*3.4. Hemolysis Test*

As RSANs are administered intravenously, the nanoparticles must not cause the rupture of RBCs. As shown in Figure 5A,B, the positive CG (Tube 1) was completely red and transparent with no cell. This confirmed that the RBCs were ruptured and causes hemolysis. For RSANs group, with the increase in concentration, RBCs sank, and the supernatants become colorless and clear from tube 1 to tube 7 similar to the negative CG (Tube 2). Further observation showed that no erythrocyte was aggregated under the inverted microscope. However, when the concentration increased gradually, slight hemolysis and hemagglutination phenomena were observed (tube 8 to tube 12). Results of hemolysis rate measurement were consistent with visual observation. The percentage of hemolysis was below 0.5% at high concentration, while it was below 0.01% at low concentration. The hemolysis test implied that RSANs were safe and suitable for the intravenous injection at final concentration of ATO lower than 0.0375 mg/mL.

**Figure 5.** In vitro hemolysis results of RSANs after incubation of 3 h ( **A**) and 24 h (**B**). Tube 1 was ultrapure water group, and tube 2 was the physiological saline group.

### *3.5. Macrophage Uptake Study*

After the preparation of RSANs, it was necessary to confirm whether the biological activity of RBCM retains. The character of immune escape is one of the essential biological characteristics. The ability of RSANs for immune evasion was verified using confocal microscopy and flow cytometry by investigating the uptake between SANs and RSANs on RAW 264.7 cells. As shown in Figure 6A, the fluorescence intensity of RSANs groups observed by confocal microscopy was significantly weaker than that of SANs group. Meanwhile, the average fluorescence intensity of SANs and RSANs determined by flow cytometry were 4158.5 and 2045 respectively (*p* < 0.01), which was consistent with confocal results. The study confirmed that encapsulation by RBCM could help the nanoparticles to avoid the uptake of macrophages that can be related to special proteins such as CD47 [36,37]. It can be speculated that RSANs can also be prevented from being ingested by macrophages in vivo, and thus avoid the premature elimination of the drug.

**Figure 6.** Cellular uptake of SANs and RSANs on RAW 264.7 macrophages. ( **A**) Confocal image 630×. The SANs were labeled with 5(6)-Aminofluorescein (green), (**B**) Fluorescence intensity detection by flow cytometry. ( **C**) Quantitative analysis of the fluorescence intensity. Data are shown as ± SD of the mean (*n* = 3). \*\* correspond to *p* < 0.01.

### *3.6. In Vitro Cellular Uptake*

To further confirm the in vitro cellular uptake and structure of RSANS, RBCM, SANs core and cell nucleus of NB4 cells as well as 7721 cells were labeled with Dil, 5(6)-Aminofluorescein and Hoechst 33342, respectively. As shown in Figure 7A,D, the red, green, and blue fluorescence represent the stained RBCM, SANs, and nuclei, respectively. After incubation of NB4 cells and 7721 cells with SANs and RSANs, it was detected that both red and green fluorescence overlapped around the nuclei. This indicated that both SANs and RSANs could be taken up by NB4 cells as well as 7721 cells. In addition, it could be preliminarily speculated the integrity of the basic shell-core structure was also maintained during the process. According to flow cytometry results, the average fluorescence intensity of SANs and RSANs ingested by NB4 cells were 813 and 941 (Figure 7B,C), while those of 7721 cells were 5111 and 5211 (Figure 7E,F) respectively. Both confocal and flow cytometry results showed that the uptake of RSANs was slightly increased without any significant di fference. It can be predicted that the change of the negative charge on the surface of the nanoparticles after being coated by RBCM may cause the slight di fference, which changed the repulsive force between membranes.

**Figure 7.** Cellular uptake of SANs and RSANs on NB4 cells and 7721 cells. ( **A**,**D**) Confocal images 630×. The nucleus of cells was labeled by Hoechst 33342 (blue), SANs were labeled by 5(6)-Aminofluorescein (green) and RBCM were labeled by Dil (red). (**B**,**E**) Fluorescence intensity detection by flow cytometry. (**C**,**F**) Quantitative analysis of the fluorescence intensity. Data are shown as ± SD of the mean (*n* = 3).

### *3.7. ATO Nanoparticles Cytotoxicity*

In vitro cytotoxicity assays were performed to evaluate the safety of nanoparticles initially. As shown in Figure 8A, the cell viability of SNs and RSNs were both higher than 95% at SA concentration of 8–60 μg/mL. It indicated that SA and RBCM were not significantly toxic and had excellent biocompatibility. Free ATO revealed potent cytotoxicity at di fferent concentration; however, the toxicity was reduced significantly after being encapsulated in nanoparticles (Figure 8B). It can be

predicted that the nano-delivery system newly designed can remarkably improve drug safety during medical treatment.

**Figure 8.** Cell survival of 293 cells after administration with different concentrations of (**A**) SNs and RSNs, (**B**) ATO, SANs and RSANs for 24 h. Data are shown as ± SD of the mean (*n* = 3).

### *3.8. In Vitro <sup>E</sup>*ffi*cacy Study*

Figure 9A,B showed that the inhibitory effects of free ATO, SANs, and RSANs on NB4 cells and 7721cells were all dose-dependent. With the increase of ATO concentration, the inhibition was significantly enhanced after 24 h. From Figure 9C,D, potent inhibition was observed at a lower concentration of 1 μg/mL. NB4 cells and 7721 cells were almost completely inhibited at 60 h and 72 h, respectively. At the same time point, the inhibitory intensity of the free ATO group, SANs group, and RSANs group decreased. However, because of the closed system of culture plate and sustained release nanoparticles, drugs in SANs and RSANs could be gradually released over time and eventually reaching the same effect as the free group. According to the results, it can be speculated that RSANs can avoid rapid elimination by the immune system in vivo with the inhibitory effect maintained.

**Figure 9.** The inhibition effects on NB4 and 7721 cells after administration with different formulations. Cell survival of NB4 (**A**) and 7721 (**B**) cells after administration with different concentrations of ATO, SANs and RSANs for 24 h. Cell survival of NB4 (**C**) and 7721 (**D**) cells after administration with multiple groups at the ATO concentration of 1 μg/mL for 72 h. Data are shown as ± SD of the mean (*n* = 3).

#### *3.9. In Vivo Toxicity and Safety Test*

In acute toxicity test, some mice in the high concentration group died, and the surviving mice were inferior to other groups in terms of mental vitality, drinking, and feeding. During continuous administration, the average body weight of the mice in each group showed an increasing trend with no abnormal change, as shown in Figure 10B,C. It indicated that all the agents have no significant systemic toxicity. All hematological analysis results (Figure 10D–F) were within the normal range except for a slight decrease in the number of WBC in the free ATO group. Analysis of the main organ tissue sections after H & E staining showed that the SANs group and RSANs group were similar to the saline group, while the ATO group developed certain lesions (Figure 10A). The specific manifestation including a large number of cardiomyocytes was cytoplasmic loosely stained. Inflammatory infiltration was observed around the local portal area of the liver. The white pulp of spleen part conglutinated to each other with irregular shape and a small number of apoptotic bodies were seen. Moreover, local interstitial congestion could be observed in the kidney. According to the above results, it can be speculated that free ATO group can cause chronic toxicity if administered continuously in this concentration as compared to SANs group and RSANs group. The transient blood concentration of the nano-groups was reduced due to the sustained release, with a lower in vivo toxicity.

**Figure 10.** In vivo toxicity and safety evaluation. ( **A**) H & E staining of primary organs under 100 μm. (**B**) Weight changes of nude mice during the experiment. ( **C**) Organ coe fficients, ( **D**) Number of white blood cells, (**E**) ALT, and (**F**) AST of each group at the end of the experiment. Data are shown as ± SD of the mean (*n* = 5).

### *3.10. In Vivo Anti-Tumor Study*

7221 tumor bearing nude mice were used to determine the anti-tumor e fficacy of formulations. Figure 11 shows the results of di fferent anti-tumor studies. Figure 11A showed the change in tumor volume during administration. In the saline group, the volume was gradually increased while the change trends of ATO group, SANs group, and RSANs group were first increased and then decreased (*p* < 0.01). The tumor mass and size at the end of the study were monitored as Figure 11C,D. According to the results, it could be speculated that the free ATO was quickly cleared in the body, the amount of drug reaching the tumor site was less than that of the other two groups, and that caused the lowest tumor inhibition e fficacy. Over the whole treatment period, SANs group and RSANs group can achieve better tumor inhibition e ffects with the drug gradually released from the nanoparticles. At the same time, because of the wrapping of RBCM, RSANs were less likely to be cleared by the immune system than SANs, allowing more drugs to reach the tumor site. Figure 11B showed changes of body weight. By the end of the study, body weight of the saline group increased significantly, while the other three groups showed little variation, which was consistent with the results of the in vivo toxicity and safety test. It was further supported that the safety and low toxicity of SANs and RSANs, and the DDS can still improve e fficacy of ATO.

**Figure 11.** In vivo anti-tumor e fficacy of formulation in the 7721 tumor bearing nude mice study. Changes of ( **A**) tumor volume and (**B**) body weight during the period of treatment. ( **C**) Average tumor weight after treatment. ( **D**) Photograph of tumors collected after treatment. Data are shown as ± SD of the mean (*n* = 5). \*\* correspond to *p* < 0.01, \* correspond to *p* < 0.05.
