The photosensitizer DTP was synthesized and provided by Prof. Tianjun Liu (Institute of Biomedical Engineering, Chinese Academy of Medical Sciences and Peking Union Medical College). The purity of DTP was over 98% by High Performance Liquid Chromatography. UV-Vis absorption profiling of DTP dissolved in dimethyl sulfoxide (DMSO) was performed on a UH5700 spectrophotometer (Hitachi, Japan). A 650-nm semiconductor laser (WSLS-650-500m-M-2; Wave Spectrum Laser Group Limited, China) via a columnar fiber was used in the in vivo and in vitro study, ensuring precise and controlled irradiation conditions. The energy density of the illuminated spot was quantified using a light power meter (Coherent Corp., PA, United States).

The murine 4T1 breast tumor cell line was sourced from the National Collection of Authenticated Cell Cultures (Chinese Academy of Sciences, Shanghai). Cells were cultured in RPMI 1640 medium containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (PS), under standard humidified incubation (37°C, 5% CO₂).

4T1 cells plated in confocal dishes underwent 24-h incubation with DTP (200 nM). Subsequently, cultures were irradiated (650 nm laser, 20 mW/cm², 5 min) or maintained as non-irradiated controls. Post-treatment, cells were fixed in 4% paraformaldehyde (PFA) (15 min, RT), permeabilized in 0.1% Triton X-100 (high mobility group box 1 (HMGB1) detection only), and blocked with 1% bovine serum albumin. Primary antibodies include anti-calreticulin (CRT) (1:200, ABclonal, A1066) or anti-HMGB1 (1:200, Wanleibio, WL03023) were applied overnight (4°C), followed by Alexa Fluor 488-conjugated secondary antibody incubation (1.5 h, RT). Cells were counterstained with DAPI (5 min, RT) and specimens imaged by confocal laser scanning microscopy (CLSM).

Bone marrow was isolated from femurs and tibias of 5–7-week-old C57BL/6 mice. Following erythrocyte lysis (Beyotime, C3702), cells were suspended in RPMI 1640 medium supplemented with 10% heat-inactivated FBS, 1% PS, 20 ng/mL recombinant murine granulocyte-macrophage colony stimulating factor (GM-CSF), and 10 ng/mL recombinant murine Interleukin-4 (IL-4). Culture medium was replaced every 48 h. On day 7, loosely adherent BMDCs were harvested by gentle pipetting and centrifugation (450 × g, 5 min).

4T1 cells pretreated with phosphate-buffered saline (PBS) or DTP (12 h) received 650 nm laser irradiation (20 mW/cm², 5 min). Then, BMDCs were co-cultured with DTP-PDT-treated 4T1 cells in Transwell^® inserts (0.4 μm pore, Corning) to enable soluble mediator exchange while preventing direct contact. Lipopolysaccharide (LPS; 1 mg/L) was used as a positive control. Following a 24-h co-culture period, the BMDCs were collected, labeled with anti-CD11c-BV421 (BioLegend, 117343), anti-CD80-PE (BioLegend, 104707), and anti-CD86-APC (BioLegend, 105011), and then subjected to flow cytometric analysis.

Cell supernatants from the transwell experiment were collected for cytokine detection using enzyme-linked immunosorbent assay (ELISA) kits specific for mouse tumor necrosis factor-α (TNF-α) (Cloud-Clone, SEA133Mu) and mouse interferon-γ (IFN-γ) (Cloud-Clone, SEA049Mu), strictly adhering to the protocols provided by the manufacturer.

Female BALB/c mice, aged 6–8 weeks (body weight 18–20 g), were purchased from HFK Bioscience (Beijing, China). Animals were maintained in a specific pathogen-free facility, regulated at constant temperature (22 ± 1°C), humidity (50% ± 10%), and a 12-h light/dark cycle. All animal experiments strictly followed the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committee of the Institute of Radiation Medicine, Chinese Academy of Medical Sciences approved all experimental procedures under Ethics Approval No. IRM/2-IACUC-2409-094.

For bilateral breast tumor establishment, 5 × 10⁵ luciferase-transfected 4T1 (4T1-Luc) cells, suspended in 100 μL PBS, were first subcutaneously injected into the right flank to generate the primary tumor site. Once the primary tumor reached 60 mm³, a secondary challenge was developed by inoculating 1 × 10⁵ cells into the contralateral (left) flank to form a distinct distant tumor. Tumor growth was determined using the formula: Volume = (width² × length) × 0.5.

When the primary tumor volumes reached 80–100 mm³, mice were randomly allocated into four treatment groups (n = 16 per group): Model (100 μL saline, intravenous), α-PD-L1 (10 mg/kg α-PD-L1 (A2115, Selleck, United States), administered intraperitoneal every 3 days), PDT (intravenous 10 mg/kg DTP followed by 650 nm laser irradiation at 100 J/cm²), and PDT + α-PD-L1 (combined DTP-PDT and α-PD-L1). Tumor dimensions and individual body weights were monitored every other day. Tumor progression was assessed weekly using bioluminescence imaging (IVIS Lumina III, Caliper Life Sciences).

To evaluate the immune response, tumors, lnguinal lymph nodes, serum, and spleens were collected using sterile procedures on day 8 post-treatment. Tumor and lymphoid tissues underwent mechanical dissociation to generate single-cell suspensions. Subsequent processing included red blood cell (RBC) lysis, filtration, and extensive washing steps. Cell viability was determined using the Zombie Aqua™ Fixable Viability Kit (Biolegend, 423101). To block nonspecific binding, cells were incubated with TruStain FcX™ (anti-mouse CD16/32) antibody (Biolegend, 101319). Cells were then stained with the following antibodies: anti-CD45 FITC (Biolegend, 103107), anti-CD11c BV421 (Biolegend, 117329), anti-CD80 PE, anti-CD86 APC, anti-CD3ε Percp-cy5.5 (Biolegend, 100327), anti-CD4 PE (Biolegend, 100511), anti-CD8a APC (Biolegend, 100711), and anti-Forkhead box protein P3 (FoxP3) BV421 (Biolegend, 126419), followed by flow cytometric analysis. Serum levels of TNF-α and IFN-γ were measured using specific mouse ELISA kits. Immunofluorescence staining was performed on the tumor sections using anti-CD8 and anti-FoxP3 antibodies.

On day 21 post-treatment, tumors were resected and fixed in 4% PFA, paraffin-embedded, and sectioned at 4 µm thickness. Following deparaffinization in xylene and graded ethanol rehydration, sections were subjected to H&E staining to evaluate the pathological changes. Images were acquired on a Leica DMILLED microscope system (Leica Microsystems). Ten fields per section were randomly selected for microscopic observation and quantitative analysis.

Data are expressed as mean ± standard deviation (SD). Data normality was confirmed via Shapiro-Wilk tests. Homogeneity of variance was validated using Levene’s test. For non-normally distributed data, Kruskal-Wallis with Dunn’s post hoc was employed. Differences between two groups were assessed via Student’s t-test, while differences among three or more groups were assessed via one-way analysis of variance (ANOVA). Statistical significance was defined as *p < 0.05, **p < 0.01, and ***p < 0.001.

The molecular structure of DTP is illustrated in Figure 2A. UV-Vis spectroscopy revealed the spectral profile of DTP, with the main peak at 420 nm and additional peaks at 516, 551, 592, and 647 nm (Figures 2B,C). While DTP exhibits a primary absorption peak at 420 nm, we specifically employed 650 nm laser irradiation during treatment to leverage its deeper tissue penetration capability in the red-light spectrum (Austin et al., 2021), following clinical PDT standards (Allison et al., 2024).

After laser irradiation, 4T1 cells treated with DTP were subjected to immunofluorescence staining targeting HMGB1 and CRT to assess whether DTP-PDT could induce ICD. As shown in Figure 3A, the distribution of HMGB1 was found to be predominantly nuclear in PBS or DTP treated cells. In contrast, HMGB1 was released into the extracellular space in DTP-PDT-treated cells. As demonstrated in Figure 3B, the laser-treated group exhibited intense green fluorescence on the cell surface in comparison to the non-irradiated group, demonstrating considerably elevated levels of CRT exposure. These results suggest that DTP-PDT triggers HMGB1 release and CRT translocation, confirming the induction of ICD.

BMDCs were co-cultured with DTP-PDT-treated 4T1 cells using the experimental design illustrated in Figure 4A. Following co-culture, BMDCs exhibited significant maturation, as evidenced by upregulated expression of co-stimulatory markers CD80 and CD86. As demonstrated in Figures 4B,C, DTP + Laser induced a significantly higher proportion of mature BMDCs (45.53% ± 4.12%) in comparison with PBS (21% ± 2.29%) and DTP (21.57% ± 2.20%) groups (p < 0.001).

Specifically, significant increases in the secretion of IFN-γ (p < 0.001, Figure 4D) and TNF-α (p < 0.05, Figure 4E) were observed in the DTP + Laser group compared to the PBS and DTP groups. These findings indicate that DTP-PDT effectively induces DAMP release, promotes DC maturation, and enhances the expression of pro-inflammatory cytokines, thereby initiating immune responses.

To address the limited efficacy of α-PD-L1 monotherapy in TNBC, we assessed the potential of DTP-PDT-induced ICD to enhance α-PD-L1 therapy and improve systemic antitumor efficacy in the 4T1 breast cancer model (Figure 5A). In the α-PD-L1 monotherapy group, local and distant tumor growth showed no significant difference compared with the model group (Figures 5B, 6A,B). In contrast, DTP-PDT alone resulted in an 81.62% reduction in primary tumor weight (p < 0.001 vs. model and α-PD-L1), but only a 35.35% reduction in distant tumor weight (p > 0.05 vs. model and α-PD-L1), highlighting that PDT has a significant therapeutic effect on irradiated primary tumors but little effect on unirradiated distant tumors. Remarkably, combining DTP-PDT and α-PD-L1 demonstrated a significant abscopal effect, with 83.31% primary tumor regression (p < 0.001 vs. model and α-PD-L1) and 82.15% distant tumor suppression (p < 0.001 vs. model, p < 0.01 vs. α-PD-L1, p < 0.05 vs. PDT) (Figures 6C,D). Importantly, across all treatment groups, the body weight change revealed no significant difference which suggests that the combined strategy has a good tolerability (Figure 6E).

Following treatment, H&E staining was performed on the primary and distant tumor tissue to observe histopathological changes. Figure 6F illustrates that tumor cells in the model and α-PD-L1 groups, as well as distant tumors in the PDT group, exhibited tight cell arrangements with rounded nuclei and well-defined nucleoli. In contrast, tumors in the PDT + α-PD-L1 group and primary tumors in the PDT group, displayed extensive nuclear consolidation, nuclear fragmentation, and visible necrotic areas. These findings provide further evidence of the remarkable inhibitory effect of the combination therapy on primary and distant tumors.

Collectively, our findings suggest that DTP-PDT-induced tumor-specific immune responses may be effective in sensitizing tumors towards PD-L1 blockade, and combining DTP-PDT with PD-L1 blockade could be a promising approach in treating metastatic TNBC.

The synergistic antitumor effects observed in the combination therapy prompted further investigation of its immunological mechanisms. In order to assess the immune response, we analyzed populations of immune cells in the lymph nodes, tumors, and spleens. In our study, DC maturation was evaluated by flow cytometry in tumor-draining lymph nodes. As shown in Figures 7A,B, the DTP-PDT and DTP-PDT+α-PD-L1 groups had a significantly higher percentage (15.8% and 21.8%, respectively) of mature DCs (CD45⁺CD11c⁺CD80⁺CD86⁺) compared to the model (8.29%) and α-PD-L1 (10.40%) groups (p < 0.001), indicating that PDT promotes DC maturation in vivo.

The analysis of tumor-infiltrating T cells was performed in distant tumors. DTP-PDT+α-PD-L1 group revealed a 9.8-fold and 6.9-fold increase in CD8⁺ and CD4⁺ T cells, respectively, compared to the model group, indicating efficient CTLs infiltration (Figures 7C–E). In contrast, neither the DTP-PDT nor the α-PD-L1 groups showed significant differences in T cell subpopulations (p > 0.05) relative to the model group, indicating that monotherapy was insufficient to elicit an immune response. Moreover, the proportion of Tregs (CD3⁺CD4⁺FoxP3⁺) in the combined treatment group significantly decreased compared to that in the model (p < 0.001) and PDT (p < 0.05) groups, confirming a reduction in tumor-associated immunosuppression (Figures 7F,G). And then the ratio of CD8⁺ T cells to Tregs were compared and a significant increase could be observed in Figure 7H (p < 0.01 vs. model, p < 0.05 vs. α-PD-L1 and PDT). Subsequently, immunofluorescence staining was performed on distant tumor sections to characterize the tumor immune microenvironment, Figure 7I revealed that more red fluorescence and less green fluorescence were observed in the combined therapy group, indicating the increased CD8⁺ T cell infiltration and decreased FoxP3⁺ expression. Quantitative analysis demonstrated a significant increase of CD8⁺/FoxP3⁺ ratio in the PDT + α-PD-L1 group (9.63 ± 2.91) relative to other groups (p < 0.001, Figure 7J).

Systemic immune activation was evaluated by splenic T cell redistribution. Figures 8A–C revealed DTP-PDT monotherapy elevated CD8⁺ and CD4⁺ T-cell proportions by 2.2-fold and 2.1-fold versus Model (p < 0.05). Notably, the combination therapy group showed a statistically higher CD4⁺ T cell proportion relative to either PDT or α-PD-L1 monotherapy. CD8⁺T cells proportion also increased considerably in the DTP-PDT group compared to both model and α-PD-L1 groups (p < 0.05), indicating that PDT enhances the differentiation of naive T cells into CD8⁺ T cells. The proportion of splenic Treg was significantly suppressed in the combined group (p < 0.001 vs. model, p < 0.01 vs. α-PD-L1 and PDT) (Figures 8D,E). Furthermore, the ratio of CD8⁺ T cells to Tregs was significant increased compared to other groups (Figure 7F, p < 0.05).

PDT-mediated tumor cell death induces local inflammation, accompanied by cytokine secretion, particularly TNF-α and IFN-γ (Evans et al., 1990; Zhang et al., 2022). TNF-α increased remarkably in the groups of PDT and combined therapy in comparison with the model and α-PD-L1 groups (p < 0.001, Figure 8G). TNF-α levels were significantly elevated in the combination treatment group compared to the PDT group (p < 0.05). IFN-γ was also upregulated in the combination group when compared to the other groups (p < 0.001, Figure 8H).

Collectively, these results suggested that combining DTP-PDT and α-PD-L1 therapy reshapes the immunosuppressive TME by enhancing T cell infiltration, promoting antitumor cytokine production and limiting immunosuppressive cells. This provides a strong biological foundation for the observed abscopal effects in the mTNBC model.