Light-Controlled mRNA Delivery Regulates Multiple Immune Signals for Targeted Cancer Therapy

IR780 was conjugated to the lysine moieties of PEG‐pLL to prepare the photosensitizing polymers (Scheme). The pLL block contains 80 lysine units, with IR780 modifying 38 of these units. (Figures). Size exclusion chromatography (SEC) confirmed the monodisperse molecular weight distribution of the polymers (Figure). S1 S1–S3 S4

The photosensitizing polymers exhibited pH‐dependent solubility. Dynamic light scattering (DLS) analysis revealed that the polymers associated into particles of around 44 nm in neutral buffer (pH = 7.4), while in acidic buffer (pH = 4.5) the polymers remained dissociated with a size around 10 nm and low light scattering intensity measured as derived count rate (DCR) (Figure). This pH‐dependent solubility was attributed to the protonation of the secondary amines generated after IR780 conjugation, which increases hydrophilicity and enhances electrostatic repulsion between the polymers at acidic pH. Consequently, the polymer stock solution was prepared in acidic buffer (pH 4.5) to prevent premature association before assembling the LITS. S5

mRNA‐loaded LITS formation was screened by mixing polymer and mRNA solutions in buffer (10 mM HEPES, pH 7.4) with varying ratios of lysine units in the polymer to phosphate groups in mRNA, i.e., L/P ratios. Gel electrophoresis demonstrated that by increasing the polymer concentration in the mixture, remaining free mRNA was diminished (Figure S6A), indicating the loading of mRNA into the LITS. DLS measurements also confirmed the formation of nanoparticles for L/P ratios higher than one (Figure 1b). At L/P ratios = 4 and above, the DCR reached a plateau, which indicates that no additional nanoparticles were formed, suggesting the complete assembly of LITS. Also, L/P ratio = 4 resulted to a complete encapsulation of mRNA (>95%) (Figure S6B). Therefore, L/P ratio = 4 was selected to formulate the LITS for the following studies. This LITS was monodisperse (PDI < 0.2) with an average hydrodynamic size of around 80 nm (Figure 1c). Transmission electron microscopy (TEM) images showed nanoparticles with an average diameter of around 60 nm (Figure 1d). Considering the PEG shell has low electron density, which makes it hard to be visualized by TEM, these particles are assumed to represent the cores of the LITS. Also, the difference of the size measured by DLS and that obtained by TEM suggests that the LITS has a 10 nm thick PEG shell, confirming the core‐shell structure. The dense PEG shell shielded the charge of the core, resulting in a nearly neutral ζ‐potential (Figure S7). Fluorescence correlation spectroscopy (FCS) further confirmed mRNA encapsulation within the LITS. Compared to free mRNA, mRNA encapsulated in LITS showed a lagged autocorrelation function (ACF) curve (Figure 1e), indicating delayed diffusion due to its relatively larger size (Table S1). Moreover, LITS and free mRNA showed similar count per molecule values, which indicates that there is one mRNA molecule loaded per LITS particle.

The stability of LITS was evaluated in various conditions and compared to mRNA/PEG‐pLL polyion complex (PIC) to understand the effects of IR780 conjugation. The PIC was prepared at the same L/P ratio of LITS (L/P = 4), forming complexes of around 120 nm. The characterization of PIC is summarized in Figure S8. Adding an electrolyte, i.e., NaCl, into the buffer destabilized the PIC, as its formation is mainly driven by electrostatic interaction between the amines in the lysine moieties and the phosphate groups in mRNA (Figure S9). In contrast, LITS were stable even in 500 mM NaCl, maintaining the monodisperse size. Moreover, PICs are prone to dissociation through polyion exchange when exposed to anionic macromolecules in physiological environment, such as glycosaminoglycans on cell surfaces.^[26, 27^] Such premature release leads to the degradation of mRNA in extracellular space. Thus, we investigated the dissociation of PICs and LITS using dextran sulphate as a model polyanion. Samples were incubated at varying concentrations of dextran sulphate to achieve different ratios between the sulphate in dextran sulphate and the phosphate in mRNA (S/P). FCS analysis revealed that PICs started to dissociate at S/P = 1, as indicated by an increase in the diffusion coefficient (Figure 1f). At S/P = 3, PIC completely dissociated to release the mRNA. In contrast, LITS remained stable even at S/P = 5, confirming its enhanced stability in the presence of polyanions. The stability of LITS was also evaluated by FCS in buffer containing 10% FBS to simulate the physiological condition. LITS displayed negligible changes in the diffusion coefficient, indicating they can maintain the structure in the buffer (Figure S10). In contrast, PICs showed a gradual increase in the diffusion coefficient, inferring their instability. The ability of the formulations to protect mRNA from nucleases in physiological environment was then evaluated by reverse transcription PCR (RT‐PCR). After 30 min incubation in 50% fetal bovine serum (FBS), both unencapsulated mRNA and the mRNA in PIC were degraded (Figure 1g), whereas LITS protected over 70% of the mRNA. These results suggest that LITS stability could be suitable for delivering mRNA under physiological conditions.

To elucidate the role of IR780 in the enhanced stability of LITS, we investigated the interactions between the polymer and mRNA by molecular dynamics simulations using three pLL derivatives with varying degrees of IR780 conjugation, i.e., 0%, 50%, and 100% of lysine residues modified. These derivatives were designated as pLL(0%IR780), pLL(50%IR780), and pLL(100%IR780), respectively. pLL(50%IR780), which represents the structure of the polymer used in LITS, exhibited more compact binding and stable interactions with mRNA than pLL(0%IR780) (Figure 1h,i and Videos S1–S3). This was evidenced by a reduced radius of gyration and stable root mean square deviation (RMSD) of the molecules (Figure S11). Increasing the IR780 conjugation to 100% lead to an extremely coiled polymer structure that is unable to interact with mRNA (Figure 1h,i and Videos S1–S3). These observations suggest that the partial IR780 substitution on the side chain of PEG‐pLL can enhance the polymer‐mRNA interactions, thereby improving complex stability.

Besides the solubility of the polymers, the photosensitizing properties also exhibit pH‐dependence. This was initially confirmed by observing the color change in the polymer solution at different pH (Figure S12). Upon acidification, the polymer solution displayed a deeper blue color, corresponding to the enhanced absorbance around 650 nm (Figure 2a,b). This shift may be due to the change of the protonation state of the secondary amines linking IR780 to the polymer.^[28^] Also, under 650 nm excitation, the polymer exhibited weak fluorescence emission at pH 7.4, likely due to the intramolecular quenching (Figure S13). However, the fluorescence of the polymer in pH 4.5 buffer was notably higher than that observed at pH 7.4.

To further understand the pH‐dependent properties, the electron distribution in the frontier molecular orbitals of the polymer under different protonation states was analyzed by density functional theory (DFT) computation (Figure 2c). The results confirmed that the energy levels of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) differed between the protonated and deprotonated states. Specifically, the HOMO–LUMO energy gap (ΔE) was reduced in the protonated state (1.89 eV) compared to de‐protonated state (2.15 eV).

Considering that a lower ΔE can enhance the nonradiation decay rate of the excited molecule,^[29^] we evaluated the photothermal and photodynamic activity of LITS under different pH conditions. Clearly, LITS demonstrated a higher photothermal conversion and ROS generation at pH 4.5 compared to pH 7.4 (Figure 2d–f). Because ROS may damage nucleic acids,^[30^] the integrity of the mRNA loaded in the LITS was also investigated after irradiation. The results showed no detectable mRNA degradation under short irradiation time (≤3 min) (Figure 2g). Importantly, the protonation of the secondary amine also alters the charge of the polymer, disrupting the electrostatic balance between polymer and mRNA, which can trigger the dissociation of LITS. To confirm this, the stability of LITS in presence of polyanion was tested under different pH by FCS (Figure 2h). The LITS at pH 4.5 exhibited a diffusion coefficient higher than that at pH 7.4 even in the absence of dextran sulphate (S/P = 0), indicating the destabilization of the particles in acidic environment. At pH 4.5 and S/P = 4, the LITS achieved a diffusion coefficient equivalent to that of free mRNA, which corresponds to a complete release of mRNA. In contrast, the LITS at pH 7.4 maintained a low diffusion coefficient across all measured S/P values. These findings indicate that LITS can sense the acidic pH of endosomal environment to facilitate mRNA release, while simultaneously activating the photosensitizing effects of the polymer without causing mRNA damage.

We first established the experimental conditions to achieve safe and efficient endosomal escape of LITS. The cellular uptake kinetics of LITS were assessed in CT26 cells by confocal laser scanning microscopy (CLSM) with LITS loading Cy5‐labeled mRNA (Figure S14). LITS showed higher cellular uptake than PIC, probably due to its enhanced stability. Moreover, the cellular uptake of LITS reached a plateau after 6 h. Thus, a 6 h incubation of the LITS with cells was selected before initiating irradiation in the subsequent experiments to ensure a sufficient internalization of LITS into cells. We also confirmed that PEG‐pLL(IR780) was not cytotoxic in dark conditions, maintaining cell viability above 90% even at a high concentration of 0.5 mg mL⁻¹ (Figure S15), which is much higher than the concentrations used for cellular transfection. Moreover, to avoid potential light‐induced ROS accumulation and cytotoxicity from prolonged irradiation of the photosensitizing polymer (Figures S16 and S17), the irradiation time and polymer concentration were screened to determine the safe irradiation time window that does not decrease cell viability (Figure S18). An irradiation time of 1.5 min was identified to be the maximum time capable of maintaining cell viability above 90% across all tested polymer concentrations. Therefore, this condition was chosen for the following investigations.

The impact of irradiation on the endosomal escape capability of LITS was then examined by CLSM using Cy5‐labeled mRNA. The cells were incubated with free mRNA, PIC and LITS for 6 h. Next, some cells incubated with LITS were irradiated for 1.5 min, then incubated for another 3 h. Comparing to free mRNA and PIC, LITS delivered more mRNA into the cells (Figure 3a,b), and decreased the co‐localization of mRNA with the endosomes (Figure 3c). Furthermore, light irradiation significantly expanded the endosomal escape of the mRNA delivered by LITS. The dynamic time lapse profile of endosome escape after light irradiation was then followed by CLSM. The results showed that light irradiation accelerated the endosomal escape with the co‐localization of LITS and the endosomes rapidly decreasing 25% in 20 min (Figure 3d). After this initial phase, the rate of endosomal escape slowed down, eventually becoming comparable to that of nonirradiated LITS. These findings indicate light irradiation induces an abrupt destabilization of the endosomal membrane to facilitate the access of mRNA into the cytosol.

To further understand the mechanism of endosomal escape, we investigated the subcellular localization of LITS using super‐resolution microscopic imaging and labelling the endosomal membranes with Rab5a‐GFP (Figure 3e). Following cellular uptake, both the polymer (blue) and mRNA (red) were detected within endosomes (green), confirming that the LITS entered cells via the endocytosis pathway.^[31^] Notably, the mRNA and polymer exhibited distinct localization patterns. mRNA was predominantly confined to the endosome lumen, whereas the polymer was able to co‐localize with the endosomal membrane. These findings suggest that at the acidic endosomal environment the LITS releases polymers that interact with the endosomal membrane. This process likely facilitates the subsequent endosome rupture via light‐induced hyperthermia and ROS generation.^[20^]

To evaluate the protein translation after irradiation, we used a model mRNA encoding enhanced green fluorescent protein (EGFP). Compared to naked EGFP mRNA and PIC, LITS loading EGFP mRNA (EGFP/LITS) induced significantly higher EGFP expression (Figures 3f and S19), which was further boosted after irradiation (Figure 3f–i). Extending the irradiation to 5 min decreased the EGFP expression level (Figure 3g–i) due to the loss of cell viability. In addition, we treated the cells with the ROS inhibitor N‐acetyl‐L‐cysteine (NAC) to investigate the roles of ROS and hyperthermia induced by irradiation in protein translation. The results showed that NAC partially reduced the transfection efficiency (Figure S20), suggesting both the ROS and the hyperthermia generated by the photosensitizing polymer contributed to the light‐induced transfection.

The in vivo performance of LITS was evaluated in mice, with the initial investigation focusing on intratumoral (i.t.) injection as a local administration route. Thus, LITS loaded with luciferase (Luc)‐encoding mRNA (Luc/LITS) was injected into CT26 tumors. 6 h after injection, a 1.5 min laser irradiation was applied to the tumor area. The expression level of Luc in the tumor was measured using bioluminescence via intravital imaging system (IVIS) (Figure 4a,b). Luc/LITS successfully transfected the tumors. The bioluminescence signal peaked at 8 h post‐irradiation, with irradiated tumors displaying around 4‐fold higher signal intensity compared to the nonirradiated group. These findings confirm the capability of LITS to boost the local protein expression with light irradiation.

LITS was then tested for intravenous (i.v.) injection applications. The blood circulation of LITS was analyzed using intravital CLSM (IVCLSM) using Cy5‐labeled Luc mRNA (Figure 4c; Videos S4 and S5). Shortly after injection, free mRNA exhibited aggregates with micrometer size, consistent with previous reports.^[32, 33^] In contrast, LITS circulated stably in the vessels without aggregation. Quantitative analysis of the fluorescence intensity in blood vessels revealed longer circulation for LITS compared to free mRNA. Furthermore, RT‐PCR analysis showed that LITS preserved the integrity of mRNA in the bloodstream, whereas free mRNA rapidly degraded after injection (Figure 4d). Following i.v. injection, LITS delivered mRNA to several major organs, including the liver, spleen, lungs, and kidneys (Figure 4e). Notably, significant accumulation of the mRNA in LITS was detected in tumors.

Considering the ability of LITS to deliver mRNA into different tissues, we then investigated the light‐induced transfection efficiency in different tissues after i.v. administration of Luc/LITS (Figure 4f). Since liver exhibited the highest mRNA accumulation, we first evaluated the effect of irradiation on the transfection efficacy in the liver of healthy mice using Luc mRNA. 6 h after i.v. injection of Luc/LITS, the liver area was externally irradiated for 1.5 min. The tissues were harvested 8 h later for bioluminescence imaging to assess the Luc expression. Comparing to nonirradiated mice, the irradiated mice displayed a stronger (∼3.5‐fold) local signal in the liver (Figure 4g), confirming that irradiation boosted the mRNA transfection in a light‐controlled manner.

The ability of systemically injected LITS to transfect CT26 tumors was then investigated (Figure 4h). The tumors were irradiated for 1.5 min 6 h after i.v. injection of Luc/LITS. Bioluminescence imaging of the tissue samples was performed 8 h later. While the bioluminescence signals from organs were comparable between the irradiated and nonirradiated mice, the signals in tumors increased by approximately 3.5‐fold upon light irradiation (Figure 4h), and the irradiated tumors exhibited higher bioluminescence intensity compared to all the organs, including the liver (Figure S21). A similar trend was observed with EGFP/LITS (Figure S22), where EGFP expression in tumor was markedly increased after irradiation. These results demonstrate the ability of LITS to achieve a precise local boost of mRNA transfection in tissues through light irradiation.

To demonstrate the therapeutic potential of LITS, we selected IL‐2 mRNA, as this cytokine exhibits dose‐dependent pleiotropic effects,^[34^] requiring precise spatial regulation throughout the body to achieve safe and effective therapeutic outcomes. The performance of LITS loaded with mRNA encoding IL‐2 (IL‐2/LITS) was first evaluated in CT26 cells in vitro (Figure S23). Following 1.5 min irradiation, IL‐2/LITS induced the highest IL‐2 expression. In vivo, i.v. injected IL‐2/LITS also demonstrated its ability to produce IL‐2 in CT26 tumors (Figure 5a). Consistent with the in vitro findings, an irradiation of 1.5 min resulted in the highest intratumoral IL‐2 levels. Thus, 1.5 min irradiation was selected for the following in vivo experiments. After systemic administration, IL‐2/LITS induced high IL‐2 levels in irradiated tumors, while maintaining moderate IL‐2 levels in blood, liver, and nonirradiated tumors (Figure 5b).

As the high IL‐2 levels may promote antitumor immunity, we studied the antitumoral efficacy of IL‐2/LITS in the CT26 tumor model (Figure 5c). 6 h after i.v. injection of IL‐2/LITS, the tumors were irradiated for 1.5 min. This treatment was repeated three times every 2 days. The tumors in the group receiving both IL‐2/LITS and irradiation (IL‐2/LITS+Light) exhibited significantly slower tumor growth compared to the other groups, including untreated mice (Blank), mice receiving only irradiation (Light), and nonirradiated mice injected with IL‐2/LITS (Figure 5d). Such antitumoral efficacy should be attributed to the priming of the tumor microenvironment (TME), as indicated by the differentiated cytokine levels in tumors after the treatments (Figure S24). Specifically, IL‐2/LITS+Light upregulated proinflammatory cytokines, including IFN‐γ, IL‐1β, and IL‐6, contributing to a shift of the TME toward a proinflammatory phenotype.^[35, 36, 37, 38^] In contrast, the other treatments failed to promote proinflammatory cytokines in the tumor. Particularly, IL‐2/LITS without irradiation increase the level of anti‐inflammatory IL‐10 in tumors, suggesting the occurrence of immunosuppression by the moderate IL‐2 expression.^[39^]

The presence of the cytokines in the TME changed the intratumoral immune infiltration landscape (Figures 5e and S25). After IL‐2/LITS+Light treatment, CD8+cytotoxic T lymphocytes (CTL) were expanded in the tumor (Figure 5e,f), while Foxp3+Tregs showed moderate change compared to control group. Notably, IL‐2/LITS without irradiation increased the Treg population in the tumor, which could be associated with the high expression of the IL‐2 receptor alpha chain (CD25) on Tregs, allowing them to respond to even small amounts of IL‐2.^[40^] Moreover, IL‐2/LITS+Light treatment showed the highest ratio between CD8+T cells and Foxp3+Tregs (CD8/Foxp3), which is a biomarker supporting the antitumoral immune response^[41, 42^] (Figure 5g). Indeed, the final tumor size and the CD8/Foxp3 ratio showed high correlation (r = −0.707, p = 0.0005) (Figure 5h). These results show that irradiated IL‐2/LITS can successfully drive the proinflammatory signals of IL‐2 for evoking an antitumoral immune response in the TME.

The systemic immune responses in blood, liver, and spleen were also analyzed. The cytokine profiling in blood revealed that the IL‐2/LITS treatment with or without tumor irradiation did not increase the inflammatory cytokines, including IFN‐γ, IL‐1β, and IL‐12 (Figure S26). However, anti‐inflammatory IL‐4 and IL‐10 were upregulated by IL‐2/LITS with or without tumor irradiation (Figures 6a and S26), suggesting the activation of systemic immunosuppressive signals.^[43^] The increase in IL‐4 and IL‐10 is known to support the expansion and function of Tregs,^[44, 45, 46^] which corresponds with the expansion of Foxp3+Tregs in blood after IL‐2/LITS treatment (Figures 6b,c and S27). On the other hand, the population of CD8+CTLs in blood remained unchanged. Analysis of immune T cell subtypes in the liver and spleen showed no significant differences between control and IL‐2/LITS‐treated groups (Figure S28). These results indicate the IL‐2/LITS avoided systemic CD8+CTL expansion and Treg suppression, which are commonly observed in high dose IL‐2 treatments and related with immune toxicity.^[47^] Indeed, the safety of the treatments was confirmed by the analysis of blood toxicity markers and histological observation, which showed no difference across all groups (Figures S29 and S30).

To directly compare the pro‐ and anti‐inflammatory signaling pathways induced by IL‐2/LITS, we studied the antitumor effects of IL‐2/LITS in a dual primary tumor model by inoculating CT26 cells on both sides of the abdomen (Figure 6d). IL‐2/LITS or Luc/LITS were i.v. injected, but the irradiation was restricted to only one tumor per mouse. Following IL‐2/LITS treatment, the irradiated tumors exhibited a high intratumoral IL‐2 concentration, successfully triggering IFN‐γ dependent inflammation (Figure S31). In contrast, the nonirradiated tumors showed increased IL‐10 levels within the TME. This differentiation in cytokine signaling led to contrasting intratumoral immune infiltration patterns within the same mouse. Irradiated tumors displayed robust infiltration by CD8+CTLs, while nonirradiated tumors were characterized by an upregulation of Foxp3+Tregs (Figures 6e and S32). These results demonstrate that the irradiation spatially confined the proinflammatory effects of IL‐2 to the treated tumor, while nonirradiated tumors remained dominated by the anti‐inflammatory signaling. This contrasting immune modulation in tumors resulted in distinct tumor progression (Figure 6f), where only the growth of the tumors at the irradiated side was suppressed (Figure 6g). Additionally, a strong correlation was observed between the final tumor volumes and the CD8/Foxp3 ratio within each tumor (Figure 6h), confirming the robust capability of IL‐2/LITS to spatiotemporally control the pleiotropic effects of IL‐2.

LITS generates ROS and hyperthermia that may directly induce damage of cells, thereby achieving photodynamic and photothermal therapies (PDT and PTT, respectively).^[48, 49^] PDT and PTT have also been reported to trigger ICD,^[50^] which can synergize with immunotherapies to enhance efficacy. Luc/LITS treated CT26 cells showed increased expression of the ICD marker calreticulin (CRT) after extending the irradiation time (Figures 7a and S33),^[51^] confirming the potential of LITS to induce ICD. Moreover, i.v. injected Luc/LITS effectively heated the CT26 tumors when irradiated (Figure 7b). Notably, just 3 min irradiation was sufficient to raise tumor temperature above 42 °C (Figure 7c), which can initiate the PTT process.^[52, 53, 54^] Histological analysis revealed that 3 min irradiation after Luc/LITS treatment upregulated CRT in tumor tissues (Figure 7d), and induced apoptosis/necrosis (Figure S34). Considering that IL‐2/LITS also provides a high IL‐2 expression level within tumors after 3 min irradiation (Figure 5a), we hypothesized that the cooperation between IL‐2 and ICD could lead to synergistic therapeutic effects. To validate this, we evaluated the efficacy of IL‐2/LITS with different irradiation time in the CT26 tumor model (Figure 7e). The combination of IL‐2/LITS treatment with light irradiation clearly inhibited tumor growth (Figure 7f). Notably, with IL‐2/LITS treatment, extending the irradiation time to 3 min significantly enhanced therapeutic efficacy compared to 1.5 min (Figure 7f), which prolonged the survival of the animals with two out of five mice achieving complete responses (CR) (Figure 7g). Immunostaining analysis revealed that IL‐2/LITS treatment with 3 min irradiation induced the highest infiltration of CD8+CTLs in CT26 tumors (Figure 7h) and significantly elevated Granzyme B level in these CTLs.^[55, 56^] As a result, IL‐2/LITS with 3 min irradiation demonstrated the strongest tumor‐damaging effects (Figure S35). Meanwhile, no significant toxicity was observed during the experiment, with all animals presenting similar bodyweight across all groups (Figure S36).

Beyond subcutaneous tumor models, the optoregulated tissue‐targeting capability of LITS enables precise treatment of disseminated tumor lesions, as shown in a lung metastasis model of triple negative breast cancer (TNBC) (Figure 7i). Upon i.v. injection of IL‐2/LITS, irradiation was applied to the chest front and back areas of the mice. This treatment successfully evoked IL‐2 expression in lungs (Figure 7j). The IL‐2/LITS+Light treatment showed effective inhibition of the lung metastases, as illustrated by macroscopic (Figure 7k) and H and E histological observations (Figure 7l). Moreover, the IL‐2/LITS+Light treatment demonstrated the longest survival of the animals, achieving CR in four out of six mice in the group (Figure 7m). In contrast, the other treatments, including IL‐2/LITS without light and Luc/LITS plus Light, failed to elicit significant survival benefits. Meanwhile, all the mice showed comparable bodyweight during the experiment (Figure S37), indicating the safety of the treatment.