Introduction

Pancreatic cancer is a leading cause of cancer death worldwide and its global burden has increased dramatically over the past years1,2. The latest statistics show that the 5-year overall survival rate of pancreatic cancer is merely 12.5% (https://seer.cancer.gov), which is much lower compared with many other cancer types, rendering it a major medical challenge2,3. Pancreatic cancer is commonly characterized by severe hypoxia regions, it represents one of the most hypoxic cancer with < 3 mmHg pO2 in portions of tumor tissues, more than a ten-fold decrease compared to normal tissues4,5. Hypoxia contributes to the aggressiveness of pancreatic ductal adenocarcinoma (PDAC) by provoking malignant epithelial-mesenchymal transition6, enrichment of cancer stem cell population7, and strengthened glycolysis8,9,10. Besides, hypoxia predicts aggressive growth in pancreatic cancer xenografts11 and the hypoxia-inducible transcription factors, HIF-1α, is considered a predictor of clinical outcome in patients with PDAC12. Consequently, hypoxia-targeted therapeutics have emerged as promising options in the precisive treatment of pancreatic cancer13,14. Although candidates targeting hypoxia-related proteins (HIF-1α15,16,17,18 or mTOR19,20) or glycolysis21,22 proved effective in laboratory investigations or early phase clinical studies, none of them has paved its way into the market as anti-PDAC drug to the best of our knowledge13,14.

When O2 is limited, certain chemical functional groups have the potential to be metabolized by enzymatic reduction23. Bioreductively activated prodrugs converting hypoxia-sensitive cytotoxins to their active forms primarily target DNA24,25 and there is already clinical evidence for their activities against pancreatic cancer24. Nevertheless, the outcomes of clinical phase III studies on bioreductive prodrugs are disappointing26. Besides the limited extravascular penetration of prodrugs, the other key weakness is that their activation is largely dependent on the reductase activity and/or cellular reduction potential in cancer cells. However, there is individual variability among clinical patients, and the clinical efficacy of the bioreductive prodrug cannot be guaranteed24,26,27.

Thereafter, hypoxia-selective medications that work through different modes of action are eagerly warranted, not only for therapeutic purposes but for the discovery of hypoxia-related mechanisms underlying the resistance and malignant progression of tumor. However, such compounds are rare. For example, after screening 20,000 different cultivated broths of microorganisms, Ikeda et al. discovered that rakicidin A (a 4-amido-2,4-pentadieneoate (APD) cyclodepsipeptide displayed moderate hypoxia-selective cytotoxicity28. In 2017, Poulsen et al. revealed that another type of APD cyclodepsipeptide BE-43547 compounds also exhibited hypoxia-selective cytotoxicity with a significant up to 60-fold decreased IC50 value against pancreatic cancer cells (PANC1) under hypoxia compared to normoxia29. Subsequently, they made a breakthrough in demonstrating that rather than the regular DNA targeting pattern, APD cyclodepsipeptides induced fast collapse of mitochondrial function and ultrastructural integrity in hypoxic cancer cells30. More recently, we synthesized the BE alkynyl probe upon which we disclosed that BE covalently binds the cysteine234 residue of eukaryotic translation elongation factor 1 alpha 1 (eEF1A1) to exert its anti-pancreatic cancer effects31. Following this, eEF1A1 was found highly expressed under hypoxia in pancreatic cancer cells32 and acts a vital role in regulating the stemness of pancreatic cancer cells31. Besides, utilizing 99 clinical specimens of pancreatic cancer patients, eEF1A1 protein levels are found positively correlated with pancreatic cancer stage but negatively correlated with patient survival31. This highlights the importance of eEF1A1 in the progression of pancreatic cancer. Specifically targeting eEF1A1, BE may serve as a promising lead for hypoxia target therapy in pancreatic cancer treatment.

We have worked out synthetic routes that could supply BE and other structure similar natural products or their derivatives33,34,35,36,37,38,39 for in vivo experiments29,40,41,42,43,44 refers to the synthetic study of other researches. However, low water solubility and toxicity of BE limited its further application as anticancer reagent and there is a clear need to develop a functional formulation of BE to reduce its side effects and improve its anticancer effects. As a supramolecular carrier, calixarenes have been used as a molecular vessel to transport therapeutic medications into tumors, thereby enhancing therapeutic efficacy and/or alleviating side effects45,46,47. In contrast to conventional nanoscale drug-delivery systems, calixarenes feature well-defined molecular structures and weight as well as operational simplicity, which could assure batch-to-batch consistency through rigorous manufacturing procedures46,47,48,49,50. Moreover, the unusual properties of tunable cavity size and convenient modification empower these macrocyclic hosts with intriguing molecular recognition ability, thereby quantitatively binding a variety of drug guests50,51,52. In our issue, an ideal calixarene host should own the following specific characteristics: (i) high binding affinity between host and BE to prevent unwanted leaking45,53, and (ii) efficient delivery to tumors by targeting the characteristics of tumor microenvironment50,54,55,56,57. As a result, the key challenge emerges to be designing promising calixarene macrocycles with good water solubility, excellent biocompatibility, excellent binding affinity to guest molecules and targeting ability.

In this work, we proposed a fusion of covalent and non-covalent drug delivery strategy for the treatment of pancreatic cancer. The active and susceptible covalent binding site of BE was temporarily masked by N-methyl-piperizine via a 1,6-conjugate addition to generate an amine-adducted prodrug N-methyl-piperizine-BE-43547A2 (NMP-BE). With similar inhibitory activities against PANC-1 cells and reduced toxicity against normal cells at certain concentrations. NMP-BE is further capable of being encapsulated non-covalently into a complex with sulfonated azocalix[5]arene (SAC5A). The azo groups of azocalixarene are hypoxia-responsive50,54, and can be reduced by overexpressed azoreductases in hypoxic cancer cells. SAC5A can promote the accumulation of payload toward the tumors. When NMP-BE is unloaded within cancer cells, it could subsequently release the hypoxia-sensitive toxin BE to achieve a host-guest dual hypoxia-responsive therapeutic purpose (Fig. 1). This innovative formulation was tested in vitro and further in vivo using a PANC-1 xenograft murine model. NMP-BE@SAC5A significantly suppressed tumor growth and boosted the antitumor efficacy of NMP-BE, accompanied by no systemic toxicity observed.

Fig. 1: Conceptual illustration of the dual hypoxia-responsive supermolecular complex for cancer therapy.
figure 1

The supermolecular prodrug NMP-BE@SAC5A is formed by the host-guest complexation of hypoxia-responsive molecular container SAC5A with NMP-BE. The azo groups of SAC5A quickly respond under hypoxic condition, leading to the release of loads within tumors. NMP-BE can subsequently release hypoxia-sensitive toxin BE to exert anticancer effects.

Full size image

Results

Design and preparation of NMP-BE, SAC5A and NMP-BE@SAC5A

We formerly used N-methyl-piperizine (NMP) for the 1,4-conjugate addition on ovatodiolide and its derivativeSamples

The pH 7.4 PBS solution was made by dissolving precise quantities of sodium phosphate monobasic dehydrate, disodium phosphate, sodium chloride, and potassium chloride in double-distilled water. The volume was then adjusted to 1000 mL with double-distilled water. The samples of NMP-BE@SAC5A were prepared by grinding.

Apparatus

1H and 13C nuclear magnetic resonance (NMR) spectra were recorded on a Bruker AV400 spectrometer or a Zhongke-Niu** BIXI-I 400 spectrometer for structural characterization of compounds. Fluorescence measurements were recorded on a Cary Eclipse for the fluorescence titrations. Fluorescence microscopy images were examined using a confocal laser scanning microscope (Leica TCS SP8). Tissue sections were observed and photographed with an optical microscope (CX41, Olympus, Japan).

In vitro release rate of BE-43547A2 (BE) from prodrug NMP-BE

NMP-BE was weighed and dissolved in HEPES (0.01 M) with pH of 7.4, then prepared into a solution with a concentration of 5 μg/mL. The buffer was extracted at different time points, and the peak areas of BE were detected by high performance liquid chromatography (HPLC), and the concentration of BE was calculated by external standard method. HPLC conditions: Shimadzu 20AT high performance liquid chromatography system; the mobile phase is 90% acetonitrile (0.1% H3PO4) and 10% water (0.1% H3PO4); the flow rate is 1 mL/min; the detector is UV detector; detection wavelength 210 nm for NMP-BE and 254 nm for BE; chromatographic column is C18 column.

Data analyses of fluorescence titrations

Fluorescence titrations of SAC5A, SC5A79, and SBE-β-CD were performed in 10 mM PBS, pH 7.4. The complexation of SAC5A with reporter dye (RhB) was measured by direct fluorescence titrations. A mixed solution containing known amounts of SAC5A and RhB was sequentially injected into 2.50 mL RhB solution in a quartz cuvette. The dye concentrations in mixed solution and cuvette are the same to keep dye concentration constant in the course of titrations. The fluorescence intensity was measured (λex = 554 nm for RhB) before the first addition and after every addition until a plateau was reached. By fitting the fluorescence intensity (λem = 575 nm for RhB) according to a 1:1 host-guest binding stoichiometry, the association constant was obtained. SC5A (LCG) and SBE-β-CD (NR) used the same method.

The complexation of SAC5A with NMP-BE was measured by competitive fluorescence titrations. A mixed solution containing known amounts of reporter dye (RhB), host (SAC5A) and competitive guest (NMP-BE) was injected into 2.50 mL RhB and SAC5A solution in a quartz cuvette. Care was taken to keep the concentrations of dye and SAC5A constant in the course of titrations. The fluorescence intensity was measured (λex = 554 nm for RhB) before the first addition and after every addition. The association constant was obtained by fitting fluorescence intensity (λem = 575 nm for RhB) according to a 1:1 competitive binding model. The fitting of data from direct titrations and competitive titrations was performed in a nonlinear manner, and the fitting modules were downloaded from the website of Prof. Nau’s group (http://www.jacobs-university.de/ses/wnau) under the column of “Fitting Functions”.

Cell culture

PANC1 and hTERT-HPNE cell lines were purchased from BeNa Culture Collection (Bei**g, China, BNCC352264, BNCC338221), authenticated by STR profiling and tested for mycoplasma contamination. All cells were tested for mycoplasma contamination and had no mycoplasma contamination. None of the cell lines used are classified as commonly misidentified lines. Cells were cultured in DMEM medium (Corning, Manassas, VA, USA) supplemented with 10% fetal bovine serum (FBS, Gibco Life Technologies, Grand Island, NY, USA) in a humidified incubator with 5% CO2 at 37 °C. A humidified atmosphere containing 5% CO2 was used as normoxic cell culture environment. The hypoxic cell culture environment was adjusted by purging gas mixture (94% N2, 5% CO2, 1% O2).

In vitro cytotoxicity assays

To determine the cell viabilities of NMP, BE, and NMP-BE, PANC1 cells (5 × 103 cells/well) were seeded into the 96-well plate and incubated overnight. Then various concentrations of NMP, BE or NMP-BE were added to cells. After 72 h, 20 μL MTT solution (5 mg/mL) was added and incubated for 4 h. The OD values were determined at 570 nm by using a micro-plate reader. And the IC50 were calculated by GraphPad Prism. To determine the hypoxic selectivity of NMP-BE, PANC1 cells (5 × 103 cells/well) were seeded into the 96-well plate. After 24 h, the cells were treated with NMP-BE of different concentrations and incubated for 24 h under normoxic or hypoxic conditions, respectively. And the cell viabilities were measured by MTT assays. To determine the cell viabilities of BE, NMP-BE, and NMP-BE@SAC5A on normal cell line, HPNE cells (5 × 103 cells/well) were seeded into the 96-well plate and incubated overnight. Then various concentrations of BE, NMP-BE or NMP-BE@SAC5A were added to cells. After 24 h treatment of compounds, cell viabilities were measured by MTT assays.

The cytotoxicity assays of SAC5A and SBE-β-CD and the anti-proliferation ability of NMP-BE, NMP-BE@SBE-β-CD, and NMP-BE@SAC5A were measured by cell counting kit-8 assays. PANC1 cells (5 × 103 cells/well) were seeded into the 96-well plate and incubated overnight. They were treated with SAC5A or SBE-β-CD of different concentrations and incubated for 24 h under normoxic or hypoxic conditions, respectively. In addition, the cells were treated with NMP-BE, NMP-BE@SBE-β-CD, and NMP-BE@SAC5A of different concentrations. After a 6 h incubation under normoxic conditions, the culture medium was exchanged with fresh medium. Subsequently, the cells were incubated at 37 °C for 18 h under either normoxic or hypoxic conditions, respectively. And then, the culture medium was replaced with fresh medium and 10 μL cell counting kit-8 solution (C6005, US Everbright, Jiangsu, China). Following a 4 h incubation, the optical density was assessed at 450 nm employing a microplate reader.

Confocal laser scanning microscopy (CLSM)

PANC1 cells (1 × 105 cells/well) were seeded into the confocal imaging chambers. After 24 h, the cells were treated with CY5-DM, CY5-DM@SBE-β-CD, and CY5-DM@SAC5A (10/10 μM). After a 6 h incubation under normoxic conditions, the culture medium was exchanged with fresh medium. Thereafter, the cells were incubated at 37 °C for 18 h under either normoxic or hypoxic conditions, respectively. Subsequently, the cells were fixed in 4% paraformaldehyde (PFA; Biosharp, Hefei, China) for 15 min, washed with PBS buffer three times and then imaged using CLSM. Cell nuclei were counterstained with DAPI (C0060, Solarbio, Bei**g, China) for 5 min.

Cellular uptake of SAC5A

To examine the cellular uptake mechanism of SAC5A complex, PANC1 cells (3 × 105 cells/well) were seeded into the six-well plates and incubated overnight. The cells were pretreated with different endocytosis inhibitors for 1 h: chlorpromazine (CHP, 20 μM, an inhibitor of clathrin-mediated endocytosis), Amiloride (AMI, 500 μM, an inhibitor of macropinocytosis), Genistein (GEN, 200 μM, an inhibitor of caveolae-mediated endocytosis), methyl-β-cyclodextrin (M-β-CD, 5 mM, an inhibitor of lipid rafts-mediated endocytosis), the cells were also pre-incubated under 4 °C for 1 h (energy-dependent endocytosis). Next, the cells were treated with CY5-DM@SAC5A (10/10 μM) for another 1 h. After that, the cells were collected and washed three times with PBS, and then analyzed by flow cytometry. All experiments were carried out four times80.

To verify the co-localization of SAC5A and lysosome, PANC1 cells (2 × 104 cells/well) were seeded into confocal dish and incubated overnight. Then the cells were treated with CY5-DM@SAC5A (10/10 μM) for 1, 3, and 6 h, respectively. After that, cells were stained with Lyso Tracker (50 nM) for 30 min at 37 °C, and then stained with DAPI. The cells were washed with PBS buffer and imaged using CLSM.

In vivo fluorescence imaging

For animals and tumor model, male BALB/c nude mice at 5–6 weeks were purchased from Vital River Laboratory Animal Technology (Bei**g, China). All animal experiments were performed according to the institutional animal care guidelines established by the Institutional Animal Care and Use Committee of Nankai University. To establish the PANC1 tumor-bearing mouse model, 1 × 106 PANC1 cells were injected subcutaneously into the right chest of BALB/c nude mice. After the tumor grows to a suitable size, the tumor was divided into small pieces and then subcutaneously incubated into the right chest of other mice. When the tumor volumes of mice were around 200 mm3, the mice were randomized into three groups and intravenously injected with 100 μL of CY5-DM (200 μM), CY5-DM@SBE-β-CD (200 μM), and CY5-DM@SAC5A (200 μM). In vivo fluorescence imaging of CY5-DM, CY5-DM@SBE-β-CD, and CY5-DM@SAC5A were imaged by IVIS Lumina imaging system (Caliper Life Science, USA) at the time of 1, 3, 6, 12, and 24 h after injection, respectively. The ex vivo fluorescence imaging of major organs at 24 h post injection was imaged by IVIS. Fluorescent images were analyzed using Living Image 3.1 (Caliper Life Sciences). (Since cyanine dyes can be distributed in many organs due to their non-targeting nature, notably in the kidney. If the tumor was inoculated in the flank, the fluorescence signals of tumor might overlap with the signals of kidney and abdominal area during the living imaging. Therefore, to observe the distribution of the fluorescence signals overspread their bodies, the tumor was inoculated in the chest).

Acute toxicity of NMP-BE fumaric acid

For acute toxicity of NMP-BE fumaric acid, the mice were intravenously administered with different dosage of compound (0, 10, 25, 50 mg/kg) for 10 days, respectively. Body weights were measured before administration and at daily intervals after administration (n = 3).

Experimental procedure of the pharmacokinetic (PK) study of NMP-BE

All UPLC-MS/MS analysis was carried out on an Ultra performance liquid chromatographic system (ExionLC, SCIEX, USA). The chromatographic column was Waters XSelect® HSS T3 C18 (2.1 × 50 mm, 2.5 μm). Mass spectrometric detection was performed on Triple QuadTM 4500 mass spectrometer with an electrospray ionization (ESI) interface operating in positive ion mode, was manufactured by SCIEX (Framingham, USA). The MS/MS system was operated at unit resolution in the multiple reaction monitoring (MRM) mode, and the monitored transitions were m/z 664.6 → 196.4 for NMP-BE, m/z 564.5 → 167.2 for BE. Male CD-1 mice at 6–8 weeks were purchased from Vital River Laboratory Animal Technology (Bei**g, China). Mouse were housed at 22 ± 2 °C and 55 ± 5% (relative humidity) under a 12 h light-dark cycle. Blood samples were collected at 2, 5, 15, 30, 45 min and 1, 2, 4, 8 and 24 h post-dose into heparinized tubes. Plasma was obtained after centrifugation and stored at –80 °C until they were analyzed. The plasma concentration-time profiles of NMP-BE and BE in mouse were plotted.

Detection of tumor hypoxia

To evaluate the tumor hypoxia at the early stage in PANC1-bearing BALB/c nude mice, immunofluorescence analysis was conducted using a HypoxyprobeTM−1 Omni Kit (HP3−100; Hypoxyprobe, Burlington, MA, USA). Pimonidazole HCl (Hypoxuprobe-1, 60 mg/kg) was injected intraperitoneally 1 h before they were killed. Staining was performed according to the manufacturer’s instructions. The tissue sections were cut and fixed in cold acetone for 10 min. The sections were rinsed and incubated overnight at 4 °C with rabbit anti-pimonidazole antibodies (PAb2627AP, 1:100). The sections were then incubated for 2 h with FITC-conjugated goat-anti-rabbit antibody (Thermo, 1:1000). Between all steps of the staining procedure, the sections were rinsed three times with for 5 min in PBS and imaged by CLSM.

In vivo anticancer efficacy study

For in vivo anticancer efficacy study, 1 × 106 PANC1 cells were injected subcutaneously into the right flanks of the BALB/c nude mice. To investigate the anticancer effects of SBE-β-CD@NMP-BE, the mice with tumor volumes at around 50 mm3 were randomized into four groups (6 mice per group) and injected intravenously via tails with 200 μL of PBS, NMP-BE (5 mg/kg), SBE-β-CD (15 mg/kg) and NMP-BE@SBE-β-CD (20 mg/kg) every two days for 8 times and the tumor volumes for 14 days were continuously monitored. Tumor were measured by using a Vernier calipers and the volume (V) was calculated to be V = d2 × D/2, where d is the shortest and the D is longest diameter of the tumor in mm respectively. To investigate the anticancer effects of NMP-BE@SAC5A, the mice with tumor volumes at around 50 nm3 were randomized into four groups (6 mice per group) and injected intravenously via tails with 200 μL of PBS, NMP-BE (5 mg/kg), SAC5A (15 mg/kg), NMP-BE@SAC5A (20 mg/kg) every two days for 10 times and the tumor volumes for 14 days were continuously monitored. Tumors were measured as mentioned above. To assess potential toxicities, mice were monitored for weight loss. Tumors were collected for H&E analysis and immunofluorescence staining. Major organs including heart, liver, spleen, lung, and kidney, were collected and stained with H&E for histopathologic analysis. To evaluate biosafety of SAC5A, blood samples were collected for blood chemistry assay and blood routine assay.

Statistical analysis

All results are presented as the mean ± standard deviation (SD) as indicated. Data were analyzed by one- or two-way analysis of variance (ANOVA) of comparison of multiple groups using the GraphPad Prism. P values less than 0.05 were considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.