INCB024360

Tumor microenvironment-responsive prodrug nanoplatform via co-self- assembly of photothermal agent and IDO inhibitor for enhanced tumor penetration and cancer immunotherapy

Abstract

Nanomedicine-based phototherapy in combination with immune checkpoint blockade therapy has been reported as a promising strategy for improved cancer immunotherapy. However, tumor penetration of nanomedicine into solid tumor is still an unresolved obstacle to an effective drug delivery, leading to limitations in their applica- tions. Here, we developed a tumor microenvironment-responsive prodrug nanoplatform for efficient penetration and photo-immunotherapy of cancer. The prodrug nanoplatform is performed by integrating PEGylated in- doleamine-2,3-dioxygenase (IDO) inhibitor (Epacadostat) and photosensitizer (Indocyanine green, ICG) into a core-shell nanostructure via intermolecular interactions, which can transform into small dual-drug complexes (< 40 nm) at tumor microenvironment. The resulting small dual-drug complexes could undergo caveolae- mediated endocytosis, enhance cellular uptake, directly kill tumor cells, in situ trigger antitumor immune re- sponse and modulate IDO-mediated immunosuppression. More significantly, the prodrug nanoplatform in combination with PD-L1 checkpoint blockade synergistically promoted the antitumor immunity and efficiently inhibited the growth of both primary and abscopal tumors. The present study provides a novel delivery strategy for nanoenabled phototherapy and IDO inhibition to combine PD-L1 checkpoint blockade for achieving more effective therapy of solid tumors. 1. Introduction Immune checkpoint blockade (ICB) therapy that aims to eliminate or inhibit immune suppressive factors on immune cell activation, such as CTLA-4 (cytotoxic T lymphocyte antigen 4) and PD-1/PD-L1 (pro- grammed cell death protein 1/programmed cell death-ligand 1) blockades or indoleamine-2,3-dioxygenase (IDO) inhibition, has de- monstrated clinical activity and becomes mainstream strategy to treat cancer [1–4]. Despite the considerable success in clinic, ICB therapy is not generally curative and often limited to a fraction of cancer patients [5]. The “hot” immunogenic tumor microenvironment (TME) and suf- ficient T cell infiltration are critical for an effective ICB therapy [6–8]. Therefore, ICB therapy in combination with other types of therapeutic strategy to modulate both immunostimulation and immunosuppression within TME may hold strong potential to improve cancer therapy. Photoenabled cancer ablation treatments, such as photodynamic therapy (PDT) and photothermal therapy (PTT), could cause immunogenic cell death (ICD) of tumor cells, thus inducing the antigen- specific antitumor immune response [9,10]. Such an “in situ vaccine” strategy could be used to stimulate a “cold” TME to become a “hot” immunogenic TME [11,12]. Recent studies by several groups have de- monstrated that nanomedicine-based phototherapy in combination with ICB therapy such as IDO inhibition, PD-1/PD-L1 or CTLA-4 blockade has a synergetic effect on inducing systemic antitumor im- munity and preventing cancer metastasis in animal models [13–16]. Compared with free photosensitizer and/or free immune checkpoint blockade molecules, these nanomedicines can improve the physiolo- gical properties of drugs, prolong blood circulation time, and enhance tumor accumulation and bioavailability. Recently, a phase 3 trial for melanoma treatment revealed the fact that the uselessness of IDO in- hibition with Epacadostat as a strategy to enhance PD-1 blockade therapy activity [17]. It was suggested that the clinical application of IDO inhibitors, such as Epacadostat or NLG919, etc. Is hindered by their high lipophilicity, limited bioavailability, and insufficient drug exposure within the tumor [18,19]. Many nanoparticle formulations of IDO inhibitors, including our group's work, have been reported to overcome these limitations [20–22]. Additionally, due to the dense extracellular matrix and high interstitial fluid pressure of 3D solid tu- mors, the limited deep penetration of nanomedicine is still great chal- lenge for thorough clearance of tumor cells in the core tumor regions, possibly resulting in unsatisfactory clinical treatment outcomes [23–25]. It is well known that the size of nanoparticles plays an important role in their penetration abilities inside tumor issues [26–29]. Nano- particles with smaller size can penetrate more deeply into 3D solid tumors. Thus, rational design of nanomedicine for effective tumor pe- netration and modulation of TME would be a promising strategy for highly effective photo-immunotherapy of cancer. Herein, our strategy is to develop a TME-responsive prodrug nanoplatform featured with good biocompatibility, prolonged circulation, enhanced tumor accumulation and deep penetration, all of which could lead to improved photo-im- munotherapeutic outcomes compared with conventional nanomedi- cines (Scheme 1). The prodrug nanoplatform is performed by PEGylated IDO inhibitor (IDOi, Epacadostat) and an amphiphilic photosensitizer (Indocyanine green, ICG), which can be co-self-assembled into a core- shell nanostructure via intermolecular interactions. Meanwhile, poly (ethylene glycol) (PEG) was conjugated to IDOi via the peptide se- quence PVGLIG, which can be selectively cleaved by matrix metallo- proteinase-2 (MMP-2) in TME [30,31]. The prodrug nanoplatform with the PEG layer could achieve high accumulation at tumor sites by the enhanced permeability and retention (EPR) effect. Upon reaching tumor sites, the PEG layer would be stripped by MMP-2 in TME. In- terestingly, smaller-sized particulate aggregates (IDOi/ICG NPs, < 40 nm) could be released from the prodrug nanoplatform with large initial size (~140 nm) owing to the change of intermolecular interactions between IDOi and ICG due to the shedding of PEG shell, which will facilitate deep tumor penetration. Finally, under near infrared fluores- cence (NIR) laser irradiation, the exposed smaller-sized IDOi/ICG NPs- based phototherapy could directly kill tumor cells and produce the tumor-associated antigens to trigger antitumor immunity. Meanwhile, IDOi would be also released in tumor sites and could further modulate T cell response in TME. The combination of phototherapy-based tumor cells killing, IDO inhibition and PD-L1 blockade at the tumor sites would synergitically promote the antitumor immunity, inhibit the growth of the irradiated tumors and the distant tumors without pho- totherapy treatment. Therefore, the strategy we developed not only performs the nanomedicine-based combination of phototherapy and ICB therapy but also provides a promising approach for enhanced photo-immunotherapeutic delivery. 2. Experimental sections 2.1. Synthesis and characterization of MMP-sensitive PEGylated IDOi prodrug MMP-sensitive PEGylated IDOi prodrug (mPEG-Pep-IDOi) was synthesized by a two-step reaction as shown in Supplementary Scheme 1. MMP-insensitive IDOi prodrug (mPEG-IDOi) was also synthesized as a control. Details of synthesis methods were shown in Supplementary Materials. 1H NMR, FTIR and matrix-assisted laser desorption and io- nization time-of-flight mass spectrometry (MALDI-TOF MS) were ap- plied to characterize the obtained product. Ultraviolet–visible (UV) spectrophotometer was used to measure the drug loading efficiency of IDOi. 2.2. Preparation and characterization of the prepared nanoparticles (NPs) To prepare mPEG-Pep-IDOi/ICG NPs, 6 mg of mPEG-Pep-IDOi and 0.25 mg of ICG were dissolved in a mixed solvent of chloroform and methanol (9:1, V/V) in a round-bottomed flask. The mixture was dried using a rotary evaporator, hydrated in distilled water for 3 h, and then sonicated over an ice bath for 5 min to obtain mPEG-Pep-IDOi/ICG NPs. The collected NPs were washed with water for three times and then freeze-dried for further use. mPEG-Pep-IDOi NPs and mPEG-IDOi/ICG NPs were prepared using the same procedure mentioned above with or without the addition of ICG. The morphology of the prepared NPs was visualized by using transmission electron microscopy (TEM) (Tecnai-F20, FEI, Netherlands). The hydrodynamic diameters of the prepared NPs were determined using dynamic light scattering (DLS) (A 90Plus particle size, Brookhaven). Fluorescence spectrum and ultraviolet, visible and NIR (UV–vis–NIR) absorption spectrum were recorded with a ThermoFisher Scientific Varioskan™ (Thermo Fisher Scientific, USA). The in vitro stability was evaluated by incubating NPs in PBS containing 10% fetal bovine serum through monitoring the change of size distribution at each time interval. To investigate MMP-2-induced transition of morphology, mPEG- Pep-IDOi NPs were incubated with MMP-2, followed by the observation with TEM and the analysis with DLS. The optical propertity of mPEG- Pep-IDOi NPs after incubation with MMP-2 was investigated by UV–vis–NIR range absorption spectra and fluorescence spectra. In vitro release of ICG and IDOi from mPEG-Pep-IDOi/ICG NPs and mPEG- IDOi/ICG NPs in PBS at pH7.4 with or without MMP-2 was also de- termined. 2.3. In vitro cytotoxicity, calreticulin (CRT) expression and dendritic cells (DCs) activation The in vitro cytotoxicity was evaluated in B16–F10 melanoma cancer cells (National Infrastructure of cell line Resource). The cells were cultured under standard medium and conditions. To evaluate MMP-2-activated cytotoxicity, mPEG-Pep-IDOi/ICG NPs and mPEG- IDOi/ICG NPs were pre-incubated with or without the addition of MMP-2. The cells were seeded into 96-well plate (1 × 104 cells/well) and then incubated with free ICG, mPEG-Pep-IDOi/ICG NPs or mPEG- IDOi/ICG NPs at the same ICG concentration of 20 μg/mL. After in- cubation for 6 h, the cells were replaced with fresh medium and ex- posed to a 1 W/cm 2 808 nm laser every 60 s for 5 min. After further incubation for 24 h, cytotoxicity was examined by a CCK-8 assay kit. For visualization of CRT expression, B16–F10 cells were incubated with free IDOi, free ICG and mPEG-Pep-IDOi/ICG NPs with/without the preincubation with MMP-2 for 6 h and then treated with NIR laser ir- radiation or not as described above. After various treatments, cells were stained with Alexa Fluor 488-CRT antibody and DAPI, and then ob- served using confocal laser scanning microscopy (CLSM). For in vitro study of DCs activation, bone marrow-derived dendritic cells (BMDCs) derived from C57BL/6 mouse were obtained as the established proto- cols [32]. BMDCs were treated with different formulations as men- tioned above for 24 h. BMDCs were then washed with PBS and labeled with fluorescent dye-labeled antibodies against CD11c, CD86 and CD80, and analyzed by flow cytometry. 2.4. Intracellular uptake For confocal fluorescence imaging, B16–F10 cells were treated with free ICG or mPEG-Pep-IDOi/ICG NPs with/without the preincubation with MMP-2 for 6 h. After washing with PBS, cells were labeled with DAPI to identify nucleus and observed by CLSM (Zeiss710, Jena, Germany) using a 40 × oil immersion objective (zoom 1.5) with ima- ging software ZEN 2008 (633 nm excitation and 730–797 nm emission for ICG). For the quantitative assay, B16–F10 cells were treated with free ICG, free IDOi or mPEG-Pep-IDOi/ICG NPs with/without the pre- incubation with MMP-2 for 6 h. For intracellular ICG content de- termination, cells were harvested and analyzed by flow cytometry. For intracellular IDOi content determination, cells were harvested, and lysed with 0.1 N NaOH. Then, acetonitrile was added, vortexed, and centrifuged (15,000 rpm, 5 min). The supernatant was collected and analyzed using HPLC method [33]. 2.5. In vitro tumor penetration B16–F10 multicellular spheroids were prepared using the liquid overlay method as previously described [34]. Briefly, cells were seeded into 2% agarose precoated 96-well plates at a density of 0.4 × 104 cells per well and incubated for 4 days. Then, the formed multicellular spheroids were treated with mPEG-Pep-IDOi/ICG NPs and mPEG-IDOi/ ICG NPs for 2 h or 24 h, respectively. Finally, B16–F10 multicellular spheroids were washed by PBS and observed by CLSM. 2.6. In vitro IDO inhibitory effect B16–F10 cells were seeded into 96-well plate (1 × 104 cells/well) and then incubated with free IDOi, mPEG-Pep-IDOi/ICG NPs or mPEG- IDOi/ICG NPs at the same IDOi concentration of 7.5 μg/mL. To evaluate MMP-2-activated IDO inhibition, mPEG-Pep-IDOi/ICG NPs and mPEG- IDOi/ICG NPs were pre-incubated with or without the addition of MMP-2. At the same time, recombinant murine IFN-γ (50 ng/mL) was added to stimulated the IDO expression. After further incubation for 24 h, the culture supernatant from each well was collected. The amounts of kynurenine in the culture supernatant were determined using mice kynurenine ELISA kits. 2.7. In vivo biodistribution and pharmacokinetic profile B16–F10 (1 × 106 cells/mouse) were inoculated subcutaneously into the right flank of female BALB/c mice. When the tumor reached 200 mm3, the mice were injected intravenously with free ICG, mPEG- Pep-IDOi/ICG NPs and mPEG-IDOi/ICG NPs at an identical ICG dose of 4 mg/kg. In vivo fluorescence imaging at various subsequent time points were recorded using Maestro imaging system (CRI, USA). For the quantitative biodistribution study of IDOi and ICG, the B16–F10 tumor-bearing female C57BL/6 mice were injected via the tail vein with free ICG, free IDOi, mPEG-Pep-IDOi/ICG NPs and mPEG- IDOi/ICG NPs at an identical ICG dose of 4 mg/kg and IDOi dose of 5 mg/kg (n = 3 per group). The concentration of ICG and IDOi in the tumor tissue and major organs at various subsequent time points was determined using fluorescent spectrophotometry and HPLC method, respectively [33]. To investigate the pharmacokinetics profile of the prepared nano- particles, Sprague Dawley (SD) rats were randomly grouped (n = 3 per group) and intravenously administrated with free ICG, free IDOi, mPEG-Pep-IDOi/ICG NPs and mPEG-IDOi/ICG NPs at an identical ICG dose of 4 mg/kg and IDOi dose of 5 mg/kg, respectively. Blood samples were collected via the fossa orbitalis vein at different time intervals post injection. The blood ICG and IDOi concentration were then determined using fluorescent spectrophotometry and HPLC method, respectively [33]. 2.8. In vivo tumor penetration B16–F10 tumor-bearing female C57BL/6 mice were injected via the tail vein with mPEG-Pep-IDOi/ICG NPs and mPEG-IDOi/ICG NPs at an identical ICG dose of 4 mg/kg. After treatment for 8 h, the mice were then injected intravenously with 100 μL of FITC-lectin (1 mg/mL) and euthanized 5 min post injection. Tumor tissues were excised, and conducted with cryostat section. Sections were subsequently stained with DAPI and observed using CLSM. 2.9. In vivo maturation of dendritic cells (DCs) induced by phototherapy treatment B16–F10 tumor-bearing female C57BL/6 mice were injected via the tail vein with free ICG, free FITC-IDOi, mPEG-Pep-IDOi/ICG NPs and mPEG-IDOi/ICG NPs at an identical ICG dose of 4 mg/kg and IDOi dose of 5 mg/kg (n = 3 per group). Tumors were irradiated with an 808 nm NIR laser at 1W/cm2 4 h post injection. Infrared thermographic maps were conducted during the 5 min laser irradiation using a FLIR Ax5 camera. To investigate phototherapy treatment-induced DCs matura- tion, the inguinal lymph nodes near the tumor sites were collected 24 h post-NIR laser irradiation for determination of the expression of CD80 and CD86 on DCs using flow cytometry. 2.10. Therapeutic evaluation in vivo To investigate in vivo antitumor effect of the prepared NPs, B16–F10 bearing mice model was established as follows: 1 × 106 B16–F10 cells were inoculated subcutaneously into the left flank (primary tumors) and 0.5 × 106 B16–F10 cells into the right flank (abscopal tumors) of each female C57BL/6 mice, respectively. When the primary tumors grew to about ~60 mm3, the mice were treated with PBS or mPEG-Pep- IDOi/ICG NPs an identical ICG dose of 4 mg/kg and IDOi dose of 5 mg/ kg by intravenous injection on day 0 and 6 (n = 5 per group). For phototherapy-treated group, the primary tumors were treated with laser irradiation for 4 min as described above and the abscopal tumors were without direct NIR irradiation. Anti-PD-L1 antibody was in- traperitoneally injected into mice at a dose of 10 mg/kg for three times every 3 days for the combinational immunotherapy. Tumour volume of both primary tumors and abscopal tumors was monitored every two days. To study immune cells in the abscopal tumors, tumor tissues of each group were ground into cell suspension. The obtained cells were stained with fluorescein-labeled antibodies: CD3, CD4, CD8a and Foxp3. Blood was collected to analyze cytokines using a kit (LEGENDplex™ Mouse Th1 Panel) via flow cytometry. In addition, major organs were har- vested and conducted with cryostat section. After hematoxylin eosin (H &E) staining, the sections were examined under a digital microscope. 2.11. Statistical analysis All data were expressed as means ± standard deviation (SD) or means ± standard errors (SEM). All figures shown in this article were obtained from at least three independent experiments (as detailed throughout the paper) with similar results. Statistical analyses between multiple groups were performed with the one-way ANOVA followed by Tukey's post-test or two-way ANOVA analysis followed by Bonferroni post-test (GraphPad Prism 5.0, GraphPad software, CA, USA). Statistical significance is denoted by p < 0.05, p < 0.01 and p < 0.001. 3. Results and discussion 3.1. Characterization of mPEG-Pep-IDOi/ICG NPs To prepare the TME-responsive prodrug nanoparticles, a MMP- sensitive PEGylated IDOi prodrug mPEG-PVGLIG-IDOi (mPEG-Pep- IDOi) was first synthesized by condensation reaction as shown in Supplementary Scheme 1. In this work, MMP-insensitive IDOi prodrug (mPEG-IDOi) was also synthesized as a control when the mPEG- PVGLIG-COOH was replaced by mPEG-COOH. The chemical structure of mPEG-Pep-IDOi was confirmed by FTIR, UV absorption spectra and 1H NMR (Supplementary Fig. 1 - Fig.3). The mass of mPEG-Pep and mPEG-Pep-IDOi was measured by MALDI-TOF MS. As shown in Supplementary Fig. 4, the predicted molecular weight of mPEG-Pep- IDOi is 5774.6 Da, indicating IDOi moiety was successfully conjugated to mPEG-Pep. In this study, amphiphilic mPEG-Pep-IDOi can self-assemble into nanoparticles using thin-film hydration method. ICG was further en- capsulated into mPEG-Pep-IDOi nanoparticles by molecular interac- tions between mPEG-Pep-IDOi and ICG for the construction of the prodrug nanoplatform (mPEG-Pep-IDOi/ICG NPs). From TEM images in Fig. 1a and b, we can see that both mPEG-Pep-IDOi NPs and mPEG-Pep- IDOi/ICG NPs showed a clear shell-core structure with an average size of about 140 nm (Supplementary Fig. 5). mPEG-Pep-IDOi/ICG NPs exhibited good size stability in 10% serum PBS buffer solution over a period of 48 h (Fig. 1c). The drug loading of IDOi and ICG for mPEG- Pep-IDOi/ICG NPs was 51.267 ± 1.733 μg/mg and 38.304 ± 2.562 μg/mg as calculated by UV detection, respectively. The molecular interaction between mPEG-Pep-IDOi and ICG in mPEG-Pep-IDOi/ICG NPs was firstly confirmed by UV–vis–NIR range absorption spectra of free IDOi, free ICG and mPEG-Pep-IDOi/ICG NPs. Compared with free IDOi and free ICG, the peak of mPEG-Pep-IDOi/ICG NPs appeared a red-shifted and broder adsoption band of ICG, together with a broder absorption band of IDOi, indicating that IDOi and ICG could interact with each other via hydrophobic and π–π stacking in- teractions (Fig. 1d) [35,36]. The obvious change in UV adsorption of mPEG-Pep-IDOi/ICG NPs further provided a possible evidence of the hydrophobic interaction between mPEG-Pep-IDOi and ICG after addi- tion of sodium dodecyl sulfate (SDS, 0.2% w/v) into to the mPEG-Pep- IDOi/ICG NPs aqueous solution (Supplementary Fig. 6a) [36,37]. Meanwhile, the electrostatic interactions were also confirmed by the disassembly of mPEG-Pep-IDOi/ICG NPs, which was reflected by the decreased turbidity of the suspension of the nanoparticles with the increasing addition of NaCl (Supplementary Fig. 6b) [38]. Moreover, as shown in Fig. 1e, the fluorescence emission of ICG in mPEG-Pep-IDOi/ ICG NPs was obviously decreased compared to free ICG, mainly due to the co-self-assembly of ICG into mPEG-Pep-IDOi/ICG NPs. These above results indicated that ICG could be successfully co-self-assembly into mPEG-Pep-IDOi/ICG NPs via the coordination of multiple inter- molecular interactions between mPEG-Pep-IDOi and ICG including electrostatic, hydrophobic, and π-π stacking interactions. As intermolecular interactions between mPEG-Pep-IDOi and ICG in mPEG-Pep-IDOi/ICG NPs played critical roles in nanoparticles in- tegrities, we suppose that the detachment of PEG layer may change the hydrophobic interactions in formed mPEG-Pep-IDOi/ICG NPs, thus re- sulting in partial mPEG-Pep-IDOi/ICG NPs dissociation and release of the small-sized paticles. To confirm the peptide cleavage-responsive- ness to MMP-2, morphology changes of mPEG-Pep-IDOi/ICG NPs were first investigated after incubation with MMP-2. TEM images in Fig. 1f showed that after incubation with MMP-2, the shell-core structure of mPEG-Pep-IDOi/ICG NPs almost completely disintegrated, and subse- quently exposed the small-sized particulate aggregates. Also, as shown by DLS analysis in Supplementary Fig. 7a, the small-sized nanoparticles (3–40 nm) were produced after incubation with MMP-2. By contrast, mPEG-IDOi/ICG NPs used as a nonresponsive control showed no changes after incubation with MMP-2, indicating that the production of the small-sized nanoparticles was caused by the cleavage of the MMP-2 substrate peptide (Supplementary Fig. 7b). In order to further evaluate MMP-2-sensitive detachment of PEG, mPEG-Pep-IDOi was analyzed after treatment with MMP-2 for 24 h by HPLC. Supplementary Fig. 8 shows that a peak at a retention time of 12.2 min appeared after the incubation of mPEG-Pep-IDOi with MMP-2, which was in accordance with free IDOi. However, the peak disappeared if mPEG-Pep-IDOi were incubated in the absence of MMP-2. All the results indicated PEG could be cleaved from mPEG-Pep-IDOi with the help of MMP-2. Meanwhile, the optical propertity of the exposed small-sized nano- particles was investigated by UV–vis–NIR range absorption spectra and fluorescence spectra (Fig. 1g and h). Compared with free ICG and free IDOi, the peak of the exposed small-sized nanoparticles appeared a red- shifted and broder adsoption band of ICG, together with a broder ab- sorption band of IDOi, indicating that IDOi and ICG in the small-sized nanoparticles (IDOi/ICG NPs) could interacte with each other via hy- drophobic and π–π stacking interactions. In addition, the fluorescence intensity of IDOi/ICG NPs also decreased compared to free ICG. When NaCl was added to the small drug aggregates aqueous sloution, changes in the turbidity of the solution was observed due to the additional electrostatic interactions by NaCl (Supplementary Fig. 9). All the above results indicated that the exposed smaller-sized IDOi/ICG NPs also formed via the intermolecular interactions between IDOi and ICG. In vitro drug release study showed that in the MMP-2 environment, mPEG- Pep-IDOi/ICG NPs exhibited an enhanced release behavior of both IDOi and ICG compared to that in PBS (Fig. 1i and j). However, both IDOi and ICG release from mPEG-IDOi/ICG NPs showed no changes after incubation with MMP-2. Thus, all the results indicated that the PEG corona of mPEG-Pep-IDOi/ICG NPs could be stripped after incubation with MMP-2, and subsequently exposed the small-sized IDOi/ICG NPs formed via the intermolecular interactions between IDOi and ICG. The small-sized IDOi/ICG NPs possibly penetrated more deeply into the whole tumor tissues and delivered drugs through synergistic effects. 3.2. In vitro tumor spheroids penetration and cellular uptake After confirmation of the MMP-2-induced cleavage of the PEG corona which resulted in the production of the smaller-sized IDOi/ICG NPs, we first evaluated the ability of intratumoral penetration of mPEG- Pep-IDOi/ICG NPs in multicellular spheroids derived from MMP-2 over- expression B16–F10 murine melanoma cells [39,40]. Multicellular spheroids were reported to contain an organized extracellular matrixs, which are more representative of tumors in vivo. As shown in Fig. 2a, the ICG fluorescence signal of mPEG-IDOi/ICG NPs mostly resided at the periphery of multicellular spheroids, and little fluorescence was observed in the center even after co-incubation for 24 h. In contrast, the ICG fluorescence signal of mPEG-Pep-IDOi/ICG NPs distributed throughout the multicellular spheroids only after 2 h of treatment and could be further increased after 24 h of incubation. The mean fluores- cence intensities of ICG in multicellular spheroids treated with mPEG- Pep-IDOi/ICG NPs were about 3 times higher than those of mPEG-IDOi/ ICG NPs at 24 h (Fig. 2b). These results indicated that the MMP-2-ac- tivated penetration of mPEG-Pep-IDOi/ICG NPs in multicellular spheroids were mainly contributed from the size reduction, suggesting the potential capability of mPEG-Pep-IDOi/ICG NPs to penetrate into the deeper parts of tumor tissue in vivo. We next investigated the cellular uptake of mPEG-Pep-IDOi/ICG NPs. For CLSM observation, B16–F10 cells were treated with free ICG and mPEG-Pep-IDOi/ICG NPs with or without the preincubation with MMP-2. CLSM results in Fig. 2c revealed that mPEG-Pep-IDOi/ICG NPs- treated cells showed enhanced fluorescence signals of ICG (red) com- pared to those incubated with free ICG. Furthermore, when mPEG-Pep- IDOi/ICG NPs were preincubated with the addition of MMP-2, dramatic increased fluorescence signals of red were observed. For quantitative assay, intracellular ICG content and IDOi content were determined using flow cytometry and HPLC method, respectively (Fig. 2d and e). Intracelluar IDOi content and ICG content in cells treated with mPEG- Pep-IDOi/ICG NPs preincubated with MMP-2 were approximated to 1.57- and 1.36-fold stronger than that treated with mPEG-Pep-IDOi/ ICG NPs without the treatment of MMP-2. These results showed that IDOi/ICG NPs, triggered by the PEG detachment of mPEG-Pep-IDOi/ ICG in the presence of MMP-2, could effectively enhance cell uptake, which was in agreement with previous studies that smaller nano- particles were internalized into cells more quickly and more efficiently than the larger nanoparticles [26–29]. We also studied the cellular uptake mechanism of mPEG-Pep-IDOi/ICG NPs using certain chemical blockers. As show in Supplementary Fig. 10, the cellular uptake of mPEG-Pep-IDOi/ICG NPs was significantly suppressed after treatment with filipin III, which suggests a caveolae-mediated endocytosis [41,42]. However, when mPEG-Pep-IDOi/ICG NPs were preincubated with the addition of MMP-2, the internalization of the smaller-sized IDOi/ICG NPs was significantly affected by the clathrin-mediated en- docytosis inhibitor (chlorpromazine). The different cellular inter- nalization pathways may be due to the different surface modification of mPEG-Pep-IDOi/ICG NPs and IDOi/ICG NPs. 3.3. In vitro phototherapeutic effect and IDO activity inhibition According the previous report, under NIR laser irradiation, ICG can generate reactive oxygen species (ROS) and local heat for PDT and PTT, respectively [43]. Here, in vitro ROS production and photothermal ef- fect of free ICG, mPEG-IDOi/ICG NPs and mPEG-Pep-IDOi/ICG NPs under 808 nm laser irradiation were evaluated. First, ROS production was detected using 1,3-diphenylisobenzofuran (DPBF) probe as an in- dicator. Supplementary Fig. 11 shows that the absorbance of DPBF at 425 nm reduced with the irradiation time, indicating the sufficient ROS generation produced by free ICG or ICG loaded nanoparticles. Ad- ditionally, the temperature change of free ICG, mPEG-IDOi/ICG NPs and mPEG-Pep-IDOi/ICG NPs under 808 nm laser irradiation was pre- sented in Supplementary Fig. 12. A final temperature elevation of 21.8 °C was observed in mPEG-IDOi/ICG NPs and mPEG-Pep-IDOi/ICG NPs groups, while free ICG group just showed an increase of 17.2 °C after laser irradiation for 5 min. All these results suggested that mPEG- Pep-IDOi/ICG NPs could effectively enhance the ROS production and improve the photothermal effect, which would facilitate a combination phototherapy of PTT and PDT for cancer. The MMP-2-activated phototherapy with mPEG-Pep-IDOi/ICG NPs were then studied in cells. We firstly measured the cytotoxicity of free ICG, mPEG-Pep-IDOi, mPEG-Pep-IDOi/ICG NPs and mPEG-IDOi/ICG NPs without light irradiation. It was found that all formulations-treated cells remained viable more than 80% under dark conditions after in- cubation for 24 h (Supplementary Fig. 13). For MMP-2-activated pho- totherapy treatment, B16–F10 cells were incubated with free ICG, mPEG-IDOi/ICG NPs and mPEG-Pep-IDOi/ICG NPs. The nanoparticles were preincubated with or without the addition of MMP-2. The all treated cells were then irradiated for different time at a power density of 1 W/cm2 (Fig. 2f and g). Compared to cells treated with free ICG, mPEG-IDOi/ICG NPs with/without MMP-2 incubation or mPEG-Pep- IDOi/ICG NPs, more effectively cytotoxicity with time-dependent in- crease was observed when they were incubated with mPEG-Pep-IDOi/ ICG NPs with MMP-2 treatment. For example, up to 97% cells death was induced by mPEG-Pep-IDOi/ICG NPs with MMP-2 treatment under laser irradiation for 4 min, which was much higher than that for free ICG (60%), mPEG-IDOi/ICG NPs (67%), mPEG-IDOi/ICG NPs (77%) with MMP-2 treatment and mPEG-Pep-IDOi/ICG NPs without MMP-2 treatment (83%), respectively. Then, the MMP-2-activated IDO inhibition in vitro with mPEG-Pep- IDOi/ICG NPs was tested in B16–F10 cells stimulated with IFN-γ to induce IDO expression [39,44]. The IDO enzyme activity was evaluated through determining the conversion of tryptophan to kynurenine in cells after treatment with different formulations. It was observed that mPEG-Pep-IDOi/ICG NPs effectively inhibited kynurenine production in B16–F10 cells (Fig. 2h). Moreover, a remarkable decrease of ky- nurenine production was observed when the cells were incubated with mPEG-Pep-IDOi/ICG NPs with MMP-2 treatment. In contrast, mPEG- IDOi/ICG NPs control group has a negligible effect on the inhibition of IDO activity after the preincubation of MMP-2. Therefore, all of the above results indicated that mPEG-Pep-IDOi/ICG NPs with the MMP-2- activated size reduction could be employed as a smart prodrug nano- platform with enhanced uptake efficiency, killing efficacy and IDO activity inhibition in cells. 3.4. In vitro DCs maturation It is known that PTT or PDT can induce immunogenic cell death (ICD), which is characterized by the expression of CRT on the surface of dying tumor cells. CRT acts as an “eat-me” signal for DCs uptake, thereby initiating an effective immune response [30]. To investigate whether mPEG-Pep-IDOi/ICG NPs-mediated phototherapy can induce ICD of tumor cells, we examined the expression of CRT on the surface of tumor cells after different treatments using CLSM. CLSM images showed that under NIR laser irradiation, free ICG slightly induced CRT exposure on the surface of B16–F10 cells membranes, while mPEG-Pep-IDOi/ICG NPs with/without MMP-2 treatment significantly promoted CRT expression (Fig. 3a). These results confirmed that all ICG formulations- mediated phototherapy would induce ICD of tumor cells. Since the elicitation of ICD is favorable for antitumor immune responses, DCs maturation was then measured by incubating BMDCs with different ICG formulations-pretreated B16–F10 cells with/without NIR laser irradia- tion using flow cytometry. Maturated DCs can exhibit an enhancement in the surface expression of costimulatory molecules CD80 and CD86. Fig. 3b shows that mPEG-Pep-IDOi/ICG NPs with MMP-2 preincubation plus NIR laser irradiation treated-B16-F10 cells induced significantly upregulation of CD80 and CD86 on BMDCs (63.6%), which was much higher than that for free ICG (56.1%) and mPEG-Pep-IDOi/ICG NPs without MMP-2 preincubation (58%) plus NIR laser irradiation, re- spectively. These results demonstrated that mPEG-Pep-IDOi/ICG NPs- mediated phototherapy could effectively induce ICD of tumor cells and promote the maturation of antigen presenting cells, which consequently facilitate the induction of antigen-specific antitumor immune reponses. 3.5. In vivo tumor accumulation, pharmacokinetics and tumor penetration of different formulations Furthermore, we investigated the effect of the MMP-responsive property on the in vivo tumor accumulation, pharmacokinetics and tumor penetration of mPEG-Pep-IDOi/ICG NPs. Tumor accumulation in vivo was first determined in the tumor bearing mice intravenously in- jected with free ICG, mPEG-IDOi/ICG NPs and mPEG-Pep-IDOi/ICG NPs with an identical IDOi dosage of 5 mg/kg and ICG dosage of 4 mg/ kg, respectively. Mice treated with mPEG-Pep-IDOi/ICG NPs showed the highest fluorescence signals at the tumor site over time, which reached maximum at 4 h post-injection and could continue to 12 h or even longer (Fig. 4a). To better demonstrate the tumor accumulation of mPEG-Pep-IDOi/ICG NPs, ex vivo quantitative biodistribution of ICG and IDOi was carried out in the major organs at different time points. As shown in Fig. 4b, mPEG-Pep-IDOi/ICG NPs group showed the highest accumulations of ICG in the tumor tissues compared to mPEG-IDOi/ICG NPs and free ICG groups at 1, 4 and 12 h post-injection. The in- tratumoral ICG distribution of the mPEG-Pep-IDOi/ICG NPs group was 6.38- and 1.65-fold higher than that of free ICG and mPEG-IDOi/ICG NPs groups at 4 h post-injection, respectively. mPEG-Pep-IDOi/ICG NPs also reduced the retention of ICG in normal tissues compared to the free ICG group. Similarly, mPEG-Pep-IDOi/ICG NPs group showed the highest accumulations of IDOi in the tumor tissues compared to mPEG- IDOi/ICG NPs and free IDOi groups at 1, 4 and 12 h post-injection (Fig. 4c and Supplementary Fig. 14). The intratumoral IDOi distribution of the mPEG-Pep-IDOi/ICG NPs group was 7.65- and 2.8-fold higher than that of free IDOi and mPEG-IDOi/ICG NPs groups at 4 h post- injection, respectively. Subsequently, the pharmacokinetics profiles of different formulations were studied using SD rats intravenously injected with different formulations with an identical IDOi dosage of 5 mg/kg and ICG dosage of 4 mg/kg, respectively (Fig. 4d and e). Compared to free IDOi and free ICG groups, both mPEG-Pep-IDOi/ICG NPs and mPEG-IDOi/ICG NPs showed prolonged blood circulation time in- dependently of their MMP2-responsive properties. Extending from the analysis of the tumor accumulation efficiency of mPEG-Pep-IDOi/ICG NPs, the tumor penetration ability of mPEG-Pep- IDOi/ICG NPs in vivo were then evaluated. Tomato lectin-FITC was used to label blood vessels by binding to glycoproteins of endothelial plas- malemma [45,46]. In the mPEG-IDOi/ICG NPs group, the red (blood vessels) and green (nanoparticles) fluorescence signals almost over- lapped (Fig. 4f). On the contrary, in the mPEG-Pep-IDOi/ICG NPs group, more green signals of nanoparticles were dispersed surrounding the red signals of blood vessels and could reach the deep location of tumor. This result indicated that the mPEG-Pep-IDOi/ICG NPs could penetrate deeply into tumor tissue. Collectively, the mPEG-Pep-IDOi/ ICG NPs with MMP-2-responsive properties showed the prolonged cir- culation time due to the stealth effect induced by PEGylation, the highest tumor accumulation and the improved tumor penetration in vivo owing to EPR effect and size reduction. All the above results de- monstrated that size reduction of mPEG-Pep-IDOi/ICG NPs as a result of MMP-2-triggered cleavage of the PEG corona in the tumor environment would certainly benefit enhancing therapeutic efficacy in vivo. 3.6. mPEG-Pep-IDOi/ICG NPs-mediated phototherapy promotes the maturation of DCs in vivo It has been reported that phototherapy (PTT and PDT) is able to induce tumor-specific immune responses by producing tumor-asso- ciated antigens in the ablated tumor site [9,10]. Therefore, we wonder whether mPEG-Pep-IDOi/ICG NPs-meidated phototherapy would be able to trigger effective immunological responses in vivo. Mice-bearing subcutaneous B16–F10 tumor cells were intravenously injected with PBS, free ICG, mPEG-IDOi/ICG NPs or mPEG-Pep-IDOi/ICG NPs (5 mg/ kg IDOi, 4 mg/kg ICG) and then treated with laser irradiation at 4 h post-injection. Temperature changes in the irradiated areas were re- corded during the 5 min laser irradiation. The tumor temperature of mice treated with mPEG-Pep-IDOi/ICG NPs had a maximum tempera- ture of ~60 °C, significantly higher than mPEG-IDOi/ICG NPs group (~53 °C) and free ICG group (~46.5 °C) (Fig. 5a and b). The higher tumor temperature observed in mPEG-Pep-IDOi/ICG NPs may be due to the effective accumulation and penetration of mPEG-Pep-IDOi/ICG NPs in tumor regions, which is consistent with previous in vivo distribution results. We further investigated mPEG-Pep-IDOi/ICG NPs-meidated phototherapy treatment-induced DC maturation. After 24 h post-NIR laser irradiation, the inguinal lymph nodes near the tumor sites were collected for determination of the expression of the co-stimulatory molecules CD80 and CD86 on DCs using flow cytometry. mPEG-Pep- IDOi/ICG NPs-meidated phototherapy treatment could significantly increased the expression of both CD80 and CD86, which was 2-fold more efficiently compared to that resulted from mPEG-Pep-IDOi/ICG NPs without NIR laser irradiation (Fig. 5c–e). Moreover, phototherapy treatment with mPEG-Pep-IDOi/ICG NPs could greatly promote DCs maturation to a level much higher than that induced by phototherapy treatment with mPEG-IDOi/ICG NPs, likely due to the enhanced tumor- specific accumulation and tumor cell uptake. Therefore, mPEG-Pep- IDOi/ICG NPs-medaited phototherapy treatment could promote DCs maturation and trigger the antitumor immune responses. 3.7. Effective antitumor immunity by phototherapy treatment with mPEG- Pep-IDOi/ICG NPs plus PD-L1 blockade Encouraged by the superior DC maturation and tumor penetration mediated by mPEG-Pep-IDOi/ICG NPs, we further accessed the anti- tumor efficacy of mPEG-Pep-IDOi/ICG NPs-based phototherapy in combination with PD-L1 blockade in a bilateral B16–F10 tumor model. Tumors on the left side designed as primary tumors were treated with NIR irradiation and tumors on the right side designed as abscopal tu- mors were untreated with direct NIR irradiation, respectively. When the primary tumors volume reached ~60 mm3, mice were randomly di- vided into six groups: (1) PBS, (2) anti-PD-L1 antibody, (3) mPEG-Pep- IDOi/ICG NPs, (4) mPEG-Pep-IDOi/ICG NPs with NIR irradiation, (5) mPEG-Pep-IDOi/ICG NPs plus anti-PD-L1 antibody, (6) mPEG-Pep- IDOi/ICG NPs with NIR irradiation plus anti-PD-L1 antibody. As shown in Fig. 6a and Supplementary Fig. 15a, treatment with PD-L1 blockade and/or mPEG-Pep-IDOi/ICG NPs without NIR irradiation showed moderate inhibition effect on the primary tumors growth, but mPEG-Pep-IDOi/ICG NPs-based phototherapy treatment with/without PD-L1 blockade completely eradicated the primary tumors, indicating that mPEG-Pep-IDOi/ICG NPs-based phototherapy treatment were effec- tively for the primary tumors. However, the abscopal tumors growth of mice treated with mPEG-Pep-IDOi/ICG NPs-based phototherapy was partly delayed (Fig. 6b and c, Supplementary Fig. 15b). In contrast, for mice treated with mPEG-Pep-IDOi/ICG NPs-based phototherapy in combination with PD-L1 blockade, their abscopal tumors exhibited significant tumor-inhibiting effect. These results indicated that mPEG- Pep-IDOi/ICG NPs-based phototherapy could completely eradicate the primary tumors growth and its abscopal antitumor effect could be further increased after combination with PD-L1 blockade. No obvious pathological damage was observed in the main organs (heart, liver, spleen, lung, kidney) of the mice treated with mPEG-Pep-IDOi/ICG NPs plus PD-L1 treatment (Supplementary Fig. 16), but cell apoptosis of the abscopal tumors in the mice treated with mPEG-Pep-IDOi/ICG NPs plus PD-L1 treatment was remarkable induced while comparing with all the other groups (Fig. 6d). To understand the mechanism underlying synergistc antitumor effect of mPEG-Pep-IDOi/ICG NPs-based phototherapy in combination with PD-L1blockade, immune cells inflitrated into the abscopal tumors were evaluated using flow cytometry. Cytotoxic T lymphocytes (CTLs, CD3+ CD8+ T cells) and help T cells (CD3+ CD4+ T cells) play critical roles in the promotion of antitumor immune responses. In our experi- ments, for mice treated with anti-PD-L1 antibody alone, either mPEG- Pep-IDOi/ICG NPs or mPEG-Pep-IDOi/ICG NPs with NIR irradiation, no significant enhancement was found in the levels of tumor-infiltrating CD8+ CTLs compared with the PBS-treated group (Fig. 6e and f). mPEG-Pep-IDOi/ICG NPs plus anti-PD-L1 antibody induced about 6.27% CD8+ CTLs infiltration, which was 3.86- and 3.13-fold higher than that of anti-PD-L1 antibody group and mPEG-Pep-IDOi/ICG NP group, respectively. Phototherapy treatment further increased the fre- quency of CD8+ CTLs to 11.44%, which appeared to 1.82-fold higher than that of mPEG-Pep-IDOi/ICG NPs plus anti-PD-L1 antibody group. Notably, the ratios of tumor-infiltrating CD8+ CTLs and CD4+ T cells in the abscopal tumors for mice treated by mPEG-Pep-IDOi/ICG NPs with NIR irradiation plus anti-PD-L1 antibody was found to be the highest among all treatment group (Fig. 6g). In addition, the combination of mPEG-Pep-IDOi/ICG NPs-based phototherapy treatment and PD-L1 blockade was most efficient at the promotion of the intratumoral in- filtration of IFN-γ positive CD8+ T cells (Fig. 6e and h). Meanwhile, the combination of mPEG-Pep-IDOi/ICG NPs-based phototherapy treat- ment and PD-L1 blockade significantly enhanced the ratio of CD8+ CTLs to regulatory T cells (Treg, CD4+FoxP3+) in the abscopal tumor compared to the all other groups (Fig. 6i). Furthermore, the production of Th1-polarizing cytokines in serum of mice after different treatments was measured using LEGENDplex™ Mouse Th1 Panel by flow cyto- metry. The results in Fig. 6j indicated that the levels of IFN-γ, TNF-α, IL- 6 and IL-2 were all obviously higher in the mice treated with mPEG- Pep-IDOi/ICG NPs-based phototherapy treatment plus PD-L1 blockade. All the results demonstrate that mPEG-Pep-IDOi/ICG NPs-based pho- totherapy treatment plus PD-L1 blockade could synergistically promote antitumor immune responses to efficiently inhibit the growth of tumors. 4. Conclusion In conclusion, we developed a self-assembled, MMP-responsive and size-changeable prodrug nanoplatform for improved photo-im- munotherapy of cancer. mPEG-Pep-IDOi/ICG NPs could efficiently ac- complish prolonged circulation, favorable tumor accumulation and penetration. mPEG-Pep-IDOi/ICG NPs (~140 nm) could transform into small IDOi/ICG NPs (< 40 nm)) at TME, mainly owing to the change of intermolecular interactions between IDOi and ICG after the shedding of PEG shell. The resulting small IDOi/ICG NPs could enhance tumor cell internalization, directly kill tumor cells, in situ trigger antitumor im- mune response and modulate IDO-mediated immunosuppression. Meanwhile, the triggered antitumor immune response can be sy- nergistically promoted by PD-L1 blockade, then inducing robust anti- tumor immunity and supressing the growth of tumors in distant sites without phototherapy treatment. The present study provides a novel delivery strategy for nanoenabled phototherapy and IDO inhibition to combine PD-L1 checkpoint blockade, which shows improved ther- apeutic efficacy as compared to conventional nanomedicine and might have a great INCB024360 potential for the more effective therapy of solid tumors.