Design of a HS-binding TfR binding BBB-transporter
The scFv8D3BBB transporter (31 kDa Mw) 14,16 consists of the heavy and light chain variable domains of the TfR binding antibody 8D317 connected to each other by a linker consisting of the amino acids SGSTS (G4S)388. The scFv8D3 and the two new mutant constructs were designed with a twin strep tag (WSHPQFEKGGGSGGGSGGSAWSHPQFEK), which was linked to the C-terminal with a TEV linker (ENLYFQS) followed by an in-house designed linker (APGSGTGSAPG). The constructs were cloned into the expression vector pcDNA3.4 (GeneArt, Thermo Fisher) with our standard signal peptide added to the N-terminal. The structure of scFv8D3 was generated by using the automated protein structure homology-modelling server SWISS-MODEL89 based on the primary amino acid structure of scFv8D3. The generated structure was analyzed in the protein structure software PyMOL90 and HS binding motifs were searched for on the surface of the protein. There were no suitable regions available on the surface of scFv8D3 for the creation of linear HS binding motifs. Instead, two positions in folded regions of scFv8D3 with two adjacent positively charged amino acids were found, allowing for the creation of HS binding motifs by introducing a third positively charged amino acid by point mutation at the correct distance of ~ 5–10 Å, which mimics a distance of one or two amino acid sin the linear HS binding motifs BB-X-B or BB-XX-B31,32. Amino acid present within the variable regions of scFv8D3 were not mutated, to avoide affecting the TfR-binding affinity. Additionally, to prevent introducing protein instability, mutating highly hydrophobic amino acids into basic amino acids, and vice versa, was avoided since hydrophobicity is one of the forces affecting protein folding44,45. Two HS binding motifs were added to scFv8D3 by point mutations, S75K and P176K, creating the HS(+)scFv8D3 mutant. Two potential HS binding motifs at the surface of scFv8D3 were removed by point mutations, K42Q, K43Q, K243Q and R244Q, creating the HS(-)scFv8D3 mutant.
Recombinant expression and purification
All three scFvs were produced as previously described91. In short, human Expi293 cells were transfected with the pcDNA 3.4 vectors containing the constructs by using the transfection reagent polyethylenimine (PEI) (Polysciences 24765-1) and the cell-cycle arrester valproic acid (VPA) (Sigma P4543). Seven days’ post-transfection, the cells were harvested and the supernatant was separated by using Celpure® P300 (Sigma-Aldrich) and filtered through low protein binding 0.22 µm filter units (Corning 430513). The scFvs were purified with an Äkta start liquid chromatography (LC) instrument with Strep-Tactin®XT 4Flow® cartridge (IBA Life Sciences 2-5021-001). The scFvs were eluted with a 50 mM biotin- Tris HCl buffer and then immediately concentrated by Amicon centrifugal filters (Sigma-Aldrich, UFC501024) and buffer exchanged to PBS (Gibco) with Zeba spin columns (ThermoFisher, A44301). The concentrations of the purified proteins were determined by measuring their absorbance at 280 nanometer (nm) and calculating their concentrations by factoring in their molecular extinction coefficients and Mw.
SDS-PAGE and purity analysis
To confirm the size and purity, the purified scFvs were analyzed using SDS-PAGE followed by coomassie blue staining. For the SDS-PAGE, the scFvs were loaded together with NuPAGE™ LDS Sample Buffer (Invitrogen™) and a pre-stained protein marker (ThermoFisher 26616) on NuPAGE™ Bis-Tris 4 to 12% 15-well precasted gels (Invitrogen™) without adding reducing agents, and run at 80 V for approximately 1.5 hours in MES running buffer (NuPAGE™ MES SDS Running Buffer, Thermo Fisher). The gels were then stained with comassie blue staining (PageBlue™ Protein Staining Solution, Thermo Scientific™) for 20 minutes, rinsed with deionized water and photographed using Image Studio software (version 5.2.5). The acquired image was analyzed with Fiji(ImageJ).
Transferrin receptor ELISA
The TfR binding ability of the scFvs was assessed by a previously described TfR ELISA92. Ninety-six wells half area plates (Corning Incorporated) were coated with 50 ng/well recombinant mouse TfR extracellular domain protein in PBS (prepared in our lab) and stored overnight at 4°C. The following day, the plates were blocked with 1% BSA in PBS for two hours at room temperature (RT) while shaking and then incubated with serial dilutions of the scFvs for two hours at RT while shaking. For detection, a mouse-anti Strep-Tag II (IBA Life Sciences) primary antibody and an HRP-conjugated goat anti-mouse secondary antibody (Sigma-Aldrich 12–349) were used. The signal development was done with K-blue aqueous TMB (Neogen Corp) and the absorbance was measured using FLUOstar Omega microplate reader (BMG Labtech) at 450 nm. The scFvs and antibody dilutions were made in ELISA incubation buffer (1x PBS with 0.1% BSA and 0.05% Tween-20). Washing steps were made done with ELISA washing buffer (1x PBS with 0.05% Tween-20) five times.
Thermal shift assay
The stability of the scFvs was investigated by Tycho nt.6 instrument (NanoTemper technologies, München, Germany) as previously described1. In short, 8 microliters of each scFv at 1.2 micromolar concentration were heated in glass capillaries with fluorescence intensities measured at 330 nm and 350 nm during a linear temperature increase from 35°C – 95°C.
Heparin column liquid chromatography
The heparan sulfate binding capacity of the scFvs was assessed by heparin column liquid chromatography with an Äkta start instrument. Prior to loading the scFvs onto the HiTrap Heparin HP 1 mL column (Cytiva), the proteins were buffer exchanged from PBS to 10 mM sodium phosphate buffer (pH 7) using Zeba™ Spin Desalting, 7KColumns (ThermoFisher Scientific) to prevent disruption of binding to the heparin column caused by the salts in the PBS buffer. The scFvs were loaded at a rate of 0.5 mL/min. After loading of the scFvs, the column was washed with 10 mL of binding buffer and the wash was collected. The elution (measured by monitoring absorbance at 280 nM) was performed with a linear gradient of elution buffer (measured by conductivity, mS/cm) over 15 mL of 0–100% elution buffer (0.5 M NaCl in 10 mM sodium phosphate buffer) at a rate of 1 mL/min and collected.
After elution, the fractions were analyzed by SDS-PAGE followed by Coomassie staining.
Animals
For all animal experiments, 3–4 months old (C57BL/6JBomTac, ordered from Taconic) wild-type male and female mice were used. The animals were housed in individually ventilated cages (4–5 animals/cage) in rooms with controlled temperature (20–22°C) and controlled humidity (50–55%). The mice were feed ad libitum with free access to water and had daily surveillance by trained personnel in an animal facility at Uppsala University. All procedures described in this paper were approved by the Uppsala County Animal Ethics board 5.8.18–20401/20 and performed according to the ARRIVE guide lines. The rules of and regulations of the Swedish Animal Welfare Agency, as well as the of European Communities Council Directive of 22 September 2010 (2010/63/E.U. ), were followed during the animal studies and all efforts were made to minimize animal suffering and to reduce the number of animals used.
Radiolabeling with iodine-125
The scFvs variants were labelled with iodine-125 (125I) using Chloramine-T as described previously93. In short, 400 nM of the three scFvs were mixed with 125I stock solution (Perkin Elmer Inc, Waltham) and 1 mg/mL of Chloramine-T (Sigma Aldrich) in PBS. After 90 seconds of incubation, the reaction was stopped with 1 mg/mL sodium meta-bisulphite (Sigma 08982). The radio-labeled scFvs were then purified from unbound and free 125I by using NAP columns (VWR 17-0853-02) and eluted in PBS. The radiolabeling procedures were done 1–2 hours before the ex vivo studies. The yield of the labeling process varied between the different studies, due to inherent variation of the method, however this was accounted for in the dose calculations. The yield was approximately 22.8 MBq/nmol for the in vivo 96-hour plasma stability study, 35 MBq/nmol for the 2-hour and 48-hour ex vivo studies, approximately 12 MBq/nmol for the NTE/CD ex vivo study (also 2-hour) and approximately 3.5 MBq/nmol for the 24-hour therapeutic ex vivo study.
In vivo 96-hour plasma stability study
Plasma stability of scFv8D3, HS(+)scFv8D3 and HS(-)scFv8D3 mutant were investigated in C57Bl/6 wild-type mice (3–4 months old). The mice were injected with a tracer dose (0.3 nmol/kg, 0.04 mg/kg) of each scFv. Mice (total n = 15, 5 mice for each protein) were intravenously injected via the tail vein with 1.10 ± 0.04 MBq [125I]scFv8D3, 1.06 ± 0.06 MBq [125I]HS(-)scFv8D3 or 1.24 ± 0.10 MBq [125I]HS(+)scFv8D3. Approximately thirty microliters blood samples were collected were obtained from the tail vein at 3-, 6-, 24-, 48- and 96-hours post-injection. The blood samples were then centrifuged at 15.000 x g for 5 min to obtain plasma and blood pellet samples. The mice were euthanized at 96-hours post-injection by transcardial perfusion with 0.9% physiological saline under terminal anesthesia. The brains were dissected and divided into two hemispheres. The right hemisphere was left intact while the left hemisphere was divided in cerebrum and cerebellum. Liver, spleen, heart, lung, kidney, pancreas, muscle, bone, skull, and thyroid were also isolated. The radioactivity in the brain, peripheral organs, blood and urine was measured using a γ-counter (1480 Wizard, Wallac Oy, Turku, Finland). Plasma and urine were analyzed using thin-layer chromatography (TLC) for analysing the ratio of [125I]-labelled scFvs vs free 125I. Briefly, the bottom of a glass jar was filled with 70% acetone. Urine samples (~ 2 µl) were applied at a baseline on a piece of silica coated aluminum plate and allowed to dry for approximately 5 min before adding the TLC-plate to the solvent-containing glass jar, ensuring the solvent line was below the sample-baseline. When the solvent front had migrated two-thirds the way up the TLC-plate, it was removed from the glass container, allowed to dry for 15 minutes and developed underneath an X-ray film for 48-hours in complete darkness. The X-ray film was then measured using a Cyclone Phosphoimager and the images obtained analyzed by ImageJ46.
Ex-vivo 2-hour biodistribution study
Brain uptake and peripheral biodistribution of scFv8D3, HS(+)scFv8D3 and HS(-)scFv8D3were investigated in C57Bl/6 wild-type mice (3 months old). The mice were injected with a tracer dose (0.3 nmol/kg, 0.04 mg/kg) of each scFv. Mice (total n = 9, 3 mice for each protein) were intravenously injected via the tail vein with 1.71 ± 0.09 MBq [125I]scFv8D3, 1.55 ± 0.07 MBq [125I]HS(-)scFv8D3 or 1.72 ± 0.05 [125I]HS(+)scFv8D3. Eight-microliter blood samples from the tail vein were collected at at 0.5-, 1-, and 2-hours post injection. Euthanasia of the animals and dissection of the animals, radioactivity measurement, TLC quality control and data presentation was done in the same manner as described above in the 96-hour in vivo plasma stability study.
Ex-vivo 48-hour biodistribution study
Half-life, brain uptake and peripheral biodistribution of scFv8D3,HS(+)scFv8D3 and HS(-)scFv8D3 was investigated in C57Bl/6 wild-type mice (3 months old). The mice were injected with a tracer dose (0.3 nmol/kg) of each scFv. Mice (n = 9, 3 mice for each protein) were intravenously injected via the tail vein with 1.82 ± 0.05 MBq[125I]scFv8D3, 1.6 ± 0.16 MBq [125I]HS(-)scFv8D3 or 1.66 ± 0.03 [125I]HS(+)scFv8D3. Eight-microliter blood samples from the tail vein were collected at 0.5-, 1-, 2-, 4,-, 6-, 24- and 48-hours post injection. Euthanasia of the animals was carried out 24-hours post-injection. Dissection of the animals, radioactivity measurement, TLC quality control and data presentation was done in the same manner as described above in the 96-hour in vivo plasma stability study.
Ex-vivo 24-hour biodistribution study
Half-life, brain uptake and peripheral biodistribution of scFv8D3, HS(+)scFv8D3 mutant and HS(-)scFv8D3 was investigated in C57Bl/6 wild-type mice (3 months old). The mice were injected with a therapeutic dose (30 nmol/kg) of each scFv. Mice (n = 9, 3 mice for each protein) were intravenously injected via the tail vein with 0.31 ± 0.07 MBq [125I]scFv8D3, 0.38 ± 0.05 MBq [125I]HS(-)scFv8D3 and 0.29 ± 0.07 MBq [125I]HS(+)scFv8D3. Eight-microliter blood samples from the tail vein were collected at 1-, 2-, 4-, 6-, 24- and 48-hours post-injection. Euthanasia of the animals was carried out 24-hours post-injection. Dissection of the animals, radioactivity measurement, TLC quality control and data presentation was done in the same manner as described above in the 96-hour in vivo plasma stability study.
Ex-vivo brain nuclear track emulsion (NTE) and capillary depletion study (CD)
Brain uptake and peripheral biodistribution of scFv8D3, HS(+)scFv8D3 and HS(-)scFv8D3 was investigated in C57Bl/6 wild-type mice (3 months-old). The mice were injected with a tracer dose (0.3 nmol/kg) of each scFv. Mice (n = 9, 3 mice for each protein) were intravenously injected via the tail vein with 0.49 ± 0.02 MBq [125I]scFv8D3, 0.62 ± 0.1 MBq [125I]HS(-)scFv8D3 or 0.62 ± 0.03 [125I]HS(+)scFv8D3. Eight-microliter blood samples from the tail vein were collected at 0.5-, 1-, and 2-hours post-injection. Euthanasia of the animals was carried out 24-hours post-injection. Radioactivity measurement, TLC quality control and data presentation were done in the same manner as described above in the 96-hour in vivo plasma stability study. The brains were dissected and divided into two hemispheres. The right hemisphere was left intact, measured and subsequently sectioned in readiness for NTE. The midbrain and cerebellum were dissected from the left hemisphere and the cortex was used for brain CD.
The ex vivo brain CD was performed as described previously50. Briefly, the brain cortices were isolated immediately after the animals were sacrificed by transcardial perfusion and then weighed and homogenized in 0.8 mL of cold physiological buffer (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM CaCl2, 1 mM MgSO4, 1 mM NaH2PO4, 10 mM D-glucose adjusted to pH 7.4) in an ice-cold Dunce homogenizer. Subsequentially, 0.8 mL of 30% Ficoll 400 (Sigma Aldrich) was added to the homogenate, which was then stroked one additional time in the Dunce homogenizer. The homogenate was transferred to a 15 mL Falcon tube and centrifuged at 5200 × g for 20 min at 4°C, resulting in a parenchymal supernatant and a capillary enriched pellet, which was carefully separated from the supernatant. The radioactivity of the parenchymal supernatant fraction and the capillary enriched pellet fraction was measured in a γ-counter (PerkinElmer. The radioactivity for each fraction was normalized to the injected dose (%ID) in MBq.
Before the ex vivo NTE and immunofluorescence CD31-staining was performed, the frozen right hemispheres were cryosectioned sagittally (20 µm) in a cryostat (CryoStar NX70, Thermo Scientific) and mounted on glass slides. CD31 staining was applied to visualize the capillaries in the brain as previously described48–50. The brain sections were fixed in ice-cold methanol for 10 min, washed in PBS and then blocked for 1- hour in 5% Normal Goat Serum. Slides were then washed again in PBS followed by washing buffer (0.1% Tween-20 in PBS) for 5 minutes while shaking. Sections were then incubated with 1.25 µg/mL rat-anti-mouse CD31 (BD, #553370) in PBS overnight. The next day sections were washed in washing buffer and subsequently incubated with the secondary antibody goat-anti-rat Alexa 647 for 1- hour at room temperature. The sections were washed and the NTE experiment was performed in darkness as previously described50. Briefly, ILFORD K5 emulsion (Oxford Instruments, Gometz la Ville, France) was prepared as a 50:50 emulsion in MQ-water under heating in a 40°C water bath. Previously CD31-stained brain sections were immersed in the ILFORD K5 emulsion for 10 s and the left to air dry for 2-hour before incubating for 5 weeks at 4°C. Sections were developed according to the to the manufacturer’s instructions, dehydrated in an ethanol concentration gradient (70%, 95% and lastly 100%) and mounted with Pertex (Histolab). Finally, the NTE and CD31-stained sections were imaged with a Zeiss Observer Z.1 microscope (Carl Zeiss Microimaging GmbH, Jena, Germany) and subsequentially the obtain images were processed equally using the ZEN software. The NTE-signal associated with capillary and parenchymal regions was quantified as previously described50. In short, 10 images from 2–3 individuals per scFv were analyzed with Fiji (ImageJ) by a standardized macro, and subsequentially the macro-generated results were quality controlled by manual inspection.
Free iodine-125 brain penetrance and blood distribution study
Nine C57Bl/6 wild-type mice (2–3 months old) were intravenously injected via the tail vein with 0.84 ± 0.09 MBq 125I in saline. Mice were euthanized at 2-hours (n = 3), 24-hours (n = 3) and 48-hours (n = 3) post-injection by a terminal blood sample from the heart, followed by transcardial perfusion with 0.9% physiological saline. The brain, whole blood, plasma, blood cell pellet (remaining after separation of plasma by centrifugation of whole blood) and major peripheral organs were isolated and the radioactivity was measured as described above in the 96-hour in vivo plasma stability study.
Statistics
Data are presented as mean ± SD. The data was tested for normality (gaussian distribution) followed by One-way ANOVA statistical test for the HS(-)scFv8D3 and HS(+)scFv8D3 mutants compared to the scFv8D3 for p-values: (*) < 0.05, (**) < 0.01, and (***) < 0.001.