Viral gene therapeutic strategies to obtain abluminal protein secretion from brain endothelial cells denoting the blood-brain barrier

doi:10.21203/rs.3.rs-4943914/v1

Download PDF

Article

Viral gene therapeutic strategies to obtain abluminal protein secretion from brain endothelial cells denoting the blood-brain barrier

https://doi.org/10.21203/rs.3.rs-4943914/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The blood-brain barrier (BBB) formed by brain endothelial cells (BECs) limits the passage of biopharmaceuticals from the circulation into the brain parenchyma. An emerging strategy suggests that viral vectors may specifically target and productively transduce BECs leading to protein secretion into the brain. To study protein secretion, different AAV-BR1 vector constructs encoding mCherry were designed containing either an endothelial specific promotor or the more ubiquitous CAG promotor. Intending to enhance abluminal secretion, a fragment of platelet-derived growth factor subunit B (PDGF-B) protein was fused to the N-terminus of mCherry. Post-translational removal of PDGF-B using a furin cleavage site was also examined. Secretion of mCherry from BECs occurred primarily without post-translational processing, hence limiting the applicability of PDGF-B fusion to proteins among which bioactivity is not abolished by N-terminal modifications. Modelling the BBB in vitro using primary BECs co-cultured with glial cells, mCherry secretion across the abluminal membrane was particularly favorable, following transduction with constructs containing the CAG promoter and PDGF-B fragment. Systemically administered AAV-BR1 vector transduced BECs in vivo, but tracing mCherry secretion into the brain was complicated by additional BBB crossing of the vector with neuronal off-target transduction. In conclusion, the AAV-BR1 vectors transduce BECs and enable abluminal protein secretion with potential for protein delivery to the brain.

Biological sciences/Neuroscience/Neuro–vascular interactions

Biological sciences/Biotechnology/Gene therapy

Brain gene therapy

AAV

polarized secretion

endothelial specific promoter

brain endothelial furin cleavage

Delivery of therapeutics through the blood-brain barrier (BBB) to the central nervous system (CNS) has been a significant bottleneck for treating neurodegenerative disorders (1–3). Brain endothelial cells (BECs), constituting the BBB, are connected by tight junctions, and closely supported by pericytes and astrocyte endfeet.

One approach to bypass the restrictive nature of the BBB involves genetically altering BECs using viral gene therapy, thereby enabling localized production of biologics and CNS-targeted secretion (4, 5). The AAV-BR1 vector, developed through directed evolution of the AAV-2 capsid-modified libraries, is known to primarily target the mouse BECs (mBECs), neurons, and pulmonary endothelial cells following intravenous (IV) injection, omitting transductions of most peripheral organs, including the liver (4, 6).

The strategy of targeting the BBB with gene therapy for the treatment of neurological diseases requires specific transduction of BECs, followed by production and secretion of the therapeutic proteins primarily in the abluminal direction to enable high concentrations in the brain parenchyma. Monitoring therapeutic protein secretion from BECs following gene therapy has however encountered several hurdles. Distinguishing between endogenous and recombinant therapeutic proteins in vivo is difficult, as well as determining whether the proteins are newly produced in response to transduction or have been absorbed after initial secretion (4). Polarized secretion has therefore mainly been examined in vitro, where previous attempts have resulted in primarily luminal secretion following gene therapy (5, 7). Another limitation lies in some potentially therapeutic peptides not having a signal sequence for secretion, such as glucagon-like peptide 1 (GLP-1) while their bioactivity is compromised by N-terminal modifications (8). Furthermore, the processes driving the polarized secretion within endothelial cells and especially BEC remain poorly understood, with 90 % of protens being secreted towards the luminal surface (9–11). Therefore, the capability of transduced BECs to secrete therapeutic proteins to the brain parenchyma remains to be assessed, which requires new vector designs to circumvent these potential pitfalls.

To investigate secretion from BECs in vivo, it is necessary to only transduce BECs, as even moderate additional transduction of neurons will make it difficult to determine the origin of the parenchymal protein. It was recently shown that substitution of the ubiquitous CAG promoter normally used in the AAV-BR1 vector constructs (4, 6), with a synthetic construct consisting of a fragment of the initial enhancer region of cytomegalovirus (CMV) and the occludin promoter; CMVenhancer-occludin (C-Ocln), enhances the vector specificity for mBECs, thereby significantly reducing the number of transduced neurons and astrocytes (12). Although polarized secretion mechanisms are unknown, providing favorable abluminal secretion is important for future therapies. Fusion of a protein of interest to a fragment of a protein normally secreted towards the abluminal membrane, like platelet-derived growth factor subunit B (PDGF-B) (13, 14), could potentially increase the abluminal secretion of the recombinant protein.

The present study aimed to develop and describe strategies for increased abluminal protein secretion in BEC following AAV-BR1-mediated gene therapy. We therefore hypothesized that using an endothelial-specific promoter and fusion of a protein to a PDGF-B fragment would promote the production and subsequent abluminal secretion of the protein from BECs. Furthermore, we hypothesized that the inclusion of a furin cleavage site would allow for the removal of the PDGF-B fusion following entry into a secretory pathway. We validated and investigated the vector design using immortalized cell lines, describing secretion and furin cleavage. The potential of abluminal secretion from BECs was additionally investigated in mice, and although the endothelial specific C-Ocln promoter was used, tracing mCherry in the brain was complicated by off-target transduction of neurons. The vectors were further examined using in vitro BBB models, based on primary cells, and their efficiency was compared to similar vector designs incorporating the CAG promoter. The present study shows that incorporating a PDGF-B fragment combined with a CAG promotor provides favorable abluminal secretion of mCherry.

Strategy for vector design

Two different vector constructs were designed based on the endothelial-specific C-Ocln promotor encoding the non-endogenous fluorescent protein mCherry (Fig. 1A). The first vector construct was designed for intracellular accumulation of mCherry, while the other vector construct was designed for secretion of mCherry. mCherry was therefore N-terminally coupled with a fragment of the PDGF-B protein, including its signal sequence to aim at normal processing of the signal peptide and sorting of mCherry towards an abluminal secretion pathway. Directly downstream of the PDGF-B fragment, a 3XFLAG tag was added for detection purposes, and a furin cleavage site was introduced to test the vector's feasibility for peptides sensitive to N-terminal modifications. mCherry was further integrated with Avitag at the C-terminus, providing a mechanism to determine if the produced protein has undergone C-terminal or N-terminal truncation. Similar vector constructs were designed with the CAG promotor for comparison. Figure 1B, displays the potential pathways for mCherry, depending on the vector design.

Vector generation and virus production

For the generation of plasmids C-Ocln-PDGF-B-3XFLAG-Furin-mCherry-Avitag-WPRE (C-Ocln-PFFmCA), and C-Ocln-mCherry-Avitag-WPRE (C-Ocln-mCA), nucleotide sequences containing the cytomegalovirus (CMV)enhancer-Occludin (C-Ocln) promoter, were combined with the nucleotide sequences of the PDGF-β signal peptide with its subsequent 10 amino acids, 3XFLAG, furin cleavage site, mCherry, and Avitag (Fig. 1A). The plasmid was ordered from GenScript (Piscataway, New Jersey, USA) as full gene synthesis with insertion into a pFB-1 vector. Vectors containing CAG-PDGF-B-3XFLAG-Furin-mCherry-Avitag-WPRE (CAG-PFFmCA), CAG-mCherry-Avitag-WPRE (CAG-mCA) were obtained by exchanging the C-Ocln promoter sequence by the CAG promoter. Packaging of DNA sequences into the viral AAV-BR1 vector was obtained by the Bac-to-Bac expression system as recently described (4).

Cell culture

HEK-293T and bEnd.3 cells (purchased from ATCC maximum passage number #P15) were cultured in T25 flasks with complete Dulbecco's Modified Eagle Medium (cDMEM) containing DMEM, high glucose (ThermoFischer Scientific, cat#11965092) supplemented with 10% Fetal Calf Serum (FCS) (Life Technology, cat#10270), 2mM L-glutamine (ThermoFischer Scientific, cat#25030081) and 5 mg/L gentamicin sulfate, (ThermoFischer Scientific, cat#15750060). For experiments, cells were seeded in 24-well plates.

Isolation of mouse brain endothelial cells (mBECs) and glial cells for culturing and BBB formation in vitro was performed as previously described (15, 16). Two weeks before co-culturing with mBECs, glial cells were, seeded into 12-well plates, and cultured without passaging, with medium change every four days to obtain confluent monolayers. mBECs were cultivated in a tailored mBEC medium (mBEC-medium), containing DMEM-F12 (Life Technology, cat#31331) supplemented with 10% plasma‐derived bovine serum (First Link, cat#60‐00‐810), 10 µg/ml Insulin transferrin sodium selenite (Merck KGaA, cat#11074547001), 10 µg/ml gentamicin sulfate, 1 ng/µl basic fibroblast growth factor (PeproTech Nordic, cat#100‐18B). For co-cultures, mBECs were seeded onto 12-well hanging filter inserts with 1 µm pore membranes (Greiner Bio-One, cat#665610) pre-coated twice with collagen IV and fibronectin at a final concentration of 250 µg/ml and 100 µg/ml respectively, at an estimated density of 100 000 cells/cm2. The subsequent day, mBECs hanging inserts were transferred into a 12‐well plate with glial cells. Tight junction formation was induced by exchanging medium in the top chamber with mBEC-medium supplemented with 250 µM CTP‐cAMP (Merck KGaA, #C3912), 17.5 µM RO (Merck KGaA, #B8279), and 550 nM hydrocortisone (Merck KGaA, #H4001), and the lower chamber medium with glial conditioned medium and mBEC-medium in a 1:1 ratio supplemented with 550 nM hydrocortisone. Daily measurements of barrier integrity were conducted via trans‐endothelial electrical resistance (TEER) measurements with a MiliCell ERS‐2 epithelial volt‐ohm Meter and an STX01 chopstick electrode (Millipore, cat#MERSSTX01). The final TEER values were calculated in Ω × cm² by subtracting TEER readings from cell‐free double-coated hanging filter inserts from co-culture TEER readings and then multiplying the outcome by the filter area (1.12 cm²).

Transfection of immortalized cell lines

HEK 293T and bEnd.3 cells were seeded in 24-well plates and allowed to reach approximately 60–70% confluency in cDMEM. Two hours before transfection, cDMEM was changed to fresh cDMEM. Polyethylenimine, Linear, MW 25000, Transfection Grade (PEI 25K™) (Polysciences, cat#23966-100) was used as transfection reagent in a 3 to 1 weight ratio of plasmid DNA, which was applied at a concentration of 1 ugDNA/one million cells. PEI and plasmid DNA were separately mixed with Opti-MEM (ThermoFischer Scientific, cat#31985062), followed by adding the PEI mixture to the DNA mixture, thoroughly vortexing and incubating at room temperature (RT) for 20 minutes. After incubation, the PEI/DNA-complex in Opti-MEM was added to the cells by carefully pipetting droplets on top of cell media, and carefully moving the 24-well plate in a circular motion. Medium was harvested 72 hours post-transfection and centrifuged at 2000 xg for 5 minutes, with the supernatant collected into a new clean tube. In experiments containing statistical analysis, cells originating from three distinct frozen cell aliquots were transfected on three separate occasions.

Transduction of primary cells

Transduction of the co-cultures was performed three times at different time points and with cultures of primary cells derived from different aliquots. Transduction of mBECs was performed by the addition of 10⁵ vg/cell of either AAV-BR1-C-Ocln-mCA, AAV-BR1-C-Ocln-PFFmCA, AAV-BR1-CAG-mCA or AAV-BR1-CAG-PFFmCA to the hanging filter containing mBECs, one day post tight junction induction. Thereby, the addition of the virus to the upper chamber of the co-culture represents a model of the systemically administered virus with mBECs on hanging filter inserts representing the BBB, and the mixed glial cells representing the brain parenchyma. Four days post-transduction, medium from the upper and lower chamber was harvested and centrifuged at 1000 xg for 5 min and the supernatant was transferred to a new clean tube. The medium was then frozen, for later analysis by ELISA. Cells on the hanging filter inserts and the bottom of the wells were washed twice with PBS and fixed with 4% paraformaldehyde for 5 min and subsequently washed twice in PBS. Fixated cells were stored in PBS at 4°C until immunocytochemical analysis.

Animals for in vivo experiments

36 eight-week-old female C57Bl/6J mice were purchased from Janvier Lab. Upon arrival at the animal facility, the mice were given a minimum of 14 days for acclimatization under local environmental conditions, which included a temperature range of 20–24°C, relative humidity between 40% and 60%, and a 12-hour light/dark cycle. The mice were housed in standard cages. Water and a commercial diet (Altromin 1324, Brogaarden, Gentofte) were provided ad libitum. Animals were monitored for weight changes every second week and daily for well-being. Humane endpoints consisted of a weight loss exceeding 20%, or indications of deteriorating health, that would result in the mice being humanely euthanized. No mice were excluded from the experiment before the final termination date. Animal experiments were conducted at Aalborg University following the Danish experimental permit 2018-15-0201-01467.

Experimental design for in vivo transduction and sample extraction

The mice were assigned randomly to three different groups with no blinding and i.v. tail injections were administered with either PBS (n = 8), AAV-BR1-C-Ocln-mCA (n = 14), or AAV-BR1-C-Ocln-PFFmCA (n = 14). Each mouse was given a virus dose of 1∙10¹¹ vg in a total volume of 100 µl. Animals were monitored for 30 min. post-injection for signs of hypersensitivity. After 28 days the mice were anesthetized via subcutaneous injections of Hypnorm/Dormicum (0.063 mg fentanyl; 2 mg fluanison; 0.1 mg Dormicum). Four mice from the PBS control group and six from each of the other groups had cerebrospinal fluid (CSF) collected by exposing the cisterna magna through an incision, followed by the insertion of a glass capillary to extract CSF. Blood samples were collected intracardially from all mice using a heparin-flushed syringe, immediately followed by transcardiac perfusion with 20 ml of potassium PBS (PPBS). The collected blood samples were kept chilled until centrifugation at 2000 xg for 10 min at 4°C, after which the plasma was collected. Two mice from the PBS control and three from each experimental group underwent transcardiac perfusion with 20 ml of 10% formaldehyde after perfusion with PPBS. Subsequently, brains were extracted and stored in 10% formaldehyde overnight. The fixated brain tissue was then washed in PPBS followed by submersion in a 30% sucrose solution for three days before sectioning into 40 µm coronal sections at -21°C in a cryostat (Leica CM3050 S, Germany), after embedding with Tissue-Tek® O.C.T.™ Compound (Sakura, cat# 4583). Brain sections were distributed into six equivalent sets and stored in anti-freeze solution at -20°C. The remaining brains not perfused with formaldehyde, were divided into the left and right hemispheres. The right hemisphere was snap-frozen on dry ice and the left hemisphere was used for capillary depletion.

Capillary depletion

The left half was placed in a dounce homogenizer with 525 µl ice-cold capillary depletion buffer (0.01M HEPES, 0.14M NaCl, 0.004M KCL, 0.0028M CaCl, 0.001M MgSO₄, 0.001M NaH₂PO₄, 0.01M D-glucose). This process was succeeded by six strokes for homogenization, after which 1000 µl of ice-cold 30% dextran solution was added, followed by six additional strokes. A 200 µl aliquot of the homogenate was immediately snap-frozen for RT-qPCR analysis, and the remainder was kept at 4°C until subsequent steps. To differentiate capillary-enriched tissue from capillary-depleted brain tissue, the homogenate samples were centrifuged at 5400 xg for 40 min at 4°C with acceleration and deceleration curves adjusted to 9 and 5, respectively. The resultant capillary-enriched and depleted brain tissues were stored at -70°C for further analysis.

Western blotting

Medium from b.End3 and HEK 293T cells transiently transfected with PEI and plasmids of interest were subjected to western blotting. The medium was transferred to a solution with 4X LDS Sample buffer (Invitrogen, #cat NP007) and 10X Sample Reducing Agent (Invitrogen, #cat NP005), followed by 5 min of denaturation at 95°C. Samples and Chameleon Duo Prestained Protein Ladder (Li-Cor, cat# 928–6000) were pipetted into a 4–12% Bis-Tris gel (Invitrogen, #cat NP0323), and subjected to 120 V for 90 minutes. Transfer of protein from gel to membrane was conducted in NuPAGE transfer buffer (Invitrogen, #cat NP0006), at 6 V for 16 hours. The membranes were blocked in a 5% milk solution in TBS-T for 1 hour, followed by the addition of the primary antibodies rabbit anti-Avitag 1:1000 (Genscript, cat# A00674) or rabbit anti-FLAG 1:1000 (Merck KGaA, cat#F7425-2MG) in a 1% milk solution in tris-buffered saline Tween® 20 (TBS-T) at 4°C overnight with on a rocking table. The membranes were then washed three times with TBS-T followed by the addition of the secondary antibody anti-rabbit HRP conjugated (Santa Cruz, cat# sc-2004) in 1% milk in TBS-T for 1 hour at RT. After three consecutive washes with TBS-T, membranes were subjected to Pierce ECL western blotting substrate (ThermoFischer Scientific, cat#32109) for five min and imaged using a LI-COR Odyssey Fc. Membranes containing medium samples from cells transfected with plasmids containing a CAG and C-Ocln promoter were measured for chemiluminescence for 3 min and 50 min respectively.

Enzyme-linked immunosorbent assay

All ELISA measurements in this study were conducted using mCherry ELISA KIT (Abcam, cat# ab221829) following the manufacturer’s protocol. In short, all samples were analyzed in duplicates, and cell culture medium and extraction reagents without contact with cells were also measured for standardization of sample absorbance. The standards contained mCherry ranged from 6.25 to 400 pg/ml, and absorbance was measured at 450 nm. Duplicate measurements varying above 0.4 in absorbance were excluded from further analysis. Concentration in the samples was determined by first normalizing the absorbance to the corresponding medium and then applying the linear regression from the standardized samples. To remain in the quantitative range of the kit, medium from transiently transfected cultures was diluted 1:50 and 1:5 for medium from transfections performed with a plasmid containing the CAG and C-Ocln promoter respectively. Medium from the co-culture transduced with AAV-BR1-CAG-PFFmCA was diluted 1:10 to remain in the quantitative range. For measurements conducted on intracellular content of mCherry in capillary depleted and enriched tissue, the samples were lysed following manufacturer protocol using Neuronal Protein Extraction Reagent (N-PER) (ThermoFisher Scientific, cat# 87792). For standardization of ELISA measurements from the lysed samples, total protein concentration was measured following the manufacturer’s protocol using Pierce BCA Protein Assay Kit (ThermoFisher Scientific, cat# 23225).

Immunohistochemistry

Free-floating cryosectioned brain slices were incubated in a blocking buffer containing 3% porcine serum and 0.3% Triton-X100, diluted in KPBS for one hour at RT. Subsequently, the brain slices were incubated with the primary antibody rabbit anti-mCherry (1:500, Rockland, cat# 600-401-P16), with either rat anti-CD11b (1:200, Bio-Rad, cat#MCA711G), mouse anti-NeuN (1:500, Chemicon, cat# MAB377) or rat anti-mouse transferrin receptor (TfR) (1:100, Fischer Scientific, cat# 15227067), diluted in blocking solution and left overnight at 4°C with gentle agitation. The following day slices were washed three times in a washing buffer (blocking buffer diluted 1:50 in KPBS). Slices were then incubated with the secondary antibodies anti-mouse Alexa Fluor 488 (1:200, Invitrogen, cat# A21202), anti-rat Alexa Fluor 488 (1:200, Invitrogen, cat# A11006) and/or biotinylated goat anti-rabbit (1:200, Vector, cat# BA-1000) diluted in blocking buffer and incubated at RT for one hour with gentle agitation. Sections stained with a biotinylated conjugate were incubated with a solution containing 2 drops of reagent A and B of a Vectastain kit in 10 ml KPBS (Vector, cat# PK-6103) for 30 min and washed three times in KPBS, then incubated with Tyramide (1:50, Perkin Elmer, cat# NEL 700A001KT), followed by three washes in KPBS and another round of Vectastain incubation for 30 min and three washes in KPBS. These sections were then incubated with streptavidin Alexa Fluor 568 (1:200, ThermoFisher Scientific, cat# S11226) at RT for one hour with gentle agitation. All sections were then washed twice in washing buffer followed by two wash steps with KPBS. The slices were then incubated with 2 µg/mL 4′,6-diamino‐2‐phenylindole (DAPI) in KPBS for 5 min. After a final wash in KPBS, sections were mounted on microscope glass slides with fluorescent mounting medium. Bioimaging was conducted on an Olympus IX83 inverted microscope with a Yokogawa CSU-W1 spinning disk unit or a confocal laser scanning microscope ZEISS LSM900 (Carl Zeiss).

Immunocytochemistry

Primary mBECs and mixed glial cells were incubated with blocking buffer containing 5% milk, 0,3% Triton-X100 (Merck KGaA, cat# X100), and diluted in PBS for 1 hour at RT. Next, primary antibodies diluted in blocking buffer were added and incubated at 4°C overnight with gentle rocking. mBECs were labeled with rabbit anti-Avitag 1:300 (Genscript, cat# A00674) and mouse anti-ZO-1 1:200 (Invitrogen, cat# 339100), while mixed glial cells were labeled with rabbit anti-Avitag 1:300 and rat anti-CD11b 1:200 (Bio-Rad, cat# MCA711G). Subsequently, cells were washed three times in PBS and secondary antibodies diluted 1:500 in blocking buffer was added for 1 hour at RT while rocking. The secondary antibodies used were anti-rabbit Alexa Fluor 568 (Invitrogen, Cat# A11011) anti-rat Alexa Fluor 488 (Invitrogen, cat# A11006), anti-mouse Alexa Fluor 488 (Invitrogen, cat# A21202). Finally, cell nuclei were stained with DAPI, at a concentration of 2 µg/ml in PBS and mounted on microscopy slides with fluorescent mounting medium (DAKO, #S3023). Bioimaging was conducted on a confocal laser scanning microscope ZEISS LSM900 (Carl Zeiss).

Bioimage analysis

All image processing was conducted in FIJI (18). Image contrasting was adjusted to minimize unspecific noise. Z-stacks were all compiled by maximum intensity projections.

Reverse transcription-quantitative polymerase chain reaction

RNA was extracted from 30mg brain homogenate from each mouse using the GeneJET RNA purification kit (ThermoFisher Scientific, cat# K0732) following the manufacturer's protocol for extraction from mammalian tissue. Following RNA extraction samples were subjected to treatment with DNase I (ThermoFisher Scientific, cat# EN0525) for 30 min at 37°C, with the reaction being terminated by the addition of EDTA and incubation at 65°C for 10 minutes. The DNase-treated RNA samples were used as templates for cDNA synthesis using the Maxima H Minus First-strand cDNA Synthesis Kit (ThermoFisher Scientific, cat# K1651) following the manufacturer's instructions. The qPCR reaction was conducted using the FastStart Essential DNA Green Master (Roche, cat#06402712001), based on SYBR green, and primers targeting mCherry (Fw; 5’-CCA AGC TGA AGG TGA CCA AG-3’, Rev; 5’-GTC CTC GAA GTT CAT CAC GC-3’) actin beta (Actb) (Fw; 5’-CTG TCG AGT CGC GTC CAC C-3’, Rev; 5’-TCG TCA TCC ATG GCG AAC TGG-3’) and hypoxanthine phosphoribosyltransferase 1 (Hprt1) Fw; 5’- GTT GGA TAC AGG CCA GAC TTT GTT G’, Rev; 5’- GAT TCA ACT TGC GCT CAT CTT AGG C-3’). Reactions were performed on 2 ng cDNA with 3 µM primer concentration. Primer efficiency was calculated on data from a tenfold serial dilution of samples or standards, using the equation E = 10(-1/slope). The thermal profile of the analysis was 10 min at 95°C followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. The melt curve thermal profile consisted of 95°C for 15 seconds, descending to 60°C for 1 minute with 1.6°C/s, and ascending back to 95°C with 0.05°C/s. All measurements were conducted in QuantStudio 6 Flex (Applied Biosystems, cat# 4485699). Samples were analyzed in triplicates for each primer, and each sample had a DNA contamination control, consisting of samples not subjected to reverse transcription. If a sample had either incoherent triplicate measurements (> 0.5 in threshold cycle), a signal from the contamination control, or abnormal melt curve peaks, samples would be excluded from further analysis. The geometric mean of the reference genes Actb and Hrpt1 was used to normalize all qPCR data, and the relative mRNA expression was calculated as previously described (17).

Statistics

All statistics in this study except power analysis were conducted using GraphPad Prism 9.5.1 (GraphPad Software, Boston, Massachusetts USA, www.graphpad.com). All datasets were analyzed for normality using a Shapiro-Wilk test and visual analysis of a QQ plot. Normally distributed data was presented with mean and standard deviation values, while non-parametric data was presented with median and 95% confidence interval (CI). Statistical analysis for a given dataset is described in figure legends. For in vivo experiments n-values equate to mice used. The sample size was determined from a power analysis conducted on data from three mice, prospecting minimum use of 6 mice per group for acquiring 80% power in statistical analysis of protein quantities in capillary and brain tissue. This analysis was conducted in IBM SPSS Statistics (Version 28). For in vitro experiments, a n-value equates to one well, and each experiment was conducted three times. Significance is set at p-values of *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 and ****P ≤ 0.0001.

To investigate the capability of BECs to secrete recombinant proteins into the brain parenchyma, two plasmid constructs were designed to encode the fluorescent protein mCherry with and without the initial segment of the PDGF-β sequence potentially allowing for abluminal protein secretion under the control of the endothelial-specific C-Ocln promoter. In addition, a furin cleavage site was included to assess the efficiency of N-terminal sequence cleavage, also indicating sorting into a secretory pathway, as furin cleavage normally takes place in the Golgi apparatus. The 3XFLAG tag included upstream of the furin cleavage site and should only be present in uncleaved mCherry proteins (Fig. 1A). The combination of the PDGF-B sequence, 3XFLAG, and Furin cleavage is referred to as (PFF). To compare our results to previous studies, analog constructs were designed replacing the endothelial-specific C-Ocln promoter by the ubiquitous CAG promotor. The expected mechanisms and locations of signal sequence recognition, furin cleavage, and secretion with resulting protein sizes of 38 kDa for the non-cleaved mCherry and 30 kDa c for the cleaved mCherry are depicted in Fig. 1B.

The presence of the PFF sequence leads to the secretion and furin processing of mCherry.

HEK 293T and bEnd.3 cells were transiently transfected with the different constructs, to confirm their function and to evaluate the secretion pathway of mCherry, and its sensitivity to the furin cleavage mechanisms. Medium from the transfected cells was subsequently collected and analyzed by western blotting for detection of secreted mCherry. The resulting protein concentrations after transduction of HEK 293T and bEnd.3 cells were very different, making it difficult to simultaneously visualize the secreted proteins in the medium from both cell types (Fig. 2A). The inclusion of the PFF sequence results in a distinct cellular processing of mCherry. FLAG-staining of medium from HEK293T revealed a 38 kDa band corresponding to the complete protein product of PFFmCA, which has not been subjected to furin cleavage (data not shown). Avitag-staining revealed three bands, at approximately 38, 30, and below 25 kDa, in medium from both HEK293T and bEND.3 cells transfected with the CAG-PFFmCA or C-Ocln-PFFmCA constructs (Fig. 2A). These correspond to the complete non-cleaved mCherry produced (38kDa), and furin-cleaved mCherry, where the FLAG tag is removed (30kDa), suggesting that both cell types can recognize the furin cleave site. The band below 25 kDa is an unknown N-terminally truncated cleavage fragment of mCherry, as suggested by the retained C-terminal Avitag (Fig. 2A). When the cells were transfected with vectors not containing the PFF sequence, it was expected that mCherry would be contained intracellularly and therefore not be detected in the medium (Fig. 1B). However, a 30 kDa band was detected in medium from cells transfected with both non-secretory vectors (CAG-mCA or C-Ocln-mCA), suggesting that byproducts of intracellular proteins are released into the medium during cell lysis or the existence of a non-specific secretory mechanism, possibly due to high intracellular concentrations of mCherry especially in the HEK293T cells. Transfection with CAG-Luc (encoding Luciferase) was included as a negative control for mCherry-Avitag staining.

Furin cleavage efficiency is not affected by protein secretion quantity in bEnd.3 cells

HEK 293T cells seem more effective in producing furin-cleaved products than bEnd.3 cells, as indicated by the ratios between the full-length 38 kDa PFFmCA product and the furin-cleaved 30 kDa PFFmCA products (Fig. 2A). With furin cleavge being possibly limited in BECs, we wanted to investigate if the ratio of cleavage was affected by the quantity of secretion. Recognizing the high expression in vitro using the CAG promoter, we wanted to determine if the higher secretion would lead to lower cleavage ratio. bEnd.3 cells were transiently transfected with either CAG-PFFmCA or C-Ocln-PFFmCA. The medium was collected, and total mCherry protein concentration was measured using ELISA (Fig. 2B). The ratio of mCherry cleavage products was analyzed using western blot with Avitag-antibodies (Fig. 2C, D). Medium collected from bEnd.3 transfected with CAG-PFFmCA resulted in significantly higher mCherry concentrations compared to the medium collected from bEnd.3 transfected with C-Ocln-PFFmCA with concentrations of 10301 ± 397 and 531 ± 39 pg/ml, respectively (Fig. 2B). When analyzing the ratios of the full-length and the furin-cleaved mCherry-Avitag protein products there was no statistical difference identified between the two promotor types (Fig. 2C). This indicates that the likelihood of furin cleavage in BECs is not affected by the quantity of the protein of interest being secreted.

Transduction of mice with AAV-BR1-C-Ocln-mCA and AAV-BR1-C-Ocln-PFFmCA leads to mCherry expression in both BECs and neurons.

To investigate mBEC protein secretion into the brain in vivo, the brain-specific AAV-BR1 viral vector was incorporated with the C-Ocln-mCA and C-Ocln-PFFmCA constructs. The C-Ocln promoter was used in vivo due to its higher mBEC specificity (12), which was expected to allow for differentiation between abluminal mCherry secretion originating from mBECs and mCherry expressed locally within the brain parenchyma. High concentrations of mCherry within the brain parenchyma would be expected if using the ubiquitous CAG promotor as this promotor allows for strong gene transcription in transduced neurons and astrocytes (12), potentially obscuring identification of mCherry derived from BEC by abluminal secretion. Mice were injected with AAV-BR1-C-Ocln-mCA or AAV-BR1- C-Ocln-PFFmCA, or PBS (CTRL), and 28 days later the brain, plasma, and CSF were collected.

To investigate the presence and potential differences in the localization of mCherry within the brains of mice transduced with the AAV-BR1 vector encoding mCherry with and without the PFF sequence, we performed immunohistochemical staining of mCherry and the transferrin receptor (highly expressed in BECs) focusing on the cerebral cortex and hippocampal area (Fig. 3A). mBECs were seen to co-localize with mCherry, showing that these cells were successfully transduced by the AAV-BR1 vector. However, comparative to previous observations (12), the transduction efficiency in BECs was lower than anticipated. Furthermore, the stained brain slices presented unspecific scattered dots of mCherry staining, irrespective of whether mice vector or only PBS (Fig. 3A, indicated by grey arrows). The unspecific staining was often seen as small circular dots and interpreted as being lipofuscin granules with a diameter in the range of few microns and hence clear distinguishable from the solid cytosolic labeling of transduced cells. The low specificity of the staining therefore did not allow for assessment of protein secretion. In addition, in the cerebral cortex and especially in the CA3 region of the hippocampus several of the mCherry-positive cells had neuronal morphology, which was confirmed when staining for both mCherry and NeuN (neuronal marker), despite using the supposedly endothelial cell-specific C-Ocln promoter. mCherry-positive neurons could also be detected in the cortex, although less prominently. This indicates that the C-Ocln promoter is not as endothelial specific as first anticipated (Fig. 3B). The images did not reveal any additional signs of mCherry adjacent to cells transduced with AAV-BR1-C-Ocln-PFFmCA compared to mice transduced with AAV-BR1-C-Ocln-mCA (Fig. 3C), providing no clear evidence of secretion or unspecific uptake by surrounding cells. To further investigate luminal and abluminal secretion, mCherry protein concentration was measured by ELISA in plasma and CSF respectively. Despite the high sensitivity of the assay (data not shown), mCherry could neither be detected, in the low volume of CSF extractable from mice, nor in the plasma, mCherry is expected to locate differently when entering the secretion pathway, however, this could not be observed in mCherry-positive mBECs and neurons in the cerebral cortex at higher resolution (Fig. 3C). Again, no differences were observed in the cells transduced by either construct.

Inclusion of the PFF sequence does not influence mCherry mRNA levels in whole brain homogenates but leads to lower protein quantities in brain capillaries as well as brain parenchyma indicating secretion of mCherry.

It was initially expected that the PFF sequence would result in higher quantities of mCherry in the brain parenchyma, as a result of abluminal secretion, compared to mice transduced without the PFF constructs. We therefore compared mRNA in whole brain homogenates as well as protein concentrations in brain capillary-enriched tissue vs. brain capillary-depleted tissue (brain parenchyma homogenates) (Fig. 4A). As expected, the relative expression of mCherry in whole brain homogenate did not show notable variation, with mean ± SD expression in mice injected with AAV-BR1-C-Ocln-mCA or AAV-BR1-C-Ocln-PFFmCA being 1.05 ± 0.34 and 1.28 ± 0.49 respectively (Fig. 4B), indicating equal transduction efficiency and gene expression between the two vectors.

Capillary tissue transduced with AAV-BR1-C-Ocln-mCA and AAV-BR1-C-Ocln-PFFmCA had median and 95% CI of mCherry quantities of 77.1 [41.0, 113.3], and 37.5 [9.5, 65.6] pg/mg total protein respectively (Fig. 4C). The reduction of mCherry protein in capillaries by inclusion of the PFF is expected and indicates secretion of mCherry from capillaries transduced with the AAV-BR1-C-Ocln-PFFmCA. In the brain parenchyma, however, mCherry protein concentrations were likewise significantly higher when mice were transduced with AAV-BR1-C-Ocln-mCA (32.6 ± 15.7 pg/mg total protein) compared to mice transduced with AAV-BR1-C-Ocln-PFFmCA (6.8 ± 4.5 pg/mg total protein) (Fig. 4D). This is unexpected but in line with the observed transduction of neurons seen in stained brain tissue (Fig. 3). Importantly, the lower quantity of mCherry in the brain parenchyma of mice injected with AAV-BR1-C-Ocln-PFFmCA strongly indicates degradation and/or clearance of secreted mCherry from the brain. This is highly important as it attests to the difficulty to prove mCherry secretion.

The PFF sequence does not promote differential microglia activation around mCherry-positive cells.

Although we did not succeed in measuring or visually observing abluminal secretion of mCherry, we wondered whether the extracellular presence of a nonendogenous recombinant protein, like mCherry might have triggered activation of microglia adjacent to transduced cells. Double staining for mCherry and CD11b (microglia marker) was therefore performed in both cortex cerebri and hippocampus (Fig. 5). No obvious differences were observed when comparing mice transduced with vectors with and without the PFF sequence. Microglia interacted with mCherry-positive cells in both conditions not suggesting increased activation by the presence of extracellular mCherry.

In vitro, the inclusion of the PFF sequence results in higher abluminal secretion of mCherry.

The transduction pattern upon using the endothelial-specific C-Ocln promotor deviated from previous studies showing overwhelming transduction of the vasculature (12). Instead, we observed transduction of neurons obscuring the detection of potential abluminal secretion of mCherry by mBECs. The vectors were therefore further investigated using an in vitro BBB model based on primary mBECs and mixed glial cells, where no neurons are present. Assuming some of the observed discrepancy in vivo between mRNA and protein quantity can be attributed to secretion, it remains unanswered whether transduction of mBECs with a vector encoding the PFF sequence contributes to increased abluminal secretion into the brain parenchyma, or if the secretion of mCherry by mBECs is mainly luminal, as previously reported using similar setups without the PFF sequence (5, 7). Since the C-Ocln was less endothelial specific than anticipated, and because comparable studies have been performed using strong ubiquitous promoters such as CMV and CAG (5, 7), we next also incorporated CAG promoter-driven mCherry constructs into the AAV-BR1 vector. Primary mBECs were transduced with the AAV-BR1 vector containing either the C-Ocln-mCA, C-Ocln-PFFmCA, CAG-mCA, or CAG-PFFmCA sequences and compared to non-transduced mBECs (CTRL). Four days post-transduction the polarized secretion of mCherry was investigated by analyzing medium collected from the upper and lower chambers in the in vitro BBB model, representing the luminal (blood) and abluminal (parenchyma) sides, respectively. mBECs were fixated to examine the intracellular sorting of mCherry, and during the experiment, the effect of transduction on the mBECs barrier integrity was examined (Fig. 6A). In general, the barrier integrity of the mBECs was not affected by the addition of AAV-BR1, however, mBECs transduced with vectors containing the CAG promoter had an initial decrease in barrier integrity on day 1, which persisted in cultures transduced with AAV-BR1-CAG-PFFmCA throughout the experiment but were normalized for cultures transduced with AAV-BR1-CAG-mCA (Fig. 4C). In control cultures, no mCherry was detected, but mCherry was observed evenly distributed through the cytosol of mBECs transduced with the vectors, not including the PFF sequence (AAV-BR1-CAG-mCA and AAV-BR1-C-Ocln-mCA), corresponding well with an intracellular distribution of mCherry. When the PFF sequence was included in the vector construct (AAV-BR1-CAG-PFFmCA and AAV-BR1-C-Ocln-PFFmCA) mCherry predominantly localized around the nucleus indicating entry into the secretory pathway and packaging in vesicles (Fig. 4B). More transduced cells were observed when the mBECs were transduced with vectors containing the CAG promotor.

No mCherry was detected in control cultures, while mBECs transduced with AAV-BR1-C-Ocln-mCA, only had low, almost non-detectable luminal secretion in 50% of the culture wells (1.3 pg 95% CI [-1.6, 7.0]), while no abluminal secretion was observed (Fig. 4D). mBECs transduced with AAV-BR1-C-Ocln-PFFmCA revealed that the combination of the PFF sequence and C-Ocln promoter resulted in secretion of mCherry not statistically different between luminal secretion (15.7 pg 95% CI [9.6, 22.7]), and abluminal secretion (7.5 pg 95% CI [1.2, 29.6]) (Fig. 4E), although it provides a more favorable ratio of abluminal secretion than what has been reported before (5, 7). mBECs transduced with the AAV-BR1-CAG-mCA resulted in higher mCherry concentrations in both the luminal (181.0 pg 95% CI [88.7, 210.5]) and abluminal (27.2 pg 95% CI [9.1, 48.5]) chambers than observed with the C-Ocln promotor, with the concentration of mCherry in the luminal chamber being significantly higher (Fig. 4F). This was surprising as this vector do not contain the PFF sequence but could suggest that high concentrations of intracellular mCherry are unspecifically released primarily into the luminal chamber or a result of cell lysis. Transduction with the AAV-BR1-CAG-PFFmCA construct led to almost twice as high mCherry quantities in the abluminal chamber (1010 pg. 95% CI [586.7, 1191.0]) compared to the luminal chamber (593 pg. 95% CI [425.9, 709.3]), although not statistically significant (P = 0.0931) (Fig. 4G). Concluding on the in vitro data shown in Fig. 6, the combination of the CAG promoter and the PFF sequence therefore resulted in an abluminal quantity of mCherry being 100 times higher than abluminal secretion from mBECs transduced with vectors encoding the C-Ocln promotor. Together, these observations indicate that the inclusion of the PFF sequence in the vector construct promotes abluminal secretion, which is amplified when combined with the strong CAG promoter.

By specifically developing and describing strategies for abluminal protein secretion, our study aimed to expand upon prior work using AAV-BR1-directed gene therapy to provide proof-of-concept measurements of abluminal secretion by BECs. BECs-specific transduction was attempted in vivo by including the endothelial-specific C-Ocln promoter in the vector design to investigate mCherry abluminally secreted by BECs, minimizing any contribution of mCherry production by off-target cells located in the parenchyma, as seen using the CAG promotor (12). To our surprise the C-Ocln promoter, was not as endothelial specific as anticipated, resulting in several transduced neurons, hindering a solid verification of abluminally secreted mCherry. Analysis of mice transduced with AAV-BR1 furthermore indicated that including a PFF sequence into the vector design likely led to secretion of mCherry from both BECs and neurons, although the methodology was insufficient to track abluminal secretion. Further studying the potential of the PFF sequence in vitro showed increased abluminal secretion of mCherry in primary mBECs, especially in combination with the CAG promotor. Approximately one-third of secreted protein was processed by the furin protease in BECs, possibly lowering the applicability of BEC-directed gene therapy with peptides that require post-translational targeted cleavage for proper functionality.

Vector design is important for high mCherry expression and abluminal-directed secretion

The CAG promoter provided higher secretion in immortalized cell lines in this study relative to the C-Ocln promoter. In investigations of polarized secretion in AAV-BR1 transducted primary cell co-cultures of mBECs and mixed glia cells, similar observations were made. Medium from cells transduced with AAV-BR1-CAG-PFFmCA had 30 times higher luminal and 100 times higher abluminal concentrations of mCherry than medium from cells transduced with the C-Ocln promotor. Applications requiring high expression can therefore benefit from the use of the CAG promoter. Importantly, incorporating a PFF sequence coupled to the CAG promoter into the vector design resulted in an almost twofold higher abluminal secretion pathway of mCherry in primary mBECs compared to luminal secretion, although not significant (P = 0.093). Of note, transduction with AAV-BR1 encoding the CAG promoter coupled with only the PDGF-B fragment of PFF fused to IL-10 has shown the same tendency in our lab, having significantly higher abluminal secretion (unpublished data). Even though preferential abluminal secretion of a transgene following gene therapy has been described in immortalized cell lines of mBECs (19), this is the first time, that a similar tendency is described in a culture of primary cells, indicating that the inclusion of a PFF sequence, can potentially enhance secretion at the abluminal surface of BECs. Future studies could expand on the use of different signal sequences or fusion proteins’ impact on secretion in mBECs, as has been done in a range of biotherapeutic production schemes e.g., monoclonal antibody production (20, 21), to possibly allow for optimal efficiency in targeted gene therapy of mBECs as therapeutic protein factories with high abluminal secretion. The fused PDGF-B sequence consists of the protein signal sequence and the following 10 amino acids. The fusion of both the signal sequence and some of PDGF-B to mCherry are added to promote correct recognition of the signal sequence by the signal recognition particle, and potential insertion of the protein into the correct polarized secretion pathway in mBECs, and subsequent cleavage by the signal peptidase. However, all the above mechanisms and their influences are unknown, whereof each would require a comprehensive study to be fully described, especially since the polarized secretion of endothelial cells is poorly understood (9–11). It is, therefore, possible that the preferential abluminal secretion observed in mBECs transduced with AAV-BR1-CAG-PFFmCA, could be achievable only using the signal sequence of PDGF-B, and that the cleavage of the sequence would be performed in the correct location when included in a sequence of a N-terminally sensitive protein, therefore not requiring a furin cleavage site at all. As approximately 70 % of secreted mCerry protein is not being cleaved by furin in bEnd.3 cells, these investigations could provide valuable information, as including the PFF sequence together with an N-terminally sensitive peptide might not be an optimal strategy, as it would lead to a high proportion of proteins not having functionality. Conversely, another strategy to investigate abluminal secretion could also include the addition of a longer PDGF-B sequence, potentially increasing abluminal secretion further, which however also holds the risk of disrupting the function of the fused protein completely, by the interaction of PDGF-B with its receptor on pericytes.

BBB integrity following AAV-BR1 gene therapy

While investigating polarized secretion in vitro in primary co-cultures of mBECs and glia cells, a decrease in barrier integrity of cells transduced with AAV-BR1-CAG-PFFmCA was observed. It's therefore important to highlight that TEER values never descended below levels that would lead to a compromised BBB integrity facilitating high passage of even small molecules (15), indicating diffusion of mCherry through the barrier as an unlikely cause of the abluminal mCherry. Furthermore, only 2 % o the total mRNA production from the vector's genetic material during AAV-BR1 transduction is attributable to transduced glia cells in the bottom chamber (4), thereby isolating abluminal secretion by mBECs as the main reason for the presence of mCherry. The underlying mechanisms contributing to reduced barrier integrity however remain elusive. Considering that in this study, the lowered TEER values were only observed in the cultures transduced with AAV-BR1-CAG-PFFmCA, and not previously experienced in cultures transduced with vectors also containing a CAG promoter and a native signal for secretion (4), a potential cause could encompass increased abluminal secretion or heightened furin protease activation, which might disrupt the natural cellular processes of BECs.

Tracking of mCherry secretion in vivo

The strength of the C-Ocln promoter was expected to be seen in vivo where reduction of off-target transduction of neurons and lung tissue was anticipated due to previous reports (12). However, neuronal expression of mCherry was still prominent using the C-Ocln promoter, why protein measurements of mCherry in capillary-depleted brain tissue were significantly higher when mice were transduced with the vector without the PPF sequence. Neurons not primed to secrete mCherry post-transduction likely resulted in an increased content of mCherry in brain parenchymal tissue, despite similar mRNA expression levels of mCherry being measured in whole brain homogenate. It is however important to note that equal levels of mRNA do not necessarily translate into equivalent protein generation, particularly when one of the proteins is destinated to undergo processing through a secretory pathway, but a strong correlation often exists between the two measures of the same gene (22). It is possible that a significant portion of the observed difference in intracellular mCherry content within both capillary-enriched tissue and brain parenchyma (capillary-depleted tissue), can be attributed to secretion. Use of Proximity Extension Assay-based Olink technologies for direct mouse CSF measurements (23), or optimized HPLC-MS methods could circumvent issues related to low volume and CSF protein content (24).

Off-target in vivo expression

Pyramidal neurons have been shown to express occludin (25), which could explain the high off-target expression seen in these cells. Pyramidal neurons are among other places located in the hippocampus, so the high transduction observed in this area could be a result of the C-Ocln promoter not being capable of alleviating off-target specificity in these neurons. Another possibility is that the targeted specificity associated with promoter types is affected by regulatory elements of gene expression functioning differently when incorporated in episomal DNA, rather than chromatin (26). The results of this study indicated that both transduced neurons and mBECs were capable of secretion post-transduction when comparing mRNA expression with intracellular protein content. The in vivo results of AAV-BR1 and C-Ocln promoter specificity in this study, also need to be considered in the context of varying results of cellular AAV-BR1 specificity between laboratories (4, 6, 12). The target receptor of AAV-BR1 remains unknown, however tropism between mice strains for LY6A, a receptor for AAV-PHP.B, has been found to heavily impact brain transduction of other brain-targeted AAV variants (27). Although C57BL/6 mice were also used in the study investigating the C-Ocln promoter (12), factors such as commercial breeding facility, laboratory environment, age at transduction, or sex of the mice could affect the expression of the AAV-BR1 receptor, leading to the observed discrepancies between studies.

In conclusion, this study demonstrates enhanced in vitro secretion of mCherry by mBECs when coupled with the PFF sequence, possibly favoring abluminal secretion when combined with the CAG promoter. Constructs requiring furin cleavage in BEC might not be optimal as only one-third of protein is cleaved. Mice transduced with the AAV-BR1 vector containing the C-Ocln promoter facilitated mCherry expression and secretion in both mBECs and neurons when coupled with the PFF sequence; however, more sensitive methods are required to verify this notion.

Blood-brain barrier (BBB), brain endothelial cells (BECs), CAG-Luciferase (CAG-Luc), CAG-mCherry-Avitag-WPRE (CAG-mCA), CAG-PDGF-B-3XFLAG-Furin-mCherry-Avitag-WPRE (CAG-PFFmCA), central nervous system (CNS), cerebrospinal fluid (CSF), CMVenhancer-occludin (C-Ocln), C-Ocln-mCherry-Avitag-WPRE (C-Ocln-mCA),, C-Ocln-PDGF-B-3XFLAG-Furin-mCherry-Avitag-WPRE (C-Ocln-PFFmCA), complete Dulbecco's Modified Eagle Medium (cDMEM), confidence interval (CI), glucagon-like peptide 1 (GLP-1), intravenous (IV), mouse brain endothelial cells (mBECs), mouse brain endothelial cells medium (mBEC-medium), phosphate buffered saline (PBS), platelet-derived growth factor subunit B (PDGF-B), PDGF-B:3XFLAG:furin (PFF), potassium phosphate buffered saline (KPBS), room temperature (RT), trans‐endothelial electrical resistance (TEER), transferrin receptor (TfR), tris-buffered saline with Tween® 20 (TBS-T), 4′,6‐diamino‐2‐phenylindole (DAPI).

ACKNOWLEDGEMENTS

We want to acknowledge the excellent technical assistance of Merete Fredsgaard, Hanne Krone Nielsen, Louise Bolther, Rikke Louise Larsen, and Helle Christiansen, Aalborg University, Denmark. Furthermore, we want to thank Julia Minert, University Medical Center Hamburg, for cloning the plasmids containing the CAG promoter, and providing guidance in the Bac-to-Bac expression system. Finally we want to thank Helene Halkjær Jensen and Anders Olsen, Aalborg University, for access to the spinning-disc microscopy facility.

CONFLICT OF INTERESTS

JK is listed as an inventor on a patent on AAV-BR1, held by Boehringer Ingelheim International.

The other authors declare no conflict of interests

FUNDING

The present study has been supported by the Independent Research Fund Denmark - Medical Sciences (Grant no. 1030-00474B), The Multiple Sclerosis Society (Grant no. R588-A40233 to T Moos), the Lundbeck Foundation Research Initiative on Brain Barriers and Drug Delivery (Grant no. R155-2013-14113 to T. Moos), Svend Andersen Fonden and Fonden for Lægevidenskabens Fremme.

AUTHOR CONTRIBUTIONS

Conceptualization: BL, TM, Funding acquisition: AB, MST, TM, Investigation: BL, Supervision: AB, JK, MST, TM, Visualization: BL, Writing – original draft: BL, Writing – review and editing: BL, AB, MST, TM

Pardridge WM. The blood-brain barrier: bottleneck in brain drug development. NeuroRx [Internet]. 2005 [cited 2023 Sep 3];2(1):3–14. Available from: https://pubmed.ncbi.nlm.nih.gov/15717053/
Heigener DF, Reck M. Crizotinib. Recent Results in Cancer Research [Internet]. 2018 [cited 2024 Mar 25];211:57–65. Available from: https://link.springer.com/chapter/10.1007/978-3-319-91442-8_4
D’Souza A, Dave KM, Stetler RA, S. Manickam D. Targeting the blood-brain barrier for the delivery of stroke therapies. Adv Drug Deliv Rev. 2021 Apr 1;171:332–51.
Rasmussen CLM, Hede E, Routhe LJ, Körbelin J, Helgudottir SS, Thomsen LB, et al. A novel strategy for delivering Niemann‐Pick type C2 proteins across the blood–brain barrier using the brain endothelial‐specific AAV‐BR1 virus. J Neurochem [Internet]. 2023 Jan 1 [cited 2023 Jul 21];164(1):6. Available from: /pmc/articles/PMC10084444/
Hede E, Christiansen CB, Heegaard CW, Moos T, Burkhart A. Gene therapy to the blood–brain barrier with resulting protein secretion as a strategy for treatment of Niemann Picks type C2 disease. J Neurochem [Internet]. 2021 Feb 1 [cited 2023 Jul 24];156(3):290–308. Available from: https://onlinelibrary.wiley.com/doi/full/10.1111/jnc.14982
Körbelin J, Dogbevia G, Michelfelder S, Ridder DA, Hunger A, Wenzel J, et al. A brain microvasculature endothelial cell-specific viral vector with the potential to treat neurovascular and neurological diseases. EMBO Mol Med [Internet]. 2016 Jun [cited 2023 Jul 19];8(6):609–25. Available from: https://pubmed.ncbi.nlm.nih.gov/27137490/
Burkhart A, Andresen TL, Aigner A, Thomsen LB, Moos T. Transfection of primary brain capillary endothelial cells for protein synthesis and secretion of recombinant erythropoietin: a strategy to enable protein delivery to the brain. Cellular and Molecular Life Sciences [Internet]. 2017 Jul 1 [cited 2023 Jul 21];74(13):2467–85. Available from: https://link.springer.com/article/10.1007/s00018-017-2501-5
Jazayeri A, Rappas M, Brown AJH, Kean J, Errey JC, Robertson NJ, et al. Crystal structure of the GLP-1 receptor bound to a peptide agonist. Nature 2017 546:7657 [Internet]. 2017 May 31 [cited 2023 Sep 4];546(7657):254–8. Available from: https://www.nature.com/articles/nature22800
Wei H, Sundararaman A, Dickson E, Rennie-Campbell L, Cross E, Heesom KJ, et al. Characterization of the polarized endothelial secretome. FASEB Journal. 2019 Nov 1;33(11):12277–87.
Wolpe AG, Ruddiman CA, Hall PJ, Isakson BE. Polarized proteins in endothelium and their contribution to function. J Vasc Res [Internet]. 2021 Mar 1 [cited 2023 Aug 24];58(2):65. Available from: /pmc/articles/PMC7954879/
Worzfeld T, Schwaninger M. Apicobasal polarity of brain endothelial cells. J Cereb Blood Flow Metab [Internet]. 2016 Feb 1 [cited 2024 Jan 26];36(2):340–62. Available from: https://pubmed.ncbi.nlm.nih.gov/26661193/
Graßhoff H, Müller-Fielitz H, Dogbevia GK, Körbelin J, Bannach J, Vahldieck CMG, et al. Short regulatory DNA sequences to target brain endothelial cells for gene therapy. J Cereb Blood Flow Metab [Internet]. 2022 Jan 1 [cited 2023 Jul 19];42(1):104–20. Available from: https://pubmed.ncbi.nlm.nih.gov/34427142/
Gerhardt H, Betsholtz C. Endothelial-pericyte interactions in angiogenesis. Cell Tissue Res [Internet]. 2003 Oct 22 [cited 2023 Jul 19];314(1):15–23. Available from: https://link.springer.com/article/10.1007/s00441-003-0745-x
Zerwes HG, Risau W. Polarized secretion of a platelet-derived growth factor-like chemotactic factor by endothelial cells in vitro. J Cell Biol [Internet]. 1987 [cited 2024 Jan 26];105(5):2037–41. Available from: https://pubmed.ncbi.nlm.nih.gov/3680370/
Thomsen MS, Humle N, Hede E, Moos T, Burkhart A, Thomsen LB. The blood-brain barrier studied in vitro across species. PLoS One [Internet]. 2021 Mar 1 [cited 2023 Jul 28];16(3). Available from: https://pubmed.ncbi.nlm.nih.gov/33711041/
Thomsen MS, Birkelund S, Burkhart A, Stensballe A, Moos T. Synthesis and deposition of basement membrane proteins by primary brain capillary endothelial cells in a murine model of the blood–brain barrier. J Neurochem [Internet]. 2017 Mar 1 [cited 2023 Jul 25];140(5):741–54. Available from: https://onlinelibrary.wiley.com/doi/full/10.1111/jnc.13747
Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res [Internet]. 2001 [cited 2024 Apr 2];29(9):E45. Available from: https://pubmed.ncbi.nlm.nih.gov/11328886/
Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji - an Open Source platform for biological image analysis. Nat Methods [Internet]. 2012 Jul [cited 2023 Aug 2];9(7):676–82. Available from: /pmc/articles/PMC3855844/
Jiang C, Koyabu N, Yonemitsu Y, Shimazoe T, Watanabe S, Naito M, et al. In vivo delivery of glial cell-derived neurotrophic factor across the blood-brain barrier by gene transfer into brain capillary endothelial cells. Hum Gene Ther [Internet]. 2003 Aug 10 [cited 2024 Jan 9];14(12):1181–91. Available from: www.liebertpub.com
Kober L, Zehe C, Bode J. Optimized signal peptides for the development of high expressing CHO cell lines. Biotechnol Bioeng [Internet]. 2013 Apr 1 [cited 2023 Sep 20];110(4):1164–73. Available from: https://onlinelibrary.wiley.com/doi/full/10.1002/bit.24776
Owji H, Nezafat N, Negahdaripour M, Hajiebrahimi A, Ghasemi Y. A comprehensive review of signal peptides: Structure, roles, and applications. Eur J Cell Biol. 2018 Aug 1;97(6):422–41.
Liu Y, Beyer A, Aebersold R. Leading Edge Review On the Dependency of Cellular Protein Levels on mRNA Abundance. 2016 [cited 2023 Jul 21]; Available from: http://dx.doi.org/10.1016/j.cell.2016.03.014
Petrova R, Patil AR, Trinh V, McElroy KE, Bhakta M, Tien J, et al. Disease pathology signatures in a mouse model of Mucopolysaccharidosis type IIIB. Scientific Reports 2023 13:1 [Internet]. 2023 Oct 4 [cited 2024 Jan 28];13(1):1–18. Available from: https://www.nature.com/articles/s41598-023-42431-4
Dislich B, Wohlrab F, Bachhuber T, Müller SA, Kuhn PH, Hogl S, et al. Label-free Quantitative Proteomics of Mouse Cerebrospinal Fluid Detects β-Site APP Cleaving Enzyme (BACE1) Protease Substrates In Vivo. Mol Cell Proteomics [Internet]. 2015 Oct 1 [cited 2024 Jan 28];14(10):2550. Available from: /pmc/articles/PMC4597136/
Morgan S V., Garwood CJ, Jennings L, Simpson JE, Castelli LM, Heath PR, et al. Proteomic and cellular localisation studies suggest non-tight junction cytoplasmic and nuclear roles for occludin in astrocytes. European Journal of Neuroscience [Internet]. 2018 Jun 1 [cited 2023 Sep 19];47(12):1444–56. Available from: https://onlinelibrary.wiley.com/doi/full/10.1111/ejn.13933
Inoue F, Kircher M, Martin B, Cooper GM, Witten DM, McManus MT, et al. A systematic comparison reveals substantial differences in chromosomal versus episomal encoding of enhancer activity. Genome Res [Internet]. 2017 Jan 1 [cited 2023 Sep 19];27(1):38–52. Available from: /pmc/articles/PMC5204343/
Huang Q, Chan KY, Tobey IG, Chan YA, Poterba T, Boutros CL, et al. Delivering genes across the blood-brain barrier: LY6A, a novel cellular receptor for AAV-PHP.B capsids. PLoS One [Internet]. 2019 Nov 1 [cited 2024 Jun 24];14(11). Available from: https://pubmed.ncbi.nlm.nih.gov/31725765/

There is NO conflict of interest to disclose.

Download PDF

Editorial decision: revise
09 Oct, 2024
Review #3 received at journal
28 Sep, 2024
Review #2 received at journal
21 Sep, 2024
Reviewer #3 agreed at journal
06 Sep, 2024
Reviewer #2 agreed at journal
03 Sep, 2024
Reviewer #1 agreed at journal
01 Sep, 2024
Reviewers invited by journal
01 Sep, 2024
Editor assigned by journal
23 Aug, 2024
Submission checks completed at journal
23 Aug, 2024
First submitted to journal
22 Aug, 2024
Unknown event
22 Aug, 2024

You are reading this latest preprint version

Viral gene therapeutic strategies to obtain abluminal protein secretion from brain endothelial cells denoting the blood-brain barrier

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Strategy for vector design

Vector generation and virus production

Cell culture

Transfection of immortalized cell lines

Transduction of primary cells

Animals for in vivo experiments

Experimental design for in vivo transduction and sample extraction

Capillary depletion

Western blotting

Enzyme-linked immunosorbent assay

Immunohistochemistry

Immunocytochemistry

Bioimage analysis

Reverse transcription-quantitative polymerase chain reaction

Statistics

Results

Furin cleavage efficiency is not affected by protein secretion quantity in bEnd.3 cells

Discussion

Vector design is important for high mCherry expression and abluminal-directed secretion

BBB integrity following AAV-BR1 gene therapy

Tracking of mCherry secretion in vivo

Off-target in vivo expression

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1