Mesoporous nanoperforators as membranolytic agents via nano- and molecular-scale multi-patterning

Abstract

Plasma membrane lysis is an effective anticancer strategy, which mostly relying on soluble molecular membranolytic agents. However, nanomaterial-based membranolytic agents has been largely unexplored. Herein, we introduce a mesoporous membranolytic nanoperforators (MLNPs) via a nano- and molecular-scale multi-patterning strategy, featuring a spiky surface topography (nanoscale patterning) and molecular-level periodicity in the spikes with a benzene-bridged organosilica composition (molecular-scale patterning), which cooperatively endow an intrinsic membranolytic activity. Computational modelling reveals a nanospike-mediated multivalent perforation behaviour, i.e., multiple spikes induce nonlinearly enlarged membrane pores compared to a single spike, and that benzene groups aligned parallelly to a phospholipid molecule show considerably higher binding energy than other alignments, underpinning the importance of molecular ordering in phospholipid extraction for membranolysis. Finally, the antitumour activity of MLNPs is demonstrated in female Balb/c mouse models. This work demonstrates assembly of organosilica based bioactive nanostructures, enabling new understandings on nano-/molecular patterns co-governed nano-bio interaction.

Fig. 1. Characterizations of MLNPs.

a Schematic illustration of the nano- and molecular-scale patterns of MLNPs. Blue beads represent silicon atoms; Red beads represent oxygen atoms; White beads represent carbon atoms. b–d TEM image of MLNPs at different magnifications. Black arrows indicate the tube-like structure of the nanospikes; while arrows indicate the lattice fringes stacked along the long axis of spikes. Each experiment was repeated three times independently with similar results. e Powder XRD spectrum of MLNPs. f ¹³C NMR and g ²⁹Si NMR spectrum of MLNPs. h Nitrogen adsorption-desorption isotherm and pore size distribution (inset) of MLNPs.

Fig. 2. MLNPs induce cytotoxicity and membrane damage.

a TEM images of MLNPs, R/O-NPs and S/A-NPs. b The anti-proliferation activity of MLNPs, R/O-NPs and S/A-NPs at 24 h (p values from left to right: 0.0408, 0.0087, 0.0004, <0.0001 and <0.0001) and 48 h (p values from left to right: 0.0083, 0.0021, 0.0016, <0.0001 and <0.0005) in 4T1 cell line (n = 3 independent experiments). c CLSM image of Calcein-AM/PI stained 4T1 cells after treatment of MLNPs, R/O-NPs and S/A-NPs (n = 3 independent samples with similar results). d CLSM image of Annexin V and PI stained 4T1 cells after treatment of MLNPs, R/O-NPs and S/A-NPs (n = 3 independent samples with similar results). e, f Flow cytometry analysis of Annexin V/PI stained 4T1 cells after treatment of MLNPs, R/O-NPs and S/A-NPs (n = 2 independent experiments, p = 0.0234). Statistical significance was determined by one-sided unpaired t-test. Data were shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig. 3. The membranolytic and perforation activity of MLNPs.

a CLSM images of 4T1 cells treated with Dextran-FITC and different nanoparticles (n = 3 independent samples with similar results). Cell nuclei were stained with Hoechst (blue). Green punctae seen in the cell cytoplasm are indicative for Dextran-FITC molecules trapped in endo/lysosomes. The diffuse cytosolic and nuclear staining indicates free cytosolic Dextran-FITC as a result of membrane perforation. b The release of LDH from 4T1 cells after treatment with different nanoparticles for 6 h (p = 0.0184) and 24 h (p = 0.0051) (n = 3 independent experiments). c The release of ATP from 4T1 cells after treatment with different nanoparticles for 24 h (n = 3 independent experiments). d The release of intracellular proteins into the supernatants after treatment with different nanoparticles for 24 h (n = 3 independent experiments). ND non-detected. e Bio-TEM images of 4T1 cells treated with MLNPs for 0.5 h, 2 h and 6 h. Each experiment was repeated three times independently with similar results. Black arrows indicate the nanoparticles that were “stuck” in the plasma membrane. Statistical significance was determined by unpaired t-test. Data were shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig. 4. Corse-grained molecular dynamics simulation for revealing the multivalent perforation behaviours of benzene-incorporated organosilica nanospikes with ordered molecular structure.

a–c Snapshots of the side view of the membrane perforation process of a a single spike and triple spikes with b cone-like and c sector-like arrangements. Snapshots of the generated membrane damage and pores after the penetration of (d) a single spike and triple spikes with e cone-like and f sector-like arrangement. g Area of the hole generated on the membrane as a function of distance between spikes and the membrane surface. Negative values of distance were defined as the distance between spikes and the membrane before penetration, while positive values of distance were defined as the distance between spikes and the membrane after penetration. h The depleted number of phospholipid molecules from the membrane after interacting with a single spike and triple spikes with cone-like and sector-like arrangement. i Radial distribution functions of the distance between Si–O bond and hydrophobic tail of phospholipid molecules (denoted as Si–O, black curve), and the distance between Si–O bond hydrophobic tail of phospholipid molecules (denoted as Benzene, red curve), respectively.

Fig. 5. Quantum chemistry calculation for the binding energy of benzene groups with different alignment towards a phospholipid molecule as a function of intermolecular distance.

The intermolecular binding energy of benzene groups positioning at a headgroup side, b left side beside tail region and c right side beside tail region of phospholipid molecule. The schemes at the right illustrate the relative positions of benzene groups to the phospholipid molecules.

Fig. 6. MLNPs suppressed tumour growth and elicited a proinflammatory tumour immune microenvironment.

a–c Tumour-growth curves (PBS vs. 10 mg kg⁻¹, p = 0.0192; PBS vs. 30 mg kg⁻¹, p = 0.0020), tumour weight (PBS vs. 10 mg kg⁻¹, p = 0.0057; PBS vs. 30 mg kg⁻¹, p = 0.0010), and optical images of tumours of mice treated with MLNPs (n = 5 biologically independent animals). d H&E staining tumours treated with PBS or MLNPs (30 mg/kg). e MHC-II on CD11c⁺F4/80^– dendritic cells within the CD45⁺ leukocyte population in tumors 24 h after intratumoral injection of MLNPs (30 mg/kg) or PBS (p = 0.0040; n = 4 biologically independent samples). f Percentage of infiltrating CD8+ T cells in tumors 96 h after MLNPs (30 mg/kg) or PBS were administered intratumorally (p = 0.0381; n = 4 biologically independent samples). g Immunohistochemical analysis of tumours 96 h after MLNPs (30 mg/kg) or PBS were administered intratumorally (Representative for 4 independent experiments). h, i The IFN-γ (p = 0.0061) and IL-12p40 (p = 0.0087) levels in the tumour tissue measured 96 h after MLNPs (30 mg/kg) or PBS were administered intratumorally (n = 4 biologically independent samples). Statistical significance was determined by one-sided unpaired t-test. Data were shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.