDX-70 kDa, Fab, and mAb, respectively. In the absence of HAVN1 peptide, the apical-to-basolateral (A->B) permeability was low and showed a rank order of FDX-70 > Fab1 > mAb1 (Fig. 1C). In contrast the basolateral-to-apical (B->A) permeability of the tested molecules were mAb1 > Fab1 > FDX-70 (Fig. 1C). Interestingly, the presence of serum resulted in decreased permeability of all molecules in the A->B direction in the absence of HAVN1 (Fig. 1C). A similar effect of serum was observed for mAb1 in the B->A direction (Fig. 1C).

In vitro selectivity HAVN1 for the 3D-BBB over 3D-PVE microfluidics system

One of the major concerns with the use of cadherin peptides as BBBM is the potential lack of selectivity between the BBB endothelial cells and the peripheral vasculature endothelial cells because they all have VE-cadherin in their adherens junctions. Therefore, the selectivity of HAVN1 for modulating endothelial cell permeability was compared in the hCMEC/D3 (human brain cell line) versus PVE (Fig. 2). Qualitatively, in the absence of HAVN1 peptide, FDX-70 kDa leak-proof barriers were observed in both the BBB (Fig. 2A; Left) and PVE microfluidic models (Fig. 2B; Left). In the presence of 0.5 mM HAVN1, FDX-70 kDa showed increased permeation across the BBB (Fig. 2A; Right) but not the pulmonary EC (Fig. 2B; Right), indicating the selectivity of HAVN1 for the BBB EC over PVE cell barrier. Quantitatively, HAVN1 significantly enhanced the permeation of both FDX-70 kDa and IRdye (1067.11 Da) across the BBB microfluidic model compared to FDX-70 kDa and IRdye alone, respectively, as reflected in their P_app (Fig. 2C). In contrast, HAVN1 did not enhance permeation of FDX-70 kDa and IRdye across the PVE cell barrier (Fig. 2C).

Figure 2.

In vitro selectivity of BBBMs in modulating the BBB endothelial cell (EC) over the pulmonary vascular endothelial (PVE) cell barrier in the absence of serum. (A) The NIRF images of the enhanced BBB permeation of FDX-70 kDa when administered with 0.5 mM HAVN1 (Right) compared to no BBB permeation of FDX-70 kDa alone (Control; Left). (B) No differences in permeation of FDX-70 kDa alone (Control; Left) compared to FDX-70 kDa with 0.5 mM HAVN1 (Right). (C) Quantitative comparison for the permeation of FDX-70 kDa and IRdye (1067.11 Da) across the PVE cell barrier in the absence and presence of various concentrations of HAVN1 (0.01, 0.1, 0.5 mM) as well as in the BBB EC with 0 and 0.5 mM HAVN1. *p<0.05; **p<0.01

HAVN1 enhanced brain delivery of mAbs with minimal impact on peripheral organs

The effect of co-formulation of HAVN1 on the distribution of mAb1 and mAb2 in the brain and other organs was investigated in mice after a 15-min circulation time. Animals were euthanized via transcardiac perfusion with HBSS to remove excess mAb from brain capillaries. Qualitative analysis of whole brain fluorescence scans showed higher brain concentrations of mAb2 when co-administered with HAVN1 compared to control when mAb2 was delivered alone (Fig. 3A). The whole brain concentrations of mice treated with mAb1 + HAVN1 were significantly higher (1.8-fold higher) than those treated with mAb1 alone (p<0.01; Fig. 3B). Similarly, the brain depositions of mAb2 were significantly higher (3.3-fold higher) in mice treated with mAb2 + HAVN1 than those treated with mAb2 alone ((p<0.01; Fig. 3B).

Figure 3.

Co-formulation of therapeutic mAbs with HAVN1 enhances mAb brain delivery. Mice were dosed with 50 nmol/kg (7.5 mg/kg) of either mAb1 or mAb2 with or without 18 μmol/kg HAVN1 peptide via an IV bolus dose. The animals were euthanized 15 minutes following dosage. (A) Whole brains of mice that received mAb2 + HAVN1 or mAb2 alone. Brains were scanned at a depth of 1 mm on a Licor Odyssey Clx tissue scanner. (B) Quantitative comparison of mAb2 and mAb1 brain concentrations when delivered with and without HAVN1 represented in %ID/g. (C) Plasma concentrations of mAb1 and mAb2 in the presence and absence of HAVN1 obtained 10 minutes after administration. (D) Tissue distributions of mAb1 with and without HAVN1 co-formulation. (E) Tissue distributions of mAb2 with and without HAVN1 co-formulation. **p<0.01

HAVN1 had no significant effect on plasma concentrations of mAb1 and mAb2 (Fig. 3C). There was a significantly higher mAb1 concentration in the kidney in mice treated with HAVN1 + mAb1 compared to mAb1 alone (Fig. 3D). Although not statistically significant, there was a possible trend of higher deposition of mAb1 in spleen, liver, lung, or heart of mice treated with mAb1 + HAVN1 compared to mAb1 alone (Fig. 3D). Similarly, tissue concentration in each organ (i.e., kidney, spleen, liver, lung, or heart) for mice treated with mAb2 + HAVN1 had a trend of higher average tissue concentrations compared to the mAb control (Fig. 3E). Overall, these studies suggest that HAVN1 opens the BBB for mAb delivery while having a minimal effect on vasculature of peripheral tissues.

Effects of HAVN1 to enhance brain delivery of Fab fragments

Next, we evaluated the potential of HAVN1 to enhance brain exposure of Fab fragments by co-administering the peptide with either Fab1 or Fab2. In this study, the delivery of Fab fragment was evaluated to determine the effects of Fc domain and size on their permeation through the BBB. It has been suggested that the permeation of full size of mAb through the BBB was via binding and uptake by neonatal Fc receptor (FcRn). The size of Fab fragment (50 kDa) was below the glomerular filtration (65 kDa); thus, the glomerular filtration on the delivery of Fab fragment to the brain was evaluated. To see the enhancement of brain deposition by HAVN1, the dose of Fab1 or Fab2 was 160 nmol/kg compared to 50 nmol/kg for full mAb; this presumably due to a fast glomerular filtration of Fab compared to a full mAb. Here, HAVN1 significantly enhanced brain concentrations of Fab1 (p=0.0073) but not Fab2 (p=0.084, Fig. 4A). Fab2 showed a trend of higher brain deposition when delivered with HAVN1 compared to control. Consequently, both Fabs in HAVN1 groups showed similar fold enhancements compared to their controls, with a 1.4-fold increase observed for Fab1 and a 1.3-fold increase observed for Fab2.

Figure 4.

HAVN1 significantly enhances brain delivery of Fab1, but not Fab2. Mice were dosed with 160 nmol/kg (8.0 mg/kg) of either Fab1 or Fab2 with or without 18 μmol/kg HAVN1 peptide via an IV bolus dose. The animals were euthanized 15 minutes following dosage. (A) Quantitative comparison of Fab1 and Fab2 brain concentrations when delivered +/− HAVN1. (B) Plasma concentrations of Fab collected 10 minutes following IV dosage. Tissue distribution of (C) Fab1and (D) Fab2 determined after a 15-minute Fab circulation time. **p<0.01

It is interesting to find that both Fabs displayed similar plasma (Fig. 4B) and tissue concentrations (Fig. 4C–D) between HAVN1-treated and control groups; this suggests that HAVN1 did not have effects in the vascular system of peripheral organs such as heart, lung, liver, spleen, and kidney (Fig. 4C–D). However, Fab1 and Fab2 had distinct overall biodistribution patterns in peripheral tissues. The deposition of Fab2 in the kidney (Fig. 4D) was about doubled compared to that of Fab1 (Fig. 4C). The second highest deposition of Fab1 was in the liver and spleen with lower accumulation observed in the lung and heart (Fig. 4C). For Fab2, the lowest distribution was found in the heart with similar distributions of Fab2 in lung, liver, and spleen (Fig. 4D). These results suggest that the physicochemical properties of each Fab could influence its organ distributions.

Effects of delayed administration of HAVN1 after mAb delivery

In this study, we evaluated the impact of delayed administration of HAVN1 after mAb2 delivery on its brain deposition in mice. The goal was to investigate whether HAVN1 could be used for multiple BBB modulation to improve the accumulation of mAb in the brain. Because the half-life of human mAb in mice is around 3 days, it was expected that sufficient mAb plasma concentrations at 24- or 72-h time points for enhancing mAb2 brain delivery by HAVN1. Therefore, the mice were initially injected with mAb2 followed by administration of HAVN1 at 24- or 72-h time point. After HAVN1 in circulation for 15 min, the mAb2 brain deposition was determined. As shown previously, a co-administration of mAb2 + HAVN1 showed significantly higher mAb2 brain deposition (Figs. 5A–B) compared to that of mAb2 alone. A significant enhancement in brain concentration (Fig. 5A) was also observed following a 24-h delay of HAVN1 administration after mAb. Furthermore, the magnitude of brain concentration enhancement of 24-delayed dosing of HAVN1 was lower than that of the co-administration arm; this is due to (a) the increase in brain deposition of mAb2 over 24 h when delivered without HAVN1 (Figs. 5A–B) and (b) the decrease in mAb2 plasma concentration after 24 h (Fig. 5C). Although a higher average brain concentration was observed for the HAVN1 group following a 72-hour delay, statistical significance was not reached compared to PBS control group at 72-h injection (Fig. 5A). The mAb brain concentration was higher at 24-h post-administration of HAVN1 compared to 72-h HAVN1 post administration. It is interesting to see that the deposition of mAb2 when delivered alone was increased at 24-h time point compared to non-delayed control (15 min time point); however, there was no further increased at 72-h time point. This suggests that after maximum accumulation of mAb in the brain at 24-h time point, the mAb begin to clear out from the brain as lower detection of mAb found in the brain at 72-h time point. These results are consistent with those reported by Chang et al., who also observed peak mAb brain concentrations at 24 h using microdialysis.⁴⁰

Figure 5.

Effect of HAVN1 peptide on the mAb2 brain deposition in mice following 24- and 72-hour delay times between initial mAb2 administration (50 nmol/kg) and HAVN1 (18 μmol/kg) or PBS vehicle injection compared to a mixture of mAb2 + HAVN1. Animals were euthanized 15 minutes following administration of peptide or PBS vehicle. (A) Whole brain concentrations were determined after HAVN1 or vehicle delay times. The enhancement in brain concentration with HAVN1 was also compared to co-dosing mAb with HAVN1. (B) Blood-to-brain ratio of mAb2 when delivered with HAVN1 peptide in mice following 24- and 72-hour delay times between initial mAb2 administration and HAVN1 or PBS vehicle injection. (C) Plasma samples were collected 10 minutes following HAVN1 or vehicle injection. Tissue mAb concentrations were determined at (D) 24 or (E) 72 hours. *p<0.05; **p<0.01\

Another way to evaluate the effect of HAVN1 was by using brain-to-plasma ratio (Fig. 5B). Using this method, the effects of mAb2 clearance, plasma concentration decrease, and intrinsic mAb2 brain accumulation can be normalized. For both co-dose and 24-h delayed dosing between mAb2 and HAVN1, the increase in mAb2 brain-to-plasma ratios were significant compared to the respective PBS controls (i.e., mAb2 administered alone) (Fig. 5B). As expected, the brain-to-plasma ratio in 72-h delayed administration had no significant enhancement compared to PBS control (Fig. 5B). This is because of there were continuous increased in inherent mAb2 brain-to-plasma ratios from of 15 min to 72 h time points in mAb control groups (Fig. 5B). In addition, the increase in inherent mAb brain accumulation after administration along with the decrease in the mAb plasma concentration caused the small difference between brain-to-plasma ratios of 72-h delayed delivery of HAVN1 and PBS control.

No differences were observed in plasma or peripheral tissue concentrations between HAVN1 and control groups at any time point (Fig. 5C–E). As expected, both groups showed a rapid ~2.5-fold drop in plasma concentrations from 15-min (co-dose) to 24-h time points, followed by a ~2-fold decrease from 24-h to 72-h time points. The lower plasma concentrations at the 24-h and 72-h time points could partly explain the less pronounced effects of HAVN1 on mAb delivery to the brain at these two time points (Fig. 5A). The patterns of mAb distribution in heart, lung, liver, spleen, and kidney were similar at both 24-h and 72-h time points. As expected, the plasma concentrations of mAb influenced the amount of mAb deposition in the brain when BBB junctions are opened with HAVN1.

Effects of mAb doses in brain deposition

To further investigate how circulating plasma mAb concentration influences mAb delivery with HAVN1, the mAb2 dose was increased from 50 to 130 nmol/kg and administered intravenously in mice with or without peptide. After administration of mAb2 + HAVN1 or mAb2 alone, animals were euthanized after a 15-min circulation time. Significant increases in brain concentrations compared to control were observed at both 130 nmol/kg mAb (p=0.049) and 50 nmol/kg mAb (p=0.0095) administration doses when delivered along with HAVN1 peptide (Fig. 6A). However, there was no difference in the brain depositions of mAb when delivered at 50 to 130 nmol/kg doses in the presence of HAVN1. This suggests that the pores with large sizes created by HAVN1 have a limit in allowing a maximum number of mAb2 molecules to cross the intercellular junctions; the result suggests that there is a saturation effect observed at 130 nmol/kg dose in a certain time duration (i.e. 15 min). In control groups, the mAb2 brain deposition when delivered at 130 nmol/kg dose was doubled than that of delivered at 50 nmol/kg dose (Fig. 6A); this is presumably due to the inherent dose dependent transport of mAb2 across the BBB that is paralleled to the increase in plasma concentration at the 130 nmol/kg dose (Fig. 6B). Administration of HAVN1 did not influence the plasma concentrations at both high or low doses of mAb. Overall, these results suggest that there is an optimal dose of mAb for enhancing mAb brain deposition by HAVN1; this is presumably due to the short time opening of large pores in the paracellular pathways created by HAVN1.

Figure 6.

A 2.5-fold increase in mAb2 dosage does not increase mAb brain delivery with HAVN1. Mice dosed with 130 nmol/kg of mAb2 with or without 18 μmol/kg HAVN1 were euthanized after a 15 min circulation time. (A) The whole brain concentrations of animals were compared with and without HAVN1 co-formulation. The effect of this antibody dose escalation was also assessed by comparing the brain concentrations of animals that had received 50 nmol/kg mAb +/− HAVN1. (B) Plasma samples were obtained 10 minutes after antibody administration. *p<0.05; **p<0.01

The effects of co-dosing vs. multiple doses of HAVN1 in mAb brain deposition

After observing that a 24-h delayed delivery of HAVN1 delivery still enhanced mAb brain deposition, a multiple dosing of HANV1 was investigated to maximize brain deposition mAb2. In this study, three different arms were designed to evaluate the HAVN1 multidose strategy (Fig. 7A). In the first arm, mAb2 was administered followed by administration of PBS at 24-h time point; then, 15-min later the mice were sacrificed to determine the mAb brain deposition. Second, the mAb and HAVN1 were co-administered and then PBS was administered 24 h later followed determination of mAb brain deposition 15 min after PBS injection. In the final arm, a combination of mAb + HAVN1 was administered followed by another administration of HAVN1 at 24-h time point; after 15 min in the circulation, the brain deposition of mAb was determined.

Figure 7.

Co-dosing and multiple doses of HAVN1 enhance mAb2 brain deposition after 24 hours. (A) On the study design, all animals were initially dosed with mAb2 (50 nmol/kg) with or without HAVN1 peptide (18 μmol/kg). At 24 h, the animals were given a second injection of either HAVN1 (18 μmol/kg) or vehicle. Plasma and select tissues were harvested 10 and 15 min following the second injection, respectively. (B) Brain, (C) select tissues, and (D) plasma concentrations of each group were determined using fluorescence and presented as percentage of injected dose per gram (%ID/g). *p<0.05; **p<0.01; ***p<0.001;****p<0.0001

The results showed that both single and multiple peptide doses in 2^nd and 3^rd arms showed significantly enhanced mAb2 brain depositions compared to control or 1^st arm (Fig. 7B). Multiple doses of peptide resulted in the most significant improvement in mAb brain concentrations and there was a 2.1-fold enhancement of mAb brain deposition compared to control (1^st arm). In comparison, a single peptide dose showed an increase of 1.7-fold of brain concentration relative to control. Although animals receiving multiple doses of HAVN1 showed a higher average brain concentration, its comparison with a single co-dose administration did not reach statistical significance (p=0.072). These findings indicate that optimizing the dosing of modulator peptide may be a viable strategy to enhance the mAb brain deposition because the long residence time of mAb in the systemic circulation.

Peripheral tissue concentrations were analyzed and compared among the three-arm groups. There was no difference in the deposition of mAb in the lung in all three different dosing arms (Fig. 7C). The HAVN1 multiple dose (3^rd arm) had significantly higher mAb deposition in the heart compared to the single peptide dose (2^nd arm); these heart concentrations values paralleled to the plasma mAb concentration values (Fig. 7D). The control group had the highest plasma concentration while a single peptide dose had the lowest plasma concentration (Fig. 7D); however, there were no statistically significant differences within all three arms. Multiple doses of HAVN1 (3^rd arm) showed significantly higher mAb concentrations in the liver compared to the single dose (2^nd arm) and control (1^st arm); the liver mAb concentration was in the order of 3^rd > 2^nd > 1^st arms. The increased liver accumulation could suggest that HAVN1 could increase liver metabolism of circulating antibody. Interestingly, the peptide had the opposite effect on the spleen than that of in the liver in which the mAb depositions were the highest in the control (1^st arm) compared to one peptide dose (2^nd arm) and two peptide doses of peptide (3^rd arm). There was a decrease in antibody deposition in the spleen upon single and multiple administration of HAVN1. Notably, the single peptide dose group had significantly higher kidney concentrations compared to the other two groups (1^st and 3^rd arms); this suggests that the increased in degradation of mAb in the liver in 3^rd group lowered the kidney deposition in this group. Taken together, these results suggest that HAVN1 may also increase the permeability in some peripheral organs at longer peptide circulation times.

DISCUSSION

HAVN1 peptide is in a group of HAV peptides that binds to the EC1 domain of cadherin to block cis-cadherin-cadherin interactions in a reversible fashion.³⁴ Upon binding to the cis-interaction region of the EC1 domain, HAV peptides inhibit the cis-interactions of the EC1 domain of one cadherin that binds to the EC2 domain of a separate cadherin that is from the same membrane of BBB endothelial cell. Thus, inhibition of the cis-EC1–EC2 domain interactions disrupts the adherens junction to increase the paracellular porosity of the BBB transiently. This enhanced porosity enables molecules to permeate from the bloodstream into the brain. Due to the effects of the cadherin peptides on BBB permeability, potential applications as BBBMs for delivery of a wide variety of drugs, macromolecules and imaging agents have been reported.^27–29 The emerging interest in biologicals, specifically mAbs and Fabs, for brain related pathologies has caued an increased interest in improving the delivery of these macromolecules across the BBB. While other transient BBBD approaches such as osmotic disruption have been reported to increase antibody delivery to the brain, there were adverse reactions noted in some patients.^{18, 41} The use of focused ultrasound and microbubbles (FUS-MB) to transiently increase the mAb BBB permeability has demonstrated safety and tolerability in AD patients.²⁶ However, BBBD using FUS-MB effects only a small portion of the brain and alternative methods for improving the delivery of mAbs and Fabs to brain are needed.

In this study, HAVN1 peptide was used to enhance the permeation of various Fabs and mAbs across the BBB in the in vitro and in vivo systems. In these studies, we used fluorescence via NIRF Dye labeling of macromolecules to determine their whole organ concentrations of animals. IRDye800 was chosen because the low autofluorescence of the dye allows for accurate measurements at the low concentrations expected in the brain. However, IRDy800-labeling of mAbs can change the physicochemical properties of the mAb, resulting in increased clearance and altered biodistribution.⁴² These changes could influence the observed concentrations within whole organs, and subsequently affect the delivery of the mAbs to the brain using HAVN1.

The permeability studies in the BBB microfluidic model was performed in the presence and absence of serum. The rationale for examining permeability in the presence and absence of serum is to account for potential effects of albumin and immunoglobulins present in the serum that could compete for intracellular transport of macromolecules across the BBB. Without HAVN1, the apical (A)-to-basolateral (B) permeability of mAb and Fab were significantly lower in the presence of serum. Interestingly, the presence or absence of serum had no effect on the permeation of mAb and Fab when perfused with HAVN1 in the microfluidics BBB model. While additional experiments are needed, these studies suggest that receptor-mediated transport may contribute to the passage of antibodies and macromolecules across the BBB. The fact that modulation of permeability with HAVN1 was not impacted by the presence of serum is consistent with its effects on paracellular routes of passage across brain endothelial cells.

The observation that HAVN1 enhanced the permeability of FDX-70 kDa in the BBB culture model, but not in PVE model (Fig. 2), suggests that HAVN1 has selectivity for VE-cadherin in brain microvessel endothelial cells over the cadherins found in the peripheral vasculature. It has been suggested previously that the VE-cadherin on the BBB has “E-cadherin-like” properties;^43–46 thus, VE-cadherins of the BBB may have different properties than those of in the peripheral vasculatures. Unlike the peripheral capillary endothelial cells, the BBB endothelial cells are anchored to basal membranes and surrounded by supporting cells such pericytes, astrocytes, interneuron cells.⁴⁷ Thus, the different structure of the BBB compared to peripheral vasculature could also contribute to selectivity of HAVN1 to modulate brain endothelial cell permeability observed in the present study. It should be noted that a similar selectivity for modulation of BBB permeability with HAVN1 was observed in vivo. However, higher doses of HAVN1 did result in enhanced liver accumulation of mAb (Fig. 7C), suggesting that HAVN1 may also influence molecular permeability across the liver. Therefore, in the future, we will investigate the mechanism of selectivity BBBM peptides (i.e., HAVN1, ADTC5) in modulating the BBB over peripheral vasculatures. While additional investigations of the mechanism of selectivity of the cadherin peptides for BBB versus other endothelial and epithelial barriers are warranted, these studies provide evidence for selective BBBM.

In the in vivo system, HAVN1 peptide enhanced the delivery of mAbs and Fabs into the brain when co-administered together or with a delayed time between mAb and HAVN1 (Figs. 2–6). One potential strategy to increase the accumulation of mAb in the brain is by multiple injections of the BBBM to allow more mAb penetration into the brain. This is needed because the number of large pores in the paracellular pathway created by our peptides is limited and short in duration.^{28, 33} Thus, pulsing the BBB with BBBM over time allows more mAb to accumulate in the brain. Our data supports this idea as demonstrated by a higher brain deposition of animals that received two doses of HAVN1 at 0 and 24-h time points after the delivery of the mAb compared to control (Fig. 7B). Increasing mAb brain deposition with a 24-h delay of HAVN1 administration was also possible due to the long circulation time of human mAb2 in mouse (t_1/2 = ~3 days). Thus, it provided a high mAb blood concentration to drive the influx of into the brain by HAVN1 at 24-h time point. One strategy to adjust the dose of mAb in the brain is through designing the BBBM dosing schedule at various timepoints following mAb administration. The longer half-life of mAb in human (~14 days) may allow for higher mAb brain accumulation with BBBM dosing regimens that can be adjusted to achieve a desired therapeutic effect.

In this study, mAb1 exhibited twice the average number of dye molecules per mAb compared to mAb2. It is worth considering that each IRDye800CW molecule carries a net charge of −3, which can impact the hydrophobicity and the potential brain uptake across the negatively charged endothelial cells of the BBB. Additionally, the different dye conjugations between mAbs may partially explain the altered liver distributions. In the future, we will further confirm the mAb brain deposition enhancement by BBBMs using MRI imaging in living animals as an orthogonal method.

Our studies have shown that the increase in paracellular porosities of the BBB by BBBMs depends on the BBBM being used, time duration of paracellular pores opening, and the size of molecule being delivered. Several cyclic peptides have been shown to deliver antibodies to the brain.^{27, 35} Although ADTC5 peptide can deliver 150 kDa mAb, it cannot deliver 220 kDa fibronectin to the brain, indicating that each BBBM has a specific size limit for effectively delivering molecules.³⁵ The time duration of BBB pores opening by BBBM is considered short compared to the disruption by osmotic BBB disruption (BBBD) method using a hypertonic mannitol solution. Two different BBBMs have been shown to have two different durations of time for the BBB pores opening in vivo. For example, linear HAV6 peptide (Ac-SHAVSS-NH₂) has a duration of BBB pore opening less than 1 h for delivering small molecules like IRdye 800cw (MW = 1067.11 Da) and gadopentetic acid (Gd-DTPA; MW = 545.56 Da) as detected by NIRF imaging and MRI, respectively.³³ In contrast, ADTC5 peptide (Cyclo(1,7)Ac-CDTPPVC-NH₂) enhanced brain delivery of Gd-DTPA following a 2-h delay between ADTC5 treatment followed by injection of Gd-DTPA; however, there was no significant enhancement following a 4-h delay as detected by magnetic resonance imaging (MRI).²⁹ Thus, ADTC5 peptide opens the paracellular pathways longer (i.e., 2 h) than HAV6 peptide (i.e., 1 h) for a small molecule such as Gd-DTPA. Finally, ADTC5 enhanced the brain delivery of galbumin (67 kDa) when the delivery of galbumin was delayed for 10 min after ADTC5; in contrast, HAV6 did not enhance galbumin brain delivery when galbumin administration was delayed for 10 min after HAV6. These data suggest that the BBB pores opening by HAV6 peptide for a large molecule (galbumin) was shorter than that of a small molecule (Gd-DTPA). When the administration of galbumin was delayed for 40 min after the ADTC5 peptide, there was no enhancement of galbumin brain deposition compared to the control. The results suggest that the large BBB pores created by ADTC5 last longer (less than 40 min) than those that created by HAV6 (less than 10 min).²⁸ To explain this observation, we propose that BBBMs mechanism of action is by simultaneously creating large, medium, and small pores in the BBB intercellular junctions with different duration of pore openings.^{27, 28, 35} The large pores collapse to medium and small pores followed by the collapse of medium pores to small pores. Finally, the small pores collapse over time until the BBB regains its integrity. This proposed mechanism can explain why the duration of transport of large molecules was shorter than the transport duration of small molecules. In summary, the delivery of molecules by BBBM depends on the BBBM and the duration of pore opening by BBBM.

The activity of BBBM to create paracellular porosities was dose dependent. In vitro, HAVN1 increased the BBB permeability of FDX-70 kDa in dose-dependent manner (Fig 1B). BBBMs such as HAV6 and cHAVc3 (Cyclo(1,6)Ac-CSHAVC-NH₂) enhanced the brain depositions Gd-DTPA in vivo in dose-dependent manner as determined by MRI.^{33, 34} We have found that the formation of cyclic peptides such as HAVN1 and cHAVc3 enhanced the BBB modulatory activity compare to the parent linear peptide (i.e., HAV6).^{27, 34} This supports the idea that the higher the activity of BBBMs in inhibiting cadherin-cadherin interactions in the BBB intercellular junctions will cause larger pore opening and long duration of opening in the BBB. Thus, modulatory activity of BBBMs can be improved by increasing their selectivity and binding affinity to VE-cadherins in the BBB.

Physicochemical properties (i.e., size, hydrophilicity, protein binding) of mAbs and Fabs influence the distribution of mAbs and Fabs throughout the body. These observations suggest that the different physicochemical and binding properties of mAbs may result in different delivery enhancement by HAVN1 peptide. These properties explain why mAb1 and Fab1 exhibit increased accumulation in the liver compared to mAb2 and Fab2. Additionally, physicochemical properties of antibodies can impact passive transport through the BBB into the brain. Comparatively, the delivery of mAbs to the brain was more efficient than that of Fab fragments; a higher dose of Fab than mAb was required to achieve a comparable brain enhancement of Fab by HAVN1. This could be due to the faster renal clearance of Fab compared to mAb. A similar observation was found when the delivery of 13 kDa lysozyme by ADTC5 was compared to 67 kDa albumin where a higher dose of lysozyme than albumin was needed to reach a similar levels of brain depositions.³⁵ It was found that administration of mAb2 + HAVN1 enhanced mAb2 brain concentrations 3-fold compared to mAb2 alone (control). A similar experiment for mAb1 + HAVN1 resulted in 2-fold enhancement compared to that of mAb1 alone.

CONCLUSION

In summary, we demonstrate the ability of our HAVN1 peptide to enhance the transport of antibodies across the BBB in both in vitro and in vivo models. HAVN1 significantly improved the ability of fluorescently labeled macromolecules through human BBB endothelial cells in an in vitro microfluidics model. Following intravenous administration in mice, coformulation of HAVN1 with a variety of mAbs and Fabs of varying physicochemical properties resulted in increased brain concentrations while having minimal effect on peripheral organs. We have shown that administering a single HAVN1 dose within 24 h of mAb injection can effectively enhance brain concentrations of the antibody and that multiple HAVN1 doses within this time period further enhances brain concentrations. Moreover, our results indicate that multiple doses of HAVN1 within this time window further amplify the brain concentrations of the antibody. In the future, we plan to optimize mAb/HAVN1 dosing schedules with aims to maximize mAb brain deposition. Additionally, we plan to assess the mAb brain kinetics following peptide delivery in vivo using longitudinal MRI imaging.

Abbreviations:

A

Apical

AD

Alzheimer’s disease

ADTC5

Ala-Asp-Thr Cyclic 5

anti-Aβ

Anti amyloid beta

B

Basolateral

BBB

Blood-brain barrier

BBBD

BBB disruption

BBBM

Blood-brain barrier modulator

BDNF

Brain-derived neurotrophic factor

CNS

Central nervous System

ECM

extracellular matrix

Fab

Fab fragment of antibody

FcRN

Fc neonatal receptor

FDX-70

Fluorescein dextran 70 kDa

FUS-MB

Focused ultrasound microbubbles

Gd-DTPA

Gadolinium diethylenetriaminepentacetate

HAVN1

His-Ala-Val N1

IV

Intravenous

PEG

Polyethylene glycol

mAb

Monoclonal antibody

MS

Multiple sclerosis

P_app

Apparent permeability

PVE

pulmonary vascular endothelial