Novel Strategy Utilizing Extracellular Cysteine-Rich Domain of

Jul 17, 2017 - The development of proteolysis-resistant d-peptide ligands for targeted drug/gene delivery has been greatly limited, due to the challen...
0 downloads 7 Views 4MB Size
Subscriber access provided by UNIV OF NEWCASTLE

Article

A Novel Strategy utilizing Extracellular Cysteine-Rich Domain of Membrane Receptor for Constructing D-Peptide Mediated Targeted Drug Delivery Systems: a Case Study on Fn14 Zhuoxuan Li, Jing Xie, Shan Peng, Sha Liu, Ying Wang, Weiyue Lu, Jie Shen, and Chong Li Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.7b00326 • Publication Date (Web): 17 Jul 2017 Downloaded from http://pubs.acs.org on July 17, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Bioconjugate Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

D-peptide that specifically binds to the Fn14 receptor, was obtained by mirror-image phage display. The biostable D-Fn14 peptide modified liposomes can target both stroma-rich pancreatic cancer and lung matastases of tripple negative breast cancer following intravenous administration. 754x271mm (300 x 300 DPI)

ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A Novel Strategy utilizing Extracellular Cysteine-Rich Domain of Membrane Receptor for Constructing D-Peptide Mediated Targeted Drug Delivery Systems: a Case Study on Fn14 Zhuoxuan Li1, Jing Xie1, Shan Peng1, Sha Liu1, Ying Wang1, Weiyue Lu2, Jie Shen3, Chong Li1,* 1 College of Pharmaceutical Sciences, Southwest University, Chongqing, 400715, PR China, 2 School of Pharmacy, Key Laboratory of Smart Drug Delivery, Ministry of Education, Fudan University, Shanghai 201203, China 3 College of Pharmacy; University of Rhode Island, Kingston, RI 02881, United States *Corresponding authors: Chong Li, College of Pharmaceutical Sciences, Southwest University, No. 2 Tiansheng Road, Beibei District, Chongqing 400715, China. Tel.: +86-23-68251225; Fax: +86-23-68251225; E-mail address: [email protected].

1 ACS Paragon Plus Environment

Page 2 of 34

Page 3 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Abstract The development of proteolysis-resistant D-peptide ligands for targeted drug/gene delivery has been greatly limited, due to the challenge lies in the chemical synthesis of membrane receptors without altering their structures. In the present research, a novel strategy utilizing self-stabilized extracellular CRD of membrane receptor was developed to construct D-peptide ligands and their mediated targeted drug delivery systems. Fn14, a cell surface receptor overexpressed in many cancers including pancreatic and triple-negative breast cancers, was selected as the model receptor. Fn14 CRD was synthesized and folded, and used to screen Fn14 binding peptides using phage display (L-peptide) and mirror-image phage display (Dpeptide) techniques, respectively. The D-peptide ligand successfully mediated targeted drug delivery to Fn14 positive tumor cells. In addition, the D-peptide possessed better targetbinding affinity, stromal barrier permeability, and tumor targeting ability in vivo when conjugated with liposomes. More importantly, D-peptide mediated liposomal paclitaxel delivery significantly inhibited pancreatic tumor growth in a subcutaneous xenograft model and drastically prolonged survival in a lung metastasis of breast cancer mouse model. This study demonstrated that mirror-image phage display based on the CRD of membrane receptor can be a promising strategy to advance active targeted drug delivery via bio-stable D-peptides.

Introduction Over the past decades, peptide-mediated delivery systems have become one of the main strategies to achieve active targeted drug/gene delivery. Compared to other kinds of targeting molecules such as biomacromolecules or small chemicals, peptide ligands offer various advantages such as high affinity and good selectivity in binding with target proteins, low immunogenicity, and relatively simple site-specific modification.1 However, peptides may be susceptible to proteolytic degradation in physiological environments, which may significantly affect the in vivo fate of peptide-mediated drug delivery systems.2 Recently, D-enantiomeric 2 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 34

peptides that are entirely composed of D-amino acids (except for glycine, which has no Denantiomers), have attracted increasing interest, owing to their high resistance to proteolytic degradation under physiological environments and hence superior ability in retaining peptide functionalities (e.g. binding with membrane receptors). 3-5

Retro-inverso isomerization is a commonly used method to design a D-enantiomeric peptide based on the sequence of its native L-peptide.

6,7

A retro-inverso peptide, also known as a

retro-enantio- or retro-all-D-peptide, is a peptide composed of D-amino acids with the reversed sequence. It has been suggested that the side chain topology of certain retro-inverso peptides resembles that of their native counterparts. Consequently, these retro-inverso peptides emulate the biological activities of the parent molecules while remaining fully resistant to proteolytic degradation.8 This strategy was originally proposed by pioneers in peptide chemistry to tackle peptide susceptibility to proteolysis over 30 years ago, and has been used in targeted drug delivery in recent decades

6,9

. For example, the retro-inverso

isomer (DFDSDMDEDGDIDHDQDGDQDHDPDKDIDRDMDIDQDMDTDI) of a peptide fragment (ITMQIMRIKPHQGQHIGEMSF) derived from vascular endothelial growth factor (VEGF), maintained the affinity to the L-peptide receptor VEGFR2. This retro-inverso isomer efficiently mediated targeted drug delivery against VEGFR2-overexpressed melanoma, tumor neovasculature, as well as vasculogenic mimicry.10 Recently, the application of retro-inverso ‘versions’ of other peptide ligands (e.g. prosaptide D4, angiopep-2, and CDX) in targeted drug delivery has also been demonstrated.3,4,11

For some malignant cancers such as pancreatic cancer and triple-negative breast cancer (TNBC), the lack of attractive therapeutic targets and their targeting ligands constitutes a significant challenge to the design and development of appropriate targeted delivery systems. In addition, the stroma surrounding pancreatic cancer cells may provide a strong barrier to 3 ACS Paragon Plus Environment

Page 5 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

limit and/or prevent therapeutics to reach cancer cells.12-15 Accordingly, screening target ligands, preferably D-enantiomeric peptide ligands, is urgently needed for improving targeted drug delivery and consequent the treatment of these life-threatening cancers. Phage display peptide library is a well-known, powerful research tool for high-throughput screening of natural peptide ligands towards target molecules, which theoretically can be any molecule. Since L-peptide ligands obtained from natural phage systems are subject to degradation by naturally occurring enzymes, Schumacher et al. developed a novel phage display technique termed mirror-image phage display, through which D-enantiomeric peptide ligands can be obtained.16 In this approach, the natural protein target is replaced by its enantiomeric counterpart, which is chemically synthesized in the D-amino acid configuration. The Denantiomeric protein is used to select a peptide ligand from a phage display library expressing random L-amino acid peptides. The selected peptide of interest is ultimately translated into its D-enantiomeric form, which can interact with the original target protein composed of natural L-amino acids, for reasons of symmetry (Scheme 1). Clearly, chemical preparation of Denantiomeric protein plays a crucial role in this process. Considering that membrane proteins or receptors are typically large proteins or proteins with complex structures, it has been very challenging to synthesize and/or fold these proteins.17 This has greatly limited the application of mirror-image phage display.

It has been found that many cell surface receptors contain cysteine-rich domains (CRD), which are usually only dozens of amino acid residues in length and with several cysteine pairs forming disulfide bonds that stabilize unique structures of the proteins.18 With the recent advances in thiol/disulfide chemistry, it has become technically possible to fold peptides/proteins with desired disulfide bond parings.17,19 Accordingly, it is envisioned that mirror-image phage display based on relatively small but stable CRD regions of the

4 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

membrane receptors rather than the full extracellular domain of receptors themselves, may be a promising approach for screening of D-peptide ligands.

In the present research, a novel strategy utilizing extracellular CRD of membrane receptors for D-peptide mediated targeted drug delivery was presented. Fibroblast growth factorinducible 14 (Fn14), also known as the TWEAK (tumor necrosis factor-like weak inducer of apoptosis) receptor, was chosen as the target membrane receptor. Fn14 can respond to its natural ligand TWEAK to promote fibroblast proliferation, and it has been found to be extensively overexpressed in pancreatic cancer cells,20 and TNBC.21 A D-peptide ligand of Fn14 was obtained using mirror-image phage display based on one Fn14 canonical CRD (EQAPGTAPCSRGSSWSADLDKCMDCASCRARPHSDFCLGCAAA, MW=4433.98 Da).22 In addition, the D-peptide mediated liposomal delivery systems were constructed and characterized, and their capability of targeting pancreatic and TNBC cancer cells, as well as their therapeutic effects were investigated.

Scheme 1. (The principle of mirror-image phage display technology.) The D-enantiomeric form of a natural target protein/receptor is synthesized and used for phage display to obtain an L-peptide ligand (indicated as yellow shapes) that binds to the D-enantiomeric target protein/receptor. Subsequently, the D-enantiomeric form of the obtained L-peptide ligand is synthesized (indicated as an orange shape), and it can bind to the natural target protein/receptor for reasons of symmetry.

5 ACS Paragon Plus Environment

Page 6 of 34

Page 7 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Results and Discussion Preparation and characterization of target protein Cysteine residues of proteins often form disulfide bonds, and linkages of multiple disulfide bonds play a crucial role in stabilizing and maintaining the distinct three-dimensional structure of extracellular motifs/domains in some membrane receptors. For example, the change in cysteine connectivity between Cys6-Cys174 of the extracellular domain of the human interleukin-6 receptor resulted in a complete loss of ligand binding.23,24 Accordingly, it was envisioned that the CRD of the membrane receptors may provide a relatively independent target region for the recognition and screening of ligands that can potentially be used to mediate targeted drug/gene delivery.

The extracellular CRD of natural Fn14 contains six cysteines arrange in the following unique pattern: Cys36-Cys49, Cys52-Cys67, and Cys55-Cys64,22 which greatly benefits the determination of its conformation. Following the chain assembly of the linear Fn14 CRD (Figure S1), the challenge lies in cross-linking the paired cysteines in a desired pattern since free cysteines can randomly pair with each other, resulting in a “misfolded” product. Typically, a conservative and time-consuming stepwise oxidation method is utilized to cross-link free cysteins in a controlled manner. In this method, cysteine pairs are protected by different protective groups during solid phase synthesis, and deprotection of each cysteine pair group is sequentially performed to form disulfide bonds.25 With the advances in peptide/protein synthesis, simultaneous folding of three disulfide bonds using sufficient amounts of denaturants in the presence of oxidized/reduced thiol pairs has become feasible.26 The folding conditions were optimized in the present research in order to obtain Fn14 CRD with its preferential conformation (three disulfide bridges), which was confirmed and is shown in Figures 1A and 1B.

The

disulfide

connectivity

was

verified

by

disulfide

mapping

aided

by

chymotrypsin/trypsin digestion coupled with LC-MS analysis.27 Both L-Fn14 and D-Fn14 6 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

CRDs were synthesized and folded as described above. As expected, the CD spectra of the two proteins were symmetrical about the X-axis for a pair of mirror image isomers (Figure 1C). Biotinylated L-Fn14 CRD and D-Fn14 CRD that designed to be immobilized to the surface of streptavidin-conjugated magnetic beads for phage display, were also prepared so that the site-specific modification of ‘‘affinity handles’’ can ensure the correct orientation of target proteins for productive phage binding.28

Figure 1. (Characterization of Fn14 CRD) Mass spectrum (A) and HPLC chromatography (B) of the folded Fn14 CRD. (C) CD spectra of D-Fn14 CRD and L-Fn14 CRD.

Screening and characterization of D-peptide ligand In order to confirm the feasibility of screening the Fn14-binding peptide (FNB peptide), a phage display of L-FNB peptide was performed prior to the mirror-image phage display. Considering the possible non-specific binding that emerges in the interaction between the phage library and L-Fn14 CRD (Figure S2A), competitive elution was conducted to avoid the elution of background-bound phage. The natural ligand of Fn14, TWEAK, was served as the eluent. After five rounds of panning, twenty plaques were randomly selected and the amplified DNA was sequenced, of which thirteen binding clones gave rise to the following sequence: DMHADHAVYTPR (Table S1). The specific binding of this L-FNB peptide to Fn14 CRD was confirmed by SPR (Figure S3A), and the D-enantiomeric isomer of the 7 ACS Paragon Plus Environment

Page 8 of 34

Page 9 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

screened L-FNB peptide was used as an eluent in the following mirror-image phage display. The insert sequence of the most promising clones that bound to D-Fn14 CRD was selected, and this phage-expressed L-peptide sequence was used a template for the chemical synthesis of the D-FNB peptide using D-amino acids as DEDGDADKDHDGDLDTDFDSDGDG (Table S2). Figure 2A shows the binding between the D-FNB peptide and L-Fn14 CRD by SPR assay. After fitting with a 1:1 Langmuir model, the affinity of D-FNB peptide against Fn14 CRD was calculated as 280 nM, suggesting its potential as a Fn14-targeting moiety.

To further understand the interactions between the obtained peptides and the Fn14 receptor, a molecular docking approach was implemented to explore the binding modes of these peptides. The peptides were found to rotate around the target protein in a serpentine manner (Figure S2B-S2D). The binding appeared to be dominated by hydrophobic, hydrogen bonding, and electrostatic interactions. For D-FNB peptide in particular (Figure 2B), the peptide binding affinity is mainly contributed by the hydrogen-bond networks, which was similar to the binding mode of L-FNB (Figure S3B). At the head of D-FNB peptide, Glu1 formed electrostatic interaction with Arg58. The backbone of Glu1 and Gly2 formed a large hydrogen-bond network with the Asp45 and Asp62 of the Fn14 protein. At the tail of the peptide, Gly12 also formed a large hydrogen-bond network with the Arg38 of the Fn14 protein. The binding energy was predicted as -8.796 kJ mol-1 and the Ki value was predicted as 352.03 nM, which was in good agreement with the experimentally determined values (Table S3).

8 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. (Interaction between D-FNB peptide and L-Fn14 CRD.) (A) Surface plasmon resonance (SPR) assay. The concentrations of injected D-FNB peptide are indicated in the overlay of the sensorgrams. (B) The predicted binding mode of D-FNB (colored yellow). The L-Fn14 protein (colored white) is shown as a ribbon. Residues involved in binding are represented by sticks and hydrogen-bond networks are denoted by black dashed lines.

Characterization of liposomes Peptide-functionalized lipids (i.e. D-FNB-PEG3400-DSPE and L-FNB-PEG3400- DSPE) were synthesized by conjugating the reactive end group of MAL-PEG-DSPE with thiolated D-FNB or L-FNB peptides via the thiol-maleimide Michael-type addition. As is shown in Figure 3A, the experimental molecular weights (MWs) of MAL-PEG3400-DSPE before and after conjugation were approximately 4,389 Da and 5,652 Da, respectively. The difference between the two values (1,263 Da) was consistent with the theoretical MW of the D-FNB peptide (1,263.4 Da), confirming that the reaction between the D-FNB peptide and MAL-PEG3400DSPE was successful. The successful conjugation of the L-FNB peptide and MAL-PEG3400DSPE was also confirmed (data not shown).

The physicochemical characteristics of the liposomes with or without the modification of FNB peptides are summarized in Table S4. Overall, all the prepared liposomes had average particle size of 90-100 nm, and peptide ligand modification resulted in slight increase in liposomal particle size. All the liposomes showed uniformly spherical shape, and 9 ACS Paragon Plus Environment

Page 10 of 34

Page 11 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

representative TEM images (PTX loaded liposomes) of the prepared liposomes are shown in Figures 3B. Tunable resistive pulse sensing confirmed that all the liposomes are uniform and with a mean particle size around 100 nm and the raw concentrations around 1.6 × 1012 particles/ml (Figure 3C). In addition, around 90% of PTX was encapsulated into the liposomes. PTX loaded liposomes (i.e. mPEG-Lip/PTX and D-FNB-Lip/PTX) showed similar release characteristics, with more than 80% cumulative PTX released within four days (Figure 3D). The modification of D-FNB did not appear to affect PTX release from the prepared liposomes.

Figure 3. (Characterization of functionalized lipids and their modifed liposomes (i.e. D-FNBLip/PTX and mPEG-Lip/PTX).) (A) MALDI-TOF mass spectrometry analysis of MAL-PEG3400-DSPE (upper panel, red color) and D-FNBPEG3400-DSPE (lower panel, blue color). (B) Transmission electron microscopy images of D-FNBLip/PTX. (C) Tunable resistive pulse sensing histograms for liposomes showing both particle size distribution and liposome concentrations. (D) PTX release profiles of mPEG-Lip/PTX and D-FNBLip/PTX in PBS containing 10% FBS (n = 3, mean ± SD).

10 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

D-FNB peptide-mediated recognition and targeting of Fn14-positive target cells in vitro For direct evaluation of the binding between peptide-modified liposomes and surface receptors on the target cells, a real-time cell-liposome interaction assay was performed at room temperature using the Ligand Tracer system. Compared with mPEG-Lip, typical binding traces of D-FNB-Lip on Bxpc-3 cells were observed (Figure 4A). For Bxpc-3 cells, the affinity constants (Ka) of D-FNB-Lip and mPEG-Lip were 1.76 × 104 and 3.64 × 103, with estimated dissociation constants (Kd) of 3.97 × 10-5 and 1.61 × 10-2, respectively. On the other hand, due to the lack of Fn14 receptor expression on BT-474 cells (Figure S4), the binding of D-FNB-Lip and mPEG-Lip appeared to be similar (Figure S5). These results demonstrated that the D-FNB peptide can mediate Fn14-targeting liposomal drug delivery.

In vitro cellular uptake of liposomes was evaluated qualitatively via fluorescence microscopy and quantitatively via flow cytometry. Quantitative fluorescence assessment showed that the intracellular RhB intensity of the D-FNB-Lip/RhB group in positively expressing Fn14 cells (i.e. Bxpc-3, NIH 3T3, and MDA-MB-231) was 2.21-fold, 1.97-fold, and 2.4-fold that of the mPEG-Lip/RhB group, respectively (Figures 4B-4D). The cellular uptake of the prepared liposomes in different cells was also visualized using fluorescence microscopy (Figure 4E), following a similar trend to that of the flow cytometry analysis. Furthermore, compared to its scrambled control (random permutation of D-FNB peptide, D-FNBS), D-FNB peptide exhibits higher and more specific uptake in two different Fn14 highly expressing cells (i.e. Bxpc-3 and Mcf-7) (Figure S6). To further demonstrate the specific Fn14 targeting effect of the D-FNB modified liposomes, Bxpc-3 cells were pretreated with Fn14 ligand TWEAK as well as the DFNB peptide. As shown in Figure 4F, both TWEAK and the D-FNB peptide significantly decreased the uptake of D-FNB-Lip in Bxpc-3 cells. These results confirmed that D-FNB can be used to mediate Fn14 targeted drug delivery via the ligand-receptor interaction between the Fn14 receptors and the D-FNB peptide ligand. 11 ACS Paragon Plus Environment

Page 12 of 34

Page 13 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Generally, liposomes release their payload (e.g. PTX) and exert anti-proliferation effects following internalization into cells. In this respect, all PTX formulations investigated showed the ability to inhibit the proliferation of cancer cells (Figure 4G-4H). For Bxpc-3 cells, the IC50 values of D-FNB-Lip/PTX, mPEG-Lip/PTX, and Abraxane were 147.5, 772.1, and 675.9 ng/ml, respectively (Figure 4G). A similar trend was found in the triple-negative MDA-MB231 breast cancer cells, in which the IC50 values of D-FNB-Lip/PTX, mPEG-Lip/PTX, and Abraxane were 7.7, 98.6, and 78.9 ng/ml, respectively (Figure 4H). Compared with the unmodified liposomes, D-FNB peptide-modified liposomes showed a remarkably better antiproliferative effect, which was mainly attributed to the enhanced cellular uptake of liposomes via Fn14-ligand recognition. Most importantly, D-FNB-Lip/PTX demonstrated better antiproliferation effect compared to the commercial product Abraxane, indicating that D-FNB modified drug delivery systems may be a promising treatment strategy for pancreatic cancers.

12 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. (In vitro targeting ability of D-FNB-Lip in different cell lines.) (A) Real-time interaction assay between liposomes and Bxpc-3 cells (over-expressing the Fn14 receptor). Red and black curves represent binding kinetic fits using TraceDrawer 1.2 for D-FNB-Lip and mPEG-Lip binding to the Fn14 receptor in Bxpc-3 cells, respectively. Flow cytometry analysis of Bxpc-3 (B), NIH 3T3 (C), and MDA-MB-231 (D) cells treated with free RhB, mPEG-Lip/RhB, and D-FNB-Lip/RhB. (E) Fluorescence images of cellular uptake of RhB-labeled liposomes in Bxpc-3, NIH 3T3 and MDA-MB-231 cells following a 2-h treatment of the liposomes. The scale bars are 50 µm. (F) Flow cytometry results of the competitive cellular uptake inhibition assay in Bxpc-3 cells with the pre-treatment of free D-FNB or TWEAK, respectively. Cell viabilities of Bxpc-3 pancreatic cancer cells (G) and MDA-MB-231 breast cancer cells (H) following the treatment with the formulations in vitro. Data are presented as the mean ± standard deviation (n = 3).

D-FNB peptide mediated tumor targeting and tumor penetration in vivo Pancreatic adenocarcinoma is characterized by a dense background of pancreatic cancerassociated fibroblasts, which is a key determinant in the malignant progression of cancer.29 13 ACS Paragon Plus Environment

Page 14 of 34

Page 15 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Accordingly, Bxpc-3/NIH 3T3 tumor, a stroma-rich subcutaneous xenograft tumor model consisting of both Bxpc-3 human pancreatic cells and NIH 3T3 fibroblast cells (with Matrigel), was used to better resemble the original tumor.30 As shown in Figures 5A-5C, DFNB-Lip accumulated in Bxpc-3/NIH 3T3 tumor regions at different time points following the systemic administration of DiR-loaded liposomes, whereas only marginal fluorescence signals were observed in the tumors of the D-FNBS-Lip and mPEG-Lip groups. Immunohistochemistry on Bxpc-3/NIH 3T3 tumor tissues from the D-FNB-Lip group showed strong co-localization of the fluorescence signals of C6 and Fn14 or C6 and α-SMA (α smooth muscle actin), which was not observed in the D-FNBS-Lip group (Figure 5D). This confirmed that D-FNB-Lip can target both tumor and tumor associated fibroblast cells in pancreatic cancerous tissues.

Figure 5. (In vivo targeting evaluation of D-FNB peptide and D-FNBS peptide modified liposomes containing DiR in nude mice bearing Bxpc-3/NIH 3T3 tumors.) (A) In vivo near infrared fluorescence imaging of model mice following intravenous administration of DiRloaded liposomes during a 24-h study period. Red circles indicate tumor locations. (B) Ex vivo images and (C) semi-quantitative results of tumors and main organs collected from the animals euthanized 24 h post administration. (D) Co-localization immunohistochemistry assay of the Bxpc-3/NIH 3T3 tumors from the D-FNB-Lip/C6 and D-FNBS-Lip/C6 treated groups (scale bars are 200 µm). Tumor tissue sections were stained with α-SMA and Fn14 antibodies.

Synchronous targeting of the tumor associated fibroblasts as well as pancreatic cancer cells may be critical for the treatment of pancreatic cancers as this strategy may facilitate the penetration of anticancer drug through the stromal barrier to reach pancreatic cancer cells. To 14 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

assess the tumor penetration capability of D-FNB-Lip, RhB loaded liposomes (i.e. D-FNBLip/RhB, and mPEG-Lip/RhB), and an RhB solution were intravenously injected into tumorbearing mice, and the tumors were harvested and sectioned 4 h after injection. Confocal microscopy images showed a gradient of fluorescence signal from the tumor edge to the tumor core that was significantly wider and stronger in the D-FNB-Lip group than in the other groups (Figure 6A), indicating the superior tumor penetration capability of D-FNB-Lip. Further analysis of the average tumor penetration depth of the liposomes investigated showed that the penetration depth of the D-FNB-Lip group was 867.4 ± 42.9 µm, approximately two times that of the mPEG-Lip group (427.3 ± 30.1 µm) and four times that of the control group (239.4 ± 18.6 µm). The tumor penetration study was also conducted using the 3D tumor spheroids. As shown in Figure 6B, much stronger fluorescence signal throughout the 3D tumor spheroids was observed for the D-FNB-Lip group compared to the mPEG-Lip group.

Figure 6. (Tumor penetration effect of the prepared liposomes.) (A) Penetration of RhB loaded liposomes into pancreatic tumor tissues (co-implanted Bxpc-3 and NIH 3T3 cells). Free RhB was studied as a control. Red color: RhB. (B) CLSM uptake images of C6-loaded liposomes in 3D tumor spheroids at different time points. The scale bars are 200 µm.

15 ACS Paragon Plus Environment

Page 16 of 34

Page 17 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

These data confirmed that D-FNB peptide modified liposomes can successfully cross the stromal barrier and subsequently increase the perfusion of the drug payload into the target tumor site, which in turn may greatly improve anticancer efficacy. A further investigation was conducted to uncover the uptake mechanism of D-FNB-Lip by Fn14-positive cells. As shown in Figure S7, chlorpromazine and filipin displayed a marked effect on the cellular uptake of D-FNB-Lip compared with mPEG-Lip (36.4% and 14.8% decreased, respectively), indicating that the clathrin and caveolin-dependent pathways are the main mechanisms of Fn14 receptormediated endocytosis in both fibroblasts and Bxpc-3 cells.

In vivo anti-tumor effect The in vivo antitumor efficacy of PTX-loaded liposomes was investigated in two tumor models (i.e Bxpc-3/NIH 3T3 subcutaneous tumor-bearing and TNBC nude mice). As shown in Figures 7A and 7B, the tumor size of the untreated group increased almost 4 fold over a period of 2 weeks in Bxpc-3/NIH 3T3 subcutaneous tumor-bearing nude mice. All PTX formulations showed anti-cancer efficacy to different extents, among which the D-FNBLip/PTX treatment almost completely inhibited tumor growth compared with the mPEGLip/PTX treatment group and the Abraxane group, a marketed PTX product (Figures 7A and 7B). The Bxpc-3/3T3 tumors excised after treatment were dissected and sectioned for TUNEL and α-SMA immunohistochemistry assays, respectively. As shown in Figure 7C, D-FNBLip/PTX exhibited the most effective pro-apoptotic effects, and induced apoptosis in 43% of cells, compared to 7% in the control group, 13% in the mPEG-Lip/PTX group, and 22% in the Abraxane group. In addition to eliciting an apoptosis-inducing effect in carcinoma cells, DFNB-Lip/PTX significantly inhibited α-SMA-positive fibroblasts in the tumor extracellular matrix,30 which could notably enhance tumor penetration and hence anti-tumor effect of drugloaded liposomes (Figure 7D). Moreover, there is no significant difference in body weight of mice bearing tumor after different treatments (Figure S8). 16 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7. (Antitumor effects of PTX-loaded formulations in Bxpc-3/NIH 3T3 bearing nude mice.) (A) Relative tumor volumes of various treatment groups (red arrows indicate the injection dates). Results are means ± SD (n = 6, *p < 0.05, **p < 0.01). (B) Morphology of tumor xenografts in different groups. (C) Evaluation of the effect on the induction of tumor cells apoptosis via TUNEL assay (black arrows indicate the positive cells). Scale bars are 40 µm. (D) Immunofluorescence analysis (a-SMA staining) of fibroblast growth inhibition. Scale bars are 500 µm.

In a sequential parallel comparison study, L-FNB-Lip showed significantly less permeability and accumulation in the target region compared to D-FNB-Lip (Figures S9 and S10A-B). More importantly, D-FNB peptide was much more stable in blood compared to the L-FNB peptide (Figure S10C). The superior stability of the D-FNB peptide in the blood circulation may contribute to the better tumor targeting and penetration of the D-FNB peptide mediated targeted delivery systems, compared to the L-FNB peptide, especially given that lysosomes are involved in the receptor-mediated endocytosis of FNB peptides (Figure S11).31,32 The remarkably improved antitumor efficacy of D-FNB peptide modified liposomal PTX further confirmed that the Fn14-mediated targeting strategy can not only target pancreatic cancer cells but also improve the permeability of therapeutics to reach pancreatic cancers. The superior bio-stability of D-enantiomeric peptide plays a key role in the entire process. 17 ACS Paragon Plus Environment

Page 18 of 34

Page 19 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

TNBC, the most lethal subtype of breast cancer with a high rate of metastatic spread, is characterized by the lack of the three most common types of receptors associated with most breast cancers: growth-hormone epidermal growth factor receptor 2 (HER-2), estrogen receptor (ER), and progesterone receptor (PR). This has impeded targeted drug/gene delivery via these receptors.33 Furthermore, TNBC is highly aggressive, with a high risk of metastasis and hence high mortality.34,35 Currently, there is an urgent need for novel biomarkers for tracking and therapy. Following injection of MDA-MB-231-luc cells into nude mice via the tail vein, a murine metastatic lung cancer model of TNBC was successfully developed within 12 days of inoculation based on bioluminescence detection and pathological analysis (Figure S12). Ex vivo imaging of the distribution of D-FNB-Lip in the lungs as well as in the frozen lung tissue sections clearly indicated the potential of D-FNB peptide as a targeting moiety to mediate specific drug delivery to the breast cancer metastatic site in the lungs (Figure S13).

The therapeutic effect of PTX-loaded formulations against metastatic lung cancer of TNBC was subsequently evaluated in the developed mouse model. Bioluminescence imaging of model mice showed the distribution of MDA-MB-231-luc cells throughout the course of the treatment. Notably, fluorescence signals in the lungs of the D-FNB-Lip/PTX group gradually decreased (Figures 8A, and 8B), suggesting the effective diminishment of tumor cells at the metastatic site. In contrast, fluorescence signals in the saline-treated group were persistently strong. The average survival time of the untreated mice was 22.5 days, and that was prolonged to 30.5 days and 35 days with the treatment of liposomal PTX without D-FNB modification and Abraxane treatment, respectively. Notably, the average survival time of the D-FNB-Lip/PTX group, was more than 50 days (Figure 8C). During a period of 60 days, the mice treated with D-FNB-Lip/PTX survived significantly longer than the untreated mice (P < 18 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

0.001) and the other treatment groups (P < 0.05). Similar findings were observed by histopathological assessment, whereby fewer tumor foci were found in the lung tissue sections of the D-FNB-Lip/PTX group after 2-week treatment compared with the other treatment groups (Figures 8D and 8E). These results confirmed that the D-FNB peptide mediated liposomal PTX delivery can identify and attack metastatic breast cancer cells, thus resulting in significantly improved therapeutic effect against lung metastases of TNBC.

Figure 8. (Therapeutic effect of the PTX formulations in MDA-MB-231-luc metastatic model mice.) (A) In vivo bioluminescence imaging of the distribution of MDA-MB-231-luc cells in MDA-MB-231-luc cell lung metastasis model mice after the treatment. (B) Visualization of the distribution of MDA-MB-231luc cells in major organs after treatment with bioluminescence imaging. Note that no data was shown for the control group (untreated) in (A) & (B) as the animals did not survive at 28 days post treatment. (C) Kaplan-Meier survival curves for lung metastasis model mice treated with paclitaxel-loaded formulations (n = 10, *p < 0.05, **p < 0.01). (D) Histological characterization of lung metastatic cancer following treatment. H&E staining of lung tissue sections revealed carcinoma cells in different groups (yellow arrows indicate tumor foci). The scale bars are 100 µm. (E) The cancer burden conditions in the lungs excised from

19 ACS Paragon Plus Environment

Page 20 of 34

Page 21 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

the nude mice bearing MDA-MB-231 tumors after treatment. a) Saline (control group, the control mice die at day 28 d), b) mPEG-Lip/PTX group, c) Abraxane group, and d) D-FNB-Lip/PTX group.

With respect to active targeted drug delivery, the in vivo fate of targeted drug delivery systems and their therapeutic effects greatly depend on targeting moieties such as peptides. Through suitable design and/or screening, a single targeting peptide can be conferred with sufficiently powerful and even multiple functionalities, thus producing efficient peptidemediated drug carrier which is easy to be prepared and characterized while improving the prospects for clinical application. For example, iRGD, a peptide which interacts with both integrin and neuropilin-1 receptors, can sequentially mediate the tumor-specific tissuepenetrating and targeting of drug delivery systems, thus significantly improving therapeutic efficacy.36,37 In the current study, a D-enantiomeric Fn14-binding peptide successfully facilitated the penetration of liposomal vehicles through the stromal barrier while showing a good targeting capability towards pancreatic and metastatic breast cancer cells. With the increasing interest in Fn14 as a novel potential target for a variety of diseases, as well as given that several Fn14-targeting agents are currently undergoing clinical trials,38 this Fn14targeting strategy may be implemented in the treatment of various diseases in the future.

Although the CRD is frequently observed in the extracellular regions of membrane receptors, some membrane receptors may lack this domain. The concept of a scaffold-based protein design could therefore be utilized by first choosing a disulfide bond-stabilized miniature protein (with a conformation similar to that of the target domain) as the scaffold, followed by grafting the key residues of the target into the topological equivalent site of the scaffold. Such product may potentially be used to substitute the original receptors for ligand screening.27,39 The natural reservoir contains large quantities of disulfide-containing peptides with different structure motifs, thus providing solid supports for scaffold selection and optimization in the case that CRD is absent for some membrane receptors. It may be possible that the side chain 20 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

topology of certain retro-inverso peptides may not resemble that of their native counterparts.10,11 If this is the case, the retro-inverso isomerization strategy is not suitable. Nevertheless, mirror-image phage display represents a useful tool for screening D-peptide ligands that target all types of membrane receptor.

Conclusion A novel strategy utilizing CRD of membrane receptor for screening bio-stable D-peptide ligands via mirror-image phage display was developed. The obtained Fn14 D-peptide showed high binding affinity to its target membrane receptor Fn14, and successfully mediated drug delivery to target both pancreatic and breast cancer cells. More importantly, D-peptide mediated PTX liposomal delivery demonstrated superior therapeutic effect in treating pancreatic cancer and lung metastases of triple negative breast cancers in animals, compared to the commercial PTX product Abraxane. This research paved the way for the development of bio-stable D-peptide based active targeted drug delivery. Furthermore, the obtained Fn14 D-peptide can potentially be used to mediate targeted drug delivery for the treatment of a variety of cancers and human diseases with positive Fn14 expression.

Experimental Section Phage display and mirror-image phage display screening Fn14 binding peptide Phage display was performed according to the manufacturer’s instructions. In brief, protein Lobind tubes (Eppendorf, 0030108116) were pretreated by washing with 1 ml of filtersterilized blocking buffer containing 0.1 M NaHCO3 (pH 8.6) and 5 mg/ml BSA (bovine serum albumin) three times. The target protein, L-enantiomeric, N-terminally biotinylated Fn14 CRD, was preincubated with 100 µl of 1 × TBS (50 mM Tris-HCl, 150 mM NaCl, pH 7.5) and mixed with 10 µl of phage suspension from the Ph.D-12 Phage Display Peptide 21 ACS Paragon Plus Environment

Page 22 of 34

Page 23 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Library Kit. Next, the solution was pipetted into a suspension of streptavidin magnetic beads, which were immobilized and trapped using a magnetic separation rack. Beads were subsequently washed four times with TBS containing 0.1% Tween-20 (TBST). Competitive elution of bound phages was then performed by adding 50 µl of TWEAK, the natural ligand of the Fn14 receptor, and reacted for 30 mins at room temperature. The solution was collected and used for amplification of the eluted phages. Determination of the output titer and phage amplification was conducted according to the manufacturer’s instructions. Briefly, phage dilutions ranging from 102-107 were mixed with 100 µl of E. coli K12 ER2738 cells (should be early-log phase, OD600 0.01-0.05). The mixture was plated together with 3 ml of top agar per dilution on LB/IPTG/X-Gal Petri dishes (90 mm × 90 mm), and incubated overnight at 37°C. The next day, plaques were counted and the titer was determined. Phage amplification was conducted according to the manufacturer’s instructions and the input titer was measured analogically to the output titration at dilutions ranging from 108-1014. Following several rounds of panning, the specificity and affinity of single phage clones for the target protein was analyzed using ELISA. The peptide with the best binding affinity to Fn14 CRD was selected and named as L-FNB peptide.

In the case of mirror-image phage display, D-enantiomeric Fn14 CRD was used as the target protein. A protocol similar to that described above was carried out, except that the competitive eluent of bound phages was replaced with a D-enantiomeric peptide solution of the L-FNB peptide obtained using phage display technique described above. After five rounds of panning, one candidate L-peptide with the best possible binding the target D-Fn14 CRD protein was selected, and its D-enantiomer isomer was subsequently synthesized and named D-FNB peptide, which is expected to bind to the natural target L-protein/receptor for reasons of symmetry.

22 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Surface plasmon resonance validation and binding affinity determination The binding affinity of the obtained peptides to their target L-protein/receptor (L-Fn14 CRD) was validated by surface plasmon resonance (SPR) using a Biacore 3000 system (GE Healthcare, Chicago, USA). Briefly, biotinylated L-Fn14 CRD was immobilized on SA sensor chips using a biotin-streptavidin coupling method. Various concentrations (0-400 nM) of the peptides were run in a HBS-P buffer (0.01 M HEPES, 0.15 M NaCl, and 0.005% surfactant P20, pH 7.4). Kinetic data were fitted to a 1:1 Langmuir binding analyte model using a BIAevaluation software.40

Molecular docking study The structural mode of the Fn14 protein was obtained from www.pdb.org (pdb code: 2rpj). Missing atoms in the protein structure were added using the protein prepare package of the Schrodinger software. The structural models of L-FNB and D-FNB peptides were constructed and optimized in an OPLS2005 force field. The designed peptides were then docked to the Fn14 protein, and the binding energy predicted using the glide package in Schrodinger with default parameters.41-44 Following analysis of the resulting binding modes, the figures presented in this article were constructed using PyMOL software.

Liposome-cell interaction assay Real-time liposome-cell interactions were investigated at 25°C using Ligand Tracer Green (Ridgeview Instruments AB, Uppsala, Sweden).45 Briefly, Bxpc-3 (over-expressing the Fn14 receptor) and BT-474 cells (that do not express the Fn14 receptor) were seeded onto a local part of an inclined cell culture dish and cultivated until the cell number was approximately 1 × 106, respectively. All measurements were conducted in 3 ml of cell culture medium and commenced with a short baseline followed by a three-step uptake study, with different concentrations of C6-loaded liposomes. Next, a retention measurement was performed in 23 ACS Paragon Plus Environment

Page 24 of 34

Page 25 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

fresh cell culture medium. Each revolution resulted in a peak on the detector, which reflected the additional activity of liposomes bound to the cells. We could obtain an estimation of the kinetics of the interaction through the peak, which was followed over time. Data evaluation and estimation of the equilibrium dissociation constant KD were performed with TraceDrawer 1.3 software using a one-to-one binding model.

Competitive cellular uptake inhibition of the liposomes in Bxpc-3 cells was also studied. Briefly, Bxpc-3 cells were seeded in 12-well cell culture plates as descried in the Cellular Uptake Study in Supporting Information. After culturing for 18 h, the Bxpc-3 cells were pre-incubated with excess amount of free D-FNB peptide or TWEAK for 30 mins. The cells were then treated with rhodamine B (RhB) loaded liposomes (i.e. mPEG-Lip/RhB and DFNB-Lip/RbB), respectively. Cells without pretreatment with either D-FNB peptide or TWEAK, as well as cells treated with PBS or free RhB solution were used as controls. After 2 h incubation, the cells were washed with cold PBS, trypsinized, and washed three times by centrifugation and resuspended in PBS. Mean fluorescence intensity of the samples was measured by flow cytometry.46

Establishment of Bxpc-3/NIH 3T3 subcutaneous tumor model and MDA-MB-231-luc metastatic lung cancer model in nude mice Briefly, to establish the xenograft subcutaneous tumor models, 4 × 106 Bxpc-3 cells and 2 × 106 NIH 3T3 cells in 100 µl of PBS with Matrigel (BD Biosciences, CA) at a ratio of 1:1 (v/v) were subcutaneously co-injected into the right flank of nude mice.30 Tumor volume was measured every 2 days (starting from the eighth day after tumor cell inoculation), and was calculated as V(mm3) = 0.5 × (longest diameter) × (shortest diameter)2. When the tumor volume reached approximately 100 mm3, the animals were used for further studies. The lung metastatic animal model of MDA-MB-231-luc tumor cells, transfected with a plasmid24 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

carrying firefly luciferase gene, was established in nude mice by tail vein inoculation of 1 × 106 tumor cells.47 Tumor growth in the lungs was visualized by bioluminescence imaging, which was performed 10 mins after the intraperitoneal administration of 200 µl of D-luciferin (15 mg/ml), using an in vivo imaging system (Xenogen IVIS Spectrum, Perkin Elmer, USA).

Tumor targeting and penetration effect studies Bxpc-3/NIH 3T3 tumor model mice were randomly assigned to two groups. When the tumor volume reached approximately 100 mm3, the animals were treated with 0.2 ml of DiR loaded liposomes (i.e. mPEG-Lip/DiR and D-FNB-Lip/DiR) (DiR concentration: 5 µg/ml) via tail vein injection, respectively. At predetermined time intervals, the mice were anesthetized with isoflurane, and the biodistribution of the liposomes was studied using an in vivo imaging system. 24 h post administration, the mice were sacrificed and tumors and tissues (i.e. hearts, lungs, spleens, livers, and kidneys) were collected. The ex vivo fluorescent signal of the tumors and organs was also detected. For the MDA-MB-231-luc animal model, we using ex vivo imaging to detect the fluorescent signal of organs 4 h and 8 h after the i.v. injection of DiR loaded liposomes (i.e. mPEG-Lip/DiR and D-FNB-Lip/DiR).

The tumor penetration of D-FNB modified liposomes was evaluated using Bxpc-3/NIH 3T3 tumor and 3D tumor spheroid models, respectively. Briefly, nude mice bearing Bxpc-3/NIH 3T3 tumors were treated with free RhB, mPEG-Lip/RhB, and D-FNB-Lip/RhB. After 4 h, mice were sacrificed and excised tumors were frozen in optimum cutting temperature medium at -80°C. Corresponding tumor slices (0.7 mm) were prepared, air-dried for 10 min, and fixed with 4% paraformaldehyde for 10 mins. The samples were analyzed using confocal microscopy, and five random fields were selected to measure the penetration depth of RhB.48 In addition, ex vivo multicellular core-shell 3D tumor spheroids mimicking the single tumor nest surrounded by fibroblasts, were developed using the lipid overlay system with some 25 ACS Paragon Plus Environment

Page 26 of 34

Page 27 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

modifications.49 Briefly, Bxpc-3 cells (1 × 104) were seeded into each well of the ultralow attachment round bottom 96 well plates (Costar, Corning, NY), then clustered to the bottom of the wells by centrifugation at 900 rpm for 2 mins. The plates were placed in incubator at 37°C for 4 days and the culture medium were replaced every 2 days. On the 4th day, NIH 3T3 (1 × 104) cells were added into each well, followed by continuous gentle rotation (200 rpm, 37°C) for 6 h. Then the formed core-shell tumor spheroids were incubated for another 2 days. For the time-lapse assay, mPEG-Lip/C6 and D-FNB-Lip/C6 were added into different wells, and fluorescence images were acquired at predetermined time points approximately 150 µm from the bottom of the spheroid.

In vivo antitumor efficacy study Bxpc-3/NIH 3T3 subcutaneous tumor bearing nude mice (n = 24) were randomized into four groups (n = 6 per group): saline (control), Abraxane (positive control), mPEG-Lip/PTX, and D-FNB-Lip/PTX. Treatment was initiated when the tumor volume reaches approximately 100 mm3, the above testing formulations were administrated to the animals at a PTX equivalent dose of 8 mg/kg body weight via the tail vein once every 4 days for six injections.50,51 Body weights and tumor sizes were monitored until mice were sacrificed. Tumors were excised and a portion of them were fixed in 10% formalin, paraffin-embedded and sectioned for terminal dexynucleotidyl transferase (TdT)-mediated dUTP nick end labeling (TUNEL) assay. Another portion of the tumors were frozen-sectioned for α-SMA immunofluorescence assay.

The metastatic MDA-MB-231-luc tumor model was used as the TNBC metastatic lung cancer animal model. After the mice developed metastatic lung cancer, the animals were randomly divided into four groups and treated with: saline (negative control), Abraxane (positive control), mPEG-Lip/PTX, and D-FNB-Lip/PTX (PTX dose: 10 mg/kg, twice a week for two weeks).52 Tumor nodules in the lungs were visualized by bioluminescence and analyzed by 26 ACS Paragon Plus Environment

Bioconjugate Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

hematoxylin and eosin (H&E) staining. The lungs obtained from mice in each group were fixed in Bouin’s fixative for 24 h, and the tumor burdens on the lungs were then observed and recorded.

Immunohistochemistry analysis Nude mice bearing MDA-MB-231-luc metastatic lung cancer were treated with mPEGLip/C6 and D-FNB-Lip/C6. Nude mice bearing Bxpc-3/NIH 3T3 tumor were treated with mPEG-Lip/C6, D-FNB-Lip/C6 and D-FNBS-Lip/C6. After 8 h, the mice were sacrificed, and the tumors and lungs were excised under dark conditions. Frozen tissue sections of tumors or lungs were used for single or double immunofluorescence labeling. In addition, an antibody against α-SMA was used for the identification of active fibroblasts in the tumor stroma. An antibody against Fn14 was used for labeling tumor cells.

Statistical analysis Results were expressed as mean ± standard deviation (SD). A student’s t-test was used for comparing two groups, and one-way analysis of variance (ANOVA) was used for multiple group comparisons. P values < 0.05 and