Annexin A1‐Targeted d‐Peptide‐Monomethyl Auristatin E Conjugate Enhanced Antitumor Effect of Monomethyl Auristatin E for Chemotherapy of Prostate Cancer

ABSTRACT

Background

N‐terminus 15 amino acid (MC16) of Annexin A1 (ANXA1) is an effective therapeutic target for tumors and tumor vasculature as it specifically localizes to the cell surface of tumor cells and tumor vascular endothelial cells. We identified the d‐type peptide (hpnevrs, dhp7) by mirror‐image phage display targeting l‐type MC16 and tested the hypothesis that intravenously injected c(vcMMAE)dhp7 eradicates prostate tumor in mice with less adverse side effects than that of mc‐Val‐Cit‐PAB‐monomethyl auristatin (vcMMAE) alone.

Methods

The binding affinity of dhp7 to MC16 of ANXA1 was analyzed by biolayer interferometry and in silico conformational analysis. c(vcMMAE)dhp7 was synthesized by coupling dhp7 to vcMMAE to evaluate the antitumor effects. A subcutaneous tumor model was generated by inoculating the luciferase‐expressing prostate cancer PC‐3 cell line, PC‐3‐hLuc‐PSMA, into the nude mice. 1.25 mg/kg vcMMAE or 2.14 mg/kg c(vcMMAE)dhp7, the equal dose as 1.25 mg/kg vcMMAE, was administered intravenously every 3 or 4 days and growth of tumors was assessed by photon count using an IVIS system.

Results

The dhp7 bound specifically to l‐type MC16 (KD 0.48 μM) but not bound to d‐type MC16 and mutant l‐type MC16. Both in vitro and in silico interaction analysis indicating the α‐helix of the l‐type MC16 may be important for dhp7 binding. Fluorescence labeled‐dhp7 accumulated in both ANXA1 positive cells and subcutaneous prostate tumor. To evaluate the antitumor effect, c(vcMMAE)dhp7 was synthesized by coupling dhp7 with mc‐Val‐Cit‐PAB‐monomethyl auristatin E (vcMMAE) as a payload. The IC50 s of vcMMAE (129.9 ng) and c(vcMMAE)dhp7 (122.7 ng) against prostate cancer cells showed almost equivalent in vitro drug sensitivity. The c(vcMMAE)dhp7‐treated mice showed significant antitumor effects including complete responses at the same dose of MMAE (1.25 mg/kg) as mice treated with vcMMAE alone. Pathological analysis of the residual tumor showed that CD31 density of the c(vcMMAE)dhp7 treated group was significantly higher than that of the vcMMAE group, and the Ki‐67 index was significantly lower than that of the vcMMAE alone. Clinical chemistry tests showed side effects of MMAE significantly improved in terms of hepatic/biliary toxicity in c(vcMMAE)dhp7‐treated group.

Conclusions

The c(vcMMAE)dhp7 conjugate exerts a better antitumor effect than vcMMAE on ANXA1‐positive prostate cancer.

Abbreviations

ADC

Antibody‐drug conjugate

Alb

Albumin

ALT

Alanine aminotransferase

ANXA1

Annexin A1

AST

Aspartate aminotransferase

BUN

Blood urea nitrogen

Cre

Creatinine

Glu

Glucose

HCT

Hematocrit

HGB

Hemoglobin count

LDH

Lactate dehydrogenase.

MCH

Mean corpuscular hemoglobin

MCHC

Mean corpuscular hemoglobin concentration

MCV

Mean corpuscular volume

MI‐Cy7.5

Maleimide‐cyanine 7.5

MI‐FAM

Maleimide‐fluorescein

PDC

Peptide‐drug conjugate

PLT

Platelet count

PSMA

Prostate‐specific membrane antigen

RBC

Red blood cell count

TG

Triglyceride

TMR

5‐carboxytetramethylrhodamine

T‐Bil

Total bilirubin

T‐Cho

Total cholesterol

T‐Pro

total protein

WBC

White blood cell count

1. Introduction

Annexin A1 (ANXA1) was identified as a specific surface protein for malignant tumor vasculature using subtractive proteomics analysis of malignant versus normal vasculature [1, 2, 3]. Previously, Fukuda et al. identified IFLLWQR (IF7) peptide which bound to the N‐terminal 1–15 residues (MC16) of ANXA1 by phage display technique, and showed that intravenously injected IF7 accumulated tumor tissue due to IF7 binds to the ANXA1 on the surface of endothelial cell and crossed tumor endothelial cells by transcytosis [4, 5, 6, 7, 8, 9]. Taking advantage of the above functions, IF7 was used to a drug delivery system targeting the tumor vasculature. Specifically, intravenous injection of IF7‐SN‐38 conjugate suppressed the growth of colon cancer in mice at low doses without side effects [8]. Nonaka et al. also reported that intravenous injection of IF7‐SN‐38 conjugate into glioma and melanoma model efficiently reduced the size of those tumors, suggesting that IF7‐conjugated anticancer drug might be able to overcome the blood‐brain barrier [9]. Although IF7 was specifically targeting tumor vasculature and suitable drug delivery system for anticancer drug, the IF7 itself is susceptible to proteases in vivo, making IF7‐anticancer drug conjugate less stable [8]. Many researchers have addressed stability issues in vivo by designing d‐peptides, retroinverso‐type peptides, and cyclizing l‐peptides [10, 11]. A further problem is that IF7 is highly hydrophobic, making dissolving the IF7‐anticancer drug conjugate difficult. Therefore, to develop a protease‐resistant and less hydrophobic IF7 that retains ANXA1‐binding function, we try to screen d‐type peptide against the chemically synthesized MC16 peptide of human ANXA1 by mirror‐image phage display screening [9, 12, 13, 14, 15]. We identified that the 7‐mer d‐type peptide (hpnevrs, dhp7) bound to the MC16. The mc‐Val‐Cit‐PAB‐monomethyl auristatin E (vcMMAE) has been used as a candidate payload for antibody‐drug conjugates (ADCs) and is therefore readily applicable to peptide‐drug conjugates (PDCs).

Prostate cancer (PCa) is the most commonly diagnosed malignancy in men globally and its mortality is also highest in United States, Western Europe, and Australia [16]. In particular, approximately 10% to 30% of PCa patients develop castration‐resistant PCa (CRPC) [17, 18], which is a lethal disease, although it is treated with taxane‐based chemotherapy and prostate‐specific membrane antigen (PSMA)‐targeting theranostics. Taxane‐based anticancer drugs for CRPC exert potent antitumor effects, but systemic side effects are a major problem. PSMA‐targeting theranostics is one of the PCa‐targeted therapies with favorable therapeutic outcomes reported [19], but treating PSMA‐negative CRPC is associated with an unfavorable response to PSMA theranostics [20, 21]. Therefore, there is an unmet need about PSMA‐independent cancer‐specific chemotherapy such as ANXA1 targeted PDC strategy.

In this study, we synthesized c(vcMMAE)dhp7 conjugating dhp7 to vcMMAE via the side chain of N‐terminal d‐cys residue and aimed to evaluate its targeting anticancer effect as a DDS of dhp7 in ANXA1 positive CRPC cell line derived tumor‐bearing mice.

2. Methods

2.1. Materials

MC16 peptide which includes the N‐terminal fifteen residues of human ANXA1 adding a cysteine residue on the C‐terminal side (MAMVSEFLKQAWFIEC) was chemically synthesized as a d‐amino acid type (d‐MC16) and l‐amino acid type (l‐MC16) by Biologica Co. Ltd. (Nagoya, Japan). Mouse anti‐MC16 antibody that recognized MC16 of ANXA1 was gifted from Prof. Motohiro Nonaka at Kyoto University. Anti‐PSMA (D718E), anti‐mouse CD31 (D8V9E), and anti‐Ki‐67 antigen (D3B5), anti‐His‐Tag (D3I10) rabbit monoclonal antibodies (mAbs) were from Cell Signaling Technology (Danvers, MA, USA).

2.2. Mirror‐Image Phage Display Screening for MC16 Binding d‐Type Peptide

Mirror‐image phage display protocol for screening of MC16 binding d‐amino acid peptide was shown in Figure 1A. The l‐cysteine (10 nmol/well) or the d‐MC16 peptide (10 nmol to 0.01 nmol/well) was coated on maleimide‐activated plates (Corning Inc. Corning, NY, USA) at 4°C for 20 h. Then, blocked with 200 μL of SuperBlock solution (Thermo Fisher Scientific Inc. Waltham, MA, USA) for 1 h. Then, 200 μL of a fully random 7‐mer l‐amino acid peptide library displaying T7 phage (3.21 × 10¹² diversity) as prepared previously [22] was added to l‐cysteine coated plate and incubated for 1 h at 37°C to deplete background binding phage. This biopanning procedure consists of association phase of the peptide library displaying T7 phage to the d‐MC16 peptide coated well, washing phase, elution phase, and amplification phase of the d‐MC16 binding phage. In brief, association phase from 1st to 4th round biopanning, phages were incubated for 5 min at 37°C with 10 nmol, 1 nmol (high stringency), 0.1 nmol (very high stringency), and 0.01 nmol (super very high stringency) of d‐MC16 coated wells, respectively. Unbound phage was washed six times with phosphate‐buffered saline (PBS). In elution phase, d‐MC16 bound phages were eluted with PBS containing 1% SDS for 5 min at 37°C. In amplification phase, the d‐MC16 bound phages were amplified with logarithmic E. coli. BL21 for 3 h at 37°C. At the final 5th round biopanning, the inoculated 4th round phage was applied to 10 nmol l‐MC16 peptide coated well to subtract l‐MC16 bound phage. Then, following procedure was conducted as same as 1st round biopanning. The genome copies per μL (gc/μL) of each round input and output T7 phage was quantified by droplet digital PCR as described previously [23] (Figure 1B) and all reagents and equipment for ddPCR were from Bio‐Rad Laboratories (Hercules, CA, USA). The primers and probe for T7 phage DNA were forward 5′‐TGGATGGGATAACTGGTAAGC‐3′, reverse 5′‐TGGTTCTTAGTGTGGATGTCG‐3′ and probe 5′‐FAM/TGGCTCACTTCATGGCTCGCTTT‐3′. A standard curve of T7 phage genome copies (gc) was prepared by serially diluted from 1 fg/μL to 10⁷ fg/μL T7 packaging control double‐stranded DNA (37314 bp, 0.1 μg/μL, Merck Millipore, MA, USA) as a template of ddPCR. To convert phage standard DNA concentration from fg/μL to gc/μL, the following equation was used: [gc/μL] = [dsDNA g/μL]/[DNA size (bp) ∗ 607.4 + 157.9] ∗ (6.02∗10²³) [23]. Triplicate assays were used for each sample.

Figure 1.

Schematic representation and results of mirror‐image phage display screen for MC16 binding d‐amino acid peptide. A, A phage library displaying l‐type peptides is represented as red and after fifth round screening is then chemically synthesized d‐type peptides are represented as blue. Chemically synthesized d‐type MC16 peptide is represented as a magenta arrow and l‐type MC16 peptide is represented as a cyan arrow. B, Output genome copy per μL of d‐type MC16 binding phage obtained after each round of pooled phage clones. C, Enriched sequences presented by representative d‐type MC16‐binding phage clones. D, Enrichment ratio of top 10 enriched peptide sequences in 5th round. All peptide sequences are listed in Table S1. [Color figure can be viewed at wileyonlinelibrary.com]

2.3. Next‐Generation Sequencing (NGS) and Data Analysis

The displayed peptide sequence of phage was determined by next‐generation sequencing using an iSeq. 100 system (Illumina Inc. San Diego CA, USA) as described previously [24]. Sequencing was performed with a single‐read mode and a read length of 150 bp. Raw Fastq files generated from the sequencing runs were subjected to quality control evaluation using FastQC. Subsequently, the DNA sequences corresponding to the T7 phage capsid‐coding region were identified. The sequences directly downstream of the capsid region were translated into amino acid sequences up to the first in‐frame stop codon. The capsid protein sequences were removed, leaving only the displayed peptide sequences for further analysis. To evaluate the diversity and copy number of displayed peptides, the amino acid sequences were grouped, and the frequency of each peptide was calculated. These peptides were then sorted in descending order of occurrence to find frequently observed sequences (Table S1). Data processing and analysis were performed using Python 3.9.12, with Biopython employed for sequence manipulation and translation. Top‐enriched amino acid sequences HPNEVRS (HP7) and the next two high‐enriched peptides QYATNLK (QY7) and STSRNTL (ST7) were chemically synthesized using d‐amino acid (denoted by lower case letter) (Figure 1C,D).

2.4. The Binding Affinity Analysis Between the Biotinylated d‐Peptide Candidate and MC16

Binding affinity of MC16 to d‐peptide candidates were performed by biolayer interferometry (BLI) on a FortéBio Octet K2 (Sartorius, Gottingen, Germany). N‐terminal biotinylated d‐peptide (dhp7, dqy7, dst7) was synthesized by Biologica Co. Ltd., and 50 μg/mL biotinylated d‐peptides were immobilized to Octet super streptavidin (SSA) biosensor (Sartorius) (Figure 2A,B,C,D). 50 μg/mL biocytin immobilized to biosensor as a reference. The l‐MC16 or d‐MC16 or l‐MC16 mutant (E6A, F7A, K9A, Q10A, W12A, F13A, E15A) peptides (Figure 2E) were serially diluted (100, 50, 25, 12.5, 6.25, 3.12, and 1.56 μM) in kinetic buffer. Then, the biosensor was dipped into each MC16 dilution series for 120 s, followed by dissociation for 120 s at 37°C. The biosensor was regenerated with a 0.1 mM glycine HCl (pH 2.0) solution and pulsed 3 times for 5 s each. The equilibrium dissociation constant (KD) values were calculated from the data fitted to a 1:1 binding model from which K _on and K _off values using Octet Data analysis HT software Version 10 (Sartorius) (Figure 2A,B,C,D,F).

Figure 2.

The binding affinity of enriched d‐peptides to l‐MC16 and d‐MC16 analyzed by biolayer interferometry. A, dhp7 versus l‐MC16 B, dhp7 versus d‐MC16. C, dqy7 versus l‐MC16. D, dst7 versus l‐MC16. E, dhp7 versus l‐MC16 mutants. Binding kinetics using several concentrations and 1:1 model with global fitting for determination of KD value. F, Iso‐affinity Graph of dhp7 versus l‐MC‐16 WT or l‐MC16 mutants. G, In silico docking analysis of the ANXA1 N‐terminus and dhp7. The left image indicated a predicted interaction summary between MC16 and dhp7. The right image indicated a predicted whole docking model. H, Calculated mutation energy for binding and stability between point mutants of MC16 and dhp7 docking. [Color figure can be viewed at wileyonlinelibrary.com]

2.5. In Silico Docking Analysis Between the ANXA1 and dhp7

The predicted structure of dhp7 peptide was constructed using Biovia Discovery Studio 2019 (Dassault Systems, Paris, France) and optimized by minimization and preparation using the CHARMM force field. The ANXA1 structure was presented by Nonaka et al [15]. The Dock Ligand LibDock program in Discovery Studio was employed to generate a predicted structure for the ANXA1 and dhp7 complex using the predicted structure of the monomeric ANXA1 and the predicted structure of dhp7 peptide (Figure 2G). The N‐terminal 14 amino acid residues (Ala2 to Glu15) of ANXA1 was allowed to interact with dhp7 peptide. The resulting ten candidate structures with higher LibDock scores were investigated using the BLI results. Using the Calculate Mutation Energy function of Discovery Studio, the N‐terminal amino acid residues of ANXA1 were screened by alanine scanning (Figure 2H). The structures were generated using a Biovia Discovery Studio Visualizer.

2.6. Preparation of dhp7 Peptide Conjugates

All dhp7 conjugates were synthesized by the Peptide Institute Inc. (Osaka, Japan). The C(MI‐FAM)‐vc‐dhp7 for intracellular uptake study of dhp7, the fluorescein‐5‐maleimide (MI‐FAM, Thermo Fisher Scientific Inc. MA, USA) was conjugated to via the side chain of the Cys residue of Cys‐vc‐dhp7. The purity of C(MI‐FAM)‐vc‐dhp7 was 99.4% (Figure 3A). The C(MI‐Cy7.5)‐vc‐dhp7 for the biodistribution study, Cy7.5‐maleimide (MI‐Cy7.5, MedChemExpress, NJ, USA) was conjugated to via the side chain of the Cys residue of Cys‐vc‐dhp7. The purity of C(MI‐Cy7.5)‐vc‐dhp7 was 98.3% (Figure 3B). The c(vcMMAE)dhp7 for the tumor targeting chemotherapy, the mc‐Val‐Cit‐PAB‐MMAE (vcMMAE, MedChemExpress) as a payload for PDCs was conjugated to via the side chain of the dCys residue of dhp7. The purity of c(vcMMAE)dhp7 was 99.4% (Figure 3C).

Figure 3.

Structure of dhp7 conjugates. Shown are chemical structures and chromatograms of dhp7 conjugates. A, C(MI‐FAM)‐vc‐dhp7. B, C(MI‐Cy7.5)‐vc‐dhp7. C, c(vcMMAE)dhp7.

2.7. In Vitro dhp7 Peptide Intracellular Uptake of ANXA1‐Expressed 293 T Cells

Human embryo kidney epithelial‐like cell line, 293 T (RRID: CVCL_0063, ATCC, CRL‐3216) was maintained in DMEM high glucose with 10% fetal bovine serum (FBS). ANXA1 gene stably expressed 293 T (293T‐ANXA1‐His) cells were established by transfection of pCMV3‐ANXA1‐His plasmid (HG‐12113‐CH, Sino Biologicals Inc., Rockville, MD, USA) to the 293 T cells with 100 μg/mL hygromycin selection. To confirm ANXA1‐His expression, 293 T and 293T‐ANXA1‐His cells were fixed with BD Phosflow Fix buffer I (BD biosciences, NJ, USA) and incubated with anti‐MC16 (1 μg/mL) or anti‐His‐Tag (1:1000 dilution) antibody at 25°C for 1 h, followed by Alexa488‐conjugated goat anti‐mouse IgG (A11001, Thermo Fisher Scientific) or APC‐conjugated goat anti‐rabbit IgG (A‐10931, Thermo Fisher Scientific) at 25°C for 1 h. Then mounted by ProLong Diamond Antifade Mountant with DAPI (Thermo Fisher Scientific). To evaluate peptide intracellular uptake study, 31.05 μg/100 μL of C(MI‐FAM)‐vc‐dhp7 or 7.44 μg/100 μL FAM) with 50 nM Lysotracker Red (Thermo Fisher Scientific) was added to 293T‐ANXA1‐His or parent 293 T cells and incubated at 37°C for 5 min. Then the cells were immediately fixed with 4% formalin for 15 min. Fluorescence images (40× objective) were captured using a BZ‐9000 microscope with BZ‐II analyzer V2.2 software (Keyence, Tokyo, Japan) and intracellular uptaked particles were analyzed by the Fiji platform (ImageJ distribution, http://fiji.sc/Fiji).

2.8. Biodistribution of C(MI‐Cy7.5)‐vc‐dhp7

Human androgen‐independent prostate adenocarcinoma cell line, PC‐3 (RRID: CVCL_0035, ATCC, CRL‐1435), was maintained in RPMI1640 with 10% FBS. To detect tumor by in vivo imaging system (IVIS lumina XRMS, Sumitomo Pharma Co. Ltd., Osaka, Japan), pReceiver‐Lv217 PSMA ORF‐hluc plasmid packaged lentiviral particles (LPP‐G0050‐Lv217, GeneCopoeia Inc., MD, USA) were transduced into the PC‐3 cells with 2 μg/mL puromycin selection and prostate‐specific membrane antigen (PSMA) gene and firefly luciferase (hLuc) gene stably expressed PC‐3 (PC‐3‐hLuc‐PSMA) cells were established. All animal studies were carried out in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Hirosaki University Graduate School of Medicine Animal Care and Use Committee (permit numbers: AE01‐2024‐005). 8‐week‐old BALBc/AJcl nu/nu male mice were purchased from CLEA Japan and were bred under specific pathogen‐free conditions in the Hirosaki University Animal Core Facility. The 2 × 10⁶ of PC‐3‐hLuc‐PSMA cells suspended in 20 μL of RPMI1640 plus Matrigel (Corning Inc. NY, USA) 1:1 mixture were injected subcutaneously using a 33‐guage neuros 1700 syringe (Hamilton, NV, USA) into the dorsal side of the upper hind limb of nude mice [25] as a location where the accumulation of fluorescently labeled dhp7 in the tumor area would not overlap with other organs during in vivo imaging. When photon number of PC‐3‐hLuc‐PSMA tumor reached 5 × 10⁴, The C(MI‐Cy7.5)‐vc‐dhp7 (37.36 μg/100 μL) were intravenously administered, respectively. After intravenous injection, Cy7.5 intensity (Radiant Efficiency [p/s/cm²/sr]/[μW/cm²]) was monitored from 15 min to 24 h after injection using IVIS. After 24 h, all mice were euthanized by cervical dislocation under 5% isoflurane for induction and 2% isoflurane for maintenance anesthesia, and accumulation of Cy7.5 fluorescence in excised tumors and organs was monitored by ex vivo imaging.

2.9. Cell Surface MC16 Expression of Cancer Cell Lines by Flow Cytometry

Mouse bladder transitional cell carcinoma cell, MBT2 (JCRB, IFO50041), and human osteosarcoma cell, MG‐63 (RRID: CVCL_0426, JCRB, IFO50108), was maintained in EMEM with 10% FBS. Human squamous carcinoma of the tongue cell, SAS (RRID: CVCL_1675, JCRB, JCRB0260) was maintained in 45% DMEM/45% Ham's F12 with 10% FBS. All human cell lines have been authenticated by STR‐PCR method performed by JCRB cell bank and all experiments were performed with mycoplasma‐free cells. Monodispersed MBT2, MG‐63, SAS, PC‐3‐hLuc‐PSMA, 293 T, and 293T‐ANXA1‐His cell lines were obtained using TrypLE Express Enzyme (1×) (Thermo Fisher Scientific) and were incubated with anti‐MC16 antibody at 4°C for 1 h, followed by APC‐conjugated goat anti‐mouse IgG (A‐865, Thermo Fisher Scientific) at 4°C for 1 h under non‐permeable conditions. Cell surface MC16 expression of each cell line was analyzed by FACSAria II cell sorter (BD Biosciences). For controls, the primary antibody was omitted. All flow cytometry data were gated by monodispersed live cells, and histograms of MC16 expression were created by KALUZA V2.1.2 software (Beckman Coulter, CA, USA).

2.10. In Vitro Drug Sensitivity of PC‐3‐hLuc‐PSMA Cells Against vcMMAE or c(vcMMAE)dhp7

PC‐3‐hLuc‐PSMA cells were cultured with vcMMAE or c(vcMMAE)dhp7 as vcMMAE concentration ranging from 0 to 12500 ng/100 μL for 48 h, and viability was assessed using Cell counting kit‐8 (Dojindo Laboratory, Kumamoto, Japan). The IC50 of each drug conjugate was determined using GraphPad Prism software.

2.11. Treatment of PC‐3‐hLuc‐PSMA Tumor‐Bearing Nude Mice With C(vcMMAE)Dhp7

The subcutaneous PC‐3‐hLuc‐PSMA tumor‐bearing mice were prepared as same as biodistribution experiment. When photon count of PC‐3‐hLuc‐PSMA tumor reached 5 × 10⁴, the mice were randomly divided into vcMMAE and c(vcMMAE)dhp7 treatment groups of six mice each. A total of four courses of 1.25 mg/kg vcMMAE or 2.14 mg/kg c(vcMMAE)dhp7, the same dose as 1.25 mg/kg vcMMAE, were administered intravenously at D0, D3, D7, and D10, and body weight was weighed after each administration. Tumor size was monitored by photon count (Average Radiance [p/s/cm²/sr]) by IVIS. After four courses of treatment, all mice were euthanized by cervical dislocation under 5% isoflurane for induction and 2% isoflurane for maintenance anesthesia and total 500 μL of blood was collected in BD microtainer EDTA/2Na tube (Becton, Dickinson and Company, New Jersey, USA) by cardiac blood collection. Blood count tests were performed using pocH^TM‐100iV Diff (Sysmex Corporation, Hyogo, Japan) and clinical biochemistry tests were performed using SPOTCHEM^TM EZ SP‐4430 (ARKRAY Inc., Kyoto, Japan) with SPOTCHEM II reagent strips (PANEL Liver‐2, Kidney‐2, Glu, T‐Cho, TG, ARKRAY Inc.). The excised tumors were weighed and fixed in 20% formalin solution for immunohistochemical analysis. No mice met the euthanasia criteria of either tumor burden > 10% of body weight or tumor size > 20 mm in any one dimension [26].

2.12. Histopathological and Immunohistochemical Analysis of Residual PC‐3‐hLuc‐PSMA Tumors

Four µm‐thick sections of formalin‐fixed paraffin‐embedded tumor tissues were mounted on silane‐coated glass slides and air‐dried for 1 h. The sections were immersed in the de‐paraffinized antigen retrieval solution pH 6 (Nichirei Biosciences Inc. Tokyo, Japan) and autoclaved at 121°C for 20 min for heat‐induced antigen retrieval. The sections were then incubated with anti‐Ki‐67 rabbit mAb (D3B5, 1:800 dilution) or anti‐MC16 mouse mAb (1:100 dilution, 1 μg/mL) or anti‐PSMA rabbit mAb (D718E, 1:100 dilution) or anti‐CD31 rabbit mAb (D8V9E, 1:200 dilution) in PBS containing 5% bovine serum albumin at 4°C overnight. The diaminobenzidine (DAB) staining following 1st antibody detection was used to the Envision/HRP rabbit mouse kit (Agilent Technologies Japan, Tokyo, Japan) according to the manufacturer's instructions. Nuclear counterstaining was incubated in sections with hematoxylin solution (Dako North America Inc. CA, USA) for 2 min at room temperature. Hematoxylin/Eosin (HE) staining was performed using hematoxylin solution and Eosin alcohol acid extract solution (Fujifilm Wako Pure Chemical Corporation Ltd.). DAB‐ or HE‐stained whole tumor section images (10× objective) were captured using a BZ‐9000 microscope with BZ‐II analyzer V2.2 software (Keyence). The white balance was adjusted for each section. The DAB‐stained cells and hematoxylin‐stained nuclei in.tif format file of each whole tumor section images were counted using color deconvolution and the particle analysis plug‐in of the Fiji platform (ImageJ distribution, RRID: SCR_003070 http://fiji.sc/Fiji). The mean number of CD31‐positive blood vessels density per tumor area and ANXA1 or PSMA staining intensity per tumor cell in each treated group's whole tumor section was used for statistical analysis. According to the recommendations of the International Ki‐67 in Breast Cancer Working Group [27], The number of Ki‐67‐positive nuclei regardless of staining intensity was counted and divided by the sum of hematoxylin‐stained nuclei, and Ki‐67 index were expressed as a percentage.

2.13. Statistical Analysis

All statistical calculations were performed using Graphpad Prism 10 ver.10.4.0 (GraphPad, CA, USA). The nonparametric Mann–Whitney U‐test was used to analyze non‐normally distributed intergroup and Kruskal‐Wallis and Dunn's post hoc test was used to multiple‐group differences. p values less than 0.05 were defined as significant. All results of animal experiments were analyzed by intention‐to‐treat analysis.

3. Results

3.1. ANXA1 Binding d‐Peptides Identified by a Mirror‐Image Phage Display

We enriched several phage clones that bound to MC16 under reaction conditions at 37°C for 5 min using a mirror‐image phage display (Figure 1A,B,C). The most enriched amino acid sequence HPNEVRS (HP7) and the next two high‐enriched peptides QYATNLK (QY7) and STSRNTL (ST7) were chemically synthesized using d‐amino acid (denoted by lower case letter) (Figure 1D).

3.2. Binding Affinity of dhp7 to MC16 In Vitro

SSA biosensor immobilized with N‐terminal biotinylated d‐peptide and then dipped into various concentrations of L‐ or D‐ MC16 solution by BLI analysis. The dhp7 bound to l‐MC16, with a KD of 0.48 μM (Figure 2A) but did not detect any signals against d‐MC16 (Figure 2B), and higher affinity than those of another candidate d‐peptides dqy7 (0.99 μM) and dst7 (0.72 μM) with l‐MC16 (Figure 2C,D). Therefore, we selected the most enriched and highest affinity sequence dhp7 against l‐MC16 and used it following the binding specificity experiment between dhp7 and mutant forms of l‐MC16. The dhp7 binding affinity to l‐MC16 point mutants E6A, F7A, K9A, Q10A, W12A, F13A and E15A was much lower than to l‐MC16WT (Figure 2E,F). An In Silico docking analysis on stability and binding energy between the point mutation in MC16 of ANXA1 and dhp7 was also performed. The Glu6, Gln10, and Phe13 residues of MC16 were predicted to be located close together in the α‐helix structure and to interact primarily with the N‐terminal d‐His residue of dhp7 (Figure 2G and additional movie file 1). Similar to the BLI results (Figure 2E,F), the calculated MC16 point mutations in Glu6Ala (E6A), Gln10Ala (Q10A), and Phe13Ala (F13A) would have negative effects on stability and binding (Figure 2H) of dhp7, suggesting the α‐helix structure of MC16 is important for dhp7 binding and stability.

3.3. dhp7 Targeting of ANXA1 Positive Cells

To examine the ANXA1‐targeting activity of dhp7 by immunofluorescence, we confirmed ANXA1 expression of parental 293 T and 293T‐ANXA1‐His cells. The immunofluorescence results in permeable and fixed conditions showed weak endogenous MC16 expression in parental 293 T cells, and significantly higher green fluorescence intensity corresponding to MC16 and also higher magenta fluorescence intensity corresponding to His‐Tag in 293T‐ANXA1‐His cells (Figure 4A). Thus, we confirmed MC16 and His‐Tag expression of 293T‐ANXA1‐His cells were significantly higher than that of parent 293 T cells (Figure 4A). We also tested cell surface MC16 expression level of both parental 293 T and 293T‐ANXA1‐His cells by flow cytometry under live cell non‐permeable conditions (Figure 5A). APC fluorescence intensity of parental 293 T cells by flow cytometric analysis showed that weak cell surface MC16 expression. The median APC fluorescence intensity of 293T‐ANXA1‐His (2.89) is 1.64‐fold higher than that of parental 293 T (1.76) (Figure 5B). Thus, we confirmed that the expression of cell surface MC16 in 293T‐ANXA1‐His is higher than in the parental 293 T.

Figure 4.

Intracellular uptake of C(MI‐FAM)‐vc‐dhp7 in ANXA1 positive 293 T cells and biodistribution of C(MI‐Cy7.5)‐vc‐dhp7. A, Immunofluorescence of MC16 (green) and His‐Tag (magenta) expression. B, Intracellular uptake of each peptide (green) and lysosome (Red) distribution in 293 T or 293T‐ANXA1‐His cells at 37°C for 5 min incubation. C, Number of FAM‐positive dots taken up into the cell. *p < 0.05, *** < 0.0005, ns: not significant (Kruskal‐Wallis and Dunn's post hoc test), error bars denote means ± SEM. D, In vivo and fluorescence distribution and E, average Radiant Efficiency [p/s/cm²/sr]/[μW/cm²]) of Cy7.5 of PC‐3‐hLuc‐PSMA tumor‐bearing mice from 0 to 24 h after intravenous injection of C(MI‐Cy7.5)‐vc‐dhp7, error bars denote means ± SEM. F, Ex vivo fluorescence distribution and Ex vivo Cy7.5 Average Radiant Efficiency [p/s/cm²/sr]/[μW/cm²] of excised tumors and organs after 24 h intravenous injection, error bars denote means ± SEM. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 5.

In vitro drug sensitivity of PC‐3‐hLuc‐PSMA cell against vcMMAE or c(vcMMAE)dhp7. A, Cell surface ANXA1 expression of MBT2 mouse bladder cancer, MG63 squamous cell carcinoma, SAS head and neck cancer, PC‐3‐hLuc‐PSMA prostate cancer and 293 T, 293T‐ANXA1‐His cell lines analyzed by flow cytometry. B, The median APC fluorescence intensity of each cell line. C, In vitro drug sensitivity of PC‐3‐hLuc‐PSMA cell against vcMMAE (blue line) or c(vcMMAE)dhp7 (red line), error bars denote means ± SEM. [Color figure can be viewed at wileyonlinelibrary.com]

C(MI‐FAM)‐vc‐dhp7 significantly accumulated in 293T‐ANXA1‐His cells at 37°C for 5 min and colocalized in the lysosome, while FAM did not in 293T‐ANXA1‐His cells and also C(MI‐FAM)‐vc‐dhp7 accumulation in 293T‐ANXA1‐His cells were higher than that of 293 T cells (Figure 4B,C), suggesting the dhp7 primarily targets ANXA1 positive cells.

3.4. dhp7 Targeting of Human Prostate Tumor‐Bearing Mice Model

A biodistribution study showed that C(MI‐Cy7.5)‐vc‐dhp7 accumulated in PC‐3‐hLuc‐PSMA tumors at 30 min after intravenous injection and was detectable at the tumor site until 120 min after injection (Figure 4D,E). Twenty‐4 h after intravenous injection, ex vivo imaging showed that C(MI‐Cy7.5)‐vc‐dhp7 was accumulated mainly in liver, kidney, and remained slightly in the intestine and tumor (Figure 4F), suggesting the dhp7 primarily targets ANXA1‐positive tumors within 120 min after injection and may remain in the liver and excreted from the kidney.

3.5. In Vitro Drug Sensitivity of PC‐3‐hLuc‐PSMA Cells Against MMAE or c(vcMMAE) dhp7

To assess the In Vitro antitumor effects of vcMMAE or c(vcMMAE)dhp7, we confirmed that ANXA1 is expressed on the cell surface in PC‐3‐hLuc‐PSMA, MBT2 bladder cancer, MG‐63 osteosarcoma, and SAS squamous cancer cell lines (Figure 5A,B). The IC50 of vcMMAE (129.9 ng) and c(vcMMAE)dhp7 (122.7 ng) against PC‐3‐hLuc‐PSMA cells showed comparable drug sensitivity In Vitro (Figure 5C).

3.6. Antitumor Effect of c(vcMMAE)dhp7 on PCa‐Bearing Mouse

Next, we compared In Vivo anticancer effect of intravenously administrated c(vcMMAE)dhp7 or vcMMAE alone in subcutaneous PC‐3‐hluc‐PSMA tumor bearing mice model. Despite being administered at the same dose as vcMMAE alone group (1.25 mg/kg every 3 or 4 days, total four courses), photon count of each tumors by IVIS monitoring (Figure 6A,B) and residual tumor weight (Figure 6D,E) showed a significant antitumor effect in the c(vcMMAE)dhp7 group (2.14 mg/kg every 3 or 4 days, total four courses). The maximum percentage of tumor versus body weight observed in the study was 1.7% (467 mg tumor/27.4 g body weight). The maximum tumor was 15 mm in the study. The 2.95%–7.87% body weight loss was observed in this study. Significant body weight changes due to ascites or cachexia were not observed in the study (Figure 6C).

Figure 6.

Antitumor effect of vcMMAE or c(vcMMAE)dhp7 in PC‐3‐hLuc‐PSMA tumor‐bearing mice and immunohistochemical analysis of PC‐3‐hLuc‐PSMA residual tumors after vcMMAE or c(vcMMAE)dhp7 treatment. A, Antitumor effect of intravenously injected c(vcMMAE)dhp7 (2.14 mg/kg) or vcMMAE alone (1.25 mg/kg) at day 0, 3, 7, 10 (total four courses) on the PC‐3‐hLuc‐PSMA subcutaneous tumor model mice. B, Photon number (Average radiance [p/s/cm²/sr]) and C, Bodyweight of vcMMAE (n = 6) and c(vcMMAE)dhp7 (n = 6) treated group, error bars denote means ± SEM. *p < 0.05 (Mann Whitney U test). D, Appearance of excised residual tumor treated with vcMMAE (n = 6) or c(vcMMAE)dhp7 (n = 6). The yellow dashed line indicates the split section of the pathological analysis. The red dashed line indicates no tumor cells observed in the HE stained sections. Scale bars: 500 μM. E, Weight of excised residual tumor treated with vcMMAE or c(vcMMAE)dhp7. F, Representative tumor tissue sections stained with HE or indicated antibodies. Yellow squares indicated enlarged images. G, The number of CD31 positive blood vessels per whole tumor area (n/mm²) of indicated groups. H, The ANXA1 intensity per whole tumor cell (n) of indicated groups. I, The PSMA intensity per whole tumor cell (n) of indicated groups. J, Ki‐67 index of indicated groups. Results are expressed as a violin plot showing all data. *p < 0.05, ** < 0.01, ns: not significant (Mann Whitney U test). [Color figure can be viewed at wileyonlinelibrary.com]

Hematoxylin/eosin staining of excised tissue showed no residual tumor in 2 of 6 mice in the c(vcMMAE)dhp7 group, indicating a complete response to treatment. CD31‐positive vascular density of the residual tumor in the c(vcMMAE)dhp7 group (Figure 6F,G) was much higher than that of vcMMAE group and a trend toward higher staining intensity of ANXA1 MC16 and PSMA in c(vcMMAE)dhp7 group (Figure 6F,H,I). It is thought that MMAE effectively reached the tumor area via tumor vessels, which may be one of the reasons for the observed remarkable antitumor effect. The Ki‐67 index of the residual tumor in the c(vcMMAE)dhp7 group was significantly lower than that of vcMMAE group (Figure 6F,J), suggesting that the proliferative potential was also effectively suppressed. Clinical chemistry blood tests showed that the predominant findings of vcMMAE group included hepatic/biliary toxicity (T‐Bil 2.2 ± 0.7 mg/dL, AST 1162.7 ± 556.1 IU/L, LDH 5030.2 ± 2431.5 IU/L), and protein changes (T‐pro 5.8 ± 2.1 g/dL and Alb 2.4 ± 1.3 g/dL) while these findings significantly decreased to the same level as those in healthy mice with the remarkable shrinkage of the tumors of c(vcMMAE)dhp7 group (T‐Bil 0.5 ± 0.2 mg/dL, AST 494.3 ± 206.6 IU/L, LDH 2161.5 ± 843.8 IU/L, T‐pro 4.2 ± 0.3 g/dL and Alb 1.6 ± 0.1 g/dL, respectively) (Table 1) [28]. In addition, both vcMMAE group and c(vcMMAE)dhp7 groups showed a significantly high level of creatinine (0.9 ± 0.3 and 0.9 ± 0.1 mg/dL, respectively) compared to that of healthy control mice (0.7 ± 0.1 mg/dL), suggesting that both drugs had renal toxicity.

Table 1.

Blood count and blood biochemistry test at sacrifice of vcMMAE or c(vcMMAE)dhp7 treated mice and healthy mice.

Blood count	vcMMAEa (n = 6)		c(vcMMAE)dhp7b (n = 6)		Healthy micec (n = 6)		p value
	Mean	SD	Mean	SD	Mean	SD	a vs. b	a vs. c	b vs. c
WBC × 102/uL	50.6	20.9	44.8	11.5	37.6	9.5	0.95	0.32	0.33
RBC × 104/uL	806.7	267.4	873.5	53.2	971.2	32.6	0.33	0.19	0.03*
HGB g/dL	12.4	4.0	13.0	0.7	14.0	0.5	0.12	0.97	0.12
HCT %	33.9	11.1	36.8	2.3	39.1	1.5	0.32	0.69	0.17
MCV fL	42.1	0.3	42.2	1.6	40.3	0.7	0.91	0.02*	0.02*
MCH pg	15.4	0.4	14.9	0.2	14.4	0.3	0.17	0.00*	0.05
MCHC g/dL	36.6	0.9	35.3	1.5	35.7	0.5	0.10	0.17	0.86
PLT × 104/uL	24.5	19.7	42.4	30.1	40.6	38.2	0.33	0.59	0.71
Blood biochemistry	Mean	SD	Mean	SD	Mean	SD	a vs b	a vs c	b vs c
T‐Pro g/dL	5.8	2.1	4.2	0.3	4.2	0.5	0.00*	0.00*	0.94
Alb g/dL	2.4	1.3	1.6	0.1	1.6	0.2	0.02*	0.00*	0.71
BUN mg/dL	21.7	2.2	23.8	5.0	33.0	7.1	0.57	0.02*	0.06
Cre mg/dL	0.9	0.3	0.9	0.1	0.7	0.1	0.77	0.08	0.04*
Glu mg/dl	189.0	22.4	157.3	25.8	146.0	20.9	0.12	0.02*	0.40
T‐Cho mg/dL	110.2	68.6	90.7	12.5	53.2	4.0	0.86	0.01*	0.00*
TG mg/dL	123.0	17.0	94.5	23.8	79.6	24.0	0.05*	0.01*	0.34
T‐Bil mg/dL	2.2	0.7	0.5	0.2	0.6	0.1	0.00*	0.01*	0.72
AST IU/L	1162.7	556.1	494.3	206.6	479.4	159.2	0.01*	0.02*	0.97
ALT IU/L	225.0	142.4	104.3	48.0	98.8	15.4	0.03	0.03*	0.84
LDH IU/L	5030.2	2431.5	2161.5	843.8	1877.2	582.1	0.02*	0.02*	0.84

Note: Asterisk * indicates significant difference between the groups (Kruskal‐Wallis and Dunn's post hoc test).

Abbreviations: Alb, albumin; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; Cre, creatinine; Glu, glucose; HCT, hematocrit; HGB, hemoglobin count; LDH, lactate dehydrogenase; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; PLT, platelet count; RBC, red blood cell count; T‐Bil, total bilirubin; T‐Cho, total cholesterol; TG, triglyceride; T‐Pro, total protein; WBC, white blood cell count.

4. Discussion

We identified 7‐mer d‐peptides targeting the N‐terminal MC16 of ANXA1 by mirror‐image peptide phage display strategy (Figure 1A). MC16 is short and flexible in solution and is unlikely to form stable 3D structures, Nonaka et al. successfully identified d‐TIT7 peptides that bound to l‐MC16 of ANXA1 utilizing mirror‐image peptide phage display screening under reaction conditions at 4°C for 24 h [15]. In our study, to obtain peptide sequences with high affinity for MC16 under In Vivo‐like conditions, we screened peptide sequences that bound to MC16 under high stringency conditions with d‐MC16 concentrations ranging from 10 to 0.01 nmol/well in 5th rounds at 37°C for 5 min. Indeed, dhp7 was specifically bound to l‐MC16, not d‐MC16 (Figure 2A–D). We also showed that dhp7 did not show any binding to the l‐MC16 mutant by BLI analysis (Figure 2E,F) and obtained negative effects on binding by in silico docking analysis (Figure 2G,H), suggesting that the α‐helix structure of l‐MC16 may be important for dhp7 binding. In this study, we have shown that C(MI‐FAM)‐vc‐dhp7 accumulated in lysosomes of ANXA1 positive cells (Figure 4B,C) and also showed C(MI‐Cy7.5)‐vc‐dhp7 accumulates in ANXA1 positive tumor, liver, and kidney in an ANXA1 positive PCa bearing mouse model (Figure 4D,E,F). Targeting of C(MI‐Cy7.5)‐vc‐dhp7 to the liver or kidney may be due to the C(MI‐Cy7.5)‐vc‐dhp7 conjugate that binds to unknown molecules expressed on the surface of benign organs. This hypothesis is supported by the fact that intravenous injection of fluorescent Alexa488‐IF7 targets brain tumors but not kidneys [15]. Nonaka et al. reported that intravenously administrated IRDye‐dTIT7 has been specifically accumulated in brain tumor [15]. From this result it was expected that intravenously injected GA‐dTIT7 into mice with brain tumors would show antitumor activity In Vivo, but no therapeutic activity was observed [15]. However, orally administrated GA‐dTIT7 inhibited tumor growth in a mouse brain tumor model [15]. This report suggests that the route of administration and absorption of PDCs depends on the type of payload bound to the peptide, which may affect drug efficacy. In ADCs, vcMMAE is often used as a payload, and once taken up into the cell via cell surface target molecules, the drug is hydrolyzed by cathepsin B in intracellular lysosomes and efficiently released into the cell. This linker‐mediated ADC strategy is well established and antibody against Nectin‐4, which is highly expressed in urothelial carcinomas, conjugated with vcMMAE (Enfortumab‐vedotin) has been approved and reported favorable clinical results [29, 30, 31].

Taxane‐based anticancer drugs for CRPC exert potent antitumor effects, but systemic side effects are a major problem. PSMA‐targeting theranostics with Lutetium‐177 (Lu¹⁷⁷)‐PSMA‐617 is one of the PCa‐targeted therapies with favorable therapeutic outcomes reported [19], but treating PSMA‐negative but FDG‐positive lesion in patients with CRPC is associated with an unfavorable response to Lu¹⁷⁷‐PSMA‐617 [20, 21]. To solve the above problem, PSMA‐independent cancer‐specific chemotherapy such as ANXA1‐targeted PDC strategy is needed. ANXA1 has been reported to be expressed in tumor vasculature of PCa [9]. In addition, ANXA1 localized in cytoplasm and membrane of PCa tissue and indicated a significant negative correlation between ANXA1 and androgen receptor expression in PCa tissue [32]. Furthermore, higher expression of nuclear ANXA1 and epidermal growth factor receptor was detected in both CRPC cell lines and in PCa tissue [33]. Another fact reported that ANXA1 and ribonucleotide reductase subunit M2 regulate PCa progression and resistance to docetaxel treatment [34]. These reports suggest that c(vcMMAE)dhp7, a PDC targeting ANXA1, may be an effective therapeutic target for CRPC. The molecular weight of c(vcMMAE)dhp7 (2257.6) is 1/66 smaller than that of ADC (about 150000) [35]. Thus, c(vcMMAE)dhp7 is more deeply absorbed than ADCs and may accumulate in target tissues faster within 30‐120 min after intravenous administration, suggesting that c(vcMMAE)dhp7 is an effective strategy to replace ADCs, eliminating the instability of conventional l‐type PDCs. We also showed that c(vcMMAE)dhp7 has in vitro cytotoxic activity comparable to that of vcMMAE against ANXA1 positive PCa cells (Figure 5C). These results indicate that neither tumor specificity nor drug sensitivity is decreased when dhp7 is bound to vcMMAE. The present study shows that c(vcMMAE)dhp7 has a significantly higher antitumor effect, including pathologically complete response, than vcMMAE alone at a clinically effective dose of 1.25 mg/kg equivalent of vcMMAE (Figure 6). Pathological observations of the residual tumors showed that the CD31‐positive vascular density of the residual tumors in the c(vcMMAE)dhp7‐treated group (Figure 6F,G) was much higher than in the vcMMAE‐treated group (Figure 6F). This suggests that c(vcMMAE)dhp7 administration effectively targeted tumor sites via intra‐tumor vessels without destroying tumor vessels. This may be the same mechanistic insight that IF7 accumulates in tumor tissue by passing through tumor vascular endothelial cells by transcytosis, as has been observed with IF7 [4, 5, 6, 7, 8, 9].

Previously, Oh et al. reported that ANXA1 is selectively concentrated in human and rodent caveolae in the tumor endothelium and that caveolin 1 and ANXA1 expression are required for intravenously injected ANXA1‐targeted antibodies to rapidly cross the endothelium and enter the tumor site [3]. In addition, delayed angiogenesis has been observed in Anxa1‐deficient mice in tumor angiogenesis [36]. These reports suggest that ANXA1 may be involved in tumor angiogenesis, but the detailed mechanisms are not fully understood within the tumor microenvironment. Several externalization mechanisms of ANXA1 in in vitro studies have been reported via activated ATP‐binding cassette transporter [37], via fusion of ANXA1‐containing granules to the plasma membrane [38], via phosphorylation of ANXA1 Ser27 by protein kinase C [39]. Hebeda et al. reported that externalized ANXA1 interacts with Formyl peptide receptor (FPR) in an autocrine manner in endothelial cells to promote the externalization of VEGF [40, 41], and it promotes the angiogenesis of endothelial cells by inducing the phosphorylation of ANXA1 mediated by p38/MAPKAP kinase‐2/LIMK1 activation [42]. Therefore, VEGF and ANXA1 may form a positive feedback loop that synergistically promotes tumor angiogenesis [43] and such a mechanism may be involved in the accumulation of ANXA1 on the tumor endothelium surface. Insights into such further mechanisms of ANXA1 function, need to be addressed in future studies focusing on the elucidation of signaling pathways related to ANXA1 externalization, VEGF signaling and Caveolin 1‐mediated mechanisms of ANXA1 enrichment into caveolae in tumor endothelium.

ANXA1 is also highly expressed on neutrophils, mast cells, and monocyte macrophages [44], and plays a role as an immunomodulatory factor in the tumor immune microenvironment [43]. ANXA1 released from cancer cells activates macrophage FPRs signaling pathways and induces M2 macrophage polarization [43]. ANXA1 released by M2 macrophages has also been enhanced with cancer cell FPRs‐PI3K/AKT and MEK/ERK axis to promote cancer progression [43]. It has been reported that the N‐terminal peptide Ac2‐26 of ANXA1 act as an FPR receptor agonist, activating downstream signaling [45]. These reports suggest that blocking the signaling of ANXA1 and FPRs between tumor cells and tumor vasculature or immune microenvironment may lead to the development of novel therapeutic way for cancer‐specific chemotherapy. Since dhp7 binds to the N‐terminal MC16 domain of ANXA1, if dhp7 inhibits ANXA1 binding to the FPR receptor, dhp7 itself could act as a signal blocker between ANXA1 and the FPR receptor and disrupt the tumor microenvironment including angiogenesis and immune response. In this study, we did not evaluate the function of dhp7 itself in the tumor microenvironment as described above, which should be investigated in future studies.

In addition, hepatic/biliary toxicity and accompanying serum protein changes due to hepatic/biliary toxicity, a side effect of MMAE, was significantly reduced with c(vcMMAE)dhp7 (Table 1). In the biodistribution study (Figure 4D) of fluorescence‐labeled dhp7, fluorescence remained in the liver and kidney at 24 h after tail vein injection. MMAE is mainly metabolized through CYP3A4‐mediated metabolism by the liver [46]. If c(vcMMAE)dhp7 shows the same biodistribution as fluorescence‐labeled dhp7, the accumulation of c(vcMMAE)dhp7 at the tumor site, kidney, and liver may improve the systemic condition by effective tumor shrinkage and decrease hepatic and biliary toxicity due to proper metabolism of the MMAE accumulated in the liver. The protective effect on the liver of the dhp7 peptide itself also cannot be ruled out. On the other hand, in the case of MMAE alone, the deterioration of the systemic condition caused by tumor growth may lead to liver damage due to impaired hepatic metabolic capacity and also lead to kidney damage. It was reported that whole‐body pharmacokinetics study of MMAE in tumor‐bearing mice showed MMAE is rapidly eliminated from plasma, but is distributed in liver and kidney tissues, blood cells, and tumors for a long time and extensively, with area under the plasma concentration curve ratios over 20 in highly perfused tissues (e.g., heart, lung, spleen, liver and kidney) and 1.3 to 2.4 in poorly perfused tissues (e.g., fat, pancreas, skin, bone, and muscle) [46]. Biodistribution study (Figure 4B) suggests that c(vcMMAE)dhp7 may be excreted by the kidneys, and indeed, creatinine levels in the dhp7‐MMAE group were similar to those in the MMAE group and higher than in healthy mice, with no significant improvement, suggesting a certain degree of nephrotoxicity (Table 1). Although the reaction conditions for the KD measurements were different and therefore cannot be directly compared, the KD of dhp7 peptide to MC16 is 48 μM, which was higher in affinity to MC16 than those of dTIT7 is 466 μM [15], and IF7 is 638 μM [9]. Since, Nonaka et al. showed that IF7‐SN38 and GA‐dTIT7 had significantly higher antitumor effects with reducing side effect than single agents in a mouse brain and subcutaneous tumor model [8, 9, 15, 47], tumor targeting function of dhp7 might be considered to be comparable or higher. A direct comparison will be needed in the future with the same payload to determine which peptide is the best. The d‐type peptides (dhp7 and dTIT7) are more hydrophilic than IF7. Therefore, the dhp7 and dTIT7 often remain soluble when bound to the payload, and they are more flexible than IF7 with regard to drug form. The dhp7 peptide also has the unique ability to actively accumulate in tumor tissue rather than simply diffuse into tumor tissue, which may explain why it reduced side effects in the mouse model and showed significant therapeutic efficacy. These results showed that c(vcMMAE)dhp7 conjugated with MMAE is a treatment that is expected to exert more antitumor effects and reduce side effects than vcMMAE.

The limitations of the present study included the evaluation of the antitumor effect of c(vcMMAE)dhp7 using a PCa cell line‐derived subcutaneous tumor mice model. Further validation of therapeutic efficacy using a patient‐derived xenograft model is needed to adapt preclinical translation. In this study, In Vitro cell‐based and In Vivo biodistribution studies using fluorescently labeled dhp7 peptides showed that dhp7 peptides are targeted to ANXA1‐positive cells and tumors, but mechanisms of drug toxicity using c(vcMMAE)dhp7 itself studies were not included. Further mechanistic studies of drug toxicity, including quantification of MMAE concentrations in various organs and tumors, are needed to better assess clinical safety. Furthermore, since dhp7 targets tumor endothelium and tumor ANXA1, the use of different payloads, such as radioisotopes (e.g., Lu¹⁷⁷, Ga⁶⁸ and F¹⁸) and angiogenesis‐inhibiting tyrosine kinase inhibitors, will lead to the development of PDCs dedicated to tumor‐specific imaging or theranostics against PSMA‐negative PCa, and tumor angiogenesis inhibition.

5. Conclusions

The development of advanced therapies such as immune checkpoint inhibitors (ICIs) and ADC‐based molecularly targeted chemotherapy has made cancer treatment increasingly expensive. d‐type PDC can be chemically synthesized more cost‐effectively than ADC or ICIs. Since ANXA1 is a highly specific tumor vasculature surface marker [2], and d‐type PDCs targeting ANXA1 should also be able to effectively eradicate tumors with few side effects. We may soon be able to evaluate the efficacy of these strategies in clinical trials.

Author Contributions

Kai Ozaki: conceptualization – equal, data curation – equal, formal analysis – equal. writing – original draft – equal. Tohru Yoneyama: conceptualization – lead, data curation – lead, formal analysis – lead, funding acquisition – lead, investigation – lead, methodology – lead, project administration – lead, supervision – lead, validation – lead, writing – original draft – Lead, writing – review and editing – Lead. Yuka Kubota: conceptualization – equal, data curation – equal, formal analysis – equal, writing – original draft – equal. Mihiko Sutoh Yoneyama: data curation – equal, formal analysis – equal, investigation – equal, writing – review and editing – equal. Motohiro Nonaka: data curation – equal, formal analysis – equal, investigation – equal, writing – review and editing – equal. Tomonori Suzuki: data curation – equal, formal analysis –equal, investigation – equal, writing – review and editing – equal. Ryuuma Tanaka: data curation – equal, formal analysis – equal, investigation – equal, writing – review and editing – equal. Chikara Ohyama: conceptualization – equal, data curation – equal, formal analysis – equal, funding acquisition – equal, investigation – equal, methodology – equal, supervision – equal, writing – review and editing – equal. Shingo Hatakeyama: conceptualization – equal, data curation – equal, formal analysis – equal, funding acquisition – equal, investigation – equal, methodology – equal, supervision – equal, writing – original draft – equal, writing – review and editing – equal.

Ethics Statement

Approval of the research protocol by an Institutional Reviewer Board: N/A. Informed Consent: N/A. Registry and the Registration No. of the study/trial: N/A. Animal Studies: All animal studies were carried out in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Hirosaki University Graduate School of Medicine Animal Care and Use Committee (permit numbers: AE01‐2024‐005).

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Table S1 Ranking of enriched peptide seqience of each biopanning rounds.

Additional movie file (mp4).

ARRIVE Author Checklist‐Full.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. NGS sequencing data set of each screening round of T7 phage sequence was submitted to NCBI BioSample. The accession numbers SAMN47328472 (1st round input of T7 library), SAMN47328473 (1st round of T7 library), SAMN47328474 (2nd round of T7 library), SAMN47328475 (3rd round of T7 library), SAMN47328476 (4th round of T7 library), and SAMN47328477 (5th round of T7 library) will be released on 30th April, 2026.