T-cell activity feedback-guided photothermal immunotherapy enabled by immune-responsive redox nanodots

Abstract

Immunogenic cell death (ICD)-based tumor therapies are fundamentally limited by the absence of real-time immune activity readouts and by dysregulated oxidative stress during treatment, which together undermine immune preservation and therapeutic consistency. To address these challenges, an ultrasmall polychlorogenic acid nanodot platform, termed PCGA-ND, was developed via Fe³⁺ and TEMPO catalyzed oxidative polymerization of chlorogenic acid. The resulting PCGA-NDs exhibited a uniform size of approximately 5 nm. Compared with conventional polyphenol-based materials, PCGA-NDs exhibited significantly enhanced resistance to spontaneous oxidation, retaining over 99% of their initial electrochemical activity and optical absorbance after 15 days of storage. A granzyme B responsive fluorescent probe was constructed by covalent conjugation of a quenched peptide substrate onto the PCGA-ND surface. Under near infrared irradiation, PCGA-NDs induced effective photothermal tumor ablation and immunogenic cell death in a 4T1 breast cancer model. Meanwhile, their quinone-catechol redox activity selectively scavenged excessive extracellular ROS from mitochondrial dysfunction after heat injury and preserved cytotoxic activities of T lymphocytes recruited and infiltrated after ICD. Granzyme B-responsive fluorescence enabled in situ quantification of cytotoxic T-cell activity with high spatiotemporal resolution, yielding a signal-to-background ratio exceeding 10-fold upon enzymatic activation, thereby providing immune feedback for photothermal dose adjustment. Collectively, the strategy resulted in a tumor growth inhibition rate of 99.42% relative to control treatments in 4T1 tumor-bearing mice, accompanied by a more than 13-fold increase in intratumoral CD8⁺ T - cell infiltration. Overall, this work establishes a generalizable theranostic framework that couples photothermal therapy with immune responsive imaging and redox regulation, providing a practical strategy for improving immune preservation and therapeutic precision in photo immunotherapy.

Graphical abstract

1. Introduction

Cancer immunotherapy has reshaped the treatment paradigm for solid tumors by activating endogenous antitumor immunity [1]. However, a large proportion of patients still fail to achieve durable responses due to insufficient immune activation and immunosuppressive tumor microenvironments [2]. To address these limitations, photothermal therapy (PTT) has been widely explored as a local immunomodulatory strategy [3]. Through the conversion of near-infrared (NIR) light into heat [4], PTT can induce immunogenic cell death and promote tumor antigen release [5]. Despite these advantages, the mismatch between the photothermal dose and the therapeutic demand limited the immune activation efficacy. Moreover, acute inflammation induced by photothermal therapy compromised the antitumor effect and led to side effects such as tumor metastasis and recurrence. Therefore, regulation of inflammation within the local tumor microenvironment and real-time monitoring of the photothermal dose for immune activation were critically important for safe and effective tumor immunotherapy.

During the process of PTT, reactive oxygen species (ROS) generated by thermal stress play a key role in regulating the microenvironment of local tissue and oxidative-reduction system. The excessive oxidative stress impairs T-cell cytotoxic activity by oxidizing the surface thiol (-SH) groups on key membrane molecules such as CD4, shifting the surface redox status towards disulfide (S-S) formation and thereby disrupting receptor signaling and immune function within the tumor microenvironment [6]. Natural polyphenolic compounds, particularly those containing catechol moieties, have attracted increasing interest as building blocks for photothermal immunotherapy owing to their photothermal conversion ability and intrinsic antioxidant properties [7]. However, translating such materials into effective theraputic platforms remains challenging at both molecular and morphological levels. At the molecular level, catechol groups are susceptible to spontaneous oxidation, which compromises antioxidant stability during storage and application [8]. Although the introduction of electron-withdrawing groups can stabilize catechol rings by reducing electron density, this strategy often impedes oxidative polymerization and complicates structural control. At the morphological level, conventional polyphenol-based polymers and aggregates typically display large and heterogeneous sizes due to strong covalent cross-linking and noncovalent interactions during synthesis [9]. Such unfavorable morphologies limit tumor penetration and homogeneous intratumoral distribution, thereby restricting both photothermal efficacy and integrated theranostic performance.

Imaging of various indicator states within the tumor microenvironment plays a critical guiding role in the implementation of tumor therapeutic strategies [[10], [11], [12]]. More importantly, real-time monitoring and evaluation of the functional status of immune cells can avoid the adverse effects of insufficient or excessive photothermal therapy on antitumor immune responses. Granzyme B (GzmB) is a key effector molecule released by activated CTLs [13] and plays a central role in mediating tumor cell apoptosis [14]. The release dynamics of GzmB closely reflect the functional status and cytotoxic activity of T cells during antitumor immune responses [15,16]. Accordingly, GzmB has emerged as a practical functional biomarker for monitoring immune activation kinetics in situ [17]. However, conventional approaches for assessing GzmB activity, including tissue biopsy–based histological analysis and blood or urine measurements [18], are invasive and lack sufficient spatiotemporal resolution [19]. These limitations hinder real-time evaluation of immune responses during photothermal therapy. Therefore, the development of intelligent nanosystems capable of coupling therapeutic intervention with in situ visualization of immune activity is critically needed.

Herein, a feedback-informed photothermal immunotherapy nanoplatform was developed to address the imbalance between thermal dosage, oxidative stress, and immune activation during photothermal therapy. Ultrasmall polychlorogenic acid (PCGA) nanodots were synthesized via Fe³⁺/TEMPO-catalyzed oxidative polymerization and were subsequently functionalized with a GzmB-cleavable fluorescent peptide (containing the GzmB-specific substrate sequence Gly-Lys-Ile-Glu-Pro-Asp-Ala-Pro-Cys) [20] to enable immune-responsive imaging. Upon near-infrared irradiation, efficient photothermal tumor ablation and immunogenic cell death were induced. Excessive ROS generated during thermal stress were adaptively scavenged by the intrinsic catechol/quinone redox activity of PCGA nanodots to preserve immune cell viability and function. Meanwhile, GzmB-triggered fluorescence recovery was employed to provide a real-time, in situ readout of cytotoxic T lymphocyte activity to inform immune response dynamics during therapy. The overall design and working principle are illustrated in Scheme 1. To mechanistically validate the proposed coordination between photothermal killing and immune protection, intracellular and extracellular ROS levels, mitochondrial oxidative stress, and T-cell viability were systematically evaluated in vitro and in vivo to delineate the temporal and spatial effects of redox modulation. In parallel, GzmB-responsive fluorescence imaging was integrated with immunological and therapeutic assessments, including dendritic cell maturation, T-cell infiltration, tumor growth inhibition, and metastasis suppression, to establish causal links between immune activity readouts and therapeutic outcomes.

Scheme 1.

A schematic diagram of the preparation of polychlorogenic acid loaded with GzmB-cleaved peptide probes and the study of anti-tumor mechanism.

2. Results and discussion

2.1. Engineering of photothermal PCGA nanodots for synergistic immunotherapy

PCGA NDs were synthesized via oxidative polymerization of chlorogenic acid (CGA), which was accomplished through the synergistic catalytic action of an electron shuttle (2,2,6,6-tetramethylpiperidine-1-oxyl, TEMPO) and a transition metal compound (FeCl₃). This catalytic system is commonly employed for the highly selective transformation of alcohols into aldehydes or ketones [21,22]. To evaluate its selectivity, control experiments were conducted using Tris buffer, FeCl₃, or TEMPO individually. During these experiments, both the color changes of the reaction solutions and the kinetic absorbance curves at 600 nm [23] were recorded simultaneously (Fig. 1a and b). No significant color variation was detected in the presence of Tris buffer or TEMPO alone, even after the reaction was completed. When only FeCl₃ was introduced, the solution turned green, though absorbance changes remained minimal. In contrast, a marked increase in absorbance was observed concurrent with the formation of a black product following the addition of TEMPO and FeCl₃, attributed to the polymerization of phenolic compounds into high-molecular-weight species [24]. Following the synthesis, the morphology of the obtained product was examined using transmission electron microscopy (TEM). As shown in Fig. 1c, PCGA NDs exhibited a spherical morphology with a uniform particle size of approximately 5.13 ± 1.06 nm. The results of dynamic light scattering (DLS) measurement (Fig. 1d) further show that the hydrodynamic diameter of PCGA in aqueous solution is about 20 nm, which is larger than the size observed by TEM. The difference is mainly due to the essential difference between the hydrated diameter and the apparent particle size: the hydrodynamic diameter measured by DLS includes the contribution of the hydrated layer on the surface of the nanoparticles and the surrounding ionic atmosphere.

Fig. 1.

Preparation, photothermal properties, and ICD induction capability of PCGA NDs. (a) Color evolution of the reaction mixture at different time points. (b) Time-dependent absorbance changes at 600 nm for a 10-fold diluted reaction solution. (c) TEM image of PCGA nanodots. (d) The hydrodynamic diameter distribution of PCGA. (e) O 1s XPS spectra with the relative ratios of C-O and C=O bonds determined by peak integration. (f) Differential pulse voltammetry (DPV) curve measured in a homogeneous system (100 μg mL⁻¹). (g) Raman spectra of CGA and PCGA. (h) ESI-MS spectrum acquired in positive ion mode. (i) UV-Vis-NIR absorption spectra of CGA and PCGA (100 μg mL⁻¹). (j) Real-time infrared thermal images of PCGA (200 μg mL⁻¹) and water under 808 nm laser (1.0 W cm⁻²) (k) Live/dead staining of 4T1 cells using calcein-AM (green, viable) and PI (red, dead) after various treatments. Scale bar: 30 μm. (l) CLSM images demonstrating CRT exposure (green) on 4T1 cell surfaces. Scale bars: 25 μm. (m) Immunofluorescence detection of HMGB1 release (red) from 4T1 cells by CLSM. Scale bars: 25 μm. (n) Representative plots of mature DCs (CD11c⁺CD80⁺CD86⁺) showing enhanced maturation in response to PCGA with NIR laser irradiation.

Then, the chemical properties of PCGA NDs were characterized through multiple analytical techniques. The X-ray photoelectron spectroscopy (XPS) survey indicated the presence of C, N, O, and Fe elements in PCGA NDs (Fig. S1a). The C 1s spectrum (Fig. S1b) resolved three components at 284.49 eV, 286.25 eV and 288.58 eV, corresponding to C=C, C–O, and C=O bonds, respectively. Analysis of the N 1s region (Fig. S1c) identified peaks at 401.80 eV and 399.67 eV, assignable to pyrrolic nitrogen (C–N–C) and N–O, confirming the successful incorporation of TEMPO [25]. In the Fe 2p spectrum (Fig. S1d), the typical peaks at 710.98 eV and 725.08 eV were indexed to the Fe^Ⅱ 2p_3/2 and Fe^Ⅲ 2p_3/2, respectively, revealing the conversion between redox pairs of iron ions [26].

In the high-resolution O 1s XPS spectra (Fig. 1e), the peak at binding energy of 288.58 eV was attributed to C=O bond, and 286.25 eV could be ascribed to C‒O bond [27]. Notably, the proportion of C=O species increased significantly from 18.45% in CGA monomers to 62.95% in PCGA. According to the differential pulse voltammetry (DPV) curves, the oxidation peak current in the presence of CGA and PCGA emerged at 0.52 V and 0.61 V (Fig. 1f), respectively. The anodic shift (0.09 V) demonstrated the successful oxidation of catechol groups. In the electron paramagnetic resonance (EPR) spectrum, the peak positioned at 3510 G was detected (Fig. S2), indicating the generation of semiquinone radicals in polyphenolic polymerization reactions. Moreover, the Raman spectrum (Fig. 1g) of PCGA NDs showed two significant characteristic bands at 1351 cm⁻¹ and 1570 cm⁻¹ that were attributed to the deformation of the catechol groups [28]. Electrospray ionization mass spectrometry (ESI-MS) (Fig. 1h) revealed that monomer and oligomers could be separated and assigned, with the most abundant oligomer being a dimer at m/z 707. These results demonstrated the formation of extended π-conjugated structures during oxidative polymerization.

In addition, the optical absorption properties were investigated by ultraviolet-visible-near-infrared (UV-vis-NIR) spectrophotometry. As depicted in Fig. 1i, free CGAʼs absorption was no more than 450 nm. In contrast, PCGA exhibited a pronounced red-shift with absorption extending beyond 850 nm, consistent with an extended conjugated structure. Moreover, the adsorption intensity increased linearly with an elevation of the PCGA NDs concentration in the NIR region, and the extinction coefficient at 808 nm was determined as 2.43 L g⁻¹ cm⁻¹ (Fig. S3) [29]. Moreover, PCGA NDs were stable in different solvents, including Tris-HCl, ethanol, PBS, and MES (Fig. S4). Given the improved light-harvesting ability of PCGA, the photothermal performance was systematically evaluated. It was found that the temperature gradually increased with the increasing concentration of PCGA (0-20°C, at 10-200 μg mL⁻¹, 1.0 W cm⁻²) (Fig. S5a) after NIR irradiation for 10 min. Moreover, a positive correlation was observed between the temperature of the PCGA suspension and the power density of laser, with the former increasing steadily as the latter was elevated (0-31°C, at 0.5-1.25 W cm⁻², 200 μg mL⁻¹) (Fig. S5b). In addition, the thermal imaging results further verified the rapid production of heat during illumination (Fig. S6a and Fig. S6b). To further assess the photothermal stability of PCGA, the cycling temperature changes were recorded during five times laser irradiation/natural cooling cycles. The negligible changes in temperature profiles indicated the NIR photostability (Fig. 1j). Furthermore, the average photothermal conversion efficiency (η) of PCGA at 808 nm was calculated to be 33.22% (Fig. S7) based on the cooling curve fitting. Above all, the rationally designed PCGA NDs with extended π-conjugated structures exhibited improved photothermal effect.

Effective cellular internalization is a prerequisite for inducing intracellular phototoxicity. To verify the cellular uptake behavior, PCGA NDs were labeled with calcein and observed using confocal laser scanning microscopy (CLSM). After 12 h of incubation, distinct green fluorescence from calcein was detected within 4T1 cells, confirming the efficient internalization of PCGA (Fig. S8a). The intracellular phototoxicity of PCGA was subsequently evaluated using the cell counting kit-8 (CCK-8) assay, with graphene oxide (GO) as a control [27]. The cytotoxicity of PCGA NDs on 4T1 and L929 cells was evaluated as shown in Fig. S8b. The results showed that the biocompatibility was high because the cell viability of both cells was high at all test concentrations. Although the cell survival rate decreased slightly in a concentration-dependent manner, even at the highest concentration of 200 μg mL-1, the survival rate of 4T1 cells remained at 93.59%, and L929 cells remained above 87.89%. This consistent performance underscores the low cytotoxicity of PCGA NDs. Meantime, under NIR irradiation (1 W cm⁻², 10 min), concentration-dependent phototoxicity was observed. The cell viability in the group treated with 200 μg mL⁻¹ PCGA combined with laser irradiation dropped below 20% (Fig. S9). Thermal imaging revealed that the combination of PCGA and NIR irradiation induced a concentration-dependent temperature increase, with the highest concentration of NDs (200 μg mL⁻¹) achieving a temperature rise of 9.7°C. Furthermore, live/dead cell staining using calcein-AM and propidium iodide (PI) visually confirmed the effective destruction of cells by PCGA upon irradiation (Fig. 1k). As illustrated in Fig. S10, the PCGA + NIR group exhibited the highest level of apoptosis, with 17.70% early apoptotic and 31.70% late apoptotic cells, reaffirming the enhanced cell mortality induced by photothermal treatment.

We further investigated whether photothermal therapy could synergistically induce the exposure of danger-associated molecular patterns (DAMPs), thereby triggering ICD. CRT and HMGB1 were examined as hallmark DAMPs. As shown in As shown in Fig. 1l, CRT signal (green fluorescence) exhibited a perinuclear distribution pattern in the control, NIR-only, and PCGA-only groups, indicating minimal translocation. In contrast, the PCGA + NIR group displayed prominent CRT exposure on the cell surface, as confirmed by semi-quantitative fluorescence intensity analysis (Fig. S11a). Similarly, HMGB1 translocation was most pronounced in the PCGA + NIR group (Fig. 1m), with semi-quantitative analysis confirming the lowest nuclear HMGB1 intensity among all groups (Fig. S11b). WB results demonstrated that PCGA + NIR treatment markedly increased CRT expression while decreasing HMGB1 expression. As shown in Fig. S12, PCGA upon irradiation induced the highest CRT expression and the lowest HMGB1 expression on 4T1 cells, representing a 5.99-fold increase in CRT and a 3.90-fold decrease in HMGB1 compared to the control group, respectively. Therefore, the WB results complement the original immunofluorescence observations by providing quantitative biochemical evidence, while the immunofluorescence data provide spatial visualization of the ICD process. Together, they offer robust support for the induction of ICD by PCGA-mediated photothermal therapy. Furthermore, both GO + NIR and PCGA + NIR treatments significantly increased the population of mature dendritic cells (CD11c⁺CD80⁺CD86⁺) by approximately 2-3 fold compared to the control (Fig. 1n and Fig. S13). Notably, the PCGA + NIR group resulted in a further 14.7% increase in mature DC proportion relative to the GO + NIR group.

2.2. Synergistic anti-tumor immunity via dual ROS regulation: Amplifying ICD while protecting immune cells

In order to evaluate the storage stability of PCGA nanodots, we systematically monitored the changes of UV-Vis-NIR absorption spectra and DPV curves of PCGA nanodots stored in ethanol for up to 15 days. The UV-Vis-NIR spectra (Fig. S14a) demonstrated that the characteristic absorption of PCGA (200-800 nm) remained virtually unchanged over 15 days of storage, with no noticeable peak shift or decay, retaining 99.04% of its initial intensity. This indicates excellent stability of the conjugated backbone and the absence of irreversible auto-oxidation or degradation. To quantitatively assess the preservation of the redox-active catechol groups, we performed DPV analysis on samples stored for different durations. As shown in Fig. 2a and Fig. S14b, the characteristic oxidation peak near +0.4 V, corresponding to catechol electro-oxidation, showed negligible current loss, corresponding to a retention of 99.0% after 15 days. This exceptional electrochemical stability aligns with the behavior of molecularly engineered catechol derivatives (e.g.,4-nitrodopamine) reported previously [8]and markedly surpasses that of the readily oxidizable reference molecule dopamine. This directly confirms that the key phenolic hydroxyl functional group in PCGA has excellent stability during storage. This rich and stable catechol group ensures that PCGA can retain sufficient phenolic hydroxyl groups after storage, which provides a reliable guarantee for PCGA to effectively remove excessive ROS generated by photothermal therapy in vivo, thereby protecting T cell function from oxidative damage.

Fig. 2.

ROS Scavenging Capacity of PCGA and Its Protective Effect on T Cell Viability Post-PTT. (a) DPV curves of PCGA NDs stored in ethanol for different days. (b-f) Quantitative analysis of radical scavenging capacity for PCGA and CGA at different concentrations against: (b) DPPH•, (c) •O₂⁻, (d) H₂O₂, (e) •OH, (f) ABTS•⁺. (g) CLSM images of 4T1 cells showing mitochondria (green) and mitochondrial superoxide (MitoSOX Red, red) after treatment with PCGA NDs or GO NPs, followed by exposure to 808 nm laser irradiation (1 W cm⁻²) or no irradiation. Nuclei are stained with DAPI (blue). Scale bar: 10 μm. (h) Extracellular ROS detection against 4T1 cells using DCFH as the indicator, scale bars: 25 μm. (i) Flow cytometric analysis of intracellular ROS levels using DCFH-DA probe in PCGA- or GO-treated cells with/without NIR irradiation. (j) The schematic diagram of separation and activation of spleen T cells in mice and the T cell protection by PCGA NDs. (k) Morphology of naive versus activated mouse T cells. Scale bar: 50 μm. (l) Biocompatibility assessment of PCGA and GO on T cells. (m) Viability of T cells treated with different concentrations of PCGA or GO in the presence of 1 mM H₂O₂. (n) DCFH-DA fluorescence intensity in GO- and PCGA-treated T cells pre- and post-irradiation. (o) Crystal violet staining of 4T1 cells co-cultured with activated T cells under different treatments. Scale bar: 100 μm. (p) Survival rate of T cells after co-culture with 4T1 cells under various treatment conditions. ***p < 0.001.

Studies have shown that the increase in ROS levels induced by PTT impairs the activity and function of anti-tumor immune cells [30]. Removing excessive ROS generated during PTT is thus essential for modulating the tumor immune microenvironment and sustaining anti-tumor immunity [31]. As shown in Fig. 2b–f, S15, and S16, the scavenging activity of PCGA against various ROS was systematically evaluated. The results demonstrate that PCGA effectively scavenged 2,2-diphenyl-1-picrylhydrazyl (DPPH•), hydroxyl radical (•OH), superoxide anion (•O₂⁻), hydrogen peroxide (H₂O₂), and ABTS cation radical (ABTS•⁺), with clearance rates of 85.26%, 99.12%, 89.28%, 99.68%, and 94.38%, respectively, indicating broad-spectrum and efficient ROS elimination. Notably, PCGA exhibited higher scavenging activity than its monomer CGA toward both DPPH• and H₂O₂, further confirming that the polymerized nanostructure enhances its antioxidant capacity.

At the cellular level, mitochondria are sensitive to high temperature and are another important source of free radicals [32]. Heat stress is intertwined with oxidative stress by affecting mitochondrial function. Mitochondrial respiration is a continuous reaction system composed of a series of hydrogen transfer and electron transfer reactions in a specific order. Electrons generated by cellular metabolism are transferred along the mitochondrial electron transport chain, creating a proton gradient across the inner mitochondrial membrane. This gradient drives ATP synthesis as protons flow back down the electrochemical gradient through ATP synthase. Heat stress induces dysfunction of mitochondrial respiratory chain complexes (complexes I, II, III) function defects, leading to mitochondrial damage. This impairs ATP synthesis and causes electron transport chain dysfunction, resulting in increased permeability of the intragranular membrane, electron leakage, impaired oxidative phosphorylation, and generation of highly reactive superoxide anion radicals [33]. To investigate the pathway of ROS production under photothermal stress, we assessed the co-localization of mitochondria and superoxide in 4T1 cells following nanoparticle endocytosis and 808 nm laser irradiation. The NIR-only control group exhibited only minimal red fluorescence, due to the lack of a photothermal agent. In contrast, the GO + NIR treatment induced excessive mitochondrial oxidative stress, generating substantial superoxide and yielding the strongest red fluorescence. Conversely, the PCGA + NIR group demonstrated a significant ROS-scavenging ability, as evidenced by markedly weaker fluorescence compared to the GO group. These results are summarized in Fig. 2g and Fig. S17 alongside the Control, GO, and PCGA-only groups. To quantitatively assess the colocalization of mitochondrial superoxide (MitoSOX, red) and mitochondria (MitoTracker, green), we performed line-scan fluorescence intensity profile analysis. As shown in Fig. S18, the GO + NIR group exhibited highly synchronized spatial distribution of red and green fluorescence signals along the scanning distance (Pearson's R = 0.82), confirming massive mitochondrial superoxide production. In contrast, the PCGA + NIR group showed markedly reduced red signal intensity that did not follow the green signal distribution, resulting in a significantly lower Pearson's correlation coefficient (R = 0.38), which quantitatively demonstrates the effective reduction of mitochondrial oxidative stress by PCGA upon NIR irradiation. We further validated the ROS scavenging ability of PCGA at the cellular level using the DCFH-DA fluorescent probe (Fig. S19). In addition, the cell-impermeable probe DCFH was employed to monitor extracellular ROS levels during the early to intermediate phases after irradiation. As shown in Fig. 2h, substantial extracellular ROS generation was observed following light exposure. After different incubation times with GO or PCGA, the fluorescence intensity in the GO group remained largely unchanged at 6 h, whereas that in the PCGA group gradually decreased. At 12 h, the GO group still exhibited strong fluorescence, while only a weak signal was detected in the PCGA-treated group, indicating that PCGA effectively scavenges extracellular ROS. Concurrently, the intracellular ROS levels were evaluated using the membrane-permeable probe DCFH-DA. Flow cytometry analysis revealed that the ROS-positive rate was 42.3% in the GO + NIR group, but only 7.33% in the PCGA + NIR group (Fig. 2i). This dual-compartment ROS-scavenging capacity of PCGA is attributed to its ability to both enter cells via endocytosis and bind to the extracellular matrix through hydrogen bonding between its phenolic hydroxyl groups and collagen [34], thereby establishing a more favorable microenvironment for subsequent immune cell infiltration and directly protecting T-cell function from oxidative damage in both compartments.

To assess the protective effect of PCGA on immune cells, T cells were isolated from mouse splenocytes (Fig. 2j) and activated with concanavalin A. In contrast to primary T cells, activated T cells exhibited pronounced cluster aggregation [35]. (Fig. 2k). While neither PCGA nor GO showed significant cytotoxicity toward T cells at concentrations up to 200 μg mL⁻¹ (Fig. 2l), T cell viability dropped to 20% in the presence of H₂O₂ when pre-treated with GO. In contrast, PCGA NDs pre-treatment maintained T cell viability above 90% under the same oxidative stress conditions (Fig. 2m), underscoring its role in protecting T cells via ROS removal. Consistent with this, the DCFH-DA signal in the GO group increased dramatically upon laser irradiation, reaching 57-fold that of the PCGA group (Fig. 2n). Finally, we evaluated T cell-mediated tumor killing (Fig. 2o and p). Crystal violet staining revealed distinct outcomes across treatment groups. In the PCGA + NIR + T cell group, the majority of 4T1 tumor cells were destroyed. This enhanced cytotoxicity is attributed to the ROS-scavenging capability of PCGA, which maintained T cell activity and potentiated tumor cell killing. In contrast, due to the lack of such ROS regulation, a substantial portion of tumor cells remained intact in the GO + NIR + T cell group, retaining 49.41% of the cell integrity relative to the control group (Fig. S20).

2.3. Real-time monitoring of therapeutic efficacy through GzmB activity imaging

The GzmB-responsive peptide, labeled with a TAMRA fluorophore, was conjugated to PCGA nanodots through a Michael addition reaction between the terminal cysteine thiol group of the peptide and the quinone moieties on PCGA, resulting in the PCGA-Pep probe. Successful conjugation was confirmed by Raman spectroscopy (Fig. 3a), which displayed the emergence of a C-S bond vibration near 600 cm⁻¹ [36]and the disappearance of the -SH stretching band at 2500 cm⁻¹. UV-Vis-NIR spectral analysis confirmed the conjugation. Both the PCGA-Pep conjugate and the Pep showed an identical absorption peak at 550 nm (Fig. 3b), ascribed to the characteristic absorption of the TAMRA label on the peptide. Furthermore, the TAMRA fluorescence was effectively quenched upon conjugation with PCGA (Fig. 3c), demonstrating a strong quenching interaction between the two components. UV-Vis-NIR spectroscopy further validated the conjugation, showing a characteristic absorption peak at 530 nm for both PCGA-Pep and the free peptide, which was absent in CGA and PCGA alone. Hydrodynamic size analysis (Fig. S21a) indicated an increase in particle diameter from approximately 20 nm for PCGA to about 60 nm after peptide attachment. Additionally, zeta potential measurements (Fig. S21b) revealed a shift in surface charge from −5.3 mV (Pep) and −2.1 mV (PCGA) to −17.4 mV for PCGA-Pep, consistent with successful covalent linkage.

Fig. 3.

Characterization of PCGA-Pep as a granzyme B-responsive probe. (a) Raman and (b) UV-Vis-NIR spectra confirmed the successful conjugation of Pep to PCGA. (c) Fluorescence Quenching following Peptide Attachment (λex/λem = 570/750 nm). (d) Schematic illustrates GzmB-specific cleavage of PCGA-Pep. (e) IVIS imaging and (f-h) fluorescence spectra demonstrated concentration- and time-dependent fluorescence recovery upon GzmB incubation (0 - 30 pM, 4 h). (i) Michaelis-Menten kinetics analyzed GzmB activity toward PCGA-Pep (10-300 μg mL⁻¹). (j,k) Specificity tests showed minimal response to interferents (Caspase-3, Cathepsin C, MMP-2, DNase I, GSH, H₂O₂). ns = not significant, ***p < 0.001.

We evaluated the fluorescence response of PCGA-Pep to GzmB, whose cleavage mechanism is illustrated in Fig. 3d. Upon incubation with GzmB (30 pM), In Vivo Imaging System (IVIS) imaging confirmed probe activation (Fig. 3e), accompanied by a marked increase in fluorescence emission at 575 nm (Fig. S22). Fluorescence titration revealed a concentration-dependent enhancement in signal intensity with increasing GzmB levels (5-30 pM, Fig. 3f), and a strong correlation was observed between fluorescence intensity and enzyme concentration (Fig. 3g). The GzmB imaging probe was incubated with the activated recombinant GzmB at 37°C, and the fluorescence activation of the quenched dye after cleavage was measured at different time points. As shown in Fig. 3h, the fluorescence intensity was observed to increase gradually over time, resulting in a 4-fold increase in fluorescence intensity within 4 h, indicating that the probe operates on a FRET-based mechanism, and can lead to the activation of the quenched dye after digestion. To quantify enzymatic activity, a standard curve was established by measuring the fluorescence of free peptide (Pep) across a concentration gradient (0-200 μg mL⁻¹, Fig. S23). Various concentrations of PCGA-Pep were then incubated with 30 pM GzmB, and the reaction rates were derived from the fluorescence data. Michaelis–Menten analysis yielded a Km of 27.18 μg mL⁻¹ and a kcat of 0.0109 s⁻¹, corresponding to a catalytic efficiency (kcat/Km) of 4.02 × 10⁴ M⁻¹ s⁻¹ (Fig. 3i). Specificity was assessed against a panel of proteases and biomolecules. While GzmB induced a 13.44-fold increase in fluorescence, no significant response was observed with Cathepsin C, Caspase-3, MMP-2, DNase I, GSH, or H₂O₂ (Fig. 3j). This selectivity was corroborated by IVIS imaging (Fig. 3k), confirming that PCGA-Pep enables specific and sensitive monitoring of GzmB activity.

Based on the excellent response of PCGA-Pep NDs to GzmB in vitro, we further investigated their capability for imaging GzmB-mediated CTLs antitumor responses. We employed PCGA-Pep NDs to visualize CTL activity against 4T1 tumor cells in a co-culture system, where T cells were used to induce an antigen-specific immune response against the tumor cells. To maximize GzmB production and subsequent PCGA-Pep activation, we first optimized the effector-to-target (E:T) ratio in the co-culture system. After co-incubating varying numbers of T cells with 4T1 cells for 12 h, PCGA-Pep NDs (200 μg mL⁻¹) were added. As shown in Fig. S24, negligible fluorescence was observed in 4T1 cells cultured without T cells. As the proportion of T cells increased, the turn-on fluorescence signal of PCGA-Pep in 4T1 cells gradually intensified, reaching a maximum at an E:T ratio of 4:1. Using this optimized ratio (T cells:4T1 cells = 4:1), we monitored the activation kinetics of PCGA-Pep NDs in 4T1 cells. Results in Fig. S25 showed that the fluorescence signal gradually increased over time, approaching a plateau at 12 h. Although substantial fluorescence was still detectable when the co-culture was extended to 24 h, the tumor cells were severely damaged by T cell attack, making them unsuitable for further fluorescence imaging analysis. Therefore, a 12 h co-culture period was selected for all subsequent experiments to monitor the PCGA-Pep activation process in 4T1 cells.

Based on the optimized co-culture conditions, we further evaluated the responsiveness of PCGA-Pep NDs to GzmB secretion in a 4T1/T-cell co-culture system. After 12 h of incubation with PCGA-Pep NDs, both 4T1 cells and culture supernatants were collected for confocal imaging and fluorescence quantification (Fig. 4a). Parallel ELISA measurements of GzmB in the supernatant were used to establish a standard curve for absolute quantification (Fig. S26a). The results indicated that GzmB release increased with higher effector-to-target ratios and longer co-culture durations (Fig. 4b and Fig. S26b), reaching maximum concentrations of 380 pg mL⁻¹ and 391 pg mL⁻¹ at a 4:1 ratio and 24 h time point, respectively. Consistent with these data, confocal imaging showed enhanced fluorescence signals from both GzmB and activated PCGA-Pep in the 4:1 co-culture group (Fig. 4c). Co-localization analysis confirmed spatial overlap between GzmB and PCGA-Pep signals (Fig. 4d). The fluorescence intensities of both signals were observed to increase over time (Fig. 4e).

Fig. 4.

Fluorescence Imaging and Detection of GzmB Produced by CTLs. (a) Schematic overview of the experimental workflow for confocal imaging and extracellular fluorescence detection. (b) ELISA quantification of GzmB in the supernatant of 4T1-T cell co-cultures at different effector-to-target ratios. (c, d) CLSM and co-localization images of 4T1 cells loaded with PCGA-Pep (200 μg mL⁻¹) after 12 h of co-incubation with T cells at varying ratios (Scale bar: 10 μm). (e-g) Time-dependent CLSM images and corresponding fluorescence intensity analysis of GzmB (red) and TAMRA-labeled PCGA-Pep (green) in 4T1-T cell co-cultures; nuclei were stained with DAPI (blue), Scale bar: 25 μm. (h, i) Fluorescence spectra of PCGA-Pep in 4T1-T cell co-cultures under different ratios and incubation times. **p < 0.01, ***p < 0.001.

Semi-quantitative analysis of the time-dependent confocal images further demonstrated this increasing trend for GzmB signal (Fig. 4f). A similar temporal pattern was also quantified for the activated PCGA-Pep signal (Fig. 4g), which was additionally supported by the analysis presented in Fig. S27. Furthermore, fluorescence measurements of the co-culture supernatant yielded concordant results (Fig. 4h and i). Together, these findings verify efficient cleavage of the GK-IEFD-APC motif in PCGA-Pep by GzmB secreted from activated T cells.

2.4. In vivo imaging and therapy monitoring of PCGA-Pep: Insights into long-term immune mechanisms

To enable precise photothermal therapy, an imaging-capable photothermal agent is essential for distinguishing tumor lesions from normal tissues. We therefore assessed the in vivo tracing ability of PCGA-Pep NDs in 4T1 tumor-bearing mice using whole-body fluorescence imaging. The biodistribution of Cy5-labeled PCGA NDs was monitored over time (Fig. 5a), revealing a significant increase in tumor fluorescence up to 6 h post-injection (Fig. S28a). Quantitative analysis showed a gradual decline in the whole-body fluorescence signal by 9 h, while a high level of ND accumulation was maintained at the tumor site (Fig. 5b). Ex vivo imaging of major organs at 24 h indicated that tumor fluorescence intensity was only lower than that in the liver and spleen (Fig. S28b). The strong hepatic and splenic signals suggested a typical hepatobiliary clearance pathway. Notably, even after 24 h, a strong fluorescence signal persisted in the tumor tissue (Fig. S28c).

Fig. 5.

In Vivo Distribution, Phototherapy, and Treatment Monitoring of PCGA NDs. (a) Schematic of in vivo fluorescence imaging following intravenous Cy5-labeled PCGA injection in tumor-bearing mice. (b) Time-dependent quantification of tumor fluorescence intensity (n = 4). (c) Photothermal imaging and temperature profiles of tumors after 10 min irradiation at 6 h post-injection (n = 3) (d) Integrated strategy for imaging-guided treatment. (e-f) Representative images of small animals after PCGA-Pep injection and 808 nm laser irradiation (1W cm⁻²) and corresponding temporal quantification of tumor fluorescence. (g) Ex vivo organ fluorescence at the terminal time point. (h) Representative in vitro fluorescence images of major organs and tumors were collected within 18 days after photothermal therapy. (i) GzmB immunofluorescence was performed on tumor sections after a single photothermal treatment (n = 6). (j, k) Representative images of small animals after three times of PCGA-Pep injection and 808 nm laser irradiation (1 W cm⁻²) and corresponding temporal quantification of tumor fluorescence. (l) Fluorescence of isolated organs at the terminal time point after three illuminations. (m) Representative in vitro fluorescence images of major organs and tumors were collected within 24 days after PTT. (n) GzmB immunofluorescence in tumor sections after a single session of PTT (n = 6). ns = not significant, **p < 0.01, ***p < 0.001.

To directly verify the advantage of the ultrasmall size for tumor penetration, we compared the intratumoral distribution of Cy5-labeled PCGA nanodots (∼5 nm) with Cy5-labeled PDA nanoparticles (∼100 nm) as a conventional-sized control [[37], [38], [39]]. PCGA nanodots achieved complete penetration throughout the tumor tissue with strong fluorescence in deep interior regions (depth >1600 μm), whereas PDA nanoparticles were predominantly localized at the tumor periphery with signal diminishing toward the interior (Fig. S29). These results demonstrate that the ultrasmall size of PCGA nanodots is critical for deep tumor penetration and homogeneous distribution, which are essential for subsequent photothermal and immunomodulatory efficacy.

This favorable biodistribution profile allows effective mild photothermal therapy within an optimal time window. To further investigate the effectiveness of PCGA-Pep in photothermal immunotherapy in vivo, PTT was administered to tumor sites using an 808 nm laser (1 W cm⁻²) at 6 h post-initial intravenous injection of NDs. The temperature changes at the tumor region were monitored with an infrared camera (Fig. 5c). Within 10 min of irradiation, the local temperature at NDs-treated tumors increased by over 17.3°C, while surrounding normal tissues exhibited minimal temperature increase. This significant thermal contrast demonstrates effective tumor-specific accumulation of the NDs and their capacity to mediate localized photothermal ablation without damaging adjacent healthy tissue. In contrast, tumors in PBS-treated mice showed temperature rises within only 5°C under identical irradiation conditions (Fig. S30), confirming the essential role of NDs in achieving therapeutic hyperthermia.

To further investigate the expression dynamics of immune effector proteins at tumor sites following combined immunotherapy and photothermal treatment, we performed systematic experiments according to the schematic timeline in Fig. 5d. PCGA-Pep was intravenously administered every two days for 20 days to mice receiving the combined treatment regimen. Longitudinal fluorescence imaging using a small animal imaging system revealed temporally dynamic activation patterns (Fig. 5e–S31a). GzmB-mediated fluorescence signals initially emerged on day 3, gradually intensified to peak intensity by day 5, and subsequently declined to minimal levels by day 18. Quantitative analysis corroborated this kinetic profile (Fig. 5f), indicating initiation of immune infiltration and cytotoxic molecule activity by day 3 post-PTT, followed by a decline phase beginning on day 5 that coincided with tumor regrowth and gradual volume expansion. Quantitative assessment of the terminal ex vivo organ fluorescence revealed no statistically significant differences in distribution among organs (Fig. 5g). Subsequent ex vivo imaging of major organs and tumor tissues (Fig. 5h and Fig. S31b) demonstrated predominant accumulation of GzmB-associated fluorescence in tumor regions, with secondary distribution observed in renal tissues. Immunofluorescence analysis of tumor sections confirmed moderate granzyme B infiltration (Fig. 5i). This established PCGA-Pep as an effective sensor for visualizing GzmB expression during immune activation, and it also exhibited favorable clearance characteristics. When we modified the treatment protocol to incorporate multiple photothermal sessions (Fig. 5j and Fig. S32a), we observed sustained enhancement of GzmB fluorescence signals from day 2 through day 8 after initial PTT, corresponding with extensive tumor ablation. This sustained signal enhancement was further quantified over time (Fig. 5k). A third irradiation administered on day 9 further augmented GzmB signals through day 12, followed by a gradual decline, despite nearly complete tumor elimination. Continuous imaging monitoring demonstrated maintained signal elevation through day 24. This biodistribution profile was quantitatively verified, showing significantly higher fluorescence intensity at tumor sites compared to major organs (Fig. 5l). Subsequent ex vivo tissue imaging (Fig. 5m and Fig. S32b) confirmed predominant tumor-localized fluorescence with minor renal accumulation. Finally, immunofluorescence analysis revealed substantial GzmB infiltration in tumor sections (Fig. 5n). These comprehensive results demonstrate that GzmB fluorescence imaging enables real-time assessment of photothermally induced immune activation, facilitating optimized treatment scheduling that achieves both efficient tumor ablation and sustained immune surveillance to eliminate residual malignant cells. This approach provides a valuable strategy for monitoring and guiding combination cancer therapies. To further evaluate the spatial relationship between T-cell infiltration and probe activation, we performed immunofluorescence colocalization analysis of CD8⁺ T cells and activated PCGA-Pep (TAMRA) on tumor sections from single- and multiple-irradiation groups (Fig. S33). In the single-treatment group, both CD8⁺ and TAMRA signals were sparse with limited colocalization. In striking contrast, the multiple-treatment group exhibited substantially enhanced CD8⁺ and TAMRA signals with prominent spatial colocalization, indicating that multiple photothermal sessions promoted sustained T-cell infiltration and prolonged GzmB activity in the tumor microenvironment.

Excellent biosafety is a fundamental prerequisite for in vivo applications. To this end, we first evaluated the hemocompatibility of PCGA-Pep NDs. Even at a concentration of 200 μg mL⁻¹, the hemolysis rate remained below 10%, which is well within the safe range (Fig. 6a). To further verify the in vivo biosafety, healthy mice were intravenously injected with PCGA-Pep NDs and major organs were harvested on days 1, 3, 5, 7, and 14 for histopathological analysis. Hematoxylin and eosin (H&E) staining revealed no significant physiological abnormalities (Fig. S34), indicating no substantial systemic toxicity associated with the NDs. Serum biochemical analysis further confirmed the absence of notable hepatorenal toxicity (Fig. S35), supporting the good biocompatibility of PCGA-Pep NDs. These results also suggest efficient clearance of NDs from blood and organs, providing a theoretical basis for their safe use in living systems.

Fig. 6.

In vivo imaging-guided evaluation of therapeutic efficacy. (a) The representative picture and hemolysis percentage of red blood cells treated with different concentrations of PCGA-Pep. Data were expressed as mean ± SD (n = 3). (b) Body weight curves measurements post-treatment. (c) Individual tumor growth kinetics across different groups. (d) Tumor weight measurements post-treatment. (e-i) H&E, TUNEL, and Ki67 staining of tumor sections collected after treatment (scale bar: 50 μm). (f, h) Quantitative analysis of Ki67-positive fluorescence ratio and TUNEL-positive cell percentage, respectively. All data are presented as mean ± SD (n = 6). (j) H&E staining of lung sections from different treatment groups in the photothermal anti-metastasis study (scale bars: 1 mm and 100 μm). (k-n) Serum levels of IFN-γ, IL-6, TNF-α, and TGF-β quantified by ELISA in each group post-treatment. ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001.

To simulate clinical decision-making based on real-time diagnostic results, female mice bearing 4T1 tumors were randomly allocated into six treatment groups: PBS, NIR alone, GO nanoparticles, PCGA nanodots, GO + NIR, and PCGA + NIR. Body weight was monitored throughout the treatment period to assess systemic toxicity and animal health status; no significant body weight reduction was observed in any group relative to the PBS control (Fig. 6b), indicating minimal systemic toxicity from all treatments. For the evaluation of tumor growth kinetics, subcutaneous tumor volumes were measured every two days during the entire treatment course. The individual volume-time curves (Fig. 6c) showed that both the NIR-alone and nanoparticle-alone (GO or PCGA) groups exhibited only modest inhibition compared with PBS. In contrast, the GO + NIR combination induced partial tumor ablation, while PCGA + NIR achieved complete ablation of primary tumors and significantly enhanced growth suppression over GO + NIR, ultimately reaching a tumor inhibition rate of 99.42%; other treatments exerted no significant inhibitory effect on tumor progression. These volume-based findings were further corroborated by the final tumor weights measured post-treatment using excised tumor tissues (Fig. 6d), which showed corresponding differences across groups and confirmed the superior efficacy of PCGA + NIR, with no overt toxicity as reflected by the stable body weights.These findings demonstrate the superior therapeutic efficacy of PCGA-mediated photothermal therapy while maintaining an favorable safety profile.

Comprehensive histopathological evaluation was performed on harvested tumor tissues using hematoxylin-eosin (H&E) staining, Ki67 proliferation marker staining, and TUNEL apoptosis assay to assess tissue damage, proliferative activity, and apoptotic levels, respectively. H&E staining of tumors from the PCGA + NIR group revealed extensive tissue destruction characterized by numerous cavities and prominent nuclear pyknosis (Fig. 6e). Furthermore, this group exhibited the lowest percentage of Ki67-positive cells (22.87%, brown staining) (Fig. 6f and g), indicating potent inhibition of cellular proliferation. The TUNEL assay demonstrated the most intense red fluorescence signal in the PCGA + NIR group (80.46% positivity) (Fig. 6h and i), confirming massive apoptotic cell death. To investigate immune activation within the tumor microenvironment, GzmB immunostaining was performed across all treatment groups. GO + NIR group showed modest GzmB infiltration compared to other control groups (Fig. S36). However, this infiltration was substantially lower than that observed in PCGA + NIR group. This attenuated immune response in the GO + NIR group can be attributed to the accumulation of excessive ROS generated during photothermal treatment, which impairs T-cell function and compromises antitumor immunity. In contrast, PCGA's superior ROS-scavenging capacity preserves T-cell viability and activity, enabling robust immune-mediated tumor clearance.

We further investigated the ability of PCGA nanodots to suppress tumor metastasis following combination treatment. Lung tissues were harvested from tumor-bearing mice at day 24 to capture circulating tumor cells effectively. H&E staining analysis of complete lung sections (Fig. 6j) revealed numerous deeply stained nuclear clusters in the PBS, NIR, PCGA, and GO + NIR groups, indicating substantial metastatic burden. In striking contrast, the PCGA + NIR group exhibited the smallest metastatic area. After Bouin's dye staining, pulmonary metastatic nodules appearing as white lesions were quantified, and the corresponding inhibition rates were calculated (Fig. S37). The PBS control group showed an average of approximately 46 pulmonary nodules, whereas PCGA + NIR treatment completely suppressed metastasis, reducing the count to zero. Compared to the negative control, the PCGA + NIR regimen achieved 94.20% inhibition of lung metastasis, substantially surpassing the efficacy of GO + NIR treatment (59.42%)(Fig. S38). These findings demonstrate the remarkable capacity of PCGA-mediated photoimmunotherapy to prevent tumor dissemination and metastatic progression. Furthermore, groups exhibiting poor therapeutic outcomes demonstrated widespread splenomegaly (Fig. S39a), likely resulting from excessive systemic inflammation - a characteristic syndrome in mice with highly metastatic tumors. In contrast, the PCGA + NIR group maintained normal splenic architecture with an average spleen weight approximately one-third of that observed in the PBS control group (Fig. S39b). This preservation of normal splenic morphology and weight underscores the favorable safety profile of PCGA-mediated photoimmunotherapy and its capacity to mitigate cancer-associated systemic inflammation.

We further investigated systemic immune responses by measuring serum levels of key pro-inflammatory cytokines (TNF-α, IFN-γ, IL-6) following various treatments (Fig. 6k–m). The GO + NIR group exhibited substantially elevated levels of these cytokines, reaching 2.7-4.6 times those of the control group, indicating the induction of an acute and widespread inflammatory state following photothermal treatment. In addition, compared with traditional photosensitizers, PCGA-mediated photothermal therapy significantly reduced the expression of pro-inflammatory cytokines by about 2-fold [40]. In contrast, the PCGA + NIR group demonstrated markedly reduced expression of pro-inflammatory cytokines (0.9-1.2 times control levels), suggesting well-controlled inflammatory activation. This moderated inflammatory environment supports the survival and function of immune effector cells, enabling sustained antitumor immunity and inhibition of tumor recurrence. Interestingly, we observed a significant increase in the immunosuppressive cytokine TGF-β (Fig. 6n), which may reflect a compensatory regulatory mechanism activated in response to robust antitumor immunity. This cytokine profile shift demonstrates that PCGA-based photoimmunotherapy successfully activates a potent yet balanced systemic antitumor immune response while avoiding excessive inflammation that could impair immune function.

To further investigate the long-term efficacy and systematic mechanisms of photoimmunotherapy, we examined key immunological processes in the tumor microenvironment. Initial assessment of intratumoral ROS levels using DHE staining revealed a significant reduction in red fluorescence signal in the PCGA + NIR group compared to the GO + NIR group (Fig. 7a). Immunofluorescence analysis of DAMPs revealed distinct patterns of ICD induction between treatment groups. The PCGA + NIR group exhibited the lowest nuclear HMGB1 signal (Fig. 7b) alongside the most intense CRT surface exposure (Fig. 7c), indicating robust ICD induction. Furthermore, infiltration of dendritic cells (CD80⁺CD86⁺) in the tumor microenvironment was visualized. The signals for CD80 (red) and CD86 (green) were substantially stronger in the PCGA + NIR group than in other groups (Fig. 7d). Semi-quantitative analysis confirmed these observations: it showed a 61.59% decrease in DHE fluorescence intensity (Fig. 7e), a 20% decrease in HMGB1 mean fluorescence intensity (MFI), and an approximately 15% increase in CRT MFI in the PCGA + NIR group compared to the GO + NIR group (Fig. 7f and g). Quantitative assessment of dendritic cell infiltration further demonstrated that the CD80/CD86 fluorescence intensity in the PCGA + NIR group was approximately twice as high as that in the GO + NIR group (Fig. 7h). Together, these data demonstrate PCGA's remarkable capacity to effectively scavenge photothermally generated ROS at tumor sites and significantly enhance ICD induction. This ROS-eliminating function enables remodeling of the tumor microenvironment, establishing favorable conditions for subsequent immune cell recruitment and activation [41]. The enhanced DC maturation suggested by immunofluorescence was further quantified by flow cytometric analysis of CD11c⁺CD80⁺CD86⁺ DCs in tumor-draining lymph nodes (TDLNs) (Fig. S40). As shown in Fig. S41 and S42, the percentage of mature DCs reached 47.9% in the PCGA + NIR group, representing a 1.33-fold increase compared to the GO + NIR group.

Fig. 7.

In Vivo Immunogenic Cell Death Induction and Immune Activation Following Photothermal Therapy. (a) ROS detection in tumor frozen sections by DHE fluorescence staining. (b) HMGB1 release (red) and (c) CRT exposure (green) visualized by immunofluorescence in tumor tissues. (d) Infiltration of dendritic cells (blue: DAPI; green: CD80; red: CD86) in the tumor microenvironment. (e-h) Semi-quantitative analysis of fluorescence intensity for DHE, HMGB1, CRT, and CD80/CD86, respectively, performed using ImageJ (mean ± SD, n = 5). (i) Infiltration of CD4⁺ (red) and CD8⁺ (green) T cells in tumor tissues post-treatment. Scale bars: 50 μm. (j) Semi-quantitative analysis of CD4/CD8 fluorescence intensity. (k) Flow cytometric analysis of mature dendritic cells (CD4c⁺CD8⁺) in tumor. Statistical significance was determined by one-way ANOVA with Tukey's post hoc test (ns = not significant, **p < 0.01, ***p < 0.001).

Subsequently, we investigated the activation of adaptive immune responses. Immunofluorescence staining revealed that PCGA + NIR treatment significantly enhanced the infiltration of both CD8⁺ (green) and CD4⁺ (red) T cells within tumor tissues (Fig. 7i). Quantitative analysis demonstrated that the proportion of CD8⁺ T cells, which primarily mediate cytotoxic immune functions, increased to 17.4% in the PCGA + NIR group (Fig. 7j). This represented a substantial enhancement compared to GO + NIR treated mice (12.2%) and the PBS control group (1.31%), and an increase of 5.51 times compared with traditional photosensitizer treatment [42]. Simultaneously, the results of flow cytometry showed (Fig. 7k) that the PCGA + NIR group exhibited the highest abundance of intratumoral CD4⁺ T cells, with counts measuring 6.80-fold, 3.5-fold, 4.96-fold, and 2.79-fold higher than those in the PBS, NIR, PCGA, and GO + NIR groups, respectively. These findings collectively demonstrate that PCGA nanodots significantly enhance ICD induction and promote antitumor immune responses through efficient DAMPs release and surface exposure, effectively promoting DC maturation and infiltration. The coordinated actions of ROS scavenging and enhanced ICD induction establish PCGA-mediated photoimmunotherapy as a potent approach for reversing immunosuppressive conditions and activating sustained antitumor immunity. The results verify that PCGA nanodots effectively potentiate the transformation of immunologically “cold” tumors, thereby amplifying the efficacy of immunotherapy through enhanced T cell recruitment and activation within the tumor microenvironment. microenvironment.

3. Conclusion

In summary, this work introduces an ultrasmall theranostic nanoplatform that integrates photothermal tumor ablation, redox-adaptive microenvironment regulation, and immune-responsive imaging within a unified design. The PCGA nanodots demonstrate that photothermal efficacy and immune preservation are not mutually exclusive but can be concurrently achieved through precise control of interfacial redox chemistry. By selectively mitigating extracellular oxidative stress while maintaining effective immunogenic cell death, the platform establishes a balanced therapeutic regime that sustains CTLs function during photothermal intervention. Importantly, embedding granzyme B-responsive activation into the nanomaterial architecture enables immune activity to be directly visualized and quantified in situ, transforming immune response from a retrospective outcome into an actionable therapeutic parameter. Beyond its specific implementation, this work defines a general design paradigm in which immune-state sensing and microenvironment regulation are intrinsically coupled. Thus, it offers a conceptual framework for the development of next-generation photo-immunotherapeutic systems with enhanced precision, controllability, and translational relevance.

4. Experimental section

4.1. Synthesis of poly(Chlorogenic acid) (PCGA)

The synthesis of PCGA was performed with slight modifications to the literature method, using TEMPO and Fe³⁺ as oxidants. The procedure was as follows: First, TEMPO (78.1 mg) was dissolved in 15 mL of deionized water, and CGA (74.34 mg) was dissolved in 4 mL of anhydrous ethanol. The two solutions were combined and vigorously stirred to form a homogeneous mixture. Then, 1 mL of an aqueous FeCl₃ solution (3.2 mg mL⁻¹) was added [43]. After reacting for 18 h, the mixture was transferred to a dialysis bag and dialyzed to remove impurities. The final product was obtained as particles after freeze-drying.

4.2. Simultaneous detection of mitochondrial quantity and mitochondrial superoxide in 4T1 cells following 808 nm laser treatment

4T1 cells were seeded into confocal dishes at a density of 8 × 10⁵ cells/mL in DMEM supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin, and cultured overnight at 37°C in a 5% CO₂ incubator. The cells were divided into six experimental groups: control, NIR only, GO only, PCGA only, GO + NIR, and PCGA + NIR. For the groups receiving near-infrared (NIR) laser treatment (NIR, GO + NIR, and PCGA + NIR), 4T1 cells were seeded and cultured overnight to allow adhesion. The cells were then incubated with 200 μg mL⁻¹ GO or PCGA for 12 h. After incubation, the medium was removed, and the cells were washed three times with PBS to eliminate non-internalized nanoparticles. Fresh medium was then added, and the cells were continuously irradiated with an 808 nm near-infrared laser at a power density of 1.0 W cm⁻² for 10 min. Subsequently, the cells were incubated with a mixture of Hoechst 33342 (1 μg mL⁻¹) and MitoTracker™ Green FM (100 nM) at 37°C under 5% CO₂ for 30 min. After removal of the staining solution, the cells were treated with MitoSOX™ Red working solution (2.5 μM) and further incubated for 15 min under the same conditions. Finally, the cells were visualized using laser scanning confocal microscopy.

4.3. DPV characterization in a heterogeneous system using a film-modified electrode

A 50 μL aliquot of a 1 mg mL⁻¹ PCGA ethanol suspension was added dropwise onto the surface of the glassy carbon electrode. After the solvent had evaporated to form a film, the electrode was placed in a three-electrode system containing 3 mL of electrolyte (20 mM Tris-HCl, pH 7.4). Among them, a glassy carbon working electrode, an Ag/AgCl reference electrode, and a platinum wire counter electrode. The DPV test was performed in the potential range of 0 to 1.0 V (relative to Ag/AgCl).

4.4. Detection of extracellular ROS

4T1 cells were seeded into confocal dishes at a density of 5 × 10⁵ cells per dish and cultured overnight to allow adhesion. The cells were then incubated with 200 μg mL⁻¹ GO for 12 h, washed three times with PBS, and irradiated continuously with an 808 nm near-infrared laser at a power density of 1.0 W cm⁻² for 10 min. After irradiation, the medium was replaced with fresh medium containing either 200 μg mL⁻¹ PCGA or GO and further incubated. At 0, 6, and 12 h post-treatment, the DCFH probe (10 μM) and Hoechst33342 were added to each dish and incubated for 30 min in the dark. Finally, the cells were observed and imaged under a laser scanning confocal microscope, and extracellular ROS levels were assessed based on green fluorescence intensity.

4.5. BMDC culture and maturation assay

Bone marrow-derived dendritic cells (BMDCs) were isolated from mouse bone marrow mononuclear cells and differentiated in RPMI-1640 medium supplemented with recombinant murine GM-CSF (10 ng mL⁻¹) and IL-4 (10 ng mL⁻¹) for 7 days. Subsequently, the BMDCs were co-cultured for 24 h with supernatants collected from 4T1 cells that had been subjected to various treatments (PBS, NIR, GO NPs, PCGA NDs, GO NPs + NIR, PCGA NDs + NIR). The maturation of BMDCs was then analyzed by flow cytometry using the following fluorescently labeled antibodies: anti-CD11c-FITC, anti-CD80-PE, and anti-CD86-APC.

4.6. T cell-mediated tumor cell killing assay

4T1 cells were seeded in 6-well plates and cultured overnight to allow adherence. The cells were then treated with PCGA NDs under the aforementioned conditions (with or without NIR laser irradiation). Subsequently, the treated 4T1 cells were co-cultured with pre-activated T cells for 24 h at an effector-to-target (E:T) ratio of 1:4. Following co-culture, non-adherent cells and debris were removed by gentle washing with PBS. The remaining adherent cells, representing viable 4T1 cells, were fixed and stained with crystal violet. The stained cells were solubilized, and the absorbance at 570 nm was measured using a spectrophotometer to quantify relative viability [44]. Additionally, the plates were imaged under a bright-field microscope for visual assessment.

4.7. T cell activation and viability assay

Mouse splenocytes were isolated and cultured in RPMI-1640 medium for 48 h. The cells were then stimulated by replacing the medium with fresh RPMI-1640 containing concanavalin A (ConA, 6 μg mL⁻¹) for 72 h. T cell activation was initially confirmed by observing characteristic lymphocyte aggregation under a bright-field microscope. To assess the protective effects of the nanodots against oxidative stress, the activated T cells were treated with GO NPs or PCGA NDs (200 μg mL⁻¹) for 12 h, followed by exposure to 1 mM H₂O₂. Cell viability was subsequently quantified using the CCK-8 assay.

4.8. Fluorescence response of PCGA-Pep to GzmB

The fluorescence-quenched PCGA-Pep nanodots (200 μg mL⁻¹) were incubated with 0.5 U GzmB, in PBS buffer (10 mM, pH 7.4) at 37°C. The recovery of fluorescence was monitored over time using a fluorescence spectrophotometer (λex/λem = 570/585 nm) at 0, 1, 2, 3, and 4 h. Additionally, aliquots of the reaction solution taken before and after the 4-h incubation were imaged using an IVIS spectral imaging system (λex/λem = 560/620 nm) to visually compare the signal enhancement resulting from the enzymatic reaction.

4.9. Sensitivity of PCGA-Pep to GzmB

A stock solution of PCGA-Pep (1.0 mg mL⁻¹) was prepared in PBS buffer (10 mM, pH 7.4). For the assay, 10 μL of this stock solution (final concentration: 10 μg mL⁻¹) was mixed with varying volumes of a GzmB solution (0-30 pM) in a 1.5 mL centrifuge tube. The total reaction volume was brought to 1 mL with PBS. After incubating the mixtures at 37°C for 4 h, the fluorescence emission spectra were recorded using a fluorescence spectrophotometer with an excitation wavelength of 575 nm and a scan range of 570-750 nm.

4.10. Selectivity of PCGA-Pep for GzmB

The selectivity of the PCGA-Pep probe (200 μg mL⁻¹) for GzmB was assessed by incubating it separately with various potential interferents under their respective optimal buffer conditions at 37°C for 4 h. These included glutathione (GSH, 100 μM) and H₂O₂ (100 μM) in PBS (10 mM, pH 7.4); Cathepsin C (0.5 U) in HEPES buffer (100 mM, pH 7.4, 0.3 M NaCl); DNase I (0.5 U), Caspase-3 (0.5 U), and GzmB (30 pM) in PBS (10 mM, pH 7.4); as well as MMP-2 (0.5 U) in Tris buffer (50 mM Tris, 10 mM CaCl₂, 150 mM NaCl, 0.05% Brij-35, pH 7.5). Following incubation, the fluorescence intensity of each mixture was quantified using a fluorescence spectrophotometer (λex = 570 nm, λem = 585 nm), and the reactions were additionally visualized using an IVIS spectral imaging system (λex/λem = 560/620 nm).

4.11. Determination of catalytic kinetics of GzmB

To determine the catalytic kinetics of granzyme B (GzmB, 0.5 U), different concentrations of PCGA-Pep nanodots (10, 25, 50, 100, and 200 μg mL⁻¹) were incubated in PBS buffer (10 mM, pH 7.4) at 37°C for 6 h. After the reaction, the fluorescence intensity of each solution was measured using a fluorescence spectrophotometer (λex = 570 nm, λem = 585 nm). The residual substrate concentration was determined based on a pre-established standard curve relating fluorescence intensity to PCGA-Pep concentration. The initial reaction rate (V, nmol min⁻¹) at each substrate concentration was then calculated. The relationship between substrate concentration ([S], μg mL⁻¹) and initial rate (V) was analyzed by nonlinear regression fitting to the Michaelis–Menten equation [14]:

$$\begin{array}{c}\mathrm{V}=\frac{\text{Vmax}·[\mathrm{S}]}{\text{Km}+[\mathrm{S}]}\end{array}$$

where Vmax represents the maximum reaction rate and Km is the Michaelis constant.

4.12. Evaluation of therapeutic efficacy in vivo

Tumor-bearing BALB/c mice were randomly allocated into six groups (n = 6 per group): (I) PBS, (II) PBS + NIR, (III) GO, (IV) PCGA, (V) GO + NIR, and (VI) PCGA + NIR. All formulations were administered via tail vein injection at a dose of 5 mg kg⁻¹ on days 0, 3, and 9. For groups receiving NIR treatment, the tumor region was irradiated with an 808 nm laser at 6 h post-injection for 10 min (1.0 W cm⁻²). Tumor size and body weight were monitored every two days from day 0 to day 14. Tumor volume was calculated as follows: Volume (mm³) = (length) × (width)²/2. On day 14, all mice were euthanized. Blood samples were collected for complete blood count and serum biochemical analysis. Major organs (heart, liver, spleen, lung, and kidney) and tumors were harvested for histopathological examination. Tumor and major organ tissues were fixed, paraffin-embedded, and sectioned for H&E and TUNEL staining to evaluate histopathology and apoptosis. Tumor single-cell suspensions were analyzed by flow cytometry for CD3⁺CD8⁺ and CD3⁺CD4⁺ T cell infiltration. Immunofluorescence staining of tumor sections was performed and imaged by CLSM.

4.13. In vivo imaging of GzmB activity

The GzmB-activated fluorescent probe (PCGA-Pep) was synthesized by conjugating a TAMRA-labeled peptide substrate to PCGA NDs. For in vivo imaging, tumor-bearing mice were intravenously injected with the probe (10 nM). After allowing for systemic distribution, the tumor region was irradiated with an 808 nm near-infrared laser (1.0 W cm⁻², 10 min). Fluorescence imaging was performed using an IVIS imaging system at selected time points post-injection. To sustain detectable signal levels throughout the 24-day observation period, an additional 5 nmol of PCGA-Pep was administered every 48 h.

4.14. Statistical analysis

Data are presented as the mean ± standard deviation (SD). Differences between two groups were compared using an unpaired two-tailed Student's t-test, and differences among multiple groups were analyzed by one-way ANOVA followed by Tukey's post hoc test. All statistical analyses were performed using GraphPad Prism (version 9.0). A p-value <0.05 was considered statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001).

CRediT authorship contribution statement

Yuxin Jin: Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Writing – original draft, Writing – review & editing. Jing Zhu: Data curation, Formal analysis, Investigation, Methodology, Software, Visualization. Huan He: Formal analysis, Methodology, Software. Yunyun Wu: Methodology, Software. Lin Li: Funding acquisition, Methodology, Software. Sheng Zhao: Methodology. Zhenqiang Wang: Conceptualization, Data curation, Formal analysis, Project administration, Writing – review & editing. Jixi Zhang: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Validation, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.