Polypeptide sono-adjuvant for ultrasound-activatable regulation of innate immunity and cancer vaccination therapy Authors: Fangmin Chen, Huijuan Zhang, Shiqin Li, Siyuan Ren, Lujia Huang, Yu Guo, Yi Lai, Zhixiong Cai, Xiaolong Liu, Zhiai Xu, Haijun Yu Published: 2025 ABSTRACT AbstractSpatiotemporally-controllable regulation of innate immunity in lymph node remains a critical challenge to implement effective cancer vaccination. Here we report an ultrasound-activatable strategy to precisely stimulate innate immunity activation in vivo. Mechanistically, ultrasound-triggered mechanical and oxidative forces synchronously activate innate immune pathways in antigen-presenting cells via calcium ion influx and mitochondria DNA release. We next design a polypeptide sono-adjuvant (SONA) library with adjustable physicochemical properties for lymph node-targeting delivery. The top-performed SONA specifically activates robust and durable innate immune responses in lymph node upon localized ultrasound stimulation. The combination of SONA-based vaccine with ultrasound stimulation induces about 3.0-fold higher antigen-specific T cell responses than conventional adjuvant-based vaccines. Moreover, in syngeneic mouse models of orthotopic breast and liver tumors, the combination of SONA-based neoantigen vaccine with ultrasound stimulation markedly boosts immune checkpoint blockade therapy to suppress tumor growth and distant metastasis. Collectively, the polypeptide sono-adjuvant offers great promise for precise regulation of innate immunity and cancer vaccine therapy.Subject terms: Cancer immunotherapy, Biomaterials - vaccines, Tumour immunology Introduction The innate immune system is the front line of host defense against infection or tumor1–3. Antigen-presenting cells (APCs), including dendritic cells (DCs) and macrophages, play crucial roles in bridging innate and adaptive immunity4–6. APCs can phagocytize and cross-present tumor-associated antigens to induce cytotoxic T lymphocytes (CTLs) response in a tumor-specific manner7,8. However, microenvironment within lymph nodes (LNs) often are ineffective to stimulate APCs maturation, migration, and cytokine secretion, thus deactivating the antigen cross-presentation ability of APCs9–11. The activation of nucleotide-binding oligomerization domain (NOD)-like receptor pathways and cytosolic DNA-sensing pathways, such as NOD-like receptor protein 3 (NLRP3) inflammasome, stimulator of interferon genes (STING), and Toll-like receptor 9 (TLR9), can stimulate APCs maturation by upregulating type I interferon (IFN-I) and pro-inflammatory cytokine responses12–14. Accordingly, various types of adjuvants targeting the innate immune pathways have been discovered and evaluated in preclinical settings. However, small-molecule adjuvants have shown limited antitumor efficacy and dose-limiting toxicity in clinical trials, which could be associated with their unsatisfactory LNs-targeting efficacy15,16. Alternatively, various adjuvant-integrated nanoparticles have been developed to improve adjuvant delivery to LNs, and generated potent antitumor immunity in preclinical studies17–20. However, most of these strategies exhibited the “always on” profile to stimulate innate immunity in vivo, and could potentially induce immune-related adverse effects due to their non-specific distribution and immune stimulation on untargeted tissues, thus limiting their future clinical translation21–23. Therefore, spatiotemporally-controllable regulation of innate immunity in LNs is of high priority, but remains a critical challenge to implement safe and robust cancer vaccination therapy. Ultrasound (US), as a clinically-accessible technology, can non-invasively deliver energy into deep tissues up to tens of centimeters. US-triggered mechanical and oxidative force have been exploited to locally activate the intratumoral adaptive T-cell immunity24–29. However, US-triggered activation of innate immunity in LNs has rarely been reported, and the underlying mechanisms of US stimulation on activating innate immune pathways in APCs remain largely unknown. In this study, we demonstrated that US-triggered mechanical and oxidative forces could simultaneously activate multiple innate immune pathways (NLRP3, STING, and TLR9) in APCs by a US-dependent manner. We next engineered a class of polypeptide-based sonosensitive adjuvant (namely sono-adjuvant, SONA) to implement LNs-specific and APCs-targeting delivery through adjusting the physicochemical parameters (e.g., size, charge, surface ligand). The top-performed SONA specifically drained into LNs and effectively targeted APCs, while displaying marginal off-target distribution in normal organs. It was worth noting that SONA remained in an ‘immune-silent’ state in the absence of US stimulation, since the sonosensitizer itself was incapable of triggering mechanical and oxidative force to activate innate immune responses. In contrast, the combination of LNs-localized US stimulation could precisely trigger the innate immunostimulatory effect of SONA, and induced robust IFN-I and pro-inflammatory cytokine responses in vivo. Moreover, the combination of SONA-based vaccine with US stimulation (SONV + US) elicited antigen-specific T-cell immunity and promoted intratumoral CTLs infiltration, which ultimately displayed more potent antitumor efficacy than the conventional adjuvant-immunotherapies in different mouse models of orthotopic tumors. Results US-triggered mechanical and oxidative forces activated innate immune pathways in APCs We selected the US stimulation frequency to 1 MHz with a 50% duty cycle, and altered the power density and duration of US stimulation to investigate the effect of different levels of mechanical and oxidative forces on activating innate immune pathways in APCs. We screened the power density of US stimulation to avoid unintended cellular damage caused by mechanical force. No obvious cytotoxicity was observed in DC2.4 cells and RAW264.7 cells at US power density below 0.2 W/cm2 (Supplementary Fig. 1). We next investigated the impact of US-triggered mechanical force on innate immune response of APCs (Fig. 1a). Flow cytometry analysis and confocal laser scanning microscope (CLSM) imaging data indicated that US stimulation affected the intracellular calcium ion (Ca²⁺) concentration of DCs as a function of power density. Specifically, 0.2 W/cm2 of US stimulation induced a 3.1-fold increase of intracellular Ca²⁺ concentration in DCs compared to PBS group (Fig. 1b–d). The mechanosensitive channel Piezo1 is generally considered to be activated by various types of mechanical stimuli and trigger Ca2+ influx to promote pro-inflammatory immune response30,31. The addition of Dooku-1, a Piezo1-specific antagonist, markedly abrogated Ca2+ influx induced by US stimulation (Fig. 1e), indicating the US-triggered activation of Piezo1 channel in DCs. Ca2+ influx can induce NLRP3 inflammasome assembly and lead to caspase 1-dependent secretion of pro-inflammatory cytokines (IL-1β, IL-18)32,33. Western blot and enzyme linked immunosorbent assay (ELISA) further validated that US-triggered mechanical force activated the NLRP3-Capase-1-IL-1β signaling in APCs (Fig. 1f, g). Specifically, 3 min of US stimulation (0.2 W/cm2, 1 MHz, 50% duty cycle) induced 3.9- and 4.3-fold increase of IL-1β release and activated caspase-1 expression in DCs compared to PBS group, respectively (Fig. 1h). Moreover, the addition of Dooku-1 significantly inhibited US-triggered active caspase-1 release in DCs (Fig. 1i), further confirming the critical role of Piezo1 in US-triggered innate immune activation. US-triggered mechanical and oxidative force activated STING, TLR9 and NLRP3 pathways in APCs in vitro. a Schematic illustration showing US-triggered mechanical force stimulated innate immune activation in DCs. b The intracellular Ca2+ concentration of DC2.4 cells upon US stimulation (n = 3 independent experiments). c Representative CLSM images to show the intracellular Ca2+ concentration of DC2.4 cells after indicated power density of US stimulation (n = 3 independent experiments). Scale bar = 10 μm. d The intracellular Ca2+ concentration of DC2.4 cells at indicated time points after US stimulation (n = 3 independent experiments). e The intracellular Ca2+ concentration of DC2.4 cells after US stimulation with or without the addition of Dooku-1 (n = 3 independent experiments). f Western blot assay of US-triggered NLRP3 activation in DC2.4 cells or RAW264.7 cells (n = 3 independent experiments). g Heat-map of US-triggered IL-1β secretion in BMDCs (n = 3 independent experiments). h Heat-map of US-triggered active caspase-1 release in BMDCs (n = 3 independent experiments). i US-triggered active caspase-1 release in BMDCs with or without the addition of Dooku-1 (n = 3 independent experiments). j Schematic illustration for the preparation of SONP and mechanism of US-triggered oxidative force to stimulate innate immune activation. k Representative CLSM images of intracellular ROS generation in DC2.4 cells after SONP + US treatment (n = 3 independent experiments). Scale bar = 10 μm. l Western blot assay of endoplasmic reticulum stress induced by SONP + US in DC2.4 cells. I: PBS, II: SONP, III: SONP + US (n = 3 independent experiments). m Representative CLSM images of mitochondrial DNA release in DC2.4 cells after SONP + US treatment (n = 3 independent experiments). Scale bar = 10 μm. n Western blot assay of US-triggered STING and TLR9 activation in DC2.4 cells (n = 3 independent experiments). o IL-6 and TNF-α secretion by BMDCs after SONP + US treatment (n = 3 independent experiments). p The ISG response produced by RAW264.7-ISG-Luc or THP-1-ISG-Blue cells after SONP + US treatment (n = 3 independent experiments). Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via one-way ANOVA with Tukey’s test (b, e, i, o) and P values were indicated. Source data are provided as a Source Data file. The elements in this figure (a, j) were created by Adobe Illustrator. On the basis of mechanical activation of innate immunity by US stimulation, we next investigated whether the combination of US-triggered oxidative force could amplify the innate immune activation of APCs. We thus synthesized a biodegradable polypeptide via ring-opening polymerization of tyrosine-N-carboxyanhydrides (NCA). A sonosensitive polypeptide was subsequently obtained by conjugating a clinically applied sonosensitizer, hematoporphyrin monomethyl ether (HMME), onto the end amino group of the polypeptide (Supplementary Figs. 2 and 3). The sonosensitive polypeptide could self-assemble into a sonosensitive nanoparticle (SONP) for precisely modulating oxidative forces in APCs (Fig. 1j). We optimized the US power density at 0.2 W/cm² and the HMME concentration at 1.0 μg/mL, and further screened the duration of US stimulation to avoid unintended cellular damage caused by the combination of mechanical and oxidative stress. The result showed that negligible cytotoxicity was observed in SONP + US-treated DC2.4 cells and RAW264.7 cells at US duration below 4 min (Supplementary Fig. 4). We next used the reactive oxygen species (ROS) indicator, 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA), to detect the generation level of oxidative force by SONP + US in DCs. CLSM images showed that only SONP incubation without US stimulation showed no ROS generation effect, while SONP + US could trigger the intracellular ROS generation in DCs as a function of US stimulation duration (Fig. 1k). Meanwhile, SONP + US was demonstrated to exert endoplasmic reticulum (ER) stress and trigger mitochondrial DNA release in DCs (Fig. 1m), and this effect could be specifically abrogated by N-Acetyl-L-cysteine (NAC)-mediated ROS scavenging (Fig. 1l). ER stress-mediated mitochondrial DNA release was demonstrated to activate both TLR9 and STING pathways34,35. We therefore evaluated whether SONP + US activated these innate immune pathways in DCs. As expected, SONP + US (1 MHz, 50% duty cycle, 0.2 W/cm2, 3 min) obviously activated TLR9-myeloid differentiation primary response 88 (MyD88)-interferon regulatory factor 7 (IRF7) and STING-TANK-binding kinase-1 (TBK1)-interferon regulatory factor 3 (IRF3) axes, as well as their downstream canonical nuclear factor κB (NF-κB) pathway in DCs by a US-dependent manner (Fig. 1n). Moreover, the addition of TLR9 inhibitor (TLR9-i) and STING inhibitor (STING-i) effectively abrogated the activation of above innate immune pathways induced by SONP + US in DCs (Supplementary Fig. 5). Next, we compared the effect of US-triggered mechanical and oxidative forces on inducing innate immune pathways activation in APCs. WB analysis showed that only US stimulation-mediated mechanical force moderately activated NLRP3 pathway in DC2.4 and RAW264.7 cells, while merely affecting STING and TLR9 pathways activation. In comparison, SONP + US could simultaneously trigger the mechanical and oxidative forces, and effectively activate STING, TLR9 and NLRP3 pathways in DC2.4 and RAW264.7 cells, as evident from the upregulation of indicated proteins expression, such as phosphorylated-TBK1 (p-TBK1), phosphorylated-IRF3 (p-IRF3) and phosphorylated-IRF7 (p-IRF7) (Supplementary Fig. 6). In addition, we evaluated the effect of mechanical and oxidative forces on stimulating DC maturation in vitro. The result showed that US-triggered mechanical force moderately enhanced DC maturation level (CD86+, 8.9%) compared with PBS group (CD86+, 5.7%). In comparison, the combination of mechanical and oxidative forces mediated by SONP + US significantly increased the frequency of CD86+ DCs (15.1%), showing a 2.6- and 1.7-fold higher than that in PBS and US groups, respectively (Supplementary Fig. 7). Furthermore, we found that SONP + US obviously elicited pro-inflammatory cytokine (TNF-α, IL-6) secretion by DCs (Fig. 1o), and upregulated type-I interferon-related genes (ISG) expression in THP1-ISG-Blue and RAW264.7-ISG-Lucia cells (Fig. 1p). In contrast, SONP incubation without US stimulation negligibly affected DCs maturation and innate immune activation. We also examined whether SONP + US affected other immune cell types beyond APCs, such as T cells and B cells. The CD3⁺ T cells and CD19⁺ B cells were isolated from untreated-mouse LNs by using fluorescence-activated cell sorting (Supplementary Fig. 8). The isolated T cells and B cells were subsequently treated in vitro with US, SONP, SONP + US for 24 h. Western blot result showed that there were no significant changes in the expression levels of p-TBK1, p-IRF3, and NLRP3 proteins in either CD3⁺ T cells or CD19⁺ B cells following SONP + US treatment compared to PBS group (Supplementary Fig. 9). These findings suggested that SONP + US may not effectively activate innate immune pathways in T cells and B cells. Collectively, these findings demonstrated that SONP + US offered a spatiotemporally-controllable approach to stimulate multiple innate immune pathways in APCs by combing US-triggered mechanical and oxidative forces. Engineering polypeptide sono-adjuvant for US-activatable activation of innate immunity in LNs It has been extensively demonstrated that the formulations to spatially confine adjuvant activity within the secondary lymphoid organs are crucial for enhancing cancer vaccine-mediated T-cell response36–39. To optimize LNs-specific and APCs-targeting delivery, we developed a set of polypeptide-based sonosensitive adjuvant (namely sono-adjuvant, SONA) with adjustable physicochemical properties, including particle size, surface charge, and ligand modification, as shown in the corresponding table in Fig. 2a. Inspired by the mannose receptor was highly expressed in APCs40–42, we introduced an active-targeting moiety, mannose, for modifying the sonosensitive polypeptide (SOP) to improve the APCs-targeting performance (Supplementary Figs. 10 and 11). Meanwhile, a series of charge-modified polypeptides (CMPn) were synthesized via ring-opening co-polymerization of lysine-NCA and tyrosine-NCA (Supplementary Figs. 12 and 13). Through the co-assembly of SOP with CMPn, we obtained SONA1-12 that displayed adjustable particle size ranging from 128.7 to 169.5 nm, surface charge ranging from −57.1 mV to +9.5 mV, and with or without mannose modification. SONA1-12 were subcutaneously injected into the tail base of C57BL/6 mice, and ex-vivo fluorescence imaging of major organs was conducted at 12 h post-injection. SONA1-12 all effectively accumulated in LNs with minimal distribution in heart, liver, spleen, lung and kidney (Fig. 2b). Semi-quantification of the fluorescence signal screened out the top-performed SONV4, with a particle size of 141.3 nm and surface charge of +8.7 mV, which exhibited the highest LNs distribution in vivo (Fig. 2c, d and Supplementary Fig. 14). In addition, compared with SONA3 that without mannose-modification, SONA4 displayed 2.1- and 2.3-fold higher internalization by DCs and macrophages in LNs, respectively (unless noted, all SONA mentioned below refers to SONA4) (Fig. 2e and Supplementary Fig. 15). SONA + US effectively targeted and activated innate immunity in LNs by a spatiotemporally-controllable manner. a Schematic illustration for developing SONA library with adjustable physicochemical property and APCs-targeting ability to achieve US-activatable innate immunity activation in LNs. b Representative fluorescence imaging of LNs and major organs examined at 12 h post SONA1-12 injection (n = 3 mice). c Semi-quantification of the fluorescence signal in LNs at 12 h post-injection (n = 3 mice). d Representative fluorescence micrographs of LNs at 12 h after SONA4 injection. Scale bar = 100 μm. e The frequency of SONA+CD11c+ DCs and SONA+CD11b+ macrophages in LNs at 6 h after SONA3 or SONA4 injection (n = 3 mice). f The absolute number of CD11c+ DCs and CD11b+ macrophages in LNs after SONA + US treatment (n = 3 mice). g The CD80 or CD86 expression level on CD11c+ DCs after SONA + US treatment (n = 3 mice). h The IFN-β and IL-1β production in LNs at different time points after SONA + US treatment (n = 3 mice). i Schematic illustration for evaluating the antitumor immunity in tdLNs elicited by SONA + US treatment in 4T1 tumor-bearing mouse model. j The frequency of matured DCs (CD80+CD86+CD11c+) in tdLNs (n = 5 mice). k The ratio of M2-type (CD86−CD206+CD11b+) macrophages to M1-type (CD86+CD206−CD11b+) macrophages in tdLNs (n = 5 mice). l The frequency of activated CD8+ T cells (CD69+CD8+CD3+) in tdLNs (n = 5 mice). m The frequency of Tregs (FOXP3+CD4+CD3+) in tdLNs (n = 5 mice). n, o WB analysis of STING, TLR9, and NLRP3 pathways activation in tdLNs. p Immunofluorescence staining of Ki67+ (green) CD3+ (red) T cells in tdLNs. Scale bars = 100 μm. Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via unpaired two-tailed Student’s t-test (e) or one-way ANOVA with Tukey’s test (j–m) and P values were indicated. Source data are provided as a Source Data file. The elements in this figure (a, i) were created by Adobe Illustrator. We investigated the influence of different US stimulation parameters (time point, power density, duration) on controlling the adjuvant activity of SONA in vivo. The biodistribution of SONA in LNs was evaluated by utilizing the fluorescence imaging capacity of porphyrin-based sonosensitizer. Upon subcutaneous injection at the tail base, SONA gradually accumulated into LNs and peaked at 6 h post-injection (Supplementary Fig. 16). We thus selected the 6 h post SONA injection as the experimental time point for LNs-localized US stimulation. We next altered the power density and duration of US stimulation to investigate the effect of different levels of mechanical and oxidative forces on innate immune activation in LNs. The LNs were locally stimulated with 5 min of indicated power density of US (0, 0.2, 0.5, 1.0, 1.5 W/cm2) at 6 h post SONA injection. The results revealed that only SONA injection induced negligible innate immunity in LNs. In contrast, the combination of LNs-localized US stimulation effectively activated innate immune responses in vivo, and 1.0 W/cm2 of US stimulation elicited the highest recruitment and maturation of DCs and macrophages in LNs (Fig. 2f, g and Supplementary Fig. 17). We subsequently optimized the duration of US stimulation to achieve the optimal LNs-specific innate immune activation in vivo. The LNs were locally stimulated with 1.0 W/cm2 of indicated duration of US (0, 1, 3, 5, 7 min) at 6 h post SONA injection. The results showed that the recruitment and maturation levels of DCs in LNs gradually increased with the increase of US stimulation duration, and it almost reached the peak value when combined with 5 min of US stimulation (Supplementary Fig. 18). Therefore, we screened out the time point (6 h post injection), power density (1.0 W/cm2) and duration (5 min) of LNs-localized US stimulation for activating innate immunity in vivo. Under these US stimulation parameters, we found the combination of SONA injection with US stimulation (namely SONA + US) could induce abundant and sustained IFN-β and IL-6 secretion in LNs for over 48 h (Fig. 2h), and these cytokine levels gradually declined toward baseline by 72 h post-treatment, indicating a transient but substantial local innate immune responses induced by SONV + US. We next examined the innate and adaptive immune responses in tumor-draining lymph nodes (tdLNs) of 4T1 breast tumor-bearing mice (Fig. 2i). SONA treatment induced no effect on DCs maturation and T cell activation compared with PBS group. In marked contrast, the frequency of matured DCs (CD80+CD86+) in SONA + US group was about 16.1%, showing 2.0- and 1.8-fold higher than that in PBS group (8.1%) and SONA group (8.7%), respectively (Fig. 2j and Supplementary Fig. 19a). The frequency of activated CD8+ T cells (CD69+) in SONA + US group was about 15.1%, showing 2.2- and 2.4-fold higher than that in PBS group (6.74%) and SONA group (6.04%), respectively (Fig. 2l and Supplementary Fig. 19c). It was worth noting that IFN-β was demonstrated to directly inhibit Tregs proliferation and reprogram M2-type macrophages43–45. Therefore, a significant decrease in the frequency of immune-suppressive cells in tdLNs, including Tregs (FOXP3+) and M2-type macrophages (CD206+), could be observed after SONA + US treatment (Fig. 2k, m and Supplementary Fig. 19b, d). Western blot analysis revealed that SONA + US activated the STING, TLR9, and NLRP3 innate immune pathways, as demonstrated by the upregulation of indicated proteins expression in tdLNs (Fig. 2n, o). Immunofluorescence staining of tdLNs further validated that SONA + US significantly increased the proliferation level of T cells than SONA treatment (Fig. 2p). These results collectively demonstrated that SONA + US provided a spatiotemporally-controllable strategy to induce region-confined activation of antitumor immunity in LNs. SONA + US-boosted cancer vaccination elicited robust antigen-specifc T cell immunity Spatiotemporally-controllable orchestration of antigen cross-presentation with efficient innate immune stimulation in APCs remains a critical challenge for inducing safe and potent antigen-specific T-cell response46–48. Herein, we sought to explore the potential of SONA as a generalizable vaccine platform for US-activatable antigen cross-presentation and T cell activation (Fig. 3a). Through the electrostatic interaction, a model antigen ovalbumin (OVA) was loaded into SONA to prepare the sonosensitive vaccine (sono-vaccine, namely SONV). Dynamic light scattering (DLS) and transmission electron microscope (TEM) analysis revealed that SONV exhibited a uniform spherical morphology, with hydrodynamic diameter of 138.6 ± 11.1 nm and surface charge of +5.7 mV (Fig. 3b, c). SONV + US elicited antigen-specific T-cell response via US-triggered innate immunity activation. a Schematic illustration showing the preparation of SONV for US-activatable innate immune stimulation with antigen cross-presentation. b, c DLS analysis and representative TEM images of SONV (n = 3 independent experiments). d Representative ex vivo fluorescence images and signal quantification of LNs and major organs after SONV-FITC injection (n = 3 mice). e Representative fluorescence micrographs of LNs at 6 h after SONV-FITC injection. Scale bar = 200 μm. f The frequency of activated macrophages (CD80+CD86+CD11b+) in LNs after indicated treatments (n = 5 mice). g The frequency of maturated DCs (CD80+CD86+CD11c+) in LNs after indicated treatments (n = 5 mice). h The frequency of antigen-presented DCs (SIINFEKL-H2Kb+CD11c+) in LNs after indicated treatments (n = 5 mice). i The frequency of antigen-specific T cells (SIINFEKL-H2Kb tetramer+CD8+CD3+) in spleen after indicated treatments (n = 5 mice). j The antigen-specific killing effect in vivo after indicated treatments (n = 5 mice). k Heatmap of the frequency of TNF-α+ CD8+ T cells and IFN-γ+ CD8+ T cells in splenocytes after re-stimulation by SIINFEKL peptide in indicated groups (n = 5 mice). l Schematic illustration of US-activatable innate immunity activation in LNs to promote antigen-specific CD8+ T cells response. m Volcano plot of DEGs in LNs after SONV + US versus PBS treatment. n GSEA for the altered gene sets in SONV + US versus PBS group. o Heatmap of DEGs in dLNs after SONV + US versus PBS treatment (n = 3 mice). p KEGG enrichment analysis of the DEGs in LNs between SONV + US with PBS group. q Protein-protein interaction network of DEGs in LNs after SONV + US versus PBS treatment. Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via unpaired two-tailed Student’s t-test (d, m, p) or one-way ANOVA with Tukey’s test (f–j) and P values were indicated. Source data are provided as a Source Data file. The elements in this figure (a, l) were created by Adobe Illustrator. Lysosomal trapping of tumor antigen usually limits the cancer vaccination efficiency49,50, which prompts us to investigate the performance of SONV for facilitating antigen lysosomal escape and cytosolic release. We performed an acridine orange staining assay to investigate the influence of US-triggered mechanical and oxidative forces on lysosomal membrane permeability of DCs. The result showed that only US stimulation-mediated mechanical force was demonstrated to increase the permeability of lysosomal membrane through the cavitation effect, while only SONV incubation showed no obvious lysosomal disruption effect (Supplementary Fig. 20). In comparison, the combination of mechanical and oxidative forces mediated by SONV + US (1 MHz, 50% duty cycle, 0.2 W/cm2, 3 min) could further increase the lysosomal membrane permeability of DCs. In addition, obviously disrupted lysosomal structures were observed in the Bio-TEM images of SONV + US-treated DCs (Supplementary Fig. 21). We employed a FITC-labelled OVA (OVA-FITC) to monitor the internalization and intracellular distribution of tumor antigen. Flow cytometry analysis revealed that SONV-FITC displayed 2.6-fold higher cellular uptake efficacy in DCs than free OVA-FITC at 12 h post incubation (Supplementary Fig. 22). CLSM examination revealed that the OVA-FITC signal in SONV group was majorly entrapped inside the lysosomal vesicles. In contrast, SONV + US significantly induced antigen cytosolic release, with most of OVA-FITC signal detected in the cytosol while not in lysosome (Supplementary Fig. 23). Specifically, the Pearson’s R value (PR) of OVA-FITC signal (green) with lysosomal signal (red) in SONV and SONV + US groups were 0.25 and −0.12, respectively. These results collectively confirmed that SONV + US offered a US-activatable approach to promote antigen cytosolic delivery by combing the US-triggered mechanical and oxidative force. We subsequently examined the ability of SONV + US to induce antigen cross-presentation and T cell activation in vitro. CpG as a TLR-9 agonist and MSA-2 as a STING agonist were selected as the control adjuvants. Only SONV incubation barely affected DCs maturation and T cell activation. In contrast, SONV + US efficiently induced DCs maturation and elicited 2.5 and 2.0-fold more IL-1β secretion compared to OVA + CpG and OVA + MSA-2 groups, respectively (Supplementary Fig. 24). Flow cytometry analysis revealed that SONV + US promoted 2.8 and 2.6-fold higher antigen presentation in DCs than OVA+CpG and OVA + MSA-2 groups, respectively (Supplementary Fig. 25a). CLSM examination further confirmed that SONV + US promoted antigen presentation on the surface of DCs (Supplementary Fig. 25b). In addition, SONV + US-pretreated DCs most efficiently activated OT-1 T cells, with 2.1 and 1.9-fold higher proliferation level than OVA + CpG and OVA + MSA-2 groups, respectively (Supplementary Fig. 26). We further validated the significance of innate immune pathways activation for DCs maturation and antigen presentation induced by SONV + US. Treatment with STING-IN2, an inhibitor of STING, was able to decrease the SONV + US-induced BMDCs maturation and antigen presentation levels (Supplementary Fig. 27). Likewise, using the TLR9 inhibitor AT791 and the NLRP3 inflammasome assembly inhibitor NLRP3-IN-10, we demonstrated that SONV + US stimulated BMDCs maturation by activating both the TLR9 and NLRP3 pathways (Supplementary Fig. 28). Collectively, above data demonstrated that SONV + US provided a US-activatable approach to orchestrate antigen cross-presentation with efficient innate immune stimulation in DCs. To investigate the performance of SONV for LNs-targeting antigen delivery in vivo, C57BL/6 mice were subcutaneously injected with free OVA-FITC or OVA-FITC-loaded SONV. Ex vivo fluorescence imaging showed minimal distribution of free OVA-FITC in LNs and major organs. In contrast, SONV-FITC effectively drained into LNs and showed approximately 2.5-fold higher fluorescence signal than free OVA-FITC at 6 h post injection (Fig. 3d). Immunofluorescence staining of LNs observed obvious co-localization of antigen (OVA-FITC, green), sono-adjuvant (SONA, red) with CD11c+ DCs (Fig. 3e). Flow cytometry analysis further demonstrated that SONV-FITC 3.8- and 4.2-fold more efficiently delivered antigen to DCs and macrophages in LNs than free OVA-FITC, respectively (Supplementary Fig. 29). These results together confirmed the great targeting ability of SONV to deliver tumor antigen into APCs in LNs. We next assessed the impact of different US stimulation parameters on the long-term T cell responses induced by SONV + US. We selected the US frequency of 1 MHz, the duty cycle of 50%, and altered the power density (0.5, 1.0, 1.5 W/cm²) and duration (0, 3, 5 min) of US stimulation. The dynamic activation of antigen-specific CD8+ T cells and memory CD8+ T cells in the peripheral blood of SONV + US-treated mice was examined by flow cytometry (Supplementary Figs. 30 and 31). The results revealed that only SONV treatment induced moderate activation of antigen-specific CD8+ T cell response, and the frequency of Tetramer+CD8+ T cells was marginally increased compared to PBS group at day 28 (Supplementary Fig. 32). In contrast, the combination of SONV with LNs-localized US stimulation effectively enhanced antigen-specific CD8+ T cell activation. Specifically, it was observed that US stimulation under the conditions of 1.0 W/cm² for 5 min or 1.5 W/cm² for 3 min elicited robust and durable antigen-specific CD8+ T cell responses, reaching the highest levels among all tested conditions. We subsequently evaluated the activation of immune memory responses in the peripheral blood. The result showed that SONV + US (1.0 W/cm² + 5 min) treatment induced high-quality and long-term T cell immune responses, showing the highest levels of central memory CD8+ T cell (CD44+CD62L+) and effector memory CD8+ T cell (CD44+CD62L−) responses among all tested conditions (Supplementary Fig. 33). The performance of SONV + US (1.0 W/cm² + 5 min) for inducing antigen-specific immune response in vivo was next examined. C57BL/6 mice were subcutaneously injected with SONV three times at a 7-day interval, followed with LNs-localized US stimulation at 6 h post per injection. We found that SONV + US treatment significantly triggered the expression of stimulatory markers (CD80 and CD86) in macrophages and DCs in the LNs (Fig. 3f, g and Supplementary Fig. 34a, b). The maturation degree of DCs (CD80+CD86+ in CD11c+ cells) in SONV + US group was 22.5%, showing 2.2- and 1.6-fold higher than that in OVA + CpG (10.1%) and OVA + MSA-2 (14.2%) groups, respectively. In addition, SONV + US treatment induced the highest level of antigen-specific peptide-MHC-I complex on the surface of DCs (8.2%, SIINFEKL-H2Kb+ in CD11c+ cells), with 2.5- and 1.9-fold higher than that in OVA+CpG and OVA + MSA-2 groups, respectively (Fig. 3h and Supplementary Fig. 34c). More importantly, the frequency of antigen-specific CD8+ T cells (SIINFEKL-H2Kb tetramer+ in CD8+ cells) in SONV + US group was about 10.1%, while that in OVA + CpG and OVA + MSA-2 groups were only about 2.5% and 3.4%, respectively (Fig. 3i and Supplementary Fig. 35a). Upon ex-vivo restimulation with the SIINFEKL peptide, the splenocytes from SONV + US group mice showed the highest IFN-γ+CD8+ T cell and TNF-α+CD8+ T cell responses among all the groups. Specifically, the frequency of IFN-γ+ in CD8+ T cells in SONV + US group was about 24.3%, showing 2.5- and 1.9-fold higher than that in OVA + CpG group (9.8%) and OVA + MSA-2 group (12.6%), respectively (Fig. 3k and Supplementary Fig. 35b, c). CTLs killing experiment confirmed that SONV + US elicited the highest antigen-specific T cell killing effect among all the groups, while no significant increase was observed after SONV treatment than PBS group (Fig. 3j and Supplementary Fig. 35d). We next examined the transcriptomic features of LNs from SONV + US group mice to investigate the underlying mechanism of innate and adaptive immunity activation. 1329 differentially expressed genes (DEGs) were identified between SONV + US with PBS group (Fig. 3m), and some innate immune pathways-relevant genes (Tlr6, Tlr7, Tlr9, Cgas, Sting1, Ifnab, Nlrp3 and Il1b) were upregulated, while some immune tolerance-relevant genes (Tgfbr1, Tgfb3, Smad1 and Ly6g6e) were downregulated after SONV + US treatment (Fig. 3n, o). Gene set enrichment analysis (GSEA) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis identified that DEGs were enriched in multiple innate immune pathways, such as Toll-like receptor signaling pathway, cytosolic DNA-sensing pathway, NOD-like receptor signaling pathway and RIG-1-like receptor signaling pathway (Fig. 3p, q). These results confirmed that SONV + US provided a US-activatable strategy to coordinatively stimulate multiple innate immune pathways in APCs of LNs and amplify antigen-specific T-cell immunity (Fig. 3l). We next evaluated and compared the immune-related side effects of SONV + US with OVA + CpG and OVA + MSA-2 by measuring the systemic levels of inflammatory cytokines (IL-6 and IL12p70). The results demonstrated that mice treated with OVA + CpG or OVA + MSA-2 exhibited pronounced systemic cytokine release in serum, with IL-6 concentrations reaching about 219.2 pg/mL (OVA + CpG), 126.0 pg/mL (OVA + MSA-2) and IL-12p70 concentrations reaching about 34.0 pg/mL (OVA + CpG), 26.2 pg/mL (OVA + MSA-2) at 1 day post each injection (Supplementary Fig. 36). These cytokine levels gradually declined but remained slightly above baseline by day 7 post each injection, indicating a transient but substantial systemic inflammatory response induced by OVA + CpG and OVA + MSA-2. In contrast, mice treated with the SONV + US showed only a modest increase in IL-6 (35.8 pg/mL) and IL-12p70 (10.1 pg/mL) secretion at 1 day post each injection. Importantly, inflammatory cytokine levels were found to decline toward baseline and remain at physiological levels by day 7 post injection. These results demonstrated that SONV + US reduced systemic immune activation and exhibited good long-term safety compared to conventional adjuvant-based vaccination strategies. SONV + US reshaped intratumoral immune microenvironment to suppress tumor growth and metastasis Given the promising effect of SONV + US to elicit antigen-specific T-cell immunity in vivo, we continued to evaluate the prophylactic efficacy of US-activatable vaccination strategy in the B16-OVA melanoma tumor model. C57BL/6 mice were subcutaneously injected with OVA + CpG, OVA + MSA-2 and SONV three times at 7-day intervals, and LNs-localized US stimulation of SONV + US group mice was conducted at 6 h post per injection. The mice were inoculated with B16-OVA cells at 7-day post final treatment, and tumor growth was subsequently monitored (Fig. 4a). SONV treatment showed little efficacy to prevent B16-OVA tumor growth and induced no obvious survival benefit compared to PBS group. In contrast, mice immunized with SONV + US maintained an evidently higher survival rate than OVA + CpG and OVA + MSA-2 groups, with 3/6 of mice had no detectable tumor mass up to 60 days (Fig. 4b and Supplementary Fig. 37a). To assess the induction of long-term immune memory effect, the cured mice in SONV + US group were rechallenged with B16-OVA cells, and the result demonstrated 100% of cured mice prevented secondary tumor growth (Fig. 4e, f and Supplementary Fig. 37b). Notably, the frequency of effector CD8+ T memory cells and central CD8+ T memory cells in SONV + US group mice were markedly increased compared to the naïve mice on day 100 (Fig. 4g). SONV + US preformed potent prophylactic and therapeutic effect to mitigate tumor burden. a Experimental schedule for evaluating the prophylactic effect of SONV + US in primary and metastatic B16-OVA tumor models. b, c Averaged B16-OVA tumor growth curves and corresponding survival curves of mice after indicated treatments (n = 6 mice). d Representative photographs of lungs and quantitative analysis of lung metastatic foci of B16-OVA tumor (n = 4 mice). e, f Averaged B16-OVA tumor growth curves and corresponding survival curves of SONV + US-cured mice after tumor rechallenge (n = 4 mice). g Representative flow cytometry plots and quantitative analysis of TEM (CD44+CD62L−CD8+CD3+) and TCM (CD44+CD62+CD8+CD3+) in the splenocytes of SONV + US-cured mice (n = 4 mice). h Experimental schedule for evaluating the therapeutic effect of SONV + US in established B16-OVA tumor models. i, j Averaged B16-OVA tumor growth curves and corresponding survival curves of mice after indicated treatments (n = 6 mice). k The ratio of tumor-infiltrating CD8+ T cells (CD8+CD3+CD45+) to CD4+ T cells (CD4+CD3+CD45+) (n = 5 mice). l The frequency of tumor-infiltrating cytotoxic T cells (IFNγ+CD8+CD3+CD45+) (n = 5 mice). m The frequency of tumor-infiltrating Tregs (CD25+FOXP3+CD4+CD3+CD45+) (n = 5 mice). n The frequency of tumor-infiltrating MDSCs (Ly6G+CD11b+CD45+) (n = 5 mice). o The ratio of tumor-inflicting M2-type (F4/80+CD206+CD11b+CD45+)-TAMs to M1-type (F4/80+CD86+CD11b+CD45+) TAMs (n = 5 mice). p Heatmap of intratumoral secretion of TNF-α and TGF-β (n = 5 mice). q Immunofluorescence staining of tumor-infiltrating CD8+ (green). Scale bars = 50 μm. r Schematic illustration of SONV + US-mediated immunosuppressive microenvironment remodeling to eliminate established tumors. Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via unpaired two-tailed Student’s t-test (e, g), or one-way ANOVA with Tukey’s test (b, d, i, k–o), or log-rank (Mantel–Cox) test (c, f, j), and P values were indicated. Source data are provided as a Source Data file. The elements in this figure (a, r) were created by Adobe Illustrator. To evaluate the potential of SONV + US to inhibit lung metastases of tumor cells, immunized mice were intravenously inoculated with B16-OVA cells, and the lung tissues were harvested to quantify the metastatic nodules at day 14 after tumor inoculation. SONV treatment induced negligible anti-metastases efficacy compared to PBS group, and OVA + CpG, OVA + MSA-2 afforded modest protection against tumor metastases. In contrast, SONV + US treatment effectively suppressed the formation of lung metastatic nodules, showing 4.0-fold higher anti-metastases efficacy than SONV treatment (Fig. 4d). The therapeutic efficacy of SONV + US to eradicate established tumors was subsequently investigated in the B16-OVA melanoma model. C57BL/6 mice were inoculated with B16-OVA cells in the right flank and treated with indicated formulations 7 days later (Fig. 4h). SONV treatment negligibly affected tumor growth and mice survival compared with PBS group. OVA + CpG and OVA + MSA-2 only moderately inhibited tumor growth and all the mice succumbed to tumors within 30 days (Fig. 4i and Supplementary Fig. 38a). SONV + US treatment markedly inhibited tumor growth and prolonged the median survival of mice from 20 days (PBS group) to 36 days, with 1/6 of mice produced complete tumor regression over 60 days (Fig. 4j). The body weight of all mice changed negligibly when exposed to the indicated treatments (Supplementary Fig. 38b). Mechanistic studies were next performed to investigate whether SONV + US treatment elicited antitumor immunity in tdLNs and tumor microenvironment (Supplementary Figs. 39 and 40). We found that SONV + US strongly promoted DCs maturation and macrophages activation in tdLNs (Supplementary Fig. 41). The maturation degree of DCs (CD80+ in CD11c+ cells) in SONV + US group was 28.8%, showing 1.8- and 1.5-fold higher than that in OVA + CpG (16.2%) and OVA + MSA-2 (19.2%) groups, respectively. Compared to the SONV treatment that elicited negligible antitumor immunity, SONV + US significantly enhanced tumor-infiltration of CD8+ T cells (Fig. 4k and Supplementary Fig. 42a), and increased the activation of CTLs. The frequency of IFN-γ+CD8+ T cells (IFN-γ+ in CD8+ T cells) in SONV + US group was about 19.8%, with 2.5 and 1.7-fold higher than that in OVA + CpG (7.9%) and OVA + MSA-2 (11.6%) groups, respectively (Fig. 4l and Supplementary Fig. 42b). In addition, compared with PBS group, SONV + US treatment increased the intratumoral progenitor exhausted (PD1+TIM3−) CD8+ T cells infiltration, while decreased the tumor-infiltration of terminally exhausted (PD1+TIM3+) CD8+ T cells. The frequencies of progenitor exhausted CD8+ T cells (PD1+TIM3− in CD8+CD3+ cells) and terminally exhausted CD8+ T cells (PD1+TIM3+ in CD8+CD3+ cells) in SONV + US group were about 28.4% and 19.4%, showing 2.2-fold higher and 3.0-fold lower than that in PBS group, respectively (Supplementary Fig. 42c–e). These results indicated that SONV + US treatment not only enhanced intratumoral CD8+ T cell infiltration and activation, but also was promising to address intratumoral CD8+ T cell exhaustion. Next, we measured the cytotoxic effect of tumor-infiltrating NK cells and the suppressive function of tumor-infiltrating Tregs in order to provide clearer insights about SONV + US-mediated antitumor immunity. The findings showed that SONV + US treatment increased the intratumor cytotoxic NK cell response (IFN-γ+ NK cells and GZMB+ NK cells) compared with PBS group. The frequencies of IFN-γ+ NK cells (IFN-γ+ in NK1.1+ cells) and GZMB+ NK cells (GZMB+ in NK1.1+ cells) in SONV + US group were about 12.5% and 11.6%, respectively, showing 2.1- and 2.0-fold higher than that in PBS group (IFN-γ+ NK, 5.96% and GZMB+ NK, 5.56%), respectively (Supplementary Fig. 43). In addition, SONV + US treatment not only decreased the frequency of tumor-infiltrating Tregs, but also inhibited the production of Tregs-derived suppressive cytokines (IL-4, IL-10). The frequencies of tumor-infiltrating IL10+ Tregs (12.6%, IL10+ in CD25+FOXP3+ cells) and IL4+ Tregs (10.6%, IL4+ in CD25+FOXP3+ cells) in SONV + US group displayed 3.1- and 3.2-fold lower than that in PBS group (IL10+ Tregs, 39.2% and IL4+ Tregs, 33.3%), respectively (Fig. 4m and Supplementary Fig. 44). Importantly, SONV + US also relived the immunosuppressive tumor microenvironment by reducing intratumoral infiltration of myeloid-derived suppressor cells (MDSCs) (Fig. 4n and Supplementary Fig. 45a), and repolarizing tumor-supportive M2-type tumor-associated macrophages (TAMs) to tumor-suppressive M1-type TAMs (Fig. 4o and Supplementary Fig. 45b, c). ELISA result revealed that SONV + US significantly increased intratumoral secretion of TNF-α, while inhibited the typical immunosuppressive cytokine (TGF-β) secretion (Fig. 4p and Supplementary Fig. 46). The intratumoral TNF-α secretion levels in OVA + CpG, OVA + MSA-2, and SONV + US group was about 1189.6 pg/mL, 1996.3 pg/mL, and 2346.3 pg/mL, respectively. The intratumoral TGF-β secretion levels in OVA + CpG, OVA + MSA-2, and SONV + US group was about 2515.0 pg/mL, 1593.6 pg/mL, and 835.8 pg/mL, respectively. Immunofluorescence staining of tumor sections further confirmed the excellent performance of SONV + US to promote the intratumoral infiltration of CD8+ T cells (Fig. 4q). These results together suggested the potent prophylactic and therapeutic efficacy of the US-activatable vaccination strategy against tumor growth and metastases (Fig. 4r). Multivalent activation of innate immunity was essential for the antitumor efficacy of SONV + US We next investigated whether SONV + US-mediated antitumor efficacy were attributed to the US-activatable activation of innate immunity in vivo. A TLR9-specific antagonist, AT79151,52, was intraperitoneally injected to block the interaction of mtDNA with TLR9 (Fig. 5a). We found that TLR9 blockade not only significantly diminished the tumor-inhibitory effect of SONV + US, but also reduced the survival time of B16OVA tumor-bearing mice (Fig. 5b, c and Supplementary Fig. 47). Further evaluation of antitumor immunity in tdLNs revealed that TLR9 blockade highly decreased co-stimulatory molecules (CD80, CD86) expression on macrophages and DCs that elicited by SONV + US treatment (Fig. 5d, f and Supplementary Fig. 48a, b). TLR9-i injection remarkably decreased the antigen presentation level on APCs and antigen-specific CD8+ T cell response in tdLNs induced by SONV + US treatment (Fig. 5e, g). Specifically, the frequency of antigen-specific CD8+ T cell in SONV + US group (6.9%) was 2.2-fold higher than that in TLR9-i+SONV + US group (3.1%) (Fig. 5h and Supplementary Fig. 48c–e). TLR9-i injection almost completely abolished SONV + US-mediated CTLs infiltration, and the frequency of tumor-infiltrating IFN-γ+CD8 T+ cell in SONV + US group (9.5%) was 4.5-fold higher than that in TLR9-i + SONV + US group (2.1%) (Fig. 5i, j and Supplementary Fig. 49a, b). Meanwhile, the ability of SONV + US to remodel immunosuppressive tumor microenvironment, including M2-type TAMs repolarization, decrease of Tregs and MDSCs, was markedly impaired by TLR9 blockade (Fig. 5k–m and Supplementary Fig. 49c–f). Activation of STING, TLR9, and NLRP3 pathways was essential for antitumor immunity induced by SONV + US. a Experimental schedule for evaluating the contribution of mtDNA-TLR9 pathway for antitumor efficacy of SONV + US in B16-OVA tumor models. b, c Averaged B16-OVA tumor growth curves and corresponding survival curves of mice after SONV + US with or without TLR9-mtDNA blockade (n = 6 mice). d–h The frequency of activated macrophages (d), antigen-presented macrophages (e), matured DCs (f), antigen-presented DCs (g) and antigen-specific CD8+ T cells (h) in tdLNs (n = 5 mice). i–k The frequency of tumor-infiltrating CD8+ T cells (i), cytotoxic CD8+ T cells (j), Tregs (k) in B16-OVA tumors (n = 5 mice). l The ratio of tumor-infiltrating M1-type TAMs to M2-type TAMs (n = 5 mice). m The frequency of tumor-infiltrating MDSCs (n = 5 mice). n Immunofluorescence staining of tdLNs to show the antigen presentation level of DCs in wild-type or IFNAR-KO tumor-bearing mice. Scale bars = 20 μm. o Immunofluorescence staining of tumors to show the infiltration of CD8+ T cells (red) and CD206+ M2-type TAMs (green) in wild-type or IFNAR-KO tumor-bearing mice model. Scale bars = 50 μm. p Experimental schedule for evaluating the antitumor effect of SONV + US+anti-IL1β in B16-OVA tumor model. q, r Averaged B16-OVA tumor growth curves and corresponding survival curves of mice after SONV + US treatment with or without anti-IL1β blockade (n = 6 mice). s Experimental schedule for evaluating the antitumor effect of SONV + US+anti-CD8 in B16-OVA tumor model. t, u Average B16-OVA tumor growth curves and corresponding survival curves of mice after SONV + US treatment with or without anti-CD8+ T cell blockade (n = 6 mice). Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via one-way ANOVA with Tukey’s test (b, d–m, q, t) or log-rank (Mantel–Cox) test (c, r, u), and P values were indicated. Source data are provided as a Source Data file. Given the fact that IFN-I play essential roles in promoting antigen presentation and CTLs tumor-infiltration53–55, we next visualized antigen-presented DCs in tdLNs and tumor-infiltrating CD8+ T cells in wild-type or type I IFNs receptor (IFNAR)-KO mice after SONV + US treatment. Immunofluorescence staining results showed that IFNAR-knockout abrogated the antigen presentation of DCs in tdLNs and the recruitment of CD8+ T cells in tumors in SONV + US-treated mice (Fig. 5n, o). These data implied that activation of STING-IFN-I signaling is crucial for SONV + US-mediated adaptive antitumor immunity activation. A monoclonal antibody against IL-1β (anti-IL-1β) was intraperitoneally injected to block the NLRP3-IL-1β signaling in vivo (Fig. 5p). We found that IL-1β neutralization not only markedly impaired the tumor-inhibitory effect of SONV + US, but also reduced the survival time of B16OVA tumor-bearing mice (Fig. 5q, r and Supplementary Fig. 50). To identify whether activation of CD8+ T cell response was required for SONV + US-mediated antitumor efficacy, we used a monoclonal antibody against CD8 (anti-CD8) to deplete CD8+ T cells in vivo (Fig. 5s). We performed transient CD8⁺ T cell depletion by using monoclonal antibodies to preserve the overall immune architecture and better mimic a functional immune system under therapeutic pressure. This approach allowed to identify and specifically assess the role of CD8⁺ T cells without introducing broader developmental or compensatory alterations that often observed in genetically immunodeficient strains. Moreover, antibody-mediated depletion was a widely used, well-established method to interrogate T cell involvement in preclinical studies32,48. The result showed that the tumor-inhibitory effect of SONV + US was completely abolished by CD8+ T cell depletion, and resulted in largely reduced survival time of tumor-bearing mice, indicating the indispensable role of CD8+ T cells in SONV + US-mediated antitumor immunity (Fig. 5t, u and Supplementary Fig. 51). These results together demonstrated that the US-activatable multivalent innate immunity stimulation and subsequent CD8+ T cell activation is essential to generate SONV + US-based antitumor immunity. Neoantigen-based SONV + US boosted ICB therapy in orthotopic tumor models The identification of tumor-specific neoantigen has attracted great interest in developing personalized therapeutic cancer vaccines56,57. Although the synthetic neoantigen peptides elicit T-cell immunity in vitro, a major issue remaining is their overall low immunogenicity and poor LNs-specific distribution in vivo58. To demonstrate the therapeutic potential of SONA platform for neoantigen vaccine immunotherapy, a triple-negative breast cancer (TNBC) cell, 4T1-specific neoantigen peptide (SHRSCSHQTSAPSPKALAHNGTPRNAI) was synthesized and assembled with SONA for preparing neoantigen-based sonosensitive vaccine (Ne-SONV) (Fig. 6a). Ne-SONV displayed a uniform hydrodynamic diameter of 127.6 ± 24.1 nm and positive surface charge of +5.4 mV (Supplementary Fig. 52). Ne-SONV + US combined with αPD-L1 therapy to suppress orthotopic TNBC growth and lung metastasis. a Schematic illustration to show the experimental timeline for evaluating the combination therapeutic efficacy of Ne-SONV + US with αPD-L1 in orthotopic 4T1-Luc TNBC model. b Representative bioluminescence images of orthotopic 4T1-Luc tumor-bearing mice at different time points (n = 5 mice). c, d Average 4T1-Luc tumor growth curves and corresponding survival curves of mice (n = 5 mice). e Representative photographs of lungs and quantitative analysis of lung metastatic foci of 4T1-Luc tumors (n = 3 mice). f Representative bioluminescence images and the corresponding H&E-stained slices of lung tissue on day 30. Scale bar = 2000 μm. g Representative immunofluorescence micrographs to show the infiltrating CD8+ T cells (green) and CD4+ T cells (red) into lung tissue. Scale bars = 100 μm. h t-SNE (t-distributed stochastic neighbor embedding) clustering of tumor-infiltrating CD45+ immune cell subsets. i, j Representative t-SNE plots and normalized PD-1 expression on CD45+ immune cells after indicated treatments (n = 5 mice). DNT, double-negative T cells. k Heatmap to show the frequency of tumor-infiltrating immune cells after the indicated treatment (n = 5 mice). Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via one-way ANOVA with Tukey’s test (c, e, f, j) or log-rank (Mantel–Cox) test (d), and P values were indicated. Source data are provided as a Source Data file. The elements in this figure (a) were created by Adobe Illustrator. Compared to the admixture of neoantigen peptide with CpG adjuvant (Neop + CpG) that moderately inhibited the orthotopic 4T1 tumor growth and all the mice lost within 30 days, Ne-SONV + US markedly inhibited tumor growth and prolonged the median survival of mice from 26 days (PBS group) to 40 days (Fig. 6b and Supplementary Fig. 53). Notably, the combo of Ne-SONV + US with programmed death-ligand 1 (PD-L1) blockade (Ne-SONV + US + αPD-L1) further extended the mice survival and 3/6 of mice survived above 60 days (Fig. 6c, d). Lung metastasis is the major cause of death in patients with TNBC. The mice of PBS or Neop+CpG group were observed with severe lung metastasis in 30 days post primary tumor inoculation (Fig. 6e). In contrast, Ne-SONV + US and Ne-SONV + US + αPD-L1 treatment dramatically increased the CD4+ and CD8+ T cells infiltration into lung tissue, and effectively inhibited the lung metastasis formation (Fig. 6f, g). We next investigated whether Ne-SONV + US elicited potent intratumoral antitumor immunity in the orthotopic 4T1 tumor model (Supplementary Fig. 54). According to the t-SNE plots, we could intuitively observe the changes in the tumor-infiltration and corresponding programmed cell death protein 1 (PD-1) expression levels of different immune cells after indicated treatments. Compared with Neop+CpG treatment, Ne-SONV + US significantly increased intratumoral immune cells infiltration, including DCs, TAMs, CD8+ T cells, CD4+ T cells, and NK cells (Fig. 6h, i). Meanwhile, Ne-SONV + US treatment downregulated the PD-1 expression on TAMs when compared with PBS group, while marginally influenced the PD-1 expression on intratumoral CD8+ or CD4+ T cells (Fig. 6j). We found that Ne-SONV + US treatment recruited 3.5- and 2.3-fold higher of IFN-γ+ CD8+ T cells (15.2%, IFN-γ+ in CD8+ T cells) and TNF-α+ CD8+ T cells (6.38%, TNF-α+ in CD8+ T cells) than Neop + CpG, respectively (Supplementary Fig. 55). Combination therapy with αPD-L1 further increased the proportion of tumor-infiltrating CTLs. The frequency of IFN-γ+CD8+ T cells and TNF-α+ CD8+ T cells in Ne-SONV + US + αPD-L1 group was about 24.9% and 11.2%, showing 5.8- and 4.0-fold higher than that in Neop + CpG treatment, respectively. Ne-SONV + US and Ne-SONV + US + αPD-L1 treatment also reversed the immunosuppressive tumor microenvironment, as indicated by the significant decrease of Tregs, M2-type TAMs and increase of M1-type TAMs (Fig. 6k and Supplementary Fig. 56). These results together demonstrated that Ne-SONV + US was able to enhance the immunogenicity of neoantigen and elicit robust cytotoxic CD8+ T cell response in solid tumors. To demonstrate the therapeutic genericity of Ne-SONV + US strategy in solid tumors, we further evaluated the performance of Ne-SONV + US for boosting ICB therapy in the orthotopic hepatocellular carcinoma (HCC) model. HCC remains one of the most diagnosed solid tumors with poor survival in the clinic59. ICB therapy have shown promising antitumor efficacy in partial HCC patients; however, the advanced HCC patients marginally benefit from ICB therapy due to limited tumor-infiltration of effector T cells60,61. Our previously identified HCC cell, Hepa1-6-specific neoantigen peptide (RWLYWQPTL), was synthesized and assembled with SONA for preparing Ne-SONV vaccine, which displayed a uniform hydrodynamic diameter of 130.6 ± 21.0 nm and positive surface charge of +3.7 mV (Fig. 7a). Ne-SONV + US boosted αPD-L1 therapy to establish durable immune memory effect in orthotopic HCC model. a Schematic illustration to show the experimental timeline for evaluating the combination therapeutic efficacy of Ne-SONV + US with αPD-L1 in orthotopic Hepa1-6-Luc HCC model. b Representative bioluminescence images of orthotopic Hepa1-6-Luc tumor-bearing mice at different time points (n = 5 mice). c, d Average bioluminescence intensity of Hepa1-6-Luc tumors and corresponding survival curves of mice (n = 5 mice). e Representative photographs and corresponding liver tissue weight examined on day 30 (n = 3 mice). f Representative bioluminescence images, intensity, and the corresponding H&E-stained slices of liver tissue on day 30. Scale bar = 1000 μm (n = 3 mice). g Representative immunofluorescence micrographs to show TUNEL and Ki-67 staining of the Hepa1-6-Luc tumor sections collected from indicated groups. Scale bars = 100 μm. h Schematic illustration to show the experimental timeline for tumor rechallenge study in Ne-SONV + US+α-PDL1-cured mice. i Representative bioluminescence images of subcutaneous Hepa1-6-Luc tumor-bearing mice at different time points (n = 5 mice for naïve mice group, n = 4 mice for cured mice group). j, k Average tumor growth curves and the corresponding survival curves of mice in the tumor rechallenge study (n = 5 mice for naïve mice group, n = 4 mice for cured mice group). l Representative flow cytometry plots and quantitative analysis of TEM (CD44+CD62L−CD8+CD3+) and TCM (CD44+CD62+CD8+CD3+) in the splenocytes of Ne-SONV + US+α-PDL1-cured mice (n = 3 mice). Unless specified otherwise, the data are presented as mean ± SD. Statistical significance was calculated via unpaired two-tailed Student’s t-test (j, l), or one-way ANOVA with Tukey’s test (c, e, f), or log-rank (Mantel–Cox) test (d, k), and P values were indicated. Source data are provided as a Source Data file. The elements in figure (a, h) were created by Adobe Illustrator. Combination of Ne-SONV + US with αPD-L1 was examined in the orthotopic Hepa1-6-Luc liver tumor model. Bioluminescence imaging in vivo indicated that αPD-L1 moderately inhibited Hepa1-6-Luc tumor growth and 4/5 of mice lost within 40 days (Fig. 7b). In contrast, Ne-SONV + US markedly inhibited tumor growth and prolonged the mice survival, with 3/5 of mice survived above 60 days. Notably, the combo of Ne-SONV + US with αPD-L1 induced 100% survival of the orthotopic liver tumor-bearing mice in 60 days post tumor inoculation (Fig. 7c, d). Ex vivo imaging of the liver tissue further confirmed that Ne-SONV + US boosted αPD-L1 therapy to suppress orthotopic HCC development (Fig. 7e, f). The Ki67 and terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) staining of tumor section indicated the obviously decreased tumor cell proliferation and increased cell apoptosis/necrosis after the combo of Ne-SONV + US with αPD-L1 treatment (Fig. 7g). To assess the elicitation of long-term immune memory effect, the cured mice in Ne-SONV + US + αPD-L1 group were rechallenged with Hepa1-6-Luc cells at 60 days post the primary tumor inoculation (Fig. 7h). Bioluminescence imaging in vivo indicated that 3/4 of cured mice prevented secondary tumor growth over 90 days (Fig. 7i–k and Supplementary Fig. 57), which was mainly attributed to the induction of high-quality and long-lived effector CD8+ T memory cells and central CD8+ T memory cells (Fig. 7l). Hematoxylin-eosin (H&E) staining of the mouse major organs displayed no inflammatory lesion or histopathological damage at 40 days post-final SONA + US treatment (Supplementary Fig. 58). Biochemical analysis of serum also indicated the favorable long-term safety of SONA + US treatment in vivo (Supplementary Fig. 59). These findings together confirmed that SONA + US treatment showed no obvious toxicity on non-target organs. Collectively, our study provided a US-activatable approach for achieving safe and effective cancer vaccination therapy. Discussion The activation of innate immune pathways in APCs is crucial for promoting tumor-associated antigen presentation to induce cytotoxic T cell response in an antigen-specific manner. However, the innate immune stimulation by small-molecule adjuvant is highly impeded by their poor intracellular delivery, serum stability, and rapid systemic clearance. Various delivery strategies, including agonist-loaded nanoparticles, antibody-agonist conjugates, and self-adjuvating biomaterials, have been developed to enhance antitumor immunity by addressing the above limitations. However, most of these strategies remain face the challenge of non-specific distribution and activation of agonists on untargeted tissues, thus inducing potential dosage-dependent immune-related adverse effects in clinical settings. With the goal of improving the safety and efficacy of adjuvant-immunotherapy of cancer, we designed a US-activatable approach to precisely elicit antitumor immunity by stimulating multiple innate immune pathways in APCs of LNs. US-triggered mechanical and oxidative force allowed to spatiotemporally tailor innate immunity activation in LNs by adjusting the irradiation site, power density, time interval, and duration of external US stimulation. Mechanistically, SONV + US activated the intracellular pattern recognition receptors, including STING, TLR9, NLRP3, via Ca2+ influx and ER stress-mediated mtDNA release in APCs, thereby producing type-I IFNs and pro-inflammatory cytokines. SONV + US promoted APCs maturation, antigen cross-presentation, and ultimately induced robust and durable antigen-specifc CD8+ T cell response. Notably, benchmark validation with the commercial TLR or STING adjuvant-based cancer vaccines, SONV + US, effectively recruited tumor-infiltrating CTLs in a US-dependent manner, thus displaying more potent tumor-inhibitory effect in murine tumor models mimicking a range of clinical scenarios. Moreover, Ne-SONV + US enhanced the immunogenicity of tumor-specifc neoantigen to suppress orthotopic breast and liver tumors growth, and further combined with ICB therapy to effectively eliminate the primary and secondary liver tumors by establishing high-quality and long-term immune memory effect (Supplementary Table 1). Various types of US-activatable immune adjuvant strategies have been developed to control the activation of immune agonists in a region-specific manner. For instance, a US-responsive polymeric pro-agonist system was reported to specifically activated immune response within the tumor microenvironment. Upon US stimulation, the pro-agonist system released bioactive STING agonist, triggered potent local immune responses while minimizing systemic toxicity17. In addition, a semiconducting polymeric nanoagonist enabling US-triggered activation of immune agonist was reported for precise cancer immunotherapy. In combination with US stimulation, the nanoagonist generated abundant singlet oxygen to induce immunogenic cell death and simultaneously release caged immune agonist via ROS-cleavable linker. This US-triggered activation enhanced innate immune activation within tumor tissue, promoted T cell infiltration, and elicited robust antitumor immunity62. Compared to the conventional vaccination strategies, the US-activatable sono-adjuvant reported herein exhibited distinct advantages for future clinical application. First, the sono-adjuvant is rational designed to avoid the use of conventional immune adjuvants (e.g., TLR agonist or STING agonist) with always-on immune stimulation property. Thereby, the elicitation and duration of innate immunity could be spatiotemporally tailored by the external US stimulation. The US-activatable turning on/off of innate immunity ensured to mitigate the immune-related adverse effects on untargeted tissues. Second, the previously reported adjuvants generally activate single innate immune pathways, and the sustained exposure to mono innate immune stimuli is intend to induce APCs dysfunction63–65. In contrast, our sono-adjuvant represented a multivalent platform to synchronously activate multiple innate immune pathways. Third, distinct from other strategies that aimed to precisely activate immune agonists within tumor tissues, our sono-adjuvant strategy specifically leveraged US stimulation to modulate innate immune responses within LNs, which directly synergized with antigen cross-presentation in APCs to boost antigen-specific T cell responses. Moreover, the sono-adjuvant represented a highly biodegradable and generalizable vaccine platform, and could have broad implications to co-deliver various tumor antigens for personalized cancer immunotherapy. It was worth noting that although the critical advantages of US-activatable sono-adjuvant in cancer vaccine immunotherapy, additional efforts are required to address the current limitations for future clinical translation. First, from the manufacturing aspect, a more precise control of polypeptides self-assembly by automated microfluidic devices should be explored to produce uniform sono-adjuvant and minimize batch-to-batch variation, which could address the critical translational challenges from bench to bedside. Second, from the innate immune activation aspect, although sono-adjuvant represented a multiply adjuvanted platform to synchronously activate STING, TLR9, and NLRP3 pathways in APCs, the underlying potential and mechanism of US-activatable innate immune stimulation strategy for addressing APCs dysfunction and exhaustion were not extensively investigated. Third, the induction of long-term antitumor immune responses was highly essential for preventing tumor recurrence and distant metastasis. Therefore, additional tumor-rechallenge assay and more extensive immune memory response monitoring should be conducted to overcome the current limitations of short-term experimental observation in this study. Moreover, the current sono-adjuvant platform was merely examined in various types of mouse tumor models. Although US therapy exhibited high tissue penetration depth and might be promising for stimulating human LNs, the tissue uptake of US stimulation may be variable in the treatment of patients compared with mouse models. Therefore, an extensive and personalized exploration of the influence of different US parameters on innate immune responses should be conducted when applied for clinical immunotherapy. Furthermore, the high-frequency ultrasound imaging (HFUS) technique should be further explored and combined to precisely visualize LNs that located at the superficial internal66, and thus can guide to perform LNs-specific US stimulation and innate immune activation. The precise monitoring of activation status of different immune cell populations during US-guided immune activation treatment was highly essential for achieving personalized cancer immunotherapy. In future studies, we could combine the multi-color flow cytometry assay to continuously monitor the activation status of different immune cell populations in the peripheral blood, such as detecting the expression of T cell activation-related markers (e.g., CD69, CCR7, CD44, PD1). Based on these results, we could adjust the treatment parameters, such as the frequency, duration and power density of US stimulation, to accordingly regulate the activation of immune cell populations. We expected the sono-adjuvant platform could be optimized in various aspects to improve its therapeutic application scopes, and implement a safe and effective cancer vaccination strategy. Methods Ethics statement All animal experiments were approved by the Institutional Animal Care and Use Committee of the Shanghai Institute of Materia Medica, Chinese Academy of Sciences (2024-09-YHJ-06). The study complied with the ethics committee’s limit on maximal tumor size (1500 mm3) and the requirement that animal body weight loss did not exceed 20%. Neither of these thresholds was exceeded at any point during the study. Materials HMME, MitoTracker Red CMXRos, Yoda-1, Dooku-1, Fluo-4 AM, and DCFH-DA were purchased from MedChemExpress. The STING agonist MSA-2, NAC (SJ-MX0166), were purchased from Shandong Sparkjade Biotechnology Co., Ltd. DAPI, Lysotracker Red were purchased from Meilun Biotech CO., Ltd. OVA257-263 peptide, FITC-labeled OVA257–264 peptide, 4T1 tumor-specific neoantigen (SHRSCSHQTSAPSPKALAHNGTPRNAI) peptide, and Hepa1-6 tumor-specific neoantigen (RWLYWQPTL) peptide were purchased from BankPeptide Biological Technology Co., Ltd. CpG-ODN 1585 (SOC-A1585) was purchased from Synbio Tech (Suzhou/NJ). The HRP-labeled goat-anti-rabbit IgG (H + L) (P8002) was purchased from New Cell & Molecular Biotech. The Cell Freezing Medium (S924JV) was purchased from BasalMedia Co., Ltd. The pre-stained protein ladder (PM008) was purchased from Novoprotein. The D-Luciferin was purchased from Yeasen. Mouse GM-CSF (GMF-M5211) was purchased from ACROBiosystems. Mouse M-CSF (PCK043) was purchased from Procell. Mouse IL-4 protein (PRP2117) was purchased from Abbkine. ELISA kit for mouse TNF-αwas purchased from Elabscience Biotechnology Co., Ltd. ELISA kit for mouse IL-6 was purchased from absin. Caspase 1 Activity Assay Kit was purchased from Abcam. ELISA kits for mouse IFN-γ, mouse IFN-β and mouse TGF-β were purchased from NeoBioscience Technology Co., Ltd. Glutamic-pyruvic Transaminase Assay Kit (BC1555) was purchased from Beijing Solarbio Science & Technology Co., Ltd. Mouse Lymphocyte Separation Medium was purchased from Shenzhen Dakewe Bio-engineering Co., Ltd. L-Tyrosine (Tyr), 2-iodoxybenzoic acid, sodium dithionite (Na2S2O4), 1-Hydroxybenzotriazole, 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide Hydrochloride (EDCI) and azodiisobutyronitrile were purchased from J&K chemical. Triphosgene was purchased from TCI. Methoxy PEG-NH2 (Mn = 5000) was obtained from SINOPEG. N, N-Dimethylformamide (DMF), dimethyl sulfoxide (DMSO), and ethyl ether were all offered by SinoPharm Chemical Reagent Co., Ltd. Cell lines and animals DC2.4 cells (cat. GNM51), RAW264.7 cells (cat. TCM13), B16-OVA cells (cat. TCM2), and 4T1 cells (cat. TCM32) were all purchased from the Cell Bank of Shanghai, Chinese Academy of Sciences. Hepa1-6-Luc cells were provided by Prof. XL Liu from Mengchao Hepatobiliary Hospital of Fujian Medical University. RAW264.7-Lucia ISG cells and THP1-Blue ISG cells were provided by Prof. MY Zheng from Shanghai Institute of Material Medica, Chinese Academy of Sciences. The construct expressing luciferase was introduced into Hepa1-6, and 4T1 cells by lentiviral transduction for bioluminescence imaging. The cell lines were authenticated by the supplier using STR analysis. All cell lines were tested negative for mycoplasma contamination. BMDCs and BMDMs were differentiated from bone marrow cells isolated from the femurs of 6-week-old C57BL/6 mice according to the established method. All the cells were maintained in DMEM medium containing 10% FBS (LONSERA, Suzhou Shuangru Biotechnology Co., Ltd.) and 1% penicillin/streptomycin, and maintained at 37 °C in a humidified incubator with 5% CO2 supply. BALB/c mice (cat. SM-003, 4–5 weeks, female) and C57BL/6 mice (cat. SM-001, 4–5 weeks, female) were purchased from Shanghai Model Organisms Center. Ifnar1-KO C57BL/6 mice (cat. C001268, female, 4–5 weeks) was purchased from Cyagen (Suzhou). All animal procedures were performed under the guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of the Shanghai Institute of Materia Medica, Chinese Academy of Sciences. According to the animal ethics of institute and animal welfare, animals were euthanized by carbon dioxide inhalation when their tumor volumes (V) exceeded 1500 mm3 in all the antitumor studies. V = (L × W × W)/2) (L, the longest axis, W, the shortest axis). The tumor size of mice was not exceeded 1500 mm3 in all the animal studies. Synthesis of the Man-mPEG113-b-P(Tyr40)-HMME sonosensitive polypeptide To synthesize the macroinitiator Man-mPEG113-NH2, NH2-mPEG5k-NH2 (100 mg, 0.02 mmol), α-D-mannopyranosylphenyl isothiocyanate (3.14 mg, 0.01 mmol), and DIEA (2.58 mg, 0.02 mmol) were reacted in anhydrous DMF for 24 h at room temperature. The product was obtained by dialyzing against ethanol and DI water. The structure was examined by 1H-NMR spectrum. Next, the Man-mPEG113-b-P(Tyr40) polypeptide was synthesized through ring-opening polymerization of L-tyrosine N-carboxyanhydride (Tyr NCA) using macroinitiator Man-mPEG5k-NH2, as described in above statement. Finally, to synthesize sonosensitive Man-mPEG113-b-P(Tyr40)-HMME polypeptide, HMME (15.9 mg, 0.013 mmol), DMAP (8.0 mg, 0.33 mmol), EDCI (12.6 mg, 0.33 mmol), and DIEA (8.8 mg, 0.33 mmol) were dissolved in anhydrous DMF. After stirring for 1 h, the DMF solution of Man-mPEG113-b-P(Tyr40) (100.0 mg, 0.0082 mmol) was added dropwise into the mixture, and stirred continued for 24 h at room temperature. Then, the reaction was precipitated with cold diethyl ether to obtain Man-mPEG113-b-P(Tyr40)-HMME as black powder. All newly generated materials are available for reuse by contacting the corresponding author upon reasonable request. Synthesis of the charge-modified polypeptides The mPEG113-b-P(Lys40), mPEG113-b-P(Tyr10-Lys30), mPEG113-b-P(Tyr20-Lys20), mPEG113-b-P(Tyr30-Lys10), and mPEG113-b-P(Tyr40) polypeptides were synthesized through ring-opening co-polymerization of lysine N-carboxyanhydride (Lys NCA) and tyrosine N-carboxyanhydride (Tyr NCA) using mPEG113-NH2 as the macroinitiator. Typically, NH2-PEG5k (100 mg, 0.018 mmol) and indicated mass of Lys NCA and Tyr NCA were dissolved in anhydrous DMF under argon. The reaction was stirred for 3 days at room temperature and precipitated with cold diethyl ether to obtain indicated polypeptides as white powder. The mPEG113-b-P(TyrP10-Lys30), mPEG113-b-P(TyrP30-Lys10), and mPEG113-b-P(TyrP40) polypeptides were synthesized via phosphorylation of the phenolic hydroxyl groups of tyrosine. For instance. mPEG113-b-P(Tyr40) (100 mg, 0.0087 mmol) was suspended in anhydrous N-methylpyrrolidone under ice bath. Anhydrous triethylamine (200 μL) and phosphorus oxychloride (200 μL) were then slowly added. The mixture was stirred in the darkness overnight and saturated NaHCO3 solution was added. After stirring for another 3 h, the mixture was dialyzed against DI water and lyophilized to obtain the mPEG113-b-P(TyrP40) as yellow solid. Preparation and characterization of the sono-adjuvant The SONA library were prepared by a nanoprecipitation method. Typically, 1.0 mg of Man-mPEG113-b-P(Tyr40)-HMME sonosensitive polypeptide or mPEG113-b-P(Tyr40)-HMME sonosensitive polypeptide in 50 μL DMSO and 0.5 mg of indicated charge-modified polypeptides was drop by drop added into 1.0 mL deionized water under supersonic vibration (85 W, on 5 s, off 2 s, 3 min). The resultant nanoparticles were purified through ultrafiltration tube (Mw = 100 kDa) and dispersed in deionized water for further use. The hydrodynamic diameter and zeta potential of nanoparticles were measured in pH 7.4 buffer solution by dynamic laser scattering (DLS). US stimulation parameters The US equipment was purchased from DJO France S.A.S. (Cat. number: DJO 2776, Frequency: 1 MHz, Power density: 0–1.5 W/cm2, Duty cycle: 50%, Transducer type: COLLIMATING, Effective radiating area: 4 cm2). We selected the US frequency of 1 MHz and the duty cycle of 50% based on the following considerations. First, the US frequency of 1 MHz has good penetration depth in biological tissues, which allows it to delivery energy into the LNs67. When combined with a 50% duty cycle, the energy is evenly distributed and can ensure sufficient energy transfer to achieve the treatment effect during the penetration process. Second, the combination of 1 MHz frequency and 50% duty cycle can avoid excessive damage to the tissue caused by continuous high-energy output, allowing to reduce adverse effects such as thermal effects on tissues68. Third, this combination was demonstrated to produce a moderate cavitation effect, and can avoid unnecessary damage to the surrounding tissues caused by overly intense cavitation69–71. Western blot assay The cells or tissues were harvested, and proteins were extracted using IntactProtein™ lysis kit (GenuIN, 415). Then the protein concentration was measured by the bicinchoninic acid protein assay kit according to the manufacturer’s instruction. Equal amounts of protein were mixed with an equal volume of 5 × loading buffer and boiled at 100 °C for 20 min. Gel electrophoresis and protein transformation were then performed. The target proteins were then probed with specific primary antibodies followed by secondary antibodies conjugated with horseradish peroxidase. Immunoreactive proteins were visualized using a Tanon 5200 chemiluminescent imaging system (Tanon). Target protein expression was semi-quantified with ImageJ software. Cytokine detection For cytokine analysis in cell supernatant, the sample was collected at 24 h after US stimulation. For cytokine analysis in tumor tissue, the sample was collected and homogenized in cold PBS in the presence of digestive enzymes to form single-cell suspensions, and the supernatant after centrifugation was collected. The cell supernatant or intratumor cytokine levels of TNF-α, IL-6, IL-1β, IFN-β, and TGF-β were measured with ELISA kits according to the manufacturer’s instructions. Analysis of antigen-specific immune response in vitro BMDCs (2 × 105 cells per well) were incubated with various formulations with the concentration of HMME (1.0 μg/mL), OVA (10 μg/mL), MSA-2 (10 μM), and CpG (2.0 μg/mL). BMDCs in SONV + US group were stimulated with 0.2 W/cm2 of US (1 MHz, 50% duty cycle, 3 min) at 6 h post-incubation. The cells were continuously cultured for 24 h and then stained with corresponding antibodies to analyze the maturation and antigen presentation levels of BMDCs. For analyzing CD8+ OT-1 T cell proliferation in vitro, BMDCs were treated with the same formulations as described above, and then co-cultured with CFSE-stained OT-1 T cells for 72 h (T cell: BMDC = 10:1). CD8+ T cell proliferation was then assessed by flow cytometry. Preparation and characterization of sono-vaccine SONA was used as the vaccine carrier, and protein antigen or neoantigen peptide was selected as the tumor antigen for preparing sonosensitive vaccine. Briefly, 100 μL of OVA protein solution (0.2 mg) was added into 1.0 mL of as-prepared SONA solution. The resultant mixture was kept stirring at room temperature for 60 min, and then purified using an ultrafiltration tube (Mw = 100 kDa). The OVA loading efficiency and capacity in SONV was determined by a bicinchoninic acid protein assay kit. To prepare Ne-SONV, 100 μL of 4T1 tumor-specific neoantigen peptide (SHRSCSHQTSAPSPKALAHNGTPRNAI) or Hepa1-6 tumor-specific neoantigen peptide (RWLYWQPTL) (0.2 mg) was drop by drop added into 1.0 mL of as-prepared SONA solution. The resultant mixture was kept stirring at room temperature for 60 min, and then purified using an ultrafiltration tube (Mw = 100 kDa). The particle size and zeta potential of resultant SONV, Ne-SONV were evaluated in pH 7.4 buffer solution by a Zeta-sizer Nano ZS90 (Malvern Instruments, England). Lymph node-targeting and cellular uptake of SONV in vivo C57BL/6 mice were subcutaneously injected at the tail base with free OVA-FITC or SONV-FITC at an identical OVA-FITC dose of 5.0 mg/kg. Fluorescence signal in LNs and major organs were examined 6 h or 12 h post-injection with an IVIS Spectrum Imaging System. To determine the cellular uptake of SONV-FITC by APCs in vivo, LNs were harvested and digested to obtain single-cell suspensions. The cells were washed with PBS three times and stained with corresponding antibodies and then analyzed by flow cytometry. Vaccination and antigen-specific immune response in vivo C57BL/6 mice were vaccinated with indicated formulations with the identical dose of HMME (1.0 mg/kg), OVA (1.0 mg/kg), MSA-2 (2.0 mg/kg), and CpG (2.0 mg/kg) three times at one week interval. The LNs of SONV + US group mice were stimulated with 1.0 W/cm2 US (1 MHz, 50% duty cycle, 5 min) at 6 h post per SONV injection. The LNs were harvested 3-day after the final vaccination and digested to obtain single cell suspensions. The cells were stained with corresponding antibodies to analyze the maturation and antigen presentation levels of DCs and macrophages. To evaluate antigen-specific CD8+ T cell response in vivo, the spleen of vaccinated mice was harvested, digested, and passed through 200-mesh nylon mesh filters to obtain single-cell suspensions. The splenocytes were then stained with corresponding antibodies to analyze the antigen-specific CD8+ T cell response using the tetramer staining assay. To evaluate the antigen-dependent cytotoxic effect of CD8+ T cells, 1 × 106 splenocytes were collected from the vaccinated mice and stimulated with 10 μg/mL SIINFEKL peptide for 24 h. The cells were then stained with corresponding antibodies to analyze the cytotoxic CD8+ T cell response. In vivo cytotoxicity killing assay We performed an in vivo cytotoxicity killing assay to detect the OVA-specific cytotoxic T cell response generated by SONV + US. Half of splenocytes from untreated mice were labeled with OVA peptide and high concentration of CFSE (5.0 μM, TargetMol, T6802), denoted as OVAp+ cells. Meanwhile, the other half of splenocytes were labeled with low concentration of CFSE (0.5 μM) and not labeled with OVA peptide, denoted as OVAp− cells. Then, the OVAp+ cells (5 × 106 per mouse) and OVAp− cells (5 × 106 per mouse) were equally mixed and injected intravenously into the SONV + US-treated mice. After 24 h, spleen from treated mice was collected and subjected to flow cytometry analysis. The numbers of CFSEhigh and CFSElow cells were determined and used to calculate the percentage of OVA-specific cell killing. Percentage of specific lysis = [1 − non-transferred control ratio/experimental ratio] × 100%. Assessment of systemic immune side effects To evaluate systemic immune activation associated with different adjuvant strategies, serum levels of IL-6 and IL-12p70 were measured following indicated treatments. Female C57BL/6 mice were randomly assigned into three groups: OVA+CpG (50 μg CpG), OVA + MSA-2 (50 μg MSA-2) and SONV + US (50 μg HMME). All formulations contained 50 μg OVA antigen and were subcutaneously injected at the tail base of mice on day 0, day 7, and day 14, respectively. For SONV + US group mice, the draining lymph nodes were stimulated with US (1 MHz frequency, 50% duty cycle, 1.0 W/cm² intensity, 5 min) at 6 h post each injection. Peripheral blood was collected via submandibular vein at 1 day and 7 day post each injection. Serum was isolated and stored at −80 °C until analysis. Levels of IL-6 and IL-12p70 were quantified using indicated ELISA kits following the manufacturer’s protocol. RNA-seq analysis The LNs from PBS or SONV + US group mice were harvested to extract the total RNA according to the manufacturer’s protocol. RNA purification, reverse transcription, library construction, and sequencing were performed at Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. according to the manufacturer’s instructions (Illumina, San Diego, CA). To identify DEGs between two different samples, the expression level of each transcript was calculated according to the transcripts per million reads (TPM) method. Differential expression analysis was performed using the DESeq2. DEGs with |log2FC| ≥ 1 and adjusted p-value ≤ 0.05 were considered to be significantly different expressed genes. In addition, functional-enrichment analysis of Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/) was performed to identify which DEGs were significantly enriched in metabolic pathways at Bonferroni-corrected p-value ≤ 0.05 compared with the transcriptome background. KEGG pathway analysis were carried out by Python scipy software. The Gene Set Enrichment Analysis (GSEA) did not use the MSigDB database. Instead, it used the annotation results from KEGG pathway analysis. The ranking algorithm used for GSEA was Signal2Noise. The protein-protein interaction network analysis on DEGs was conducted by using the STRING database (combined score ≥ 0.4, http://string-db.org/). Antitumor efficacy in multiple murine tumor models To evaluate the prophylactic effect of SONV + US against tumor growth and metastasis, C57BL/6 mice were vaccinated with indicated formulations with the identical dose of HMME (1.0 mg/kg), OVA (1.0 mg/kg), MSA-2 (2.0 mg/kg), and CpG (2.0 mg/kg) three times at one week interval. The LNs of SONV + US group mice were stimulated with 1.0 W/cm2 US (1 MHz, 50% duty cycle, 5 min) at 6 h post per SONV injection. 1 × 106 B16-OVA cells were subcutaneously or intravenously injected into vaccinated mice on day 7 after the final vaccination. The body weight and tumor volume of group mice were monitored every two days. To evaluate the therapeutic effect of SONV + US against established tumor growth, 1 × 106 B16-OVA cells were subcutaneously injected into the right flank of C57BL/6 mice on day 0, and mice were randomly grouped on day 7 and treated with indicated formulations as described above three times at one week interval. The body weight and tumor volume of group mice were monitored every two days. To evaluate the combinational antitumor effect of Ne-SONV + US with PD-L1 blockade therapy in orthotopic TNBC model, 1 × 106 4T1-Luc cells were injected into the breast pad of BALB/C mice on day 0, and mice were randomly grouped on day 7, and treated by subcutaneous injection of Neop + CpG or Ne-SONV with identical dose of HMME (1.0 mg/kg), neoantigen peptide (1.0 mg/kg), CpG (2.0 mg/kg), and intraperitoneal injection of αPDL1 (5.0 mg/kg). The LNs of Ne-SONV + US and Ne-SONV + US + αPDL1 group mice were stimulated with 1.0 W/cm2 US (1 MHz, 50% duty cycle, 5 min). The body weight, tumor volume of group mice were monitored every two days and the bioluminescence imaging of group mice were conducted every ten days. To evaluate the combinational antitumor effect of Ne-SONV + US with PD-L1 blockade therapy in orthotopic HCC model, 5 × 106 Hepa1-6-Luc cells were injected into the liver of C57BL/6 mice on day 0, and mice were randomly grouped on day 7, and treated by Ne-SONV + US or αPDL1 as described above. The bioluminescence imaging of group mice were conducted every ten days. mtDNA-TLR9 blocking assay For the mtDNA-TLR9 blocking, 200 µg of AT-791 (TargetMol, USA) was administered to the C57BL/6 mice by intraperitoneal injection 1 day before SONV + US treatment and every third day thereafter for five times. IL-1β blocking assay For the IL-1β blocking, 100 µg of anti-IL-1β (BE0246) (BioXcell) was administered to the C57BL/6 mice by intraperitoneal injection 1 day before SONV + US treatment and every third day thereafter for five times. CD8+ T cell blocking assay For the CD8+ T cell blocking, 100 μg of anti-CD8α (53–6.7) (BioXcell) was administered to the C57BL/6 mice by intraperitoneal injection 1 day before SONV + US treatment and every third day thereafter for five times. The depletion of CD8+ T cells in spleen was confirmed by flow cytometry. Analysis of tumor-infiltrating immune cells in vivo B16-OVA, Hepa1-6 and 4T1 tumor-bearing mice model were established and treated as described above. Tumor tissues were collected 3-days post final treatment, and processed through mechanical disruption and digested by collagenase IV (0.2 mg/mL), hyaluronidase (0.2 mg/mL) and DNase I (0.1 mg/mL) for 90 min at 37 °C. The cells were then filtered through 200-mesh nylon filters and washed three times with PBS to obtain single-cell suspensions. The collected cells were stained with corresponding antibodies to analyze tumor-infiltrating CD4+ T cells, CD8+ T cells, NK cells, IFN-γ+CD8+ T cells, Gzmb+CD8+ T cells, IFN-γ+ NK cells, Gzmb+ NK cells, Tregs, IL-10+ Tregs, IL-4+ Tregs, MDSCs, M2-TAM, and M1-TAM. Statistics and reproducibility For quantitative analysis, a minimum of three biological replicates were analyzed. Western blotting and immunofluorescent staining experiments were repeated three times independently with similar results. All data were presented as mean ± SD unless otherwise indicated. Statistical analysis was performed using GraphPad Prism software 8.2.1. Two-tailed Student’s t test and One-way analysis of variance (ANOVA) with a Tukey post-hoc test were used for the statistical comparison between the two groups and among multiple groups, respectively. The survival study was analyzed through log-rank (Mantel–Cox) test. No statistical method was used to predetermine sample size. No data were excluded from the analyses. The investigators were not blinded to allocation during experiments and outcome assessment. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting Summary Transparent Peer Review file Source data Source Data Acknowledgements This work was supported by the Major Project of the Study on Pathogenesis and Epidemic Prevention Technology System by the Ministry of Science and Technology of China (2021YFC2302501 to H.J.Y.), National Natural Science Foundation of China (U22A20328, W2412035, 32411530049 to H.J.Y., and 82203041 to L. Y.), Strategic Priority Research Program of the Chinese Academy of Sciences (XDB1060000 to H.J.Y.), Science and Technology Commission of Shanghai Municipality (23490712700 to Z.A.X., 24430713500 and 23ZR1475000 to H.J.Y.), Shanghai Oriental Talents Program (BJKJ2024046 to Z.A.X.), and State Key Laboratory of Chemical Biology, Open Funds of State Key Laboratory of Drug Research (SKLDR-2023-KF-09 to H.J.Y.). The Mass Spectrometry System and the cell sorter BD Influx of the National Facility for Protein Science in Shanghai (NFPS), Shanghai Advanced Research Institute, CAS are gratefully acknowledged. All animal procedures were carried out under the guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of the Shanghai Institute of Materia Medica, CAS. All schematic illustrations were created by the Adobe Illustrator 2022 software. Author contributions F.C., Z.X., and H.Y. designed the study. F.C. performed the study and wrote the primary manuscript. F.C., H.Z., and S.L. conducted the experimental studies and analyzed the data. S.R., L.H., Y.G., Y.L., and Z.C. assisted with the experimental studies. X.L., Z.X., and H.Y. revised the manuscript. Peer review Peer review information Nature Communications thanks Jingfei Liu, Sadman Labib, Lijia Yu, Xing-Jie Liang, Ningqiang Gong, and the other anonymous reviewer for their contribution to the peer review of this work. [A peer review file is available]. Data availability The RNA-seq data generated in this study have been deposited in the NCBI Sequence Read Archive database under accession code: PRJNA1224287. All OTHER data are included in the Supplementary Information. The raw numbers for charts and graphs are available in the Source Data file. Source data are provided with this paper. Competing interests The authors declare no competing interests Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Zhiai Xu, Email: zaxu@chem.ecnu.edu.cn. Haijun Yu, Email: hjyu@simm.ac.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41467-025-66976-2. References Associated Data Supplementary Materials Supplementary Information Reporting Summary Transparent Peer Review file Source Data Data Availability Statement The RNA-seq data generated in this study have been deposited in the NCBI Sequence Read Archive database under accession code: PRJNA1224287. All OTHER data are included in the Supplementary Information. The raw numbers for charts and graphs are available in the Source Data file. Source data are provided with this paper.