Preparation and characterization of LDH@ZnPc
LDH@ZnPc was synthesized through isolation of ZnPc by layered double hydroxides (LDH). Firstly, LDH nanoparticles with Mg2+ and Al3+ as metal cations, and chloride as interlayer anions were constructed by hydrothermal-assisted coprecipitation method. ZnPc was then incorporated onto the LDH with ZnPc loading of 3.17% ± 0.36% detected by UV − Vis spectrophotometer. The absorption spectrum of LDH@ZnPc was similar with that of ZnPc, indicating the integrity of ZnPc after incorporation by nonabsorbent LDH (Fig. 1F). Additionally, elemental mapping of LDH@ZnPc reveals that in addition to Mg, Al and Cl in LDH, N from ZnPc is also measured and uniformly distributed throughout the nanosheets of LDH, which well demonstrated the isolation of ZnPc using LDH (Fig. 1E). Transmission electron microscope (TEM) and scanning electron microscope (SEM) images showed that LDH possessed a uniform orthohexagonal morphology with particle size ranging from 80 to 150 nm, and LDH@ZnPc maintained the same morphology as LDH (Fig. 1A, B). Moreover, the equivalent hydrodynamic diameters of LDH and LDH@ZnPc in PBS were measured to be 120 ± 10.8 nm (polydispersity index, PDI = 0.13 ± 0.04) and 130 ± 18.2 nm (PDI = 0.19 ± 0.06) respectively (Fig. 1C), which are consistent with their corresponding images of TEM and SEM. The naked LDH displayed a highly positive zeta potential of + 44.1 ± 1.2 mV, while LDH@ZnPc still exhibited a relatively high value at + 23.9 ± 0.7 mV (Fig. 1D), which is sufficient to target mitochondria. The XRD patterns (Fig. 1G) of LDH@ZnPc showed the typical peaks (003), (006), (009), (018), and (110) of LDH, indicating that the incorporation of ZnPc did not affect the crystal structure of LDH. Besides, LDH@ZnPc elemental speciation was examined by XPS. There were six characteristic peaks in the XPS spectra of LDH@ZnPc (Fig. 1H), of which four of them were located at 1303.64, 74.29, 284.56, and 197.80 eV, which corresponded to Mg 1s, Al 2p, Cl s, and Cl 2p, respectively. We also observed two peaks at 1021.33 and 399.2 eV that corresponded with Zn 2p and N 1 s from ZnPc, providing important evidence that we successfully synthesized LDH@ZnPc.
The stability of LDH@ZnPc in storage condition and simulated physiological conditions was assessed by measuring the changes of absorbance. Figure 1I illustrates that LDH@ZnPc displayed negligible variation within 48 h of incubation in PBS or FBS, and that it remained substantially stable in water at 4 °C for 7 days. Moreover, a release pattern of the entrapped ZnPc from LDH@ZnPc formulation was observed in Fig. 1J. Only 20% of ZnPc was released from LDH@ZnPc after 48 h at pH value of 7.4. However, under pH of 6.5 and 5.5, LDH@ZnPc showed sustained drug release within 48 h. To further evaluate the blood safety of LDH and LDH@ZnPc, the hemolytic assay was conducted according to our previously reported method. When hemolysis occurs, the hemoglobin of RBCs will be released to make the solution turn red. As displayed in Fig. 1K, the control group treated with H2O exhibited robust red solution with no RBCs precipitated in the bottom, whereas the PBS-treated group as well as various concentration-treated groups of LDH and LDH@ZnPc revealed nearly red-free solution. Even the concentration of LDH and LDH@ZnPc was up to 1600 µg mL−1, no obvious hemolysis was observed. Meanwhile, the corresponding hemolysis rates were provided in Fig. 1L. Therefore, these results demonstrated LDH and LDH@ZnPc possess excellent in vitro blood safety.
Cellular uptake and intracellular localization
We first investigated cellular uptake of LDH@ZnPc in TC-1 cells at various time by confocal laser scanning microscopic (CLSM). As shown in Fig. 2A and Additional file 1: Figure S1, LDH@ZnPc was rapidly internalized by TC-1 cells in 1 h and exhibited a much higher cellular uptake than ZnPc. In addition, fluorescence signals were barely observed in TC-1 cells incubated with ZnPc within 6 h. Overall, the fluorescence intensity exhibited a time-dependent manner within investigated times, and LDH@ZnPc displayed desirable intensity at 24 h, which is obviously stronger than that of ZnPc. These results indicated the efficient uptake of cationic LDH@ZnPc in TC-1 cells, which is consistent with our previous reports of cellular uptake using positively charged LDH [33].
The mitochondria-anchored ability of photosensitizers is a critical parameter evaluating the therapeutic efficacy of PDT [27, 50, 56]. Therefore, the subcellular organelle localization behavior of LDH@ZnPc was then investigated by CLSM and further quantified using Image J. As exhibited by the fluorescence in Fig. 2B, LDH@ZnPc was quickly taken up and trapped within 1 h in lysosomes. Notably, the fluorescence of LDH@ZnPc in lysosomes almost disappeared at 24 h (Fig. 2B), whereas the fluorescence in mitochondria enhanced significantly (Fig. 2C). Besides, the colocalization Pearson’s coefficients (R) between ZnPc and mitochondria was only 0.36, while the value between LDH@ZnPc and mitochondria was up to 0.91 (Fig. 2C). Furthermore, the colocalization behaviors of LDH@ZnPc with golgi apparatus and endoplasmic reticulum were also investigated. As reflected in Additional file 1: Figure S2, LDH@ZnPc overlapped partially with Golgi-Tracker and ER-Tracker with Pearson’s coefficients of 0.63 and 0.61, respectively, indicating that LDH@ZnPc did not have obvious preference on golgi apparatus and endoplasmic reticulum. Collectively, these results demonstrated the excellent mitochondria-anchored ability of LDH@ZnPc.
ROS generation and depolarization of mitochondrial membrane potential
The produced ROS in PDT play a significant role in killing tumor cells, and therefore the intracellular ROS generation is closely associated with the therapeutic efficacy. To examine whether isolation of ZnPc with LDH can facilitate ROS generation, we then evaluated the intracellular ROS levels with various doses of LDH@ZnPc in TC-1 cells. 2', 7'-dichlorodihydrofluorescein diacetate (H2DCFDA) was used as ROS indicator, which can be deacetylated by endocellular esterases and further oxidized by generated ROS within cells. The oxidized product is retained intracellularly and thus can be monitored through detecting its green fluorescence (525 nm). As displayed in Fig. 2D, E, cells incubated with 10 ng mL−1 of ZnPc upon irradiation exhibited a ROS-positive cell population of merely 1.77%, which was similar with that of the control. In comparison, cells demonstrated significantly increased ROS levels upon irradiation with a LED light (670 nm, 3 J cm−2) in the presence of 10 ng mL−1, 40 ng mL−1, and 160 ng mL−1 of LDH@ZnPc, which demonstrated cell population percentages of 10.0%, 16.8%, and 33.5%, respectively. It is noteworthy that the prominent results were achieved under such an ultralow light dose of 3 J cm−2 in PDT, which well testified the potent phototoxicity of LDH@ZnPc. Besides, these results also indicated that LDH@ZnPc-mediated ROS generation displayed a dose-dependent pattern in TC-1 cells within the investigated concentration ranges.
The membrane potential of mitochondria is a critical indicator reflecting its normal function, and the potential can be disrupted when ROS-mediated PDT occurred in mitochondria. To investigate the change of mitochondria membrane potential, the assay was thus performed after PDT using JC-1 dye as the indicator. In energized mitochondria, JC-1 can accumulate in the matrix of mitochondria and form a polymer emitting strong red fluorescence (~ 590 nm). However, red JC-1 polymer transforms to green JC-1 monomer (~ 529 nm) when mitochondrial depolarization occurs. As shown in Fig. 2F and Additional file 1: Figure S3, the normal cells displayed strong red fluorescence, whereas LDH@ZnPc incubated cells showed intense green fluorescence upon irradiation with nearly no observed red light, indicating that membrane potential of mitochondria was severely destructed by LDH@ZnPc-mediated PDT.
Cytotoxicity evaluation in vitro
Construction of supramolecular photosensitizer with high phototoxicity and simultaneously low dark toxicity is of great importance during PDT. Therefore, we further examined the cell toxicity of LDH@ZnPc. TC-1 cells were incubated with different dilutions of ZnPc and LDH@ZnPc followed by irradiation, and cell viability was evaluated by cell counting kit-8 (CCK-8) assay and flow cytometry assay, respectively. Meanwhile, the cellular dark toxicity was determined as well using parallel studies with no irradiation. As shown in Fig. 2H, after 24 h incubation of ZnPc, the viability of TC-1 cells remained normal, suggesting that only incubation with ZnPc was not sufficient to kill tumor cells after irradiation from LED light (670 nm, 3 J cm−2). In contrast, the phototoxicity of LDH@ZnPc was significantly enhanced after light irradiation, resulting in a concentration-dependent manner toward cell killing, which also supported mitochondria localization for enhanced therapeutic efficacy. In particular, the TC-1 cells were almost completely eradicated when the dose of LDH@ZnPc rose to 80 ng mL−1. On the contrary, negligible change in TC-1 cell viability was observed without LED light irradiation, which indicated that both LDH@ZnPc and ZnPc displayed negligible dark toxicity (Fig. 2I). Furthermore, the phototoxicity of LDH@ZnPc at various irradiation times was also performed in TC-1 cells, which exhibited a time-dependent manner within investigated time (Fig. 2G). To further testify the obtained results, flow cytometry assay was performed and confirmed the potent phototoxicity of LDH@ZnPc than ZnPc alone or a mixture of LDH and ZnPc (Additional file 1: Figure S4). Besides, we also investigated the phototoxicity behavior of our photosensitizer in HeLa cells, which displayed similar trends to that observed in TC-1 cells, with LDH@ZnPc showed the most potent efficacy (Additional file 1: Figure S5). Taken together, these results demonstrated that LDH@ZnPc presented high phototoxicity and low dark cytotoxicity within examined cells.
PDT-induced pyroptosis in vitro
Photosensitizers with mitochondria-anchored ability can induce photodamage on mitochondria and the subsequently mitochondrial oxidative stress, disrupting its own structure and leading to changes in cell morphology [51]. We have demonstrated the severe destruction of mitochondrial membrane potential induced by LDH@ZnPc-mediated PDT. Herein, to provide more insight of mitochondria targeting photodynamic cytocidal action, the changes of TC-1 cell morphology were recorded by CLSM. After being cultured with photosensitizers and exposed to light irradiation (670 nm, 3 J cm−2) or in the absence of light, the TC-1 cells were cultured for another 2 h before the cell morphology was observed.
As reflected in Fig. 3A, cells treated with PBS and LDH upon irradiation showed a relatively intact cell membrane without obvious morphological changes. Upon treatment with ZnPc or a mixture of LDH and ZnPc under the same irradiation, only a minor quantity of cells swelled with multiple small bubble-like protrusions (as indicated by yellow arrows) around cell membranes. In contrast, for the group incubated with LDH@ZnPc, an increased number of cells swelled and blew out large or small bubbles (as indicated by yellow arrows) from the plasma membrane. Furthermore, in the absence of light, the cell morphology of each group remained unchanged (Fig. 3D). Interestingly, this classic bubble-like cell morphology is considered as the typical characteristic of pyroptosis, a newly disclosed manner of programmed cell death. Therefore, the cell morphological change of group LDH@ZnPc may indicate the occurrence of pyroptosis.
According to recent research, the intracellular ROS was identified to active NOD-like receptor family pyrin domain protein 3 inflammasome (NLRP3), which subsequently recruit the adaptor protein apoptosis associated speck-like protein containing a caspase recruitment domain (ASC) to further recruit proCaspase-1, ultimately forming of inflammasome complex [17]. Caspase-1 is then activated within the inflammasome, and active Caspase-1 processes gasdermin D (GSDMD) to form its N-terminal domain (GSDMD-N), which is known as the canonical pyroptosis pathway. Therefore, to further ascertain the occurrence of pyroptosis and its mechanism after LDH@ZnPc-mediated PDT, we performed quantitative real-time PCR (qPCR) and western blot assay. In addition to Asc, the mRNA expression of Gsdmd, Caspase-1, and Nlrp3 was not significantly changed in the absence of light (Fig. 3C), while under light conditions (Fig. 3B), LDH@ZnPc-treated groups showed significantly higher expressions of Gsdmd, Caspases-1, Nlrp3 and Asc compared to control groups. Additionally, we noticed that the ZnPc and LDH + ZnPc reduced Asc expression at the mRNA level more significantly than LDH@ZnPc without light, which may be possible reason for the failure of ZnPc and LDH + ZnPc to increase Asc obviously when irradiated. More importantly, western blot analysis indicated that GSDMD-N and Cleaved-Caspase-1 levels were significantly increased in the LDH@ZnPc-treated group upon irradiation (Fig. 3E), providing the strongest evidence for the occurrence of canonical GSDMD-dependent pyroptosis.
Evaluation of ICD efficacy
Previous studies have confirmed the occurrence of ICD can release DAMPs, such as calreticulin (CRT) and high-mobility group box 1 (HMGB1), which then evoke antitumor immune response systematically. CRT can act as an “eat me” signal when translocated to the cell surface, which further leads to the uptake of dying tumor cells and their debris by antigen presenting cells (APCs). HMGB1 as a chromatin-binding protein can be released from the nucleus in the advanced stages during cell death. In general, CRT exposure on the cancer cell surface and HMGB1 release from the nucleus were regarded as emblematic hallmarks indicating the generation of ICD. Pyroptosis was recently demonstrated as a novel type of ICD, which can induce antigens release and subsequently robust antigen specific immune responses. Moreover, the photosensitizer with excellent mitochondria-anchored ability evokes prominent mitochondria oxidative stress, and consequently induces ICD. Therefore, we investigated whether our photosensitizer could massively evoke ICD generation using CLSM. As shown in Fig. 3F, the LDH@ZnPc treated group displayed apparent surface CRT exposure, whereas the other groups failed to exhibit such a phenomenon. Meanwhile, the ability of HMGB1 release was also investigated, which indicated that the LDH@ZnPc treated group revealed significant decrease of fluorescence in the nucleus, but no obvious changes were observed for the other groups (Fig. 3G). Taken together, these results verified the robust ICD generation triggered by pyroptosis and mitochondria anchoring in TC-1 cells after LDH@ZnPc-mediated PDT.
In vivo fluorescence imaging and antitumor activity
To investigate the retention and metabolism of LDH@ZnPc in tumor, the fluorescence imaging was conducted before in vivo antitumor experiments. As reflected in Fig. 4A, after intratumorally injecting the LDH@ZnPc, fluorescence intensity gradually increased and reached a maximum around 24 h in living mice. In contrast to LDH@ZnPc, ZnPc reached a maximum within 6–12 h. Additionally, LDH@ZnPc is more permeable to tumor tissue and can infiltrate the entire tumor tissue. Even 48 h after injection, there was an obvious fluorescence signal on the tumor tissue of LDH@ZnPc-treated group but not that of ZnPc-treated group, which indicated an ameliorative tumor retention ability when ZnPc was incorporated onto LDH.
An excellent photosensitizer should possess potent in vivo antitumor ability, we thus investigated the antitumor efficacy of LDH@ZnPc on C57BL/6 mice inoculated with TC-1 tumors. PBS- and ZnPc-treated groups with light irradiation were used as positive control, meanwhile groups of that without light irradiation were provided as negative reference. The light-treated groups were given a light dose of 120 J cm−2 from a 660 nm laser. To demonstrate the potent efficacy of LDH@ZnPc-mediated PDT, only single-time of irradiation was performed in irradiation-treated groups (Fig. 4B). As illustrated in Fig. 4D, E, LDH@ZnPc-treated group exhibited significant retardation of established TC-1 tumors in 3 of 5 mice (60%) and no trends of bounce was observed within investigated period. The weights of tumors were provided accordingly in Fig. 4F. ZnPc-treated group displayed certain inhibition against tumor growth, however, it was so poor that the tumors still rebounded on 9th day. As for the other groups, they displayed a similar trend of soared tumor growth without obvious inhibitory effects. Furthermore, data of body weight changes suggested that all the groups exhibited certain increase (Fig. 4C), indicating the low toxicity of LDH@ZnPc.
As depicted in Fig. 4G, hematoxylin and eosin (H&E) staining demonstrated that LDH@ZnPc-treated group exhibited severe destruction of tumor tissues with obvious spaces in comparison with other groups, which further verified its robust antitumor effect. Meanwhile, TUNEL staining was also performed with potent fluorescent intensity for LDH@ZnPc-treated group and certain intensity for ZnPc-treated group, which correlated well with their in vivo PDT antitumor effect. In addition, indexes of liver and kidney function revealed the security of various treatments (Additional file 1: Table S1). Besides, H&E staining of major organs (Additional file 1: Figure S6) confirmed the low toxicity of LDH@ZnPc. Collectively, these results manifested the robust in vivo antitumor efficacy and security of LDH@ZnPc-mediated PDT.
Primary and distant tumor inhibition
Encouraged by these obtained results, we then investigated whether LDH@ZnPc-mediated PDT could work effectively towards primary and distant tumors. This assay was performed using a bilateral model of TC-1 tumors on C57BL/6 mice. The tumors on the right side (primary) were treated with LDH@ZnPc and irradiated, while the left side tumors (distant) were untreated (Fig. 5A). To well illustrate the immunity efficacy induced by PDT, group underwent combined therapy of LDH@ZnPc-mediated PDT and αPD-1 (PDT & αPD-1) was also investigated as reference. Besides, the mice treated with PBS or αPD-1 served as controls. The therapeutic efficacies of various treated regimens were evaluated by the growth volumes of both primary and distant tumors.
As revealed in Fig. 5C–F, both PDT and combined therapeutic groups significantly inhibited the tumor growth, which exhibited tumor growth inhibitory of 40% and 60%, respectively. The αPD-1 group displayed certain inhibition against tumor growth, however, it was so poor that the tumor growth rates still exhibited a similar trend as the PBS group. For the effect towards untreated distant tumors, groups of PBS or αPD-1 failed to show inhibitory effect with tumor growth rates of similar trend. To our delight, group of single LDH@ZnPc-mediated PDT displayed effective inhibition towards distant tumor (p < 0.01 compared with the PBS group). Furthermore, PDT & αPD-1 group displayed the most potent abscopal antitumor effect, which can well demonstrate that LDH@ZnPc has a significantly synergistic effect compared with group of single αPD-1. Besides, the body weights were steady with slight increment, indicating no systemic toxicity (Fig. 5B). Overall, the in vivo antitumor results manifested that LDH@ZnPc-mediated PDT can not only kill primary tumors, but also exhibited apparent inhibition toward abscopal tumors.
PDT-induced pyroptosis in primary tumor
To further investigate the death patterns of primary tumor and immune state of distant tumor, the bilateral tumor tissues were conducted for transcriptomic analysis after treatment with PBS, αPD-1, PDT and PDT & αPD-1. In all, 1835 genes of primary tumor and 2007 genes of distant tumor were analyzed, and the genes expression was displayed by a hierarchically clustered heatmap (Additional file 1: Figure S7A, S8A). In addition, the distributions of differentially expressed genes (DEGs, fold change > 1 and p < 0.05) are illustrated via a standard volcano plot. Besides, the DEGs in the PBS vs PDT-treated groups were also performed using Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis as well as gene ontology (GO) enrichment analysis.
For primary tumor, a total of 897 DEGs were screened. As shown in Fig. 6A, the volcano plot showed that 730 genes were up-regulated and 167 genes were down-regulated after treatment with PDT. As identified by GO analysis of these DEGs (Additional file 1: Figure S7B), KEGG pathway proteins were enriched (Fig. 6B), including cytokine-cytokine receptor interaction, cell adhesion molecules, T cell receptor signaling pathway, natural killer cell mediated cytotoxicity, and NOD-like receptor signaling pathway (Additional file 1: Figure S7C). Recent studies have manifested that intracellularly generated ROS can trigger the pyroptosis, which involves the activation of multiple inflammasome signaling pathways including NOD-like receptor family pyrin domain protein 3 inflammasome (NLRP3) [17]. In our study, we have ascertained the occurrence of GSDMD-dependent pyroptosis after LDH@ZnPc-mediated PDT in vitro (Fig. 3A–E). To further confirm the occurrence of pyroptosis after LDH@ZnPc-mediated PDT in vivo, DEGs related to pyroptosis were screened on KEGG enrichment analyses. The results indicated that PDT increased levels of DEGs encoding proteins, such as gasdermin family proteins (GSDMD), cysteine aspartic acid-specific protease (Caspase-1), inflammasome (NLRP3), interleukin-1β (IL-1β), and interleukin-18 (IL-18), all of which are associated with pyroptosis (Fig. 6C, D). Furthermore, we examined the expression levels of these DEGs via qPCR, and the results were coincided with that obtained by transcriptomic analysis (Fig. 6E–J). Accordingly, the results of western assay also demonstrated the enhanced expression levels of GSDMD-N, Cleaved-Caspase-1, NLRP3, and IL-1β (Fig. 6K). Taken together, these data as well as cellular results jointly indicated that LDH@ZnPc-mediated PDT activated the canonical inflammasome pathway of GSDMD-dependent pyroptosis.
Immune response in vivo
To further investigate the immune state of distant tumor, DEGs of 483 up-regulated and 734 down-regulated were screened (Additional file 1: Figure S8B) with a result of significant difference in gene expression between PBS and PDT groups. Subsequently, the DEGs were analyzed by GO enrichment and KEGG pathway enrichment analyses (Fig. 7A, Additional file 1: Figure S8C). In particular, the differentially expressed genes related to “immune response” and “adaptive antitumor immunity” were screened to evaluate the antitumor immune effect of LDH@ZnPc-mediated PDT (Fig. 7B, Additional file 1: Figure S8D). Compared with the PBS group, αPD-1 failed to obviously up-regulate these immune-related genes, suggesting an occurrence of poor immune response. On the contrary, the distant tumor in PDT group displayed an increased up-regulation of these immune-related genes, indicating the effective immune activation triggered by LDH@ZnPc-mediated PDT.
To well examine the immune response in vivo, we further investigated the immune-related cell populations in distant tumors as well as spleens at day 15 after various treatments. TC-1 tumor-bearing mice were euthanized to obtain the tumor and spleen tissues, which were then performed for flow cytometry analysis. As illustrated in Fig. 7C, the percentages of CD4 + and CD8 + T lymphocytes in the distant tumors were significantly increased in PDT-treated groups compared with other groups. Concretely, single LDH@ZnPc-mediated PDT and PDT&αPD-1 group exhibited cell percentages of 6.40% ± 0.32% and 7.87% ± 0.43% respectively in terms of CD4 + T cells, which is ~ 2 − 3 folds compared with those of PBS and αPD-1 groups. As for CD8 + T cells, the data of PDT-treated groups still demonstrated similar differentiation in comparison with those of other groups (Fig. 7D). Meanwhile, the levels of CD4 + and CD8 + T cells in distant tumor tissues were further investigated using immunofluorescence staining (Fig. 7E), revealing that the fluorescence intensities in groups of PDT treatment were obviously stronger than the other groups, which is correlated well with the results of flow cytometry.
Moreover, the levels of immune cells in spleens were further investigated as well. As DCs was vital in initiating adaptive anti-tumor immunity, we therefore assessed the maturation of DCs in spleen. As displayed in Fig. 7F, G, the percentage of CD80 + CD86 + DCs in PDT and PDT&αPD-1 groups were 28.00% ± 2.14% and 29.93% ± 1.71% respectively, which is significantly increased compared with that of PBS (19.38% ± 2.30%) and αPD-1 group (20.03% ± 0.92%), suggesting the enhanced level of DCs maturation after LDH@ZnPc-mediated PDT treatment. As depicted in Fig. 7H, I, PDT-treated groups all demonstrated increased levels of CD4 + and CD8 + T cells compared with PBS and αPD-1 groups in spleen. In both distant tumor and spleen analysis, all PDT&αPD-1 groups obviously enhanced the blockade efficacy of αPD-1 therapy, which can be ascribed to the reason that LDH@ZnPc-mediated PDT elicited efficient ICD and further enhanced the efficacy of immune checkpoint blockade therapy.
Taken together, these results demonstrated all PDT-treated groups triggered potent immunological response. More importantly, LDH@ZnPc-mediated PDT significantly enhanced the efficacy of anti-PD-1 therapy, which consequently may act as a promising immune adjuvant to boost antitumor immunity.