Boosting immunogenic tumour cell death via nanotherapeutic targeting of the Stanniocalcin 1 phagocytosis checkpoint for enhanced cancer immunotherapy

Abstract

Eliciting calreticulin (CRT) surface exposure is essential for triggering immunogenic cell death (ICD). However, stanniocalcin 1 (STC1) suppresses CRT translocation by sequestering it within mitochondria, limiting ICD induction in tumours. Here, we show that silencing STC1 enhances CRT surface exposure in Lewis lung carcinoma (LLC) cells when combined with paclitaxel (PTX), converting dying tumour cells into an in-situ vaccine that drives immunoprevention of tumour growth. To maximize this therapeutic synergy, we engineered a nanoplatform co-delivering siSTC1 and PTX, in which PTX is covalently conjugated to a sphingolipid and siSTC1 is electrostatically encapsulated (siSTC1/LNP-PTX). This system improves pharmacokinetics, synchronizes co-delivery to tumours, and enhances intratumoral exposure. Consequently, it amplifies CRT expression, promotes antigen-presenting cell-mediated phagocytosis and antigen presentation, and elicits robust cytotoxic T cell responses in LLC models. Moreover, siSTC1/LNP-PTX sensitizes tumours to PD-1 blockade. Our nanosystem, which unlocks ICD potential by silencing STC1, represents a paradigm-shifting approach to cancer immunotherapy.

Calreticulin-mediated activation of immunogenic cell death (ICD) is hindered by stanniocalcin-1 (STC1), which blocks its surface translocation. Here, the authors report the antitumor immune benefits of STC1 silencing and present a nanoplatform co-delivering siSTC1 and the ICD inducer paclitaxel in Lewis lung carcinoma models.

Introduction

Immunogenic cell death (ICD) instigated by select chemotherapy and radiation etc. has been emerging as one of the most powerful means to transform the cancer treatment paradigm, particularly in the realm of cancer immunotherapy¹. Distinguished from regular apoptosis or necrosis, which is commonly not immunogenic or even tolerogenic, ICD promotes immune activation to fortify the antitumour T cell immunity. This is accomplished through inducing the hallmarks of damage-associated molecular patterns (DAMPs), especially the translocation of the calreticulin (CRT) from the endoplasmic reticulum to the surface of the dying cancer cells, which serves as the “eat-me” signal for antigen presenting cells (APCs) uptake to initiate the ICD process, facilitating the processing and presentation of tumour-associated antigens, thereby eliciting a robust T cell mediated antitumour immune response^2–4. ICD not only targets and eradicates primary and metastatic tumour cells but also establishes long-term immunological memory effects, enabling durable protection against cancer recurrence^5,6. Additionally, ICD can convert the tumour microenvironment into a more immunogenic state to synergize with immune checkpoint inhibitors for amplified therapeutic efficacy^7,8.

Stanniocalcin 1 (STC1), a hormone-like glycoprotein involved in carcinogenesis, has recently been identified as a crucial phagocytosis immune checkpoint to arouse immune evasion. Specifically, STC1 binds to CRT and actively sequesters CRT within mitochondria, which hinders the translocation of CRT to the cell surface, dampening membrane CRT-directed phagocytosis by APCs such as dendric cells and macrophages^9,10. Consequently, this leads to impaired antigen processing and suboptimal presentation to T cells^9,10. The entrapment of CRT by STC1 disrupts the paramount step in the ICD cascade, weakening activation of cytotoxic T cells and diminishing anti-tumour immune responses. Thus, STC1-mediated immune resistance presents a formidable challenge to the ICD-enabled immunotherapy and targeting STC1 and its interaction with CRT to unleash the therapeutic potential of ICD for cancer immunotherapy is of significant impact.

We discovered genetic silencing tumour STC1 indeed synergizes with a series of classic ICD inducers (PTX, camptothecin, docetaxel, irinotecan, and oxaliplatin) to potentiate CRT surface expression in LLC cells that overexpress STC1, which was achieved by abrogating STC1-mediated mitochondrial sequestration of CRT, thereby reinstating pivotal role of CRT as the “eat me” signal for APC phagocytosis. However, the enhanced CRT expression was not observed in STC1-low MC38 cells. Interestingly, this synergy was also not achievable when combing with non-ICD agents such as gemcitabine (GEM), cisplatin, topotecan, ceritinib, and canertinib. Further, combining the silencing of STC1 and ICD inducer (PTX) augmented the ICD potential by eradicating majority of LLC tumours in a vaccination-based model, but which was not observed when paired with non-ICD inducer GEM, nor in STC1-low MC38 tumours. These findings highlight the importance of the STC1-CRT axis in modulating ICD.

To effectively translate this combination approach into in vivo setting, we engineered a lipid nanoparticle (LNP) system to synchronize co-delivery of STC1 silencing and ICD induction in tumours. For proof of concept, PTX and GEM were chosen as the ICD and non-ICD inducer, respectively, and were covalently attached to a backbone phospholipid. To enable the co-encapsulation of an siRNA targeting STC1 (siSTC1), DLin-MC3-DMA (DMA), an ionizable lipid used in an FDA-approved siRNA LNP Onpattro¹¹, was used to electrostatically entrap siSTC1. The siRNA/LNP-drug improved the pharmacokinetic performance, allowed more distribution to tumours, and imparted the co-localization of PTX and siSTC1. siRNA/LNP-PTX yielded the better anticancer efficacy than the co-administration regimen through fortifying the surface CRT exposure, enhancing phagocytosis and tumour-associated antigen presentation, and subsequently activating the CD8⁺ T cell response in LLC tumours model but not in STC1-low MC38 tumour model or when siSTC1 was co-delivered with a non-ICD inducer GEM. Notably, siSTC1/LNP-PTX sensitized tumours unresponsive to checkpoint blockade to anti-PD-1 therapy, eliminating 60% tumours. Our nanotechnology-empowered targeting of STC1-CRT unlocks the ICD potential, representing a paradigm-shifting cancer immunotherapy platform.

Results

Silencing STC1 synergized with ICD inducers to fortify the CRT surface exposure

To evaluate the role of STC1 in modulating CRT surface exposure, we selected LLC cells, known for high STC1 expression, and MC38 cells, which exhibit low STC1 levels¹². Consistent with literature, our Western blot results further verified their STC1 expression levels in these two cancer cell lines (Fig. 1a, e). We examined a series of classic ICD (PTX, camptothecin, docetaxel, irinotecan, and oxaliplatin)^13–15 and non-ICD (GEM, cisplatin, topotecan, certinib, canertinib)^16–19 inducers on their ability to trigger CRT translocation. First, through cytotoxicity study (Fig. 1b, f and Supplementary Figs. 1, 2), we identified the IC20, IC35, and IC50 doses for these drugs, which were used to treat the LLC and MC38 cells. The flow cytometric analysis demonstrated that the IC20 doses of the ICD inducers increased the CRT surface expression compared to no treatment control, which was further enhanced by IC35 doses in both LLC and MC38 cells (Fig. 1c, d, g, h). However, further increasing the doses to IC50 for ICD inducers failed to further boost the CRT translocation, indicating the ICD-induced CRT expression reached plateau (Fig. 1c, g). Thus, the IC35 was used for further studies. ELISA further confirmed that STC1 expression in LLC cells was significantly reduced by the combination of ICD inducers and siSTC1/LNP compared with ICD inducers alone (Supplementary Fig. 3). Of note, combining ICD inducers with STC1 silencing leveraging the siSTC1 delivered by a lipid nanoparticle (LNP) markedly augmented CRT surface expression in the LLC cells (Fig. 1c). To confirm the necessity of the LNP delivery system and the specificity of the siSTC1, flow cytometric analysis showed that CRT surface exposure was significantly enhanced by PTX in combination with siSTC1/LNP, but not when combined with free siSTC1 or LNPs loaded with a negative control siRNA (Supplementary Fig. 4). In congruence with literature, non-ICD inducers were unable to stimulate noticeable CRT levels on cancer cells surface²⁰, even with the co-treatment of siSTC1/LNP (Fig. 1d, h). These findings suggest that the suppression of STC1 overexpression can effectively potentiate CRT membrane translocation in response to ICD inducers, but not in the presence of non-ICD agents.

Fig. 1. Silencing STC1 synergized with ICD inducers to potentiate CRT expression.

Representative Western blots depicting STC1 protein levels in LLC (a) and MC38 (e) cells following treatment with siSTC1/LNP (50 nM, lipid composition DLin-MC3-DMA, DSPC, Cholesterol, PEG_2K-C-DMG for LNP) for 24 h⁴⁹ (n = 3 independent experiments, similar results were observed). The table delineating the IC20, IC35 and IC50 of various drugs in LLC (b) and MC38 (f) cells based on MTT assay after 24 h treatment. Flow cytometry analysis of calreticulin (CRT) surface expression on LLC (c, d) and MC38 (g, h) cells treated with ICD inducers (c, g) and non-ICD inducers (d, h) at their respective IC20, IC35 and IC50 concentrations or in combination with siSTC1/LNP at 50 nM for 24 h, respectively. Data are presented as mean ± s.d. (n = 3 independent experiments). Statistical significance was determined by one-way ANOVA followed by Tukey’s multiple comparisons test. Source data are provided as a Source Data file.

Combination of siSTC1 and Paclitaxel enhanced APC phagocytosis through upregulating the surface CRT by reducing CRT trapped in mitochondria

To delve deeper into the mechanism on how the siSTC1 synergized with ICD inducer to potentiate the CRT surface expression, PTX was chosen as the ICD inducer for proof-of-concept study and GEM was selected as the non-ICD inducer control. To visualize the surface CRT, confocal imaging was performed. We found that CRT surface translocation was significantly enhanced in PTX treated LLC cells, particularly when combined with siSTC1/LNP (Fig. 2a, c). Mechanistically, via comprehensive Western blot analysis of the CRT proteins in various cellular organelles, the synergy was determined to be accomplished by reducing the CRT trapped in mitochondria so that more membrane CRT translocation was realized (Fig. 2e and Supplementary Fig. 7). To ensure the reliability of these organelle-specific findings, Western blot analysis of marker proteins from non-target organelles was conducted to verify the high purity of the isolated fractions (Supplementary Fig. 5). Furthermore, to explore the physical relationship between these proteins, LLC cell lysates were subjected to immunoprecipitation (IP) using an anti-STC1 antibody. Subsequent Western blot analysis for CRT revealed an interaction between STC1 and CRT, which was absent in negative controls (IgG, beads only) and diminished in cells treated with siSTC1/LNP (Supplementary Fig. 6). Interestingly, silencing STC1 alone cannot decrease the mitochondria CRT levels, thereby, membrane CRT levels were not impacted. However, this synergistic CRT enhancement was not achieved in MC38 (STC1-low) cells (Fig. 2b, d). Consistently, non-ICD inducer, GEM could not markedly increase the surface CRT levels, even in combination with siSTC1/LNP in both cell lines.

Fig. 2. Intracellular delivery of siSTC1 synergized with PTX to heighten APC phagocytosis.

a, b Illustrations to show tumour-derived STC1 impairs immune responses by acting as an intracellular “eat-me” signal blocker. STC1 sequesters CRT within mitochondria, preventing its surface expression on tumour cells, thus, hindering APC phagocytosis and subsequent T cell activation. Confocal imaging to visualize the membrane CRT in LLC (c) and MC38 cells (d) treated with siSTC1/LNP (50 nM), PTX (0.64 µM or 11.34 µM), PTX + siSTC1/LNP, GEM (0.03 µM or 1.12 µM), GEM + siSTC1/LNP for 24 h (n = 3 independent experiments, similar results were observed). Scale bars: 20 µm. e Western blots showing CRT distribution in different organelles in LLC cells after the same treatments as (c, d). β-actin was whole CRT loading control. Na⁺, K⁺-ATPase were membrane CRT loading control. LAMP1 was lysosome CRT loading control. TOM20 was mitochondria CRT loading control. Bip was Endoplasmic reticulum CRT loading control (n = 3 independent experiments, similar results were observed). f Effect of STC1 on DCs-mediated phagocytosis. DCs were incubated with dying cells from CFSE-labelled LLC or MC38 cells. The mean fluorescence intensity (MFI) in DCs was determined by flow cytometry (n = 3 independent experiments). g–k Cartoon to depict the vaccination approach. C57BL/6 mice received 2 times (one week apart) of s.c. injections of dying LLC or MC38 cells (1 × 10⁶) treated as (c,d). On day 0, Mice (n = 5) were inoculated with living LLC or MC38 cells subcutaneously on contralateral flank (g). Individual tumour growth curves (h, j) to show tumour developed on the contralateral side. Percentage of tumour-free mice (i, k). The tumour-free mice in PTX + siSTC1/LNP group in h–k were rechallenged with fresh LLC cells at the day 35 or MC38 cells at the day 18 and the tumour growth were monitored (Supplementary Figs. 14,15). Data in f are presented as mean ± s.d. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test. Source data are provided as a Source Data file.

Since membrane CRT serves as an “eat me” signal that facilitates APC-mediated phagocytosis^21,22, we then investigated the impact of CRT exposure on dendritic cells (DCs)-mediated phagocytosis²³. We co-cultured DCs with CFSE-labelled dying LLC cells and observed that PTX treatment significantly enhanced DCs phagocytosis. This effect was further amplified when PTX was combined with siSTC1/LNP (Fig. 2f and Supplementary Fig. 8).

To elucidate how the improved CRT surface exposure and DCs-enabled phagocytosis affected the ICD potential, we performed the gold-standard vaccination approach using the dying cancer cells^19,24–26 (Fig. 2g–k). Dying LLC or MC38 cells induced by various treatments were subcutaneously (s.c.) injected twice into one flank of immunocompetent C57BL/6 mice at one week apart. Subsequently, live LLC or MC38 cells were injected into the contralateral flank on day 0 (Fig. 2h, j and Supplementary Figs. 9–13). Vaccination with PTX-treated LLC or MC38 cells led to significant tumour reduction and eradicated 1 or 2 out of 5 tumours, respectively, confirming the potent ICD-elicited antitumour immune response of PTX. Remarkably, co-administration of siSTC1/LNP further enhanced the tumour growth inhibition and eliminated 4 or 3 out of 5 LLC and MC38 tumours, respectively (Fig. 2h–k). Noteworthily, the tumour-free mice in PTX+siSTC1/LNP-treated group (Fig. 2h–k) successfully rejected the rechallenged fresh cancer cells growth, corroborating a robust memory immunity was established by ICD activation (Supplementary Figs. 14, 15).

In accordance with literature¹⁶, non-ICD inducer GEM-treated cells failed to produce marked preventive vaccination effect, albeit in combination with siSTC1/LNP in both LLC and MC38 models (Fig. 2g–k). Collectively, our both in vitro and in vivo findings demonstrated that silencing STC1 overexpression can significantly fortify ICD effects elicited by ICD inducers, further bolstering anti-tumour immune responses.

Development of sphingomyelin (SPH)-derived nanotherapeutic vesicles for co-delivering siSTC1

SPH-derived drug nanovesicles have been proven to increase drug loading, improve the pharmacokinetics and tumour delivery, as well as enhance the therapeutic efficacy compared to unconjugated drug and drugs physically loaded in nanoparticles^27,28. To boost the therapeutic delivery of the ICD induced by PTX, we have synthesized the SPH-conjugated PTX bridged by a labile ester bond (Supplementary Figs. 16–19), which can self-assemble into nanovesicles with the other helper lipids (DMA, SPH, Chol, PEG), producing a LNP-PTX size of ~78 nm and good monodispersity (PDI = 0.17) (Supplementary Fig. 21). Of note, the optimized LNP-PTX can also effectively co-deliver the siSTC1 (at the N/P ratio of 6) based on the electrostatic interaction due to the use of the DMA ionizable lipid that was used in FDA-approved siRNA nanoformulation Onpattro (Fig. 3b)²⁹. Co-loading of siSTC1 did not markedly affect the size, zeta, morphology, and stability of the LNP-PTX (Fig. 3b, d, e and Supplementary Figs. 21–23). To evaluate the drug release profile, cumulative release studies were performed under physiological (pH 7.4) and acidic lysosomal (pH 5.5) conditions. The results for PTX and GEM from their respective siSTC1/LNP platforms (Supplementary Fig. 24a, b) demonstrated pH-responsive release. Additionally, consistent Cy5-siSTC1 release patterns were confirmed across both ICD and non-ICD inducer formulations (Supplementary Fig. 24c). A gel retardation assay revealed that free siRNA in serum was quickly degraded. However, siSTC1/LNP was able to significantly increase the siRNA stability as evidenced by the no degradation even after 48 h in serum (Fig. 3f). We further evaluated the Stc1 gene knockdown efficiency in LLC (Fig. 3g) and MC38 (Fig. 3h) cells. Western blot analysis confirmed that siSTC1/LNP-PTX significantly reduced STC1 levels in tumour cells, similar to the results observed with siSTC1/LNP. To assess the functional impact of this knockdown on secretion, ELISA was performed to quantify secreted STC1 in culture supernatants from LLC cells. Treatment with siSTC1/LNP or siSTC1/LNP-PTX significantly reduced secreted STC1 levels compared with controls (Supplementary Fig. 25). The therapeutic advantage of the LNP delivery was further demonstrated through CRT quantification. Flow cytometry analysis showed that LNP-loaded ICD inducers significantly increased cell-surface calreticulin (CRT) expression compared with free ICD inducers (Supplementary Fig. 26a, b). Notably, LNP-PTX induced higher surface CRT levels in LLC cells than free PTX or PTX/LNP (Supplementary Fig. 27a), likely due to enhanced cellular uptake (Supplementary Fig. 27c). This enhancement was specific to ICD inducers, as LNP-GEM did not significantly increase surface CRT expression (Supplementary Fig. 27b).

Fig. 3. Development of sphingomyelin (SPH)-derived lipid nanoparticle (LNP) platform for co-delivering siSTC1.

a A schematic diagram illustrating the preparation of siSTC1/LNP-PTX or siSTC1/LNP-GEM, composed of SPH-PTX or SPH-GEM conjugate, DLin-MC3-DMA (DMA), SPH, PEG (2000)-C-DMG, and cholesterol (Chol), for the co-delivery of siRNA (siSTC1) and paclitaxel (PTX) or gemcitabine (GEM). b Physicochemical characterization of various LNP formulations. DLS, dynamic light scattering; d.nm, diameter values in nanometres. Representative DLS size histogram (c), size (d) and zeta potential (e) monitoring. f Serum stability analysis of free siSTC1 or siSTC1/LNP, siSTC1/LNP-PTX and siSTC1/LNP-GEM (mixed with FBS, v/v = 1:1, incubated at 37 °C) by gel retardation assay with 1% agarose gel electrophoresis (n = 3 independent experiments, similar results were observed). Western blots depicting STC1 protein levels in LLC (g) and MC38 (h) cells following treatment with various formulations for 24 h (n = 3 independent experiments, similar results were observed). Data in b, d and e are presented as mean ± s.d. (n = 3 independent experiments). Source data are provided as a Source Data file.

Additionally, LNP-GEM and siSTC1/LNP-GEM were developed to serve as the negative ICD inducer nanoparticle controls, which showed similar physicochemical property, siRNA stability protection and gene silencing efficiency (Fig. 3a–e).

LNP-PTX increased the circulation time and tumour distribution of siSTC1 and PTX

To determine if the co-delivery LNP could facilitate the in vivo delivery of siSTC1 and PTX, pharmacokinetics and biodistribution studies were performed in LLC tumour-bearing mice (Fig. 4 and Supplementary Figs. 28–31). To monitor the siSTC1, a Cy5-labeld siSTC1 was used. Free Cy5-siSTC1 was rapidly eliminated after 4 h with little tumour accumulation. Nevertheless, Cy5-siSTC1-laden LNP significantly improved the pharmacokinetic profile by prolonging the blood circulation time, increasing the area under curve (AUC) and mean residence time (MRT), and attenuating the clearance (CL) and volume of distribution (V), as well as delivering drastically more siSTC1 to tumour (Fig. 4a–d and Supplementary Figs. 28, 29). Similar phenomena were observed in Cy5-siSTC1/LNP-PTX or Cy5-siSTC1/LNP-GEM. To visualize delivery efficiency, we examined siSTC1 extravasation and intratumoural penetration in subcutaneous LLC tumours. Confocal imaging of tumour cryosections 24 h post-injection revealed minimal signal from free Cy5-siSTC1, whereas Cy5-siSTC1/LNP-PTX showed robust deep tumour penetration (Supplementary Fig. 30).

Fig. 4. siSTC1/LNP-PTX improved the pharmacokinetic profiles and fortified the tumour delivery.

a–c LLC tumour-bearing mice (n = 3 mice, tumour, ~300 mm³) were intravenously administered by various Cy5-siSTC1 formulations at 1 mg Cy5 kg⁻¹. a siSTC1 blood fluorescence imaging at various time points. b Blood kinetics of siSTC1 based on the fluorescence intensity as shown in a. c The pharmacokinetic parameters of siSTC1 analysed by PKSolver software (n = 3 mice). d Ex vivo Lago imaging of tumours and major organs at 24 h. e–h LLC tumour-bearing mice (n = 3 mice, tumour, ~300 mm³) were intravenously administered by various formulations at 20 mg PTX kg^-1. e Blood kinetics of PTX. f The pharmacokinetic parameters of PTX analysed by PKSolver software. PTX intratumoural release (g) and tissue distribution (h) at 24 h. The percent injected doses in LNP-PTX represent the released PTX and SPH-PTX. Drug contents in plasma and major tissues were measured by HPLC (Supplementary Fig. 31). Data in b, c and e–h are expressed as mean ± s.d. (n = 3). Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test. Source data are provided as a Source Data file.

Moreover, we found that compared to PTX, LNP-PTX with or without siSTC1 was able to markedly increase the half-life, AUC, MRT and reduce the CL of PTX, and rendered much higher tumour uptake rate while efficiently releasing PTX inside tumours (Fig. 4e–h and Supplementary Fig. 31). Taken together, siSTC1/LNP-PTX not only can improve the PK and tumour accumulation, but it can also synchronize the in vivo delivery and enable tumour colocalization of PTX and siSTC1, vital for enhancing the antitumour immunity and efficacy.

Silencing STC1 enhances the ICD response and antitumour efficacy of LNP-PTX

We then evaluated the therapeutic efficacy of siSTC1/LNP-PTX in LLC tumour mouse model first following the i.v. injections of various formulations at eq. 20 mg PTX/kg and 1 mg siSTC1/kg (Fig. 5a). Tumour progression in PBS treated group was uncontrolled, underscoring the aggressive nature of this tumour model. Free siSTC1 administration had a negligible effect on tumour growth, while siSTC1/LNP moderately suppressed tumour progression, indicating that siSTC1, delivered via LNPs, effectively inhibited tumour growth by targeting and silencing overexpressed STC1 (Fig. 5b and Supplementary Fig. 32), aligning with previously reported findings¹². To confirm in vivo efficacy, Western blot analysis of LLC tumour tissue lysates demonstrated that treatment with siSTC1/LNP-PTX, LNP-PTX + siSTC1/LNP, or siSTC1/LNP-PTX + siSTC1/LNP markedly reduced STC1 protein levels compared with controls (Supplementary Fig. 33). Notably, LNP-PTX demonstrated superior antitumour efficacy compared to PTX. Furthermore, combining siSTC1/LNP with LNP-PTX significantly enhanced the therapeutic effect. However, the co-delivery formulation siSTC1/LNP-PTX outperformed the co-administration regimen (LNP-PTX + siSTC1/LNP), achieving the most pronounced tumour inhibition without significant body weight loss (Supplementary Fig. 34), likely due to its significantly higher cellular uptake in LLC cells (Supplementary Fig. 35), and was further corroborated by a notable extension in overall survival (Fig. 5c).

Fig. 5. Co-delivery siSTC1/LNP-PTX enhanced antitumour efficacy and immune responses compared to co-administration regimen in LLC mouse model.

a Drug administration timeline scheme in s.c. LLC tumour mice (n = 5 mice; tumours, ~100 mm³). Mice were intravenously injected at eq. 1 mg siSTC1 kg⁻¹ and 20 mg PTX kg^-1 on days 15, 17 and 19. b Average tumour growth curve. c In a parallel study, the mice Kaplan–Meier survival curves were monitored (n = 5 mice). d Representative flow histogram (left panel) and quantitative analysis (right panel) for intratumoural CRT expression on LLC tumour cells in (b) on day 25. e Representative flow cytometric plots (left panel) of intratumoural CD8⁺/Granzyme B⁺ T cells, CD8⁺/IFN-γ⁺ T cells, CD80⁺/CD86⁺ DCs and CD11c⁺/CD103⁺ DCs and respective quantitative analysis (right panel) in (b). (n = 3 tumours, which were randomly chosen from b on day 25). f Drug administration timeline scheme in s.c. MC38 tumour mice (n = 5 mice; tumours, ~100 mm³). Mice were intravenously injected at eq. 1 mg siSTC1 kg⁻¹ and 20 mg PTX kg⁻¹ on days 8, 10 and 12. g Average tumour growth curve were monitored. Representative flow histogram (h) and quantitative analysis (i) for intratumoural CRT expression on MC38 tumour cells in (g). j Representative flow cytometric plots of intratumoural CD8⁺/Granzyme B⁺ T cells, CD8⁺/IFN-γ⁺ T cells, CD80⁺/CD86⁺ DCs and CD11c⁺/CD103⁺ DCs and respective quantitative analysis (right panel) in (g) (n = 3 tumours, which were randomly chosen from g on day 18). Data in b, d, e, g, i and j are presented as mean ± s.d. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test; survival curves were compared using the log-rank Mantel-Cox test. Source data are provided as a Source Data file.

To elucidate the underlying mechanisms for the improved efficacy achieved in siSTC1/LNP-PTX, we investigated CRT expression on tumour cell surface using flow cytometry. The results revealed that siSTC1/LNP-PTX had the highest level of CRT surface expression, the initiator of ICD (Fig. 5d and Supplementary Fig. 36a). Subsequent analysis of diverse immune cell populations, including IFN-γ⁺ and granzyme B⁺ CD8⁺ T cells, as well as CD80⁺/CD86⁺ and CD103⁺ dendritic cells (DCs), unveiled that siSTC1/LNP-PTX significantly fortified the tumour infiltration of these immune cells compared to PTX, LNP-PTX, PTX + siSTC1/LNP, LNP-PTX + siSTC1/LNP (Fig. 5e and Supplementary Fig. 36b–e). To evaluate whether the siSTC1/LNP-PTX system could induce systemic anti-tumour immunity, we established a bilateral tumour model. As shown in Supplementary Fig. 37, siSTC1/LNP-PTX treatment significantly inhibited the growth of both left and right tumours, indicating that this therapy induced a robust systemic anti-tumour immune response. Collectively, these findings support that the improved therapeutic efficacy of siSTC1/LNP-PTX could be attributed to the boosted ICD instigation, and augmented antigen presentation and ultimately the enhanced cytotoxic T cell antitumour immunity. To assess potential off-target effects, we performed H&E histological analysis on major organs including the liver and spleen. All examined tissues exhibited well-preserved structures across treatment groups, with no evidence of hepatocyte degeneration, necrosis, or inflammatory infiltration. The integrity of the splenic white and red pulp further indicated that immune organ structures remained intact. These results suggest that systemic administration does not induce detectable histopathological toxicity or disrupt immune homeostasis (Supplementary Fig. 38).

Nonetheless, such antitumour efficacy and immune response improvements over co-administration of LNP-PTX + siSTC1/LNP were not observed for siSTC1/LNP-PTX in STC1-low MC38 tumour mice (Fig. 5f–j and Supplementary Figs. 39, 40), indicating the critical role of the high STC1 expression in tumours.

In addition to ICD inducer, LNP-PTX, we also investigated the therapeutic potential of non-ICD inducer, LNP-GEM in co-delivering siSTC1 in both LLC and MC38 tumour models (Supplementary Figs. 41–46). While GEM showed marked tumour reduction in both models, LNP-GEM outperformed free GEM in delaying tumour growth, demonstrating the merit of leveraging a nanodelivery platform. Unfortunately, regardless of the expression levels of the siSTC1 in tumours, siSTC1/LNP-GEM failed to produce the benefits over the co-injection of the LNP-GEM + siSTC1/LNP on anticancer efficacy and immunity in both LLC and MC38 models, underpinning the fact that triggering the ICD is the indispensable step to synergize with siSTC1 for improved efficacy.

siSTC1/LNP-PTX potentiates PD-1 immune blockade therapy

Since siSTC1/LNP-PTX can induce a more immunogenic phenotype in the tumour microenvironment, we hypothesize that it may enhance the effectiveness of immune checkpoint blockade (anti-PD-1, BioXcell, Clone RMP1-14) therapy, which is otherwise ineffective in treating LLC tumours (Fig. 6a, b and and Supplementary Fig. 47). We confirmed that α-PD-1 monotherapy failed to elicit marked tumour reduction compared to vehicle control. While the combination of PTX + siSTC1/LNP markedly improved the efficacy of PD-1 blockade, co-administration of LNP-PTX + siSTC1/LNP-PTX with α-PD-1 further enhanced antitumour effects. The therapeutic improvements were more pronounced in siSTC1/LNP-PTX with further prolonged mouse survival (Fig. 6b–d), demonstrating the advantage of using a nanoparticle-enabled co-delivery strategy. The enhanced antitumour activity was found to be arising from the significantly boosted ICD effects as well as the fortified tumour-infiltrating cytotoxic T cell immunity demonstrative of the upregulated CRT and HMGB-1 and increased levels of the CD8, IFN-γ, Granzyme B and Perforin in tumour tissues (Fig. 6e and Supplementary Fig. 48).

Fig. 6. siSTC1/LNP-PTX sensitized PD-1 immune checkpoint blockade therapy.

a Drug administration timeline scheme in s.c. LLC tumour mice (n = 5 mice; tumours, ~100 mm³). Mice were intravenously injected at eq. 1 mg siSTC1 kg^-1, 20 mg PTX kg^-1 and/or i.p. α-PD-1 (100 μg per mouse) on days 15, 17 and 19. Individual tumour grow curves (b) and average tumour growth curves (c). d In a parallel study, the mice Kaplan–Meier survival curves were monitored (n = 5 mice). e Immune phenotypic analysis of the tumours in c using immunohistochemistry (n = 3 independent experiments, Scale bar = 50 µm). Data in b and e are presented as mean ± s.d. Statistical significance in b and e was determined using one-way ANOVA followed by Tukey’s multiple comparisons test; survival curves were compared using the log-rank Mantel–Cox test. Source data are provided as a Source Data file.

Discussion

Despite the significant potential of ICD-enabled antitumour immunotherapy, its therapeutic applications remain underutilized. CRT translocation to the cell surface triggers the “eat-me” signal that stimulates ICD-induced antitumour immune responses^30–34. However, recent studies have shown that tumour STC1 acts as an intracellular “eat-me” signal blocker by sequestering CRT in mitochondria, inhibiting APC phagocytosis, and impeding T cell activation and responses¹². Elevated tumour STC1 levels have been negatively correlated with the efficacy of immunotherapies (e.g., checkpoint blockade) and patient survival^12,35–37. Therefore, targeting STC1 offers a strategy to maximize the therapeutic potential of ICD-enabled immunotherapy.

We confirmed that genetically silencing the Stc1 gene synergized with 5 commonly used ICD inducers (PTX, Camptothecin, Docetaxel, Irinotecan, Oxaliplatin) to potentiate the CRT membrane exposure on tumour cells (Fig. 1), Mechanistically, this effect was achieved by reducing the CRT trapped in mitochondria by STC1, without affecting its levels in the lysosome or endoplasmic reticulum (Fig. 2). However, STC1 silencing did not increase CRT expression when combined with non-ICD agents, suggesting that STC1 primarily interferes with CRT translocation to the cell membrane rather than directly impacting CRT expression levels. These results highlight the critical importance of incorporating ICD inducers into siSTC1-based therapeutic regimens to effectively enhance antitumour immune responses. Moreover, the enhanced ICD response was observed only in immune-cold LLC tumours with high STC1 expression, but not in immune-hot MC38 tumours with low STC1, nor when siSTC1 was combined with non-ICD agents. This underscores that the synergistic increase in CRT surface exposure is contingent upon ICD induction and requires a significant reduction of STC1 expression, which is elevated in immunosuppressive tumour tissues.

We also unveiled that silencing STC1 alone is insufficient to promote CRT translocation to the cell membrane, as this process requires an external trigger, such as ICD inducers. In addition, the translocation of CRT to the membrane positively regulated DCs-mediated phagocytosis of dying tumour cells, and that inhibiting STC1 augmented this immunological process. The improved synergistic ICD observed in the LLC tumour model can be attributed to (1) abnormally higher STC1 levels, (2) heightened membrane CRT expression, and (3) enhanced DC phagocytosis (Figs. 1a, e, and  2c, d, f). These findings underscore the importance of targeting the STC1-CRT axis to modulate ICD effectively.

Leveraging SM-derived nanotechnology^{27,28,38–41}, via systemically screening of various lipid compositions/ratios (Supplementary Fig. 21), we engineered an optimal co-delivery LNP nanosystem that effectively incorporated siSTC1 and the ICD inducer, PTX simultaneously (Fig. 3). The siSTC1/LNP-PTX nanosystem unleashed the antitumour effects by outperforming monotherapy or co-administration regimen through minimizing the tumour growth and extending the mouse survival in LLC tumour model (Fig. 5). Notably, siSTC1/LNP-PTX sensitized immune-cold LLC tumours to anti-PD-1 therapy, which showed minimal response to PD-1 inhibition alone (Fig. 6). The enhanced efficacy and synergy can be attributed to two key factors. First, physiochemically, the co-delivery LNP system improved the in vivo stability of siSTC1, maximizing its gene knockdown efficiency, while significantly improving pharmacokinetics by increasing half-life, AUC, and MRT, and reducing volume of distribution (V) and clearance (CL). This also resulted in greater drug/siRNA accumulation in tumours (Figs. 3, 4). Also, siSTC1/LNP-PTX synchronized the co-delivery of PTX and siSTC1 to the tumour (Fig. 4), facilitating temporo-spatial synergistic effects.

At the molecular level (Figs. 5d, e, 6e), the ICD machinery was amplified, as evidenced by increased intratumoural CRT, which serves as the “eat-me” signal for antigen-presenting cells (APCs) to uptake dying tumour cells, and HMGB-1, an adjuvant stimulus for APC activation⁴².

Additionally, comprehensive immune phenotyping of the tumours revealed: (1) a notable upregulation of mature antigen-presenting DCs, as indicated by elevated levels of CD80 and CD86, co-stimulatory molecules essential for T cell activation, expansion, and differentiation; (2) increased expression of CD103 in CD11c⁺ DCs, which is critical for promoting CD8⁺ T cell responses and antitumour immune responses^43,44; (3) enhanced IFN-γ cytokine expression in CD8⁺ T cells, a key subset of cytotoxic T lymphocytes (CTLs) that directly kill cancer cells through the release of perforin and granzyme B. Collectively, the stimulated and strengthened “ICD-APC phagocytosis-recruited and activated CTLs immunity” signalling switched the immunosuppressive LLC tumour into immunogenic and inflammatory phenotype, unlocking the therapeutic efficacy of immune checkpoint blockade.

However, we did not observe improved efficacy with siSTC1/LNP-GEM compared to the combination of LNP-GEM and siSTC1/LNP (Supplementary Figs. 41, 42). This can be attributed to GEM, a non-ICD inducer, being unable to enhance CRT expression in tumour tissues, even with STC1 silencing (Supplementary Figs. 41f, 42e, f). As a result, insufficient CRT membrane expression likely impaired the recruitment and activation of antigen-presenting DCs and, consequently, CTLs, which are critical for strong antitumour immunity (Supplementary Figs. 41g, 42g). These findings highlight the limitations of non-ICD inducers, like GEM, in overcoming tumour immunosuppression and unµµderscore the necessity of using an ICD inducer in combination therapies.

While this study focuses on mechanistic validation of STC1 as an intracellular phagocytosis checkpoint, the siSTC1/LNP-PTX platform is designed with clinical translation in mind. Notably, the LNP architecture incorporates the ionizable lipid DLin-MC3-DMA, which is the same clinically validated component used in the FDA-approved siRNA therapy Onpattro (patisiran). This shared formulation element supports translational feasibility with respect to manufacturability, scalability, and regulatory familiarity.

From a specificity standpoint, the therapeutic effect of STC1 silencing is shown to be context-dependent, occurring predominantly in STC1-high tumours (e.g., LLC) but not in STC1-low models (e.g., MC38). This tumour-selective dependency mitigates concerns regarding systemic immune activation and suggests a precision-medicine approach in which STC1 expression could serve as a predictive biomarker for patient stratification. Importantly, the platform co-delivers siSTC1 with an ICD inducer, ensuring synchronized tumour exposure and minimizing off-target gene silencing in non-ICD contexts.

Potential translational challenges include siRNA-associated innate immune activation, nanoparticle-related toxicity, and off-tumour gene knockdown. However, our in vivo safety studies show no detectable histopathological toxicity in major organs and preserved immune homeostasis following systemic administration. In addition, the use of pH-responsive drug release and tumour-preferential accumulation further enhances therapeutic index.

Looking forward, this nanoplatform could be clinically positioned as (i) a combination immunotherapy to sensitize immuno-cold tumours to checkpoint blockade or (ii) a biomarker-guided therapy for STC1-high malignancies. Collectively, these features support the translational promise of STC1-targeted ICD amplification using a clinically precedent LNP system.

In summary, our systematic evaluation confirmed the detrimental role of the STC1 phagocytosis checkpoint in tumour ICD immunity and immunotherapy. We demonstrated that genetically silencing STC1 synergized with ICD inducers like PTX to enhance CRT-mediated immunogenicity by attenuating the CRT trapped in mitochondria. The sphingolipid-derived nanoplatform we developed, which targets STC1 and its interaction with CRT while co-delivering PTX, fully unlocks the therapeutic potential of this combination regimen. This strategy offers a approach to enhancing the efficacy of tumour immunotherapy and provides a solution to overcome the limited responsiveness of tumours to immune checkpoint blockade therapy.

Methods

Ethics statement

This research complies with all relevant ethical regulations. The animals were maintained under pathogen-free conditions and all animal experiments were approved by The University of Arizona Institutional Animal Care and Use Committee (IACUC). Tumour size was periodically assessed using a digital caliper and calculated using the formula: 0.5 × length × width². The maximal permitted tumour size was 2000 mm³ according to the animal ethics guidelines of IACUC and animal welfare regulations, and the mice were sacrificed once the tumour volume grew to ≥ 2000 mm³ or the status of the mice became moribund. Nevertheless, the tumour size of some mice has grown greater than 2000 mm³ by the final day of measurement, and the mice were sacrificed immediately.

Cells and mice

The LLC cell line (ATCC) and the MC38 cell line (Kerafast) were cultured in complete DMEM medium, while the DC 2.4 cell line (Sigma Aldrich, SCC142) was maintained in RPMI-1640 medium. All media were supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine at a constant temperature of 37°C in a CO₂ incubator. Female C57BL/6 mice, aged 5–6 weeks and acquired from The Jackson Laboratory, were employed in the study. Standard Individually Ventilated Caging (IVC) system was used to maintain the mice under pathogen-free conditions. The animal house was kept at a temperature of 68–72 °F and indoor humidity of 30–70% to abide by the NIH Guide and in accordance with the guidelines of 12 h light/12 h dark by 7 am on-7pm off.

ICD biomarkers analysis

In 12-well plates, a total of 1 × 10⁵ LLC or MC38 cells were initially plated. After 12 h, the cell culture medium was aspirated and replaced with media containing different concentrations of siSTC1/LNP (50 nM), PTX (IC20, IC35 and IC50), PTX (IC20, IC35 and IC50) plus siSTC1/LNP (50 nM), free GEM (IC20, IC35 and IC50), and free GEM (IC20, IC35 and IC50) plus siSTC1/LNP (50 nM) for a 24 h incubation period.

For the assessment of cellular surface calreticulin expression via flow cytometry, cells were trypsinized and subjected to three cold PBS washes. They were then stained with calreticulin antibody [Alexa Fluor® 647] (#NB600-101AF647, 1/500, Novus Biologicals) in staining buffer (#00-4222-26, eBioscience) at 4 °C for 30 min. After washing twice with PBS at 4 °C, cells were resuspended in 500 μL of staining buffer and analysed using a BD FACSCanto™ II flow cytometer (BD Bioscience). The data were presented as a fold increase in the mean positive percentage ratio compared to the no-treatment control group.

For the assessment of cellular surface calreticulin expression via immunofluorescence, LLC or MC38 cells were fixed with 4% paraformaldehyde and stained with calreticulin Antibody [Alexa Fluor® 647] (#NB600-101AF647, 1/500, Novus Biologicals). The cell nuclei were stained with Hoechst 33342 (#3570, 1/2000, Invitrogen™) for 5 min. Slides were sealed with ProLong™ Glass Antifade Mount (#P36980, Invitrogen™). Fluorescence images were obtained using a Leica DMI6000B microscope with Leica LAS X 3.7 software. The images were analysed using LAS X 3.7 software (v. 3.7.3.23245).

Vaccination experiment

LLC or MC38 cells were treated with PTX (IC35), PTX (IC35) plus siSTC1/LNP (50 nM), free GEM (IC35), and free GEM (IC35) plus siSTC1/LNP (50 nM) for 24 h. Afterward, 1 × 10⁶ dying LLC or MC38 cells in 100 μL of serum-free DMEM medium (4 °C) were subcutaneously inoculated twice into the left flank of C57BL/6 mice (n = 5) at an interval of 7 days. Control group mice were injected with 100 μL of serum-free DMEM medium. 7 days after the second inoculation, 1 × 10⁶ living LLC or MC38 cells in 100 μL of serum-free DMEM medium were subcutaneously injected into the right flank of the mice. Tumour incidence, tumour growth curves, and mouse body weights were closely monitored.

Preparation siSTC1/LNP-PTX and siSTC1/LNP-GEM

To efficiently encapsulate siSTC1, the ionizable lipid DLin-MC3-DMA (DMA, WuXi AppTec), which is used in the FDA-approved siRNA nanotherapeutic Onpattro, was added to the lipid bilayer. Briefly, DMA, SPH, Chol, and PEG2000-C-DMG (#B2699-358145, BOC Sciences) at a molar ratio of 49.29/10.19/39.03/1.5, as in Onpattro¹¹, were used as the control lipid and dissolved in ethanol. For the preparation of siSTC1/LNP-PTX and siSTC1/LNP-GEM. siSTC1 was dissolved in 1.85 volumes of 25 mM sodium acetate buffer (pH 4) according to a nitrogen-to-phosphate (N/P) ratio of 6. The N/P ratio describes the ratio of amine groups (N) within the ionizable lipid to the phosphate groups (P) of the siRNA backbone. The two solutions were mixed under vigorous stirring, resulting in spontaneous LNP formation in an aqueous buffer containing 35% ethanol. Afterward, LNPs were dialysed against PBS (pH 7.4) overnight at 4 °C. Finally, the LNPs were passed through 0.2 µm Nuclepore™ Track-Etched Membranes (#0306048, Whatman) 10 times using an extruder (#610000, Avanti).

The size, zeta potential, and morphology of the LNPs were characterized using dynamic light scattering (DLS) with a Zetasizer Nano (Nano-ZS, Malvern Panalytical) and analysed using Zetasizer software (v. 7.13). To assess siSTC1 encapsulation efficiency, The siRNA concentration in the LNPs was determined by Ribogreen assay after the NPs were disrupted by adding 0.5 w/v% Triton X-100. siRNA encapsulation efficiency was calculated by comparing siRNA concentration in the presence and absence of Triton X-100^45,46.

Transmission electron microscopy (TEM)

Lipid nanoparticles (LNPs) were characterized by transmission electron microscopy. LNP samples (total lipid concentration: 0.5 mg mL⁻¹) was deposited onto carbon-coated copper grids and incubated for 60 s. Excess liquid was removed by gently blotting the grid edge with filter paper. The grids were then negatively stained with 5 μL of 1% (w/v) phosphotungstic acid (PTA, pH 7.0) for 60 s, followed by removal of excess stain and air-drying at room temperature. TEM images were acquired using an FEI Tecnai G2 Spirit BT transmission electron microscope equipped with an AMT Image Capture Engine V6.02 (4-Mpix) digital camera at the University of Arizona ORP Imaging Cores-Electron (RRID:SCR_023279).

siSTC1 degradation assay

siSTC1 solutions, siSTC1/LNP, siSTC1/LNP-PTX, and siSTC1/LNP-GEM were incubated with fetal bovine serum (FBS) at a 1:1 (v/v) ratio at 37 °C to evaluate serum stability. At selected time points (0, 2, 4, 8, 12, 24 and 48 h), within a timeframe in which naked siSTC1 undergoes complete degradation, aliquots were collected and immediately snap-frozen at −80°C to halt further degradation. After thawing at 4°C, samples were lysed by adding Triton X-100, and siSTC1 was extracted using a chloroform:phenol:isoamyl alcohol (25:24:1) mixture. The extracted siSTC1 was then subjected to 1% agarose gel electrophoresis in TBE buffer (10.80 g/L Tris base, 5.5 g/L boric acid, and 0.58 g/L EDTA). Samples were mixed with Gel Loading Buffer (#B7024S, Biolabs) before separation at 100 V for 30 min. After electrophoresis, siSTC1 was visualized using UV trans-illumination and imaged with a GelDoc System (Bio-Rad, USA).

Western blot

For immunoblot analysis, whole-cell lysates were prepared using RIPA lysis buffer (#06182116, GenDEPOT). Protein concentrations were quantified using BCA Protein Assay Kits (#23227, Thermo Fisher) (Supplementary Fig. 20). Cell surface proteins were isolated using the Cell Surface Protein Isolation Kit (#89881, Thermo Fisher) and subsequently analysed via immunoblotting. Lysosomal and ER fractions were isolated following previously established protocols^12,47. Cells were homogenized in ice-cold isolation buffer (250 mM sucrose, 5 mM Hepes, pH 7.4) containing protease inhibitors. The homogenate was first centrifuged at 1000 g for 10 min to remove the nuclear fraction. The resulting supernatant was then subjected to ultracentrifugation at 100,000 g for 1 h to separate the cytosolic fraction from the post-nuclear particulate fraction, which was collected as a pellet. The pellet was resuspended in 0.5 mL of isolation buffer and loaded onto a discontinuous OptiPrep gradient (25%, 20%, 15%, and 10%) (#AXS-1114542, CoSPHo Bio). Ultracentrifugation was performed at 100,000 g for 2 h at 4 °C. Gradient fractions were sequentially collected from the top, and lysosomal and ER enrichment was verified by Western blotting using specific organelle markers. Mitochondria were isolated using the Mitochondria Isolation Kit (#89874, Thermo Scientific) according to the manufacturer’s protocol. The isolated proteins were collected by centrifugation, mixed with loading buffer (#AR1112, Boster), and heated to 95 °C for 5 min for denaturation. Protein samples were then loaded onto SDS-PAGE gels for electrophoresis and transferred onto membranes. Membranes were blocked with 5% milk in TBST buffer for 1 h and incubated overnight at 4 °C with the following primary antibodies: Anti-calreticulin (#ab92516, 1/400, Abcam), Anti-β actin (#4970, 1/1000, Cell Signalling Technology), Anti-Na⁺/K⁺-ATPase (#3010, 1/1000, Cell Signalling Technology), Anti-Tom20 (#42406, 1/1000, Cell Signalling Technology), Anti-BiP (#3177 T, 1/1000, Cell Signalling Technology), Anti-LAMP1 (#3243, 1/1000, Cell Signalling Technology). Following primary antibody incubation, membranes were thoroughly washed with TBST and incubated for 1 h at room temperature with an HRP-conjugated anti-Rabbit IgG secondary antibody (#W401B, 1/2000, Promega). After additional TBST washes, protein bands were visualized using the Azure Biosystems 600 imaging system (v.1.9.0.0406).

Phagocytosis assay

Flow cytometry was employed to investigate the kinetic effects of tumour-derived STC1 expression on DCs-mediated phagocytosis. LLC or MC38 cells were exposed to siSTC1/LNP, PTX, PTX + siSTC1/LNP, GEM, or GEM + siSTC1/LNP for 24 h. Following treatment, LLC or MC38 cells were fluorescently labelled with CFSE. DCs (2 × 10⁵) were subsequently incubated with CFSE-labelled LLC or MC38 cells (1 × 10⁶) for 12 h. To assess DC-mediated phagocytosis, cells were stained with an anti-CD11b antibody (#101208, BioLegend, 1/100 dilution) and analysed by flow cytometry (BD FACSCanto™ II, BD Biosciences). By gating on CD11b-positive DCs, phagocytic activity was quantified by measuring the mean fluorescence intensity (MFI) of CFSE.

Pharmacokinetics, biodistribution, and release rate in LLC tumours

To investigate the pharmacokinetics and biodistribution of Cy5-siSTC1 and PTX formulations, LLC tumour-bearing C57BL/6 mice (n = 3, tumour volume: ~300 mm³) were intravenously administered either free Cy5-siSTC1 (1 mg/kg) or Cy5-siSTC1-loaded LNPs (1 mg/kg) via the tail vein. Blood samples were collected at predefined time points (0.083, 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h post-administration), and plasma was isolated using BD Microtainer tubes followed by centrifugation at 12,000 g for 10 min. The fluorescence intensity of Cy5-siSTC1 in the serum was measured using a SpectraMax M3 reader (SoftMax Pro, v. 7.1.0), and its concentration was quantified based on a pre-established standard curve. Pharmacokinetic parameters were analysed using PKSolver software (version 2.0)⁴⁸. All plasma samples were stored at −80 °C in the dark until further analysis. At 24 h post-injection, serum samples were imaged using a Lago optical imager (Aura 64-bit analysis software, v. 3.2.0), and tumours along with major organs (heart, liver, spleen, lung, and kidneys) were harvested for ex vivo imaging to assess biodistribution.

For HPLC-based pharmacokinetics of PTX formulations, C57BL/6 mice bearing LLC tumours (n = 3, tumour volume: ~300 mm³) received intravenous injections of PTX (20 mg/kg), either as free PTX, LNP-PTX, PTX + siSTC1/LNP, or siSTC1/LNP-PTX. Blood was collected at the same predefined time points, and plasma was separated via centrifugation at 12,000 g for 10 min. Plasma samples were treated with methanol (90 μL methanol per 10 μL serum) for HPLC analysis to quantify PTX and its SPH-derived conjugates. At 24 h post-injection, tumours and major organs were excised, homogenized in methanol (900 μL per 100 mg tissue), and subjected to HPLC quantification to determine drug accumulation and tissue distribution.

Therapeutic efficacy investigation of siSTC1/LNP-PTX and siSTC1/LNP-GEM in LLC tumour model

C57BL/6 mice (n = 5) bearing subcutaneous LLC tumours were established by injecting 1 × 10⁶ LLC cells in 100 μL of serum-free DMEM into the right flank. Once tumours reached ~100 mm³, mice were intravenously administered PBS (vehicle control), free siSTC1, siSTC1/LNP (siSTC1, 1 mg/kg), PTX, LNP-PTX, PTX + siSTC1/LNP, LNP-PTX plus siSTC1/LNP, or siSTC1/LNP-PTX (PTX, 20 mg/kg) on days 15, 17, and 19, with tumour growth curves and body weight monitored throughout. On day 25, tumours were harvested for flow cytometry analysis. A separate cohort of LLC tumour-bearing mice (n = 5, tumour volume ~100 mm³) underwent the same treatment regimen in an independent study to assess survival, which was analysed via Kaplan–Meier plots. In a parallel experiment, another cohort of C57BL/6 mice (n = 5) with subcutaneous LLC tumours of ~100 mm³ received PBS (vehicle control), free siSTC1, siSTC1/LNP (siSTC1, 1 mg/kg), PTX, LNP-GEM, GEM + siSTC1/LNP, LNP-GEM plus siSTC1/LNP, or siSTC1/LNP-GEM (GEM, 20 mg/kg) on days 16, 18, and 20, with tumour progression and body weight monitored. On day 26, tumours were excised for flow cytometry analysis, and a separate survival study was conducted under the same treatment conditions, with Kaplan–Meier plots used to evaluate survival outcomes.

Therapeutic efficacy investigation of siSTC1/LNP-PTX in MC38 tumour model

C57BL/6 mice (n = 5) bearing subcutaneous MC38 tumours were established by injecting 1 × 10⁶ MC38 cells in 100 μL of serum-free DMEM into the right flank. Once tumours reached ~100 mm³, mice were intravenously administered PBS (vehicle control), free siSTC1, siSTC1/LNP (siSTC1, 1 mg/kg), PTX, LNP-PTX, PTX + siSTC1/LNP, LNP-PTX plus siSTC1/LNP, or siSTC1/LNP-PTX (PTX, 20 mg/kg) on days 8, 10, and 12, with tumour growth curves and body weight monitored throughout. On day 18, tumours were harvested for flow cytometry analysis.

Therapeutic efficacy investigation of siSTC1/LNP-PTX plus α-PD-1 in LLC tumour model

C57BL/6 mice (n = 5) bearing subcutaneous LLC tumours were injected with 1 × 10⁶ LLC cells in 100 μL of serum-free DMEM into the right flank. When tumours grew to ~100 mm³, mice were treated with PBS (vehicle control), α-PD-1 (100 μg/mouse, i.p.), PTX + siSTC1/LNP, PTX + siSTC1/LNP + α-PD-1, LNP-PTX + α-PD-1, siSTC1/LNP + LNP-PTX + α-PD-1, siSTC1/LNP-PTX, and siSTC1/LNP-PTX + α-PD-1 (PTX, 20 mg/kg; siSTC1, 1 mg/kg) on days 15, 17, and 19. Tumour growth and body weight were monitored, and on day 26, mice were dissected, and tumours were collected for Immunohistochemistry (IHC) analysis. Kaplan–Meier survival curves were generated to assess the impact on survival in each group.

Flow cytometry analysis

The dissected tumours were cut into all pieces by scissors on ice, and then digested in DMEM medium (0.5 mg/mL collagenase type I, Worthington Biochemical Corporation) for 1 h at 37 °C. To obtain single cell for analysis, the samples were meshed by a 70 μM cell strainer twice. The cell solution was incubated with Ack lysing buffer (Gibco, 2217610) to lyse the red blood cells according to the manufacturer’s protocols. The cell samples were washed with 4 °C PBS twice and resuspended in 4 °C staining buffer. After counting and aliquoting the cell, the cell suspensions were pre-incubated with FcBlock (TruStain fcXTM anti-mouse CD16/32, clone 93, 101320, BioLegend, 0.5 μg/100 μL) at 4 °C for 30 min to avoid nonspecific binding. Cells were stained with fluorescence-labelled antibodies at 4 °C for 30 min and live cells were gated using Zombie Violet™ (#423114, 1/100), CD45-PerCP/Cyanine5.5 (#103132,1/100), CD4-Brilliant Violet 510™(#100559, 1/100), mouse I-Ab Antibody- PE/Cyanine7 (#116420, 1/100), Granzyme B-APC(#396408, 1/100), IFN-γ-APC (#505810, 1/100) from BioLegend, CD3-APC-eFluor 780 (#47-0032-82, 1/100), CD8a-PE (#561095, 1/100), CD45-APC-Cy™7 (#557659, 1/100), CD11c-percp/Cyanine 5.5 (#560584, 1/100), CD86-PE (#553692,1/100), CD80-APC (#553766, 1/100), CD103-Alexa Fluor 647 (#566717, 1/100) from BD Biosceinces. Multi-parameter staining was used to measure the T cells: (1) IFN-γ⁺ T cells (CD3⁺/CD8⁺/IFN-γ⁺), (2) Granzyme B⁺ T cells (CD3⁺/CD8⁺/GranzymeB⁺), DCs cells: (1) CD80⁺/CD86⁺ DCs (CD45⁺/CD11c⁺/MCH-II⁺/CD80⁺/CD86⁺), (2) CD103⁺ DCs (CD45⁺/CD11c⁺/MCH-II⁺/CD103⁺), For intracellular staining, cells were fixed and permeabilized using a staining buffer set (eBioscience, #00-5523-00) followed by intracellular staining of IFN-γ, granzyme B. After washing, cells were measured on flow cytometry and analysed by FlowJo software (version 10.0.7, TreeStar). The numbers presented in the flow cytometry analysis images were percentage based.

Immunohistochemistry (IHC)

Tumour samples from therapeutic efficacy studies were harvested from sacrificed mice, fixed in 4% paraformaldehyde overnight, and subsequently processed and embedded in paraffin. The paraffin-embedded blocks were sectioned into 4 μm slices, which were mounted onto positively charged glass slides by the UACC TACMASR Core facility. Immunohistochemistry (IHC) staining procedures were performed according to established protocols. After staining, the sections were dehydrated and visualized using a Leica DMI6000B microscope, with a Leica DFC450 colour camera, and Leica LAS X 3.7 software. Histological images were analysed using LAS X 3.7 software (v. 3.7.3.23245). A board-certified veterinary pathologist evaluated the slides, and semi-quantitative analysis of IHC staining intensity was performed. Specifically, the mean intensity of 3,3’-diaminobenzidine (DAB) staining was divided by the total number of nuclei and quantified using ImageJ Fiji software (v. 1.2), according to established protocols. The semi-quantitative data were then normalized to vehicle control samples for comparison. The following primary antibodies were used for IHC staining anti-interferon gamma (#MM700, 1/200) was come from Thermo fisher, anti-HMGB1 (#ab18256, 1/400), anti-CD8α (#ab209775, 1/100), anti-granzyme B (#ab255598, 1/100) and anti-calreticulin (#ab92516, 1/400) were come from Abcam (Cambridge, UK) and anti-Perforin (#31647, 1/200) was come from Cell Signalling. All antibodies were diluted in Bond Primary Antibody Diluent (AR9352, Leica Biosystems) according to the manufacturer’s instructions.

Hematoxylin and eosin (H&E) staining

Liver and spleen samples were collected from euthanized mice at the endpoint of the therapeutic efficacy studies. Tissues were fixed overnight in 4% paraformaldehyde, followed by routine processing and paraffin embedding. Paraffin-embedded blocks were sectioned at a thickness of 4 μm and mounted onto positively charged glass slides by the UACC TACMASR core laboratory. Hematoxylin and eosin (H&E) staining (#H-3502, Vector labs) was performed according to standard histological protocols. Following staining, sections were dehydrated and examined using a Leica DMI6000B microscope equipped with a Leica DFC450 colour camera. Image acquisition and histological analysis were performed using Leica LAS X software (version 3.7.3.23245).

Statistical analysis

The level of significance in all statistical analyses was set at P < 0.05. Data are presented as mean ± s.d. and were analysed using the two tailed, unpaired Student’s t test for two groups or one-way analysis of variance (ANOVA) for three or more groups followed by Tukey’s multiple comparisons test using Prism 8.0 (GraphPad Software). Kaplan–Meier survival curves were compared with the log-rank Mantel-Cox test.

Reporting summary

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

Author contributions

J.L. conceived and supervised the project. J.L. and W.L. designed the experiments, analysed the data and wrote the manuscript. Z.W. assisted in the synthesis. M.L., Y.J., S.W., L.C. and M.K. assisted in nanoparticle preparation, sample preparation for HPLC analysis and in vivo animal studies. All authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks Nhiem Tran, Longguang Tang and the other anonymous reviewers for their contribution to the peer review of this work. [A peer review file is available].

Data availability

All the data supporting the findings of this study are available within the article and its Supplementary Information. The full image dataset is available from the corresponding author upon request. Source data are provided with this paper.

Competing interests

J.L. has applied for patents related to Sphingomyelin-derived drug nanotechnology. The other authors have no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-72526-1.