Introduction

Vascular disrupting agents (VDAs) effectively disrupt blood vessels and induce vascular shutdown in solid tumors1,2. The most studied VDA, combretastatin A-4 phosphate (CA4P, Fosbretabulin)3, induces microtubule depolymerization in vascular endothelial cells, resulting in catastrophic shutdown of the tumor vasculature within minutes of drug exposure, causing necrosis and secondary tumor cell death2,3. In addition, CA4P activates crucial pathways, such as RhoA-GTPase and its target kinase, Rho kinase (ROCK)-dependent signaling cascade, disrupting focal adhesions and membrane blebbing4. Despite its dramatic effects, clinical studies have shown that CA4P monotherapy has a limited response5,6,7,8. Consequently, combinations of CA4P with other drugs to overcome VDA treatment resistance, including anti-angiogenic agents, immune checkpoint inhibitors, and drugs targeting pro-survival pathways, have been evaluated9,10,11,12. A challenge with these secondary therapies is that VDA-induced vascular shutdown results in tumors becoming increasingly hypoxic and nutrient-depleted, conditions known to render tumors resistant to many anticancer therapies13,14,15. Another major drawback of VDA treatment resistance is that it leaves a layer of malignant cells at the outer tumor rim, which contribute to tumor regrowth15,16,17. Solid tumors also exhibit resistance to non-canonical VDAs, including vascular-disrupting nanosystems18. Interestingly, treatment of breast tumor mice with one such nanosystem, iron oxide nanoparticles coated with a tumor-homing peptide (CGKRK) fused to a pro-apoptotic peptide [D(KLAKLAK)2], was found to enhance the expression of angiogenesis-related genes and integrin β3 in treatment-resistant tumor blood vessels18. The molecular signatures imposed by VDA treatment could potentially be targeted with secondary affinity ligand-guided therapeutics to overcome drug resistance and achieve a sustained therapeutic response.

Peptides and peptidomimetics featuring the RGD motif are developed as potent cell adhesion modulators and affinity-targeting ligands for drugs, contrast agents, and nanoparticles19,20. Systemic RGD-containing peptides target integrins, particularly αvβ3 integrins21,22, which are commonly overexpressed in vascular and malignant cells across different solid tumors. The tumor-penetrating peptide iRGD (sequence: CRGDKGPDC) incorporates the tumor-homing RGD module in the context of the conditional neuropilin-1 (NRP-1) interacting motif, RGDK, which is activated by proteolytic cleavage and then transforms the solid tumor microenvironment into a temporary drug conduit23,24. RGDK and other peptides with consensus R/KXXR/K engage with NRP-1 only when present at the C-terminus of the (poly)peptide chain, a feature that inspired the designation of this family of C-end Rule, or CendR, peptides25. CendR motifs are involved in different biological and pathological processes, spanning the modulation of growth factor signaling to facilitate cellular entry of pathogenic viruses26. Importantly, the CendR peptide in VEGF-A165 promotes cellular internalization and vascular leakage27, and short peptides with CendR sequences have similar activity28. iRGD effectively enhances the accumulation and spreading of conjugated and coadministered cargoes within solid tumors23,24 and has demonstrated promise in clinical studies29,30. iRGD is able to penetrate solid tumors not only through the vasculature, but also locally and independently of circulation, expanding its applications to locoregional cancer therapy31,32.

Peritoneal carcinomatosis (PC), an intraperitoneal (IP) dissemination of cancer, is a major cause of mortality in patients with advanced abdominal neoplasms33. IP administration of chemotherapeutics has evolved into a clinically efficacious treatment modality for patients with PC34. Preclinical studies have demonstrated that in addition to cytotoxic small anticancer drugs, IP treatment with VDAs can effectively suppress the development of PC35. IP drug perfusion provides a pharmacokinetic advantage: a high IP dose generally results in only moderate systemic drug exposure due to peritoneal clearance across the peritoneal-plasma “barrier”, which is much slower than systemic clearance36. However, high interstitial fluid pressure, deposition of dense extracellular matrix (ECM), and factors such as drug sequestration, metabolism, and degradation collectively impede the uptake of IP drugs by tumors. These obstacles restrict drug penetration into peritoneal tumors to a range of 3–5 mm, even when heated drug solutions are used to facilitate the breakdown of tissue barriers and enhance drug diffusion37. The IP iRGD peptide was found to significantly augment intratumoral entry of the coinjected anthracycline doxorubicin and suppress the growth of large peritoneal tumors32. In another study, we observed that iRGD functionalization of paclitaxel-loaded polymeric nanoparticles (polymersomes, PS) increased their intraperitoneal accumulation in PC lesions and their therapeutic efficacy31. These studies suggest that iRGD-based precision delivery strategies can be used to enhance the therapeutic impact of intraperitoneal cancer drugs and nanosystems. These studies also raised the possibility that because of the known dependence of the CendR tissue penetration pathway on nutrient deprivation38, the iRGD peptide may be particularly well suited for targeting hypoxic and nutrient-depleted tumors post-VDA treatment.

Here, we hypothesized that the VDA treatment-imposed gene expression signature in tumor tissue provides an opportunity for targeting with a secondary iRGD-guided nanotherapy. We show that CA4P treatment upregulates the expression of iRGD receptor integrins and NRP-1 in the tumor tissue, enhancing the effect of the iRGD peptide on tumor accumulation and the therapeutic efficacy of coadministered polymersomes loaded with a novel anthracycline, Utorubicin. These observations suggest that VDA-treated tumors can be treated with secondary iRGD-potentiated nanotherapy and warrant further studies on the precision targeting of VDA-induced molecular signatures.

Materials and methods

Materials

UTO was provided by ToxInvent LLC, Estonia. The copolymers polyethylene glycol-polycaprolactone (PEG5000-PCL10000, with MW of 5000 and 10,000, respectively; hereafter PEG-PCL) and maleimide-PEG5000-PCL10000 (mal-PEG-PCL) were acquired from Advanced Polymer Materials Inc., Canada. 5(6)-carboxyfluorescein-Cystein (FAM-Cys) was purchased from TAG Copenhagen, Denmark. Tween-20, paraformaldehyde (PFA), and Triton X-100 were acquired from Sigma–Aldrich, Germany. Dulbecco's Modified Eagle Medium (DMEM) and phosphate-buffered saline (PBS) with a pH of 7.4 were acquired from Lonza, Belgium. Trypan Blue was procured from Gibco, Thermo Fisher Scientific, USA. Bovine serum albumin (BSA) and fetal bovine serum (FBS) were obtained from Capricorn Scientific, Germany. The iRGD peptide (sequence CRGDKGPDC with N-terminal acetylation and C-terminal amidation, as well as a C–C disulfide bond, MW 948.05, also known as LSTA1, formerly known as CEND-1) was obtained from Lisata Therapeutics, USA. Combretastatin A4 phosphate (CA4P, Fosbretabulin) was purchased from Adooq Bioscience, USA. Anti-NRP-1 (in-house made), anti-integrin Alpha V + Beta 3 (bs-1310R), anti-integrin alpha V/CD51 (bs-2250R), and anti-integrin alpha V + beta 5 (bs-1356R) antibodies were obtained from Bioss Antibodies, USA. Anti-fluorescein/Oregon Green (2400907), goat anti-rabbit Alexa Fluor 546 (A11010), goat anti-rat Alexa Fluor 546 (A11081), and goat anti-rabbit Alexa Fluor 647 (A21245) antibodies were purchased from Thermo Fisher Scientific, USA. Anti-CD31 antibody (cat. no. 553370) was purchased from BD Biosciences, USA. Anti-integrin β3 (ab203122) and anti-CD105 (ab137389) antibodies were purchased from Abcam, UK.

Fluorescein-PEG-PCL (FAM-PEG-PCL) synthesis

The copolymer polyethylene glycol (5 kDa)—polycaprolactone (10 kDa) functionalized with maleimide in the PEG block (mal-PEG-PCL) (20 mg) was dissolved in 300 µL of dimethylformamide (DMF), which had been purged with nitrogen. Two equivalents (eq.) of FAM-Cys were dissolved in 100 µL of nitrogen-purged DMF and added to the mal-PEG-PCL solution. The resulting mixture was continuously stirred at room temperature (RT) for 2 h and then at 4 °C overnight (ON). Subsequently, the solution was diluted in 2 mL of deionized water and subjected to dialysis for 2 h and 2 h at RT and ON at 4 °C against water using a dialysis cassette (Thermo Scientific, USA) with a molecular weight cut-off of 10 kDa, and the water was changed after each dialysis step. This step was performed to eliminate excess FAM-Cys from the solution. The resulting suspension was freeze-dried to obtain a yellow powder.

Synthesis of FAM-labeled polymersomes (FAM-PS) and UTO-loaded polymersomes (UTO-PS)

Polymersomes (PS) were formed using the film hydration method following a protocol that was optimized based on previous research39. To prepare FAM-PS, 1 eq. of FAM-PEG-PCL, and 9 equiv. PEG-PCL were then dissolved in acetone. The acetone was then evaporated using nitrogen flow, resulting in the formation of a thin polymeric film on the inner surface of a glass vial purchased from Sigma-Aldrich, Germany. The film-coated vial was subsequently hydrated with sterile phosphate-buffered saline (PBS) at pH 7.4. The vial was then heated for 30 s in a water bath maintained at 65 °C and sonicated for 30 s. This heating and sonication cycle was repeated until PS formation was achieved, ensuring that no polymer aggregates were observed in the suspension. The final polymer concentration in each sample was 10 mg/mL.

UTO-PS were synthesized according to a previously described protocol40. Briefly, UTO (MW 627.2) was dissolved in acetone and added to the PEG-PCL-containing acetone solution to obtain a final concentration of 100 µM UTO. The resulting UTO/polymer solution in acetone was evenly distributed among glass vials. Each vial contained 50 mg of PEG-PCL and 500 nmol of UTO. Acetone was evaporated to form a thin film of polymer and drug within each vial, following the same procedure as described above for PS formation. The vials containing the drug/polymer film were tightly sealed, shielded from light, and stored at -20 °C. Before injecting UTO-PS, 5 mL of sterile PBS was added to each vial, as previously described, in order to guarantee their freshness.

Characterization of PS samples

The average hydrodynamic diameter of the prepared PS samples was determined using dynamic light scattering (DLS) on a Zetasizer Nano ZSP instrument (Malvern, USA). To measure the diameter, PS samples were diluted with PBS (pH 7.4) to a concentration of 1 mg/mL. DLS measurements were performed by scanning the samples for 10 s at 22 °C and averaging the results over 10 runs, with a total of three measurements. The hydrodynamic diameter results are shown in Supporting Information (Figs. S1 and S2).

To quantify the amount of encapsulated UTO, a Nanodrop 2000c UV–VIS spectrophotometer (Thermo Scientific, USA) was used. For UTO quantification, previously reported calibration curve was used40. The absorbance of the drug-PS samples was measured at 490 nm using 5 µL of the PS sample blank with PBS.

Cell culture

MKN45P human gastric carcinoma cells isolated from parental MKN-45 cells41 and MCF10CA1a human triple-negative breast cancer cells (obtained from the Erkki Ruoslahti laboratory at the Cancer Research Center, Sanford Burnham Prebys Medical Discovery Institute) were cultured in DMEM supplemented with 100 IU/mL streptomycin, penicillin, and 10% FBS. Cell culture was maintained at 37 °C in an incubator with 5% CO2. A trypsin solution (TrypLE Express, Gibco, USA) was used for cell detachment. When a specific number of cells needed to be used, they were mixed with 0.4% Trypan Blue (1:1) and counted using the Bio-Rad TC 10 automated cell counter (Bio-Rad, USA).

In vivo experiments

The animal experiments conducted in this study followed protocols approved by the Estonian Ministry of Agriculture Committee of Animal Experimentation (projects #159 and #160). All mouse experiments were performed in accordance with the relevant guidelines and regulations by the authors. The authors confirm that the study was conducted in accordance with the ARRIVE guidelines.

To induce MKN45P tumors, athymic nude female mice (Hsd/Athymic Fox1 nu Harlan; 7–8 weeks old) were intraperitoneally injected with 1 M MKN45P cells in 200 µL of PBS. Tumors developed over a period of 4–7 days. For MCF10CA1a tumor induction, athymic nude mice were injected with 2 M MCF10CA1a triple-negative breast tumor cells in 50 µL PBS in the mammary gland with tumors developing for 14 days.

Receptor-modulating effects in triple-negative breast tumors

When MCF10CA1a tumors reached approximately 40 mm3 in volume, the immunodeficient mice were divided into two groups maintaining an equal average size for each group. One group received daily IP injections of PBS, whereas the other group received IP injections of CA4P every other day (100 mg/kg). On day 33 after tumor induction, the mice were sacrificed by perfusion with 10 mL of PBS, and the organs and tumors were excised and fixed in 4% PFA solution. Subsequently, the tissues were embedded in OCT and cryosectioned to obtain 10 µm-thick sections. An immunostaining protocol without heat-induced antigen retrieval (HIAR) was used, as described below. The tumor sections were immunostained with anti-integrin β3, anti-NRP-1, and anti-CD31/mCD105 primary antibodies (see the Materials section for additional information). The tumor sections were then incubated with goat anti-rabbit Alexa Fluor 647 and goat anti-rat Alexa Fluor 546 secondary antibodies. DAPI staining was performed for nuclear visualization. Finally, the samples were imaged using a fluorescence confocal microscope (LSM 710, Zeiss). The acquired images were processed and analyzed using ZEN lite 2012 image software, and fluorescence signal quantification was performed as described below.

FAM-PS homing to MKN45P tumors in combination with CA4P and iRGD

Mice bearing orthotopic MKN45P tumors were randomly distributed into two groups that received seven alternate IP injections of CA4P and PBS as a control. Seven IP injections of CA4P was administered at a dose of 100 mg/kg, dissolved in saline solution at a concentration of 5 mg/mL, every other day. As a control, 500 µL of PBS was administered daily. Following a 14-day interval, tumor-bearing mice received IP injections of FAM-PS at a dose of approximately 160 mg/kg polymer. This injection was performed with or without the coadministration of iRGD, which was administered IP at a dose of 14 µmol/kg 15 min prior to FAM-PS injection. Six hours after FAM-PS injection, the animals were sacrificed by perfusion with 10 mL of PBS. The tumors and organs were excised and preserved in a 4% PFA solution in PBS at 4 °C. For FAM-PS homing quantification, immunofluorescence staining was performed on dissected tumor sections (20 µm) following the procedure described below without performing the HIAR step. Rabbit anti-fluorescein/Oregon Green and rat anti-CD31 antibodies were used as the primary antibodies, and goat anti-rat Alexa Fluor 546 and goat anti-rabbit Alexa Fluor 647 were used for staining.

Tissue immunofluorescence stainings

Immunofluorescence was selected as the method for quantifying receptor expression and assessing the spatial distribution of both receptors and FAM-PS. This choice was driven by the solubility of FAM-PS in methanol. Consequently, PFA-fixation was employed to facilitate the staining and quantification of FAM-PS homing to tumors. Therefore, this fixation method introduced methodological constraints for further evaluation of receptor expression.

For extracted PC tumors, PFA-fixed tissues were embedded in OCT, and 20 µm-thick sections were mounted on Epredia SuperFrost Plus Adhesion slides (Fischer Scientific) and stored at -20 °C before staining. Prior to staining, slides were air-dried for 2 h at room temperature (RT). The slides were then washed three times for five minutes each with PBS, and heat-induced antigen retrieval (HIAR) was performed as follows: HIAR solution (10 mM Tris base, 1 mM EDTA, 0.05% Tween-20, 0.9% NaCl, pH 9 in mQ) was heated in a microwave until boiling occurred. Subsequently, the slides were submerged in the solution in a Coplin jar, placed on a thermoblock at 120 °C for 20 min, covered with foil, and cooled in the same solution. The slides were then placed in cassettes (Ted Pella, #36105, USA) and washed three times for five minutes each with PBST (0.05% Tween-20 in PBS). Tissue sections were further permeabilized with PBS + 0.2% Triton X-100 for 20 min at RT. The tissues were then blocked with 5% bovine serum albumin (BSA) + 5% goat serum (GS) + 50 nM glycine in PBST for 1 h. The sections were further incubated with primary antibodies (anti-integrin alpha V/CD51 and anti-NRP-1; dilution 1:200) in an antibody buffer at 4 °C overnight. The next day, the slides were washed three times for five minutes each with PBST and incubated with secondary antibody solution (goat anti-rabbit Alexa Fluor 546; dilution 1:500) for 1 h at RT in the dark. After incubation, the slides were washed thrice for five minutes each with PBST and counterstained with DAPI solution (1 mg/mL in PBS; dilution 1:500) for 5 min at RT. Next, the slides were washed with PBS for 5 min and mounted with mounting medium. The stained tumors were scanned using an Aperio VERSA slide scanner with a 20 × objective.

The fluorescence signals from the scans and confocal microscopy were analyzed using Fiji (ImageJ 1.54f.) as follows: channels of the opened scan were split, each channel inverted and then converted to an 8-bit image, threshold adjusted until the red mask covered the true signal, images were converted to binary, and the signal intensities were measured from the whole images/scans. The mean signal intensity was normalized to the DAPI signal to avoid necrotic tumor areas. n = 3 mice per group for FAM signal quantification and n = 6 for integrin and NRP-1 quantification.

To quantify the FAM-positive area within the tumor section scans using Fiji, the channels were first split and duplicates were created for each channel. A median filter with a size of 1.0 pixel was then applied to each channel. The images were thresholded, ensuring the true fluorescence signal was covered, and selections of the two signals were made to generate representative masks for both DAPI and FAM. The areas of these masks were measured and subsequently converted into percentages to represent the FAM-positive areas of the tumors. For this analysis, n = 3 mice were included in each group.

To assess the receptor fluorescence intensity in the tumor periphery, a Python code42 was used to analyze changes in tissue intensity relative to the DAPI signal by comparing two sets: 'baseline as DAPI' and 'integrin av or NRP-1'. Intensity ratios were computed between paired images to detect immunostained tumor periphery using the Sobel operator, and contrast was enhanced for visualization. The receptor/DAPI signal ratio, represented on a logarithmic scale, provided a normalized representation of receptor expression relative to cellular density. The resulting gradient images provided a visual representation of the changes in tissue fluorescence intensity in the rim of the tumors.

Drug combination treatment studies

For the survival study, on day 6 after tumor induction, PC-bearing mice were divided into six groups, with seven mice per group (PBS group, n = 10) receiving different treatment combinations of CA4P, iRGD, and UTO-PS for 20 consecutive days. On days 7, 9, 11, 13, 15, 17, 19, 21, 23, and 25, mice were IP injected with CA4P (100 mg/kg; 5 mg/mL in saline solution; 10 IP injections every other day) or PBS (500 µL). On days 8, 10, 12, 14, 16, 18, 20, 22, 24, and 26, mice received IP injections of UTO-PS (1.4 mg/kg; 10 injections every other day) with or without iRGD (14 µmol/kg; 10 IP injections every other day) or PBS (500 µL). On the 27th day after tumor induction, injections were terminated and survival was monitored over time. Mice were sacrificed by CO2 inhalation when severe ascites developed, weight reduction was > 20%, or severe behavioral changes occurred (shivers, shaking, impaired movement, or severe cachexia).

To assess tumor progression in terms of tumor weight and total tumor count in the abdomen, another treatment study was conducted following the regimen used for the survival study. Due to the challenges in accurately quantifying and assessing the size of intraperitoneally growing PC tumors, tumor burden was estimated in end-point treatment studies. On day 4 after tumor induction, PC-bearing mice were divided into six groups, with six mice per group. On days 4, 6, 8, 10, 12, 14, 16, 18, and 20, the mice received IP injections of CA4P (100 mg/kg; 5 mg/mL in saline solution, 9 IP injections every other day) or PBS (500 µL). On days 5, 7, 9, 11, 13, 15, 17, 19, and 21, the mice received IP injections of UTO-PS (1.4 mg/kg; 9 IP injections every other day) with or without iRGD (4 µmol/kg; 9 IP injections every other day) or PBS (500 µL). On day 23 after tumor induction, all mice were sacrificed by CO2 inhalation and all tumors were extracted, weighed, and counted.

Toxicological studies

Healthy 8-week-old female Balb/c mice were divided into three groups with three mice per group. The mice received IP injections of CA4P (100 mg/kg, 5 mg/mL solution in saline, 5 IP injections every other day) or PBS (500 µL) every other day. On alternate days, mice received IP injections of UTO-PS (1.4 mg/kg, five IP injections every other day) with iRGD (14 µmol/kg; five IP injections every other day) 15 min prior to UTO-PS injection. After 10 days of treatment, 2–3 mL of blood was collected from deeply anesthetized mice through retro-orbital bleeding into Lithium Heparin tubes (BD Vacutainer, no. 368494). The blood samples were then centrifuged for 10 min at 1800g at 4 °C, and the plasma was analyzed for the concentration of glucose (GLUC), creatinine (CREAT), and activity of alanine aminotransferase (ALT) using a Cobas 6000 IT-MW machine (Roche Diagnostics Gmbh) and reagents for creatinine (CREP2, REF 03263991) and for ALT (ALTLP, REF 04467388). These measurements were conducted at the Tartu University Hospital. Following blood collection, heart tissues from the treated mice were removed and frozen at 80 °C for further histological analyses, as described below.

H&E stainings

Paraffin-embedded tissue sections (2 µm) were stained using the ST 5020 device with the "Ready-to-Use H&E Staining System Leica ST Infinity" kit (ST-1 HemaLast, ST-2 Hematoxylin, ST-3 Differentiator, ST-4 Bluing Agent, and ST-5 Eosin) and other chemicals according to the manufacturer's protocol. The staining protocol included sequential steps with specific durations of tissue immersion in chemical solutions. Initially, the sections underwent three consecutive xylene treatments (2, 2, and 1 min) to remove paraffin. Dehydration was performed using absolute ethanol (1 min), followed by 80% ethanol (1 min). The tissue sections were then stained with ST-1 HemaLast (30 s) to enhance affinity, ST-2 Hematoxylin (5 min) for nucleus staining, and rinsed with tap water (2 min). The ST-3 Differentiator (45 s) was applied, followed by another water rinse (1 min). An ST-4 Bluing Agent (1 min) was used for contrast enhancement, followed by another rinse with water (1 min). Finally, ST-5 Eosin (1 min) stained the cytoplasm and structures, and the sections were dehydrated (30 s, 30 s, 2 min) and xylene cleared (2, 2 min). H&E staining was performed at the Pathology Department of the Tartu University Hospital. Stained heart sections were scanned using a slide scanner (Leica SCN400) with a 20 × objective lens. Images were analyzed using ImageScope × 64 program.

Statistical analyses

All statistical analyses were carried out using Student’s t-test for two-sample comparisons, and one-way ANOVA and Fisher LSD (Least Significant Difference) using GraphPad Prism (version 10) with a 0.05 significance threshold. The same program was used to plot Kaplan–Meier survival curves, and the Gehan-Breslow-Wilcoxon test was used for the statistical analysis.

Results

CA4P treatment upregulates expression of receptors of the iRGD peptide in tumors

We have previously reported that treatment of breast tumor mice with vascular disrupting iron oxide nanoworms imposes an angiogenic gene expression signature on treatment-resistant tumors that involves upregulation of integrins and other factors, such as FGF2, CXCL10, and MMP1318. Further investigation of these treatment-resistant tumors using in vivo peptide phage display revealed an overrepresentation of phages displaying peptides containing integrin-targeting RGD motifs. Building on this insight, we hypothesized that in breast cancer models, a conventional VDA such as CA4P induces a molecular signature within the tumor microenvironment that can be effectively targeted using the clinical-stage tumor-penetrating peptide iRGD.

We first examined the impact of VDA treatment on the expression of iRGD receptors in solid tumors. Immunofluorescence imaging showed that treatment of mice bearing orthotopic MCF10CA1a triple-negative breast cancer (TNBC) with CA4P (100 mg/kg, six IP injections) resulted in a significant upregulation of β3 integrins and NRP-1 (Fig. 1a). Quantification of the fluorescent signal in CA4P-treated tumors showed a 2.4-fold increase in β3 integrin immunoreactivity (Fig. 1b) and a 7.3-fold increase in NRP-1 antigen expression (Fig. 1c) compared to the control group. This upregulation was particularly pronounced with both receptors being induced in CD31-positive blood vessels and to a lesser degree in the extravascular tumor parenchyma (Fig. 1a), suggesting the availability of receptors in this model for systemic iRGD-mediated targeting.

Figure 1
figure 1

CA4P treatment upregulates the expression of β3 integrins and NRP-1 in MCF10CA1a breast tumors. (a) Representative confocal fluorescence images of the tumors treated with CA4P or PBS. Tumors were collected on day 34 post tumor induction, sectioned, and immunostained for integrin β3 subunit, CD31 or CD105 (blood vessels), and NRP-1. The nuclei were counterstained with DAPI. Green, CD31 or CD105; red, integrin β3 or NRP-1; blue, nuclei (DAPI). Scale bar = 200 µm. Integrin β3 (b) and NRP-1 (c) fluorescence signals were normalized to the DAPI signal from the corresponding tumor sections and subsequently quantified. Statistical analysis was performed using Student's t-test, *p < 0.05; **p < 0.001. Error bars: ± standard error of the mean (SEM); n = 4 for the CA4P group quantification of integrin β3, n = 2 for the quantification of NRP-1, and n = 3 for the PBS group in both quantifications.

Full size image

We next investigated whether IP CA4P treatment upregulates the expression of receptors of the iRGD peptide in peritoneal tumors to enable direct tissue-penetrative targeting during IP therapy. As a PC model, we used an in vivo xenograft model of PC based on MKN45P cells, a sub-line of highly metastatic MKN45 cells derived from liver metastasis of poorly differentiated human gastric adenocarcinoma43. Upon IP inoculation, MKN45P cells exhibit widespread and rapid abdominal dissemination and the formation of tumor nodules in the liver, kidney, spleen, and across the mesentery. PC-bearing mice were treated with CA4P (100 mg/kg; seven IP injections every other day), followed by tumor collection, cryosectioning, and immunostaining with antibodies for iRGD receptors. Analysis of immunostained tumor sections showed that CA4P treatment resulted in a mild upregulation of αv integrins and NRP-1 (Fig. 2). The increase in expression was most evident in the periphery of the tumor (Fig. 2a,b; Fig. S4), which survives the VDA treatment and drives therapeutic resistance3,17,44,45,46. In contrast, CA4P treatment did not affect the expression of p32 protein, a receptor for the tumor-penetrating peptides LyP-147 and TT1/LinTT48,49 (Fig. 2c,d).

Figure 2
figure 2

CA4P treatment upregulates the expression of αv integrins and NRP-1 in peritoneal tumors. MKN45P tumor-bearing mice were IP-treated with CA4P (100 mg/kg, seven IP injections every other day) and PBS starting on day seven after tumor induction. After treatment, the mice were perfused, and tumors were collected, sectioned, immunostained for (a) integrin αv subunit, (b) NRP-1, and (c) p32 (red), and counterstained with DAPI for the nuclei (blue). Scale bar = 200 µm, n = 6. (d) Quantification of αv integrin, NRP-1, and p32 fluorescence signals normalized to DAPI signal in CA4P treated and non-treated (PBS) tumors. Statistical analysis was performed using the Student's t-test; error bars indicate ± SEM; *p < 0.05, **p < 0.01.

Full size image

These studies demonstrated a consistent upregulation of iRGD receptors in solid tumors after treatment with CA4P and suggest that iRGD-based strategies may facilitate the delivery of secondary therapeutics to tumors that have undergone CA4P treatment.

CA4P treatment sensitizes IP tumors to iRGD-potentiated nanoparticle delivery

The ongoing clinical development of iRGD leverages its bystander effect, with the peptide being tested as an adjunct to standard-of-care therapies for pancreatic cancer and other solid tumors29. Previously, we demonstrated that IP iRGD increased PC accumulation and penetration of coinjected small-molecule drugs. We next tested the effect of coadministered iRGD on the PC delivery of polymersomes (PS)—nanovesicles formed by the self-assembly of amphiphilic block copolymers consisting of biocompatible, biodegradable, and FDA-approved PEG and PCL polymers50,51. To simulate UTO-PS homing, FAM-labeled PS (FAM-PS) were utilized, circumventing UTO's weak fluorescence by virtue of their comparable physicochemical properties, such as hydrodynamic diameters of approximately 90 nm (Fig. S1 and S2) and zeta-potentials of − 1.5 mV for FAM-PS and − 1.1 mV for UTO-PS40.

MKN45P-bearing PC mice were treated with IP CA4P (100 mg/kg; seven injections every alternate day). Subsequently, the animals were injected IP with iRGD peptide (14 µmol/kg) and FAM-PS (160 mg of polymer/kg), either in combination or individually. Tumors were collected after 6 h, and biodistribution analyses were conducted to assess the presence and dispersion of FAM-PS in the tumor tissues (Fig. S3). Following coadministration of iRGD and FAM-PS, we observed a ~ twofold increase in the accumulation of polymersome nanovesicles in PC lesions in mice pretreated with CA4P (Fig. 3a,b). Moreover, compared to the CA4P + FAM-PS group, mice treated with CA4P + iRGD + FAM-PS showed a ~ threefold increase in PS homing, highlighting that iRGD enhanced tumor accumulation and penetration of the nanoparticles. In both CA4P-treated and control tumors, iRGD coadministration appeared to increase the tumor spreading of PS, seen as a diffuse signal in the tumor parenchyma (Fig. 3a). Quantification of the FAM-PS distribution in tumor tissue revealed that tumors treated with CA4P and targeted with iRGD exhibited the largest area of FAM-positive signal (Fig. 3c). This underscores the efficacy of sequential targeting, not only in facilitating a higher accumulation of potentially drug-loaded PS in the tumor, but also in achieving a more homogenous distribution of these particles within the tumor tissue. While VDAs are known to increase vascular permeability in tumors52, the distribution of PS in CA4P-treated tumors without iRGD was significantly lower than that in tumors treated with CA4P + iRGD + FAM-PS (30% vs. 3%, respectively). Thus, the iRGD peptide enhanced the distribution of FAM-PS in CA4P treated tumors by almost tenfold, addressing the challenge of the low penetrability of FAM-PS in tumors, which is a critical concern for effective drug delivery37,53,54. These findings underscore the applicability of a receptor amplification strategy, particularly in the case of PC tumors, for achieving an improved pharmacokinetic profile through sequential targeting.

Figure 3
figure 3

CA4P facilitates iRGD-mediated delivery of FAM-PS and improves their distribution in PC tumors. (a) Representative slidescanner images of immunostained peritoneal tumors treated with CA4P or PBS. PC mice were treated with CA4P or PBS (CA4P, 100 mg/kg; seven IP injections every alternate day) before administering FAM-PS with iRGD or PBS. After 6 h of circulation, the tumors were collected, cryosectioned, immunostained for FAM-PS (FAM), and counterstained with DAPI. Green, FAM-PS (FAM); blue, nuclei (DAPI). Scale bar = 200 µm. (b) Quantification of the FAM-PS fluorescence signal normalized to the DAPI signal in tumors. (c) Quantification of FAM-PS distribution in tumors by representing the percentage of FAM-signal-positive areas in the tumors. All tumors from the three different planes were used for analysis, with n = 3 mice per group. Statistical analyses were performed using ANOVA. Error bars indicate ± SEM; *p < 0.05, **p < 0.001.

Full size image

These findings indicate that tumors with upregulated αv integrins and NRP-1, induced by CA4P, can be targeted by nanoparticles coadministered with iRGD.

iRGD-potentiated secondary nanotherapy reduces peritoneal dissemination of cancer and increases survival of PC mice

We next performed a survival study using IP CA4P treatment in combination with iRGD-enhanced nanotherapy (groups, dosing, and schedule are outlined in Fig. 4a; and the proposed mechanism for the combination therapy is illustrated in Fig. S5). For comparative assessment of tumor burden and dissemination, we conducted an additional experimental therapy study in which the mice were sacrificed on 23rd day post tumor induction.

Figure 4
figure 4

Combined therapy attenuates metastasis of disseminated gastric tumors without causing systemic toxicity. (a) Outline of treatment regimens. Immunodeficient mice with PC of gastric carcinoma origin were IP-injected with combinations of the vascular disrupting agent CA4P (100 mg/kg, ten IP injections every other day), the tumor-penetrating peptide iRGD (14 µmol/kg, ten IP injections every other day), and nanoformulated UTO (UTO-PS) (1.4 mg/kg, ten IP injections every other day). n = 7 mice per group; PBS n = 10. Survival curves for mice treated with CA4P (b), UTO-PS (c), CA4P + UTO-PS (d), CA4P + iRGD + UTO-PS (e), or iRGD + UTO-PS (f). Panels B-F show graphs from the same treatment. Black dashed line: control group mice treated with PBS only; colored line: group treated with the corresponding therapy. In an endpoint study, mice bearing PC were treated following the same regimen and sacrificed on the 23rd day after tumor induction. (g) Body weight loss (%) of mice was monitored throughout the treatment period to assess the systemic toxicity of the therapies. CA4P + iRGD + UTO-PS combination treatment significantly reduced tumor weight (h) and total tumor burden (i) (Panels a–g and h,i represent different treatments). Statistical comparisons were performed using the Gehan–Breslow–Wilcoxon test and ANOVA. Error bars indicate ± SEM; *p < 0.05.

Full size image

Despite causing fast and dramatic shutdown of the tumor vasculature, CA4P monotherapy has not proven to be sufficiently effective for treating solid tumors55,56. In our study, the treatment of PC mice with CA4P (Fig. 4b), UTO-PS (Fig. 4c), or their combination (Fig. 4d) did not result in a significant extension of survival compared to the PBS-treated control group, underscoring the need for strategies to potentiate the accumulation and penetration of therapies. Compared to the control group, both the CA4P + iRGD + UTO-PS and iRGD + UTO-PS combination treatments improved survival by almost five days (Fig. 4e,f). Although the survival differences among mice treated with these combination therapies did not reach statistical significance, analysis of the tumor burden revealed that the CA4P + iRGD + UTO-PS combination treatment significantly decreased the overall tumor weight (Fig. 4h), and this was the only treatment that resulted in a reduction in the number of tumor nodules (Fig. 4i). CA4P and UTO-PS monotherapies also led to a reduction in the total tumor weight (Fig. 4h). However, these treatments did not result in a significant decrease in tumor dissemination, as indicated by the number of tumor nodules. Interestingly, compared to CA4P + UTO-PS treatment, experimental therapy with UTO-PS resulted in a significant reduction in tumor weight, suggesting the potential interference of CA4P with the ability of UTO-PS to penetrate tumors. This hypothesis is supported by biodistribution studies that demonstrated lower accumulation and distribution of FAM-PS following CA4P treatment (Fig. 3).

During treatment, mouse body weight decreased by 10–15% in all groups owing to the progression of PC; however, the difference in body weight between the treated and control groups was not significant, suggesting that all mice tolerated the treatments without showing signs of systemic toxicity (Fig. 4g). The safety of the therapies was confirmed by additional toxicological studies in healthy mice treated with five IP injections of 100 mg/kg CA4P, 14 µmol/kg iRGD peptide, and 1.4 mg/kg UTO-PS. None of the treatments had a significant impact on mouse body weight, thus providing additional evidence for the safety of the intervention (Fig. 5a). The post-treatment levels of the hepatic and renal enzymes alanine transaminase (ALT), creatinine (CREAT), and glucose (GLUC) showed that the levels of these metabolic and endocrine markers were not significantly altered by CA4P + iRGD + UTO-PS or iRGD + UTO-PS compared with those in the control group (Fig. 5b–d). A ~ 40% decrease in ALT after iRGD + UTO-PS treatment does not necessarily suggest hepatotoxicity, given that a ≥ threefold increase in ALT is the established threshold for indicating hepatotoxicity57. These results suggest that the treatments did not cause acute hepatic or renal toxicity or glucose-related metabolic dysfunction. As cardiotoxicity is a dose-limiting side effect of anthracycline-based chemotherapy58,59, we also evaluated the effect of this therapy on myocardial tissue. Histological examination did not reveal any differences between the treatment and control groups (Fig. 5e).

Figure 5
figure 5

Combination therapy does not induce overt toxicities. Healthy Balb/c mice received IP PBS or CA4P (100 mg/kg; five IP injections every other day) on one day, and iRGD (14 µmol/kg; five IP injections every other day) with UTO-PS (1.4 mg/kg UTO; five IP injections every other day) the other day for ten days. (a) Body weight was monitored throughout the treatment, and the percentage loss from the initial body weight is shown. On day 11, blood was collected through retro-orbital bleeding and serum was tested for alanine transaminase (b), glucose (c), and creatinine (d); n = 3. (e) Histological analysis of the extracted myocardial tissue. Scale bars = 2 mm in the upper row and 100 µm in the lower row. Representative images of three mice are shown. Statistical analyses were performed using ANOVA. Error bars indicate ± SEM; *p < 0.05.

Full size image

Discussion

Our aim was to investigate whether VDA-induced molecular changes could prime tumors for homing peptide-guided nanotherapies. We show that: (1) the administration of CA4P in mice bearing breast cancer and PC xenografts increases tumor expression of αv integrins and NRP-1, receptors of clinical-stage tumor-penetrating peptide iRGD; (2) administration of CA4P in PC mice enhances the accumulation and distribution of IP-administered nanoparticles coadministered with iRGD peptide; (3) the therapeutic efficacy of CA4P treatment, particularly in addressing peritoneal dissemination, is enhanced by subsequent treatment involving a combination of iRGD and nanoparticles loaded with a novel anthracycline, UTO; and (4) the combined treatment regimen of CA4P, iRGD, and UTO is well tolerated and does not induce overt toxicities. Collectively, these observations show that VDA treatment sensitizes tumors to iRGD tumor-penetrating peptide-based precision therapies and supports the feasibility of CA4P/iRGD-based multistage therapy.

Utilizing combination cancer therapies sequentially shows promise for achieving enhanced disease control. This concept involves employing a second drug to target a molecular change or vulnerability induced by the initial therapeutic agent. Strategies employed to sensitize tumors for a second therapeutic include modulating the tumor microenvironment to amplify drug levels and distribution60,61. In the case of CA4P treatment, the striking disruption of pre-established vessels coupled with subsequent tumor necrosis is likely to induce significant transcriptomic and proteomic shifts in surviving tumor cells. An immediate outcome of CA4P treatment is the activation of the hypoxic response, a known trigger that transcriptionally activates > 300 genes62,63. This response plays a pivotal role in angiogenesis, tumor development, and metastatic dissemination observed across various solid tumors, including breast64 and gastric cancer65,66. Whereas both αv integrins and NRP-1 are commonly expressed in solid tumors, HIF-1α-dependent mechanisms have been shown to lead to further induction of αvβ3 integrins67 and upregulation of NRP-168. Here, we observed an upregulation of NRP-1 and αv integrins in tumors following CA4P treatment. Significantly, the treatment increased tumor accumulation and penetration of polymersomes coinjected IP with iRGD peptide. This underscores that the upregulated receptors were accessible from the IP space, initiating a tumor penetration cascade. CA4P has been found to be well tolerated with antitumor effects observed at 1/10 of the maximum tolerated dose3. Further optimization of the treatment schedule and dosage of CA4P is necessary to achieve optimal induction of tumor NRP-1 and αv integrins for secondary chemotherapy. The impact of VDA treatment on the expression of αv integrins and NRP-1 likely varies across distinct tumor types and models. Further investigations in follow-up studies are needed to pinpoint tumors displaying the highest upregulation in order to identify those most suitable for CA4P combination treatment.

One concern with VDA-based interventions is the potential for increased metastatic dissemination of malignant cells post-treatment. CA4P treatment can elevate vascular permeability both directly by disrupting endothelial cell microtubules, compromising their barrier function52, and indirectly through a hypoxia-driven angiogenic response69. This heightened vascular permeability may facilitate tumor cell extravasation and metastasis70,71. Conversely, potent tumor blood vessels are essential for metastasis72, and vessel shutdown following VDA treatment limits this pathway at the tumor core. Thus, vascular disruption can both promote and inhibit tumor metastatic spread. While our findings did not show a significant increase in tumor metastasis after CA4P treatment (Fig. 4i), future studies are needed to thoroughly address this issue.

Given that the clinical standard-of-care for PC involves cytoreductive surgery (CRS) followed by IP chemotherapy73, we chose the IP administration route for CA4P, iRGD, and NPs in our study. Nanoencapsulation can enhance the drug residence time in the abdominal cavity and slow systemic absorption to target PC lesions74. Our current study builds on previous studies, where we have demonstrated that PS were effective in IP therapy because of their preferential accumulation in tumors at levels higher than that achievable by IV injections31,75. In another study, we showed that IP-administered iRGD and iRGD-functionalized and pH-sensitive PS accumulated in MKN45P tumors via direct penetration from the peritoneal cavity and circulation-mediated access31,32. The IP route of CA4P administration aligns well with the treatment approach for peritoneal dissemination, a condition also associated with ovarian cancer, for which CA4P has been evaluated in clinical trials76. Moving forward, the goal is to evaluate the effectiveness of our strategy across a broader spectrum of tumor types to validate its potential applicability across diverse cancer models.

For fluorophore-labeled iRGD, ~ 75% of IP-administered peptide reaches tumors through the IP space, whereas the rest reach tumors systemically after absorption from the IP cavity32. This allowed balanced targeting of early-stage avascular tumor nodules (through direct peritoneal penetration) and advanced vascularized tumor nodules (through a combination of direct and circulation-mediated targeting) for better disease control. In the current study, reduced total tumor count and thereby metastasis was observed only with the CA4P + iRGD + UTO-PS combination therapy. Integration of such a combination treatment with CRS may hold promise for eliminating microscopic tumor nodules left behind during CRS to limit the recurrence of postoperative PC77. The anticancer drug used in this study, UTO, is a novel potent anthracycline designed to minimize the risk of cardiotoxicity40, which is a significant concern in anthracycline treatment58. Nanoencapsulation of hydrophobic UTO protects the prodrug from premature degradation and slow release, leading to an extended drug exposure time to peritoneal lesions following IP chemotherapy74,78,79. Notably, UTO release from PS remained below 2% over 48 h, preventing premature drug release40. In the current study, combination treatment demonstrated a safe pharmacodynamic profile and no signs of acute cardiotoxicity. Here, we used direct targeting of malignant PC lesions from the intraperitoneal space, bypassing the need for tumor perfusion. To broaden our understanding of the sequential targeting approach and optimize intervention regimens for tumors with different vascularization statuses, systematic future studies on various tumor models with different perfusion statuses and administration routes (IV vs. IP) are required.

Among tumor-homing peptides80, iRGD stands out for its ability to facilitate the tumor-specific delivery of anticancer agents in both coinjected and tethered forms23,24. The coadministration mode, used here and currently under evaluation in phase I and II clinical studies on pancreatic cancer and other solid tumors, offers advantages, such as simplicity and the ability to enhance standard-of-care therapies without limitations on the delivery capacity due to the number of accessible receptors81. On the other hand, conjugation of iRGD to nanoparticles results in multivalent peptide presentation, potentially enhancing cellular transcytosis and penetration of tissue barriers through receptor clustering. The multivalent presentation of targeting ligands can amplify ligand affinity by several orders of magnitude because of the simultaneous engagement of receptor-binding sites on the cell surface82,83. Tethering peptides onto nanoparticles also provides the opportunity to incorporate two or more different homing peptides, aiming to boost affinity and selectivity for cells expressing multiple target receptors. It can be envisioned that in the context of CA4P-potentiated delivery, the relevant additional peptides recognize targets, such as PL184 and PL385 peptides that bind to Tenascin-C, and LinTT1 and LyP-1 peptides that target cell surface p3247,48, and mUNO86 binding to tumor-associated macrophages87. Future research is needed to investigate the biodistribution and efficacy of iRGD-functionalized nanoparticles in combination with VDA therapy.

Conclusions

This study shows that VDA treatment induces signatures that can be targeted by iRGD tumor-homing peptide-potentiated nanotherapy. Our study highlights the potential of combination therapy with CA4P, iRGD, and UTO-PS to achieve more efficient and better tolerance for the management of PC. Further studies to optimize the mode of targeting (coadministration vs. conjugated), treatment schedule, and dosage are necessary to achieve the optimal anticancer effect and allow for full exploitation of the VDA-enhanced precision nanotherapy.