Twins‐like nanodrugs synchronously transport in blood and coalesce inside tumors for sensitive ultrasound imaging and triggerable penetrative drug delivery

Nanodrugs capable of aggregating in the tumor microenvironment (TME) have demonstrated great efficiency in improving the therapeutic outcome. Among various approaches, the strategy utilizing electrostatic interaction as a driving force to achieve intratumor aggregation of nanodrugs has attracted great attention. However, the great difference between the two nanodrugs with varied physicochemical properties makes their synchronous transport in blood circulation and equal‐opportunity tumor uptake impossible, which significantly detracts from the beneficial effects of nanodrug aggregation inside tumors. We herein propose a new strategy to construct a pair of extremely similar nanodrugs, referred to as “twins‐like nanodrugs (TLNs)”, which have identical physicochemical properties including the same morphology, size, and electroneutrality to render them the same blood circulation time and tumor entrance. The 1:1 mixture of TLNs (TLNs‐Mix) intravenously injected into a mouse model efficiently accumulates in tumor sites and then transfers to oppositely charged nanodrugs for electrostatic interaction‐driven coalescence via responding to matrix metalloproteinase‐2 (MMP‐2) enriched in tumor. In addition to enhanced tumor retention, the thus‐formed micron‐sized aggregates show high echo intensity essential for ultrasound imaging as well as ultrasound‐triggered penetrative drug delivery. Owing to their distinctive features, the TLNs‐Mix carrying sonosensitizer, immune adjuvant, and ultrasound contrast agent exert potent sonodynamic immunotherapy against hypovascular hepatoma, demonstrating their great potential in treating solid malignancies.

poor tumor retention as a predominant hurdle for the in vivo delivery of nanodrugs.Even though nanodrugs with elaborately regulated particle size, morphology, and surface chemistry may enter the solid tumors, [5][6][7] the high interstitial fluid pressure (IFP) in there may generate a strong washout effect shortening the tumor retention time to significantly affect the anti-cancer effect of nanodrugs. [8]oteworthily, nanoparticle aggregation inside tumors has proven highly effective for enhancing the tumor retention of nanodrugs.11][12][13] S C H E M E 1 Schematic illustration of matrix metalloproteinase-2 (MMP-2) sensitive twins-like nanoparticles (TLNs) synchronously transport in blood and coalesce inside tumor for sensitive ultrasound imaging and triggerable penetrative drug delivery.
Up to now, various strategies have been proposed for achieving nanoparticle aggregation inside tumors.For example, Cheng et al. developed a "grafting/peeling-off" strategy for modifying iron oxide nanoparticles with a polyethylene glycol (PEG) layer through an enzyme-responsive peptide.The PEG coating was selectively removed in the enzyme-enriched tumor microenvironment (TME) to turn the nanoparticles hydrophobic, which induced their aggregation for prolonged tumor retention and enhanced magnetic resonance imaging. [14]Moreover, recent studies have found that two nanodrugs pre-modified with the alkynyl group and the azide group separately may undergo a click reaction inside the tumor to induce in situ aggregation. [15,16]Due to the absence of a catalyst in the tumor site, the click reaction was subject to a low reaction rate and efficiency, which may be major obstacles to forming the desirable large aggregates inside the tumor.In comparison, since two nanoparticles with similar sizes and opposite surface charges may assemble into higher-order structures, [17] coulombic interactions may provide a facile solution for intratumor aggregation of nanodrugs.Unfortunately, nanoparticles with different physicochemical properties usually show great differences in blood circulation time, meaning that an equal-opportunity tumor uptake for efficient intratumor aggregation is impossible.For example, it is well-known that a negatively charged surface may increase the blood half-lives of nanoparticles, while a positively charged surface may behave oppositely. [18]herefore, to achieve intratumor aggregation of two oppositely charged nanodrugs, one may need to modify the surface of the positively charged nanodrug with negative groups which are sheddable in TME in order to prolong its blood circulation for better tumor accumulation. [19]However, the different chemical compositions on the surface of two nanodrugs would still impede an equal-opportunity tumor accumulation to affect the intratumor aggregation, thus detracting from the resultant beneficial effects.
On the other hand, nanodrug aggregation inside tumors also brings about new challenges, chief among them being decreased tissue penetration, especially in hypovascular tumors. [20]It is well-known that micron-sized particles carrying bioinert gases such as perfluorocarbons (PFCs) have high echo intensities, which makes them suitable to act as ultrasonographic contrast agents, e.g., the three microbubbles (Optison, Definity, and SonoVue) approved for clinical applications. [21]Furthermore, recent studies have revealed the great performance of these echogenic microparticles as ultrasound-triggerable drug delivery systems to achieve controlled drug release and tissue-penetrating drug delivery in diseased sites such as solid tumors. [22]For example, Zheng et al. reported that a perfluoropropane-filled microbubble showed high echo intensities, owing to which the externally applied ultrasound broke the microbubble into smaller nanoparticles delivering drugs deep into tumors. [23]Based on these advances, intratumor aggregation of PFCs-encapsulated nanodrugs is likely an effective approach to ultrasoundtriggered drug penetration in solid tumors in addition to sensitive ultrasound imaging.
Herein, we put forward a new strategy for constructing a pair of extremely similar nanodrugs, referred to as "twinslike nanodrugs (TLNs)", to achieve equal-opportunity tumor uptake and intratumor aggregation.As shown in Scheme 1, the TLNs@C/R, that is, TLNs-1@C/R (amide inward oriented) and TLNs-2@C/R (amide outward orientated), carrying sonosensitizer Chlorin e6 (Ce6), toll-like receptor 7 (TLR-7) agonist imiquimod (R837) as an immune adjuvant and ultrasound contrast agent 1,1,1,3,3-pentafluorobutane (PFB) were prepared respectively from two MMP-sensitive block copolymers whose molecular structures only differ in the orientations of the MMP-2-sensitive peptide linking the mPEG block and the PAsp(MEA/DIP/PFA) block (Figure S1).Since the two block copolymers share the same mPEG, peptide, and PAsp(MEA/DIP/PFA) blocks, the TLNs would have the same physicochemical properties to ensure that both their blood circulation and tumor entrance are in lockstep.Moreover, the tumor-enriched MMP-2 would specifically break the oppositely oriented amide bonds in the peptide linker to leave carboxyl groups on one nanodrug (TLNs-N@C/R, negative) and amino groups on the other one (TLNs-P@C/R, positive), yielding two oppositely charged surfaces to initiate electrostatic interaction which would drive nanodrug aggregation in situ.Consequently, the formation of large aggregates, ideally micron-sized, would not only increase the tumor retention of nanodrugs but also boost their echogenicity to enable ultrasound imaging as well as ultrasound-triggered penetrative drug delivery inside a tumor.The potential of our new strategy to realize ultrasound imaging and potent sonodynamic immunotherapy was explored in hypovascular hepatoma.

Polymer synthesis and nanodrug preparation
Two amphiphilic copolymers having exactly the same triblock structures but differing in the orientations of the MMP-2-sensitive peptide junction blocks, mPEG-Pep(+)-PAsp(MEA/DIP/PFA) (P-PPAEDP) and mPEG-Pep(−)-PAsp(MEA/DIP/PFA) (N-PPAEDP), were synthesized via multistep reactions (Figure S1).As shown in Figure 1A,B and Figure S2,S3, analyses using 1 H nuclear magnetic resonance spectrum, Fourier-transform infrared spectroscopy, and gel permeation chromatography (GPC) demonstrated the successful synthesis of the two triblock copolymers.After treating the triblock copolymer with MMP-2, two prominent peaks were shown in the GPC chromatogram for the PEG block and PAsp(MEA-DIP-PFA) block, respectively, which verified the enzymatic cleavage of peptide linker by MMP-2 (Figure 1B).
The detailed information about block lengths of copolymers was summarized in Table 1, under which the two triblock copolymers in aqueous solution may self-assemble into the aimed twins-like nanoparticles (TLNs@C/R), that is, TLNs-1@C/R and TLNs-2@C/R, incorporating Ce6, R837, and PFB.Dynamic light scattering analysis showed that TLNs-1@C/R and TLNs-2@C/R possessed approximately the same nanosize around 140 nm (Table S1), which is important for their synchronous blood circulation and equal-opportunity tumor uptake.TLNs-1@C/R, TLNs-2@C/R, and their equimolar mixture (shown as TLNs-Mix@C/R unless otherwise noted) were all stable in the bloodstream-mimicking physiological environment containing 10% fetal bovine serum (FBS) (pH 7.4, 37 • C), showing stable particle size and drug-loading capacity over the experimental time of 4 days as displayed in Figure S5.At pH 6.5, TLNs-1@C/R, TLNs-1@C/R, and TLNs-Mix@C/R all expanded to about 350 nm due to the partial protonation of N,N-diisopropylamino ethylamine (DIP) groups, while they were not disassembled because the introduced pendant MEA groups crosslinked the membrane of polymersomes by forming disulfide bonds. [24]Notably, after adding 10 nM MMP-2 into the solution of TLNs-Mix@C/R, the formation of microscale particles up to 1400 nm was detected at pH 6.5 (Figure 1C and Table S1).The transmission electron microscopic (TEM) observation further revealed that nanoparticle aggregates were formed upon the MMP-2 treatment (Figure 1D).TLNs-1@C/R, TLNs-2@C/R, and TLNs-Mix@C/R without MMP-2 treatment all exhibited weak negative charge at pH 7.4 and weak positive charge at pH 6.5 (Table S2).After MMP-2 treatment at pH 6.5, TLNs-1@C/R and TLNs-2@C/R displayed a fairly strong positive charge of +23.6 ± 0.5 mV and a fairly strong negative charge of −13.7 ± 3.2 mV, respectively (Figure 1C).The above results were reasonable because cleavage of the peptide linker left different chemical groups on the nanoparticle surface, that is, amino groups on the new nanodrug TLNs-P@C/R resulted from TLNs-1@C/R but carboxyl groups on another new nanodrug TLNs-N@C/R resulted from TLNs-2@C/R.Moreover, when TLNs-Mix@C/R consisting of equimolar TLNs-1@C/R and TLNs-2@C/R were incubated with 10 nM MMP-2 in a solution of pH 6.5, a zeta potential of +7.7 ± 1.8 mV was detected.Obviously, the formation of nanodrug aggregates was driven by the electrostatic interactions because the cleavage of oppositely oriented peptide linkers turned the TLNs-Mix@C/R into two oppositely charged nanodrugs (Scheme 1).A near-infrared dye cyanine7.5(Cy7.5) was incorporated into TLNs for assessing the synchronous transport in blood.The clearance rates of nanodrugs from blood after tail-vein injection were determined by analyzing the Cy7.5 fluorescence intensities of mouse orbital blood at different time points post-injection.As shown in Figure 1E, the TLNs-1@C/R, TLNs-2@C/R, and TLNs-Mix@C/R exhibited similar blood clearance rates, implying an equal-opportunity tumor entrance of the two nanodrugs.However, the MMP-2-pretreated TLNs mixture possessed distinctly shortened blood circulation time, obviously because the enzyme treatment triggered TLNs-Mix aggregated into micron-sized clusters to decrease their blood half-lives. [25]

In vitro drug release and ultrasound imaging of TLNs
The TLNs-Mix@C/R solutions containing equimolar TLNs-1@C/R and TLNs-2@C/R (Table S3) were used for assessing the drug release and US imaging potential in vitro.As shown in Figure S4,S6, treatment with 10 nM MMP-2 resulted in rapid release of the encapsulated drug (Ce6) under low-frequency ultrasound (LFUS).Consequently, the MTT assay showed that LFUS irradiation significantly increased the cytotoxicity of the MMP-2-pretreated TLNs-Mix@C/R against Hepa1-6 cells (Figure S8).Meanwhile, in comparison with the TLNs-Mix@C/R solutions without MMP-2 pretreatment, the TLNs-Mix@C/R solutions with MMP-2 pretreatment displayed much stronger US imaging signals (Figure 1F).Nevertheless, because an LFUS pre-irradiation broke the nanodrug disassembly, the US imaging signals of TLNs-Mix@C/R solutions were dramatically weakened in this event.These results imply that TLNs-Mix@C/R may aggregate inside the MMP-2-enriched tumor tissue to boost their echogenicity, based on which not only a sensitive US imaging but also LFUS-triggered drug release and tissue-penetrating drug delivery may be achievable. [26,27]

LFUS-promoted drug diffusion and sonodynamic therapy in vitro
The drug diffusion promoted by LFUS was assessed in a three-dimensional cell sphere model mimicking the solid tumor.When the LFUS was not applied, Ce6 fluorescence was only detected on the surface of the cell sphere in the TLNs-Mix@Ce6 + MMP-2 group (Figure 2A), meaning that nanodrugs were endocytosized only by the peripheral cells and could not penetrate the cell sphere.However, when the LFUS was applied, a much more uniform distribution of Ce6 fluorescence throughout the whole sphere was observed.The analysis of the Ce6 fluorescence profile along the Z axis showed more clearly the difference resulting from an LFUS irradiation.Additionally, quantitative analysis of monodisperse cells from the above cell spheres with flow cytometry indicated that the percentage of Ce6-positive cells was significantly increased to 68.22 ± 6.18% of the MMP-2 + LFUS group from just 16.47 ± 2.70% of the MMP-2 group (Figure 2B,C).Most likely, as the MMP-2-induced electrostatic aggregation of nanodrugs greatly increased the US sensitivity, an LFUS irradiation could generate a strong cavitation effect, [28,29] which broke the aggregates into small pieces and drove a cell sphere-penetrating delivery. [30,31]n consideration that the TLNs-Mix loaded with the photosensitizer Ce6 upon LFUS irradiation may exert sonodynamic therapy (SDT) to eradicate solid tumors (Figure 2D), we conducted several experiments to test the therapeutic effect of the nanodrugs in vitro.The level of intracellular reactive oxygen species (ROS) was analyzed using 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) as a fluorescent probe.As shown in Figure 2E, cells not incubated with nanodrug showed no green fluorescence regardless of the MMP-2 addition and LFUS irradiation.In addition, MMP-2 treatment and LFUS irradiation only led to fairly weak green fluorescence in the TLNs-2@C/R group, indicating the production of very limited ROS.34] Notably, cells treated with TLNs-Mix@C/R with MMP-2 pretreatment and LFUS irradiation showed the highest level of ROS.Most likely, owing to the echo-enhancing aggregation of TLNs-Mix induced by MMP-2 in this event, the LFUS irradiation generated the strongest ultrasonic cavitation effect for a much more efficient cellar uptake of Ce6, thereby generating abundant ROS inside the cells.Quantitative analysis with flow cytometry showed consistent results.That is, the TLNs-1@C/R treatment group with MMP-2 pretreatment and LFUS irradiation showed a significantly higher intracellular ROS level than all other groups (Figure 2F).As ROS accounts for cell killing in SDT, dead tumor cells were quantified using a double staining kit for apoptosis.In comparison with other groups, the TLNs-Mix@C/R group with MMP-2 pretreatment and LFUS irradiation exhibited the highest level of tumor cell apoptosis (64.41 ± 9.26%), including the early and late apoptosis (Figure 2G,H).

Detection of damage-associated molecular patterns and activation of immune cells in vitro
It is known that excessive ROS causes endoplasmic reticulum (ER) stress resulting in up-regulated expressions of a series of related "chaperone" proteins. [35]For instance, calreticulin (CALR) is one of the key related proteins that is up-regulated as a typical member of damage-associated molecular patterns (DAMPs) under the condition of ER stress. [36]DAMPs can induce dendritic cell (DC)-mediated innate immunity of cancer, in which other modulators may be needed in order to provoke potent adaptive immune responses and to avoid chronic inflammations.Herein, a TLR agonist R837 as an immune modulator was incorporated into TLNs-Mix to boost the immune response elicited by SDT.
The expression and location of CALR in tumor cells receiving nanodrug treatment were observed under a confocal laser scanning microscope (CLSM).As shown in Figure 3A, there was no obvious expression of CALR (green) in the TLNs-Mix@C/R group and the TLNs-Mix@R + MMP + LFUS group, indicating that both Ce6 and LFUS are essential for inducing effective SDT.In contrast, CALR was significantly increased in cells treated with Ce6-loaded nanodrugs under LFUS, i.e., the TLNs-Mix@C + MMP + LFUS group and the TLNs-Mix@C/R + MMP + LFUS group.In addition, migration of CALR from cytoplasm to cell membrane was observed.Two other well-known proteins of DAMPs, the high mobility group box-1 (HMGB1) and adenosine triphosphate (ATP), were evaluated as well.The above-mentioned two treatment groups with effective SDT also showed higher levels of ATP and HMGB1 released by cells than other groups (Figure 3B,C).These up-regulated DAMPs enhance the immunogenicity of tumor sites and activate the innate immune system through pattern recognition receptors. [37,38]e further investigated whether combination therapy of R837 and SDT could promote bone marrow-derived DCs (BMDCs) maturation and T cell activation using a cell co-culture model (Figure 3D).After incubating the tumor cells with nanodrugs in different conditions, the cell culture medium was collected for further DC incubation.As shown in Figure 3E,F, an effective SDT was essential for inducing DC maturation.The culture medium supernatant collected after the tumor cells were incubated with TLNs-Mix@C/R only induced maturation in a small number of BMDCs (11.84%) due to the absence of SDT without LFUS.On the contrary, the supernatant collected after the tumor cells were incubated with TLNs-Mix@C/R in the presence of MMP-2 and LFUS caused maturation in a large number of BMDCs (50.62%) as well as up-regulated expression of MHCII promoting the antigen-presenting of DCs.Besides, the expressions of proinflammatory cytokines such as interferon (IFN)-β and interleukin (IL)-12 were significantly increased in the TLNs-Mix@C/R treatment group with MMP-2 and LFUS (Figure 3G).These results implied that DCs exposed to R837 and DAMPs were activated to mature phenotype for antigen-presenting, which promoted the secretion of proinflammatory cytokines to boost the immune response.Taken together, owing to the MMP-2-induced aggregation of TLNs-Mix@C/R, LFUS induces an efficient SDT on tumor cells, and then these damaged cells release DAMPs as "eat me" sig-nals to facilitate DCs maturation and elicit a strong adaptive anti-tumor immune response (Figure 3H).

In vivo tumor accumulation, penetrative drug delivery, and US imaging
The LFUS-triggered tumor-penetrating delivery of drugs based on the MMP-2-induced TLNs-Mix aggregation was explored using C57BL/6J mice bearing subcutaneous Hepa1-6 tumor.TLNs-Mix was labeled with a near-infrared fluorescent dye DiR for the in vivo fluorescence imaging.As shown in Figure 4A, the fluorescence imaging indicated that DiR was delivered to the tumor site at 2 h after injecting TLNs-Mix (1:1 molar ratio unless otherwise noted).The DiR fluorescence at tumor sites reached the highest level at 8 h and then lasted above 72 h, showing an excellent tumor accumulation and retention effect of nanodrugs.However, administration of the same dose of unpaired nanodrug (TLNs-1 or TLNs-2) resulted in obviously weaker DiR fluorescence at tumor sites after 72 h post-injection (Figure S7A,B).The percent injected dose per gram (%ID/g) in the tumors of the TLNs-Mix group was higher than the TLNs-1 or TLNs-2 group (Figure S7C).These results demonstrated an enhanced tumor retention of TLNs-Mix which may provide an extended time window for SDT and US imaging.Most likely, the MMP-2-enriched TME caused the PEG shedding of TLNs-Mix to yield oppositely charged nanoparticles driving the formation of large aggregates, which was in line with the aforementioned in vitro results of TEM observation.According to the tumor accumulation profile of TLNs-Mix, the tumor sites were exposed to LFUS at 8 h after tail vein injection, which dramatically increased the DiR fluorescence in the tumor (Figure 4B).Quantitative data revealed that the fluorescence intensity at tumor sites of mice receiving TLNs-Mix showed a 5-fold increase upon the LFUS irradiation (Figure 4C).In contrast, the fluorescence intensity at tumor sites of mice receiving unpaired TLNs-2 and TLNs-1 only exhibited a 1.3-fold increase and a 1.5-fold increase upon LFUS irradiation, respectively (Figure S7D,E).Obviously, dequenching of DiR fluorescence was induced by the LFUS irradiation in tumor sites of mice receiving TLNs-Mix.Most likely, after TLNs-Mix was injected, the two nanodrugs TLNs-1 and TLNs-2 accumulated in tumor sites, and then MMP-2 enriched therein broke the oppositely oriented amide bonds of their same peptide linker, which left carboxyl groups on one resulted in nanodrug TLNs-N and amino groups on another resulted in nanodrug TLNs-P to yield two oppositely charged surfaces.Consequently, electrostatic interaction drove the nanodrug aggregation in situ.According to the in vitro results, these newly formed large aggregates would have higher echogenicity enabling the LFUS to break the aggregates and trigger an effective DiR release.
To further explore whether the in situ aggregation of TLNs-Mix would assist the LFUS in promoting tumor-penetrating drug delivery, we analyzed the co-localization of blood vessels and drugs using immunofluorescence imaging.Prior to the LFUS irradiation, the TLNs-Mix@C/R accumulated in extravascular tumor tissue much more efficiently than the unpaired TLNs-2@C/R or TLNs-1@C/R, which was attributed to the enhanced tumor retention effect resulting from the in situ nanoparticle aggregation (Figure 4D).Upon the LFUS irradiation, the drug penetration inside the tumor was promoted for all three nanodrug treatments, especially Ce6 delivered by TLNs-Mix@C/R diffused most distantly in the hypovascular hepatoma model featuring dense tumor tissue (Figure 4D,E).Most likely, the formation of large aggregates with higher echogenicity enabled LFUS to trigger more effective Ce6 release and generate more potent cavitation as an energic effect promoting drug diffusion (Figure 4F).
It is well-known that microbubbles with high echogenicity enhance ultrasound imaging. [39,40]Here, we detected the US signals at tumor sites of mice receiving intravenous injections of TLNs-Mix@C/R.The clinically available contrast agent, SonoVue with a particle size of 2.5 μm, was used for comparison.Since microscale particles were unable to cross the tumor vasculature, [41][42][43][44][45] no obvious US signal was detected in tumor tissue (circled by a dotted line) of mice receiving SonoVue injection (Figure 4G).On the contrary, strong US signals were detected at tumor sites of mice injected with TLNs-Mix, which was attributed to the in situ aggregation increasing echogenicity for US imaging.Moreover, the echo signals were obviously weakened upon a pre-irradiation of LFUS, which was reasonable because LFUS irradiation broke the aggregates to release PFB before the US imaging.These results indicated that our TME-responsive TLNs-Mix possesses great potential as a US imaging probe for solid tumors.

In vivo therapeutic effect
The in vivo antitumor effect of TLNs-Mix@C/R was assessed using a murine model bearing subcutaneous hypovascular hepatocellular carcinoma, [46][47][48] following the therapeutic schedule shown in Figure 5A.Mice bearing Hepa1-6 allografts randomly divided into four groups received different treatments of phosphate-buffered saline (PBS), TLNs-1@C/R + LFUS, TLNs-2@C/R + LFUS, and TLNs-Mix@C/R + LFUS.In comparison with the unpaired TLNs-1@C/R + LFUS treatment and the TLNs-2 + LFUS treatment, the TLNs-Mix@C/R + LFUS treatment suppressed tumor growth much more effectively, reaching a remarkable tumor growth inhibition of 90% on day 16 (Figure 5B,C).Moreover, the TLNs-Mix@C/R + LFUS treatment significantly prolonged the survival time of mice without causing obvious side effects (Figure 5D and Figure S9).Microscopic observation of tumor sections further confirmed the therapeutic effect of the TLNs-Mix + LFUS treatment.As shown in Figure 5E, the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay demonstrated that the TLNs-Mix@C/R + LFUS treatment resulted in the most prominent tumor cell apoptosis, showing abundant TUNEL-positive cells (green fluorescence) across the entire tumor region.By contrast, apoptotic cells were mostly found in the tumor periphery both in the TLNs-1@C/R + LFUS treatment group and the TLNs-2@C/R + LFUS treatment group.These results are in line with the intratumor drug distribution.When the TLNs-Mix formulation was administered, the TLNs-Mix efficiently entered the tumor and then aggregated in situ to enhance echogenicity, which enabled an LFUS-induced tumor-penetrating drug delivery to boost the anti-tumor effects in deep tumor tissue.The immunohistochemical analysis of nuclear protein Ki67, as a cellular proliferation indicator, and the hematoxylin and eosin (H&E) staining assay provided consistent results that the TLNs-Mix@C/R + LFUS treatment most effectively suppressed the tumor cell proliferation (Figure 5E).
The in vivo immune responses induced by various treatments were compared by detecting the mature DCs and infiltrated T cells in the tumor. [49,50]The TLNs-Mix@C/R + LFUS treatment resulted in the release of the highest levels of DAMPs as well as the highest proportion of DC maturation (about 50%) (Figure 5F and Figure S10).These results were in line with the previous reports that, with the aid of the TLR7 agonist R837, the DAMPs released upon SDT [51] or photothermal therapy [52] could effectively activate the DC maturation.Consequently, the TLNs-Mix@C/R + LFUS treatment group showed the highest proportion of CD8 + T cells (30%) in the total T cells (CD45 + CD3 + ), which meant a 1.3-fold increase and 2.5-fold increase as compared to the TLNs-1 + LFUS treatment group and the TLNs-2 + LFUS treatment group, respectively (Figure 5F).The tumor infiltration of cytotoxic T cells as a key point for antitumor immunity was investigated via immunohistochemical assay.As shown in Figure 5G, the TLNs-Mix@C/R + LFUS treatment promoted the tumor infiltration of CD8 + T cells more effectively than other treatments, showing more CD8 + T cells both in the peripheral and the center of tumors.Then, the expressions of multiple cytokines including IL-12, tumor necrosis factor-alpha (TNF-α), and IFN-γ were analyzed to assess the immune activation.Consistently, the TLNs-Mix@C/R + LFUS treatment resulted in the highest levels of IL-12, TNF-α, and IFN-γ (Figure 5H).
Finally, the therapeutic effect of nanodrugs is also related to adaptive immunity in secondary lymphoid organs.Since sentinel lymph nodes serve as the primary sites for antitumor response, the analysis of humoral and cellular immune responses was conducted. [53]As shown in Figure S11A, the percent of mature DCs and total T lymphocytes were increased in the TLNs-Mix@C/R + LFUS group, implying the improved antigen cross-presentation and T cells homing into the lymph nodes.Additionally, the obvious fluorescence signal of B220 and GL7 (germinal center) indicated the B cell activation and humoral immune response (Figure S11B). [54]Therefore, the remarkable therapeutic effect was attributed to the unique twins-like design of nanodrugs.In other words, the intravenously administered TLNs-Mix@C/R efficiently entered the tumor and then aggregated in situ.As a result, the local LFUS irradiation of the tumor-induced a strong cavitation effect for tumor-penetrating delivery of the sonosensitizer Ce6 and immune adjuvant R837, which effec-tively enhanced the immunosonodynamic therapy against the hypovascular hepatoma.

DISCUSSION
The efficient tumor internalization, retention, and tissue penetration of nanodrugs have been demonstrated critical for achieving potent anticancer therapy in recent years.However, achieving tumor accumulation and intratumoral penetration with nano-sized drug delivery systems poses significant challenges due to their contradictory requirements. [46,55]On one hand, small-sized nanodrugs possess great advantages in extravasating the leaky vasculature and penetrating the heterogeneous tissue of solid tumors, thus being favorable for improving the therapeutic effect.For instance, while the ideal particle size for tumor accumulation is under 100 nm, nanodrugs smaller than 30 nm are necessary for efficient penetration in tumor tissue, notably in poorly permeable hypovascular tumors such as pancreatic cancer and hepatocellular carcinoma. [20,47,48]On the other hand, the high IFP of solid tumors may force the small-sized nanodrugs back into blood circulation, which hinders tumor retention.Therefore, strategies for in situ TME-triggered aggregation of nanodrugs inside tumors, which enhances their tumor retention without detracting from their tumor internalization and penetration, have been intensively explored thus far, with some showing tremendous potential to improve the therapeutic efficacy. [14,56,57]In particular, microscale aggregates filled with PFCs possess high echo intensities, based on which both a sensitive US imaging and an LFUS-triggered cavitation effect for rapid release and tumor-penetrating delivery of drugs may be realized. [21,58,59]To date, hydrophobic interaction, [60] crosslinking upon chemical bonds, [61] DNA hybridization, [62] and temperature change [63] have been proposed as driving forces to make small-sized nanodrugs coalesce inside solid tumors.In addition, coulombic interaction has appeared as a facile solution for intratumor aggregation of nanodrugs. [64,65]Unfortunately, nanoparticles with different physicochemical properties (sizes and surface charges, and so on) usually show great differences in their blood circulation time, making an equal-opportunity tumor uptake for efficient intratumor aggregation impossible.This still remains a great hurdle for utilizing electrostatic interaction to assemble small-sized nanodrugs in situ.
Here, a formula of TLNs with the same physicochemical properties was developed to ensure their synchronous transport in blood circulation and tumor entrance.The surface charges of TLNs-Mix were transformed oppositely in situ under the enzyme digestion of MMP-2 in TME, that is, one surface exhibited a positive charge while the other had a negative charge, which would initiate electrostatic interaction to drive the nanodrugs aggregation in situ.Interestingly, the formation of large aggregates being ideally micron-sized (∼1400 nm), would not only increase the tumor retention of nanodrugs but also boost their echogenicity to enhance ultrasound imaging as well as LFUS-triggered drug release and penetration inside the tumor.In vitro and in vivo data proved that the in situ aggregation based on mildly coulombic electrostatic interaction led to an enhanced SDT and improved infiltration of cytotoxic T lymphocytes for a highly efficient cancer therapy.
To sum up, increasing tumor retention of nanodrugs still remains a tremendous challenge, which has driven great research efforts using electrostatic interaction for nanoparticle assembly inside tumors.Unfortunately, nanodrugs with different physicochemical properties usually show varied blood half-lives leading to nonequivalent tumor uptake for poor nanodrug coalescence inside the tumor.Through a simple means of changing the orientation of the MMP-2sensitive peptide linker between the micellar core and shell, we developed the TLNs as a new strategy to address the problems.

4.1
Preparation of nanoparticles loaded with Ce6, PFB, and R837 (TLNs-1@C/R or TLNs-2@C/R) R837 (5 mg), Ce6 (5 mg), and mPEG-Pep(+)(or Pep(−))-PAsp(MEA/DIP/PFA) (50 mg) were dissolved in dimethyl sulfoxide (DMSO, 0.5 mL).The solution was slowly dropped into PFB (200 μL) under sonication in an ice bath.Then the emulsions were slowly dropped into deionized water (DI water) (20 mL) under sonication in an ice bath again.After the large aggregates were removed by a syringe filter (pore size: 0.45 μm), the mixture was transferred into a dialysis bag (MWCO: 14 kDa) and dialyzed in phosphate-buffered saline (PBS, pH 7.4) at 4 • C for 12 h to remove the DMSO and free polymer.After that, the supernatant of the solution inside the dialysis bag was collected by discarding the unencapsulated PFB droplets at the bottom.Then the obtained solution was concentrated and washed three times with PBS (pH 7.4) via a MILLIPORE Centrifugal Filter Device (MWCO: 10 kDa), centrifuged at 500 rpm and 4 • C, and then passed through a syringe filter (pore size: 0.45 μm) to eliminate large aggregates, which finally yielded the solution of TLNs-1@C/R or TLNs-2@C/R.

Cell culture
A murine hepatoma cell line of Hepa1-6 was purchased from the American Type Culture Collection (ATCC).Hep1-6 cells were cultured in DMEM medium supplemented with 10% (v/v) FBS and 1 × penicillin-streptomycin.The cells were maintained in a humidified atmosphere of 5% CO 2 at 37 • C.

Cellular uptake
Hep1-6 cells were seeded at 2 × 10 4 cells/well in a confocal dish with a glass bottom and cultured overnight.The nanodrugs were labeled with Ce6 and added into the cells for 4 h in the presence of MMP-2.After being washed with PBS twice, the cells were treated with US irradiation.The cells were stained with 1 mM LysoTracker, fixed with 4% paraformaldehyde, and then counterstained with 4′,6diamidino-2-phenylindole (DAPI).The cellular uptake was observed under confocal laser scanning microscopy (CLSM; Nikon C2).

Penetration of TLNs@Ce6 in multicellular tumor spheroids
The tumor cells and L929 cells were seeded at 8 × 10 3 cells/well at a 1:3 ratio in a 96-well plate that was precoated with sterilized agarose.The cells were cultured in a DMEM complete medium for 2 days to form cell spheroids.Half of the medium was removed and replenished with fresh medium every 3 days after the formation of cell spheroids.When the spheroids grew up to about 1.2 mm, the multicellular tumor spheroids were incubated with TLNs@Ce6 (concentration of Ce6, 2 μM) overnight.MMP-2 and LFUS irradiation were performed if required.Then the multicellular tumor spheroids were transferred into the confocal dishes for CLSM observation.After CLSM observation, the corresponding multicellular tumor spheroids were transferred to a new 96-well plate.Trpsin-EDTA solution was used to detach the spheroids at 37 • C for 20 min.The single-cell suspension was obtained for FCM analysis.

Detection of ROS in vitro
The generation of ROS was determined using the FCM and CLSM.Briefly, Hepa1-6 cells were seeded at 1 × 10 6 cells/well in a 12-well plate overnight.The cells were incubated with various nanodrugs for 4 h and then received different treatments.The cells were washed three times with PBS and stained with an intracellular ROS dye 2′−7′dichlorofluorescin diacetate (DCFH-DA).The cells were collected and resuspended in PBS for FCM detection.
For CLSM observation, the cell nuclei were labeled with Hoechst solution.

Expression levels of calreticulin, ATP, and high mobility group B1 in vitro
Calreticulin expression was analyzed using immunocytochemistry and observed under CLSM.Hepa1-6 cells were seeded at 2 × 10 4 cells/well in a confocal dish with a glass bottom and cultured overnight.The cells were incubated with various samples for 4 h and then received different treatments.After being cultured for another 24 h, the cells were fixed with 4% paraformaldehyde, blocked, and incubated with the anti-calreticulin antibody at 4 • C overnight.The cells were stained with the FITC secondary antibody for 1 h at room temperature.Finally, the Actin-Tracker was used to stain the actin at a 1/100 dilution for 1 h at room temperature and DAPI was used to label the cell nuclei.The images were captured using the CLSM.The expression levels of ATP and HMGB1 were evaluated by enzyme-linked immunosorbent assay (ELISA).The cells were treated in a similar way, and the cell supernatants were collected for detection using an ATP ELISA Kit and HMGB1 ELISA Kit by following the manufacturer's protocol.

BMDCs maturation
The C57BL/6j mice were euthanized and sterilized with 70% ethanol.The hind leg was cut using the sterilized scis-sors and the surrounding muscles were removed.The bone marrow cells were isolated by inserting the 29G × ½ needle into the femur and tibia and flushing into the pre-cold RPMI-1640 medium.The cell slumps and red blood cells (RBC) were removed using a 70-μm cell strainer and RBC lysis buffer.After being rinsed by cold PBS, the cells were seeded in 12-well plates at a density of 10 × 10 6 cells/mL and cultured in RPMI-1640 supplemented with 10% (v/v) FBS, 1× penicillin-streptomycin and 20 ng/m granulocytemacrophage colony-stimulating factor (GM-CSF).On day 3, half of the medium was changed and the non-adherent cells were removed.At day 6, the non-adherent and loosely adherent cells were harvested for purity assessment using FCM (CD11c + cells).Note that, 60%-80% of CD11c-positive BMDCs were obtained using this method for subsequent stimulation.The BMDCs were incubated for 24 h with the supernatants after tumor cells received various treatments.
After being rinsed and the Fc receptor was blocked, the cells were stained with the DCs markers including CD11c, CD80, CD86, and MHC II for FCM analysis.After being gated on CD11c + cells, the DCs maturation was determined by the proportion of the population co-expressing CD80 and CD86, and the expression of MHC II was indicated in the histogram.

Animal model
The C57BL/6j mice (female, 4-6 weeks of age, 16-18 g) were purchased from the Guangdong Medical Laboratory Animal Center.All experimental procedures involving animals were performed under the Guidelines for Care and Use of Laboratory Animals of Sun Yat-sen University (SYXK 2016-0112).A murine hepatoma cell line of Hep1-6 was used to establish the allograft mouse model.Briefly, 1 × 10 6 Hep1-6 cells in 100 μL of PBS were subcutaneously transplanted into the right flank of the mouse.The length (L) and width (W) of the tumor were measured using a caliper and the tumor volume was calculated by the formula: Volume = 0.5 × L × W 2 .

In vivo tumor accumulation and tissue-penetrating drug delivery
For in vivo fluorescence imaging, different nanodrugs were labeled with the near-infrared fluorescent cyanine dye DiR to form DiR/TLNs-1@C/R, DiR/TLNs-2@C/R and DiR/TLNs-Mix@C/R.The tumor-bearing mice were injected with nanodrug solutions via the tail vein.The in vivo fluorescence imaging was performed at various time points.
According to the distribution profile in vivo, 8 h after the intravenous (i.v.) injection of nanodrugs was chosen for studying the US responsiveness.After the tumor sites were exposed to the LFUS irradiation (2 MHz, 2.0 W/cm 2 , 20% duty cycle), [66] the changes in fluorescence intensities were detected at 10 min and 1 h after LFUS irradiation.For in vivo US imaging, the clinical used contrast agent SonoVue and TLNs-Mix@C/R nanodrugs were i.v.administered to the tumor-bearing mice, respectively.The US images were acquired with or without irradiation using clinical US equipment (GE Logiq E9).The tumor-penetrating drug delivery was assessed by the immunofluorescent (IF) analysis.
TLNs-Mix@C/R was i.v.administered to the tumor-bearing mice.The tumor tissues were harvested at 8 h after nanodrug treatments (with or without LFUS) embedded in Tissue-Tek O.C.T. Compound and frozen at −80 • C. The specimens were cut into 10 μm-thick sections and mounted onto poly-L-lysine coated slides.After being fixed, permeabilized, and blocked, the tissue sections were incubated with primary antibody (anti-CD31 antibody) at 4 • C overnight.After being rinsed, the sections were labeled with FITC secondary antibody, followed by counterstaining with DAPI.After staining, the sections were mounted onto the slides for CLSM observation.

Antitumor effect in vivo
The Hep1-6 tumor-bearing mice were randomly divided into four groups.When the tumor volume reached about 50 mm 3 , the four groups of mice received the treatments of PBS, TLNs-1@C/R + LFUS, TLNs-2@C/R + LFUS, and TLNs-Mix@C/R + LFUS, respectively.The dose of nanodrug was 10 mg/kg body weight.The LFUS irradiation (2 MHz, 2.0 W/cm 2 , 20% duty cycle) was conducted at 8 h after i.v.injection of nanodrugs. [66]The treatment was performed every 3 days five times.During the treatment, the tumor volume and body weight were recorded and analyzed.After the treatment, the tumor tissues of animals were excised for histology analysis and FCM detection.

Evaluation of immune cells within tumors and lymph nodes
The excised tumor tissues were placed in ice-cold PBS.The tumor tissues were mechanically dissociated using a scalpel and passed through a 70-μm cell strainer to prepare the single-cell suspension.The obtained cell suspensions were counted, blocked for the Fc receptors, and stained with Zombie Viability dye.Without washing, the cell suspensions were subjected to extracellular staining with an antibody cocktail (anti-CD45/CD3/CD4/CD8 cocktail for T lymphocyte subset; anti-CD45/CD11c/CD80/CD86 cocktail for DCs maturation) for 30 min at 4 • C in the dark.The unbound antibody was removed by centrifugation at 350× g for 5 min.The cells were resuspended in 400 μL of FACS staining buffer for FCM analysis.Slimily, the sentinel lymph nodes were harvested and ground.The single-cell suspension was stained with an antibody cocktail for T lymphocytes and DCs.

Histological analysis
The tumor tissues of animals receiving various treatments were excised, fixed with 4% paraformaldehyde, dehydrated, embedded in paraffin, and sectioned using a microtome.After the paraffin sections (5-μm thickness) were prepared, staining was performed.The hematoxylin and eosin (H&E) staining was carried out.Briefly, the paraffin sections were deparaffinized in xylene, re-hydrated in graded ethanol (100%, 95%, and 80%, two washes 10 min each), washed with deionized water, and stained with hematoxylin and eosin.After being rinsed, the sections were dehydrated and mounted onto the slides.The terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was performed following the manufacturer's protocol.For the immunohistochemical assay, the paraffin sections were pretreated with 3% hydrogen peroxide and heat-induced epitope retrieval (HIER, pH 6.0).Permeabilization (0.1% Triton X-100 in PBS, 10 min) and non-specific blocking (5% BSA in PBS, 30 min) were performed prior to antibody incubation.Without washing, sections were incubated with corresponding primary antibodies (Ki-67, CD8, and calreticulin) at 4 • C overnight, followed by incubation with the conjugated secondary antibody.Finally, the sections were counterstained with DAPI for fluorescence detection.For chromogenic detection, the sections were incubated with a diaminobenzidine (brown) working solution and counterstained with hematoxylin.For cryosection of the lymph node, the obtained frozen section was fixed with 4% PFA and blocked with 5% BSA.Then the section was stained with Alexa Fluor 647-labeled anti-mouse GL7 antibody and Alexa Fluor 488-labeled anti-mouse B220 antibody for the germinal center and B cells, respectively.

F I G U R E 4
In vivo tumor accumulation of nanodrugs, low-frequency ultrasound (LFUS)-triggered penetrative drug delivery and ultrasound imaging.(A) In vivo fluorescence imaging of mice bearing Hep1-6 allograft at different time points after tail vein injection of DiR-labeled twins-like nanodrugs (TLNs)-Mix@C/R.(B) In vivo fluorescence imaging of mice bearing Hep1-6 tumor at 8 h after tail vein injection of DiR-labeled TLNs-Mix@C/R with or without LFUS irradiation.(C) Quantification of DiR fluorescence intensities in tumor sites of mice based on Figure 4B.(D) Immunofluorescence images of Hep1-6 tumor tissue sections from mice after tail vein injection of TLNs-1@C/R, TLNs-2@C/R, and TLNs-Mix@C/R, labeled with Nile Red (red fluorescence).LFUS irradiation of mouse tumor sites was conducted at 8 h after injection.Tumor blood vessels were stained with FITC-CD31 (green fluorescence) and nuclei were stained with DAPI (blue fluorescence).Scale bar, 25 μm.(E) Quantitative analysis of the DiR fluorescence intensities along the yellow dotted lines shown in Figure 4D for different groups.(F) Schematic illustration of nanodrug aggregation in matrix metalloproteinase-2 (MMP-2)-enriched tumor and LFUS-assisted drug release and penetrative drug delivery.(G) US images showing tumor sites (white circles) of mice intravenously injected with SonoVue and TLNs-Mix@C/R.Left: B-mode.Right: Contrast mode.

F I G U R E 5
In vivo therapeutic effects and antitumor immune response.(A) Therapeutic schedule of Hep1-6 tumor-bearing mice receiving treatments of different nanodrugs.Low-frequency ultrasound (LFUS) irradiation of tumor sites (2 MHz, 2.0 W/cm 2 , 20% duty cycle) was conducted for 10 min at 8 h after intravenous injection of twins-like nanodrugs (TLNs)-1@C/R, TLNs-2@C/R or TLNs-Mix@C/R.(B) Tumor growth during the treatments (n = 5; ***p < 0.001).(C) Tumor weight at the end of the treatment (n = 5; ***p < 0.001).(D) Cumulative survival of mice in different treatment groups (n = 5).(E) Immunohistochemical and histological analyses for terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), Ki67 expression, and hematoxylin and eosin (H&E) of tumor tissues from mice receiving different treatments.For the TUNEL assay, the whole sample images and magnified images (yellow rectangles) were captured by a confocal laser scanning microscope (CLSM).TUNEL-positive cells were labeled with FITC secondary antibody (green fluorescence) and nuclei were stained with DAPI (blue fluorescence).Ki67 (middle row, brown) and H&E (bottom row, purple and pink) results were recorded by the optical microscope.(F) FCM analysis of DCs maturation (gated on lived CD45 + CD11c + cells) and tumor-infiltrating lymphocytes (gated on lived CD45 + CD3 + cells) within tumors in different groups.(G) Immunohistochemical assay showing the CD8 + T cells (brown) in the peripheral and center sites of the tumor.The arrows indicate the CD8 + positive cells.Scale bar, 100 μm.(H) Expression levels of IL-12, TNF-α, and IFN-γ in tumors determined by ELISA.Data represents mean ± SD (n = 5; **p < 0.01, ***p < 0.001).
Characteristics of the synthesized block copolymers.
TA B L E 1