Hyaluronate-based Self-stabilized Nanoparticles for Reverting Immunosuppression and Immunochemotherapy in Osteosarcoma Treatment

Background: Chemotherapy and immunotherapy are the mainly non-surgical treatment for osteosarcoma in clinic presently. However, serious side effects, short blood circulation time, and immunosuppressive tumor microenvironment have extremely limited their application. To generate an onsite combined chemotherapy and immunotherapy regimen, we designed a self-stabilized hyaluronic acid (HA) nanoparticle for the tumor-targeting delivery of doxorubicin (DOX), cisplatin (DDP), and resiquimod (R848) in osteosarcoma treatment, referred to as DDP NP DOX&R848 . Methods: Here, we tested the physicochemical properties of and the ability of DDP NP DOX&R848 to induce immunogenic cell death (ICD) and revert immunosuppressive cells were characterized in vitro. The therapeutic ecacy of DDP NP DOX&R848 in vivo was evaluated in an orthotopic osteosarcoma mouse model. Furthermore, we also veried the immune memory evoking effect of DDP NP DOX&R848 in an osteosarcoma rechanllenge mouse model. Results: DDP NP DOX&R848 possesses a uniform size (similar to 190 nm) and ideal pH responsibility, and its ability to induce ICD and modulate immune cells was also proven. In vivo, regardless of the mouse model used (i.e., orthotopic osteosarcoma or rechanllenge xenografted osteosarcoma), DDP NP DOX&R848 showed signicantly enhanced tumor inhibition and prolonged survival. Evoked tumoricidal immune memory responses were also demonstrated in the rechanllenged osteosarcoma mouse model.

chemotherapy drugs, these DDSs loaded with chemotherapy drugs demonstrated unsatisfactory performance for long-term tumor inhibition and tumor recurrence prevention. Moreover, the tumorcombating process based on chemotherapy seems to be a passive process of the body, which usually fails to activate its own "defense system". In fact, these issues can hardly be solved for DDSs with single chemotherapy drug loading and will extremely limit their further application.
Recently, immunotherapy has been regarded as a more e cient and thorough method to treat cancer by activating the innate and adaptive immune systems [11,12]. In contrast to the toxicity-dependent tumor inhibition of chemotherapy, immunotherapy is devoted to activating the dampened immune system to induce durable and cascading antitumor immune responses [13][14][15]. Furthermore, increasing positive evidence has shown that an ingenious combination of chemotherapy and immunotherapy could further enhance antitumor e ciency. However, similar to chemotherapy agents, with increasing clinical administration, a growing number of immune-related adverse events have attracted attention, some of which are serious or result in fatal outcomes [16,17]. Fortunately, the diverse nanosized drug carriers available might be promising prospects for the development of immunotherapy [18][19][20]. Therefore, it is of vital signi cance to encapsulate appropriate chemotherapy and immunotherapy agents in a welldesigned drug nanocarrier to deliver an in situ combination of chemotherapy and immunotherapy.
In this study, resiquimod (R848), an agonist of Toll-like receptor 7 (TLR7) and TLR8, was used as the immunotherapy agent to incorporate in our proven chemotherapy drug-loaded DDS based on HA. Thus, an HA-based, DDP-crosslinked, DOX-and R848-loaded nanocarrier (named DDP NP DOX&R848 ) was designed for drug delivery in osteosarcoma immunochemotherapy. DOX was rst assembled with HA as described in our previous report (NP DOX ) [21]. To incorporate R848 in this platform, R848 was bound to poly-Lhistidine (pHis) through hydrophobic interactions to obtain a positive surface potential (NP R848 ). The heavy potential difference between NP DOX and NP R848 drives the encapsulation of NP R848 . Next, DDP was loaded into the system through chelation. Additionally, the crosslinking process of DDP is similar to a "locking" procedure, which can markedly enhance the stability of the nanocarrier. However, when DDP NP DOX&R848 is exposed to an acidic environment, such as tumor microenvironment (TME), the reduced ionization of the HA carboxyl groups and the pHis imidazole group will induce accelerated drug release.
DOX and DDP are two of the most common chemotherapy drugs in osteosarcoma treatment [22][23][24], and they were used as a chemotherapy drugs couple in this study in comparison with the immunochemotherapy agents (R848). Moreover, as a representative anthracycline antibiotic, DOX is able to induce the immunogenic cell death (ICD) of tumor cells, which can facilitate the exposure of tumorassociated antigens (TAAs) from cells [25,26]. Interestingly, R848 is reported to be an immunomodulator due to its ability to remodel the TME by dendritic cell (DC) maturation and tumor-associated macrophage (TAM) reeducation [27,28]. Matured DCs (mDCs) can present released TAAs to native T cells to trigger the immune response. Then, the stimulated immune system will induce the proliferation of immune cells (mainly T cells) effectively and recruit these cells to the tumor site for tumor suppression. Additionally, through the R848-mediated reeducation process, TAMs can be converted from a tumor-supportive phenotype to a tumoricidal phenotype. These tumoricidal TAMs can also enhance antitumor immunotherapy e cacy by releasing proin ammatory factors, inducing stromal destruction and normalizing the tumor vasculature [29,30]. Furthermore, the activated immune system can help the body systemically eliminate tumors and develop a speci c immune memory. These ndings are of vital signi cance in preventing tumor metastasis and recurrence.
To con rm our hypothesis, we performed a series of experiments in vitro and in vivo. DDP NP DOX&R848 demonstrated a uniform and stable structure, speci c pH responsiveness, and robust induction of DCs maturation and TAM reeducation in vitro. In vivo, DDP NP DOX&R848 exhibited obviously increased accumulation at the tumor site compared with free drugs. We also tested the effect of DDP NP DOX&R848 against osteosarcoma in a K7M2 orthotopic osteosarcoma mouse model. As we hypothesized, DDP NP DOX&R848 greatly diminished tumor growth, increased the survival rate, and reduced lung metastasis. The induced anticancer immune responses were clearly investigated by immuno uorescence (IF) analyses. Moreover, the excellent performance of DDP NP DOX&R848 on immune memory evocation has also been veri ed in a K7M2 tumor rechallenging experiment. Overall, DDP NP DOX&R848 is a potential strategy to treat osteosarcoma through the synergy of chemotherapy and immunotherapy.

Results And Discussion
Preparation and characterization of DDP NP DOX&R848 . DDP NP DOX&R848 was fabricated in three steps: 1. synthesizing NP R848 by a nanoprecipitation method; 2.
assembling HA, DOX, and NP R848 through electrostatic interactions; and 3. introducing DDP to complete the crosslinking process. In the rst step, R848 was modi ed with pHis. Because of its branched imidazole group and amidogen group, pHis is a suitable material with pH-dependent amphoteric properties and a positive surface charge. As shown in Fig. 1a and c, the resulting NP R848 nanoparticles had a uniform size of 115 nm. Next, to verify whether there was a su cient potential difference between NP R848 and NP DOX to drive their assembly in the second step, the zeta potentials of NP R848 and NP DOX were measured. As shown in Fig. 1e, the zeta potentials of NP R848 and NP DOX were 43.7 ± 2.8 and − 21.5 ± 1.6 mV, respectively. The large charge difference ensured a stable combination of NP DOX and NP R848 .
The zeta potentials of DDP NP DOX&R848 were also tested (-15.6 ± 2.8 mV). The slightly negative surface charge of DDP NP DOX&R848 contributed to a decrease in its nonspeci c interactions with serum proteins, which would facilitate its longer circulation time and tumor accumulation [31]. For a more compacted and stable construction, DDP was added to accomplish the third crosslinking step, which helped to effectively secure the cargos. As shown in Fig. 1c, the hydrodynamic diameter (D h ) of DDP NP DOX&R848 was 192 nm. The morphology of DDP NP DOX&R848 was also veri ed by transmission electron microscope (TEM) (Fig. 1b). To evaluate the stability of DDP NP DOX&R848 under physiological conditions, its size and zeta potential changes were measured in 10% FBS solution (Fig. 1d). Within 3 days, no signi cant size or zeta potential changes of DDP NP DOX&R848 were observed, which indicated that DDP NP DOX&R848 could maintain its structure under physiological conditions for a long time. The outstanding stability under physiological conditions makes DDP NP DOX&R848 a potential nanoplatform for in vivo applications.
Then, to evaluate the drug release pro le of DDP NP DOX&R848 in vitro, PBS at pH 7.4, 6.5, and 5.5 was used to mimic the pH of physiological, intertumoral, and intracellular microenvironments, respectively. As shown in Fig. 1f, at pH 7.4, only 48.3 ± 6.7% of DOX was released within 72 h. At pH 6.5 and 5.5, the percentage of released DOX was 73.3 ± 4.1% and 88.1 ± 5.7%, respectively, within 72 h. Compared with DOX release at physiological pH, the release of DOX in an acidic environment was accelerated. Moreover, the release of R848 from DDP NP DOX&R848 also demonstrated pH-dependent behavior (Fig. 1h). At pH 5.5, 96.5 ± 4.0% of R848 was released within 72 h, which was 3.0 and 1.1 times higher than that at pH 7.4 and 6.5, respectively. Unsurprisingly, the loaded DDP was released in a pH-dependent manner similar to that of DOX or R848 (Fig. 1g). The promising pH response of DDP NP DOX&R848 was due to its chemical structure.
On the one hand, the ionization degree of HA carboxyl groups was reduced in acidic environments, which disrupts the electrostatic adsorption between HA and DOX. On the other hand, as pHis is a pH-dependent amphoteric biomaterial, the abundant imidazole groups result in a maximum response to a hydrophobicto-hydrophilic switch at pH 6.5. This characteristic bene ts R848 release from the NP R848 core under acidic conditions. Generally, DDP NP DOX&R848 was proven to have a uniform and stable structure, valid drug coencapsulation, and an appropriate pH sensitivity for drug release.
Tumor cell uptake and cytotoxicity.
The uptake of DDP NP DOX&R848 by K7M2 tumor cells was identi ed by FCM. As Fig. 2b shows, after 1 h of incubation with DOX or DDP NP DOX&R848 , the uorescence of DOX could be detected in tumor cells in each group. However, the signal intensity in the DOX-treated group was obviously higher than that in the DDP NP DOX&R848 -treated group. When the incubation time was extended to 6 h, the intensity of DOX uorescence in the two groups showed a conspicuous increase, and more importantly, the uorescence intensity of DOX in the DDP NP DOX&R848 group was very close to that in the DOX group. This effect might be attributed to the distinct uptake methods of free small molecule drugs and the drug-loaded nanocarrier.
Free DOX, as a small molecular drug, is taken up by cells through free diffusion and has an ultrahigh uptake e ciency before reaching a steady state. However, DDP NP DOX&R848 is taken up by tumor cells by endocytosis, which is characterized by its slow uptake rate. These properties led to the different uptake e ciencies in the early and later periods. Moreover, the uptake of drugs was also analyzed by uorescence microscopy (Fig. 2a). Consistent with the FCM results, compared to DOX uptake, the uptake process of DDP NP DOX&R848 was more moderative and complex. The DOX uorescence intensity in the DDP NP DOX&R848 group was obviously lower than that of the DOX group after 1 h of incubation. Moreover, DOX uorescence was observed in the cell nucleus in the DOX groups, while it was mainly located in the cytoplasm in the DDP NP DOX&R848 group. However, when the incubation time was prolonged to 6 h, the DOX uorescence intensity increased in all groups, and the DOX signal in each group was detected in the nucleus. The spatial and temporal disparities in the DOX signal might be attributed to the different methods by which tumor cells take up free small molecule drugs and drug-loaded nanocarriers.
The cytotoxicity of DDP NP DOX&R848 was also analyzed in vitro. As shown in Fig. 2c, R848 failed to in uence the proliferation of tumor cells when used alone or loaded in pHis nanoparticles (NP R848 ). DDP NP DOX&R848 exhibited an effect similar to that of DOX + DDP, DDP NP DOX , or DOX + DDP + R848 in suppressing K7M2 cell proliferation, which indicated that the co-delivery of DOX, DDP, and R848 have no obviously in uence on the cytotoxicity of DOX and DDP ( Fig. 2d and e).
ICD inducing in vitro.
The ICD of tumor cells is a special autophagy pathway that triggers TAA release and antigen-speci c immune response [25]. ICD is characterized by the exposure of calreticulin (CALR) on the cell surface, which serves as an "eat me" signal and can signi cantly promote the phagocytosis of TAAs by DCs [32,33]. Moreover, the release of high mobility group protein 1 (HMGB1) from the nucleus is also an indicator of ICD and can facilitate the maturation of DCs by activating TLR4 [34,35]. The expression of CALR and HMGB1 in tumor cells after the different treatments was analyzed by IF. As shown in Fig. 3a, not only signi cantly upregulated CALR expression on the cell surface but also increased HMGB1 release from the cell nucleus could be observed in tumor cells stimulated by DOX or DDP NP DOX&R848 . These results indicate that DOX and DDP NP DOX&R848 perform well in inducing ICD.

Immunosuppressive cell reversion in vitro.
To demonstrate the e cacy of DDP NP DOX&R848 in altering the immunosuppressive TME, the R848-loaded component of DDP NP DOX&R848 was used to test the ability of DDP NP DOX&R848 to promote DC maturation and TAM reeducation in vitro by FCM. DC maturation is a precondition of TAA presentation that initiates the immune response. Bone marrow-derived dendritic cells (BMDCs) were used for the DC maturation experiment in this study. mDCs are characterized by the upregulated coexpression of CD11c and CD86 and increased secretion of IL-6 and TNF-α. As shown in Fig. 3b, after 24 h of stimulation by R848 or NP R848 , the percentage of cells coexpressing CD11c and CD86 (66.7% and 57.5%, respectively) was signi cantly higher than that of the PBS-treated group (14.2%). Additionally, the signi cantly increased secretion of IL-6 and TNF-α by BMDCs was detected after stimulation by R848 or NP R848 (Fig. 3d) [36]. TAM reeducation is intended to reverse the high M2 macrophage proportion in the TME. Bone-marrow-derived macrophage cells (BMDMs) were induced to differentiate into M2 macrophages (F4/80 + CD206+) for the macrophage reeducation experiment. As shown in Fig. 3c, the percentage of M2 macrophages in total macrophages (F4/80+) was 89.5% in the control group treated with PBS, while this percentage was signi cantly reduced to 20.3% and 23.9% after stimulation by R848 or NP R848 , respectively. Additionally, the expression of two proin ammatory cytokines (IL-6 and TNF-α) was reported to be signi cantly upregulated in M1 macrophages. Thus, the levels of IL-6 and TNF-α secreted by M2 macrophages treated with R848 or NP R848 were both detected (Fig. 3e). Therefore, the ability of NP R848 to re-educate TAMs in vitro has also been proven. Therefore, based on valid TME remodeling, DDP NP DOX&R848 is anticipated to induce tumor inhibition.

Biodistribution in vivo.
To evaluate the superior properties of DDP NP DOX&R848 in tumor-targeted accumulation, the in vivo biodistribution of DDP NP DOX&R848 after administration was evaluated in an orthotopic osteosarcoma model. Because DOX is a uorescent drug, IVIS imaging was used to monitor the drug uorescence intensity of DDP NP DOX&R848 ( Figure S1). The DOX uorescence intensity at the tumor site in the DDP NP DOX&R848 group was about 3.8 times higher than that in the DOX group 12 h after injection, respectively. This nding demonstrated the outstanding ability of DDP NP DOX&R848 to accumulate in the targeted tumor, which is of vital signi cance for reducing side effects and improving therapeutic e cacy.
Furthermore, the biodistribution of DDP NP DOX&R848 in vivo was also evaluated by measure the amount of DOX in major organs after administration. As shown in Figure S2, DDP NP DOX&R848 had a signi cant tumor site accumulation of 4.0 ± 0.7% ID g − 1 24 h after the injection, which is 3.8 times higher than that of DOX. The increased tumor accumulation signi es a high drug concentration in site and tumor inhibition e ciency. In contrast, the uptake of DDP NP DOX&R848 by the liver and kidney was less than that of DOX.
This could be attributed to the decreased clearance rate of DDP NP DOX&R848 and was conducive to reduce the side effects. The great feature of prolonged circular time prolongation and tumor-targeting accumulation make DDP NP DOX&R848 a potential nanocarrier for in vivo applications.
Antitumor effect in vivo.
The in vivo antitumor effect of DDP NP DOX&R848 was evaluated in a K7M2 tibio bular osteosarcoma mouse model (Fig. 4a). As shown in Fig. 4b Fig. 4g, the TIR in the DDP NP DOX&R848 -treated group was 80.5 ± 7.6%, which was signi cantly higher than that in the other treatment groups (P < 0.001). Then, immunohistochemical staining (TUNEL) were also used to assess the improved antitumor effect of DDP NP DOX&R848 . As shown in Figure S3a, the most positive signal was observed in the DDP NP DOX&R848 -treated group. The uorescence intensity of TUNEL ( Figure S3b) in DDP NP DOX&R848 group was signi cantly higher than that in PBS-, DOX + DDP-, R848-, DOX + DDP + R848-, NP R848 -, and DDP NP DOX -treated groups (P < 0.05). Compared to other treatments, DDP NP DOX&R848 induced the highest degree of tumor cell apoptosis. The lightest tumor weight was also detected in mice treated with DDP NP DOX&R848 (Fig. S4a). These results clearly show that DDP NP DOX&R848 exhibits the most robust tumor growth inhibition effect. The security of DDP NP DOX&R848 was evaluated by body weight recording and histology changes observation of four major organs (heart, liver, spleen, kidney). As shown in Fig. 4c, there is no obviously body weight loss could be detected in DDP NP DOX&R848 group. On the contrary, the mice in free drugs treated group (DOX + DDP, R848, and DOX + DDP + R848) were all observed a signi cant body weight loss compared to the control group (P < 0.001). The histology changes of organs were analyzed by H&E staining ( Figure S5). Consistent with the result of body weight changes, some obvious histology damage, such as ballooming degeneration of liver cell, myo bril degradation, or glomerulus degradation could be found in the group treated with DOX + DDP, R848, or DOX + DDP + R848.
However, in the groups treated with drug-loaded nanocarriers ( DDP NP DOX , NP R848 , and DDP NP DOX&R848 ), no obviously organ damage or only slightly organ damage could be observed. These results demonstrate DDP NP DOX&R848 could effectively reduce side effects and possess su cient security compared with free drugs. The survival rate of each group was also recorded (Fig. 4d). Compared with other treatments, DDP NP DOX&R848 displayed the greatest performance in prolonging survival time, which is consistent with its enhanced tumor inhibition e ciency and reduced side effects.
Furthermore, mouse tibio bular osteosarcomas were also assessed by micro-CT scanning and 3D reconstruction to evaluate bone destruction, which is considered to be positively related to the progression of osteosarcoma [37,38]. As shown in Fig. 4e, the most serious bone destruction was observed in the PBS-treated group, which was characterized by osteolysis and heterotopic ossi cation. In contrast, only mild morphological variations in the tibio bular joint were detected in the DDP NP DOX&R848treated group, and the damage was apparently alleviated compared with that of the other groups. To evaluate bone destruction quality, bone volume/tissue volume (BV/TV) and bone mineral density (BMD) were measured by CTan software. The BV/TV was 56.1 ± 9.1% in the DDP NP DOX&R848 -treated group, which was obviously higher than that in the other groups (P < 0.05) (Fig. 4h and i). The highest GMD was also found in the DDP NP DOX&R848 -treated group, which was approximately 2.2 times higher than the control group (P < 0.01). DDP NP DOX&R848 was the most robust agent with superior abilities in not only inhibiting osteosarcoma but also preventing bone destruction.
Moreover, the lung is the most common location of osteosarcoma metastases, and it is an in uential factor that leads to unsatisfactory outcomes [39]. H&E staining of the lungs in each group was performed to assess lung metastases. As shown in Fig. 4f, in the PBS-treated group, the most serious lung metastases could be observed. After different treatments, lung metastases could be inhibited to some degree. Surprisingly, no signi cant lung metastases were detected in the DDP NP DOX&R848 -treated group.
Thus, DDP NP DOX&R848 not only demonstrated the highest tumor inhibition e cacy, longest survival time, and least bone destruction but also the greatest performance on lung metastasis resistance.
Immune response.
After assessing the therapeutic effect, it is clear that DDP NP DOX&R848 might exhibit the most robust tumorinhibitory effects. However, whether this antitumor effect contributed to activation of the immune response still requires further study.
First, we tested the effect of DDP NP DOX&R848 on DC maturation. As shown in Fig. 5a, in the DDP NP DOX&R848 group, the signal of CD86 is obviously stronger than that in other groups. The uorescence intensity of CD86 in DDP NP DOX&R848 group was about 32 times higher than control group (Fig. 5b), which means DDP NP DOX&R848 could effectively mature DCs in TME. The polarization of TAMs has also been measured by IF. Obviously, as shown in Fig. 5a and c, the highest inducible nitric oxide synthase (iNOS) expression has been observed in the DDP NP DOX&R848 group, which indicated that TAMs in tumors could be effectively reestablished to M1-prevailing TAMs by DDP NP DOX&R848 . Furthermore, the secretion levels of two important immune-promoting cytokine (IFN-γ and IL-6) in tumor have also been quali ed ( Figure S6).
Comparing to the other treatments, DDP NP DOX&R848 induced the highest secretion levels of IFN-γ and IL-6 (P < 0.05). DDP NP DOX&R848 showed great potential for TME changeover in vivo, which facilitated activation of the antitumor immune response. It can be concluded that, compared with the other treatments, DDP NP DOX&R848 demonstrated a signi cantly enhanced effect on DC maturation and TAM polarization.
The proliferation and recruitment of CD8 + T cells in the tumor site was also analyzed by IF to directly evaluate immune activation. As shown in Fig. 5a, an obviously increased signal of CD8 + T cells was detected in the DDP NP DOX&R848 group. The immuno uorescence images were semiquantitatively analyzed, and the CD8 signals in the DDP NP DOX&R848 -treated group were 16.9, 8.1, 4.4, 3.3, 3.0, and 2.1 times higher than those in the PBS, DOX + DDP, R848, DOX + DDP + R848, DDP NP DOX , and NP R848 groups, respectively ( Fig. 5d). Clearly, DDP NP DOX&R848 induced the most powerful immune response after administration, and the activated immune system played an essential role in eliciting a promising antitumor effect.
Long-term immune memory effect.
Antigen-speci c immune memory is one of the distinct superiorities of tumor immunotherapy [40,41]. To investigate whether the administration of DDP NP DOX&R848 could induce effective immune memory, we performed an immune memory evocation experiment (Fig. 6a). The rst inoculation of K7M2 cells and the subsequent drug administration were intended to induce primary immunization in vivo. As shown in Fig. 6b, DDP NP DOX&R848 -treated group was demonstrated the slowest tumor growth when compared with the groups treated with PBS, DDP NP DOX , and NP R848 . As shown in Fig. S4bf, the average tumor weight in the DDP NP DOX&R848 -treated group was also signi cantly lighter than that in the other groups (P < 0.001).
Lung metastasis in the DDP NP DOX&R848 -treated group has also been restrained effectively compared with groups of the other treatments (Fig. 6d). Furthermore, the strongest immuno uorescence intensity of CD8 was also observed in tumor tissues of mice with DDP NP DOX&R848 administration. (Fig. 6e). This excellent anti-tumor e ciency is consistent with the excellent tumor inhibition e ciency and effective immune activation ability of DDP NP DOX&R848 in K7M2 tibio bular osteosarcoma mouse model. Then, we analysed the ability of DDP NP DOX&R848 to evoke immune memory by effector memory T (T EM ) cells quanti cation [42]. As shown in Fig. 6f, compared to NP R848 , DDP NP DOX , or PBS, DDP NP DOX&R848 exhibited a signi cantly enhanced ability for T EM cell formation in spleen. This means that DDP NP DOX&R848 could effectively induce the immune memory response in vivo. Moreover, whether the well-established tumor-speci c immune memory could help the mice resist the secondary tumor invasion or not was veri ed by the following tumor rechallenged experiment. As shown in Fig. 6g, the rechallenged tumor in control group grew at a high rate. However, the tumor growth of mice in DDP NP DOX&R848 -treated group has been retarded effectively. Three mice in DDP NP DOX&R848 -treated group were observed resist the secondary tumor challenge completely and no sign of tumors could be detected at the rechallenged site. Compared with other treatments, DDP NP DOX&R848 has been demonstrated the best performance on the secondary tumor resistance. Additionally, within the observation period, the primary tumor recurrence ratio of mice in DDP NP DOX&R848 group was obviously reduced when comparing with groups with other treatments (Fig. 6c). Thus, it can be concluded that DDP NP DOX&R848 could activate the immunoreaction effectively and induce valid, tumor-speci c immune memory.

Conclusions
In summary, we have successfully engineered a pH-sensitive nanocarrier for the effective loading and

ICD inducing in vitro
K7M2 osteosarcoma cells were seeded in 6-well plates at a density of 10 5 cells/well and cultured for 12 h. Then, the original medium was replaced with DMEM containing DOX or DDP NP DOX&R848 at the same DOX concentration of 2 µg·mL − 1 . After 12 h incubation, the supernatants were discarded, and the cells were washed with PBS 3 times for subsequent immuno uorescence analyses of HMGB1 and CALR.
DC activation and TAM re-education in vitro BMDCs and BMDMs were prepared as previously described [43][44][45]. Brie y, bone marrow cells collected from marrow cavities of tibias of C57BL/6 mice (6-8 weeks) were cultured in culture dishes containing 10 mL RPMI 1640 medium containing streptomycin (100 mg mL − 1 ), penicillin (100 IU·mL − 1 ), and 10% fetal bovine serum (FBS). For BMDCs generation, GM-CSF (20 ng mL − 1 ) and 2-mercaptoethanol (50 µM) were added to the culture medium. On day 3, another 10 ml culture medium containing GM-CSF and 2mercaptoethanol was added. On days 6 and 8, half of the culture supernatant was collected and centrifuged, and the cell pellet was resuspended in 10 ml of fresh culture medium containing GM-CSF and 2-mercaptoethanol. On day 10, the nonadherent cells were collected as BMDCs and used in subsequent experiments. BMDMs were generated by incubating bone marrow-derived cells with medium containing M-CSF (20 ng mL − 1 ). On days 3 and 5, the medium was replaced with fresh M-CSF-containing medium. Adherent cells were harvested on day 7 as BMDMs.
BMDCs were treated with R848 or NP R848 at the same R848 concentration of 0.

Analysis Of Tme Re-establishment After Treatment
The re-establishment of TME after different treatments were analysed by IF. The para n-embedded tumors were cut into ~ 3 mm sheets for immunohistochemical analyses. CD86 and iNOS were used to the marker to display the DCs activation and TAM re-education in tumors, respectively. CD8 was stained to evaluated the cytotoxic T cells proliferation in tumors after different treatments. The histological and immunohistochemical alterations were detected by a microscope and subsequently analyzed with ImageJ software. To determine cytokine concentrations in tumors, the separated tumors were homogenized in a protein extraction buffer containing a protease inhibitor (1 mL

Consent for publication
Not applicable Availability of data and material Most of the datasets supporting the conclusions of this article are included within this article and the additional les. The datasets used or analyzed during the current study are available on reasonable request.

Figure 5
Immune response in vivo a Immuno uorescence (CD86, iNOS, and CD8) analyses and b -d the relative optical densities of tumor tissue sections after different treatments. Scale bars: 50 µm. Data are presented as the mean ± SD (n = 3 for b -d; **P < 0.01). Data are presented as the mean ± SD (n = 8 for b and n = 5 for c; **P < 0.01).