Boosting Antitumor Drug Efficacy with Chemically Engineered Multidomain Proteins

Abstract A facile chemical approach integrating supramolecular chemistry, site‐selective protein chemistry, and molecular biology is described to engineer synthetic multidomain protein therapeutics that sensitize cancer cells selectively to significantly enhance antitumor efficacy of existing chemotherapeutics. The desired bioactive entities are assembled via supramolecular interactions at the nanoscale into structurally ordered multiprotein complexes comprising a) multiple copies of the chemically modified cyclic peptide hormone somatostatin for selective targeting and internalization into human A549 lung cancer cells expressing SST‐2 receptors and b) a new cysteine mutant of the C3bot1 (C3) enzyme from Clostridium botulinum, a Rho protein inhibitor that affects and influences intracellular Rho‐mediated processes like endothelial cell migration and blood vessel formation. The multidomain protein complex, SST3‐Avi‐C3, retargets C3 enzyme into non‐small cell lung A549 cancer cells and exhibits exceptional tumor inhibition at a concentration ≈100‐fold lower than the clinically approved antibody bevacizumab (Avastin) in vivo. Notably, SST3‐Avi‐C3 increases tumor sensitivity to a conventional chemotherapeutic (doxorubicin) in vivo. These findings show that the integrated approach holds vast promise to expand the current repertoire of multidomain protein complexes and can pave the way to important new developments in the area of targeted and combination cancer therapy.

Eagle Medium (DMEM, high glucose). Media contained 10% heat inactivated FCS, Lglutamate (2 mM) and 100 U ml -1 penicillin and 100 µg ml -1 streptomycin. Cells were routinely trypsinized and reseeded twice per week. For the experiments, cells were seeded in plastic dishes and incubated with the respective compounds in the medium at 37°C and 5% CO2. For studies of C3-mediated effects in cells, the pictures of the cells were taken after the indicated incubation periods with the toxins using an Axiovert 40CFl microscope from Zeiss (Oberkochen, Germany) connected to a progress C10 CCD camera from Jenoptik (Jena, Germany). For internalization studies into cells, images were obtained using a LSM 710 laser scanning confocal microscope (LCSM) system (Zeiss, Germany) coupled to XL-LSM 710 S incubator and equipped with a 63x oil immersion objective. The acquired images were processed with ImageJ software (NIH, Bethesda). For in vivo experiments, A549-Red-Fluc cells (Perkin Elmer, BW119266) were cultured in RPMI 1640 supplemented with 10% heatinactivated FCS and puromycin (2 µg ml -1 ).
Cellular uptake studies of transport proteins, SST(N)-Avi. A549 cells were pre-cultured in high glucose DMEM medium fortified with 10% fetal bovine serum, 1% penicillin/streptomycin and seeded at 6,500 cells/well in a white 96-well (half-area) plate. The cells were left to adhere overnight at 37°C, 5% CO2. The media were removed and different concentrations of the transporters SST(N)-Avi (50 nM, 100 nM, 200 nM, 500 nM) in 50 μl DMEM were added into each well. Avi was added as a control. The treated cells were subsequently incubated separately for 4 h at 37°C, 5% CO2. After incubation, the cells were washed (3 times) with Dulbecco's PBS buffer to remove non-specific binding followed by incubating the cells for a further 24 h in 50 μl per well of cell lysis buffer. Emission measurements (λex = 558 nm, λem = 585 nm) were recorded using TECAN M1000 microplate reader to determine the uptake efficiencies. The values were given as mean ± SD (n=4) and data were analyzed by one-way analysis of variance (ANOVA) with Bonferroni correction for multigroup comparison at *p<0.05, NS = not significant. with Origin Pro 9.1. Figure S1. Full statistical analysis of the cellular uptake of SST(N)-Avi transporters into A549 cells and concentration dependency shown in Fig. 2b in main text. The values are given as mean + SD. Data were analyzed by one-way ANOVA with Bonferroni correction for multi-group comparison at *p<0.05, NS = not significant.
Internalization of the transporter SST4-Avi into mammalian cells. A549 cells were seeded at a density 30,000 cells/well in a µ-Slide 8-well chambered coverslip (ibidi, Munich, Germany) in 300 µl DMEM medium with 10% FCS. The cells were cultured overnight to allow adhesion at 37°C, 5% CO2. Subsequently, the medium was removed and 2 µM of Avi and SST4-Avi were added, respectively. The cells were then further incubated for 4 h at 37°C, 5% CO2. Before imaging, cells were washed with DMEM medium for 3 times. Cell membrane was stained with 0.5 μl of CellMask Deep Red (0.5 mg ml -1 , ex/em: 649/666) for 5 min. The live cell imaging was performed using a LSM 710 laser scanning confocal microscope system (Zeiss, Germany) coupled to an XL-LSM 710 S incubator and equipped with a 63x oil immersion objective. The emission of SST4-Avi was recorded using a 525-759 nm filter and a 514 nm argon laser for excitation. The acquired images were processed with ZEN 2011 software ( Figure S2). h. Cell membranes were stained (red) to confirm internalization. Scale bar: 10 µm.
Functional assay for SST agonist activity against SSTR2. Cell-type selectivity was demonstrated by applying SST4-Avi to wild type CHO-K1/Gα15 40 cells and CHO-K1/Gα15/SSTR2 cells overexpressing SSTR2 for a functional calcium flux assay conducted by GenScript. The assay was performed in the agonist mode in duplicate. Internalization of SST3-Avi into different cancer cell types. SSTR2-positive A549 and LiSa-2 cells as well as SSTR2-negative SK-UT1-cells were seeded in 8-well plates (ibidi GmbH, Munich, Germany) with a density of 30,000 cells per well in 300 µl. The cells were incubated at 37°C and 5% CO2 for 24 h. SST3-Avi (fluorescein-labeled) with a final concentration of 400 nM was added to a total volume of 300 µl. As control, cells were treated only with DMEM.
After 24 h, the cells were washed with PBS and reconstituted with complete DMEM and imaging was performed using LSCM. Scale bars correspond to 10 µm.
Expression and purification of the Cys-mutant C3bot1-A1C. To insert a cysteine-mutation, we constructed a mutant of C. botulinum C3bot1-exoenzyme by site-directed mutagenesis using appropriate PCR primers. By doing so, we replaced Ala-1 by cysteine. Afterwards, the enzymatically active C3bot mutant C3bot1-A1C (Cys-C3) was overexpressed as GST-tagged protein and purified by affinity chromatography. E. coli BL21 transformed with pGEX2T-Cys-C3bot1 were grown in Luria-Bertani medium at 37°C to an optical density of 0.6 -0.8. The LB-medium was added with 100 µg ml -1 ampicillin. After reaching the desired optical density, protein expression was induced by adding 200 µM isopropyl-β-D-thiogalactopyranoside (IPTG) and the cultures incubated overnight at 29°C. The bacteria were harvested by centrifugation (5000 rpm, 10 min, 4°C) and resuspended in lysis-buffer containing 10 mM NaCl, 20 mM Tris-HCl, 1% Triton X-100, 1 mM PMSF, pH 7.4. After harvesting, the bacteria were disrupted by sonification, cellular debris were centrifuged for 10 min at 12000 rpm and 4°C and the clear supernatant was incubated overnight at 4°C glutathione-agarose beads (Macherey-Nagel, Düren, Germany). After incubation, the toxin-bead-mixture was centrifuged for 5 min at 2200 rpm and 4°C, the beads were then washed two times with wash buffer containing 150 mM NaCl, 20 mM Tris, pH 7.4 and one time with PBS. The bound toxin was then incubated with thrombin (20 NIH-units l -1 bacteria culture) for 1 h at room temperature to cleave the GST-tag. The toxin-containing supernatant was then obtained by centrifugation for 30 s at 4°C and 10000 rpm. Thrombin was removed by incubation of the supernatant with benzamidine beads (GE Healthcare, München, Germany) for 10 min at 21°C. The purity of the were cultivated at 37°C and 5% CO2 in Dulbecco's modified Eagle medium (DMEM), containing 10% heat-inactivated (30 min at 56°C) FCS. The medium contained L-glutamate (4 mM), penicillin (100 U ml -1 ) and streptomycin (100 µg ml -1 ). Cells were split at least twice per week. Cells were incubated with C3bot1-A1C, C3bot1 or medium only at a final concentration of 3 µg ml -1 for 6 h and C3-induced morphological changes were recorded (Fig. S5c). Expression of the protein was confirmed by detecting with a specific antibody against C3bot1.

Synthesis of biotinylhydrazine linker (1).
Biotin (200 mg, 0.82 mmol, 1 eq.), EDC (188 mg, 0.98 mmol, 1.2 eq.) and DMAP (12 mg, 98 μmol, 0.12 eq.) were dissolved in 5 ml of anhydrous DMF at 0°C under argon atmosphere and thereafter tert-butyl carbazate (130 mg, 0.98 mmol, 1.2 eq.) was added. The reaction mixture was stirred at RT overnight and the solvent was removed by high vacuum. The residue was purified by column chromatography by using eluting solvents 10% MeOH in CHCl3 to afford Boc protected biotinylhydrazine. This compound was further dissolved in 4 ml of DCM and 1 ml of TFA was added. The resulting mixture was stirred overnight and the solvents were removed by high vacuum to afford 274 mg of biotinylhydrazine 1 (0.74 mmol, 90% yield). 1

Zetapotential measurements.
A solution of 1 mg/mL avidin or 0.5 mg/mL unmodified C3 was prepared in 25 mM HEPES buffer and diluted 1:1 with 1 mM KCl. The zetapotential values were acquired using a Malvern Nano Zetasizer over min. 10 runs. As a control, the zetapotential of a solution with final concentration of 0.5 mg/mL human serum albumin (pI < 5) was determined. Zeta deviation (mV) 5.  After 18 h, the solutions prepared above were aliquoted (5 µl) and 0.2 µl of 1 M HCl was added to each solution for acidification. MST measurements were obtained after 4 h and there was no binding observed (Fig. S8, blue), which indicated the acid-induced C3 release.
A similar procedure was repeated for B-C3 in pH 4 buffer as a negative control and no binding was observed upon incubation with SST3-Avi (Fig. S8, magenta).