Obesity enhances ovarian cancer chemotherapy efficacy through C1q-mediated tumor targeting and immune activation

Abstract

Personalized protein corona significantly influences the biodistribution and therapeutic efficacy of nanomedicines, generating unique profiles that can impact treatment outcomes. Here, we demonstrate that pegylated liposomal doxorubicin (PLD) exhibits increased tumor accumulation and enhanced antitumor immunity in obese mice bearing ovarian tumor, inducing a greater capacity to inhibit tumor growth compared to normal mice. Mechanistically, the protein corona, particularly enriched with complement component 1q (C1q) in the plasma of obese mice, significantly enhances the internalization of PLD by ovarian cancer cells and elicits strong immunogenic cell death (ICD) effects. Concurrently, C1q adsorbed on PLD promotes the engulfment of apoptotic tumor cells by dendritic cells (DCs), activating T cell-mediated antitumor immune responses and amplifying the overall antitumor efficacy of PLD in obese mice. Our findings provide new insights into the role of the personalized protein corona in modulating the therapeutic response to chemotherapy and highlight the potential of targeting C1q for enhancing the efficacy of nanomedicines in cancer treatment.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-025-03635-0.

Introduction

Nanomedicines hold significant promise in enhancing cancer chemotherapy due to the enhanced permeability and retention (EPR) effects [1, 2]. However, their clinical benefits have not yet shown substantial superiority over traditional treatments [3, 4]. A key challenge is the inadequate tumor targeting of chemotherapeutic agents [5], leading to suboptimal therapeutic outcomes and considerable side effects [6, 7]. Additionally, responses to these nanomedicines vary widely among individuals, with patient-specific factors such as gender [8], age [9] and body mass index (BMI) [10] influencing drug effectiveness. Thus, enhancing the precision of tumor targeting and elucidating the mechanisms behind individual responses to nanomedicines are crucial for boosting the efficacy of chemotherapeutic agents.

Nanomedicines, once in the bloodstream, adsorb proteins from the surrounding media to form protein corona through electrostatic, van der Waals and hydrophobic interactions [11, 12]. This protein corona can alter the physicochemical properties of nanomedicines, thereby affecting their pharmacokinetics, biodistribution and therapeutic efficacy [13, 14]. For example, the adsorption of opsonins such as IgG can promote phagocytosis by the reticuloendothelial system (RES) and the eventual removal of nanomedicines from systemic circulation, while the adsorption of functional proteins such as albumin can enable active tumor targeting [15]. The composition of plasma proteins in patients with different pathological status might be different, leading to significant alterations in the protein corona composition on the nanoparticle surface (named as personalized protein corona) and generating different therapeutic efficacy [16, 17]. Therefore, an in-depth understanding of personalized protein corona can elucidate the mechanism on different therapeutic efficacy of the same nanomedicines and serve as a potential tool for advancing personalized nanomedicines or overcoming limitations in the clinical translation of nanoparticle-based therapies.

Obesity, characterized by a BMI of 30 kg m⁻² or higher, is a global epidemic and a significant risk factor for at least 13 types of cancer, including ovarian cancer [18, 19]. It often triggers inflammation and elevates levels of leptin and insulin-like growth factor-1 (IGF-1), which can drive tumor development and progression [20, 21]. Interestingly, despite its role in promoting tumorigenesis, recent clinical analyses have revealed a paradoxical association between obesity and improved cancer therapeutic outcomes [22, 23]. Obese patients tend to show better responses to chemotherapy and immune checkpoint blockade therapy [24, 25]. For example, retrospective analyses of ovarian cancer patient data have revealed that obese patients have higher progression-free survival rates after treatment with chemotherapeutic agents such as PLD, paclitaxel, and carboplatin [24]. However, the mechanisms underlying this enhanced therapeutic efficacy in obese patients remain to be fully understood.

Ovarian cancer is one of the most lethal cancers in the female reproductive system, with chemotherapy being a primary treatment [26]. However, unsatisfactory treatment efficacy and chemotherapy-related side effects significantly impact patients’ quality of life [27]. While emerging chemotherapy-related drugs, including chemokinetic therapeutic agents [28], immunotherapeutic agents [29, 30], and aptamer drugs [31], show promise, their clinical translation remains challenging. Here, we demonstrate that PLD exhibits stronger anticancer activity in obese mice with ovarian tumors compared to normal mice. PLD adsorbs complement component 1q (C1q), a key component of the classical complement activation pathway [32, 33], from the serum of obese mice, which facilitates its uptake by tumor cells expressing high levels of C1q receptors [34], thereby efficiently killing tumor cells. Furthermore, C1q adsorbed on PLD promotes the phagocytosis of released tumor antigens by dendritic cells (DCs) via DC-SIGN receptor [35], activating DC maturation and T cell-mediated antitumor immune responses and significantly enhancing the anticancer efficacy of PLD in obese mice (Scheme 1). This work highlights a potential mechanism for leveraging personalized protein corona to enhance PLD efficacy and offering a more realistic and clinically translatable approach compared to other nanomedicines.

Scheme 1.

Schematic representation of the enhanced therapeutic sensitivity of PLD in obese mice with ovarian tumors. C1q highly present in plasma of obese mice adsorbs to PLD, leading to increased tumor accumulation and potentiating the ICD response in tumor cells. Concurrently, C1q engagement is essential for facilitating the phagocytosis of apoptotic tumor cells by DCs through DC-SIGN receptor, which activates tumor-specific CD8⁺ T cells and significantly improves the anticancer efficacy of PLD in DIO mice

Results

Obesity enhances the therapeutic efficacy of PLD and improves tumor immune microenvironment in ovarian cancer

To assess the impact of obesity on chemotherapeutic efficacy, a diet-induced obesity (DIO) model was first developed by administering a high-fat diet (60% fat diet) to female C57BL/6 mice [36]. Concurrently, a control group was maintained on a normal-fat diet (10% fat diet) over the same period. After 13 weeks, the DIO mice gained more than 20% of their initial body weight (Fig. S1A and B) and exhibited a significant increase in Lee’s index, an indicator of obesity calculated as weight^1/3/Naso-Anal length [37], compared to the normal mice (Fig. S1C). Additionally, the DIO mice had significantly more gonadal and inguinal fat pads than the control group (Fig. S1D), with a 8.6-fold increase in gonadal fat (Fig. S1E) and a 11.4-fold increase in inguinal fat (Fig. S1F). Blood biochemistry analyses also revealed that the DIO group had substantially higher levels of glucose (Fig. S1G), cholesterol (Fig. S1H), high-density lipoprotein cholesterol (HDL-C, Fig. S1I) and low-density lipoprotein cholesterol (LDL-C, Fig. S1J) compared to normal mice. These findings collectively verify the successful creation of the DIO model.

Clinical data have indicated that ovarian cancer patients with a BMI over 30 exhibit improved survival rates following PLD [24]. To determine if the obese mice showed a similar response, ID8 ovarian cancer subcutaneous tumor models were established in both normal and DIO mice. These mice were treated with PLD via tail vein injection. The particle size, zeta potential and drug loading of PLD were 81.0 ± 0.1 nm (Fig. S2A), − 1.1 ± 0.1 mV and 11.1%, respectively. Transmission electron microscopic (TEM) images revealed that PLD has a homogeneous spherical structure (Fig. S2B). Although obesity is generally associated with increased ovarian cancer growth, the DIO mice demonstrated a more effective tumor suppression effect after PLD treatment (Fig. 1A). Specifically, PLD treatment led to a 74% reduction in tumor weight in obese mice, a significantly greater decrease compared to the normal mice group (Fig. 1B). Additionally, the survival period for the PLD-treated obese mice was extended to 44 days, which was notably longer than the 34 days observed in the PLD-treated normal mice group (Fig. 1C). These data suggest that obesity might enhance the effectiveness of PLD treatment in ovarian cancer. Specifically, PLD treatment was found to induce cardiotoxicity in the normal group of mice, as demonstrated by hematoxylin and eosin (H&E) staining of key organs (Fig. S3), serological analysis (Fig. S4) and reduced body weight (Fig. S5). In contrast, obesity appeared to reduce the systemic toxicity associated with PLD (Fig. S3-S5), confirming a protective role of obesity against the adverse effects of chemotherapy [38].

Fig. 1.

Effects of obesity on the anticancer activity of PLD in ID8 tumor-bearing mice. A Tumor growth curves of ID8 tumor-bearing normal and DIO mice after intravenous injection of PBS or PLD at the DOX dosage of 4 mg kg⁻¹ once every other day for 3 times. Data are presented as means ± SD (n = 8). B Tumor weights at day 13 of the treatment indicated in (A). Data are presented as means ± SD (n = 6). C Kaplan–Meier survival plot of ID8 tumor-bearing normal and DIO mice after treatment indicated in (A) (n = 8). D, E In vivo CellVue^® Claret fluorescence images (D) and intensities (E) of ID8 tumor-bearing normal and DIO mice at different time intervals after intravenous injection of CellVue^® Claret-labeled PLD at the DOX dosage of 80 µg per mouse. Data are presented as means ± SD (n = 3). F DOX contents in tumor tissues of ID8 tumor-bearing normal or DIO mice at 24 h after intravenous injection of PLD at the DOX dosage of 80 µg per mouse. Data are presented as means ± SD (n = 5). G-M Numbers of CD11c⁺CD80⁺CD86⁺ (G), CD8⁺ T (H), CD8⁺CD69⁺ T (I), CD8⁺Ki67⁺ T (J), CD8⁺IL-2⁺ T (K), CD8⁺IFN-γ⁺ T (L) and CD8⁺GzmB⁺ T (M) cells in tumor tissues of ID8 tumor-bearing mice at 13 days after treatments indicated in (A). Data are presented as means ± SD (n = 6 mice in each group). *P < 0.05, **P < 0.01, ***P < 0.001

To determine the potential mechanism by which obesity enhanced the therapeutic efficacy of PLD in ovarian cancer, we first assessed the in vivo biodistribution of PLD in normal and DIO mice bearing ID8 ovarian tumors by intravenously injecting CellVue^® Claret-labeled PLD. Whole-animal fluorescence imaging analyses demonstrated a significant increase in CellVue^® Claret fluorescence intensity in tumor tissues of the DIO mice compared to normal mice (Fig. 1D and E). Whole-animal fluorescence imaging analyses showed a significant increase in CellVue^® Claret fluorescence intensity in tumor tissues of DIO mice compared to normal mice (Fig. 1D and E). Ex vivo imaging of tumor tissues 24 h after intravenous injection further revealed the enhanced accumulation of PLD in tumors of DIO mice (Fig. S6A and B). To further confirm the enhanced tumor accumulation of PLD in obese mice, PLD at a DOX dosage of 80 µg per mouse were intravenously injected. 24 h after injection, the DOX content in major organs, including hearts, livers, spleens, lungs, kidneys and tumor tissues, was quantified using high-performance liquid chromatography (HPLC). PLD accumulation in tumor tissues of DIO mice was approximately 3.5-fold higher than in normal mice, although no significant difference was detected in other tissues (Fig. 1E), suggesting that obesity may facilitate the delivery of PLD to ovarian tumors. Additionally, the tumor immune microenvironment was assessed in normal and DIO mice bearing ID8 tumors following intravenous injection of PLD bi-daily for a total of 3 times. Consistently, PLD treatment significantly increased the number of mature DCs, identified as CD11c⁺CD80⁺CD86⁺, in tumors of DIO mice compared with that of normal mice (Fig. 1G), revealing that PLD efficiently promoted intratumoral DC maturation in DIO mice. Moreover, the DIO group exhibited a significant upregulation in the numbers of CD8⁺ T cells (Fig. 1H), as well as functional CD8⁺CD69⁺ T cells (Fig. 1I) and proliferating CD8⁺Ki67⁺ T cells (Fig. 1J) in tumors after PLD treatment. Meanwhile, the highest percentages of CD8⁺ T cells expressing key effector molecules, including interleukin (IL)−2 (Fig. 1K) which promotes T cell proliferation and differentiation, interferon (IFN)-γ (Fig. 1L) which enhances antigen presentation and T cell activation; and granzyme B (GzmB) (Fig. 1M) which directly induces tumor cell apoptosis, were observed in PLD-treated DIO mice [39]. Altogether, these results indicated that obesity can alter the tumor immune microenvironment in response to chemotherapy, leading to an enhanced therapeutic response to chemotherapy.

Protein corona may be the key factor in regulating PLD efficacy in DIO mice

Protein corona significantly influences the in vivo transport processing of nanoparticles, thereby affecting their biological function. Given that PLD had different anticancer effects when administered to normal and DIO mice, it promotes our hypothesis that distinct protein coronas formed around PLD might modulate its in vivo behavior. To test this, we generated PLD with adsorbed protein coronas by incubating PLD with plasma from normal or DIO mice in vitro (denoted as PLD@NP and PLD@DP, respectively). Dynamic light scattering (DLS) analysis revealed that both PLD@NP and PLD@DP exhibited a significant increase in diameter (Fig. 2A) and a shift in zeta potential from nearly neutral to negatively charged (Fig. 2B) compared to PLD alone, indicating successful adsorption of plasma proteins onto PLD, a phenomenon also confirmed in other studies [40, 41]. However, the diameter and zeta potential of PLD@NP and PLD@DP were comparable (Fig. 2A and B). Although PLD@NP and PLD@DP adsorbed the same amount of plasma proteins, the specific plasma protein species adsorbed onto their surfaces were significantly different, as demonstrated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 2C). Additionally, we recovered and isolated PLD directly from the plasma of normal and DIO mice that had undergone intravenous PLD administration to obtain PLD@NP and PLD@DP. Consistently, PLD@NP and PLD@DP obtained in vivo showed similar diameter (Fig. S7A), zeta potential (Fig. S7B) and protein contents (Fig. S7C), but displayed different protein profiles (Fig. S7D). These findings indicated that the protein corona might be involved in regulating the efficacy of chemotherapy and the tumor microenvironment in DIO mice.

Fig. 2.

In vitro antitumor immunity induced by PLD@DP. A, B The diameter (A) and zeta potential (B) of PLD after preincubation with normal or DIO mice plasma. Data are presented as means ± SD (n = 3). C Representative silver-stained SDS-PAGE of PLD@NP and PLD@DP. D DOX mean fluorescence intensity (MFI) in ID8 cells after treatment with PLD@NP or PLD@DP at the DOX concentration of 10 µg mL⁻¹ at 37 °C for 4 h by flow cytometry. Data are presented as means ± SD (n = 3). E Cell viability of ID8 cells at 24 h after treatment with PLD@NP or PLD@DP at DOX concentration of 10 µg mL⁻¹. Data are presented as means ± SD (n = 5). F CRT MFI in ID8 cells at 12 h after treatments indicated in (E) by flow cytometry. Data are presented as means ± SD (n = 4). (G, H) ATP (G) and HMGB1 release (H) from ID8 cells at 12 h after treatments indicated in (E). Data are presented as means ± SD (n = 4). (I) Percentages of CD80⁺CD86⁺ cells in CD11c⁺ BMDCs at 12 h after immature BMDCs were co-cultured with ID8 cells treated as indicated in (E) by flow cytometry. Data are presented as means ± SD (n = 4). (J-L) Percentages of IL-2⁺ (J), GzmB⁺ (K) and IFN-γ⁺ (L) cells in CD8⁺ T cells after CD3⁺ T cells were incubated with the matured BMDCs treated as indicated in (I) for 3 days by flow cytometry. Data are presented as means ± SD (n = 4). *P < 0.05, **P < 0.01, ***P < 0.001

To ascertain the role of protein corona in the modulation of antitumor efficacy of PLD, ID8 cells were first treated with PLD@NP or PLD@DP at 37 °C for 4 h, after which intracellular DOX fluorescence was measured using flow cytometry. Expectedly, PLD@DP demonstrated greater internalization into ID8 cells than PLD@NP (Fig. 2D and Fig. S8). Accordingly, PLD@DP exhibited stronger cytotoxicity against ID8 cells compared to PLD@NP (Fig. 2E). Consistently, PLD@DP treatment resulted in enhanced exposure of calreticulin (CRT) (Fig. 2F) and increased release of adenosine triphosphate (ATP) (Fig. 2G) and high mobility group box 1 (HMGB1) (Fig. 2H) in ID8 cells compared to PLD@NP, suggesting that PLD@DP induced a more potent immunogenic cell death (ICD) effects in ID8 cells. When ID8 cells treated with PLD@NP or PLD@DP were co-cultured with bone marrow-derived dendritic cells (BMDCs), the PLD@DP-treated group exhibited a higher level of BMDC maturation compared to the PLD@NP-treated group (Fig. 2I). Additionally, flow cytometric analysis demonstrated that when naïve T cells were co-incubated with BMDCs treated with PLD@DP, there was a significant increase in the percentage of CD8⁺ T cells producing effector molecules IL-2 (Fig. 2J), GzmB (Fig. 2K) and IFN-γ (Fig. 2L), indicating that BMDCs matured by tumor cells treated with PLD@DP have a stronger capacity to activate CD8⁺ T cells. However, we noticed that PLD@NP and PLD@DP did not remarkably affect cell viability of BMDCs (Fig. S9) at the used DOX concentration, suggesting that PLD@NP and PLD@DP do not exert cytotoxicity against BMDCs. These data collectively indicated that the protein corona, particularly that derived from DIO mice, significantly influenced the antitumor activity of PLD by enhancing its internalization into tumor cells and promoting a more effective antitumor immune response.

C1q is adsorbed onto PLD in DIO mice

To further investigate the specific proteins influencing the in vivo behavior of PLD, we analyzed the protein components of PLD@NP and PLD@DP using liquid chromatography-mass spectrometry (LC-MS) and referenced the Protein Data Bank. Identification of all proteins revealed that PLD@NP and PLD@DP mainly recruited proteins with smaller molecular weights (10–60 kDa) (Fig. 3A). A volcano plot was used to depict these differences (Fig. 3B), and a Venn diagram highlighted the distinctions in adsorbed proteins between PLD@NP and PLD@DP. With a threshold of fold change > 1.2 and P ≤ 0.05, a total of 102 differentially expressed proteins were detected (Fig. 3C). Gene ontology (GO) functional enrichment analysis revealed that the differential proteins adsorbed on PLD@DP were predominantly associated with human diseases and organismal systems pathways, particularly those related to the immune system (Fig. 3D). Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis of the top twenty differential protein pathways showed that the “complement and coagulation cascades” pathway had the highest number of differential proteins, indicating that PLD@DP adsorbed more complement system-related proteins than PLD@NP (Fig. 3E). Among the complement proteins, C1q exhibited the most significant change, with a 2.4-fold increase in adsorption on PLD@DP compared to PLD@NP (Fig. 3F). Consistently, ELISA assay confirmed higher C1q levels in the plasma of DIO mice than in normal mice (Fig. S10). At 100 µg DOX, PLD@DP adsorbed 236 µg of C1q protein, while PLD@NP adsorbed 189 µg of C1q protein (Fig. 3G), confirming a significant increase in C1q content on PLD@DP compared to PLD@NP. These findings revealed that PLD@DP-adsorbed proteins, especially C1q, might modulate the biological activity of PLD in DIO mice.

Fig. 3.

Proteomic analysis of protein components on PLD@NP and PLD@DP. A Classification of proteins of PLD@NP and PLD@DP according to molecular weight. B Volcano plot for identifying differentially expressed proteins on PLD@NP and PLD@DP. The P values were adjusted using the Benjamini and Hochberg’s approach for controlling the false discovery rate. P ≤ 0.05 and Foldchange > 1.2 were set as the thresholds for significant differential expression. (C) Number of differentially expressed proteins adsorbed by PLD@NP and PLD@DP. D GO enrichment analysis of immune-related biological processes among differentially expressed proteins between PLD@NP and PLD@DP. E KEGG pathway analysis of differentially expressed proteins between PLD@NP and PLD@DP, with the top 20 terms were selected. Statistical significance was assessed using a two-sided hypergeometric test, with P values adjusted for multiple comparisons using the Benjamini-Hochberg method. F Complement proteins with significant quantitative differences between PLD@NP and PLD@DP. G C1q content on PLD@NP and PLD@DP by ELISA assay. Data are presented as means ± SD (n = 3). *P < 0.05

C1q is involved in the enhanced tumor cell targeting and improved antitumor immunity of PLD in DIO mice

C1q, a hexameric molecule consisting of a spherical head (gC1q) and a collagenous tail (cC1q), binds to two distinct receptors, gC1qR and CRT (Fig. 4A), respectively [34]. gC1qR is highly expressed on tumor cells [34]. To determine if C1q on PLD enhanced tumor targeting by binding to gC1qR on tumor cells of DIO mice, we created ID8 cells with stable gC1qR knockdown (ID8^KO cells) using the CRISPR/Cas9 system (Fig. 4B and Fig. S11). DIO mice were implanted with ID8 or ID8^KO cells subcutaneously and then treated with PLD via tail vein injection. Tumor accumulation of PLD was assessed 24 h post-injection. As expected, knocking down gC1qR significantly reduced PLD accumulation in tumors of DIO mice (Fig. 4C), highlighting C1q’s role in obesity-mediated tumor targeting of PLD.

Fig. 4.

Role of C1q in enhancing tumor cell targeting and improving ICD effects induced by PLD. A Structure of C1q. C1q consists of a spherical head (gC1q) and a collagenous tail (cC1q) that binds to two different receptors (gC1qR and CRT). B gC1qR expression in ID8, ID8^vector (ID8 cells stably expressing empty vector) and ID8^KO cells by western blotting analysis. C DOX contents in tumor tissues of ID8 or ID8^KO tumor-bearing DIO mice at 24 h after intravenous injection of PLD at the DOX dosage of 80 µg per mouse. Data are presented as means ± SD (n = 4). D DOX MFI in ID8 or ID8^KO cells after treatment with PLD@NP or PLD@DP at 37 °C for 4 h by flow cytometry. Data are presented as means ± SD (n = 3). E Cell viability of ID8 or ID8^KO cells at 24 h after treatment with PLD@NP or PLD@DP. Data are presented as means ± SD (n = 5). F CRT MFI in ID8 or ID8^KO cells at 12 h after treatments indicated in (E) by flow cytometry. Data are presented as means ± SD (n = 4). G, H ATP (G) and HMGB1 release (H) from ID8 or ID8^KO cells at 12 h after treatments indicated in (E). Data are presented as means ± SD (n = 4). I-K Intracellular DiO MFI (I), ratio of DiO⁺ cells (J) and percentages of CD80⁺CD86⁺ cells (K) in CD11c⁺ BMDCs at 12 h after the immature BMDCs were co-cultured with DiO-labeled ID8 or ID8^KO cells pretreated as indicated in (E) by flow cytometry. Data are presented as means ± SD (n = 4). L-P Percentages of CD8⁺ T cells in CD3⁺ T cells (L) and percentages of CD69⁺ (M), IL-2⁺ (N), GzmB⁺ (O) and IFN-γ⁺ (P) cells in CD8⁺ T cells after CD3⁺ T cells were incubated with the matured BMDCs treated as indicated in (K) for 3 days by flow cytometry. Data are presented as means ± SD (n = 4). **P < 0.01, ***P < 0.001

To verify this, ID8 and ID8^KO cells were treated with PLD@NP or PLD@DP, and then intracellular DOX content was determined. Consistently, more PLD@DP was internalized into ID8 cells than PLD@NP. The knockdown of gC1qR significantly decreased the internalization of PLD@DP, but not PLD@NP, into ID8 cells (Fig. 4D and Fig. S12), indicating that C1q facilitated the uptake of PLD@DP by ID8 cells. Correspondingly, the knockdown of gC1qR also significantly mitigated the cytotoxicity of PLD@DP against ID8 cells (Fig. 4E). Moreover, ID8^KO cells exhibited reduced CRT exposure (Fig. 4F) and decreased release of ATP (Fig. 4G) and HMGB1 (Fig. 4H) following PLD@DP treatment compared to ID8 cells, while PLD@NP treatment resulted in similar CRT exposure and HMGB1 and ATP release in both ID8 and ID8^KO cells (Fig. 4F-H), suggesting that gC1qR deletion specifically diminished ICD effects induced by PLD@DP. Furthermore, fewer PLD@DP-treated ID8^KO cells were phagocytosed by BMDCs (Fig. 4I and J), resulting in reduced maturation of BMDCs induced by PLD@DP (Fig. 4K). Additionally, when naïve T cells were co-cultured with these BMDCs, those matured from PLD@DP-treated ID8 cells showed significantly higher numbers of CD8⁺ T (Fig. 4L), CD8⁺CD69⁺ T (Fig. 4M), CD8⁺IL-2⁺ T (Fig. 4N), CD8⁺GzmB⁺ T (Fig. 4O) and CD8⁺IFN-γ⁺ T cells (Fig. 4P) than those from PLD@DP-treated ID8^KO cells. Collectively, these results revealed that C1q significantly influenced the targeting of PLD to tumor cells, potentially modulating the antitumor immune response to exert anticancer effects induced by PLD. To further substantiate the role of C1q in regulating tumor cell targeting of PLD, we used adiponectin (ADPN), a protein containing a C1q globular head structure that can compete with C1q [42]. ID8 cells were pretreated with PBS or ADPN, followed by treatment with PLD@NP or PLD@DP. Expectedly, ADPN pretreatment significantly decreased cellular uptake of PLD@DP (Fig. S13A and Fig. S14) and inhibited its cytotoxicity against ID8 cells (Fig. S13B). Additionally, ADPN pretreatment efficiently abolished PLD@DP-triggered ICD effects in ID8 cells (Fig. S13C-S13E) and the subsequent T cell-mediated antitumor immune response (Fig. S13F-S13I). In summary, the findings collectively highlight that C1q is a key regulator in directing tumor cell targeting and shaping the immune response elicited by PLD.

C1q contributes to the increased phagocytosis of PLD-triggered apoptotic tumor cells to activate antitumor immunity

A hallmark of ICD is the CRT exposure on the surface of tumor cells, which serves as a receptor for cC1q [43]. Immature DCs express DC-SIGN on their surface, which binds gC1q [35]. When C1q binds to apoptotic tumor cells, C1q increases DC uptake of apoptotic tumor cells and promotes T cell-mediated anti-tumor immunity by binding to DC-SIGN [35, 44]. To investigate the role of C1q in the phagocytosis of PLD-induced apoptotic cells by DCs, DiO-labeled PLD-triggered apoptotic ID8 cells were pretreated with PBS or ani-C1q antibody (denoted as αC1q), which specifically binds to the cC1q segment of C1q to inhibit the interaction of C1q and CRT. Subsequently, these cells were treated with PLD@NP or PLD@DP and then co-incubating with BMDCs. As anticipated, BMDCs showed increased phagocytosis of DiO-labeled apoptotic ID8 cells following exposure to PLD@DP compared to PLD@NP (Fig. 5A and B). Pretreatment with αC1q significantly diminished the enhanced phagocytosis of apoptotic ID8 cells by BMDCs induced by PLD@DP (Fig. 5A and B), while no significant changes were observed in PLD@DP-treated group. These findings suggested that the presence of C1q on PLD@DP was essential for the enhanced phagocytosis of apoptotic tumor cells by DCs. Furthermore, αC1q significantly reduced the maturation of BMDCs enhanced by PLD@DP (Fig. 5C). When these BMDCs were co-cultured with splenic T cells, the PLD@DP-treated group exhibited higher frequencies of IL-2⁺ (Fig. 5D), GzmB⁺ (Fig. 5E) and IFN-γ⁺ cells (Fig. 5F) within CD8⁺ T cells compared to PLD@NP-treated group. In contrast, αC1q significantly decreased PLD@DP-triggered CD8⁺ T cell activation (Fig. 5D-F). Collectively, these results indicated that C1q on PLD@DP contributed to the phagocytosis of PLD-induced apoptotic tumor cells by DCs, leading to increased DC maturation and CD8⁺ T cell activation.

Fig. 5.

Role of C1q on PLD@DP in increasing phagocytosis of apoptotic cells by DCs. A, B Intracellular DiO MFI (A) and ratio of DiO⁺ cells (B) in CD11c⁺ BMDCs at 12 h after the immature BMDCs were co-cultured with DiO-labeled ID8 cells which were pretreated with PBS or αC1q (2 µg mL⁻¹)for 1 h and then treated with PLD@NP or PLD@DP at the DOX concentration of 10 µg mL⁻¹ for 24 h by flow cytometry. Data are presented as means ± SD (n = 4). C Percentages of CD80⁺CD86⁺ cells in CD11c⁺ BMDCs at 12 h after immature BMDCs were co-cultured with ID8 cells treated as indicated in (A) by flow cytometry. Data are presented as means ± SD (n = 4). D-F Percentages of IL-2⁺ (D), GzmB⁺ (E) and IFN-γ⁺ (F) cells in CD8⁺ T cells after CD3⁺ T cells were incubated with the matured BMDCs treated as indicated in (C) for 3 days by flow cytometry. Data are presented as means ± SD (n = 4). ***P < 0.001

C1q significantly enhances the antitumor effects of PLD in DIO mice

To delve deeper into how C1q influences the antitumor effects and immune reactions triggered by PLD in obese mice, DIO mice that had either subcutaneous ID8 tumors or ID8^KO tumors were were subjected to treatments with PBS or PLD in the presence or absence of αC1q (Fig. 6A). Consistently, PLD significantly slowed the growth of ID8 tumors compared to PBS (Fig. 6B), resulting in a higher tumor suppression rate (Fig. 6C), reduced tumor weight (Fig. 6D) and extended survival time (Fig. 6E). Both knocking down gC1qR in ID8 tumors and treatment with αC1q significantly reduced the antitumor effects of PLD (Fig. 6B-E), suggesting that the regulation of PLD’s antitumor efficacy was significantly influenced by C1q in DIO mice. Moreover, knocking down gC1qR in ID8 tumors or treatment with αC1q markedly reduced PLD-induced increase in the numbers of CD11c⁺CD80⁺CD86⁺ DCs (Fig. 6F), CD8⁺ T cells (Fig. 6G), proliferating CD8⁺Ki67⁺ T cells (Fig. 6H), and activated CD8⁺CD69⁺ T cells (Fig. 6I), as well as CD8⁺IL-2⁺ T (Fig. 6J), CD8⁺IFN-γ⁺ T (Fig. 6K) and CD8⁺GzmB⁺ T cells (Fig. 6L), indicating that C1q participated in the antitumor immune response triggered by PLD in DIO mice. Notably, treatment with αC1q further decreased PLD-induced antitumor activity and antitumor immunity in DIO mice bearing ID8^KO tumors (Fig. 6B-L), indicating that PLD-triggered enhanced antitumor activity and improved antitumor immunity were regulated by C1q, which contributed to increased tumor accumulation of PLD and promoted the phagocytosis of PLD-induced apoptotic tumor cells by DCs for improved antitumor immunity.

Fig. 6.

PLD-induced anticancer activity and enhanced antitumor immunity in a C1q-dependent manner in DIO mice. A Schematic timeline for anticancer activity and tumor immune microenvironment analysis in ID8 or ID8^KO tumor-bearing DIO mice after intravenous administration of three doses of PLD at the DOX dosage of 4 mg kg⁻¹ once every other day with or without of intratumoral administration of αC1q antibody at the dosage of 150 µg 24 h after PLD injection. B Tumor volume curves of ID8 or ID8^KO tumor-bearing DIO mice after treatments indicated in (A), respectively. Data are presented as means ± SD (n = 7). C, D Tumor suppression rate (C) and tumor weight (D) at day 13 after treatments indicated in (A). Data are presented as means ± SD (n = 7). (E) Kaplan–Meier survival plots of ID8 or ID8^KO tumor-bearing DIO mice after treatments indicated in (A). Data are presented as means ± SD (n = 8). (F-L) Numbers of CD11c⁺CD80⁺CD86⁺ (F), CD8⁺ T (G), CD8⁺Ki67⁺ T (H), CD8⁺CD69⁺ T (I), CD8⁺IL-2⁺ T (J), CD8⁺IFN-γ⁺ T (K) and CD8⁺GzmB⁺ T (L) cells in tumor tissues of ID8 or ID8^KO tumor-bearing DIO mice at 13 days after treatments indicated in (A). Data are presented as means ± SD (n = 7). *P < 0.05, **P < 0.01, ***P < 0.001

Discussion

The relationship between obesity and cancer outcomes has been a subject of intense debate, with recent clinical data suggesting a paradoxical association between obesity and improved cancer therapeutic outcomes. Ovarian cancer patients with higher BMIs have shown improved progression-free survival after chemotherapy treatments, including PLD [24]. Similar trends have been noted in diffuse large B-cell lymphoma [22] and melanoma patients [10] undergoing chemotherapy or immunotherapy. As obesity rates continue to rise globally, it is imperative to investigate the mechanisms by which obesity influences chemotherapy and to develop targeted therapies for cancer patients.

The formation of the protein corona significantly influences the properties of nanomedicines. Strategies to design and manipulate the protein corona can be categorized into two main approaches. The first involves minimizing protein corona formation by employing the “stealth effect,” where a hydrophilic polymer such as PEG is used to coat nanomedicines, reducing non-specific interactions with blood proteins and enhancing circulation times [45, 46]. The second strategy involves intentionally adsorbing specific plasma proteins to enhance the targeting efficiency of nanomedicines through surface engineering [47, 48]. However, the range of adsorbed proteins is currently limited, with few studies exploring the roles of proteins beyond the apolipoprotein family and albumin in the in vivo behavior of nanomedicines. In this study, we uncovered that the chronic inflammatory environment associated with obesity led to the upregulation of C1q protein in mouse plasma, which adsorbed to PLD with high affinity. The interaction between C1q and gC1qR on tumor cells not only conferred tumor-targeting capabilities to PLD but also amplified the ICD effects, leading to significant tumor suppression in DIO mice.

The ICD phenomenon allows DCs to recognize and eliminate apoptotic tumor cells through “eat me” signals such as CRT, thereby promoting T cell-mediated antitumor immunity [49]. However, a delay in the quick removal of apoptotic cells can result in their disintegration and the release of immunogenic molecules like TNF-α into the surrounding tissue, triggering a pro-inflammatory response and potentially leading to autoimmunity [50, 51]. Previous works have indicated that certain collagen family members, including the complement protein C1q, serve as co-stimulatory molecules that expedite the clearance of apoptotic cells by DCs, preventing the release of inflammatory factors [52]. Additionally, Andrea et al. demonstrated that C1q bound to apoptotic cells inhibited the proliferation of Th17 and Th1 T cell subsets mediated by macrophages and DCs, thereby remodeling the adaptive immune system [53]. This highlights the dual role of C1q in promoting the clearance of apoptotic cells by phagocytes and in regulating inflammatory responses within the autophagic system. In this work, we discovered that C1q adsorbed onto PLD not only intensified the ICD effects but also engaged in the phagocytosis of apoptotic tumor cells by DCs, which promoted antitumor immunity to exert strong anticancer efficacy in DIO mice. However, the exact mechanism by which C1q on PLD promoted DC phagocytosis of apoptotic tumor cells required further clarification. In addition, C1q serves a critical role in the classical complement pathway, recognizing and contributing to the lysis of tumor cells [54]. The extent to which C1q influenced complement-dependent cytotoxic effects in the anticancer activity of PLD in DIO mice warranted additional investigation.

Conclusion

In summary, our study reveals that PLD in obese mice adsorbs a higher quantity of C1q protein, which effectively targets tumor cells and induces greater ICD effects. The binding of C1q to apoptotic tumor cells facilitates DC phagocytosis, enhancing the antigen-presenting capacity of DCs and activating CD8⁺ T cells. This C1q-mediated process significantly promotes the efficacy of PLD in cancer therapy in DIO mice. Collectively, this strategy highlights the importance of the personalized protein corona in regulating the therapeutic efficacy of nanomedicines, offering innovative avenues for the personalized treatment of cancer patients.

Methods and experimental

Materials

RPMI 1640 and DMEM medium were purchased from HyClone (GE Healthcare, South Logan, USA). Phosphate-buffered saline (PBS) were purchased from Servicebio (Wuhan, China). Trypsin EDTA, fetal bovine serum (FBS), penicillin and streptomycin were provided by Vazyme Biotech Co., Ltd (Wuhan, China). Recombinant Adiponectin (ADPN) protein (RP01461) was purchased from ABclonal Biotechnology Co., Ltd. (Wuhan, China). Anti-C1q antibody (JL-1, ab71940) was purchased from Abcam Company (Cambridge, UK). C1q ELISA kit was purchased from Fengbin Technology Co., Ltd. (Wuhan, China). CellVue^® Claret kit was purchased from Merck (Darmstadt, German). 10 kcal% normal-fat diet (AIN-93G) and 60 kcal% high-fat diet (D12492i) were purchased from Jiangsu Xietong, Inc. (Nanjing, China) and Shanghai Renbang Pharmaceutical Technology Co., Ltd., respectively. PLD was obtained from CSPC Ouyi Pharmaceutical CO., Ltd. (Shijiazhuang, China). The size and zeta potential of PLD was determined by DLS analysis. The morphology of PLD was observed by TEM (Tecnai G2 20; FEI) with an accelerating voltage of 60 kV.

Cell culture

ID8 cells were obtained from Shanghai Hongshun Biological Technology Co. Ltd (Shanghai, China). The cells were maintained in DMEM medium enriched with 10% FBS and 100 µg mL⁻¹ penicillin-streptomycin. The culture was incubated at 37 °C in an environment with 5% CO₂. BMDCs were extracted from femurs and tibiae of female C57BL/6 mice aged 8 weeks and maintained in RPMI 1640 medium enriched with 10% FBS, 100 µg mL⁻¹ penicillin-streptomycin, 10 ng mL⁻¹ recombinant mouse granulocyte-macrophage colony-stimulating factor (GM-CSF, PeproTech, Rocky Hill, USA) and 10 ng mL⁻¹ IL-4 (PeproTech). CD3⁺ T cells were purified from the splenocytes of female C57BL/6 mic aged 8 weeks using a MokoSort™ Mouse CD3⁺ T cell isolation kit (Biolegend, San Diego, USA) and expanded in RPMI 1640 medium containing 10% FBS, 100 µg mL⁻¹ penicillin-streptomycin and 20 ng mL⁻¹ IL-2 (PeproTech) for 5 days. Regular mycoplasma testing was conducted on all cell cultures using the MycAway-Color detection kit (Yeasen Biotechnology Co., Ltd, Shanghai, China) to confirm their mycoplasma-free status.

Animals

Four-week-old female C57BL/6 mice (12 ± 2 g) were procured from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). These mice were used to establish a subcutaneous ID8 ovarian tumor model by inoculating 5 × 10⁶ ID8 cells into the left lower flank region. All animal studies conducted complied with ethical standards for animal research, and were approved by the Institutional Animal Care and Use Committee at Tongji Medical College, Huazhong University of Science and Technology (Wuhan, China).

Construction of obesity models

week-old female C57BL/6 were randomly separated into normal and DIO group. The normal group and DIO group were constructed in the same living conditions except that normal mice were maintained on a normal-fat diet, while DIO mice were provided with a high-fat diet. The mouse weight was detected weekly using an electronic balance. After 13 weeks of continuous feeding, the blood was collected by removing the eyeballs and then centrifuged at 3,000g for 10 min to obtain mouse serum. The contents of glucose, cholesterol, HDL-C and LDL-C in the mouse serum were measured with automatic biochemical analyzer (Beckman, Pasadena, USA). After the mice were executed, the body length was evaluated and their gonads and inguinal fat were dissected. The Lee’s index was calculated according to the following formulation

Characterization of PLD@NP and PLD@D

PLD at the DOX concentration of 2 mg mL⁻¹ (100 µL) was incubated with 100 µL normal mouse plasma or DIO mouse plasma for 1 h at 37 °C on a homogeneous shaker. The mixtures were centrifuged at 100,000g for 30 min to pellet PLD@NP and PLD@DP, followed by washing with cold PBS. Furthermore, 100 µL of PLD was intravenously injected into normal or DIO mice at the DOX dosage of 80 µg per mouse. After 2 h, mouse plasma was collected by removing mouse eyeballs, and then centrifuged at 500g for 5 min and 100,000g for 30 min to isolate PLD@NP and PLD@DP, which were rinsed three times with PBS. The diameter and zeta potential of PLD@NP and PLD@DP were measured employing a Malvern Nano-ZS Particle Sizer (Malvern Instruments Ltd., Malvern, UK). The DOX and protein contents in PLD@NP and PLD@DP were quantified by HPLC and the BCA protein assay kit (Beyotime, Shanghai, China), respectively. The protein profiles of PLD@NP and PLD@DP were examined through SDS-PAGE and further chromatographed by using the rapid silver stain kit (Beyotime, Shanghai, China).

Proteomics analysis

Proteins from PLD@NP and PLD@DP were extracted and assessed for quality control using the Bradford quantification assay and SDS-PAGE. The proteins were then digested with trypsin to generate peptides, which were labeled using the iTRAQ Reagent-8PLEX Multiplex Kit (Applied Biosystems Sciex, Carlsbad, USA). Specifically, peptides from PLD@NP were labeled with 118-, 119-, and 121-tags, while peptides from PLD@DP were labeled with 115-, 116-, and 117-tags. Liquid chromatographic separation was carried out using a Shimadzu LC-20AD system and a Thermo UltiMate 3000 UHPLC. Following separation, peptides were ionized by nanoESI and introduced into a Q-Exactive HF X tandem mass spectrometer (Thermo Fisher Scientific, San Jose, USA) for detection in DDA mode. Raw mass spectrometry data were converted to mgf format and analyzed using the Mascot protein identification software against relevant databases. Protein identification was performed with iQuant software (Beijing Genomics Institute, Beijing, China), and proteins identified by at least two unique peptides were considered for quantification analysis with a false discovery rate of less than 0.01.

Stable knockdown of gC1qR in ID8 cells

To achieve stable knockdown of gC1qR in ID8 cells, ID8 cells were transfected with a lentivirus that contained CRISPR/Cas9 components. The lentivirus was produced by transfecting HEK 293 T cells with the lentiCRISPR v2 vector (Addgene plasmid 52961), the envelope plasmid pMD2.G (Addgene plasmid 12259) and the packaging plasmid psPAX2 (Addgene plasmid 12260) in a ratio of 4:1:3. Two guide RNA sequences were used to target gC1qR: sgRNA1 with sense sequence CACCGGCGCAGCAGAGGCAGCATCG and antisense sequence AAACCGATGCTGCCTCTGCTGCGCC, sgRNA2 with sense sequence CACCGGCGTGCGCGCAGGTTCCGAG and antisense sequence AAACCTCGGAACCTGCGCGCACGCC, which were cloned into the lentiCRISPR v2 vector. Lentiviral particles were isolated 48 h after transfection and further infected ID8 cells. ID8 cells were cultured and selected with medium containing 2 µg mL⁻¹ puromycin for 7 days, and then expanded in DMEM medium containing 1 µg mL⁻¹ puromycin. The efficiency of gC1qR knockdown in ID8 cells was verified via western blot. The following primary antibodies were used: mouse anti-β-actin (Proteintech, cat. No 60008-1-1 g, clone 7D2C10, 1/2000 dilution) and anti-gC1qR (ABclonal, cat. No A11292, clone ARC2753, 1/1000 dilution).

Cellular uptake by tumor cells

ID8 and ID8^KO cells were exposed to PLD@NP or PLD@DP under conditions where the concentration of DOX was 10 µg mL⁻¹, the pH was maintained at 7.4, and the temperature was kept at 37 °C for a duration of 4 h. The cells were harvested, rinsed with PBS, and then DOX fluorescence was measured by CytoFLEX S flow cytometry (Beckman Coulter, Fullerton, USA).

In vitro cytotoxic effects on cancer cells

ID8 and ID8^KO cells were treated with PLD@NP or PLD@DP under conditions where the concentration of DOX was 10 µg mL⁻¹, the pH was maintained at 7.4, and the temperature was kept at 37 °C for a duration of 24 h. After removing the drug and washing with PBS, 10 µL CCK-8 (Biosharp Company, Shanghai, China) was added to each well and incubated at 37 °C for 2 h in the dark. The adsorbance at 450 nm was measured on a Labsystems iEMS microplate reader (Helsinki, Finland).

In vitro assay for ICD, BMDC maturation and T cell activation

ID8 and ID8^KO cells were incubated with DOX, PLD@NP or PLD@DP at DOX concentration of 10 µg mL⁻¹ for 12 h. The supernatants were collected for analysis of HMGB1 and ATP levels using the HMGB1 ELISA kit (Moshake Biotechnology Co., Ltd, Wuhan, China) and enhanced ATP assay kit (Beyotime, Shanghai, China), respectively. The cells were then stained with a anti-mouse CRT antibody for 30 min and then incubated with FITC-conjugated goat anti-mouse antibody for 60 min. The ratio of CRT-positive cells was determined using CytoFLEX S flow cytometry. The above treated ID8 cells were stained with DiO and co-cultured with immature BMDCs for 12 h. BMDCs were subsequently stained with fluorescence-conjugated antibodies against CD11c, CD80 and CD86. Phagocytosis of tumor cells by BMDCs (CD11c⁺DiO⁺ cells) and BMDC maturation (CD11c⁺CD80⁺CD86⁺ cells) were assessed using CytoFLEX S flow cytometry.

For T lymphocyte activation analysis, BMDCs from the above treatment were co-cultured with murine CD3⁺ T cells at a ratio of 1:4 for three days. Activated T cells were collected and surface-stained with fluorescence-conjugated antibodies against CD3 and CD8. For intracellular cytokine analysis, the cells were treated with fixation buffer (Biolegend, San Diego, USA) and permeabilizing with permeabilization wash buffer (Biolegend, San Diego, USA), and then restained with fluorescence-conjugated antibodies against IL-2, IFN-γ and GzmB before flow cytometric analysis. The specific multi-color antibody combinations employed for the analysis of CRT-positive cells, BMDC maturation and T cell activation are outlined in Table S1.

In vivo biodistribution

ID8 tumor-bearing normal or DIO mice were intravenously injected with CellVue^® Claret-labeled PLD at a dose of 80 µg DOX. The biodistribution of CellVue^® Claret was monitored at different time intervals using a Caliper IVIS Lumina II in vivo imaging system (PerkinElmer, Waltham, MA, USA). At 24 h post-injection, tumor tissues were excised and examined to assess the distribution of CellVue^® Claret within the tumor. Furthermore, mice bearing ID8 and ID8^KO tumors received an intravenous injection of PLD at DOX dosage of 80 µg per mouse. At 24 h post-administration, tumors and key organs (heart, liver, spleen, lungs and kidneys) were obtained after euthanasia of mice. The tissues were were first disrupted in 100 µL of PBS, followed by the addition of 200 µL of methanol. The mixture was then incubated at 4 °C for 30 min to enhance the extraction of DOX. Subsequently, 100 µL of the supernatant was centrifuged at 10,000g for 10 min. The fluorescence of DOX was detected using the FlexStation 3 Multi-Mode Microplate Reader (Molecular Devices, Sunnyvale, USA), with excitation set at 488 nm and emission detected at 585 nm. A calibration curve was established by spiking tissue extracts with varying DOX concentrations, enabling the calculation of DOX content in the samples based on the standard curve.

In vivo anticancer activity

Mice bearing ID8 or ID8^KO tumors received intravenous administration of PLD at 4 mg kg⁻¹ DOX dosage every three days for three times. Intratumoral injections of αC1q were given the day after PLD injections. The tumor size was detected daily by vernier forceps. On the indicated time intervals, part of the mice were sacrificed. Major organs (heart, liver, spleen, lungs and kidneys) were fixed in 4% paraformaldehyde and preserved for H&E staining. Plasma was collected for serological analysis. The remaining mice were used for survival experiments and monitored for tumor size, and were executed when the tumors reached 1500 mm³.

Analysis of tumor immune milieu

After treatment, the tumors were collected, finely minced and then digested in RPMI 1640 medium with recombinant DNaseI (100 U mL⁻¹) and collagenase I (0.8 mg mL⁻¹) at 37 °C for 1 h. Subsequently, the cells were centrifuged at 300g for 5 min, and then lysed with erythrocyte lysate (Biosharp, Hefei, China) for 5 min. The cells were then rinsed with PBS and filtered through a 40 µm filter. To analyze DCs, the cells were stained with fluorescence-labeled antibodies against CD11c, CD80 and CD86. To analyze T cells, the cells were stained with fluorescence-labeled antibodies against CD3, CD8 and CD69. To stain intracellular cytokines, the cells were treated with Fix/Perm solution (Biolegend, San Diego, USA) and re-stained with fluorescence-labeled antibodies against IL-2, IFN-γ and GzmB. For Ki67 staining, the surfaced-stained cells were first treated with pre-cooled 70% ethanol for 1 h and then stained with anti-Ki67-FITC. The cells were subjected to CytoFLEX S flow cytometric analysis. Gating strategies for flow cytometric analysis of cell types involved in this study have been presented in Fig. S15. The detailed multi-color antibody panels utilized for each specific cell type are provided in Table S1.

Statistical analysis

Experiments were conducted a minimum of three times, and results were expressed as means ± SD. GraphPad Prism (version 9.0) was utilized for all statistical evaluations. Comparisons between two groups were made using unpaired two-tailed Student’s t-tests, while comparisons between multiple groups for univariate variables were assessed via one-way ANOVA with Tukey’s HSD post-hoc test and comparisons between multiple groups for multivariate variables were assessed via two-way ANOVA with Bonferroni’s post-test. Significance was set at P < 0.05 for all statistical tests.

Abbreviations

EPR

Enhanced permeability and retention

BMI

Body mass index

DCs

Dendritic cells

BMDCs

Bone marrow-derived dendritic cells

PLD

Pegylated liposomal doxorubicin

RES

Reticuloendothelial system

DIO

Diet-induced obesity

PLD@NP

PLD preincubated with normal mouse plasma

PLD@DP

PLD preincubated with DIO mouse plasma

HDL-C

High-density lipoprotein cholesterol

LDL-C

Low-density lipoprotein cholesterol

C1q

Complement component 1q

αC1q

Anti-C1q

ADPN

Adiponectin

IL-2

Interleukin-2

GM-CSF

Granulocyte-macrophage colony-stimulating factor

IL-4

Interleukin-4

ICD

Immunogenic cell death

CRT

Calreticulin

ATP

Adenosine triphosphate

HMGB1

High mobility group box 1

TEM

Transmission electron microscopic

DLS

Dynamic light scattering

LC-MS

Liquid chromatography-mass spectrometry

SDS-PAGE

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis

HPLC

High-performance liquid chromatography

GO

Gene ontology

KEGG

Kyoto encyclopedia of genes and genomes

ID8^KO

Stable knockdown of gC1qR in ID8 cells

ID8^vector

ID8 cells stably expressing empty vector

IFN-γ

Interferon-γ

GzmB

Granzyme B

H&E

Hematoxylin and eosin

Author contributions

L.G., X.Y., Q.C., and S.X. designed the project. S.X., N.B., X.S., X.Z., S.L., and H.L. performed the experiments. S.X., N.B., T.Y., Q.C., X.Y., and L.G. analyzed and interpreted the data. S.X., and L.G. wrote the manuscript.

Funding

This work was supported by National Basic Research Program of China (2020YFA0710700, 2022YFA1206000, 2021YFA1201200 and 2022YFA1206100) and National Natural Science Foundation of China (82272844).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal experiments received approval from the Institutional Animal Care and Use Committee at Tongji Medical College with an approval number: 4510.

Consent for publication

All authors approved the final manuscript and the submission to this journal.

Competing interests

The authors declare no competing interests.