Application of Bld-1-Embedded Elastin-Like Polypeptides in Tumor Targeting

Expression of various molecules on the surface of cancer cells compared to normal cells creates a platform for the generation of various drug vehicles for targeted therapy. Multiple interactions between ligands and their receptors mediated by targeting peptide-modified polymer could enable simultaneous delivery of a drug selectively to target tumor cells, thus limiting side effects resulting from non-specific drug delivery. In this study, we synthesized a novel tumor targeting system by using two key elements: (1) Bld-1 peptide (SNRDARRC), a recently reported bladder tumor targeting peptide identified by using a phage-displayed peptide library, and (2) ELP, a thermally responsive polypeptide. B5V60 containing five Bld-1 peptides and non-targeted ELP77 with a thermal phase-transition over 37 °C were analyzed to determine their bioactivities. Further studies confirmed the superior binding ability of B5V60 to bladder tumor cells and the cellular accumulation of B5V60 in cancer cells was dependent on the expression level of sialyl-Tn antigen (STn), a tumor-associated carbohydrate antigen. Additionally, B5V60 displayed excellent localization in bladder tumor xenograft mice after intravenous injection and was strictly confined to sialyl-Tn antigen-overexpressing tumor tissue. Thus, our newly designed B5V60 showed high potential as a novel carrier for STn-specific targeted cancer therapy or other therapeutic applications.

SCIeNtIfIC RepoRts | (2018) 8:3892 | DOI: 10.1038/s41598-018-21910-z abbreviated as STn), a tumor-associated carbohydrate antigen that is overexpressed in various tumor cells [24][25][26] . Overexpression of these antigens has been correlated with cancer progression, poor prognosis, and an immunosuppressive microenvironment, which suggests they are important therapeutic targets. So far therapeutic vaccination used for clinical trials and antibodies against this antigen have limited success due to low immunogenicity or specificity, thereby development of innovative targeted delivery system is needed for effective cancer treatment.
To create a targeted system with multiple targeting sites, we can exploit thermally responsive recombinant elastin-like polypeptide (ELP), as it can be readily tailored with desirable biological and mechanical properties. ELPs consist of Val-Pro-Gly-Xaa-Gly pentapeptide repeats (with "guest residue" Xaa is any amino acid except Pro) derived from a structural motif found in mammalian elastin 27 . Compared to other polymeric drug delivery systems, ELPs are biodegradable, biocompatible, and less toxic 28,29 . ELPs undergo an inverse temperature phase transition, in which they are soluble at temperatures below their transition temperature (Tt) but become insoluble and aggregate at temperatures above their Tt 29 . ELP polymers can be synthesized at the genetic level by recombinant DNA methods, which means that their sequences, compositions, and molecular weights can be precisely tuned. ELP polymers can be easily expressed and purified at high yield simply by exploiting the inverse temperature cycling (ITC) method 27 . Studies have shown that genetically encoded synthesis of ELPs can be used to specify the location at which a biological drug, peptide, or protein is attached to an ELP sequence. Functionalization of ELP with targeting and internalization peptides has been found to improve accumulation and intracellular delivery of drugs at disease sites 30 . Fusion of ELP with a cell-penetrating peptide such as peptide derived from Drosophila Antennapaedia transcription factor (penetratin), HIV transactivation of transcription (TAT), and Kaposi fibroblast growth factor signal peptide (MTS) could increase intracellular delivery towards therapeutic targets, thereby enhancing drug efficacy 30 . Recently, it was demonstrated that penetratin-functionalized ELP-based delivery of kinase inhibitor peptide p21 induces enhanced cancer cell death. ELP diblock copolymers designed with cell-penetrating peptide (CPP) domain at the hydrophilic end and a therapeutic domain at the hydrophobic end form a thermally responsive micelle-like structure and induce the multivalent display of CPP on the exterior to enhance cellular internalization 31 .
In a previous study, AP1-ELP polymers containing multiple IL-4 receptor-targeting peptides were shown to increase intracellular localization into tumor tissue 32 . Multivalent display of targeting peptide AP1 along the backbone of ELP polymer increased affinity towards its target, IL-4 receptor, by approximately 10,000-fold compared to free peptides. We performed further studies to create another multivalent targeted based ELP, containing bladder tumors targeting peptide Bld-1. Thus, to improve the binding avidity and specificity of Bld-1, B 5 V 60 was prepared by introducing multivalent Bld-1 into ELP polymer by genetic engineering, after which it's physical and bio-chemical properties were analyzed. Tumor targeting activities of B 5 V 60 were examined both in vitro and in vivo. Notably, the correlation of STn expression and B 5 V 60 binding with cancer cells was investigated. Competitive inhibition of cell binding induced by B 5 V 60 towards anti-STn antibody could be useful in STn-based targeting of cancer cells as a drug delivery system as well as in enhancing immune responses against immunotolerant tumor cells highly expressing STn.

Results and Discussion
Design and Preparation of Bld-1 ELP. As ELPs can be synthesized at the genetic level by the recombinant DNA method, their sequences, compositions, and molecular weights can be precisely controlled 27 . In addition, ELP polymer can accommodate any target-specific ligands or functional groups as well as support multivalent presentation without any physiological changes or biological activities. Accordingly, in this study, we modified the coding sequence of ELP (VGVPG; with guest residue Valine) by incorporating bladder tumor-specific peptide (Bld-1; SNRDARR). The monomer gene referred to as B 1 V 12 was designed with one Bld-1 sequence (SNRDARR) in its N-terminal region, followed by 12 pentapeptide repeats of ELP. Using the recursive directional ligation method, targeted polymer of variable lengths [B 1 V 12 ]n were generated with periodic repetition of Bld-1 throughout the polypeptide sequences. The pentapeptide ELP sequence VGVPG was used to construct the Bld-1 ELP library. Since Valine is a hydrophobic guest residue with a low Tt, it is expected that the increase in Tt due to incorporation of hydrophilic Bld-1 peptide is moderated and maintained within a relevant temperature range suitable for clinical application. The control ELP was constructed by ligation of V 21 with [V 3 G 3 A 3 ] 8 (Fig. 1A, Figure S1). Bld-1 ELP containing five Bld-1 peptides (B 5 V 60 ) was used as a targeted polymer (Fig. 1B, Figure S1). V 21 -(V 3 G 4 A) 8 referred to as ELP 77 was used as a non-targeted polymer for further in vitro and in vivo experiments. B 5 V 60 and control ELP 77 proteins were expressed by IPTG induction and purified using the inverse transition cycling (ITC) method. After four rounds of ITC, B 5 V 60 and ELP 77 were analyzed by SDS-PAGE ( Figure S2). The sizes of B 5 V 60 and ELP 77 were approximately ~30 kDa. Determination of accurate molecular weight by MALDI-TOF/MS confirmed the protein sizes of ELP 77 (31341.8 Da) and B 5 V 60 (30649.2 Da) (Fig. 1C).
Thermal and Secondary Structure Characterization. Transition temperatures (Tt) of B 5 V 60 and ELP 77 proteins were monitored by measuring optical density at 350 nm as a function of temperature with 1 °C/min increments. Transition temperature (Tt) of ELP protein was defined as the temperature at 50% of the maximum of ELP aggregation. The Tt of B 5 V 60 was in range of 37~40 °C (Fig. 1D). Due to the presence of charge residues in Bld-1 peptide, the Tt of B 5 V 60 was elevated nearly 11 °C compared to the control ELP without targeting peptide. The Tt of B 5 V 60 protein (39.77 °C) was just higher than physiological body temperature, whereas the same ELP without targeting peptide (negative control) had a Tt of ~28 °C (data not shown). Substitution of hydrophilic amino acids such as Glycine and Alanine at the fourth guest residue of the ELP pentapeptide repeat in the new non-targeted control ELP 77 further increased Tt up to 38.67 °C (Fig. 1D), approximately similar to that of targeted B 5 V 60 . Measurement of turbidity profile at different concentration clearly reveal the dependency of Tt according to concentration ( Figure S4). Despite lowering of Tt with increased concentration, their applicability was considered to be unaffected since relatively small change in Tt obtained by large increase in concentration 33 . Later SCIeNtIfIC RepoRts | (2018) 8:3892 | DOI:10.1038/s41598-018-21910-z determination of particle size at different temperatures (24, 37, and 50 °C) using DLS revealed the size increment with increased temperature in consistence with turbidity profile. But at physiological body temperature the size of ELP 77 and B 5 V 60 were 415.5 nm and 413.4 nm respectively ( Figure S3). As tumor vessels are predicted to be leakier due to irregular development of vasculature and uncontrolled angiogenesis with pores ranging in size from 200 nm to 2 µm, thus both polypeptides will be well penetrable in tumor tissue 34 . Further, circular dichroism (CD) spectra confirmed changes in secondary structure along with an increase in temperature in both ELP 77 (Fig. 1E) and B 5 V 60 ( Fig. 1F) respectively. Both polypeptides appeared to structurally consist of a helix and ß-turn in a random coil conformation. Helix content increased at higher temperature in B 5 V 60 (Table S2), whereas helix formation decreased in ELP 77 (Table S1). Both polymers displayed an increase in ß-turn content and slight decrease in random coil content at higher temperature. The increase in helix content in B 5 V 60 was due to the presence of ligands Bld-1. However, incorporation of targeting ligands did not change the physical and chemical properties of ELP, clearly indicating its versatility in accommodating any functional peptide or protein.

In vitro Cell Binding Analysis.
To analyze cell binding activity, Alexa 488-labeled ELP 77 and B 5 V 60 proteins were incubated with 5637, HT-29, and HEK293 cells, and cellular binding was accessed using flow cytometry. Targeted B 5 V 60 polymer clearly revealed higher cellular binding activity compared to non-targeted ELP 77 polymer after 1 h of incubation at 4 °C ( Fig. 2A). B 5 V 60 polymer showed 17.44 ± 2.08-fold greater cell binding activity compared to ELP 77 in 5637 cells. Further, B 5 V 60 polymer showed 1.29 ± 1.08-fold higher cell binding activity than free Bld-1 peptide (Fig. 2B). On the other hand, both polymers showed minimum binding activity in HT-29 ( Fig. 2C,D) and HEK-293 cells (Fig. 2E,F). Thus, multivalent presentation of Bld-1 peptide along the ELP backbone increased cellular accumulation of tumor cells. In addition, B 5 V 60 polymer showed 5.18 ± 1.28-fold greater cell attachment in 5637 cells compared to HT-26 cells. These results clearly suggest that B 5 V 60 is highly specific to bladder tumors. Consistent with the flow cytometry data, confocal microscopy of adherent cells demonstrated that B 5 V 60 polymer was localized more efficiently onto the surface of 5637 cells (Fig. 3A) at 4 °C. In contrast, minimum localization was observed in HT-29 ( Figure S5A) and HEK293 ( Figure S6A) cells upon incubation with respective polymers. Further, B 5 V 60 polymer displayed improved cellular uptake towards 5637 cells (Fig. 3B) compared to HT-29 ( Figure S5B) and HEK-293 ( Figure S6B) cells upon incubation at 37 °C for 30 min. Neither ELP 77 nor B 5 V 60 showed significant cellular uptake by HT-29 and HEK-293 cells. These results clearly indicate that multiple Bld-1 peptide incorporation by the ELP polymer backbone resulted in greater tumor specific accumulation and uptake compared to the non-targeted ELP control. Together, it was confirmed that B 5 V 60 showed no accumulation in normal cells, which is crucial for drug delivery systems. STn Expression Determination and Competition Assay. In previous study it was specified through homology search that Bld-1 peptide shows similarity with SIRDARR motif found in human sialic acid binding immunoglobulin-like lectin 6 and 9 (Siglec 6 and 9) which interact with Neu5Aca2-6GalNAca1 (sialosyl-Tn, STn) a tumor-associated carbohydrate antigen, overexpressed in various tumor cells [24][25][26] . Thus, we investigated the level of STn expression in 5637 and HT-29 cells by flow cytometry. Higher expression (up to 45 ± 5%) of STn was observed in 5637 cells (Fig. 4B) in contrast to lower expression (around 15 ± 3%) in HT-29 cells (Fig. 4A). Increased binding of B 5 V 60 to 5637 cells as well as lower binding to HT-29 cells may be correlated with STn expression. To confirm STn-dependent binding of B 5 V 60 , competition assay was performed where 5637 cells were pre-incubated with different concentrations of anti-STn antibody (5 and 10 µg), and binding of respective polypeptide was measured by flow cytometry. Binding of B 5 V 60 was remarkably reduced in a concentration-dependent manner upon pre-incubation with anti-STn antibody (Fig. 4D,E). At a higher antibody concentration of 10 µg, binding of B 5 V 60 was reduced by two-fold in comparison with the isotype control. Minimum or no change in binding was observed when cells were incubated with ELP 77 (Fig. 4C,E). This result demonstrates that accumulation of B 5 V 60 on cells is highly dependent on the level of STn expression by cancer cells.
Next, we performed co-localization assay to confirm specific B 5 V 60 binding to STn, which is highly expressed in some tumor cells. Confocal microscopic images clearly revealed a greater accumulation of B 5 V 60 and anti-STn antibody on the surface of 5637 cells (Fig. 5B). Minimum accumulation was observed when cells were incubated with ELP 77 (Fig. 5A). Merged image showed improved co-localization, which further proves that both proteins shared the same binding target. On the other hand, B 5 V 60 and anti-STn antibody displayed lower binding in HT-29 cells ( Figure S7). These results demonstrate that binding of B 5 V 60 to tumor cells is dependent on STn expression by cancer cells.  with excitation and emission wavelengths of 675 nm and 698 nm, respectively, after ITC confirmed that nearly 18-20% of dye was released over 24 h, confirming the dye was firmly conjugated to its respective polypeptides ( Figure S8A). Further measurement of the fluorescent intensities of serially diluted labeled B 5 V 60 and E 147 polypeptides showed a dose-dependent decrease in intensity with no significant difference ( Figure S8B). Thus, after confirming labeling efficiency, athymic mice bearing 5637 bladder tumors were injected intravenously with respective polymers labeled with FNR-675. Near infrared fluorescence (NIRF) images taken at different time intervals showed that B 5 V 60 polymers were rapidly distributed within 10 min and accumulated time-dependently in tumor tissue (Fig. 6A). At 2 h post-injection, high fluorescence intensity in target tumors was observed in mice injected with B 5 V 60 polymer and persisted longer than 24 h ( Figure S9). In contrast, the ELP 77 control showed low accumulation towards tumor tissue compared to high accumulation in other organs. Thus, despite having the same chemical characteristics, B 5 V 60 showed superior tumor accumulation than the ELP 77 control due to the presence of multiples copies of targeting ligands. Ex vivo fluorescence images of excised tumors and organs collected at 24 h post-injection showed a 2.3-fold increase in fluorescence intensity in target tumors of B 5 V 60 -injected mice compared to ELP 77 injection (Fig. 6B). Higher accumulation in kidneys was observed in both ELP 77 and B 5 V 60 -injected mice due to rapid metabolism. Since protein molecular weight has a strong effect on biodistribution in vivo, higher kidney accumulation may be attributed to the lower molecular weights of both polypeptides 35 . However, fluorescence intensity in the liver was stronger in ELP 77 -injected mice compared to B 5 V 60 (Fig. 6C). Together, immunohistological examination of tumor tissue showed that B 5 V 60 was highly confined to STn-expressing tumor tissue, consistent with in vivo and ex vivo imaging results (Fig. 7). Collectively, these results further confirm the potential of B 5 V 60 as a candidate macromolecular drug carrier for cancer therapy.

Conclusion
In this study, we demonstrated the potential of genetically encoded synthesis of ELP polypeptide in which multiple tumors targeting peptides are randomly incorporated into the polypeptide backbone without the need for covalent attachment chemistry. Multivalent presentation of bladder tumor-targeting peptide onto the ELP backbone can increase cellular uptake compared to monovalent Bld-1 peptide and the non-targeted ELP control. Consistent with previous findings demonstrating the multivalent presentation of ELP with IL-4 receptor-targeting peptide has efficient targeting ability both in vitro and in vivo, this study further confirms our strategy of seamlessly incorporating functional peptides. An in vivo study revealed that B 5 V 60 polymer accumulated in tumor tissue and was retained for over 24 h. Accumulation of B 5 V 60 on tumor cells was highly correlated with the expression level of STn (sialyl-Tn). Since high expression of STn on cell surfaces is related with advanced-stage tumor and malignancy, our newly design B 5 V 60 polymer offers an approach for application of novel therapeutics such as selective drug-delivery or STn-based immunotherapy. Subsequently, this strategy can be further optimized to study the association of tumor-associated carbohydrate antigens with immune cells.

Methods
ELP Nomenclature. ELPs are designated as ELP [X a Y b Z c ]n where X, Y, and Z specify the guest residue, a, b, and c are the numbers of corresponding guest residue repeats, and 'n' denotes the number of monomer gene repeats for RDL. For example, V 3 G 4 A consists of seven pentapeptide XGVPG repeats with Valine, Glycine, and Alanine as guest residues (X). In this experiment, B 5 V 60 consisted of B as Bld-1 peptide and V as VGVPG with 60 repeats. [V 21 (V 3 G 3 A) 8 ] refers to an ELP 77 consisting of 77 pentapeptide repeats with Valine, Glycine, and Alanine as guest residues (VGVPG, GGVPG, AGVPG) used as a control.

ELP Gene Oligomerization and Expression.
Synthetic oligonucleotides encoding monomer genes of V 7 , (V 3 G 3 A) and VGSNRDARRG-V 5 containing BamH I, PflM I, Bgl I, and HinD III enzyme sites were obtained from Macrogen Inc. Seoul, Korea. Corresponding oligonucleotides were annealed and ligated into BamH I and HinD III double-digested pRSET B vector. The pRSET B containing V 7 was linearized with PflM I, enzymatically dephosphorylated with Calf intestinal alkaline phosphatase (CIP), and ligated with VGSNRDARRG-V 5 double-digested with PflM I and Bgl I. The resulting (VGSNRDARRG-V 5 ) 2 designated as B 1 V 12 was used as a monomer gene to synthesize Bld-1 ELPs with various lengths by RDL. On the other hand, V 7 and V 3 G 4 A were used as monomer genes to synthesize ELPs with various length by RDL. After gene oligomerization by RDL, Bld1-ELP with various lengths (n = 2, 4, 5, and 6) and different transition temperatures was obtained, and B 5 V 60 containing five Bld-1 peptides was used for further experimentation. In order to construct ELP 77 control [V 21 (V 3 G 4 A) 8 ], pRSET B containing V 21 was doubled-digested with PflM I and Bgl I and ligated into (V 3 G 3 A) 8 containing vector linearized with PflM I. Positive colony were confirmed through restriction digestion with BamH I and HinD III, followed by gene sequencing (Macrogen Inc. Seoul, Korea).

ELP Gene Expression.
For protein expression, expression vector pET 25 b+ vector was modified by ligation with annealed oligonucleotides encoding sense 5′-TATGAGCGGGCCGGGCTGGCCGTGCTAAA-3′ and anti-sense 5′-AGCTTTTAGCACGGCCAGCCCGGCCCGCTCA-3′ containing Nde I, Sfi I, and HinD III restriction enzyme sites. After confirmation through DNA sequencing, modified pET25b+ vector was digested with Sfl I, and ELP 77 or B 5 V 60 gene was ligated and transformed into DH5α competent E. coli. Positive colonies were Four rounds of ITC were conducted to remove all contaminants. ELP expression and purity were analyzed by SDS-PAGE, followed by Copper chlorite staining. Protein concentration was measured by Cary UV-Vis spectroscopy using an extinction coefficient of 5690 M −1 cm −1 for both ELP 77 and B 5 V 60 .

MALDI TOF/MS Analysis.
Accurate molecular weights of ELP 77 and B 5 V 60 were determined using an UltrafleXtreme (Bruker). For the measurement, proteins were dissolved with 0.1% trifluoroacetic acid and mixed with an equal volume of matrix solution (1:1). Resultant mixture (1 µl) was then applied to a standard steel target for drying at room temperature. The spectra were obtained after calibration with standards.
Thermal Characterization. Transition temperature (Tt) of ELP 77 and B 5 V 60 were determined by monitoring the turbidity profiles of protein solutions at a wavelength of 350 nm as a function of temperature using a UV-visible spectrophotometer (Agilent Technologies, CA, USA). The absorbance was monitored from 20 °C to 50 °C with 1 °C/min increments. The first derivative of the turbidity profile with respect to temperature was numerically calculated, and the Tt was defined as the solution temperature at 50% of the maximum turbidity gradient. The Tt of ELP 77 and B 5 V 60 protein were monitored at 10 μM concentration.   Confocal Microscopy. To test binding specificity, 5637, HT-29, and HEK293 cells were seeded on a four-chambered slide (8 × 10 4 /well) and grown to 80% confluence. After 24 h, cells were incubated with 10 μM Alexa 488-labeled ELP 77 and B 5 V 60 for 1 h at 4 °C or for 30 min at 37 °C. Cells were then fixed with 4% paraformaldehyde (Sigma Aldrich) and cell membrane marker, Wheat germ agglutinin (WGA) Alexa Fluor 594 conjugate (Molecular Probes, Inc., Eugene), and cell nuclei were stained with DAPI (Sigma Aldrich). Images were captured by a Zeiss LSM-510 Meta confocal microscope.
STn Expression analysis. 5637 and HT-29 (1 × 10 6 ) cells were incubated with anti-Sialyl Tn antibody (Abcam, Seoul, South-Korea) and IgG isotype antibody as a control for 1 h at room temperature. Cells were further washed with PBS to remove excess or unbound antibody and subjected to flow cytometry to measure the level of STn receptor expression. A total of 20,000 events were collected for each sample. Co-localization Assay. A total of 8 × 10 4 5637 and HT-29 cells were seeded on a four-chambered slide. After 24 h, cells were incubated for 1 h with anti-Sialyl Tn (1:100) antibody labeled with FITC at room temperature. Cells were further incubated with 10 μM FNR 675-labeled E77 and B 5 V 60 for 1 h at 4 °C. After several washes with PBS, cell nuclei were stained with DAPI (Sigma Aldrich) for 3 min. Images were captured by a Zeiss LSM-510 Meta confocal microscope.
In vivo Fluorescence Imaging. All animal experiments were reviewed and approved by the Committee on the Ethics of Animal Experiments of the Kyungpook National University (Permit Number KNU 2016-0083). This study strictly followed the recommendations of National Institute of Health (NIH) for the Care and Use of Laboratory Animals. Athymic nude mice (BALB/c nu/nu) were housed in a specific pathogen-free environment at 22 ± 2 °C, 55 ± 5% relative humidity with light. Tumors were created by subcutaneously injecting 5637 cells (5 × 10 6 cells) into the right flanks of 5 week-old female mice. Tumors size (3-5 mm in diameter) usually develop within 1 month. Actually, the tumor sizing around 5 mm in diameter is more effective for our peptide delivery studies rather than tumor size below 3 mm in diameter. Mice bearing a subcutaneous tumor were anesthetized with 1.5% isoflurane inhalation and injected intravenously with approximately 3 mg/kg of FNR675-labeled ELP 77 (n = 10) and B 5 V 60 (n = 10). In vivo fluorescence images were taken at different time points after anesthetization (0 min, 1 h, 2 h, 4 h, 6 h, 12 h, and 24 h) using Optix eXplore (ART, Advanced research technologies Inc., Montreal, Canada).

Ex vivo Fluorescence Imaging and Tissue Preparation.
Twenty-four hours after intravenous injection, animals were euthanized with CO 2 , and tumor and organs were collected. Ex vivo fluorescence images were then taken. Tumor tissues were fixed with 4% paraformaldehyde overnight and frozen for cryosectioning. Tissues were sectioned with 8 mm thickness and incubated with anti-STn antibody (1:100) overnight. Tissues were stained with Alexa 488-labeled goat anti-mouse IgG secondary antibody (1:200), whereas nuclei were stained with DAPI and observed under a confocal microscope.
Statistical Analysis. Statistical significance of differences between experimental and control groups was analyzed by Student's t-test for two groups or one-way analysis of variance (ANOVA) for more groups. P < 0.05 was set as statistical significance, was denoted by asterisks in the figures.