Tumor-killing nanoreactors fueled by tumor debris can enhance radiofrequency ablation therapy and boost antitumor immune responses

Abstract

Radiofrequency ablation (RFA) is clinically adopted to destruct solid tumors, but is often incapable of completely ablating large tumors and those with multiple metastatic sites. Here we develop a CaCO₃-assisted double emulsion method to encapsulate lipoxidase and hemin with poly(lactic-co-glycolic acid) (PLGA) to enhance RFA. We show the HLCaP nanoreactors (NRs) with pH-dependent catalytic capacity can continuously produce cytotoxic lipid radicals via the lipid peroxidation chain reaction using cancer cell debris as the fuel. Upon being fixed inside the residual tumors post RFA, HLCaP NRs exhibit a suppression effect on residual tumors in mice and rabbits by triggering ferroptosis. Moreover, treatment with HLCaP NRs post RFA can prime antitumor immunity to effectively suppress the growth of both residual and metastatic tumors, also in combination with immune checkpoint blockade. This work highlights that tumor-debris-fueled nanoreactors can benefit RFA by inhibiting tumor recurrence and preventing tumor metastasis.

Fig. 1. A scheme illustrating the mechanism of tumor debris fueled tumor-killing HLCaP NRs in enhancing RFA treatment and boosting antitumor immunity.

Upon being fixed inside the residual tumors post incomplete RFA with adhesive glue, such HLCaP NRs in responsive to the acidic tumor microenvironment will gradually release LOX and hemin, and synergistically cause continuous lipid peroxidation from these PUFA containing phospholipids inside the tumor debris to trigger ferroptosis of residual tumor cells. Meanwhile, the released HMGB1 molecules from these ferroptotic cancer cells will recruit immature DCs to the residual tumor site and prime specific antitumor immune response featured in increased infiltration of effector T cells and secretion of effector cytokines, to further inhibit the growth of both residual tumors and metastatic (distant) tumors, especially in the combinational use of anti-PD-1 immunotherapy.

Fig. 2. Preparation and characterization of HLCaP NRs.

a Schematic illustration of the preparation procedure of HLCaP NRs via the CaCO₃-assisted double emulsion process. b A representative TEM image of HLCaP NRs from three independent experiments. c Hydrodynamic diameters of HLCaP NRs determined using a Malvern Zetasizer. d Loading efficiencies of Hemin and LOX in HLCaP NRs and HLP NPs, the latter of which were prepared via the conventional double emulsion method without CaCO₃. e The relative enzymatic activity changes of free LOX and LCaP NPs treated by protease K digestion assay. f, g Cumulative release of Hemin and LOX from HLCaP NRs and HLP NPs incubated at pH 7.4 and 6.8. Data in Fig. d–g were represented as mean ± standard error of mean (SEM), n = 3 biologically independent samples.

Fig. 3. pH-responsive catalytic capacity and in vitro therapeutic efficacy of HLCaP NRs.

a, b pH-responsive lipid peroxidation generation capacities of LCaP NPs, HCaP NPs, and HLCaP NRs incubated with linoleic acid (a) or cell lysates (b) at pH 6.8 and 7.4, respectively. c, d Confocal images (c) and flow cytometric analysis (d) of 4T1 cells incubated with HCaP NPs, LCaP NPs, and HLCaP NRs in the presence (w/) or absence (w/o) of cancer cell lysates, followed by being stained with lipid peroxidation probe of BODIPY-C11. e, f Relative cell viabilities of 4T1 cells incubated with HCaP NPs, LCaP NPs, and HLCaP NRs in the presence (e) or absence (f) of cancer cell lysates for 24 h before being determined by MTT assay. g, h Confocal imaging of intracellular lipid peroxidation (g) and relative cell viabilities (h) of 4T1 cells post various treatments as indicated. i Confocal images of 4T1 cells incubated with HCaP NPs, LCaP NPs, and HLCaP NRs in the presence of cancer cell lysates, followed by being stained with HMGB1 and CRT antibodies, respectively. j Schematic illustration of HLCaP NRs mediated propagation of lipid peroxidation and subsequent immunogenic cell death (ICD) induced in the presence of cell lysates. Data in Fig. a, b, d, e, f, and h were represented as mean ± SD, n = 3 biologically independent samples in Fig. a, b, d, n = 4 biologically independent samples in Fig. e, and n = 6 biologically independent samples in Fig. f, h. A representative image of three biologically independent samples from each group is shown in Fig. c, g, i. P values calculated by the two-tailed student’s t-test in Fig. h are indicated in the figure.

Fig. 4. In vivo study of intratumoral lipid peroxidation and HMGB1 release post sequential RFA and HLCaP NRs fixation.

a Schematic illustration of the experimental schedule. b IR thermal images of 4T1 tumor-bearing mice during RFA treatment recorded at different time points as indicated. c In vivo fluorescence imaging of 4T1 tumor-bearing mice with intratumoral injection of Cy5.5 labeled HLCaP NRs in the presence or absence of adhesive glue at indicated time points post RFA treatment. d Confocal images of tumor slices collected from 4T1 tumor-bearing mice with intratumoral injection of Cy5.5 labeled HLCaP NRs in the presence or absence of adhesive glue at indicated time points post RFA treatment. Scale bar was 1 mm. e, f Confocal images of tumor slices collected from 4T1 tumor-bearing mice after different treatments as indicated for 24 h and stained with DCFH-DA (green, e), as well as HMGB1 primary antibodies and corresponding Alexa 488-conjugated secondary antibodies (green, f). A representative image of three biologically independent animals from each group is shown in Fig. d–f.

Fig. 5. In vivo antitumor therapeutic efficacy of sequential RFA and HLCaP NRs fixation.

a Schematic illustration of the in vivo therapeutic schedule on mouse 4T1 tumor model. b In vivo representative bioluminescence imaging of different groups of mice post different treatments as indicated. c, d Tumor growth curves (c) and corresponding mobility-free survival rate (d) of 4T1 tumor-bearing mice post different treatments as indicated. The mice were set as dead when their tumor volume was larger than 1000 mm³. e–g Schematic illustrations and corresponding tumor growth curves of murine H22 tumors (e), human liver cancer PDX tumors (f), and rabbit VX2 tumors (g) post different treatments as indicated. Data of Fig. b, e–g were represented as mean ± SEM, n = 5 biologically independent animals in Fig. c–f, n = 4 biologically independent rabbits in Fig. g. P values calculated by the two-tailed student’s t-test are indicated in the figure.

Fig. 6. In vivo antitumor study and corresponding immune mechanism study of combined RFA, HLCaP NRs fixation, and anti-PD-1 immunotherapy.

a Schematic illustration of the inoculation of the bilateral tumor model for in vivo antitumor and immune mechanism studies. b–d Tumor growth curves of primary (b) and distant tumors (c), as well as corresponding mobility-free survival rate (d) of mice with bilateral tumor models post different treatments as indicated. The mouse was set as dead when its tumor volume was larger than 1000 mm³. e DC maturation status in the drain lymph nodes adjacent to the primary tumors post various treatments as indicated. f–h The frequencies of CD3⁺CD8⁺ T cells (f), and CD3⁺CD4⁺FoxP3⁺ Tregs (g), as well as their ratios (h) inside the distant tumors post various treatments as indicated. i, j The secretion levels of TNF-α and IFN-γ inside the distant tumors post various treatments as indicated. Data in Fig. b, c were represented as mean ± SEM, n = 10 or 15 biologically independent animals, data in Fig. e–j were represented as mean ± SD, n = 5 biologically independent animals. P values calculated by the two-tailed student’s t-test are indicated in the figure.