INTRODUCTIONProtein drugs play a significant role in the treatment of many diseases, like cancer, inflammatory diseases, autoimmune diseases, and metabolic diseases.[1] Currently, more than 130 different protein and peptide drugs have been approved for clinical use by the US Food and Drug Administration. Compared with small-molecule drugs, protein drugs have several advantages, such as high specificity, low toxicity, and high therapeutic effect.[2] However, clinical use of protein drugs is still limited by their fragile tertiary structure, low stability, and poor membrane penetrability.[3,4] To solve these problems, various protein delivery vehicles have been developed, such as inorganic nanocarriers,[5,6] metal-organic frameworks,[7,8] liposomes and polymer nanoparticles.[9−12] However, most of the delivery systems simply aimed at proteins functioning outside cells.[2] It is well-known that many protein drugs only work after entering cells.[13] Therefore, development of intracellular protein delivery systems is of great importance for protein therapy.Combining protein drugs with conventional chemotherapeutics provides an appealing approach to enhance anti-cancer effect.[14,15] It has been proven that combination therapy of protein drugs with chemotherapy is more effective and safer than single therapeutics in clinical applications.[16] However, co-delivery systems for efficient delivery of these two kinds of drugs are lacking because of their different properties. Common chemotherapeutics (DOX and PTX) are hydrophobic and are usually encapsulated in nanocarriers via hydrophobic interaction. Protein drugs are often hydrophilic and charged, and they are generally loaded by nanocarriers through electrostatic interaction, hydrogen bonding and coordination interaction under hydrophilic condition. It is difficult to find a strategy to simultaneously load protein drugs and chemotherapeutics with distinct properties efficiently. When these two kinds of drugs are loaded into a nanocarrier, hydrophobic interaction is favorable for chemotherapeutics but adverse to protein drugs. This may not only result in a low loading capacity but also lead to aggregation, denaturation, and inactivation of protein drugs. If protein drugs are loaded under hydrophilic conditions, the loading capacity of chemotherapeutics may be very low due to their poor solubility. Therefore, the different properties of protein drugs and chemotherapeutics hinder their co-encapsulation and co-delivery via nanocarriers. There are some systems which have been designed for co-delivery of these two payloads.[17] For instance, Phua et al. synthesized hyaluronic-acid-based nanoparticles for co-delivery of catalase and Chlorin e6, which afforded enhanced photodynamic therapy efficacy.[18] Kim et al. developed gold nanoparticle-stabilized nanocapsules for co-delivery of caspase-3 and paclitaxel with synergistic cytotoxic effects.[17] The abovementioned systems loaded drugs mainly through covalent bond and non-covalent interactions including hydrophobic and electrostatic interaction, etc. Nevertheless, the covalent bonds generally involve in complicated chemistry that may affect the structure and bioactivity of protein drugs.[19,20] Additionally, it is difficult for the covalently conjugated drugs to be released timely. The non-covalent interactions are not strong enough. The drugs may be released prematurely during blood circulation.[21] Accordingly, it is of great significance to develop intracellular co-delivery systems that can stably load and timely release protein drugs and chemotherapeutics.In recent years, dynamic covalent bonds (DCBs) have received significant attention in the field of drug delivery. There are many kinds of DCBs, such as imine bond, acylhydrazone bond, and boronic ester bond, which have been used as linkers between drugs and carriers.[22−24] For example, Edgar et al. prepared polysaccharide-based Schiff base hydrogels which showed promise in drug delivery.[25] Phenylalanine as an amine containing model drug was linked to the hydrogel through imine bond. Cheng et al. proposed a novel strategy to deliver proteins of different molecular sizes and isoelectric points.[26] Polyphenol-functionalized proteins were combined with boronic acid polymers through boronic ester bond. This strategy not only showed high efficiency in the delivery of various proteins but also maintained the protein bioactivity after intracellular release. It can be summarized from above examples that the advantages of drug-loading via DCBs mainly include the follow two aspects. On the one hand, combination of drugs with carriers via DCBs can effectively improve loading stability and prevent premature drug release during blood circulation. On the other hand, the drugs can be timely and specifically released when DCBs are broken in response to heat, light, and pH.[22] Additionally, protein drugs which are combined with carriers via DCBs can restore their initial states after intracellular release, thus their bioactivity will be fully preserved.[27] Although DCBs have so many advantages, they are mainly used to load a single drug at present, and very few delivery systems have been reported to simultaneously load protein drugs and chemotherapeutics via DCBs. Therefore, simultaneous binding of protein drugs and chemotherapeutics via DCBs may be used to develop a promising strategy for their co-delivery.In this work, we developed a well-designed system based on DCBs for intracellular co-delivery of protein drugs and chemotherapeutics. 2-APBA, a bifunctional molecule with active carbonyl and PBA groups, was used for the combination of drugs with nanocarriers. The carbonyl group of 2-APBA can react with primary amines to form N-B coordinated imine bond, which can be broken by glutathione (GSH).[28] The PBA group can form phenylboronic ester bond with cis-diols.[29,30] For the preparation of RNase A and DOX co-loaded nanoplatform, RNase A and DOX were first modified with 2-APBA by reacting with their amino groups. Subsequently, PBA-containing block polymer PAE-b-P(Asp-co-AspPBA), 2-APBA-functionalized RNase A (2-APBA-RNase A), and 2-APBA-functionalized DOX (2-APBA-DOX) were assembled with DA-containing block polymer PEG-b-P(Asp-co-AspDA) via PBA-catechol bond to form mixed-shell nanoparticles (RNase A/DOX@MNPs) with RNase A/DOX/PAsp as the core and PAE/PEG as the mixed shell (Scheme 1). Under normal physiological condition (pH 7.4), the stable PBA-catechol bonds prevented drugs from premature release in blood circulation. Meanwhile, the PEG shell and collapsed PAE domains protected RNase A away from enzymatic degradation, extending their half-life effectively. Upon arriving at tumor acidic environment (pH 6.5), RNase A/DOX@MNPs were positively charged because of the protonation of PAE chains, promoting their endocytosis by enhancing the electrostatic interaction with negatively charged cell membranes. Upon cellular uptake, the PAE chains further protonated in more acidic endosomes (pH 5.0), leading to the rupture of endosomes through the proton sponge effect. The drugs were released due to the cleavage of the pH-responsive PBA-catechol bond. In cytoplasm, 2-APBA could drop from 2-APBA-RNase A/DOX because N-B coordinated imine bond was broken in response to GSH. The restored RNase A and DOX showed a synergistic and enhanced anti-cancer effect both in vitro and in vivo. This work provides a promising platform for intracellular co-delivery of protein drugs and chemotherapeutics.Fig 1Schematic illustration of the preparation of RNase A/DOX@MNPs and the intracellular co-delivery of RNase A and DOX. PAE-b-P(Asp-co-AspPBA), 2-APBA-RNase A, 2-APBA-DOX and PEG-b-P(Asp-co-AspDA) assembled into stable nanoparticles by DCBs. At the slightly acidic tumor environment (pH 6.5), RNase A/DOX@MNPs were positively charged because of the protonation of PAE chains, promoting their endocytosis. Upon cellular uptake, further protonation of PAE chains led to the rupture of endosomes through the proton sponge effect, and the cleavage of the pH-responsive PBA-catechol bond promoted the release of two drugs (pH 5.0). Finally, the high concentration of GSH in cytoplasm converted 2-APBA-RNase A and 2-APBA-DOX to native RNase A and DOX, leading to a synergistic and enhanced anti-cancer effect.EXPERIMENTALSynthesis and characterizations are included in the electronic supplementary information (ESI).RESULTS AND DISCUSSIONCharacterization of 2-APBA-RNase A and 2-APBA-DOXThe amino groups on RNase A and DOX could react with the carbonyl group on 2-APBA to form N-B coordinated imine bond, thereby PBA groups were modified on RNase A and DOX. Matrix assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) was used to investigate the combination of 2-APBA with RNase A. As shown in Fig. 1(a), the mass spectrum peak of native RNase A is at 13679 Da. For 2-APBA-RNase A, a new characteristic peak appears at 13825 Da and the molecular weight is 146 Da higher than that of RNase A. By calculation, the increased molecular weight exactly came from the combination with one 2-APBA molecule. Fourier transform ion cyclotron resonance ultrahigh resolution mass spectrometer (FT-ICR MS) was applied to investigate the combination of 2-APBA with DOX. As shown in Fig. S9(a) (in ESI), the fragment ions of 2-APBA-DOX after the ionization (Fig. S9b in ESI) can be observed (659.2823), which indicated that 2-APBA-DOX was synthesized successfully. After reaction with 2-APBA, RNase A and DOX were modified with PBA groups which could combine with the catechol group on alizarin red S (ARS), generating fluorescent complex.[31] As shown in Fig. S10 (in ESI), the fluorescence intensity of ARS gradually increases with the increasing concentration of 2-APBA-RNase A or 2-APBA-DOX, indicating the formations of fluorescent complexes (ARS-2-APBA-RNase A and ARS-2-APBA-DOX) and the successful modifications of RNase A and DOX by 2-APBA. Finally, the modification of 2-APBA to RNase A and DOX was studied by 11B-NMR. As shown in Fig. 1(b), 2-APBA alone shows a typical narrow peak at 29.34 ppm, corresponding to the planar trigonal form of boron atom. After reaction with RNase A or DOX, the narrow peak at 29.34 ppm disappears and the new broad peak appears at 1−3 ppm. This was because that 2-APBA formed N-B coordinated imine bond with the amino groups of RNase A or DOX, which made the chemical environment of boron atom changed.[32] Therefore, the results of 11B-NMR further confirmed the combination of 2-APBA with RNase A and DOX.Fig 1Characterization of RNase A/DOX@MNPs. (a) MALDI-TOF mass spectra of native RNase A and 2-APBA-RNase A; (b) 11B-NMR spectra of 2-APBA, 2-APBA-RNase A and 2-APBA-DOX; (c) Size distributions of nanoparticles with different compositions; (d) TEM image of RNase A/DOX@MNPs (scale bar=100 nm); (e) Zeta potentials of nanoparticles with different compositions at pH 7.4, pH 6.5 and pH 5.0; (f) Average diameters and PDIs of RNase A/DOX@MNPs at 4 °C in PBS at pH 7.4 after different time. (g) Size distributions of RNase A/DOX@MNPs incubated at 25 °C in PBS at pH 7.4, pH 6.5, pH 5.0 and 5 g/L GSH for 2 h; (h) In vitro release of RNase A from RNase A-Cy5/DOX@MNPs at 37°C at pH 7.4, pH 6.5, pH 5.0 and 5 g/L GSH; (i) In vitro release of DOX from RNase A-Cy5/DOX@MNPs at 37 °C at pH 7.4, pH 6.5, pH 5.0 and 5 g/L GSH; (j) CD spectra and (k) bioactivity analysis of native RNase A, 2-APBA-RNase A and released RNase A; (l) Relative cell viability of 4T1 cells treated with PEG-b-P(Asp-co-AspDA), PAE-b-P(Asp-co-AspPBA) and blank MNPs at different concentrations (n=5). Data are shown as mean±SD (n=3).Characterization of RNase A/DOX@MNPsIn this study, PAE-b-P(Asp-co-AspPBA), 2-APBA-RNase A, and 2-APBA-DOX were assembled with PEG-b-P(Asp-co-AspDA) through PBA-catechol bond between PBA and DA moieties to form drug-loaded nanoparticles. The combination of PAE-b-P(Asp-co-AspPBA) with PEG-b-P(Asp-co-AspDA) through PBA-catechol bond had been reported in our previous study.[33] To prove the combinations of 2-APBA-functionalized drugs (2-APBA-RNase A and 2-APBA-DOX) with PEG-b-P(Asp-co-AspDA), ARS assay was adopted again. Fluorescence quenching of ARS fluorescent complexes would occur upon competitive replacement of ARS by catechol-containing substance.[31] As shown in Fig. S11 (in ESI), the fluorescence intensity of ARS in the solutions of ARS-2-APBA-RNase A and ARS-2-APBA-DOX gradually decreases with increasing concentration of PEG-b-P(Asp-co-AspDA). This was because that PEG-b-P(Asp-co-AspDA) replaced ARS to combine with 2-APBA-RNase A and 2-APBA-DOX. The formations of PBA-catechol bonds between two polymers and between 2-APBA-functionalized drugs and PEG-b-P(Asp-co-AspDA) ensured that drugs could be effectively and stably loaded in the nanoparticles.Drug-loaded nanoparticles with different formulations were prepared as shown in Table S1 (in ESI). RNase A and DOX co-loaded mixed-shell (PAE/PEG) nanoparticles and single-shell (PEG) nanoparticles were named as RNase A/DOX@MNPs and RNase A/DOX@SNPs, respectively. The encapsulation efficiency (EE) and loading capacity (LC) of nanoparticles were determined by monitoring the fluorescence intensity of Cy5-RNase A and DOX. As shown in Table S1 (in ESI), the EE (RNase A 24%, DOX 30%) and LC (RNase A 7%, DOX 3%) of RNase A/DOX@MNPs are lower than that of RNase A/DOX@SNPs (EE (RNase A 33%, DOX 44%), LC (RNase A 9%, DOX 5%)). This was because a part of the catechol groups on PEG-b-P(Asp-co-AspDA) were used to combine with PBA groups on PAE-b-P(Asp-co-AspPBA) in RNase A/DOX@MNPs, which reduced the binding sites for 2-APBA-RNase A and 2-APBA-DOX. For the same reason, the EE and LC of RNase A @MNPs and DOX@MNPs which only loaded one kind of drug are higher than that of RNase A/DOX@MNPs. Dynamic light scattering (DLS) results show that the average hydrodynamic diameters of DOX@MNPs and RNase A/DOX@MNPs are about 40−110 nm, and RNase A/DOX@SNPs is about 45−160 nm. The largest diameter can be observed in RNase A@MNPs and was about 80−300 nm (Fig. 1c), which may be attributed to its highest LC (20%) for RNase A.[21,34] Transmission electron microscopy (TEM) image (Fig. 1d) shows that RNase A/DOX@MNPs has a uniform spherical morphology with the diameter of about 100 nm, which was consistent with the DLS result. The nanoparticles with other different compositions also have uniform spherical morphologies (Fig. S12 in ESI).The surface potential of nanoparticles has a significant impact on their endocytosis behavior. In this study, the protonation of PAE chains in the mixed-shell nanoparticles (DOX@MNPs, RNase A@MNPs and RNase A/DOX@MNPs) may significantly alter their surface potentials. Therefore, zeta potentials of these nanoparticles were measured at pH 7.4, 6.5 and 5.0 using zeta potential analyzer. As shown in Fig. 1(e), four types of nanoparticles are all negatively charged at pH 7.4. When the pH decreases to 6.5, the potentials of three kinds of mixed-shell nanoparticles (DOX@MNPs, RNase A@MNPs and RNase A/DOX@MNPs) show a negative to positive conversion. However, RNase A/DOX@SNPs whose shells were only consist of PEG were still negatively charged. When the pH further decreases to 5.0, the potentials of all nanoparticles become positive. This pH-responsive variation of surface potentials of the mixed-shell nanoparticles could be explained by the protonation of PAE chains which were positively charged when the pH was less than its pKa (6.8).[35] It could be inferred that the pH-responsive charge conversion of the mixed-shell nanoparticles was conducive to endocytosis in acidic tumor tissue (pH 6.5).RNase A/DOX@MNPs was assembled through pH-responsive PBA-catechol bond which was usually stable at pH 7.4 but was broken at acidic environment. Besides, the N-B coordinated imine bond between 2-APBA and drugs (RNase A and DOX) was GSH-responsive.[28] Therefore, RNase A/DOX@MNPs was stable at pH 7.4 but may be destroyed by acidity and GSH. The stability of RNase A/DOX@MNPs was studied by measuring diameters and polydispersity index (PDI) at different time. There were negligible changes in the size of RNase A/DOX@MNPs and their PDI had been maintained below 0.15 at 4 °C in PBS within 21 days, which implied that RNase A/DOX@MNPs had excellent stability (Fig. 1f). When the pH decreases or the concentration of GSH increases, the size of RNase A/DOX@MNPs changed (Fig. 1g). The diameter was about 60−200 nm at pH 6.5 (mimicking tumor acidic tissue), and was larger than that at pH 7.4 (40−110 nm). This was because the protonation of PAE induced the swelling of RNase A/DOX@MNPs. Then the size variation of RNase A/DOX@MNPs as a function of time at pH 6.5 in 24 h was monitored. As shown in Fig. S13 (in ESI), the sizes of RNase A/DOX@MNPs are about 150 nm and almost not changed with time, and their PDIs were below 0.2 in 24 h. The results indicated that RNase A/DOX@MNPs was stable and could keep its morphology intact under acidic tumor environment (pH 6.5). When the pH was 5.0 or the concentration of GSH was 5 g/L, mimicking acidic endosomes and tumor cytoplasm respectively, the diameters increased to 1000 nm. The variation of size may be attributed to the disintegration of RNase A/DOX@MNPs caused by the breakage of PBA-catechol bond and N-B coordinated imine bond in response to acidity and GSH.[28,29] The results indicated that RNase A/DOX@MNPs could keep relatively stable under normal physiological condition (pH 7.4) and slightly acidic tumor environment (pH 6.5), while underwent disassociation at acidic endosomes (pH 5.0) and at tumor cytoplasm (5 g/L GSH).Based on the pH and GSH-responsiveness of RNase A/DOX@MNPs, the release behavior was studied in PBS at pH 7.4, 6.5, 5.0, and 5 g/L GSH. As shown in Figs. 1(h) and 1(i), the release profiles of RNase A and DOX are similar. Less than 20% of RNase A and DOX were released in 48 h at pH 7.4, which was consistent with the size stability of RNase A/DOX@MNPs at pH 7.4. The release rate was significantly accelerated by acidity and GSH. The cumulative release of RNase A and DOX increased to 40% at pH 6.5 in 48 h. With further decrease of pH to 5.0, the cumulative release of RNase A and DOX increased to 60% and 55% respectively in 48 h. The accelerated drug release under acidic environment may be attributed to the cleavage of PBA-catechol bonds and N-B coordinated imine bonds, which not only resulted in the swelling/dissociation of the nanoparticles but also caused 2-APBA-functionalized drugs to detach from PEG-b-P(Asp-co-AspDA). Similarly, the GSH also dramatically accelerated the drug release. The cumulative release of RNase A and DOX reached 80% at 5 g/L GSH within 48 h. This was because that GSH broke the N-B coordinated imine bond between 2-APBA and drug, which removed 2-APBA from 2-APBA-RNase A/DOX and caused them to detach from PEG-b-P(Asp-co-AspDA), thus resulted in fast release of two drugs.[28] Besides the responsive release of drugs, it was also important for protein delivery system to reserve the bioactivity of protein drugs. Since the bioactivity of protein was closely related to its secondary structure, circular dichroism (CD) measurement was used to investigate the secondary structure of RNase A in different states. As shown in Fig. 1(j), the CD spectra of 2-APBA-RNase A and released RNase A were similar with native RNase A, all of which showed the characteristic peaks of α helix at 208 and 222 nm, which suggested that their secondary structures remained unchanged. Moreover, the bioactivities of 2-APBA-RNase A and released RNase A were very close to native RNase A (Fig. 1k), which indicated that the modification by 2-APBA and responsive release would not influence the bioactivity of RNase A. Furthermore, the cytotoxicity of the used polymers was tested by Cell Counting Kit-8 (CCK-8) assay. About 80% cell viability was found after incubating 4T1 cells with PEG-b-P(Asp-co-AspDA), PAE-b-P(Asp-co-AspPBA), and blank MNPs for 24 h, suggesting the negligible toxicity and good biocompatibility of polymers (Fig. 1l).Intracellular Co-delivery of RNase A and DOXTo evaluate the intracellular co-delivery efficiency of RNase A and DOX, 4T1 cells were treated with RNase A, DOX, RNase A/DOX@SNPs and RNase A/DOX@MNPs (RNase A was labeled with Cy5) at pH 6.5 for 6 h at 37 °C, respectively. Confocal laser scanning microscopy (CLSM) and flow cytometry analysis were used to investigate the intracellular distribution and fluorescence intensity of RNase A and DOX. As shown in Fig. 2(a), free RNase A shows poor intracellular delivery efficiency because of its poor permeability across cell membrane. Nevertheless, when 4T1 cells were treated with RNase A/DOX@SNPs or RNase A/DOX@MNPs, the endocytosis of RNase A significantly increased. The flow cytometry results demonstrated that RNase A/DOX@MNPs shows higher intracellular delivery efficiency than RNase A/DOX@SNPs (Fig. 2b). This was because that RNase A/DOX@MNPs were positively charged due to the protonation of PAE chains at acidic environment, which enhanced the electrostatic interaction with negatively charged cell membranes and promoted endocytosis. The intracellular delivery efficiency of DOX was also studied using CLSM and flow cytometry analysis. As shown in Fig. 2(c), DOX, RNase A/DOX@SNPs and RNase A/DOX@MNPs all show high intracellular delivery efficiencies because DOX itself has good cell membrane permeability. Compared with RNase A/DOX@SNPs and RNase A/DOX@MNPs, 4T1 cells treated with free DOX show higher fluorescence intensity (Fig. 2d), because the diffusion of free DOX was easier than nanoparticles with diameters about 100 nm. It could be concluded from above results that RNase A/DOX@MNPs showed high intracellular co-delivery efficiency of both RNase A and DOX.Fig 2Intracellular delivery of RNase A. (a) CLSM images of 4T1 cells treated with RNase, RNase A/DOX@SNPs and RNase A/DOX@MNPs at pH 6.5 for 6 h at 37 °C. RNase A was labeled with Cy5 (red) and cell nuclei were stained with DAPI (blue); (b) Mean fluorescence intensity of 4T1 cells incubated as described in (a). Intracellular delivery of DOX. (c) CLSM images of 4T1 cells treated with DOX, RNase A/DOX@SNPs, and RNase A/DOX@MNPs at pH 6.5 for 6 h at 37 °C. DOX itself emitted red fluorescence and cell nuclei were stained with DAPI (blue); (d) Mean fluorescence intensity of 4T1 cells incubated as described in (c); (e) Mean fluorescence intensity of 4T1 cells incubated with RNase A/DOX@MNPs at pH 6.5 for 1 h, 4 h, and 8 h at 37 °C; (f) Mean fluorescence intensity of 4T1 cells incubated with RNase A/DOX@MNPs at pH 7.4 and pH 6.5 for 4 h at 37 °C; (g) Three-dimensional CLSM image of 4T1 cells treated with RNase A/DOX@MNPs at pH 6.5 for 6 h at 37 °C. RNase A were labeled with Cy5 (red), cell nuclei were stained with Hoechst 33342 (blue), and endosomes were stained by LysoTracker green (green). (h) Analysis on the colocalization between RNase A (red) and endosomes (green). ****p0.0001. Data are shown as mean±SD (n=3). Scale bar=20 μm.Subsequently, the effect of incubation time and pH on endocytosis of RNase A/DOX@MNPs was investigated by flow cytometry analysis. 4T1 cells were treated with RNase A/DOX@MNPs for different periods. As shown in Fig. 2(e), the mean fluorescence intensities of RNase A and DOX significantly increase when incubation time was extended from 1 h to 4 h and 6 h. This indicated that the endocytosis efficiency could be enhanced by extending incubation time. When 4T1 cells were treated with RNase A/DOX@MNPs at pH 7.4 and pH 6.5, respectively, the results in Fig. 2(f) show that the more effective endocytosis happened at pH 6.5. This was because RNase A/DOX@MNPs was positively charged under acidic conditions, which was more conducive to endocytosis.Endosomal Escape of RNase A/DOX@MNPsUpon endocytosis, to successfully reach the target subcellular compartments and normally function, the drug-loaded nanoparticles must be able to escape from the endosomes. Therefore, it was necessary to investigate the endosomal escape behavior of RNase A/DOX@MNPs. The endosomes were stained by LysoTracker green and fluorescence colocalization of endosomes and RNase A was analyzed by CLSM. When 4T1 cells were treated with RNase A/DOX@MNPs at pH 6.5 for 2 h at 37 °C, apparent co-localization of RNase A and endosomes could be observed (Fig. S14a in ESI). The quantitative analysis implied a high colocalization index (Pr=0.68) (Fig. S14b in ESI). This indicated that RNase A/DOX@MNPs entered the endosomes at this point. When the incubation time was prolonged to 6 h, RNase A does not overlap with endosomes (Fig. 2g). Quantitative analysis showed a very low colocalization index (Pr=0.09) (Fig. 2h), which suggested that RNase A successfully escaped from acidic endosomes and dispersed in the cytoplasm. This could be attributed to the protonation of the PAE chains, which induced the rupture of endosomes because of the proton sponge effect.Study on the Mechanism of EndocytosisThe endocytosis pathway of RNase A/DOX@MNPs was further investigated using CLSM and flow cytometry analysis. As shown in Fig. 3(a), the endocytosis of RNase A/DOX@MNPs is almost completely suppressed at 4 °C, which suggested that it was an energy-dependent process.[36] Among the five inhibitors applied, chlorpromazine and methyl-β-cyclodextrin effectively inhibit endocytosis and the endocytosis efficiencies decrease by 30% and 50%, respectively (Figs. 3b, 3c and 3d). The endocytosis efficiencies of 4T1 cells pretreated with genistein, amiloride and chloroquine diphosphate were almost not affected. Obviously, the endocytosis of RNase A/DOX@MNPs was mainly mediated by clathrin and caveolin.[37] Under the slight acidic environment (pH 6.5), the positively charged RNase A/DOX@MNPs would adsorb onto the negatively charged cell membranes via electrostatic interaction. Subsequently, the nanoparticles entered cells by clathrin- and caveolin-mediated endocytosis. In the endosomes (pH 5.0), the protonation of PAE chains led to the rupture of endosomes because of the proton sponge effect. The cleavage of pH-responsive PBA-catechol bond promoted the release of drugs.[38]Fig 3Effects of temperature and inhibitors on the endocytosis of RNase A/DOX@MNPs. (a) CLSM images of 4T1 cells incubated with RNase A/DOX@MNPs at 37 and 4 °C for 4 h and CLSM images of 4T1 cells pretreated with genistein, chlorpromazine, methyl-β-cyclodextrin, amiloride and chloroquine diphosphate for 2 h and then incubated with RNase A/DOX@MNPs at 37 °C for 4 h. RNase A was labeled with Cy5 (red), DOX itself emitted green fluorescence and cell nuclei were stained with DAPI (blue). Fluorescence intensity of (b) RNase A-Cy5 and (c) DOX of 4T1 cells treated as described in (a); (d) Relative endocytosis efficiencies of 4T1 cells treated as described in (a) analyzed using flow cytometry. *: Compared with 37 °C. ****p0.0001. Data are shown as mean±SD (n=3).Enhanced In vitro Anticancer Efficacy of RNase A/DOX@MNPsTo investigate in vitro anticancer efficacy of RNase A/DOX@MNPs, 4T1 cells were treated with different samples for 24 h, following which the cell viabilities were determined by CCK-8 assay. As shown in Fig. 4(a), 4T1 cells treated with native RNase A show a negligible reduction in cell viability because of the poor cell membrane permeability. When 4T1 cells were treated with RNase A@MNPs, the cell viabilities decrease significantly, which suggested that RNase A@MNPs effectively promoted RNase A to enter cells and induced cell apoptosis (Fig. S15a in ESI). However, 4T1 cells treated with DOX@MNPs display 10%−20% higher cell viabilities than those treated with DOX (Fig. 4b). This was because that the cellular uptake of free DOX is much easier than DOX@MNPs (Fig. S15b in ESI). When 4T1 cells were treated with RNase A/DOX@MNPs at pH 6.5, the cell viability is significantly lower than those treated with RNase A@MNPs and DOX@MNPs. The results indicated a synergistically enhanced anticancer efficacy by the combination therapy (Fig. 4c). The influence of pH on cell viability was investigated by changing the pH of cell culture medium. As shown in Fig. 4(d), the cell viabilities of 4T1 cells treated with RNase A/DOX@MNPs at pH 7.4 are higher than those at pH 6.5 and pH 5.0. At pH 7.4, the IC50 values of RNase A and DOX are 2.08 and 1.03 μmol/L, respectively. With the decreasing of pH to 6.5, the IC50 values of RNase A and DOX significantly decrease to 0.73 and 0.36 μmol/L, respectively. When the pH reduces to 5.0, the IC50 values of RNase A and DOX further decrease to 0.49 and 0.25 μmol/L, respectively. This was because that the positively charged PAE in acidic environment promoted the endocytosis of the mixed-shell nanoparticles, thus induced cell apoptosis more effectively. Therefore, it could be summarized that RNase A/DOX@MNPs exerted an enhanced anticancer effect by efficient intracellular co-delivery RNase A and DOX, which laid the foundation for the antitumor application in vivo.Fig 4In vitro cytotoxicity of different samples on 4T1 cells as determined by CCK-8 assay. 4T1 cells were incubated with the samples for 24 h at 37 °C; (a) Relative cell viabilities of 4T1 cells treated with different concentrations of native RNase A and RNase A@MNPs at pH 6.5; (b) Relative cell viabilities of 4T1 cells treated with different concentrations of DOX and DOX@MNPs at pH 6.5; (c) Relative cell viabilities of 4T1 cells treated with different concentrations of RNase A@MNPs, DOX@MNPs and RNase A/DOX@MNPs at pH 6.5; (d) Relative cell viabilities of 4T1 cells treated with RNase A/DOX@MNPs at pH 7.4 and pH 6.5. ****p0.0001. Data are shown as mean±SD (n=3).Antitumor Efficacy of RNase A/DOX@MNPs In vivoThe in vivo antitumor efficacy of RNase A/DOX@MNPs was then evaluated with the 4T1 tumor-bearing model mice. 25 Mice were randomly divided into five groups (n=5) and intravenously injected with different samples (PBS, DOX, RNase A@MNPs, DOX@MNPs, RNase A/DOX@MNPs) every two days (Fig. 5a). As shown in Figs. 5(b) and 5(c), within the observation period of 13 days, for PBS group without therapy, the tumor growth rate is the fastest and the tumor is the heaviest at the end of treatment. Comparatively, the tumor growth is partially inhibited by RNase A@MNPs and DOX@MNPs with only one kind of drug loaded. For the free DOX group, the tumor growth rate is slower than that of DOX@MNPs group, which may be attributed to the easier cellular uptake of free DOX than DOX@MNPs as discussed above. When RNase A/DOX@MNPs was used, the tumor growth rate is the slowest and the tumor weight is the lightest at the end of the treatment, which indicated that RNase A/DOX@MNPs could significantly enhance the antitumor efficacy through the synergistic co-delivery of RNase A and DOX. It should be noted that the tumor growth rate in the DOX group was slower than that of RNase A/DOX@MNPs group in the first 8 days, but later it was faster until the end. This was because that DOX was pumped outside cells due to its drug resistance, resulting in the decrease of antitumor efficiency in the latter phase of treatment.[39] After treatment, the mice were euthanized and the tumor tissues were subjected to terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and H&E staining to evaluate the cell apoptosis. Consistent with the tumor suppression results, tumor sections of RNase A/DOX@MNPs group shows the strongest green fluorescence signal, displaying the highest level of cell apoptosis (Fig. 5d). From the H&E images (Fig. 5e), typical apoptotic characteristics including condensed and hyperchromatic nuclei were observed in RNase A/DOX@MNPs group, which revealed that RNase A/DOX@MNPs caused severe damage to tumor tissues.Fig 5In vivo antitumor efficiency. (a) Schematic illustration of the experimental protocol. BALB/c mice were injected with 1×106 4T1 cells subcutaneously and treated with different samples on day 1, 3, 5, 7, 9 and 11 via intravenous injection; (b) Average tumor growth profiles obtained from 4T1 tumor-bearing mice treated with PBS, DOX, RNase A@MNPs, DOX@MNPs, RNase A/DOX@MNPs, respectively; (c) Weights of tumors from the mice after treatment. Inset: Representative photographs of the excised tumors from the mice treated as described in (a) on day 13. (d) TUNEL and (e) H&E staining of 4T1 tumor tissues from the mice treated as described in (a) on day 13. *: compared with PBS. ****p0.0001. Data are shown as mean±SD (n=5). Scale bar=100 μm.During the treatment, systemic toxicities of the tested samples were assessed by monitoring the change in body weight of mice. As shown in Fig. S16 (in ESI), no evident body weight loss can be observed in RNase A/DOX@MNPs group, suggesting good biosafety of RNase A/DOX@MNPs. The systemic toxicity was also evaluated by H&E staining of heart, liver, spleen, lung and kidney from treated mice (Fig. S17 in ESI). The organs from RNase A/DOX@MNPs treated mice present no pathological changes when compared with the PBS group, which suggested that RNase A/DOX@MNPs not only had good biocompatibility but also reduced the side effects of DOX. Furthermore, no abnormal changes in the hearts can be observed in DOX treated group. This was because that the DOX dosage used (2 mg/kg) was lower than the commonly reported toxic dosage (5−7 mg/kg),[40] thus avoided significant side effects.CONCLUSIONSIn summary, we developed a kind of well-designed nanoplatform for intracellular co-delivery of RNase A and DOX. PAE-b-P(Asp-co-AspPBA), 2-APBA-RNase A, and 2-APBA-DOX were assembled with PEG-b-P(Asp-co-AspDA) through dynamic PBA-catechol bond to form RNase A/DOX co-loaded mixed-shell nanoparticles (RNase A/DOX@MNPs). RNase A and DOX were encapsulated in RNase A/DOX@MNPs with good stability and satisfied LC/EE under physiological conditions. The protonation of PAE chains under acidic environment not only promoted the endocytosis of nanoparticles but also helped drugs to escape from endosomes. Both RNase A and DOX could be timely released due to the cleavage of the pH-responsive PBA-catechol bond and the GSH-responsive N-B coordinated imine bond. RNase A/DOX@MNPs could effectively induce tumor cell apoptosis and inhibit tumor growth, thus showed a synergistic and enhanced antic-cancer effect. Considering the diversity of amino containing protein drugs and chemotherapeutics, this nanoplatform may be potentially applied in co-delivery of other drug combinations and may promote the development of anticancer combination therapy.