Photoactivatable nanoagonists chemically programmed for pharmacokinetic tuning and in situ cancer vaccination

Significance

Cancer immunotherapy based on T cell checkpoint blockade only shows response in a subset of patients due to immunosuppressive tumor microenvironment. Poor pharmacokinetic properties and uncontrolled systemic dissemination of small-molecule immunoadjuvants also hamper their clinical implementation. We here developed a photoactivatable nanoagonist platform that imparts near-infrared (NIR) light-induced cytotoxicity and immunogenic cell death (ICD) in synchrony with NIR light-triggered agonist release for immunotherapy. The administration of structurally optimized nanoagonists with improved pharmacokinetics followed by NIR irradiation effectively generated reactive oxygen species not only to ablate tumors and induce the ICD cascade but also to trigger the on-demand release of agonists for cancer vaccination. This strategy represents a facile and generalizable approach for drug pharmacokinetic tuning and immunotherapy improvement.

Abstract

Immunotherapy holds great promise for the treatment of aggressive and metastatic cancers; however, currently available immunotherapeutics, such as immune checkpoint blockade, benefit only a small subset of patients. A photoactivatable toll-like receptor 7/8 (TLR7/8) nanoagonist (PNA) system that imparts near-infrared (NIR) light-induced immunogenic cell death (ICD) in dying tumor cells in synchrony with the spontaneous release of a potent immunoadjuvant is developed here. The PNA consists of polymer-derived proimmunoadjuvants ligated via a reactive oxygen species (ROS)-cleavable linker and polymer-derived photosensitizers, which are further encapsulated in amphiphilic matrices for systemic injection. In particular, conjugation of the TLR7/8 agonist resiquimod to biodegradable macromolecular moieties with different molecular weights enabled pharmacokinetic tuning of small-molecule agonists and optimized delivery efficiency in mice. Upon NIR photoirradiation, PNA effectively generated ROS not only to ablate tumors and induce the ICD cascade but also to trigger the on-demand release of TLR agonists. In several preclinical cancer models, intravenous PNA administration followed by NIR tumor irradiation resulted in remarkable tumor regression and suppressed postsurgical tumor recurrence and metastasis. Furthermore, this treatment profoundly shifted the tumor immune landscape to a tumoricidal one, eliciting robust tumor-specific T cell priming in vivo. This work highlights a simple and cost-effective approach to generate in situ cancer vaccines for synergistic photodynamic immunotherapy of metastatic cancers.

Cancer immunotherapy has emerged as a powerful therapeutic paradigm that stimulates the patient immune system to achieve durable tumor control and manage metastatic cancer recurrence and treatment escape (1, 2). Despite the remarkable clinical success of immune checkpoint blockade therapy (e.g., anti-PD-1/PD-L1 and anti-CTLA-4 inhibitors), the responses occur only in a small subset of patients (3, 4). Growing evidence has suggested that antitumor responses primed by these immunotherapies correlate with the host’s inherent immunological status, and multiple mechanisms of immune resistance exist in tumors (5). The majority of cancers are poorly immunogenic (so-called “cold” tumors), characterized by the abundant presence of immune suppressors together with a paucity of tumor-infiltrating lymphocytes, making them refractory to conventional immunotherapies (6). Consequently, approaches capable of transforming an immunologically cold tumor microenvironment (TME) into a “hot” one are essential to potentiate cancer immunotherapy (7–9).

Toll-like receptor 7/8 (TLR7/8) agonists are involved in innate immune stimulation, activating antigen-presenting cells (APCs) and promoting the subsequent proliferation of T cells (10–12). Currently, several small-molecule TLR agonists are clinically approved only for topical formulations (13). Parenteral dosing of these agonists is impeded by systemic distribution and severe immune-related adverse events. Moreover, the pharmacokinetic shortcomings of small molecules, including rapid metabolism and elimination from the systemic circulation and insufficient transport to lymph nodes or tumors, compromise their efficacy (10). Delivery strategies based on particulate platforms have been developed to address these issues, including covalent conjugation or physical encapsulation within nanocarriers (14, 15). Unfortunately, inefficient activation of parent TLR7/8 agonists in cellular organelles where the corresponding TLR receptors are located might impede the biological activity. In this context, activatable nanoparticles have been engineered to sense and respond to endogenous/exogenous stimuli, enabling effective restoration of the activity at the desired destination, along with lowering systemic toxicity for various immunotherapeutics (16–18).

Photodynamic therapy (PDT), which employs light-absorbing photosensitizers transforms molecular oxygen into reactive oxygen species (ROS) under laser irradiation, usually near-infrared (NIR) light, to kill cancer cells (19, 20). As a minimally invasive therapeutic modality, PDT has the merits of high spatiotemporal selectivity, favorable efficacy, and low toxicity, making it widely explored in clinical practice. Light can also serve as an exogenous trigger to control the pharmacological actions of drugs, but this property has been scarcely exploited for immunomodulators. More significantly, PDT has the potential to induce immunogenic cell death (ICD) in cancer cells, which changes immunogenicity by locally releasing tumor-associated antigens (TAAs), damage-associated molecular patterns (DAMPs), and proinflammatory cytokines (21, 22). These signals further stimulate the maturation of dendritic cells (DCs) and the presentation of antigens to T cells, provoking a specific antitumor immune response (23, 24). Unfortunately, PDT-triggered antitumor immunity is often not sufficiently potent. Immunoadjuvants are critical to the efficacy of vaccines. Therefore, through the combination of tumor-specific antigens with an additional adjuvant, such as a TLR agonist, vaccine-like function can be achieved in situ, eliciting an optimal antitumor immune response (25, 26).

Here, we describe a photoactivatable TLR7/8 nanoagonist (PNA) platform that imparts NIR light-induced cytotoxicity and ICD in synchrony with light-triggered release of a potent immunoadjuvant for synergistic photodynamic immunotherapy (Fig. 1). The shortened in vivo duration of the TLR7/8 agonist resiquimod (RESQ, also named R848) was first overcome by rational engineering of polymer-drug conjugates (Fig. 1A). By varying the chain length of poly(ε-caprolactone) (PCL) for drug derivatization, we systemically tailored the pharmacokinetics and biodistribution of the RESQ agonist to improve intratumoral delivery. Polymer-derived photosensitizers were further coassembled to form PNAs. Upon NIR laser irradiation, the optimized PNA platform not only potently elicited the ICD cascade but also facilitated the release of active adjuvant at tumor lesions for DC maturation, serving as an in situ tumor vaccine (Fig. 1B). In multiple preclinical cancer models, intravenous (i.v.) PNA administration followed by NIR tumor irradiation achieved durable antitumor activity and elicited a potent immune abscopal effect on drastic suppression of metastatic escape postsurgery and distant tumor growth.

Fig. 1.

In vitro characterization of photoactivatable TLR 7/8 nanoagonist. (A) Chemical structures of PCL-derived resiquimod prodrug (PCL-TK-RESQ) and photosensitizer (PCL-PPa). (B) Schematic illustration of immune remodeling mediated by photoactivatable TLR 7/8 nanoagonist under 660 nm laser irradiation. TAAs, tumor-associated antigens; DAMPs, damage-associated molecular patterns; iDC, immature DC; mDC, mature DC. (C) Representative TEM images of PCL_n-TK-NA (n = 8, 16, or 34). (D) DLS profiles and (E) zeta potentials of PCL_n-TK-NA (n = 8, 16, or 34). Inset, photograph of the prepared nanoparticle solutions at a RESQ-equivalent concentration of 0.1 mg/mL. (F) Representative HPLC chromatograms of the prodrug PCL₁₆-TK-RESQ in the presence or absence of 200 mM H₂O₂ (Left) and quantification of free RESQ release from PCL₁₆-TK-RESQ (Right). (G) Representative HPLC chromatograms of the PCL₁₆-TK-NA with incubation of 0.1 M NaOH solution (Left) and the release profile of RESQ (Right). (H) Time-dependent decomposition of ICG in the presence of free PPa or PCL-PPa-NPs upon laser irradiation (0.3 W cm⁻²), recorded by ultraviolet-visible spectrometry. (I) Morphology change of PNA₁₆ following NIR irradiation, as examined by TEM observation. Data are presented as the mean ± SD.

Results

Synthesis and Characterization of Photoactivatable TLR7/8 Nanoagonist (PNA).

The effect of different moieties for prodrug conjugates on in vivo nanoparticle performance has not been systematically and thoroughly explored to date. Hence, the TLR7/8 agonist RESQ was covalently ligated to PCL (a Food and Drug Administration (FDA)-approved biocompatible and biodegradable polymer) with varying chain lengths (Fig. 1A). To ensure the efficient release of active RESQ, a thioketal (TK) linker, which is susceptible to ROS, was used for drug conjugation (27, 28). Owing to the overall structural similarity, the resulting PCL_n-TK-RESQ prodrugs (where n = 8, 16, and 34; synthetic protocols and characterizations are given in SI Appendix, Materials and Figs. S1–S9) were successfully assembled with the amphiphilic polyethylene glycol (PEG)-PCL copolymer (e.g., PEG_10k-PCL_10k) to form colloidally stable nanoagonist (NA) suspensions (referred to as PCL_n-TK-NA). Transmission electron microscopy (TEM) observation revealed the formation of a reproducible spherical structure (Fig. 1C). The particle sizes of PCL_n-TK-NA increased with increasing PCL chain length, as measured by dynamic light scattering (DLS) analysis, with values of 84.3 ± 43.9, 76.0 ± 32.0, and 125.8 ± 49.6 nm for PCL₈-, PCL₁₆-, and PCL₃₄-TK-RESQ, respectively (Fig. 1D). The zeta potentials depicted in Fig. 1E indicated the slightly positive surface charge of PCL_n-TK-NA. ROS-responsive drug activation was assessed after the exposure of the nanoagonist to hydrogen peroxide (H₂O₂). After 2 h of incubation, a distinct peak ascribed to free RESQ was observed in the high-performance liquid chromatography (HPLC) spectra (Fig. 1F). In contrast, free RESQ was not released in the absence of H₂O₂ (Fig. 1F). Moreover, the treatment with sodium hydroxide hydrolyzed only the ester bond in the PCL_n-TK-RESQ prodrug but not the TK linker (Fig. 1G), distinguishing the mechanism of ROS-mediated drug activation.

For in vivo activation of the TLR7/8 agonist RESQ, we attempted the incorporation of the photosensitizer pyropheophorbide a (PPa) into the nanoagonist. To efficiently load PPa into nanoparticles, PPa was similarly conjugated to the PCL fragment via a noncleavable linker, which resulted in PCL_n-PPa conjugates (n = 16 or 34) with high yields. Upon irradiation with an NIR (660 nm) laser, the neighboring PPa photosensitizer is expected to produce ROS, which can spontaneously cleave the TK linker followed by the esterase-mediated release of active RESQ. Encapsulation of these PPa conjugates in the PEG_10k-PCL_10k matrix produced PCL-PPa conjugate-loaded nanoparticles (PCL-PPa-NPs), which were characterized by DLS (SI Appendix, Fig. S10). To assess the capacity of PCL-PPa-NPs for ROS generation, indocyanine green (ICG) was employed as an ROS indicator (29), and the reduction of the absorbance at 779 nm assigned to ICG can indicate the ROS-generating capacity. As shown in Fig. 1H and SI Appendix, Fig. S11, PCL-PPa-NPs exhibited comparable ability to produce ROS under irradiation compared to free PPa in dimethyl sulfoxide but was higher than free PPa dispersed in water, suggesting that the nanoformulation of PPa did not impair its activity. Finally, the PCL-ligated photosensitizer and agonist were coassembled to form a photoactivatable TLR7/8 nanoagonist (PNA, Fig. 1 A and I, Left). After NIR irradiation, a significant morphological change with irregular nanoaggregates and particle fusion was observed, indicating that ROS generated by the coassembled photosensitizer triggered TK bond cleavage in the PCL-TK-RESQ prodrug and disrupted the nanostructures (Fig. 1 I, Right).

Pharmacokinetics, Biodistribution, and Tumor Penetration.

Prolonged circulation and enhanced tumor accumulation are indispensable for favorable in vivo efficacy of therapeutic nanoparticles (30). First, the effect of PCL on the pharmacokinetic characteristics of RESQ was investigated. After i.v. injection of each PCL_n-TK-NA (n = 8, 16, and 34) into Sprague–Dawley (SD) rats, the plasma drug concentration was analyzed to extrapolate the time-dependent profiles. In contrast to the rapid elimination of free RESQ (t_1/2: 0.87 ± 0.09 h), RESQ delivered by PCL_n-TK-NA had significantly extended half-lives in the blood circulation (i.e., t_1/2: 9.71 ± 0.42, 16.17 ± 1.00, and 18.60 ± 1.29 h for PCL₈-, PCL₁₆-, and PCL₃₄-TK-NA, respectively, Fig. 2A). We noted some differences in pharmacokinetic properties, and this observation was also supported by the data of the area under the concentration-time curve (AUC_0–t). As summarized in SI Appendix, Table S1, the AUC_0–t values for PCL₁₆- and PCL₃₄-TK-NA were 4.9-fold and 4.8-fold greater than that of PCL₈-TK-NA, respectively. These data suggest that the attachment of small-molecule RESQ agents to macromolecular promoieties followed by nanoparticle delivery enabled systemic tuning of the pharmacokinetic characteristics. We further assessed the delivery efficiency (% of injected dose per gram of tumor) after a single injection of each nanoparticle at an identical dose in multiple mouse models, including orthotopic 4T1 breast carcinoma, B16F10 melanoma, and MC38 colorectal adenocarcinoma (Fig. 2B). The results consistently showed that PCL₁₆- and PCL₃₄-TK-NA had comparably higher delivery efficiencies than PCL₈-TK-NA and free RESQ, supporting that extended blood circulation was beneficial to intratumoral delivery.

Fig. 2.

Pharmacokinetic tuning by polymer conjugation, altered intratumoral drug delivery, and deep penetration. (A) Plasma RESQ concentration-time profiles in SD rats after a single i.v. injection of free RESQ or PCL_n-TK-NA (n = 8, 16, or 34) via the tail vein. The drug dose was 2 mg/kg (RESQ equivalence). (B) Intratumoral RESQ-equivalent concentration at 24 h postadministration in 4T1, B16-F10, and MC38 tumor-bearing mice, dosing with free RESQ or PCL_n-TK-NA at a dose of 5mg/kg (RESQ equivalence). (C) Real-time in vivo fluorescence imaging of tumor-bearing mice after i.v. injection of PNA₁₆ (10 mg PPa kg⁻¹, 3.3 mg RESQ kg⁻¹). NIRF images were acquired at 710 nm upon excitation at 640 nm. (D) Ex vivo NIRF imaging (Left) and quantitative analysis (Right) of the PPa amount in tumors excised at different time points postadministration. (E) Ex vivo NIRF imaging (Left) and quantitative analysis (Right) of the PPa amount in organs postadministration. He, heart; Li, liver; Sp, spleen; Lu, lung; Ki, kidneys; LN, lymph node; Tu, tumor. (F) Drug accumulation after PNA₁₆ dosing, as determined by HPLC. (G and H) Representative fluorescence images showing tumor extravasation and penetration of PNA₁₆ (red) at 6, 12, 24, and 48 h postadministration. Blood vessels were marked with Alexa Fluor® 594 anti-mouse CD31 antibody (green), and nuclei were stained with DAPI (blue). Time-dependent fluorescence intensity profiles derived from PNA₁₆ as a function of distance from vessels (0 to 60 µm) were plotted with Image J. Data are presented as the mean ± SD. The statistical significance was examined by one-way ANOVA with Tukey’s, Game–Howell’s, or Dunnett’s T3 multiple comparisons test.

To confirm the maximal accumulation time of PNA₁₆, real-time near-infrared fluorescence (NIRF) imaging was conducted after systemic administration of PNA₁₆, owing to the existence of fluorescent PPa. As shown in Fig. 2 C and D, the NIRF signal in tumors reached a maximum of ~8.0 × 10¹⁰ p/s/cm²/sr at 24 h postadministration, providing an optimal time point for phototherapy to maximize the treatment efficacy. Ex vivo imaging further showed that PNA₁₆ had high accumulation in tumors, followed by liver and other tissues (Fig. 2E). Moreover, it was further degraded in the liver over time, which was consistent with the resident concentration of RESQ as measured by HPLC (Fig. 2F). Tumor tissues were cryosectioned and observed with a fluorescence microscope. PNA₁₆ had deep tumor penetration, as evidenced by extensive intratumoral NIRF signals throughout the tumor sections (Fig. 2G). The extravasation and subsequent interstitial penetration of PNA₁₆ varied in a time-dependent manner, and the fluorescence signal could be detected even 60 µm from the vessel at 24 h postadministration, indicating the superior tumor penetration capability of PNA₁₆ (Fig. 2H). These results suggest that prodrugs with higher molecular weights prevented premature drug release during blood circulation, leading to better tumor accumulation.

Photocytotoxicity Against Cancer Cells.

In the PNA scaffold, the polymer-derived photosensitizer PPa was simultaneously coassembled to induce the PDT effect. Thus, the ROS generation efficiency of PNA₁₆ was first measured in the 4T1 murine breast cancer cells using 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) as an ROS generation indicator, which can be oxidized by ROS to emit green fluorescence (31, 32). Upon 660 nm laser irradiation, strong green fluorescence from the oxidized DCF was found in 4T1 cells, which was 4.1-fold higher than that in the control without laser irradiation (Fig. 3 A and B). Unexpectedly, PNA₁₆ exhibited a higher cellular uptake efficiency than PNA₃₄ (P = 0.0012, SI Appendix, Fig. S12), which might be due to the smaller size of PNA₁₆. The in vitro cytotoxicity of PNA₁₆ was further evaluated against 4T1 cells. Consistently, PNA₁₆ showed higher cytotoxicity against 4T1 cells in the presence of laser irradiation than PNA₃₄ (SI Appendix, Fig. S13A). In contrast, negligible cytotoxicity of PNA₁₆ against 4T1 cells was found in the absence of irradiation even after coincubation for 72 h (SI Appendix, Fig. S13B). For better validation of the potency of PNAs, a further calcein acetoxymethyl ester (AM)/propidium iodide dual-staining assay was conducted. After NIR irradiation, a larger area of cell death was found in PNA₁₆-treated cells than in PNA₃₄-treated cells (Fig. 3C). Moreover, pinpoint laser irradiation led to selective cell death observed only in the NIR-irradiated region (Fig. 3D), suggesting that ROS generated by irradiating coassembled PPa could specifically induce photocytotoxicity while sparing unirradiated cells. Therefore, upon exposure to NIR light, ROS production can induce damage to cellular compartments, resulting in irreversible cell apoptosis.

Fig. 3.

ICD induced by PNA₁₆ treatment and photoirradiation and amplifies immune response. (A and B) Intracellular ROS production by photosensitizer PPa-integrated PNA was detected with the fluorescent indicator H₂DCFDA. Cells were treated with PNA₁₆ and subsequent laser irradiation (660 nm, 0.3 W/cm², 5 min). The fluorescence signal was observed under fluorescence microscopy (A) or flow cytometric analysis (B). (C) Fluorescence microscopy images for dead/live cell staining of 4T1 cells. After drug treatments, cells were costained with calcein-AM (green, live cells) and propidium iodide (red, dead cells). (D) Locoregional photocytotoxicity induced by PCL₁₆-PPa-NP. 4T1 cells were treated with or without PCL₁₆-PPa-NP. The right side of the dotted line was irradiated with an NIR laser, while the left side was in the dark. (E and F) Induction of ICD in 4T1 cells. Representative fluorescence images showing CRT exposure on the cell membrane (E) and HMGB1 release from the nucleus to the cytoplasm (F) upon laser irradiation (660 nm) for 5 min. Cell membrane was marked with DiI (red), and the nuclei were stained with DAPI (blue). (G) Changes in intracellular and extracellular ATP concentrations following treatment with PCL₁₆-PPa-NPs and laser irradiation. (H) Illustration of the protocol of the photoactivation of PNA and stimulation of iDCs in vitro. (I–K) Flow cytometric analysis of BMDC maturation after treatments (gated as CD11c⁺ DCs). Data are presented as the mean ± SD. The statistical significance was examined by Student’s t test (G) or one-way ANOVA with Tukey’s multiple comparisons test (B, J, and K).

PDT Induces ICD and Triggers DC Maturation.

Photodynamic damage is capable of activating ICD, which is characterized by calreticulin (CRT) exposure, high mobility group box 1 (HMGB1) release, and adenosine triphosphate (ATP) secretion (33). To test whether PNA₁₆-mediated PDT triggers the ICD cascade, DAMPs were first analyzed. After laser irradiation for 5 min, CRT translocation from the endoplasmic reticulum (ER) to the cell membrane was observed, as evidenced by colocalization with the lipophilic membrane-specific red dye DiI (Fig. 3E) (34). In contrast to the intact membrane of untreated or unirradiated cells, the cell membrane after PNA₁₆ exposure and photoirradiation was entirely disrupted, which could be attributed to the potent photocytotoxicity of PPa. Moreover, the nuclear protein HMGB1 was released from the nucleus to the cytoplasm after photodynamic treatment (Fig. 3F). Furthermore, PNA₁₆ treatment followed by photoirradiation reduced the intracellular ATP concentration while upregulating the extracellular ATP concentration (Fig. 3G). Hence, these results suggest that PNA₁₆-mediated PDT can induce ICD in cancer cells, as evidenced by CRT exposure, HMGB1 release, and ATP secretion.

In addition to photodynamic action, photoirradiation of the PNA platform also triggers a spontaneous cascade reaction to release the active adjuvant RESQ. DCs are the primary APCs and an important target of adjuvants in vaccine therapeutic efficacy (35). We thus examined to what extent the activatable nanoagonist could mature bone marrow-derived dendritic cells (BMDCs) in vitro (Fig. 3H). For this purpose, the expression of the costimulatory surface molecules CD80 and CD86, biomarkers for the maturation of DCs, was first measured (33). NIR-irradiated PNA₁₆ treatment resulted in 43.6% of DC maturation, which was 2.7-fold higher than that of untreated DCs. Moreover, the culture medium of PNA₁₆-treated tumor cells with NIR irradiation elicited markedly higher expression of CD80 and CD86 (~57.8%) than free RESQ or NIR-irradiated PNA₁₆ (Fig. 3 I and J). We further analyzed the major histocompatibility complex (MHC II), which is a DC activation marker (Fig. 3K and SI Appendix, Fig. S14). A similar trend in the upregulation of MHC II was observed; treatment with photoirradiated PNA₁₆ significantly increased the MHC II proportion relative to that in the untreated group (Fig. 3K). Therefore, using the PNA platform, the ROS produced by NIR light irradiation can not only induce the ICD cascade but also trigger the release of the TLR7/8 agonist RESQ, both of which can synergize to produce the most robust immune activation. These data also confirmed the superior PNA₁₆-mediated photodynamic immunotherapeutic efficacy.

Photoactivatable Nanoagonist Suppresses Primary Tumor Growth and Postsurgical Tumor Metastases and Induces Abscopal Effect.

We next assessed the therapeutic potential of the PNA platform in a clinically relevant breast carcinoma model. To this end, BALB/c mice were orthotopically implanted with 4T1 syngeneic cancer cells in the breast pad. When tumors reached a size of ~60 mm³, mice were administered a dose of 3.3 mg/kg RESQ and 10 mg/kg PPa i.v. at 3-d intervals in nanoformulations or soluble form, followed by NIR irradiation at 24 h postadministration (Fig. 4A). The tumor grew quickly in the saline-treated group, whereas the combination of free RESQ and PPa with laser irradiation produced only a moderate tumor inhibition, presumably due to the rapid elimination in the blood and limited tumor accumulation (Fig. 4 B–D). In contrast to the continued tumor growth observed for PCL₁₆-TK-NA or PCL₁₆-PPa-NP monotherapy, PNA₁₆-mediated photodynamic immunotherapy resulted in substantial tumor shrinkage (~31.6 mm³, P < 0.0001 versus PCL₁₆-PPa-NP) (day 15; Fig. 4B), with one out of six mice cured (SI Appendix, Fig. S15). The superior antitumor activity of PNA₁₆ compared with monotherapies was probably ascribed to the augmented immune activation derived from RESQ. However, in the absence of laser irradiation, PNA₁₆ showed limited in vivo activity in terms of reduction in tumor volume, indicating that light-triggered ROS generation was indispensable for RESQ activation and irreversible cell apoptosis. Moreover, PNA₁₆ did not cause any body weight loss in comparison to the other monotherapy groups (Fig. 4E). Hematoxylin and eosin (H&E) staining and immunochemistry analysis of tumor sections also revealed significantly extensive cell apoptosis and a lower level of cell proliferation after PNA₁₆-mediated therapy (SI Appendix, Figs. S16–S18).

Fig. 4.

In vivo photodynamic immunotherapy to treat orthotopic 4T1 breast cancer, postsurgical metastasis, and abscopal effect. (A) Scheme of orthotopic 4T1 breast tumor inoculation and treatment. (B) Primary tumor growth curves for each group. (C) Individual tumor growth curves as indicated in B. (D) Weights of tumors excised from each treatment group on day 15 at the endpoint of observation. (E) Body weights of mice after different treatments (n = 6). (F) In vivo bioluminescence imaging of mice to track the metastases of 4T1-Luc breast cancer cells postsurgery. (G) Survival of mice (n = 5). Histological analysis of tumor metastases to lymph nodes (H) and lungs (I) from each treatment group on day 47. Black dotted circles indicate metastasized tumors. (J) Scheme of bilateral 4T1 tumor mouse model to examine the immune abscopal effect induced by PNA₁₆ vaccination. The growth kinetics of primary (K) and distant (L) tumors (n = 7). Data are presented as the mean ± SD. The statistical significance was examined by one-way ANOVA (D) or two-way ANOVA (B, K, and L) with Tukey’s multiple comparisons test.

Metastasis is the leading cause of cancer-related mortality worldwide, and the treatment of metastatic cancer remains a major clinical challenge (36, 37). Previous studies have shown that inoculated 4T1 breast cancer is invasive and has high metastatic potential to distant organs (38, 39). Therefore, we studied whether PNA₁₆ could suppress postsurgical tumor recurrence and metastasis. After the surgical resection of orthotopic 4T1 breast tumors that expressed luciferase (4T1-Luc), in vivo bioluminescence imaging was performed to assess the metastases of cancer cells (Fig. 4F). PCL₁₆-TK-NA monotherapy failed to prevent metastasis, and no mice were alive on day 85. PCL₁₆-PPa-NP-mediated PDT alone only partially inhibited tumor recurrence, with a survival rate of 60% up to 100 d. Of significant note, dosing with PNA₁₆ followed by photoirradiation achieved superior survival benefit, with all mice surviving up to 100 d, and 80% of the mice were free of tumor metastases after treatment (Fig. 4 F and G). On day 47, mice were sacrificed, and the organs were subjected to histopathological analysis (Fig. 4 H and I and SI Appendix, Figs. S19–S23). In the saline and PCL-TK-NA groups, breast cancer cells were found to metastasize readily to lymph nodes and internal organs, including the heart, lung, liver, and spleen, exhibiting large metastatic foci in these organs. Encouragingly, no signs of metastases in any organs were observed in mice treated with PNA₁₆ with laser irradiation. These data suggest that PNA₁₆ administration followed by local PDT produced a systemic regression of metastatic lesions through the abscopal effect, which significantly contributed to the extended survival of animals.

We further tested the abscopal effect of the PNA therapy in a bilateral subcutaneous tumor mouse model (Fig. 4J). Mice bearing two tumors were treated i.v. with PNA₁₆, followed by local NIR irradiation on the primary tumor. PNA₁₆ therapy nearly restrained primary tumor growth and was also effective in reducing distant tumor burden (Fig. 4 K and L), suggesting the induction of systemic immune responses. However, treatments with PNA₁₆ alone or the soluble drug form had a limited abscopal effect, with comparable tumor growth kinetics to the saline treatment. Overall, these experimental pieces of evidence demonstrated that PNA₁₆-mediated photodynamic immunotherapy not only induced in situ photodynamic ablation but also elicited a potent immune abscopal effect on drastic inhibition of metastatic tumor burden postsurgery and distant tumor growth.

Immune Remodeling Mediated by Photoactivatable Nanoagonist.

To gain insight into the high activity of PNA-mediated therapy, the immunological responses were investigated. PDT holds the potential to elicit ICD cascade, releasing TAAs, DAMPs, and proinflammatory cytokines, which can generate in situ tumor vaccine effects in combination with TLR7/8 agonist (22). Hence, the level of mature DCs (CD80⁺CD86⁺) in tumors was first assessed (33). As shown in Fig. 5A and SI Appendix, Fig. S24, PNA₁₆ treatment followed by photoirradiation substantially activated DCs in tumor tissues; the ratio of mature DCs was increased from 18.0% (untreated mice) to 48.3%, while less maturation of DCs was observed in other treatments. The intratumoral infiltration of cytotoxic CD8⁺ T lymphocytes (CTLs) plays a critical role in tumor killing (40); thus, we measured the tumor-infiltrated CTLs. PNA₁₆ therapy significantly increased the overall number of tumor-infiltrating CTLs. Moreover, a marked increase in the CD8⁺/CD4⁺ ratio was observed, which was 2.9-fold and 3.4-fold greater than that of photodynamic monotherapy and the free drug combination, respectively (Fig. 5 B and C and SI Appendix, Fig. S25). We further characterized the composition of tumor-associated macrophages (TAMs) and the proportion of natural killer (NK) cells in the TME of 4T1 tumors (41). Obviously, PNA₁₆ was capable of repolarizing TAMs from protumorigenic M2 (CD206^hiCD11b⁺F4/80⁺) to antitumorigenic M1 (CD86^hiCD11b⁺F4/80⁺) (Fig. 5 D–F and SI Appendix, Fig. S26) and significantly promoting the infiltration of NK cells (Fig. 5G and SI Appendix, Fig. S27). PNA₁₆ also induced robust DC maturation in tumor-draining lymph nodes (tdLNs) (Fig. 5H and SI Appendix, Fig. S28). In addition, a similar increase in the population of CTLs was observed in the spleens of the PNA₁₆-treated group with irradiation (Fig. 5 I and J and SI Appendix, Fig. S29). Matured DCs exhibited functional stimulation characterized by interleukin (IL)-6^high and IL-1α^high (42, 43). Thus, the cytokines in serum were further evaluated via the LEGENDplex™ Mouse Inflammation Panel (13-plex) with a V-bottom Plate kit (Fig. 5K). Note that the serum level of IL-6 was markedly elevated in the mice treated with PNA₁₆, consistent with the levels of DC maturation. The activated immune response was further confirmed by evaluating the cytotoxic cytokine (tumour necrosis factor-α, TNF-α) level in the serum collected from 4T1 tumor-bearing mice after various treatments. PNA₁₆ treatment produced a higher serum level of TNF-α on day 9 than the other treatments. Collectively, after photoirradiation, PNA₁₆ triggered the release of a TLR7/8 agonist, which collaborated with PDT to activate DCs and subsequently stimulated both innate and adaptive immunity, thereby remodeling the immunosuppressive TME to improve the therapeutic outcomes.

Fig. 5.

In vivo PNA₁₆-mediated photodynamic immune remodeling. Mice bearing 4T1 tumors were treated as indicated in Fig. 4A. On day 8, mice were sacrificed, tdLNs, spleens, and tumors were collected for immune analysis. (A–G) Analysis of immune landscape by flow cytometry after treatment. Frequencies of tumoral mature DCs (CD45⁺CD11c⁺CD80⁺CD86⁺ DCs, A), CD8⁺ T cells (B and C), CD86⁺ M1-, CD206⁺ M2-like macrophages (D–F) and NK cells (NK1.1⁺G) (n = 3 to 5). (H) Frequency of mature DCs (CD45⁺CD11c⁺CD80⁺CD86⁺ DCs) in tdLNs (n = 3). Flow cytometric quantification of CD8⁺ T cells (I) and the ratios of CD8⁺/CD4⁺ T cells (J) in spleens (n = 3). (K) Levels of IL-6, IL-1α, IL-17A, and TNF-α in the serum of mice on day 9 after cessation of different treatments. Data are presented as the mean ± SD. The statistical significance was examined by one-way ANOVA with Tukey’s multiple comparisons test.

Therapeutic Efficacy Against Multiple Tumor Types.

The potential of PNA₁₆-mediated photodynamic immunotherapy was further extended to different tumor models (Fig. 6A). In the MC38 (colorectal carcinoma) syngeneic mouse model, similar results were observed (Fig. 6 B–D). PNA₁₆ administration with NIR irradiation substantially triggered sustained regression of the tumors, and four out of nice treated mice showed complete tumor regression, prolonging the median survival time over 50 d (Fig. 6E). In contrast, the treatment with soluble free drug combination followed by NIR irradiation had negligible antitumor effect over saline control (P = 0.1221). Photodynamic monotherapy with PCL₁₆-PPa-NP exhibited only partial tumor regression, and the tumors grew quickly after cessation of treatment (Fig. 6D), indicating that PDT-triggered antitumor immunity was not sufficiently potent. Immune profiling of the tumors after 2 d of the treatment revealed the induction of a robust DC maturation (Fig. 6F and SI Appendix, Fig. S30) and recruitment of tumor-infiltrating CTLs by the PNA₁₆ therapy (Fig. 6 G and H and SI Appendix, Fig. S31). We also observed higher frequencies of central (CD62L⁺CD44⁺, T_CM) and effector (CD62L⁻CD44⁺, T_EM) memory T cells in the CTL populations in the PNA₁₆-immunized cured mice compared to those in the healthy mice (Fig. 6I) (44).

Fig. 6.

PNA₁₆ administration followed by NIR irradiation elicited robust therapeutic effects in multiple tumor models. (A) Treatment scheme for mice with established tumors. (B) Representative images of MC38 syngeneic tumors photographed on day 8 after treatment initiation. Individual tumor growth curves (C) and average tumor growth curves (D) in each group. (E) Mouse survival in the MC38 tumor-bearing model. Mice were sacrificed when tumor volume exceeded 2,000 mm³. (F) Frequency of tumoral mature DCs (CD11c⁺CD80⁺CD86⁺ DCs). Frequency of tumoral CTLs (CD8a⁺ T cells, G) and the ratio of CD8⁺/CD4⁺ T cells (H) (n = 5). (I) Representative flow cytometric plots and corresponding quantification of central (T_CM) and effector (T_EM) memory CTLs among PBMCs from immunized mice on day 50 (n = 4). (J–L) Therapeutic effect of PNA₁₆ against B16F10 melanoma. (J) Representative images of B16F10 tumors photographed on day 8 after treatment initiation. Individual tumor growth (K) and average tumor growth (L) in each group. (M) Tumor antigen-specific T cell expansion in B16F10-OVA tumor-bearing mice. Representative flow cytometric analysis and corresponding quantification of OVA (peptide sequence: SIINFEKL)-specific CD8⁺ T cells in tumor tissues on day 8 (n = 5). (N) ELISPOT analysis of IFN-γ spot-forming cells within splenocytes restimulated with H-2K^b-OVA_257–264 peptide for 24 h. Splenocytes were harvested from B16F10-OVA or B16F10 tumor-bearing mice on day 8 following various treatments. Data are presented as the mean ± SD. The statistical significance was examined by one-way ANOVA (F–I, M, and N) with Tukey’s multiple comparisons test or two-way ANOVA (D and L) with Bonferroni’s multiple comparisons test or log-rank test (E). *P in E indicates the significance relative to the saline group.

We also found that the PNA₁₆ therapy was effective at impairing tumor growth in mice harboring the B16F10 melanoma tumor, which is aggressive and immunologically silent (45). Dosing with PNA₁₆ followed by photoirradiation dramatically inhibited the growth of B16F10 tumors (Fig. 6 J–L) and induced extensive intratumoral cell apoptosis (SI Appendix, Fig. S32). Finally, an ovalbumin (OVA)-expressing B16F10 melanoma mouse model was included to investigate whether PNA₁₆-mediated T cell priming is tumor-specific. After three-dose vaccination, tumors were harvested and OVA-specific CD8⁺ T cell numbers were determined by H-2K^b-OVA tetramer staining (46). Mice vaccinated with PNA₁₆ had a ~2.9-fold increase in OVA-specific CD8⁺ T cell expansion in the TME than mice treated with saline (Fig. 6M). In addition, a higher frequency of OVA-tetramer⁺ CD8⁺ T cells was observed in the peripheral blood monocytes (PBMCs) from PNA₁₆-treated mice (SI Appendix, Fig. S33). Furthermore, we evaluated the numbers of interferon (IFN)-γ-secreting CD8⁺ T cells within splenocytes after PNA₁₆ vaccination using the enzyme-linked immunospot (ELISPOT) assay. Following the antigenic restimulation with H-2K^b-OVA_257-264 peptide, OVA-tetramer⁺ CD8⁺ T cells from the B16F10-OVA mice treated with PNA₁₆ and NIR irradiation were significantly activated compared to other treatments, as evidenced by the increased levels of IFN-γ-secreting T cells (Fig. 6N). However, in B16F10 tumor mice immunized with PNA₁₆, the OVA-specific CD8⁺ T cell response was negligible. These results indicate that the treatment of PNA₁₆ followed by NIR irradiation can serve as in situ vaccine to prime robust tumor-specific T cell responses.

Discussion

Short half-lives in circulation and poor delivery to the desired in vivo compartments impede the utility of small-molecule compounds acting as potent immunotherapeutic agonists against cancer (10). Moreover, uncontrolled dissemination of these therapeutics following systemic administration also produces severe side effects (47). Existing approaches that control the pharmacokinetics involve physical formulation of agonists into a broad set of delivery carriers to extend their half-lives (48, 49). However, the extent to which these noncovalent forces can prevent the mixing-induced dissociation upon i.v. injection and to improve the pharmacokinetics remain questionable (50). We here described an alternative strategy that enables pharmacokinetic tuning of the immunomodulatory agent RESQ by chemically engineering the macromolecular conjugates via ROS-susceptible linkage. Structure–activity relationships of these nanoagonists were elucidated through experimental characterization of in vitro and in vivo behavior. As a result, RESQ tethered to the PCL₁₆ fragment and delivered by polymeric nanoparticles was optimized for systemic administration in terms of pharmacokinetic properties and tumor accumulation efficiency (Fig. 2).

The optimal RESQ nanoparticles were further coassembled with a polymer-derived photosensitizer to form PNAs that can be locoregionally activated. Local NIR irradiation produces ROS and facilitates the bond cleavage to spontaneously release active RESQ at the tumor sites, which is expected to, in part, achieve tumor-restricted pharmacology. Furthermore, the local PDT by in situ generated ROS induced ICD cascades, subsequently triggering the production of proinflammatory cytokines, DC activation, and cross-presentation of antigens to immune cells. In combination with the on-demand activation of the immunoadjuvant, the PNA platform acts as a tumor vaccine. We tested the efficacy of this nanoagonist against multiple syngeneic tumor models, including B16F10 melanoma, MC38 colorectal carcinoma, and orthotopic 4T1 breast tumor models (Figs. 4 and 6). Of significant note, we observed the abscopal effect and long-term antitumor memory endowed by photoirradiated PNA, protecting mice after surgical resection of tumors from metastatic onset with prolonged survival up to 100 d (Fig. 4). Detailed analysis of tumor-infiltrating immune cells showed that following the PNA₁₆ treatment, there were increased numbers of effector T cells, NK cells, and activated DCs as well as a change in the polarization of macrophages from protumorigenic M2 to antitumorigenic M1 phenotypes, all of which contributed to robust immune responses in solid tumors (Fig. 5). In contrast to reported cancer cell vaccines (51, 52), our strategy is convenient; the in situ vaccination conferred by the nanoadjuvant platform does not require any other ex vivo processing, and only injection and local photoirradiation are needed.

In summary, we have reported a series of activatable nanoagonists differing in their molecular characteristics for the pharmacokinetic tuning of a small-molecule TLR7/8 agonist and combinational cancer immunotherapy. Photoirradiation promoted the in situ generation of ROS and induced ICD cascades, including the production of proinflammatory cytokines, the cross-presentation of TAAs to immune cells, and the on-demand activation of an immunoadjuvant. Given the feasibility of synthesizing polymeric prodrugs, this study may provide a facile and generalizable approach for the delivery of other TLR ligands or immunostimulants for tuning pharmacokinetics and improving immunotherapy outcomes.

Materials and Methods

Compound synthesis, preparation and characterization of PNAs, NIR irradiation-triggered ROS generation and drug activation, in vitro photocytotoxicity and ICD induction, BMDC maturation, pharmacokinetic assays, biodistribution, therapeutic efficacy investigation in mouse models, and immune profiling are described in detail in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.