Paclitaxel is widely used in cancer treatments, but poor water-solubility and toxicity raise serious concerns. Here we report an RNA four-way junction nanoparticle with ultra-thermodynamic stability to solubilize and load paclitaxel for targeted cancer therapy. Each RNA nanoparticle covalently loads twenty-four paclitaxel molecules as a prodrug. The RNA-paclitaxel complex is structurally rigid and stable, demonstrated by the sub-nanometer resolution imaging of cryo-EM. Using RNA nanoparticles as carriers increases the water-solubility of paclitaxel by 32,000-fold. Intravenous injections of RNA-paclitaxel nanoparticles with specific cancer-targeting ligand dramatically inhibit breast cancer growth, with nearly undetectable toxicity and immune responses in mice. No fatalities are observed at a paclitaxel dose equal to the reported LD50. The use of ultra-thermostable RNA nanoparticles to deliver chemical prodrugs addresses issues with RNA unfolding and nanoparticle dissociation after high-density drug loading. This finding provides a stable nano-platform for chemo-drug delivery as well as an efficient method to solubilize hydrophobic drugs.

Historically, chemotherapy plays a significant role in cancer treatment. Many chemotherapeutic drugs, including paclitaxel (PTX), have been used clinically in cancer patient care, but their toxicity and water-insolubility have raised serious concerns1,2. Effective therapeutic dosages of PTX cause potential organ damage, myelosuppression, immunotoxicity, hypersensitivity, and neurotoxicity2,3. These undesired effects have challenged cancer chemotherapy. Due to low water solubility and permeability, PTX was categorized as the lowest bioavailability Class IV drug according to the Biopharmaceutical Classification System (BCS). The first formulation of PTX for intravenous (i.v.) administration was a blend of Cremophor EL (castor oil derivative) and dehydrated ethanol at 1:1 (v/v) as a solvent4. However, subsequent studies revealed that castor oil and its derivatives are toxic and caused severe side effects such as non-linear plasma pharmacokinetics due to entrapment and delayed clearance in the liver5,6.

Functional nanomaterials constructed via molecular self-assembly show enormous potential in biotechnology and biomedicine applications7,8,9,10. Nanoparticle-based delivery of small chemotherapeutics via biochemical conjugation, complexation, or prodrug formulation has demonstrated to be a promising strategy for cancer therapy, due to their capacity of refining the poor physicochemical properties of chemo-drugs11. Over the years, the field of RNA nanotechnology has emerged rapidly7,12,13. RNA was widely used as a unique and biocompatible material for constructing nanoparticles via bottom-up self-assembly14,15,16,17,18,19. The versatility of RNA nanostructures, including composition, structure, and functionality, can be precisely controlled. Typically, RNA nanoparticles can be conveniently functionalized with tumor-targeting ligands to achieve selective targeting and specific delivery to tumor14,20,21. This advantage allows the maximum utilization of therapeutics after systemic administration, as well as minimizing side effects caused by non-specific accumulation in other organs. As a result, therapeutic modules such as siRNA and anti-miRNA were conjugated to RNA nanoparticles and exhibited efficient tumor inhibition in various cancer models20,21. In addition, recent studies show that the immune responses of RNA nanoparticles, based on the phi29 packaging RNA, are size, shape, and sequence-dependent22. Thus, RNA nanoparticles can be tuned not to trigger immunostimulation, or to induce strong immune responses that can be potentially used for cancer immunotherapy or as vaccine adjuvant22,23,24. Taking these advantages into consideration, small chemotherapeutic drugs utilizing RNA nanoparticles as a drug delivery platform could exhibit tumor-specific therapeutic effects and reduce the side effects and toxicity that is currently observed.

The concept of thermodynamic stability, defined as the free energy, ΔG°, required for complex formation, is important in RNA nanotechnology12. We previously discovered a thermodynamically stable phi29 pRNA three-way junction (3WJ) motif which has been used as a scaffold to construct multifunctional RNA nanoparticles25. This multivalent RNA can be modified with various chemical moieties26, which makes the covalent conjugation of chemo-drugs to RNA nanoparticles achievable27. However, increasing the drug loading capacity remains challenging for the current pRNA-3WJ nanoparticles. The limited drug loading capacity can be caused by steric hindrance that interferes with assembly, unfolding of RNA 2D or 3D structures after conjugation, or disassociation of RNA nanoparticles due to reduction of thermostability.

To address these concerns, we deliberately redesign our RNA vehicle to further elevate its thermostability and maximize the number of drugs that can be incorporated in the nanoparticle. An RNA four-way junction (4WJ-X) is constructed to achieve high-yield loading of drugs, as well as significantly solubilize the hydrophobic PTX for tumor inhibition. The engineered RNA 4WJ-X nanostructure is evolved from the pRNA-3WJ motif25 with much higher thermodynamic stability, displaying an extraordinary annealing temperature (Ta) higher than 80 °C, far superior to pRNA-3WJ. Therefore, it completely addresses the concerns of reduced assembly efficiency as well as unfolding and dissociation over time after drug conjugation. The 4WJ-X nanostructures covalently conjugated with and without PTX are successfully formed with high structural stability and rigidity, as revealed by single-particle cryo-EM and 3D reconstruction. The resulting RNA-PTX complex retains its tertiary folding and independent functionalities both in vitro and in vivo, as demonstrated in the inhibition of triple negative breast cancer (TNBC) growth in animal trial. Currently, the major challenge in cancer therapy using PTX is not the anticancer drug efficacy, but its high toxicity. In this study, the RNA-PTX nanoparticles display undetectable toxicity or immune stimulation in mice. Our finding highlights the strong potential for using thermostable RNA-PTX nanoparticles as an efficient anticancer drug carrier.

Construction of ultra-thermostable RNA nanostructure

Attachment of chemicals or bioconjugation of foreign payloads to an RNA oligomer tends to induce steric constraints, and physical hindrance leading to annealing temperature (Ta) alteration. Thus, using the previous RNA nanoparticles to carry sufficient chemo-drugs for cancer regression is challenging, due to reduced assembly efficacy and thermostability of the anticipated nanostructure after high-density loading with the drugs. As a means to improve RNA nanotechnologies’ capability in delivering high payloads of chemotherapeutic drugs, we intentionally evolved our nanostructure from pRNA-3WJ25 to 4WJ28 to maximize its thermodynamic stability as well as its drug loading capacity. The RNA nanoparticle used in the present work is a synthetic RNA 4WJ motif, with a hypothetical “X” shaped structure, composed of four RNA oligomers (Fig. 1a, b). It contains four 20 base-pair (bp) helices that are joined together at the junction domain. Each helix contains two functional domains: the core domain, and the payload domain. The nanoparticle core, the interior 8 bp of the helices, is designed to optimize the nanoparticle’s thermodynamic stability. The payload domain, the outer 12 bp of the helices, is designed to carry a high capacity of therapeutics with minimal effect on nanostructure stability. As a result, the RNA 4WJ-X displayed a larger hydrodynamic average diameter of 8.7 ± 1.9 nm, compared with 4.9 ± 1.4 nm for pRNA-3WJ (mean ± SD of one size distribution) (Fig. 1c). The Ta of 4WJ-X nanostructure (80.9 ± 2.7 °C at 2.5 μM, 66.1 ± 0.6 °C at 1 μM) was found to be 22 °C higher than that of pRNA-3WJ (58.4 ± 0.7 °C at 2.5 μM, 44.5 ± 1.6 °C at 1 μM) (n = 3 independent samples over three independent measurements, mean ± SD; representative annealing curves are shown in Fig. 1d). Next, 2′-propargyl modifications are incorporated at optimal intervals throughout the payload domain to introduce maximum alkyne moieties on the RNA oligomers for “click” coupling of azide (N3)-modified prodrugs without significant influence on assembly efficiency and thermostability. We then evaluated the capability of this ultra-thermodynamically stable RNA 4WJ-X nanostructure to carry high payloads of drug.

a Hypothetical 3D model of RNA 4WJ-X nanostructure (left), redesigned from pRNA-3WJ motif (right)25,29. b Step-wise self-assembly of 4WJ-X, evaluated by native polyacrylamide gel electrophoresis (PAGE) (size marker: ultra-low range DNA ladder; M, D, and T indicate RNA monomer, dimer, and trimer, respectively). c Size comparison of 4WJ-X (purple) and pRNA-3WJ (pink) by dynamic light scattering (DLS) (mean ± SD of one size distribution). d Ta comparison of 4WJ-X (purple) and pRNA-3WJ (pink) in representative annealing curves, measured by real-time-PCR (RT-PCR) (n = 3 independent samples over three independent measurements). Source data are provided as a Source Data file.

An RNA 4WJ-X nanostructure with ultra-thermodynamic stability was constructed for high-yield loading of PTX for targeted cancer therapy. Each RNA nanoparticle loaded with 24 copies of covalently linked PTX was used as a prodrug, while remaining ultra-thermostability for in vivo applications. The water-solubility of PTX with RNA as the carrier was greatly enhanced in comparison to free PTX. Single-particle imaging by cryo-EM clearly showed the formation of “X” shaped RNA nanoparticles at the sub-nanometer scale. Animal trials in triple negative breast cancer model revealed that the RNA nanoparticles incorporated with anti-EGFR aptamer can specifically target tumor and significantly inhibit tumor growth after systemic injections. Currently, the hurdle in using PTX for cancer patient care is not the issue of PTX’s efficacy but its toxicity. The RNA-PTX nanoparticles here displayed extremely low or undetectable toxicity or immune responses in mice. This work demonstrates an effective method to address the insolubility of hydrophobic drugs in patient care. More importantly, it demonstrates a promising nano-platform for chemotherapeutic drug delivery by using thermostable RNA nanoparticles with the capabilities of high-yield drug payload, tumor-specific targeting, and reduced adverse effects for cancer therapy.

Synthesis and conjugation of RNA-PTX

PTX-N3 prodrug was synthesized by a reaction mixture of paclitaxel (Alfa Aesar), 6-azido-hexanoic acid (Chem-IMPEX), N,N′-dicyclohexyl-carbodiimide (Acros Organics), and 4-(dimethylamino) pyridine (Sigma-Aldrich) at 1:2:2:1 equivalent ratio in 20 mL dichloromethane as solvent. The reaction was run at room temperature with stirring for 36 h, followed by filtration and rotary evaporation to yield crude product. The crude product was purified by silica gel chromatography using n-Hexane:ethyl acetate as eluent. RNA-6 alkynes oligomers (one terminal 5′-alkyne and five internal 2′-propargyl) were synthesized via standard solid-phase RNA synthesis26 using 5′-hexynyl phosphoramidites (Glen Res. Corp.) and 2′-propargyl phosphoramidites (ChemGenes Corp.). RNA oligomers were purified by desalting using Glen Pak purification cartridges (Glen Res. Corp). The RNA sequences (lower-case letters indicate 2′-F nucleotides; letters labeled with * indicate 5′-hexynyl or 2′-propargyl nucleotides; underlined letters indicate anti-EGFR aptamer) are:

1. (1)

   3WJ a: 5′-uu*G ccA uGu* GuA uGu* GGG-3′

   
   

2. (2)

   3WJ b: 5′-c*cc AcA u*Ac uuu Gu*u GAu cc*-3′

   
   

3. (3)

   3WJ c: 5′-GGA u*cA Auc* AuG Gc*A A-3′

   
   

4. (4)

   4WJ-X a: 5′*-uuA GGu* AAA Gc*c Acc uGc AGG uGc uAc c*GA uGu* AAu uc*A A-3′

   
   

5. (5)

   4WJ-X b: 5′*-uuG AAu* uAc Au*c GGu AGc AcG GGc uGu Gc*G AGG c*uG AAc* AG-3′

   
   

6. (6)

   4WJ-X b-5′NH2: 5′NH2-C6-u*uG AAu* uAc Au*c GGu AGc AcG GGc uGu Gc*G AGG c*uG AAc* AG-3′

   
   

7. (7)

   4WJ-X c: 5′*-cuG uuc* AGc cu*c GcA cAG ccA GcA c*Gc Acc* uGA Au*A GG-3′

   
   

8. (8)

   4WJ-X c-EGFRapt: 5′*-cuG uuc* AGc cu*c GcA cAG ccA GcA c*Gc Acc* uGA Au*A GGu Gcc uuA GuA AcG uGc uuu GAu Guc GAu ucG AcA GGA GGc-3′

   
   

9. (9)

   4WJ-X d: 5′*-ccu Auu* cAG Gu*G cGu Gcu GGG cuG cAG Gu*G Gcu u*uA ccu* AA-3′

   
   

Conjugation of PTX to RNA was performed using copper(I)-catalyzed alkyne-azide cycloaddition (“Click chemistry”). RNA-6 alkynes resuspended in diethyl pyrocarbonate-treated water (DEPC-H2O) were thoroughly mixed with PTX-N3 that dissolved in 3:1 (v/v) dimethyl sulfoxide/tert-butanol, followed by adding freshly prepared “click solution” (a blend of copper(I) bromide/Tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine (Sigma-Aldrich) at a 1:2 molar ratio) at a final 1:15:20 molar ratio of RNA:PTX:Cu+. After reaction at room temperature overnight, the product was characterized by 16% (w/v) 8 M urea PAGE in TBE buffer (89 mM Tris base-borate, 2 mM EDTA) at 200 V for 1 h, and stained by ethidium bromide (EtBr, Sigma-Aldrich) and visualized on a Typhoon FLA 7000 (GE Healthcare). Gel images were analyzed by ImageQuant TL. The reaction was subsequently diluted with 1/10 volumes of 0.3 M sodium acetate and 2.5 volumes of 100% ethanol and incubated at −20 °C overnight for RNA precipitation. The precipitates were re-dissolved in DEPC-H2O and purified by Ion-Pair Reverse Phase HPLC in an Agilent PLRP-S 4.6 × 250 mm 300 Å column. PTX-labeled RNA was separated from unreacted RNA-6 alkynes in an H2O/Acetonitrile (ACN) ramp. HPLC data was collected by OpenLAB CDS. Fractions of RNA-6 PTXs were combined, dried and resuspended in DEPC-H2O for nanoparticle assembly.

Construction of 4WJ-X-24 PTXs nanoparticles

The 3D computational model of 4WJ-X nanostructure was generated using Swiss PDB Viewer and PyMOL Molecular Graphics System. The 4WJ-X composed of four RNA strands is designed to lock into a four-strand assembly stabilized by base-pairing complementarity which generates four helix domains arranged around a nanoparticle core. 4WJ-X-24 PTXs nanoparticles were assembled by mixing four RNA-6 PTXs oligomers at equimolar concentrations in TES buffer (50 mM Tris pH = 8.0, 50 mM NaCl, 1 mM EDTA), followed by denaturing at 95 °C for 5 min and gradual cooling to 4 °C over the course of 1 h. Nanoparticle assembly was confirmed using a 12% (w/v) native PAGE in TBE buffer (89 mM Tris base, 200 mM boric acid and 2 mM EDTA) at 150 V for 1 h, and stained and visualized as described above. To construct 4WJ-X-24 PTXs-EGFRapt nanoparticles, one of the RNA oligomers (4WJ-X c) was incorporated with an anti-EGFR aptamer20,21,37 at the 3′-end during synthesis.

DLS measurement

The apparent hydrodynamic diameter and zeta potential of the RNA nanoparticles in 1 μM in TES buffer were determined using a Zetasizer Nano-ZS (Malvern Instrument) at 25 °C. The laser wavelength was 633 nm. Results were plotted with data points using Origin. For size comparison of 4WJ-X, 4WJ-X-24 PTXs, and PTX, three independent samples were measured and average particle size were obtained.

T a measurement by RT-PCR

Preassembled RNA nanoparticles were added to a 96-well plate with a final concentration of 2.5 μM and 1 μM. SYBR Green II dye (Invitrogen) was added to each well as a reporter dye for RNA nanoparticles formation at a final concentration of 20×. Samples were heated to 95 °C for 5 min as a denaturing process followed by a slow cooling ramp to 20 °C at a rate of 0.11 °C s−1 as an annealing process. Annealing curves of RNA nanoparticles were measured by Roche LightCycler® 480 RT-PCR machine, and were plotted with data points using Origin. Three independent samples were measured over three independent experiments with similar results, and representative annealing curves are shown. Ta was determined by LightCycler 480 Software using the first derivative of the annealing profile.

Cryo-electron microscopy data acquisition

Two samples of 4WJ-X nanoparticles, apo state (4WJ-X) and PTX-bound state (4WJ-X-24 PTXs), were diluted to a final concentration of ~20 µM. Three microliters of sample were applied onto glow-discharged 200-mesh R2/1 Quantifoil grids. The grids were blotted for 3 s and rapidly cooled in liquid ethane using Vitrobot Mark IV (ThermoFisher) at 22 °C and 100% humidity. The samples were screened using Talos Arctica cryo-electron microscope (ThermoFisher) operated at 200 kV and then imaged in a Titan Krios cryo-electron microscope (ThermoFisher) with GIF energy filter (Gatan) at a magnification of ×215,000 (corresponding to a calibrated sampling of 0.65 Å per pixel) for apo state and at a magnification of ×130,000 (corresponding to a calibrated sampling of 1.06 Å per pixel) for PTX-bound state using EPU (ThermoFisher). Micrographs were recorded with a Gatan K2 Summit direct electron detector, where each image is composed of 30 individual frames with an exposure time of 6 s and a specimen dose rate of 10 and 7.6 electrons per second per Å2 for apo and PTX-bound state, respectively. A total of 2780 movie stacks for apo state and 3800 movie stacks for the PTX-bound state were collected with a defocus range of 1.5–3.5 μm.

Single-particle image processing and 3D reconstruction

All micrographs were motion-corrected using MotionCor239 and CTF was determined using CTFFIND440. All particles were autopicked using NeuralNet option in EMAN241 and further checked manually, yielding 336,340 particles from selected 2618 micrographs for the apo state, and 406,408 particles from selected 3496 micrographs for the PTX-bound state. The particle coordinates were then imported to RELION42, where four rounds of 2D classification were performed to remove poor 2D class averages. The initial model was built using the “Ab initio 3D reconstruction” option in CryoSPARC43, and the 3D classification was performed in RELION. The final 3D refinement was performed in RELION using 106,775 particles for apo state and 125,909 particles for PTX-bound state, resulting in the corresponding maps with resolutions of 9 and 7 Å, respectively, based on the FSC of two independent particle data sets at a threshold of 0.143. Figures and movie were prepared using Chimera44.

TGGE gel shift assay

Preassembled RNA nanoparticles were run in a 12% (w/v) native PAGE in TBE buffer at 100 V for 10 min at room temperature. TGGE (Biometra Gmbh) was subsequently performed by applying a gradient temperature perpendicular to the electrical current and run for 60 min at 20 W. The gel was stained and visualized as described above. Quantified values of bands for each nanoparticle were divided by the sum of the total values in corresponding lanes, calculated by ImageJ. Melting curves were plotted with quantified data points using Origin. Tm values were defined as the temperature at which 50% of the loaded nanoparticles dissociated.

Enzymatic stability assay

Preassembled RNA nanoparticles were incubated at the final concentration of 1 μM in cell culture medium containing 50% (v/v) FBS at 37 °C for different time points through 48 h. The resulting samples were examined using a 3% agarose gel run at 100 V for 40 min and stained and visualized as described above. Quantification analysis was performed using ImageJ to calculate the percentage of intact nanoparticles (intensity of the band at a time point/intensity of the band at 0 h) for each time point. Enzymatic degradation curve was plotted with quantified data points using Origin.

Cell culture

Human MDA-MB-231 cells were obtained from American Type Culture Collection (ATCC). Cells were grown and cultured in DMEM/F-12 medium (ThermoFisher Scientific) containing both 10% (v/v) FBS and 100 U ml−1 penicillin at 37 °C in humidified air environment containing 5% CO2.

Confocal microscopy imaging

MDA-MB-231 cells were seeded on glass coverslips and cultured at 37 °C overnight. Alexa Fluor 647 labeled RNA nanoparticles were incubated with cells at a final concentration of 100 nM for 4 h at 37 °C. After washing twice with PBS buffer, cells were fixed with 4% formaldehyde and washed again, followed by treatment with 0.1% Triton X-100 (Sigma-Aldrich) in PBS buffer for 5 min and subsequent cytoskeleton staining with Alexa Fluor 488 phalloidin (ThermoFisher Scientific) for 30 min at room temperature. After rinsing with PBS buffer, the cells were mounted with ProLong@ Gold Antifade Reagent (Life Technologies Corp.) containing DAPI for cell nucleus staining and assayed on Olympus FV3000 confocal microscope (Olympus Corp.). Data were collected using Fluoview FV31S-SW.

In vitro apoptosis assay

Cell apoptosis was studied using FITC Annexin V Apoptosis Detection Kit (BD Pharmingen) following manufacturer instructions. Briefly, 5 × 104 MDA-MB-231 cells were seeded on a 24-well plate overnight. RNA nanoparticles and free PTX were added into cells at a final concentration of 1 μM. After 24 h incubation at 37 °C in a humidified 5% CO2 environment, cells were trypsinized to single cell suspension, washed with PBS buffer twice, and resuspended in 100 μL 1 × Annexin V-FITC binding buffer. Five microliters of Annexin V-FITC and 5 μL of PI were added into each sample and incubated at room temperature for 20 min. The samples were then added into flow tubes that contained 200 μL of 1 × binding buffer for FACS analysis by FACSCalibur™ flow cytometer within 1 h. Data were analyzed by FlowJo.

In vitro cytotoxicity assay

Cell cytotoxicity was studied using CellTiter 96 Non-Radioactive Cell Proliferation Assay (Promega) following manufacturer instructions. Briefly, 5 × 103 MDA-MB-231 cells were seeded on a 96-well plate overnight. RNA nanoparticles and free PTX were added into each well in triplicates at indicated concentrations. After incubation at 37 °C for 48 h in a humidified 5% CO2 environment. Fifteen microliters of MTT Dye Solution was added to each well and the plate was incubated at 37 °C in the dark for 4 h. Next, 50 μL Solubilization Solution/Stop Mix was added to each well for dissolving the crystal in the dark. The plate was incubated for 2 h at room temperature. Finally, the crystal in each well were fully dissolved to a uniformly colored solution and their absorbance at 570 nm was measured by Synergy 4 microplate reader (Bio-Tek). Data were collected by Gen5, and normalized data was plotted using Origin.

Tumor xenograft animal model

All animal procedures were performed in accordance with the Subcommittee on Research Animal Care of The Ohio State University guidelines approved by the Institutional Review Board. Mice are housed in sterile environment, with all supplies (cages, bedding, feed, water bottles) autoclaved or irradiated. To generate the xenograft model, female athymic nu/nu mice, 4–8 weeks old, were purchased from Taconic BiosciencesFarm. TNBC tumor xenografts were established by injecting 2 × 106 MDA-MB-231 cells/site resuspended in sterile PBS into the mammary fat pads of nude mice. When the tumor nodules had reached a volume of 75 mm3, the mice were used for tumor inhibition studies.

In vivo biodistribution study

Alexa Fluor 647 labeled 4WJ-X-EGFRapt nanoparticles (15 mg kg−1, RNA per body weight) were systemically administered via the tail vein into MDA-MB-231 tumor bearing mice. PBS-injected mice were used as fluorescence negative controls. The whole-body imaging of mice was conducted at 8 h using an IVIS system (XMRS) with excitation at 640 nm and emission at 680 nm. The mice were sacrificed at 8 h post-injection by the inhalation of CO2 followed by cervical dislocation, and major organs were collected and subjected to fluorescence imaging for the assessment of biodistribution profiles. The fluorescence imaging data of average radiant efficiency ([p s−1 cm−2 sr−1] [μW cm−2]−1) were quantitative by IVIS system (XMRS) program.

In vivo tumor inhibition by 4WJ-X-24 PTXs nanoparticles

Mice with established tumor nodules were randomly divided into four groups (n = 5 biologically independent animals). Samples were administrated by i.v. injection in a total of 5 doses (8 mg kg−1, PTX per body weight) every other day. Tumor volume, calculated as (length × width2)/2, and mouse weight were monitored every other day. Curves were plotted with data points using GraphPad Prism. On day 14, the mice were sacrificed followed by tumor extraction. Tumors were weighted, and result was plotted with data points using Origin. Data were statistically analyzed by two-tailed unpaired t-test and presented as mean ± SD; *p < 0.05; **p < 0.01; ****p < 0.0001.

In vivo cytokines induction evaluation

CD-1 mice (4–5 weeks old) were purchased from Charles River Laboratories. 4WJ-X-24 PTXs nanoparticles and control groups were administered into mice (n = 3 biologically independent animals) via i.v. injection at 5 mg kg−1 (PTX per body weight). Blood samples were harvested from mice 3 h post-injection by cardiac puncture and centrifugated at 12,800 × g for 10 min. Concentrations of cytokines in serum supernatant were examined in triplicates using Mouse ELISA MAX Deluxe sets (BioLegend) for TNF-α, IL-6, and IFN-γ, and using Mouse IFN-α ELISA Kit (R&D Systems) for IFN-α, following manufacturer provided protocols (1:200 dilution for all capture antibodies and detection antibodies). Results were plotted using Origin. Data were statistically analyzed by two-tailed unpaired t-test and presented as mean ± SD; *p < 0.05; **p < 0.01; ***p < 0.001.

Statistics

Each experiment was repeated independently for at least three times for each sample tested, unless otherwise indicated. The results are presented as mean ± standard deviation (SD). Statistical differences were evaluated using two-tailed unpaired t-test with GraphPad software, and statistically significant differences are denoted as *p < 0.05, **p < 0.01, ***p < 0.01, and ****p < 0.0001. No adjustments were made for multiple comparisons.

Reporting summary

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

The source data underlying Figs. 1–7 and Supplementary Figs. 1–14 are provided in the source data file. Cryo-EM structures have been deposited to the Electron Microscopy Data Bank under accession codes EMD-20699 for 4WJ-X nanoparticles and EMD-20697 for 4WJ-X-24 PTXs nanoparticles, respectively. Other relevant data that support the findings of this study are available from the corresponding author upon reasonable request.