Circadian rhythm of dendritic cells directs the time-dependent efficacy of photodynamic therapy

Abstract

Background

While cancer treatment efficacy exhibits circadian differences, the role of circadian rhythms in photodynamic therapy (PDT) remains unclear. Understanding this relationship is crucial for treatment optimization.

Methods

Using B16F10 and 4T1 tumor-bearing mice, PDT was administered at different Zeitgeber times (ZT). Immune cell infiltration was analyzed using flow cytometry and immunofluorescence. We also compared the effects of PDT treatment between NSG mice and CD11c-DTR mice. Transcriptomic profiling of sorted CD11c⁺MHCII⁺ DCs was performed via RNA-seq. In vitro assays measured immunogenic cell death (ICD) markers and dendritic cells (DC) maturation. A PDT vaccine model was also evaluated.

Results

PDT efficacy showed strong circadian dependence, optimal at ZT10 and weakest at ZT22 in immunocompetent mice—a rhythm maintained in constant darkness but inverted with light-cycle reversal, confirming endogenous control. The effect was abrogated in immunodeficient NSG mice, pinpointing an immune-mediated mechanism. Further analysis revealed circadian oscillations in the infiltration of CD4⁺, and CD8⁺ T cells into tumors and spleens, peaking at ZT10. Crucially, the rhythmic abundance and function of CD11c⁺MHCII⁺ DCs were identified as the pivotal regulator. Specific DC ablation eliminated both the circadian efficacy of PDT and the rhythmic T-cell infiltration. Transcriptomics revealed ZT10 DCs were polarized toward adaptive immunity, while ZT22 DCs favored pro-tumorigenic pathways. PDT-induced ICD promoted DC maturation in vitro and enhanced antitumor responses in vivo.

Conclusion

The intrinsic circadian rhythm of DCs is a master regulator of PDT efficacy, rhythmically governing adaptive antitumor immunity. These findings advocate chrono-immunotherapy through precise treatment timing to maximize clinical benefits.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12964-026-02973-2.

Introduction

Cancer has become a major threat to human life and health, with its incidence increasing annually. According to projections from the World Health Organization (WHO) and the International Agency for Research on Cancer, global cancer incidence and mortality rates will continue to rise. By 2050, the cancer burden is projected to increase by approximately 77% [1]. Besides traditional surgery, radiotherapy, and chemotherapy, an increasing number of cancer treatment methods have emerged. PDT, as a new non-invasive modality for treating cancer, was approved for clinical use in the treatment of esophageal cancer in the early 1990s [2].

PDT is based on the selective accumulation of photosensitizers in tumors. Following illumination with light of a specific wavelength and in the presence of molecular oxygen within cells and tissues, the photosensitizer is activated, thereby triggering two types of reactions: (Ⅰ) Energy is transferred to molecular oxygen, generating highly reactive singlet oxygen. (Ⅱ) The photosensitizer directly transfers energy to cellular molecules, producing free radicals and other reactive species. Both pathways generate reactive oxygen species (ROS), ultimately leading to cell death and tumor tissue destruction [3, 4]. In addition, a growing number of studies have found that PDT-induced dying tumor cells release damage-associated molecular patterns (DAMPs) and tumor-associated antigens (TAAs), thereby triggering ICD. This subsequently leads to the maturation of DCs and activates tumor-specific immune responses, demonstrating its significant potential for application in enhancing anti-tumor immunotherapy [5–9].

The tumor-specific immune response resulting from ICD is associated with the tumor’s immune status, particularly DCs. Meanwhile, it has been increasingly demonstrated that the overall function of the immune system is regulated by circadian rhythms.

Circadian rhythms are endogenous timing systems developed by organisms over the course of long-term evolution. They generate and maintain periodic physiological fluctuations of approximately 24 h to adapt to external day-night changes and coordinate physiological functions [10–12]. Their regulatory center is located in the suprachiasmatic nucleus (SCN), and their biological basis is the “molecular clock”. This clock is constituted by core genes (such as CLOCK, BMAL1, PER, and CRY) that generate 24-hour oscillations through transcription-translation feedback loops: the CLOCK/BMAL1 complex activates and induces the expression of PER/CRY, which in turn inhibits the former, establishing a cycle [13]. Core clock genes regulate a vast array of downstream genes (approximately 10%-43% of genes exhibit rhythmic expression) and various physiological functions [14, 15].

Recent studies have shown that immune function exhibits significant circadian characteristics and that this rhythmicity is precisely regulated by the circadian system [16]. Specifically, circadian rhythms influence the quantity, activity, and phenotypic changes of most immune cells, such as DCs, macrophages, neutrophils, NK cells, T cells, and B cells [17–23]. For instance, Wang et al. demonstrated that tumor volume in murine models is contingent upon the timing of cell inoculation. Specifically, tumor growth was most significantly attenuated when implantation occurred at ZT9, a phenomenon primarily driven by circadian regulation of dendritic cell recruitment and functionality [17]. Circadian system regulates the immune system through the periodic expression of core clock genes, thereby enabling immune responses to adapt to external diurnal changes and perform defense tasks more efficiently [24]. Concurrently, the immune system can also exert feedback regulation on the circadian system by releasing signals such as immune factors, forming a bidirectional regulatory network between the two [25]. A recent statistical study of 329,261 human samples from the United Kingdom demonstrated rhythmicity in human peripheral blood leukocyte counts, with the lowest levels occurring in the early morning. Neutrophils, monocytes, and lymphocytes reached their peak numbers at 15:00 and 19:00, respectively. Due to differing rest-activity patterns, this rhythm is inverted in mice [26]. Importantly, synchronizing treatment timing with immune cell rhythms has been proven to enhance the therapeutic efficacy against tumors, significantly reducing tumor volume and prolonging survival. However, current research primarily focuses on treatments such as immune checkpoint inhibitors and chimeric antigen receptor T cell therapy [22, 27–29]. Currently, there is limited research on the correlation between PDT and circadian rhythms, with existing studies primarily focusing on the circadian pharmacokinetics of photosensitizers and the relationship between PDT and the circadian expression of its target molecules [30, 31]. Exploration of the correlation between tumor immunity triggered by PDT-induced ICD and circadian rhythms has been extremely limited.

Therefore, this study, by observing the differential effects of PDT treatment administered at different time points, revealed that circadian rhythms can significantly influence the therapeutic efficacy of PDT through their effects on the immune system. Key participants in this process include DCs and CD8⁺ / CD4⁺ T cells. Specifically, following PDT treatment, tumor tissue undergoes induced ICD. As circadian rhythms influence not only the tissue infiltration of immune cells but also regulate DC function genetically, administration at ZT10 enhances DC activation and subsequently promotes the adaptive immune process compared to administration at ZT22, resulting in superior therapeutic efficacy. Furthermore, due to the rhythmicity of immune cells, the therapeutic outcome of PDT-based vaccines is also influenced by the timing of vaccine administration.

In summary, this study provides a novel research perspective for selecting optimal PDT treatment timing. Specifically, by administering PDT during periods of heightened immune cell infiltration and activation—guided by circadian rhythmic variations in immune cells—the ICD effect can be enhanced, thereby augmenting the therapeutic efficacy of PDT.

Materials and methods

Cell culture

B16F10 cells and 4T1 cells were purchased from the National Infrastructure of Cell Line Resource and have been confirmed to be free of mycoplasma contamination. B16F10 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (AC01L043, Life-iLab) supplemented with 10% fetal bovine serum (FBS) (HK-CH500, HUANKE), 100 U/mL penicillin, and 100 µg/mL streptomycin (03.12001 A, EallBio). 4T1 cells were cultured in RPMI-1640 medium (C11875500BT, Gibco) supplemented with 10% FBS (HK-CH500, HUANKE), 100 U/mL penicillin, and 100 µg/mL streptomycin (03.12001 A, EallBio). The cells were then maintained in a 5% CO₂ incubator at 37 °C until the medium was replaced after 2–3 days. Cell passaging was performed when the cells reached approximately 80% confluence. Cells harvested during the logarithmic growth phase were used for subsequent experiments. Unless otherwise specified, tumor cells (5 × 10⁵ cells in 100 µL of PBS) were subcutaneously injected into the right axilla of mice under isoflurane anesthesia, while the control group received 100 µL of 1× PBS. Tumor volume was monitored every 1–2 days using a vernier caliper, and calculated using the formula: (length × width²) / 2. To minimize potential confounding factors, all tumor implantations for mice within the same batch were performed simultaneously using cells from the same passage and at the same time of day.

Animals

All animal experiments were conducted in accordance with protocols approved by the Peking University Animal Ethics Committee (LA2021274) and followed the guidelines for the care and use of laboratory animals established by Peking University. Female C57BL/6J mice (6 weeks old) and female Balb/c mice (6 weeks old) were purchased from the Department of Animal Sciences, Peking University Health Science Center. NSG mice (female, 6 weeks old; NOD-Prkdcscid IL2rgem9) were obtained from HFK BIOSCIENCE (Beijing, China). B-Cd11c-EGFP-DTR-Luc mice(C57BL/6JNifdc-Itgam^{tm1(EGFP−DTR−Luc)Bcgen}/Bcgen) were provided by BIOCYTOGEN (Beijing, China). All mice were housed under specific pathogen-free conditions with free access to food and water and maintained in an artificially controlled light–dark cycle (12 h light–12 h dark). In the experiments involving mice in the circadian reversal group and constant darkness group, all mice were first acclimatized to a standard light–dark cycle (12 h light–12 h dark)for 2 weeks to achieve stable adaptation. Subsequently, the mice were housed in a light-controlled darkroom facility, where they were subjected to acute circadian reversal or constant darkness in accordance with the experimentally designed conditions. After another 1 weeks of acclimatization under the corresponding light conditions, subsequent experimental procedures were carried out. Treatment times are indicated in Zeitgeber time (ZT), defined relative to the time of lights-on in the animal facility. For example, ZT4 corresponds to 4 h after lights-on (morning), ZT10 to 10 h after lights-on (evening), ZT16 to 4 h after lights-off (midnight), and ZT22 to 10 h after lights-off (early morning). In the case of acute circadian reversal, DL ZT10 corresponding to the biological ZT22 (LD ZT22) of the original cycle prior to entrainment.

For systemic depletion of DCs, B-Cd11c-EGFP-DTR-Luc mice were intraperitoneally injected with diphtheria toxin (DTx; List Biological Laboratories, #150, USA) at a dose of 50 ng per gram of body weight.

PDT treatment

After establishing the tumor-bearing mouse models as described above, PDT was administered to each group at predetermined time points on days 7, 9, and 11 post-tumor implantation. Verteporfin (CL 318952, MedChemExpress) was first injected via the tail vein at a dose of 2 µg per gram of body weight. After injection, a 30-minute interval was allowed to ensure systemic distribution and equilibrium of the drug. Subsequently, the tumor region was irradiated using a fiber optic probe connected to a 300 W xenon arc lamp (CEAULIGHT) equipped with a 600 nm band-pass filter. The light fluence rate was maintained at approximately 1.5 mW/cm², and irradiation was performed for 5 min.

Preparation of single-cell suspensions

Tumor tissues and spleens were collected at predetermined time points. Under sterile conditions, tissues were minced into approximately 1 mm³ fragments using surgical scissors. The fragments were then placed on a 70 μm cell strainer (NEST, 258368) and gently ground using the plunger of a sterile syringe to achieve thorough dissociation. The homogenate was rinsed with PBS through the strainer to collect the cell suspension. For splenic cell suspensions, the collected cells were centrifuged and resuspended in 3 mL of 1× RBC lysis buffer (Solarbio, R1010). After gentle pipetting, the suspension was incubated on ice for 5 min to lyse red blood cells. The cells were then centrifuged again and resuspended in PBS. All cell suspensions were filtered through a 70 μm strainer to obtain single-cell suspensions.

Flow cytometry analysis

Single-cell suspensions were first incubated with TruStain FcX™ (anti-mouse CD16/32) (101319, BioLegend) at a concentration of 1 µg per 10⁶ cells (in 100 µL) for 10 min on ice to block Fc receptors. After washing, cells were resuspended in Zombie Aqua™ Fixable Viability Kit (423101, BioLegend) and incubated for 10 min at room temperature in the dark. An antibody cocktail was then added directly to the cell suspension and incubated for 15 min at 4 °C. The following anti-mouse antibodies were used for immunostaining: APC/Cyanine7 anti-mouse CD45 (103115, BioLegend); PerCP/Cyanine5.5 anti-mouse CD3 (100217, BioLegend); FITC anti-mouse CD4 (100509, BioLegend); PE/Cyanine7 anti-mouse CD8a (100721, BioLegend); PE anti-mouse CD11c (117307, BioLegend); APC anti-mouse I-A/I-E (107613, BioLegend); Samples were analyzed using a CytoFlex S flow cytometer (Beckman Coulter), and data were processed with FlowJo 10 software (BD). To clearly visualize differences between groups, cell counts were normalized to the baseline. Specifically, the average cell number of the first group was set to 1.0, and the cell numbers of all other groups were expressed as a ratio relative to this control value.

Meanwhile, in order to verify the accumulation of the drug and eliminate the interference of pharmacokinetics as a factor that might affect the observed diurnal differences in the efficacy of photodynamic therapy, we intravenously injected VPF into B16F10 tumor-bearing mice at ZT10 and ZT22. Thirty minutes after the injection (tumor tissues were collected, separated into single-cell suspensions, and analyzed by flow cytometry to quantify the accumulation of VPF within the tumors.

Immunofluorescence imaging

Tumor tissues collected at ZT10 and ZT22 were paraffin-embedded and sectioned. Sections were deparaffinized in an eco-friendly dewaxing solution (G1128, Servicebio), rehydrated through graded ethanol and distilled water, and subjected to antigen retrieval in citrate-based antigen retrieval buffer (pH 6.0) (Servicebio, G1202, China) via water bath at 95 °C for 30 min. After cooling to room temperature, sections were circled with an immunohistochemical pen (G6100, Servicebio), incubated with 3% hydrogen peroxide in the dark for 25 min to block endogenous peroxidase, and then blocked with 3% BSA (GC305006, Servicebio) for 30 min. Primary antibodies were applied overnight at 4 °C, followed by incubation with HRP-conjugated secondary antibodies and tyramide signal amplification (TSA) reagents. The antibodies and reagents used included: Recombinant Anti-CD8α antibody (Rabbit mAb) (GB15068, Servicebio); Recombinant Anti-CD4 antibody (Rabbit mAb) (GB15064, Servicebio); Anti-CD11c Rabbit pAb (GB11059, Servicebio); HRP-conjugated Goat Anti-Rabbit IgG (H + L) (GB23303, Servicebio); iF488-Tyramide (G1231, Servicebio); iF555-Tyramide (G1233, Servicebio); iF647-Tyramide (G1232, Servicebio). After PBS washes, slides were stained with DAPI (G1012, Servicebio) for 10 min in the dark at room temperature. Autofluorescence was quenched using an autofluorescence quenching reagent (G1221, Servicebio) for 5 min. Slides were mounted with anti-fade mounting medium (G1401, Servicebio) and imaged using a Pannoramic MIDI slide scanner (3DHISTECH). Image analysis was performed using ImageJ.

Sorting of CD11c⁺MHCII⁺ cells and RNA transcriptome sequencing

To isolate DCs, inguinal lymph nodes (draining lymph nodes) were collected 48 h post-PDT at ZT10 and ZT22, corresponding to the treatment time points. Single-cell suspensions were prepared as described earlier. CD45⁺CD11c⁺MHCII⁺ cells were sorted using a FACS Aria SORP flow cytometer (BD) and collected in TRIzol reagent (beyotime, R0016).

Total RNA was extracted using the MJZol Total RNA Extraction Kit (T01-200, MAJORBIO) and purified with an RNA Purification Kit (Transgen, ER701-01). Library construction was performed using the Illumina NovaSeq Reagent Kit. RNA concentration and purity were assessed using a Nanodrop 2000, integrity was evaluated via agarose gel electrophoresis, and RIN values were determined with an Agilent 5300 system. mRNA was isolated using oligo(dT) beads for polyA selection. cDNA was synthesized, adaptors were ligated, fragments were size-selected, and libraries were enriched before sequencing on a NovaSeq X Plus platform.

Raw sequencing data were quality-controlled using fastp (v0.23.4) to remove adapter sequences, reads with > 10% N bases, and low-quality reads (Phred score < 20). Clean reads were aligned to the mouse mm10 UCSC genome using HiSat2 (v2.2.1). Gene expression quantification was performed with RSEM (v1.3.3). Differential gene expression analysis was conducted using DESeq2 (v1.42.0), with significantly differentially expressed genes (DEGs) defined as those with p < 0.05 and |Fold Change| ≥ 1.5. Subsequent data analysis was performed on the Majorbio cloud platform.

RNA isolation and quantitative real time PCR (qRT-PCR)

The qRT-PCR was performed as described by Zong et al. [32]. Total RNA of colon was extracted with Trizol reagent (Beytime, R0016). The quality of the extracted RNA was evaluated using NanoDrop2000 (Thermo Fisher Scientific. USA) and the absorbance rates (260/280 nm) of all samples ranged from 1.8 to 2.0. RNA was reverse transcribed using the reverse transcription kit SweScript All-in-One RT SuperMix for qPCR (One-Step gDNA Remover) (Servicebio, G3337) according to the manufacturer’s instructions, first strand cDNA was obtained. Subsequent qRT-PCR was performed on qTOWER³ Real Time PCR system (analytik-jena, Germany) with TB Green Premix Ex Taq II (Takara, RR820A) and gene-specific primers (Supplementary Table 1). The target genes relative mRNA expression was analyzed by the 2-ΔΔCT method. GAPDH gene was used as normalization control.

Measurement of ATP and HMGB1 release and CRT translocation

B16F10 cells (5 × 10⁵) were seeded in 6-well plates and incubated with varying concentrations of verteporfin (VPF; 1 µM or 2 µM) for 12 h. After washing with PBS to remove unbound photosensitizer, cells were irradiated using a 300 W xenon arc lamp (CEAULIGHT) with a 600 nm band-pass filter (∼1.5 mW/cm²). Supernatants were collected 4 h post-treatment, centrifuged at 15,000 rpm for 3 min at 4 °C, and immediately assessed for ATP release using an Enhanced ATP Assay Kit (Beyotime, China) and for HMGB1 release using an HMGB1 ELISA Kit (EHRC01, Neobioscience).

For calreticulin (CRT) translocation analysis, 5 × 10⁵ B16F10 cells were seeded in confocal dishes and treated as above. After PDT irradiation (3 min), cells were incubated for 2 h, fixed with 4% paraformaldehyde for 15 min at room temperature, and incubated overnight with calreticulin polyclonal antibody (27298-1-AP, Proteintech). After washing, cells were incubated with CoraLite488-conjugated Goat Anti-Rabbit IgG (H + L) (SA00013-2, Proteintech) for 90 min in the dark at room temperature. Nuclei were stained with DAPI (C1006, Beyotime) for 10 min, and images were acquired using a TCS-SP8 confocal microscope (Leica).

BMDC maturation analysis

Bone marrow-derived dendritic cells (BMDCs) were obtained by flushing tibiae and femora of mice with RPMI-1640 medium supplemented with 2% FBS. Bone marrow cells were plated at 2 × 10⁵/mL in 100 mm Petri dishes with 200 U/mL recombinant mouse GM-CSF (315-03-5UG, Thermo Fisher) and cultured at 37 °C under 5% CO₂. On day 3, 10 mL of fresh complete medium containing 200 U/mL GM-CSF was added. Half of the medium was replaced on days 6 and 8. On day 10, floating and semi-adherent cells (BMDCs) were harvested.

BMDCs were replated and stimulated with recombinant mouse GM-CSF (100 U/mL) for 6 h, followed by co-culture with PDT-treated dying B16F10 cells or untreated B16F10 cells at ratios of 8:1 or 2:1 (tumor cells: BMDCs). LPS (BL1386A, Lanjieke) was used as a positive control. After 24 h, cells were harvested, stained with PE anti-mouse CD11c (117307, BioLegend), APC anti-mouse I-A/I-E (107613, BioLegend), FITC anti-mouse CD86 (159219, BioLegend), and TruStain FcX™ (101319, BioLegend), and analyzed by flow cytometry to assess maturation.

Vaccine preparation and vaccination

B16F10 cells were induced to undergo apoptosis in vitro by PDT with 1 µM VPF as described. Apoptotic cells were collected, washed, and resuspended in PBS at 5 × 10⁶/mL. Mice were subcutaneously inoculated in the right flank with 100 µL (5 × 10⁵ cells) of dying B16F10 cells or PBS (control) at ZT4, ZT10, ZT16, or ZT22. On day 8 post-vaccination, live B16F10 cells (1 × 10⁵) were injected subcutaneously into the contralateral flank at the same Zeitgeber time. Tumor dimensions were measured every 1–2 days for 4 weeks using a vernier caliper. Mice were euthanized when the largest tumor volume exceeded 1500 mm³.

Statistical analysis

Unless otherwise specified, all data were plotted from independent biological replicates. All statistical analyses were calculated using GraphPad Prism 9.3.0 software and the significance was analyzed by Student’s t-test, one-way ANOVA or two-way ANOVA. *P < 0.05,**P < 0.01,***P < 0.001,****P < 0.0001, ns, not significant.

Results

Circadian rhythms influence the therapeutic efficacy of PDT

To investigate whether the efficacy of PDT is influenced by circadian rhythms, we treated four groups of B16F10 tumor-bearing mice at four different ZT (ZT4, ZT10, ZT16, and ZT22). Three PDT sessions were administered, and tumor size was monitored throughout the treatment period. Tumors and spleens were harvested at the respective treatment times 7 days after the final session (Fig. 1g, Supplementary Fig. 1a). Tumor volume changes post-PDT were significantly correlated with treatment timing: the most pronounced tumor shrinkage occurred with evening treatment (ZT10), while morning treatment (ZT22) resulted in minimal reduction (Fig. 1a-b, Supplementary Fig. 1c, e). Similar results were observed in the 4T1 mouse model (Supplementary Fig. 1d), indicating that the circadian effect on PDT efficacy is a general phenomenon independent of tumor type. Meanwhile, As shown in Supplementary Fig. 1b, there was no significant difference in the mean fluorescence intensity (MFI) of VPF within the tumor cells between the ZT10 and ZT22 groups. This result indicates that the accumulation of VPF in the tumor does not exhibit diurnal variation, thereby confirming that the diurnal difference in PDT efficacy is not due to variations in drug concentration.

Fig. 1.

The antitumor efficacy of PDT exhibits a strong circadian rhythm. a Growth curves of B16F10 tumors in mice receiving PDT at four different ZT (n = 6). b Tumor volumes on day 16 from (a). c Tumor growth in mice maintained under constant darkness and treated with PDT at CT10 or CT22 (n = 6). d Final tumor volumes on day 16 from the experiment in (c), compared by Student’s t test. e, f Tumor growth curves and endpoint volumes following PDT administered under a standard light-dark cycle (LD, n = 6) or an inverted dark-light cycle (DL, n = 6). g Schematic illustrating the normal (LD: Light-Dark) and reversed (DL: Dark-Light) lighting conditions. The number of days indicated in the process represents the duration of exposure to the corresponding light conditions. Red dots indicate PDT treatment times; green dots indicate tumor harvest times. Data are presented as mean ± SEM. Tumor growth curves (a, c, e) were analyzed using two-way ANOVA. Tumor volume (b, f) statistical significance was determined by one-way ANOVA followed by Tukey’s post-hoc test

Next, we transferred mice to constant darkness, as true circadian rhythms persist under constant conditions. We found that the circadian pattern of PDT efficacy remained consistent under continuous darkness, with morning treatment (CT22) still yielding poorer outcomes than evening treatment (CT10) (Fig. 1c, d, Supplementary Fig. 1f). However, when mice under normal light cycles were subjected to a 12-hour inverted dark-light cycle, the efficacy pattern of PDT reversed compared to that under normal light conditions (Fig. 1e, f, Supplementary Fig. 1g). This indicates that the differential effects of PDT are not directly light-dependent but can be entrained by light—a key feature of circadian regulation. In summary, these results demonstrate that the variation in PDT efficacy is driven by endogenous circadian rhythms and can be modulated by light entrainment.

The immune system mediates circadian differences in PDT efficacy

To determine whether the immune system underlies these circadian differences, we treated NSG mice bearing B16F10 tumors with PDT at ZT4, ZT10, ZT16, and ZT22. No time-dependent differences in therapeutic outcomes were observed, indicating that the circadian variation in PDT efficacy is immune-mediated (Fig. 2a-d).

Fig. 2.

Circadian immune rhythms determine the time-dependent efficacy of PDT. a–c Tumor growth curves (a), final volumes (b), and weights (c) of B16F10 tumors in NSG mice after three PDT treatments at different ZT times (n = 3). d Representative ex vivo images of B16F10 tumors from NSG mice. e Normalized counts of total infiltrating immune cells in spleens and B16F10 tumors harvested at four time points (n = 6); analyzed using cosinor regression. Shaded areas indicate dark phases. f–i Representative images (left) and quantification (right) of CD4⁺ T cells, CD8⁺ T cells, and CD11c⁺ cells in B16F10 tumors collected at ZT10 or ZT22 (n = 3). j, k Gating strategy for CD11c⁺MHCII⁺ cells, CD4⁺, and CD8⁺ T cells in spleens of 4T1 tumor-bearing mice after three PDT sessions. l Quantification of CD11c⁺MHCII⁺ cells in tumors and spleens from mice under constant darkness or reversed light cycle (n = 6). m Gating strategy for CD4⁺ and CD8⁺ T cells in B16F10 tumors under reversed light-dark (DL) conditions. Data are presented as mean ± SEM. Tumor growth curves (a) were analyzed using two-way ANOVA. Tumor volume (b) and tumor weight (c) statistical significance was determined by one-way ANOVA followed by Tukey’s post-hoc test. Other statistical comparisons were analyzed by Student’s t-test

We also statistically analyzed immune cells in tumors and spleens of wild-type mice. Seven days after PDT, the infiltration of key adaptive immune cells(CD4⁺ T, CD8⁺ T, DC cells) in both B16F10/4T1 tumors and spleens exhibited clear circadian rhythms. Total CD45⁺ leukocyte infiltration also showed rhythmic trends, though it did not reach statistical significance in the 4T1 model. The highest levels of infiltration and smallest tumor volumes were observed with ZT10 treatment, while the lowest levels occurred at ZT22 (Fig. 2e-k, Supplementary Fig. 2a, b). Immune cell analysis in mice maintained in constant darkness also revealed rhythmic infiltration patterns, and acute inversion of the light-dark cycle resulted in a corresponding inversion of immune cell rhythms (Fig. 2l, m, Supplementary Fig. 2c-f).

DCs determine circadian variation in PDT efficacy

To identify which immune cell type plays a decisive role in the circadian response to PDT, we focused on early immune events that could explain the observed effects. Flow cytometry and immunoimaging revealed that CD11c⁺MHCII⁺ antigen-presenting cells (APCs) in tumors exhibited clear circadian variation, with peak numbers at ZT10 and the lowest at ZT22.

To further examine the role of DCs, we used DTR mice (B-Cd11c-EGFP-DTR-Luc). Intraperitoneal injection of diphtheria toxin (DTx) selectively depleted DCs without affecting other immune cells. In DC-depleted mice, no significant difference in tumor size was observed between ZT10 and ZT22 PDT treatments (Fig. 3a-c). Flow cytometry confirmed complete ablation of CD11c⁺MHCII⁺ cells (Supplementary Fig. 2g), and the circadian rhythm of CD4⁺/CD8⁺ T cell infiltration was abolished (Fig. 3d-g). These results indicate that DCs regulate the rhythmic response of CD4⁺/CD8⁺ T cells and collectively determine the circadian efficacy of PDT.

Fig. 3.

DC depletion abolishes circadian differences in PDT efficacy. a, b. Tumor growth curves (a), and final volumes (b) in DTx-treated B-Cd11c-EGFP-DTR-Luc mice bearing B16F10 tumors after PDT administered at ZT10 or ZT22 (n = 6). c. Representative ex vivo images of B16F10 tumors from DTx-injected DTR mice. d–g. Quantification of CD4⁺ T cells (d, f) and CD8⁺ T cells (e, g) in spleens (d, e) and tumors (f, g) of DTR mice following three PDT sessions (n = 6). Data are presented as mean ± SEM. Tumor growth curves (a) were analyzed using two-way ANOVA. Other statistical comparisons were analyzed by Student’s t-test

Intrinsic mechanisms drive circadian regulation of DC function

To investigate whether circadian differences are driven by cell-intrinsic mechanisms, we collected draining lymph nodes 48 h after PDT at ZT10 and ZT22 and sorted CD11c⁺MHCII⁺ DCs. Principal component analysis (PCA) revealed distinct global gene expression profiles between ZT10 and ZT22 (Fig. 4a). A total of 824 DEGs were identified: 555 were upregulated and 269 downregulated at ZT10 compared to ZT22 (Fig. 4b). Notably, genes associated with circadian regulation exhibited significant temporal variation (Fig. 4c, Supplementary Fig. 3a).

Fig. 4.

DCs exhibit circadian patterns of gene expression. a PCA of transcriptional profiles of DCs isolated at ZT10 versus ZT22. Colors and shapes indicate groups; ellipses represent 95% confidence intervals. b Volcano plot showing DEGs in DCs (fold change ≥ 1.5). c Expression levels of Nr1d2, ID3, and Egr1 in CD11c⁺MHCII⁺ cells from both groups; analyzed by Student’s t test. d, e KEGG functional enrichment analysis of the top 30 biological pathways among significantly up- (d) and down-regulated (e) genes. f Heatmap of five immune-related pathways enriched among up-regulated genes. g Network diagram of KEGG pathways enriched among up-regulated DEGs. h qRT–PCR validation of selected DEGs identified by RNA sequencing (n = 3). All data are represented as mean ± SEM

KEGG enrichment analysis showed that pathways related to adaptive immune activation—such as Th1 and Th2 cell differentiation, Leukocyte transendothelial migration, and C-type lectin receptor signaling pathway—were upregulated at ZT10 (Fig. 4d, f). These pathways promote robust, coordinated, and specific adaptive immune responses that effectively clear tumor cells. In contrast, pathways associated with innate immunity and inflammation—including Toll-like receptor signaling, Neutrophil extracellular trap formation, and IL-17 signaling pathway—were downregulated (Fig. 4e, Supplementary Fig. 3b). Overactivation of these pathways can produce neutrophil extracellular traps (NETs) and an inflammatory microenvironment that suppresses anti-tumor T cell function and may promote tumor metastasis and immune evasion.

qRT-PCR validation of 10 DEGs related to DC activation showed expression changes consistent with RNA-seq results: 8 genes (Cd5, Notch3, Dll4, Flt3, Map3k14, Ccl22, Il12b, NR1d2) were upregulated, and 3 genes (Serpinb2, Cfh, Csf3) were downregulated (Fig. 4h). Interaction network analysis highlighted crosstalk between up- and downregulated pathways (Fig. 4g, Supplementary Fig. 3c, d). These data suggest that, in addition to rhythmic variation in DC numbers, functional regulation of DCs contributes to circadian differences in PDT efficacy.

PDT induces ICD

Growing evidence indicates that PDT not only induces apoptosis via ROS production but also triggers ICD, eliciting immune responses through the release of DAMPs from dying tumor cells [33]. Hallmarks of ICD include surface exposure of calreticulin (CRT), cytoplasmic release of high mobility group box 1 (HMGB1), and extracellular release of adenosine triphosphate (ATP). These DAMPs emit “find me” and “eat me” signals that promote DC migration and maturation, leading to antigen presentation to T cells and eventual recruitment of cytotoxic T lymphocytes (CTLs) to tumors for immune-mediated killing [34].

We measured DAMPs after VPF treatment and found that even a low dose (1 µM) induced ICD: ATP and HMGB1 levels increased by 3.419- and 1.584-fold, respectively, compared to controls (Fig. 5a, b). Higher VPF concentrations enhanced ICD. Immunoimaging showed negligible CRT surface exposure in controls but prominent ectopic expression with increasing VPF concentration (Fig. 5c, d). ATP release was more pronounced than HMGB1 release, suggesting that ATP—a key “find me” signal—may be particularly effective in recruiting APCs like DCs. Because VPF accumulation remains stable across administration times (Supplementary Fig. 1b), PDT induces a consistent baseline of ICD and DAMP release regardless of the time of day. Therefore, the observed circadian variations in efficacy must be dictated by the temporal responsiveness of the host immune system to these signals.

Fig. 5.

PDT induces ICD and promotes the maturation of DCs. a, b Release of ATP (a) and HMGB1 (b) from B16F10 cells following PDT with VPF. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test. c, d Representative confocal images (c) and quantitative analysis of fluorescence intensity (d) showing calreticulin (CRT) exposure in B16F10 cells after PDT treatment, Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test. e, f Bone marrow-derived dendritic cells (BMDCs) show enhanced maturation upon co-culture with PDT-treated B16F10 cells, as indicated by upregulation of MHCII (e) and CD86 (f). LPS (100 ng/mL) served as a positive control, BMDCs followed by co-culture with PDT-treated dying B16F10 cells or untreated B16F10 cells at ratios of 8:1 or 2:1 (tumor cells: BMDCs), Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test. Data are presented as mean ± SEM

ICD stimulates DC maturation

To determine whether PDT-treated dying cells stimulate DC maturation, we co-cultured BMDCs with PDT-treated dying cells. After 24 h, flow cytometry analysis revealed significantly elevated levels of maturation markers (MHCII and CD86) in CD11c⁺ BMDCs co-cultured with PDT-treated cells compared to those cultured with untreated live cells (Fig. 5e, f). This effect was dose-dependent, indicating that ICD-positive cells promote BMDC maturation.

Injection timing affects the efficacy of PDT-based vaccination

To further evaluate the impact of circadian rhythms on immunotherapy, we investigated the effect of vaccination timing in a tumor vaccine model. Specifically, based on the premise that PDT induces ICD in B16F10 cells, we inoculated mice with PDT-treated dying B16F10 cells at ZT4, ZT10, ZT16, or ZT22. Seven days later, live B16F10 cells were injected subcutaneously on the contralateral side at the same ZT. This design isolated vaccination time as the sole variable, eliminating potential confounders from tumor cell injection timing (Supplementary Fig. 4a).

Tumor growth monitoring showed that vaccination at ZT10 significantly suppressed tumor volume compared to ZT22 (Fig. 6a-c). Flow cytometry analysis of tumors and spleens 13 days post-inoculation revealed the highest levels of DCs, CD4⁺, and CD8⁺ T cells in the ZT10 group (Fig. 6d-f). Parallel experiments in nude mice showed no time-dependent differences in vaccine efficacy (Fig. 6g, Supplementary Fig. 4b, c). Thus, vaccination timing critically influences the amplitude of immune activation and subsequent tumor suppression.

Fig. 6.

Vaccination timing determines the efficacy of a PDT-based tumor vaccine. a–c Tumor growth curves (a), final tumor volumes (b), and representative ex vivo tumor images (c) in mice vaccinated with PDT-treated B16F10 cells or PBS (control) at four different Zeitgeber times, followed by rechallenge with live B16F10 cells (n = 6). d, e Normalized counts of total infiltrating immune cells in spleens (d) and tumors (e) from the vaccine model, harvested at four time points (n = 6). f Flow cytometry analysis of CD11c⁺MHCII⁺ cells in spleens of nude mice following vaccination with PDT-treated cells, comparing Ctrl (ZT10), ZT10, and ZT22 groups. g Tumor volumes on day 4 and day 10 after rechallenge in nude mice vaccinated with PDT-treated B16F10 cells; data are plotted per mouse (n = 4). Data are presented as mean ± SEM. Tumor growth curves (a) were analyzed using two-way ANOVA. Other statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test

Discussion

Herein, we confirm that the timing of treatment influences both the therapeutic efficacy of PDT and its vaccine effect, which is primarily attributable to the circadian regulation of dendritic cell-mediated adaptive immunity. Specifically, the ICD induced by PDT activates systemic immune responses. However, DCs and CD4⁺/CD8⁺ T cells infiltrate the tumor and spleen in a circadian manner, leading to temporal variations in the intensity of adaptive immune responses elicited by PDT at different times of the day. Moreover, temporal variations in gene expression within DCs further modulate the adaptive immune processes. Our findings demonstrate that the circadian oscillations in DC infiltration and function are critical to the overall efficacy of PDT. This aligns with the emerging paradigm of “cancer chronotherapy,” which posits that the timing of treatment is as critical as the dose [35–37]. While circadian optimization has been extensively explored in chemotherapy and immune checkpoint blockade [38, 39], our work extends this concept to PDT, highlighting a previously underappreciated dimension of photodynamic intervention.

PDT elicits a pronounced immune response that significantly influences treatment outcomes. PDT generates singlet oxygen and reactive oxygen species, which induce oxidative stress, resulting in photo-oxidative damage, activation of cell signaling pathways, and induction of stress proteins. This oxidative damage exhibits potent pro-inflammatory properties, and the inflammatory processes triggered by PDT initiate and propagate antitumor immune responses [40–43]. However, existing studies have indicated that the homing and recruitment of immune cells, as well as their activity levels, are modulated by circadian rhythms. In terms of immune cell quantities, research has demonstrated that in rats, B cells, total T cells, and various T cell subsets(Tγδ, Th, and Ts/c)in the blood exhibit 24-hour rhythms, with peak levels occurring during the rest phase [44]. Similarly, studies in mice have shown that white blood cell and granulocyte counts in the blood are higher at ZT10:00 compared to ZT22:00 [45]. In our study, we found that the infiltration of immune cells, including DCs, CD4⁺ T cells, and CD8⁺ T cells, into tumors and spleens following PDT treatment also exhibits a circadian rhythm. This leads to variations in immune effects after PDT, consequently resulting in differences in PDT efficacy. Interestingly, our data show that the abundance of dendritic cells in the spleen and tumor tissues exhibits synchronous fluctuations. This synchronicity indicates a mechanism of systemic immune synchronization. We hypothesize that systematic rhythmic signals, such as fluctuations in glucocorticoids or adrenaline signals would simultaneously mobilize immune cells to be activated in multiple locations [22, 46–50]. Alternatively, the spleen may act as a dynamic supply station; at its peak, the high density of mature DCs in the spleen is likely to cause an “overflow” effect, resulting in more cells entering the bloodstream and rapidly entering the tumor microenvironment. This is consistent with recent findings that migration ability peaks in synchrony with cell maturation [17]. Given this synchronous pattern, it is crucial to distinguish the central system-driven factors (affecting the spleen and tumor separately) from the spleen-dependent supply chain. Although the current research focus is on establishing this rhythmic phenomenon to guide chronotherapy, we emphasize splenectomy as our priority for subsequent mechanism studies to precisely depict the source of tumor-infiltrating dendritic cells.

Furthermore, our study revealed that circadian rhythms also influence the function of DCs. Through transcriptomic sequencing, we observed variations in the expression of circadian clock genes within DCs. At ZT10, the expression of Nr1d2 was significantly upregulated. Among immune-related pathways, DEGs were primarily enriched in specific adaptive immune responses such as ‘Th1 and Th2 cell differentiation’ and ‘Leukocyte transendothelial migration’ et al. These changes enable DCs to efficiently recognize tumor antigens (e.g., via the ‘C-type lectin receptor signaling pathway’), promote the differentiation of T cells toward the anti-tumor Th1 phenotype, and enhance the recruitment of immune cells to tumor sites, thereby effectively facilitating tumor cell clearance [51–53]. In contrast, at ZT22, the DEGs were predominantly enriched in immune pathways such as the ‘Toll-like receptor signaling pathway’ and ‘Neutrophil extracellular trap formation’ et al. This response is not only suboptimal in efficacy but also results in the production of NETs and a pro-inflammatory microenvironment that suppresses critical anti-tumor T cell functions, potentially facilitating tumor metastasis and immune evasion, thereby compromising therapeutic outcomes [54–58]. Notably, among the immune pathway-related genes, CD5 exhibited the most significant upregulation at ZT10 compared to ZT22. Studies indicate that CD5⁺ DCs serve as critical regulators of anti-tumor immune responses, directly enhancing antigen presentation by DCs and T cell activation capacity, thereby promoting potent anti-tumor immunity [59]. This further elucidates the pivotal role of DCs in PDT-associated immune responses. The absence of DCs leads to the disappearance of circadian variations in immune outcomes, though the precise underlying mechanisms require further investigation.

While our results demonstrate that VPF-mediated PDT universally induces profound ICD and promotes DC maturation, this baseline capacity alone does not fully explain the observed circadian variations in therapeutic efficacy. Given the stable intratumoral accumulation of VPF across different administration times, the consistent release of DAMPs triggered by PDT must intersect with the rhythmic fluctuations of the host immune system. We propose that this chronotherapeutic advantage is primarily driven by the temporal alignment between the immediate, transient release of short-lived DAMPs and the natural circadian peak of tumor-infiltrating DCs. ATP, a potent but highly transient “find-me” signal for DCs, is particularly relevant in this context. When PDT is administered at the optimal time (e.g., ZT10), the immediate burst of ATP (or potentially other undiscovered short-lived molecules) precisely coincides with the peak abundance and optimal functional state of host DCs, creating a “golden window” that maximizes the anti-tumor immune response. This mechanism effectively links a constant exogenous ICD trigger with a rhythmic endogenous host response.

Moreover, this dynamic aligns with our previous findings that PDT initiates a spatiotemporal cascade, transitioning from extremely rapid reactions—such as immediate necrosis in areas of high photosensitizer concentration—to slower downstream biological events [60]. The rapid release of transient molecules during this initial phase likely acts as the primary driver for the circadian phenomena observed here. Although identifying and characterizing other potential short-lived molecules within the in vivo tumor microenvironment is crucial for a complete mechanistic understanding, real-time spatiotemporal tracking of these rapid events remains technically challenging.

Furthermore, the specific immune response elicited by PDT treatment is highly specific and durable [61–63]. Additionally, the timing of vaccination significantly influences the immune outcome. A systematic review and meta-analysis of 17 studies revealed that 11 studies demonstrated a statistically significant impact of vaccination timing [64–66]. Similarly, we confirmed that the prophylactic efficacy of PDT-generated vaccines is time-dependent. This suggests that the initial interaction between the immune system and tumor antigens sets the magnitude of the long-term memory response [67]. By simply selecting the optimal time window, the immunogenicity of PDT can be leveraged more effectively without increasing dosage or toxicity. However, the precise mechanisms underlying these phenomena require more detailed investigation.

Supplementary Information

12964_2026_2973_MOESM1_ESM.jpg ^{(841.1KB, jpg)}

Supplementary Material 1: Supplementary Fig. 1. Circadian rhythm influences the efficacy of PDT. a Schematic diagram of the PDT treatment protocol. b. Average fluorescence intensity in different groups. c. Tumor growth curves in mice implanted with B16-F10 cells at four different Zeitgeber times; data are plotted per mouse (n = 6). d. Tumor growth in 4T1 tumor-bearing mice treated with PDT three times at ZT10 or ZT22; data are plotted per mouse. e–g. Excised tumor weights from mice treated with PDT under normal light-dark conditions (e), constant darkness (f), or an inverted light cycle (g) (n = 6). Data are presented as mean ± SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test (b, e, g), Other statistical significance was determined by Student’s t-test.

12964_2026_2973_MOESM2_ESM.jpg ^{(1.3MB, jpg)}

Supplementary Material 2: Supplementary Fig. 2. Circadian immune rhythms determine the time-dependent efficacy of PDT. a, b Normalized counts of total infiltrating immune cells in spleens (a) and 4T1 tumors (b) harvested at four Zeitgeber times (n = 6). c, d. Normalized counts of total infiltrating immune cells in spleens (c) and B16F10 tumors (d) from mice kept under constant darkness and sampled at two time points (n = 6). e, f. Normalized counts of CD45⁺ cells (e) and CD11c⁺MHCII⁺ cells (f) in spleens and B16F10 tumors from mice under an inverted light cycle, sampled at two time points (n = 6). g. Representative flow cytometry plots demonstrating nearly complete depletion of CD11c⁺MHCII⁺ cells in the spleen and tumor of DTR mice following diphtheria toxin administration. Data are presented as mean ± SEM. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post-hoc test (a, b), Other statistical significance was determined by Student’s t-test.

12964_2026_2973_MOESM3_ESM.jpg ^{(2.3MB, jpg)}

Supplementary Material 3: Supplementary Fig. 3. Dendritic cells exhibit circadian patterns of gene expression. a Transcriptional expression levels of three circadian rhythm-related genes in individual CD11c⁺MHCII⁺ cell samples. b. Heatmap showing four immune-related pathways significantly enriched among down-regulated genes. c. Network diagram of KEGG pathways enriched among up-regulated DEGs. d. Top 10 significantly enriched GO terms and associated genes identified through GO enrichment analysis of up-regulated DEGs.

12964_2026_2973_MOESM4_ESM.jpg ^{(1,017.1KB, jpg)}

Supplementary Material 4: Supplementary Fig. 4. Vaccination timing influences the efficacy of a PDT-based vaccine. a Schematic diagram of the PDT-based vaccination and tumor rechallenge experimental timeline. b, c. Final tumor volumes (b) and representative excised tumors (c) in nude mice vaccinated with PDT-treated B16F10 cells and subsequently rechallenged with live tumor cells (n = 4). Data are presented as mean ± SEM. Statistical significance was determined by Student’s t-test.

12964_2026_2973_MOESM5_ESM.xlsx ^{(11KB, xlsx)}

Supplementary Material 5: Supplementary Table 1. Primer Sequences for qRT-PCR.

Authors’ contributions

Rui Zhang: Validation, Data curation, Conceptualization, Writing - Original Draft; Sihan Dong: Validation, Data curation; Yinghao Zheng: Validation, Data curation; Yuting Fu: Validation, Data curation; Tianzhen Sun: Conceptualization; Junyu Zhou: Data curation. Yuemeng Liu: Data curation; Guoquan Liu: Writing - Review & Editing. Xunbin Wei: Resources, Project administration, Funding acquisition.

Funding

This work was supported by the National Key Research and Development Program of China (Grant No. 2021YFF0502900) and the Special Fund for Research on National Major Research Instruments of China (Grant No. 62027824).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal experimental procedures were approved by the Institutional Animal Care and Use Committee of Peking University Health Science Center. All experiments were performed in accordance with relevant guidelines and regulations.

Competing interests

The authors declare no competing interests.