Virus-based IFNγ gene delivery and photodynamic therapy cooperate to remodel the tumor microenvironment and suppress breast cancer.

Introduction
Virus vector-based strategies are emerging as a promising area of breast cancer immunotherapy research focused on enhancing combined treatment outcomes.1 These approaches aim to reprogram the tumor microenvironment (TME) to stimulate a systemic anti-cancer immune response.2 Among immune cells within the TME, tumor-associated macrophages (TAMs) play a central role in promoting tumor progression and immune evasion.3 TAMs can polarize into either an anti-tumor M1 phenotype or pro-tumor M2 phenotype; however, breast cancer cells often shift this balance toward the M2 type, which supports tumor growth and suppresses immune responses.4
Cytokine gene delivery provides a potential strategy to modulate TAM polarization and restore anti-tumor immunity.5 Viral vectors that deliver cytokines such as interferon-γ (IFNγ) are able to reprogram TAMs and the TME in general.6,7 Alphaviruses, enveloped (+) ssRNA viruses belonging to the Togaviridae family, are promising vectors for cancer immunotherapy due to low preexisting immunity and efficient transgene expression in cancer cells.2,8 The Semliki Forest virus vector SFV1 is replication-deficient, which improves its safety profile by preventing unintended viral replication and dissemination. The replication-deficient particles are generated by co-electroporating in vitro transcribed RNA from an expression vector and helper constructs (single or split) that carry virus structural genes.8,9 Previous studies have tested the ability of the SFV vector to deliver IFNγ to murine mammary tumors.10 Although the SFV/IFNγ therapy significantly reduced mouse breast tumor burden and enhanced immune cell infiltration into the tumor, including CD4+ and CD8+ T cells, a complete tumor remission was not achieved.
Photodynamic therapy (PDT), which involves light-activated photosensitizers, such as chlorin e6 (Ce6), has demonstrated therapeutic efficacy in solid tumors, including breast cancer, through the induction of immunogenic cell death (ICD).11,12 Ce6 is activated by near-infrared (NIR) light, enabling localized tumor destruction while limiting off-target effects.13 Due to the limited tissue penetration depth of NIR light (5–10 mm, depending on tissue type and vascularization), this therapeutic approach is particularly well suited for treating surface-accessible malignancies such as skin cancers, certain breast cancers, and for use in surgeon-guided interventions.14,15 Recent findings suggest that Ce6-PDT not only kills tumor cells but may also modulate macrophage function toward anti-tumor phenotypes.16,17,18 Furthermore, combining oncolytic viruses with PDT has proven effective in treating primary and metastatic tumors.19,20,21
Despite these promising effects, only a few studies have explored the combination of PDT with cytokine therapy.22,23 To our knowledge, no previous research has evaluated PDT in combination with virus-delivered immunomodulatory cytokines. This novel approach could offer enhanced therapeutic potential by integrating local tumor cell destruction with systemic immune activation. In this study, we aimed to address this gap by investigating the therapeutic potential of combining Ce6-PDT with SFV-based IFNγ delivery in two-dimensional (2D) and three-dimensional (3D) in vitro 4T1 mouse breast cancer models. We hypothesized that the combination would exhibit a synergistic anti-tumor effect, which we further evaluated in vivo in an immune-competent 4T1 model.

Results

Evaluation of Ce6/LED treatment efficacy in 2D and 3D 4T1 cell cultures
The cytotoxic effect of Ce6 photodynamic treatment on 4T1 murine mammary cancer cells was evaluated under 2D conditions. The treatment involved adding Ce6 (0–3 μg/mL) to the cells, followed by irradiation with a 650 nm light-emitting diode (LED) at energy densities of 1–5 J/cm². The Ce6/LED treatment exhibited an IC50 of 2.3 μg/mL at 3 J/cm2 and 1.4 μg/mL at 5 J/cm2 (Figure 1A). At a lower irradiation intensity of 1 J/cm2, photodynamic treatment with 3 μg/mL Ce6 resulted in 76.5% cell viability, indicating insufficient cytotoxicity to reach IC50. Ce6 concentrations of 3, 5, and 10 μg/mL, followed by irradiation at 5 J/cm2, increased levels of reactive oxygen species (ROS); however, a statistically significant elevation was observed only at 10 μg/mL (p = 0.007; Figure 1B). Thus, the Ce6/LED treatment induced cell death in a dose-dependent manner, accompanied by increased production of ROS. For subsequent experiments, 3 μg/mL Ce6 and 3 J/cm2 irradiation were used, as this combination resulted in moderate cytotoxicity, allowing for the assessment of sublethal effects and cellular responses in surviving cells.
Cell uptake of Ce6 was evaluated by flow cytometry, with fluorescence measured in the APC channel corresponding to Ce6 emission spectrum. Upon treatment with Ce6 (3 μg/mL), cancer cells showed high uptake of the drug, both with and without irradiation with NIR (650 nm, 3 J/cm2) light (Figure 1C). Specifically, (31.5 ± 0.6)% of cells were positive for Ce6 uptake without NIR light irradiation, and this increased significantly to (98.7 ± 0.1)% after the irradiation (p < 0.0001). The increase in Ce6 uptake following irradiation can be associated with cell membrane disruption. Following Ce6/LED treatment, cells were analyzed for ICD by staining for surface calreticulin (CRT) and phosphatidylserine (PS) exposure via Annexin V. Annexin V staining revealed PS exposure in (79.0 ± 0.8)% of cells, indicating apoptosis, whereas CRT surface expression, a hallmark of ICD, was observed in (31 ± 2)% of cells.
Given that conventional 2D cell cultures fail to replicate the in vivo TME, including cell-cell interactions, spatial architecture, and nutrient gradients, a 3D spheroid model was employed to evaluate the effects of Ce6. 4T1 cancer cells (6×103) were cultured in an Ultra-Low Attachment plate to form 3D spheroids. Compared with monolayer cultures, the 3D cultures displayed reduced Ce6 penetration into the cells at a concentration of 3 μg/mL (Figure 1D). No increase in the fraction of Ce6-positive cells was detected after Ce6/LED treatment (p > 0.05), measuring (8.2 ± 1.2)% without LED irradiation and (6.5 ± 0.4)% after Ce6/LED treatment. Compared with 2D models, Ce6/LED treatment resulted in lower levels of Annexin V and CRT exposure ([21.9 ± 1.3]% and [4.77 ± 0.17]%, respectively). To achieve effects similar to those observed in 2D models, it was found that Ce6 should be added at higher doses. When 5 μg/mL Ce6 was used (Figure 1D), a higher percentage of cells exhibited Ce6 penetration compared with 3 μg/mL (p < 0.0001), with (33 ± 5)% of cells being positive without LED irradiation and (42.8 ± 1.5)% after Ce6/LED treatment. Furthermore, the increase in Ce6 concentration resulted in an increased percentage of cells exhibiting high uptake of Ce6 (Ce6high; p < 0.0001). Without LED irradiation, (0.24 ± 0.04)% of cells exhibited Ce6high, while Ce6/LED treatment increased this percentage to (0.63 ± 0.02)%. Moreover, treatment with 5 μg/mL Ce6/LED resulted in increased exposure of PS (73 ± 2)% and CRT (29.2 ± 1.2)% markers. These results demonstrate that Ce6 uptake, as well as CRT and PS exposure, is reduced in 3D cultures compared with 2D, indicating that higher Ce6 concentrations are necessary to achieve comparable cellular responses.
To investigate the impact of Ce6/LED treatment on the size of cancer cell spheroids, 4T1 cells expressing the green fluorescent protein (GFP) gene were employed. The spheroid size was evaluated through GFP signal intensity measurement using fluorometry. The spheroids (6×103 cells) were treated with Ce6 at concentrations ranging from 0 to 5 μg/mL and exposed to NIR light (0–5 J/cm2) at 8 h and 48 h after addition of Ce6. Results showed that spheroids did not exhibit any size reduction (Figure 1E, left). However, significant size reduction was observed when smaller spheroids (formed from 3×103 cells) were subjected to daily LED irradiation for seven consecutive days (3 μg/mL: 49% reduction, p < 0.0001; 5 μg/mL: 29% reduction, p = 0.0006). Interestingly, the largest effect on spheroid size was achieved with 3 μg/mL Ce6, suggesting that a moderate dose may be more effective in this context. These findings indicate that 4T1 cancer cells undergo ICD following Ce6/LED treatment, which potentially may enhance cancer cell recognition and clearance by macrophages.

Ce6/LED-induced cancer cell phagocytosis by macrophages
We investigated whether Ce6/LED treatment of cancer cells enhances macrophage-mediated phagocytosis. Macrophages were polarized into M0, M1, and M2 phenotypes and stained with CMAC fluorescent dye. 4T1 cancer cells were treated with Ce6/LED (5 μg/mL, 3 J/cm2), then washed to remove Ce6, stained with CMFDA fluorescent dye, and subsequently co-cultured with the macrophages for phagocytosis assessment. The percentage of macrophages that had phagocytosed cancer cells was calculated as a percentage of CMAC+CMFDA+ cells in the CMAC+ population. Co-culturing treated cancer cells with M0, M1, and M2 macrophages led to efficient phagocytosis of 4T1 cells by all phenotypes both after 2 and 24 h of co-cultivation (Figures S1, and 2A). After 24 h of co-cultivation, the enhancement in phagocytosis was statistically significant in M0 (p = 0.001) and M2 (p = 0.009) macrophages, but not in M1 (p > 0.05), compared with co-cultures with untreated 4T1 cells (Figure 2B).

Ce6/LED treatment in combination with SFV/IFNγ
PDT combination with alphavirus therapy was expected to have a synergistic anti-tumor effect as both therapies induce ICD. First, we tested whether SFV infection induces ROS production. Viruses with and without an immunomodulatory gene (SFV/IFNγ and SFV/Luc) were used to compare ROS levels at different multiplicities of infection (MOI 2.5, 5, 10) of 4T1 cells. ROS production was tested 6 and 20 h after the infection. No ROS production was detected 6 h after infection with either SFV/IFNγ or SFV/Luc; however, it was observed 20 h post infection with SFV/IFNγ (Figure 3A). Moreover, ROS production was not dose dependent and was observed exclusively in response to SFV/IFNγ infection.
To model a combination of Ce6/LED treatment with alphavirus infection in a 2D model, 4T1 cells were treated with different concentrations of Ce6, irradiated with NIR LED (5 J/cm2) twice (after 8 h and 24 h), and then infected with a model virus without an immunomodulatory gene SFV/RFP (MOI = 5) for fluorescence microscopy. The next day, a viability assay was performed. Even though SFV/RFP infection reduced cell viability by 20%, the combined effect was challenging to evaluate due to the high efficacy of Ce6/LED treatment at 3 and 5 μg/mL (Figure 3B).
Therefore, to find the most effective combination strategy, three different treatment combinations were tested on a 3D model of 4T1 cancer cells (spheroids plated at 3×104 cells per well 24 h before treatment). Specifically, Ce6/LED (5 μg/mL, 5 J/cm2) treatment was performed before (strategy I), after (strategy II) or simultaneously with SFV/IFNγ (5×105 infectious units or iu) treatment (strategy III). Combined treatment had a synergistic effect in all strategies (Figure 3C); however, the spheroid viability reduction was more significant in strategies II and III (86% and 92%), where Ce6-PDT was performed after SFV infection. Therefore, it was concluded that Ce6-PDT treatment reduces the effectiveness of SFV infection.
To quantify infection effectiveness, the spheroids were treated with Ce6/LED and then infected with viruses encoding the reporter gene (either SFV/GFP or SFV/Luc) as in strategy I. GFP signal was detected by confocal microscopy while luciferase activity was determined by luciferase assay. Consistent reduction in transgene expression was observed in both SFV/GFP and SFV/Luc systems following Ce6/LED treatment (Figures 3D and 3E).

Impact of Ce6/LED treatment on macrophage viability and activation
We have tested how Ce6/LED affects bone marrow-derived macrophages (BMDMs) in a dose-dependent manner. Ce6 was added to M0 macrophages at several concentrations (0–7 μg/mL) and irradiated with 3 J/cm2 NIR light. Cell viability gradually decreased upon increasing the concentration of Ce6 (Figure 4). Macrophages incubated with Ce6 but not irradiated remained viable despite the increase in Ce6 concentration. The measured Ce6 IC50 value is 2.8 μg/mL. Ce6/LED treatment induced a dose-dependent increase in ROS production, with a significant rise observed at concentrations of 5 μg/mL (p = 0.0098) and 10 μg/mL (p < 0.0001) compared with the untreated control. However, despite the elevated ROS levels, there was no evidence of phenotypic reprogramming toward a pro-inflammatory M1-like state. Specifically, nitric oxide (NO) production, a key functional marker of M1 macrophages, remained unchanged at a background level across all treatment conditions.
We have also investigated whether Ce6 treatment could potentiate non-polarized (M0) macrophages to achieve a more pro-inflammatory phenotype after programming into M1 or M2 states. Due to the strong cytotoxic effect of Ce6/LED treatment, it is technically challenging to reliably analyze surface markers on viable macrophages under conventional 2D (monolayer) conditions, as most cells underwent cell death after photodynamic exposure. To overcome this, a 3D cultivation model was employed, as described previously.24 Even though the uptake of Ce6 reached approximately 30%, no phenotypic shifts toward an M1 profile were detected among M0, M1, or M2 macrophages based on CD206, MHC II, and NO production following irradiation (see Figure S2). Therefore, no evidence of pro-inflammatory phenotype induction was observed in the macrophages following this treatment. These findings indicate that, while Ce6/LED treatment activates macrophages at the oxidative level, it is insufficient on its own to induce full M1 polarization.

Evaluation of the therapeutic efficacy of combined treatment in vivo
First, the experiment was conducted to determine the appropriate time point at which the concentration of Ce6 in the 4T1 breast tumors was highest. Ce6 was administered intravenously at a dose of 2.5 mg/kg, and its tissue accumulation was measured at three time points: 1.5 h, 4 h, and 18 h post injection. A PBS-injected group served as a negative control. In this experiment, mice did not receive NIR light treatment. At respective time points, the animals were humanely euthanized and the tumor, liver, kidney, lung, and spleen were analyzed using in vivo imaging system (IVIS), while serum and lysed tumor samples were examined by fluorometry. The highest Ce6 signal was observed 1.5 h after injection in all analyzed organs, including tumor and serum, as confirmed by IVIS imaging and fluorometric analysis (Figures 5A and S3B). Tumors accumulated a significant amount of Ce6 at 1.5 h post injection, with signal intensities reaching 1.8 × 109 p/s. This decreased to 1.2 × 109 p/s at 4 h, and by 18 h, the signal had returned to control levels. Remarkably, the highest total flux signal was observed in the liver, reaching up to 2.2 × 1010 p/s at the 1.5 h time point.
The experimental design to evaluate the effectiveness of combined therapy in vivo is presented in Figure 5B. The experiment included six mouse groups: (1) control (PBS, untreated tumors), (2) PDT (Ce6 + NIR), (3) SFV/Luc (virus without immunomodulatory gene), (4) SFV/IFNγ (therapeutic virus), (5) PDT + SFV/Luc, and (6) PDT + SFV/IFNγ (Figure S3C). Briefly, mice were first treated with PDT twice (days 4 and 6). Ce6 was injected intravenously at a concentration of 6.25 mg/kg, and after 1.5 h, tumors were irradiated with 650 nm NIR light. As we expected PDT to induce ROS-mediated inflammation and kill TAMs but not program them, SFV/IFNγ (1.65×107 iu per mouse) was administered intratumorally after PDT to program newly attracted immune cells. The virus was injected 48h after PDT (day 8) to reduce the PDT effect on infection effectiveness. Moreover, the next PDT treatment was performed 72h after infection (day 11) to prevent inhibition of virus-induced effects. The therapy consisted of two PDT-PDT-SFV cycles. On day 19, the animals were humanely euthanized and the tumor weight and intratumoral immune infiltrate were analyzed.
PDT alone showed a 40.8% reduction in tumor burden while SFV/IFNγ alone showed a 40.5% reduction; however, this did not reach statistical significance (p = 0.093 and 0.090, respectively). The combination of PDT with SFV/IFNγ reduced tumor weight by 87.5% (p = 0.0001), demonstrating a synergistic anti-tumor response (Figure 5B). Nevertheless, complete tumor remission occurred in only one animal. Interestingly, SFV/Luc did not show a significant effect on tumor weight (5%, p > 0.05, one tumor was excluded from the analysis because the corresponding mouse exhibited abnormal body weight), and its combination with PDT resulted in a smaller reduction compared with PDT alone (23.5%, p > 0.05).
Tumor infiltrate was further analyzed using spectral flow cytometry (16 markers listed in Table S1). Five major immune cell populations were analyzed: dendritic cells (DCs), natural killer cells (NKs), T cells, macrophages, and other myeloid cell subsets (Figures S4, S5, S6, and S7 for gating strategy). DCs (F4/80–CD11c+MHC II+) were significantly increased in the PDT + SFV/IFNγ group (p = 0.016, Figure 5C). Moreover, an increasing trend in the proportion of mature DCs (F4/80–CD11c+MHC IIhigh) was observed following treatment with PDT, SFV/Luc, and SFV/IFNγ individually; however, this increase reached statistical significance only in the combined treatment groups, PDT + SFV/Luc (p = 0.011) and PDT + SFV/IFNγ (p = 0.0012), with the highest number of mature DCs in the PDT + SFV/IFNγ group. A similar trend was observed for NKs (NKp46+): although SFV/Luc and SFV/IFNγ showed a tendency to increase NK cell numbers, statistically significant increases were detected only in the PDT (p = 0.003) and combined treatment groups (PDT + SFV/Luc: p < 0.0001; PDT + SFV/IFNγ: p = 0.004), with the highest NK cell counts found in the combined therapies. The increase in mature DCs and NKs suggests that Ce6-PDT mediates its anti-tumor effects through activation of innate immunity, which is further enhanced when combined with alphavirus therapy, indicating a synergistic immune response.
All treated groups exhibited a higher proportion of T cells (CD3+), with the greatest increase observed in the PDT + SFV/Luc group (p < 0.0001). PDT-treated groups showed elevated levels of T helpers (CD4+) and Tregs (CD4+CD25+Foxp3+), except for the PDT + SFV/IFNγ group. Interestingly, we did not observe an increase in CTLs (CD3+CD8+ cytotoxic T cells) in any of the groups; instead, there was a tendency for CTLs to decrease in the PDT groups (p > 0.05). Therefore, even though T cells are present in the PDT + SFV/IFNγ group, they are not T helpers, CTLs, or Tregs, suggesting the presence of unconventional T cell subsets, such as double-negative (CD4−CD8-) T cells. Although duPre et al. previously described that most of the CD3+ cells in the 4T1 tumor had a CD4−CD8− phenotype,25 this finding requires additional validation, considering the fact that double-negative lymphocytes were shown to suppress tumor cell proliferation in an MHC-independent manner in other tumor models.26 Furthermore, analysis of Ly6C-positive CTL cells, which are known to play a key role in mediating anti-tumor immune responses,27 revealed an increased number of these CD3+CD8+Ly6C+ cells in the SFV/IFNγ treatment group (p = 0.025). Nevertheless, the identification of such rare cell populations warrants further investigation.
The number of macrophages (Ly6G–F4/80+) was slightly increased in PDT-treated groups (PDT, PDT + SFV/Luc, and PDT + SFV/IFNγ; p = 0.005, 0.056, and 0.0098; Figures 5D and 6A). To efficiently identify and visualize broad macrophage cell populations, the Ly6G–F4/80+ cells were further analyzed by FlowSOM clustering (Figure 6B). FlowSOM clustering of tumor-infiltrating macrophages identified 15 metaclusters based on 6 marker expression (Ly6C, CD38, Arg1, iNOS, MHC II, and CD206). UMAP projection of arcsinh-transformed expression data revealed considerable phenotypic heterogeneity among tumor-infiltrating macrophages (Figures 6B–6D, S6B, and S6C).
Relative frequency of each metacluster across treatment groups was then analyzed (Figure 6C). Ly6C– mature, anti-inflammatory tumor-promoting macrophages (Metaclusters 3, 12, and 15) were either unchanged or decreased after treatment (Figures 7A and S6E). Metaclusters 3 (Ly6C–iNOSmedMHC IImed) and 15 (Ly6C–iNOS–ArghighMHC II+CD206highCD38high) were unchanged in all treated groups (Figures S6E and 6D). Metacluster 12 (Ly6C–iNOSmedArghighMHC II+CD206highCD38high) tended to decrease following alphavirus treatment (SFV/Luc and SFV/IFNγ, p = 0.068 and 0.020) and combination with PDT (PDT + SFV/Luc: p = 0.061; PDT + SFV/IFNγ: p = 0.020). Interestingly, pro-tumorigenic CD206 marker expression was enriched in M2-like Metaclusters 12 and 15 in the PDT-treated groups (PDT, PDT + SFV/Luc); however, this enrichment was diminished in the PDT + SFV/IFNγ group (Figure S6F).
Metaclusters 4 (Ly6C+iNOS–MHC II–CD206lowCD38high), 5 (Ly6ChighArglow), and 13 (Ly6C+iNOS–CD38high) likely represent newly infiltrating, immature macrophages characterized by high Ly6C expression and low M1/M2 polarization marker expression (Figures 6D and 7A). Metacluster 4 and 5 cells tended to increase by SFV/Luc (p4 = 0.060, p5 > 0.05), SFV/IFNγ (p4 = 0.018, p5 = 0.008), PDT (p4 = 0.018, p5 = 0.008), and PDT + SFV/Luc (p4 = 0.018, p5 = 0.008) but not PDT + SFV/IFNγ. Metacluster 13 cells (Ly6C+iNOS–CD38+) were specifically increased by SFV/IFNγ (p = 0.024), but this increase was reversed in the PDT + SFV/IFNγ group.
Both alphavirus and PDT therapies increase the proportion of pro-inflammatory iNOS-positive cells (Metaclusters 1 and 2, Figures 7B and S6G). iNOShigh cells (Metacluster 2) were enriched in all treated groups, while iNOShighMHC II– cells (Metacluster 1) were increased by PDT and PDT + SFV/Luc (p = 0.050 and 0.025).
Lastly, PDT + SFV/IFNγ shifts the density gradient toward MHC IIhighCD206–Arg– cells (Figure 7C). Metaclusters 6–10 represent an MHC IIhighArg–/low M1-like macrophage subset, distinguished by elevated MHC II expression and varying levels of iNOS, Ly6C, CD38, and CD206 (Figures 6D and 7C). Metaclusters 6 (Arg–MHC IIhigh) and 7 (Ly6CmediNOSmedArg–MHC IIhighCD206+CD38+) remained unchanged across all treatment groups (Figure S6H). In contrast, Metacluster 8 cells (MHC IIhigh) were decreased by PDT and PDT + SFV/Luc treatments (p = 0.022 and 0.011) but not in the PDT + SFV/IFNγ group (Figure 7C). Interestingly, Metacluster 9 (Ly6C+iNOSmedArg–MHC IIhighCD206+CD38–) was absent in the untreated group and enriched in treated groups, with the greatest tendency after combined treatment. Moreover, Metacluster 10 (MHC IIhighCD38–) was significantly increased in all groups (Figures 7C and S6D). Collectively, these data suggest that the PDT + SFV/IFNγ treatment leads to the strongest enrichment of M1-like macrophages.
M2-like cells accumulate in Metaclusters 11 and 14 (Figure 7C). Metacluster 11 (Ly6C+iNOSmedArghighMHC II+CD206+CD38+) was decreased in all treated samples. Metacluster 14 cells (iNOS–ArghighMHC II+CD206+CD38+) tended to increase by SFV/IFNγ (p = 0.13), which was reversed in PDT + SFV/IFNγ. Interestingly, CD206 expression was enriched in Metaclusters 11 and 14 in all treated groups; however, this enrichment was diminished in the PDT + SFV/IFNγ group (Figure S6I).
Thus, even though we observe that PDT + SFV/IFNγ reduces the number of mature MHC II-positive macrophages, it also increases the number of M1-like and iNOShigh macrophages and decreases CD206+ macrophages.
To explore the heterogeneity of tumor-infiltrating non-macrophage myeloid cells we focused our analysis on the CD11b+ subset, excluding T, NK, macrophages, and dendritic lineages (F4/80−MHC II−CD3−NK−CD11b+). The number of non-macrophage myeloid cells was slightly increased in the SFV/IFNγ-treated group (p = 0.009; Figure S8A). FlowSOM clustering of tumor-infiltrating macrophages identified 8 metaclusters based on 7 marker expression (Ly6C, Ly6G, CD11b, CD11c, CD38, Arg1, iNOS). Dimensionality reduction using UMAP on arcsinh-transformed marker expression revealed a continuous spectrum of phenotypes, including neutrophil-like and monocyte-like populations (Figures S8B, S8D, S7B, and S7D).
Relative frequency of each metacluster across treatment groups was then analyzed (Figure S8C). The majority of non-macrophage myeloid cells were Ly6G+ (Metacluster 1), indicating a predominance of polymorphonuclear cells (PMNs), such as neutrophils, PMN-like monocytes, or PMN-MDSCs. The amount of these cells was significantly increased by SFV/IFNγ (p = 0.011), but this effect was reversed in PDT + SFV/IFNγ (Figure S9A).
Metaclusters 4 (low cell count) and 5 did not show significant expression of Ly6C or Ly6G with CD38med and Arg1med. Both metaclusters did not change significantly. However, these metaclusters were absent in the SFV/IFNγ group (p = 0.14 and 0.047, Figure S9B).
Metaclusters 2, 3, 6, 7, and 8 represent Ly6Chigh monocyte populations, including both pro-inflammatory monocytes and monocytic myeloid-derived suppressor cells (M-MDSCs). These clusters likely represent newly recruited or activated monocyte populations based on high Ly6C expression and variable levels of inflammatory and activation markers (Figure S9C). Metacluster 2 (Ly6C+iNOS+) tended to increase by PDT (p = 0.26). On the other hand, Metacluster 3 (Ly6C+Ly6Glow) was decreased by SFV/IFNγ (p = 0.025) but increased by PDT + SFV/IFNγ (p = 0.043). Metacluster 7 (Ly6ChighLy6G−) did not change significantly (Figure S7C). Metaclusters 6 and 8 likely represent M-MDSCs (Ly6ChighArg+CD38+). Metacluster 6 (Ly6Chigh) tended to be induced by PDT (p = 0.15) but decreased in PDT + SFV/IFNγ (p = 0.038). Metacluster 8 (Ly6ChighArg1+CD38high, low cell count) was absent in the SFV/IFNγ group (p = 0.009). The heterogeneity among Ly6Chigh subsets highlights the complexity of monocyte recruitment and functional polarization in response to therapy.

Discussion
Alphaviral vectors offer a promising approach for cancer therapy by delivering macrophage reprogramming factors to the TME. Our previous studies have examined the efficacy of the SFV vector in delivering IFNγ to cancer cells.6,10,28 The application of SFV/IFNγ with TLR2/1 ligand PAM3SCK4 effectively induced the polarization of macrophages toward the M1 phenotype, as demonstrated by the production of IL-12 and NO, along with upregulation of MHC II and iNOS. While intratumoral administration of SFV/IFNγ in the 4T1 breast cancer model resulted in a reduction of tumor volume and increased anti-tumor immune response,10 complete tumor remission was not achieved, suggesting the need for further optimization or combination with other strategies.
SFV replication induces ICD in infected cells,29 a process that enhances the phagocytosis of dying cells by macrophages.30,31 This mechanism is particularly relevant in the context of cancer, as it facilitates the elimination of tumor cells and the activation of downstream antigen presentation pathways, thereby contributing to systemic anti-tumor immunity targeting metastases. Furthermore, IFNγ by itself promotes the ROS production (Figure 3A), consistent with a previous study.32 ROS are known to enhance inflammatory immune responses within the TME.33 Nevertheless, the SFV vector used in our study is replication-deficient, meaning it can only perform one round of infection, because progeny virions are not produced. Although this approach is safer compared with oncolytic vectors, it results in insufficient ICD induction to achieve complete tumor eradication. To address this limitation, we combined SFV-mediated delivery of IFNγ with a complementary treatment that strongly induces ICD, such as PDT.34
First, we investigated the effects of Ce6-PDT on 3D cultures of 4T1 breast cancer cells. The observed anti-proliferation effect of PDT in the 3D model was lower than in the 2D model, which is attributed to the insufficient Ce6 penetration and light absorption. These results highlight the need for higher doses of the photosensitizer Ce6 to achieve optimal effects in 3D cultures. Furthermore, in densely packed 3D structures, NIR light attenuation poses a significant challenge due to limited NIR light penetration in deeper regions of the spheroid.35 Prior studies have noted the importance of drug delivery approaches to enhance the effectiveness of PDT in 3D systems. For instance, Kumari et al. utilized Ce6-loaded micelles to improve penetration into A549 spheroids, resulting in an increased number of apoptotic cells after a single PDT treatment. However, the reduction in spheroid size was only modest, around 15%.36 Another promising formulation for the photodynamic treatment is Ce6-loaded ethosomes, which demonstrated a significant reduction in the cell viability of squamous cell carcinoma spheroids in vitro.37 Without the use of delivery systems, Ce6 failed to penetrate murine melanoma B16F10 cell spheroids.38 Therefore, delivery systems could enhance PDT therapy efficiency. However, for the investigation of the combination strategy in vitro, our aim is not to achieve maximal therapeutic efficacy, as this could mask insights into how each individual component contributes to spheroid growth inhibition.
Research on combining PDT with cytokine-based viral gene delivery remains very limited. While PDT has been paired with several oncolytic viruses, including vaccinia virus, reovirus, herpes simplex virus, and vesicular stomatitis virus, to enhance anti-tumor activity across different cancer models,19,20,21,39,40,41 only a few studies have evaluated PDT together with cytokines, such as IFNα (protein formulation) in cervical low-grade squamous intraepithelial lesions.22 To our knowledge, no studies have investigated PDT in combination with virus-delivered IFNγ. This represents an important conceptual advance, as IFNγ has potent immunomodulatory properties that differ from those of IFNα and oncolytic viruses. Based on this rationale, integrating PDT with alphavirus-mediated IFNγ delivery was anticipated to yield synergistic anti-tumor effects.
Nevertheless, due to cytotoxicity, PDT may potentially reduce SFV infection. In the spheroid model, we demonstrated that the spheroid inhibition was more significant when the SFV was applied prior to Ce6-PDT. This is consistent with the fact that PDT induces ICD, which can interfere with viral entry and replication.42,43 Moreover, in spheroids, where the cell surface accessible for infection is more limited than in monolayers, damage to the outer cell layer by PDT likely prevents the virus from reaching inner cells, further reducing infection efficiency. The outer cell layers have previously been shown to play a protective role by shielding the core of the spheroid.44
Previous studies have demonstrated the ability of PDT to program macrophages into anti-cancer phenotypes.16,17,18 In our study, similar to the findings shown by Yu et al.,16 Ce6/LED treatment significantly affected the viability of macrophages (M0) and induced the generation of ROS (Figure 4). The observed severe effect on macrophage viability can be viewed as a positive outcome, as reducing the population of TAMs is one of the strategies in cancer treatment.45,46 However, no evidence of macrophage reprogramming was detected, as indicated by unchanged NO levels (Figure 4)—one of the key features of classically differentiated M1 macrophages.47 Remarkably, NO was not detected in previous studies, whereas the elevated level of inducible NO synthase (iNOS) was demonstrated by western blot analysis of RAW 264.7 cells, a mouse monocyte-macrophage tumor cell line.17 It is possible that the differences observed in our study, such as the lack of macrophage reprogramming and altered responses to PDT, could be attributed to distinct macrophage origins. Previous in vitro studies have commonly employed RAW264.7 cells, which differ from primary BMDMs.48,49
We expected combined therapies to (i) kill cancer cells, (ii) induce ICD, and (iii) program macrophages to M1-like phenotype. As was noted above, PDT alone did not program macrophages toward the M1 phenotype. The combined therapy in in vitro 3D models showed synergistic effects, significantly reducing the size of cancer cell spheroids (Figure 3C). However, the combination of PDT with alphaviral therapy was challenging due to the negative effect of PDT on macrophage viability and alphaviral infection. We suppose that using PDT after an alphaviral infection would be senseless, as it would kill IFNγ-reprogrammed M1-like macrophages. On the other hand, PDT prior to infection would decrease the efficiency of alphaviral therapy. Therefore, for in vivo studies, PDT was performed first to kill tumor cells and tumor-supporting cells. SFV/IFNγ was administered 48 h after PDT to (i) reduce the PDT effect on infection effectiveness and (ii) program newly attracted immune cells. This experimental design resulted in remarkable tumor weight reduction in the PDT + SFV/IFNγ group (Figure 5B). A similar strategy was applied for the treatment of subcutaneous murine NXS2 neuroblastoma and human FaDu head and neck squamous cell carcinoma xenografts in nude mice by a combination of PDT with oncolytic (replicating) vaccinia virus, which was administered 12h after PDT, since delivery of the virus before PDT treatment led to virus inactivation.20 To achieve the highest anti-tumor efficacy, in contrast to oncolytic virotherapy, we introduced a two-day interval between PDT and SFV treatment, aiming to establish more favorable conditions for immune cell programming. This sequence was chosen to prevent PDT-induced cytotoxicity from compromising alphavirus entry and transgene expression, as demonstrated in our in vitro experiments. Allowing a 48h recovery period before SFV/IFNγ administration was expected to permit virus infection, whereas the 72h interval before the next PDT ensured that IFNγ-mediated immune programming could occur before additional tumor cell disruption. However, further optimization of the PDT/virus treatment strategy regarding dosing, timing, and treatment sequence is still required.
The mechanism of immune reactions caused by PDT is less explored and remains largely speculative.42 PDT results in the release of intracellular components known as damage-associated molecular patterns, which are recognized by pattern recognition receptors expressed on immune cells, leading to activation of innate immunity within TME. In our study, PDT enhances the maturation of DCs and NK cells, indicating that the Ce6-PDT anti-tumor mechanism clearly involves innate immunity (Figure 5C). These data are indirectly supported by several in vitro findings.50,51 Moreover, previous reports have demonstrated that PDT increases the proportion of mature DCs and F4/80+ cells.52,53,54
On the other hand, T cell-mediated immunity is not always observed after PDT,55 and low-dose PDT has been reported to induce immune escape of tumors (increase Tregs, decrease CTLs).56 Interestingly, in our study, PDT appears to decrease CTL numbers while increasing T helper (CD4+) and Treg cells. Nonetheless, SFV/IFNγ increased Ly6C+ CTLs, which were previously shown to be important in mediating anti-tumor immune responses.27,57 The low number of classical CD8+ CTLs can be attributed to the intrinsic characteristics of the 4T1 model, which is marked by weak, or “cold,” T cell immunity in BALB/c mice.58,59,60,61 This is associated with aggressive tumor growth and limited responsiveness to therapies targeting T cell-mediated responses. Furthermore, our analysis was performed 6 days after the last PDT and 4 days after the final SFV/IFNγ administration (day 19), representing a relatively late phase of immune remodeling. CTL activation may therefore have occurred earlier and was not captured at the chosen time point. Other studies with a similar experimental design have also reported no CTLs induction following Ce6-PDT monotherapy in the 4T1 tumor model.62 In contrast, Gurung et al. demonstrated that Ce6-PDT induced T cell activation and anti-tumor immunity, potentially through the suppression of PD-1 and PD-L1 pathways, in C57BL/6 syngeneic melanoma and pancreatic tumor mouse models,54 which are known for robust T cell-mediated immune responses. Interestingly, we previously demonstrated that SFV/IFNγ induces PD-L1 expression in macrophages.6 Hence, we suppose that the synergistic effect of PDT and SFV/IFNγ may be attributed to PDT-mediated PD-1/PD-L1 inhibition. Nevertheless, the observed low CTL responsiveness and high proportion of CD3+ T cells lacking both CD4 and CD8 markers warrant further investigation to elucidate the nature and functional role of T cell subsets under Ce6-PDT.
Analysis of tumor immune infiltrate revealed that SFV therapy (represented by the SFV/Luc vector) reduced the number of mature macrophages and M2-like macrophages, while increasing the number of iNOS-positive macrophages (Figure 7). The addition of an immunomodulatory gene (SFV/IFNγ) further increased the number of iNOS-positive cells. This is in line with our previous studies demonstrating that SFV/Luc and SFV/IFNγ increased macrophage Nos2 expression, polarizing macrophages toward M1 profile.6 The Ly6C–iNOS+ArghighMHC II+CD206highCD38high population (Metacluster 12, Figure 7A), which likely represents a transitional macrophage subset that exhibits both inflammatory (M1-like) and anti-inflammatory (M2-like) features, decreased following both PDT and SFV treatments, with the most pronounced reduction observed in the PDT + SFV/IFNγ combination group. These transitional or mixed-polarization TAMs are common in complex TMEs and may play roles in immune regulation, angiogenesis, and resistance to therapy.63
Ly6C+iNOS−CD38+ cells most likely represent a subset of monocytic MDSCs with immunosuppressive potential, particularly in the TME. Their suppressive mechanism may be iNOS-independent, relying on metabolic or cytokine pathways such as CD38-mediated NAD+ depletion or arginase activity.64,65 Interestingly, SFV/IFNγ specifically increased the Ly6C+iNOS–CD38+ population (Metacluster 13, Figure 7A), but this was reversed in the PDT + SFV/IFNγ group. Furthermore, although SFV/Luc and SFV/IFNγ-mediated decrease of the CD206 gene Mrc1 was previously observed,6 while decreasing the majority of CD206 cells (Metacluster 11, Figure 7C), Ly6C+iNOSlowArg+MHC II+CD206+CD38+ cells, SFV/IFNγ increased iNOS–CD206+CD38+Arg+MHC II+ M2-like cells (Metacluster 14, Figure 7C). The use of IFNγ in cancer therapy is complex and remains controversial due to its dual roles in both promoting and inhibiting tumor growth depending on the TME status, tumor stage, and the concentration of the IFNγ.66,67,68 It is likely that PDT enhances the pro-inflammatory and tumor-inhibiting effects of SFV-delivered IFNγ, because this population (Metacluster 14) decreased under combined treatment.
Similar to SFV therapy, Ce6-PDT increased M1-like and iNOS-positive cells and decreased M2-like cells. Combined PDT + SFV/IFNγ therapy amplified both PDT and SFV therapy effects by further increasing PDT- and SFV-primed DCs, mature DCs, NKs, and iNOS-positive cells and counteracted potentially negative effects of monotherapies. For example, while PDT alone decreased MHC IIhigh, SFV/IFNγ restored MHC II expression in macrophages under the combined treatment (Metacluster 8, Figure 7C). MHC II gene H2-Ab1 upregulation by SFV was reported previously.6 Furthermore, combined therapy modulated certain effects by reducing PDT-enhanced Th and Treg populations (Figure 5C), as well as SFV/IFNγ-increased newly infiltrated macrophages, M2-like cells, and Ly6G PMNs (Metacluster 14 Figures 7C and S9A). Therefore, the combined treatment can be regarded as both complementary and synergistic.
Although neutrophils are considered to play a key role in the induction of anti-tumor immunity following PDT,42 cell markers used to define neutrophils (CD11b and Ly6G) are also expressed by PMN MDSCs.69 Potentially, the CD11b+Ly6G+Ly6C− population (Metacluster 5, Figures S8D and S9B) can be classified as pro-tumorigenic PMN MDSCs. This Metacluster was absent in the SFV/IFNγ group. Furthermore, the CD11b+Ly6GlowLy6Chigh population (Metacluster 6, Figures S8D; and S9C), which can be classified as tumor-promoting monocytic MDSCs, was induced under PDT treatment but decreased with SFV/IFNγ and PDT + SFV/IFNγ treatments. Both observations indicate anti-tumor potential of SFV/IFNγ.
Overall, PDT boosts innate immune activation (DCs, NKs, iNOS+), while SFV/IFNγ diminishes immune suppression (Th, Tregs, and CD206+ macrophages, and MDSCs), creating a more immunostimulatory TME. However, even though the combined treatment enhances innate immune cells (mature DCs, NKs, and iNOS, and reduces CD206) and decreases Tregs, it cannot yet activate CTLs in this tumor model. More effective outcomes could potentially be achieved by employing the TLR2/1 ligand PAM3CSK4, as we previously demonstrated the synergistic effect of TLR2 agonists and SFV vectors.10 Additionally, targeted Ce6 delivery systems, such as nanoparticles or micelles, could potentially enhance the efficacy of the therapy by improving tumor accumulation and reducing off-target effects.70,71,72,73,74
One of the principal limitations of PDT is the limited light penetration into biological tissues. In this study, we applied NIR (650 nm) light, which provides deeper penetration than light in the visible range.75 Nevertheless, further improvement in both penetration and biosafety can be achieved by using longer-wavelength NIR light (750–1350 nm). Moreover, PDT efficacy in deep-seated tumors can be enhanced using strategies such as intraoperative or interstitial light delivery or novel approaches like upconversion nanoparticles and X-ray-activated systems.76,77,78,79 Importantly, the systemic toxicity of Ce6-based PDT has been reported to be low in multiple studies,52,80,81 and our own observations are consistent with this. Throughout the treatment period, no significant changes in mouse body weight were detected among the experimental groups. Furthermore, previous studies have demonstrated that replication-deficient SFV vectors, which undergo only a single round of infection and produce no progeny virions, do not induce systemic adverse effects.8,9,82 These findings indicate that the combined therapy is well tolerated and support its favorable safety profile, strengthening its translational potential.
4T1 is a well-established model for studying breast metastatic tumors83,84; however, we did not assess metastasis formation in this study, as it could interfere with the detection of the Ce6 signal. Future studies should include dedicated experimental setups to evaluate the therapy’s impact on metastasis, for example, by combining Ce6-PDT with histological analysis of secondary organs. The abscopal, or systemic, effect primarily mediated by adaptive immune system activation, as demonstrated in previous studies,54,85,86 is expected to enhance the efficacy of combined PDT treatment with SFV vectors delivering TME-modulating cytokines. Nevertheless, evaluating the overall impact of the combined therapy on the immune microenvironment becomes challenging when treatment leads to very small or absent tumors, as key tumor–immune interactions may be minimal or entirely lacking altogether. Furthermore, the observed immune changes may represent a later phase of immune remodeling (day 19) rather than an immediate response to treatment. To better understand the mechanisms behind successful combination therapy, it is important to analyze the tumor and its immune context during the acute treatment phase. Nonetheless, this study offers valuable insights into the core mechanisms of the combined therapy, particularly its capacity to induce a robust innate immune response. Although new strategies and challenges of phototherapy in cancer were recently outlined,87 combining PDT with virus-based delivery of immune-modulating genes may offer a novel and promising direction for both PDT and virus-based gene therapy, representing a synergistic approach to cancer treatment.

Materials and methods

Cell lines
Murine mammary gland adenocarcinoma 4T1 cells (ATCC CRL-2539) were cultured in DMEM-GlutaMAX (Cat. No. 31966-021; Gibco, Life Technologies, Thermo Fischer Scientific, USA) supplemented with 10% fetal bovine serum (FBS; Cat. No. FBS-HI-12A; Capricorn Scientific) and 40 μg/mL gentamicin (Cat. No. 00-0442; Sopharma, Bulgaria).
4T1 cells expressing GFP and firefly luciferase (Luc) were obtained from Imanis Life Sciences. 4T1-Fluc-Neo/eGFP-Puro (4T1/GFP) cells were cultivated in RPMI-1640 medium (Cat. No. 12-115F; Lonza BioWhittaker) containing 10% FBS, 2 mM L-glutamine (Cat. No. 25030-024; Gibco, Life Technologies), 1% penicillin/streptomycin (PEST; Cat. No. 15070-063; Gibco, Life Technologies), 0.1 mg/mL G418 (Cat. No. 10131-027; Gibco, Life Technologies), and 2 μg/mL puromycin (Cat. No. ant-pr; InvivoGen).
Fibroblast-like L929 cells (ATCC CCL-1) were cultivated in RPMI-1640 medium supplemented with 10% FBS, 1% PEST, and 2 mM L-glutamine.
Baby hamster kidney fibroblasts (BHK-21) were obtained from ATCC (ATCC CCL-10) and cultured in Glasgow’s MEM (Cat. No. 21710025; Gibco, Life Technologies) supplemented with 5% FBS, 10% Tryptose Phosphate Broth solution (Cat. No. 18050039; Gibco, Life Technologies), 20 mM HEPES (Cat. No. 15630056; Gibco, Life Technologies), 2 mM L-glutamine, and 1% PEST.
All cells were incubated in a humidified 5% CO2 incubator at 37°C and passaged using 0.05–0.25% Trypsin solution (Cat. No. 15400-054; Gibco, Life Technologies).

Formation of spheroids
Cancer cell (4T1, 4T1/GFP) spheroids were generated using a 96-well Black Round Bottom Ultra-Low Attachment plate (Cat. No. CLS4515; Corning, Life Sciences). Cells were collected from monolayers by trypsin treatment, filtered through a 40 μm cell strainer (Cat. No. CLS431750; Corning, Life Sciences), counted, and resuspended in cultivation medium supplemented with 10% FBS, 1% PEST, and 2 mM L-glutamine. The cell suspension was added to a 96-well Ultra-Low Attachment plate (3⋅103 cells in 100 μL per well) and incubated in a humidified 5% CO2 incubator at 37°C. The size of the GFP-expressing spheroids was measured by fluorometry assay using Victor3V 1420-040 Multilabel HTS Counter (PerkinElmer, emission filter, 485 nm; detection filter, 535 nm). The viability of the 4T1 spheroids was measured with CellTiter-Glo 3D Cell Viability Assay (Cat. No. G9682; Promega).

Isolation, generation, and polarization of Bone Marrow-Derived Macrophages
Mouse macrophages were differentiated from bone marrow cells according to an established protocol using the L929 cell-conditioned medium as a source of macrophage colony stimulating factor (M-CSF).88 Bone marrow from the femurs and tibiae of the hind legs of 8- to 12-week-old female BALB/c mice was isolated as described previously.24
Non-polarized BMDMs or M0 phenotype was achieved by cultivating cells in RPMI-1640 medium supplemented with 10% FBS (Cat. No. ES-009-B; Millipore), 10% M-CSF, 1% PEST, and 1% L-glutamine. To polarize cells into M1 or M2 phenotype, the medium was supplemented either with 50 ng/mL recombinant mouse IFNγ (Cat. No. BMS326; eBioscience) and 100 ng/mL Pam3SCK4 (Cat. No. tlrl-pms; InvivoGen) to achieve M1 phenotype or with 20 ng/mL IL-4 (Cat. No. RP-8666; Invitrogen) to acquire M2 phenotype.

Photodynamic treatment
Chlorin e6 (Ce6; Cat. No. Ce6; Frontier Scientific, Logan, UT, USA) was dissolved in DMSO to 10 mg/mL. For the in vitro photodynamic treatment, Ce6 was added to the medium to obtain a final concentration of 0.5–10 μg/mL and 8 h later activated with a 650 nm laser or multi-well plate LED Array System consisting of LEDA-x array and LAD-1 array driver (Amuza Inc, USA) to achieve an irradiation dose of 1–5 J/cm2.

Viability assay
To measure the viability of the cancer cells, the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) cell viability assay was used as described by Zajakina et al.89 Briefly, the medium was replaced with 0.5 mg/mL MTT (Affymetrix, Cleveland, USA) solution in DMEM/F-12 without phenol red (Cat. No. 11039021; Gibco, Life Technologies) supplemented with 5% FBS. The cells were incubated for 2 h in a humidified 5% CO2 incubator at 37°C. After incubation, MTT solubilization solution consisting of 10% Triton X-100 (Cat. No. T8787-100ML; Sigma-Aldrich, Merck) and 0.1 M HCl in anhydrous isopropanol was added to dissolve the formed formazan crystals. The absorbance was measured using a microplate spectrophotometer (BioTek Instruments, Winooski, USA) at a test wavelength of 570 nm and a reference wavelength of 620 nm. Cell viability (%) was calculated using the following equation: Cell viability (%) = (570 nm–620 nm)/(control 570 nm–620 nm) × 100%, where the control is the value obtained from not treated cells.
To determine the viability of BMDMs, Cell Counting Kit–8 (CCK–8; Cat. No. 96992; Sigma-Aldrich, Merck) was used. In summary, the medium was replaced with 10 μL of CCK-8 reagent in 100 μL of fresh medium. The cells were incubated in a humidified 5% CO2 incubator at 37°C for 1–4 h. The absorbance at 450 nm was measured using a microplate spectrophotometer. Cell viability was calculated as % of non-treated cell absorbance.

Nitric oxide assay
A Nitric Oxide Assay Kit was used (Cat. No. EMSNO; Invitrogen) to determine the level of NO in the cell culture medium. NO is measured by the determination of nitrites. Briefly, 50 μL of cell culture medium was collected from each well, centrifuged, and used for NO quantification. The optical density was measured at 540 nm using a spectrophotometer.

Reactive oxygen species assay
The levels of ROS were measured using 2′,7′-Dichlorofluorescin Diacetate (H2DCFDA; Cat. No. 287810; Sigma-Aldrich, Merck). Upon exposure to ROS, this compound is oxidized and converted into a highly fluorescent molecule. The intensity of the fluorescent signal is proportional to the amount of ROS present in the cells.
To measure ROS production after SFV infection, 10 μM H2DCFDA was added to the cells, and the cells were incubated in a humidified 5% CO2 incubator at 37°C for 1.5 h. The fluorescence (ex/em 485/535) was measured using Victor3V 1420-040 Multilabel HTS Counter.
To measure ROS production after Ce6/LED treatment, the cells were washed with PBS supplemented with 1% FBS (1% PBS) twice. Next, a solution containing both Ce6 and 10 μM H2DCFDA was added to the cells, and the cells were incubated in a humidified 5% CO2 incubator at 37°C for 2 h. After the incubation, the cells were washed with 1% PBS and cultivated in DMEM/F12 for 30 min before being exposed to 3–5 J/cm2 NIR light. Lastly, the cells were incubated for 2 h, and the fluorescence (ex/em 485/535) was measured using Victor3V 1420-040 Multilabel HTS Counter.

Phagocytosis assay
In this experiment, the BMDMs were labeled with CellTracker Blue CMAC Dye (Cat. No. C2110; Invitrogen) and the cancer cells were labeled with CellTracker Green CMFDA Dye (Cat. No. C2925; Invitrogen). When macrophages engulfed and digested cancer cells through phagocytosis, the macrophages became labeled with CMFDA. The process of phagocytosis was assessed by counting cells that were positive for both CMAC and CMFDA dyes.
BMDMs were plated to a 12-well plate with a density of (8–10)×104 cells per well; 24 h later, BMDMs were polarized as described above. On the following day, M0, M1, and M2 macrophages were stained with CMAC Dye following the manufacturer’s instructions. Briefly, the cells were incubated with 5 μL/mL CMAC solution in a serum-free medium at 37°C. After 30 min, the cells were washed and used for experiments.
4T1 cancer cells were stained with CMFDA Dye following the manufacturer’s instructions. In brief, the cells were incubated with 0.5 μL/mL CMFDA solution in a serum-free medium at 37°C. After 30 min, the cells were washed and collected, and (8–10)×104 cells were added to BMDMs. The cells were co-cultivated for 2 or 24 h and then collected for flow cytometry analysis. The percentage of macrophages that had phagocytosed cancer cells was calculated as a percentage of CMAC+CMFDA+ cells in the CMAC+ population.

Flow cytometry
For flow cytometry analysis, the cell suspension was centrifuged at 410 g for 5 min and washed with PBS. Spheroids were treated with 0.05% trypsin solution for 5 min. Then, the cells were washed with PBS supplemented with 10% FBS (10% FBS) and centrifuged at 410 g for 10 min.
For staining of cancer cells, they were incubated with 1% bovine serum albumin in 10% FBS for 30 min at 4°C. After the incubation, 800 μL of 10% FBS was added, and the suspension was centrifuged at 410 g for 10 min. Next, the cell pellet was resuspended in 50 μL of 10% FBS and primary anti-Calreticulin antibodies (Cat. No. PA3900; Thermo Fisher Scientific) or the Rabbit IgG Isotype Control was added. Once the cells were incubated with primary antibodies for 40 min at 4°C, they were washed twice with 10% FBS. Then the cells were incubated with Alexa Fluor 488-coupled goat anti-rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody (Cat. No. A11008; Invitrogen) for 40 min at 4°C. The incubation was followed by two washing steps with 10% FBS and staining with Annexin-PE (Cat. No. 640947; BioLegend) according to the manufacturer’s instructions.
After staining, the cell pellet was resuspended in 300 μL of 0.5% paraformaldehyde in PBS. The experiment was repeated twice, each staining in duplicate. To create the compensation matrix, UltraComp eBeads (Cat. No. 01-2222; Invitrogen, Thermo Fisher Scientific) were used. The stained cells were kept at 4°C and analyzed the next day using a FACSAria BD Hardware flow cytometer using FACSDiva Software (BD Biosciences, San Jose, CA, USA). The data were analyzed by FlowJo v10.0.7.

Production of replication-deficient viral particles
The pSFV1 vectors encoding, accordingly, the Discosoma sp. red fluorescent protein (dsRed or RFP), GFP, Luc, and IFNγ genes were generated in our lab as described previously.28,89 Prof. Henrik Garoff (Karolinska Institute, Sweden) has generously provided the pSFV1 and pSFV-Helper1 plasmids.90
Replication-deficient viral particles were produced as described by Vasilevska et al.82 Briefly, the recombinant RNAs were packaged into viral particles by co-electroporation with SFV-Helper1 RNA in BHK-21 cells. After 48 h, the cell cultivation medium containing the infectious viral particles was concentrated according to the procedure of Hutornojs et al.91 The concentrated viral particles were stored at −70°C until further use. The infectious virus titers and infection efficacy of SFV/RFP and SFV/GFP were determined by fluorescence microscopy. The infectious virus titers of SFV/Luc and SFV/IFNγ were determined by immunostaining with anti-SFV nsp1 antibody (a kind gift from Prof. A. Merits, Tartu, Estonia) as previously described.28 Further infection of cancer cell monolayers and spheroids was performed according to the previously described protocols.10,92
Infection efficacy of SFV/Luc was assessed based on the luciferase activity measured with a luciferase assay (Cat. No. E1501; Promega, Madison, WI, USA). For the luciferase assay (Cat. No. E1501; Promega, Madison, WI, USA), the cells were lysed in 100 μL of cell culture lysis buffer (Cat. No. E1531; Promega) and centrifuged at 600 rcf for 5 min, and 2 μL of the cell lysate was used immediately for the measurement of the luciferase enzymatic activity with the Victor3V 1420-040 Multilabel HTS Counter (PerkinElmer).

Confocal microscopy
The samples have been studied under confocal laser scanning microscopy using a Leica TCS SP8 (Leica Microsystems, Wetzlar, Germany, with a 400 Hz scan speed). The fluorescence was detected as follows: a HeNe (633 nm) laser was used for the excitation of Ce6, and an argon (458 nm) laser was used for GFP. For fluorescence detection, HyD was used for Ce6 (640 nm–696 nm) and GFP (464 nm–504 nm). The images were analyzed using Leica Application Suite X 3.1.5. software (Leica Microsystems, Germany).

Experiments with animals
Female 8-week-old BALB/c mice were purchased from the Laboratory Animal Center, University of Tartu (Tartu, Estonia). The mice were housed 5 per cage in a climate-controlled room (temperature 22 ± 2°C and humidity 50 ± 10%) under a 12 h light/dark cycle and provided a standard diet and water ad libitum. All animal experimental protocols were approved by the Latvian Animal Protection Ethical Committee of Food and Veterinary Service (Permit Nr. 153/2024, from 7 June 2024, Riga, Latvia).

Ce6 distribution in vivo
Ce6 distribution was tested in the orthotopic 4T1 model of breast cancer. 4T1 cells were suspended in PBS, and 20 μL of cell suspension containing 1×105 cells was injected into the right thoracic mammary gland fat pads (day 0). When the tumors reached 4–5 mm in diameter, on day 4, mice were intravenously injected with pure Ce6 (in PBS) at a concentration of 2.5 mg/kg (100–110 μL depending on mouse weight). After 1.5 h, 4 h, and 18 h, the mice were anesthetized with isoflurane/oxygen, placed in an in vivo imaging system (IVIS Spectrum, PerkinElmer), visualized, and then humanely sacrificed. The lungs, livers, spleens, tumors, and kidneys were removed and placed in 24-well plates to quantify Ce6 fluorescence (IVIS, PerkinElmer). The data were analyzed using the total photon flux emission (photons/second) in the regions of interest using Living Image version 4.5 (PerkinElmer).
The mice blood was taken for the Ce6 quantification. Blood was incubated for 30 min at 37°C, transferred to 8°C for another 4 h, and centrifuged twice at 6,000 rpm for 20 min each to obtain serum. Ce6 concentrations in serum were measured by fluorimetry (485/665 nm) using a Victor 3V spectrophotometer and normalizing the resulting RFU to μg/mL of the Ce6 standards curve.

Combined treatment in vivo
Treatment of 4T1 tumors was tested in the orthotopic 4T1 model of breast cancer. Twenty microliters of cell suspension containing 2×104 cells was injected into the right thoracic mammary gland fat pads (day 0). On day 4, PDT was performed: mice were intravenously administered Ce6 at a concentration of 6.25 mg/kg (or PBS as a control), and after 1.5h, tumors were irradiated with 650 nm NIR light (120 J/cm2, 50 mW) as demonstrated in Figure S3A. On day 6, PDT was repeated. On day 8, an intratumoral injection (50 μL) of SFV vectors (or PBS as a control) was performed with 1.65×107 iu of SFV/IFNγ or SFV/Luc per tumor. The PDT was repeated on days 11 and 13, and SFV therapy was repeated on day 15. On day 19, the animals were anesthetized and sacrificed, then the tumors were removed, weighed, and subjected to immune cell isolation for flow cytometry.
The general protocol used for flow cytometry analysis of 4T1 tumors has been described previously.10 Tumors were homogenized with collagenase A (Cat. No. 10103586001; Roche, Basel, Switzerland) at 1.5 mg/mL and DNase at 15 μg/mL (Cat. No. ENZ-417; Prospect Medical Holdings, Los Angeles, CA, USA). Then, the cells were treated with 1X RBC Lysis Buffer (Cat. No. 00-4333-57; Thermo Fischer Scientific). The cells were counted, and 4×106 cells per staining were used for further procedures. Cells were first stained with Live-or-Dye Fixable Viability Staining Kit 350/448 (Cat. No. 32002; Biotium) to determine cell viability. Then, the cells were resuspended in a blocking solution containing 12.5 μg/mL mouse IgG diluted and incubated for 30 min on ice. After blocking, the cells were stained with fluorochrome-labeled monoclonal antibodies. The used monoclonal antibodies are summarized in Table S1. For intracellular staining, the BD Cytofix/Cytoperm Fixation/Permeabilization Solution (Cat. No. 554722) and BD Perm/Wash Perm/Wash Buffer (Cat. No. 554723) were used according to the provided instructions. Stained cells were analyzed with Cytek Aurora 5L. UltraComp eBeads (Cat. No. 01-2222; Invitrogen, Thermo Fisher Scientific) were used for reference controls.

Flow cytometry analysis
Flow cytometry data were first analyzed using FlowJo (v10.0.7). To perform high-dimensional analysis, gated .fcs files were imported into R (v4.4.1) using the flowCore package.93 Non-marker and non-cell parameters (e.g., Time, FSC/SSC, Viability dye) were excluded. Fluorescence intensity values were transformed using the arcsinh function with a cofactor of 150.
Each sample was downsampled to a maximum of 5,000 cells to ensure balanced representation. For datasets with substantial variation in cell counts, downsampling was performed to match the sample with the lowest cell number, ensuring equal representation from all groups and preventing sampling bias. The data were then concatenated into a single matrix for downstream analysis.
Clustering was performed using the FlowSOM package.94 A self-organizing map (SOM) was built using relevant marker channels, with 100 nodes further aggregated into metaclusters via consensus hierarchical clustering. To visualize high-dimensional structure, UMAP was applied using the uwot package95 with the parameters: n_neighbors = 15, min_dist = 0.1, and metric = “euclidean.” To examine the spatial relationship between marker expression and UMAP dimensions, Pearson correlation coefficients were calculated between marker intensities and UMAP1/UMAP2 coordinates.
For phenotypic profiling, the median expression of each marker was calculated for each metacluster and visualized using heatmaps (pheatmap package96). The relative frequency of each metacluster in individual samples was also computed and used for group-wise comparisons. To quantify treatment effects, log2 fold changes (log2FC) in metacluster abundance were calculated relative to the control group. Differences in metacluster frequencies between groups were tested for statistical significance using the Wilcoxon rank-sum test.

Statistical analysis
Data are presented as the mean ± standard deviation (SD) from at least three independent experiments. Statistical analysis of flow cytometry frequencies and functional assays was performed using GraphPad Prism (version 8.02). Normality of data distribution was assessed using the Shapiro-Wilk test. Datasets that followed a normal distribution were analyzed using one-way or two-way ANOVA, followed by the Holm-Sidak multiple comparisons test. For datasets that did not meet the assumption of normality, non-parametric tests such as the Wilcoxon rank-sum test with Benjamini-Hochberg correction (for pairwise comparisons) or the Kruskal-Wallis test followed by the Dunn multiple comparisons test (for multiple groups) were applied. p-values less than 0.05 were considered statistically significant.

Data and code availability
The data are available upon request from the corresponding author; the flow cytometry data are available in the Zenodo online database (https://doi.org/10.5281/zenodo.17457137).

Acknowledgments
The authors are thankful for the “Mikrotīkls” doctoral scholarship provided by the University of Latvia Foundation and Mikrotīkls enterprise in the field of exact and medical sciences. The authors sincerely thank Aiva Plotniece and Karlis Pajuste for the Ce6 solution preparation and Fēlikss Rūmnieks for his support with confocal microscopy.
This research was funded by the 10.13039/501100005375Latvian Council of Science, grant number lzp-2021/1-0283 “Programming of breast cancer microenvironment to create a ‘hot’ immune-responsiveness in cancer” and lzp-2024/1-0254 “Exploring tumor-associated myeloid cell dynamics in response to alphavirus-driven immunomodulation.”
During the review process, K.K. was supported by the ERDF Project No. 1.1.1.8/1/24/I/003 “Strengthening the Research and Development Capacity of Doctoral Studies at the 10.13039/501100005414University of Latvia in the Fields of Smart Specialisation.”
All animal experimental protocols were approved by the Latvian Animal Protection Ethical Committee of Food and Veterinary Service (Permit Nr. 153/2024, from 7 June 2024, Riga, Latvia).
Icons were obtained from bioicons by Servier https://smart.servier.com/, licensed under CC-BY 3.0 Unported https://creativecommons.org/licenses/by/3.0/.

Author contributions
K.K. performed investigation, formal analysis, and data visualization; contributed to methodology development; and drafted the manuscript. Z.R., D.L., O.N., and K.S. contributed to experimental methodology and investigation. D.S. and J.J. conducted in vivo experiments and contributed to the investigation. A.Z. contributed to methodology development and experiment planning, provided critical manuscript revisions, and secured funding for the study. All authors read and approved the final manuscript.

Declaration of interests
The authors declare no conflicts of interest.