Nanoconjugates to enhance PDT-mediated cancer immunotherapy by targeting the indoleamine-2,3-dioxygenase pathway

Background Photodynamic therapy (PDT) may elicit antitumor immune response in addition to killing cancer cells. However, PDT as a monotherapy often fails to induce a strong immunity. Immune checkpoint inhibitors, which selectively block regulatory axes, may be used in combination with PDT to improve treatment outcomes. Indoleamine 2,3-dioxygenase (IDO) is an immunoregulatory enzyme and an important meditator of tumor immune escape. Combination therapy with PDT and IDO-targeted immune checkpoint blockage is promising but has been seldom been explored. Methods Herein we report a composite nanoparticle that allows for simultaneous delivery of photosensitizer and IDO inhibitor. Briefly, we separately load ZnF16Pc, a photosensitizer, and NLG919, an indoleamine 2,3-dioxygenase (IDO) inhibitor, into ferritin and poly(lactide-co-glycolic)-block-poly(ethylene glycol) (PEG-PLGA) nanoparticles; we then conjugate these two compartments to form a composite nanoparticle referred to as PPF NPs. We tested combination treatment with PPF NPs first in vitro and then in vivo in B16F10-tumor bearing C57/BL6 mice. Results Our results showed that PPF NPs can efficiently encapsulate both ZnF16Pc and NLG919. In vivo studies found that the combination treatment led to significantly improved tumor suppression and animal survival. Moreover, the treatment increased tumor infiltration of CD8+ T cells, while reducing frequencies of MDSCs and Tregs. 30% of the animals showed complete tumor eradication, and they successfully rejected a second tumor inoculation. Overall, our studies introduce a unique composite nanoplatform that allows for co-delivery of photosensitizer and IDO inhibitor with minimal inter-species interference, which is ideal for combination therapy. Supplementary Information The online version contains supplementary material available at 10.1186/s12951-021-00919-z.


Background
Photodynamic therapy or PDT has been extensively tested for cancer treatment. During PDT, a large amount of reactive oxygen species such as singlet oxygen are produced, which cause physical damage to cancer cells or tumor vasculature [1]. PDT also induces immunogenic cell death (ICD), eliciting an antitumor immunity that benefits long-term tumor management [2]. However, PDT as a monotherapy is associated with a relatively high recurrence rate. While the reasons are multi-fold, one factor is that PDT may cause excessive inflammation that induces immunosuppressive mechanisms [2][3][4], leading to cancer cell immune escape and repopulation. Immune checkpoint inhibitors, which can selectively block related regulatory axes, may be used in combination with PDT to improve therapy outcomes [5][6][7]. For instance, several groups, including us, have demonstrated that PDT in conjugation with anti-PD-1/PD-L1/CTLA4 antibodies augments immune response [8,9], facilitating suppression of both primary and distant tumors [10]. Indoleamine 2,3-dioxygenase (IDO) is another immune checkpoint frequently implicated in tumor immunosuppression [2]. IDO is a monomeric, heme-containing enzyme that metabolizes tryptophan (Trp) to kynurenine (Kyn). IDO is frequently upregulated in both cancer cells and host antigen-presenting cells (APCs) [11]. IDO overexpression leads to the depletion of Trp that is essential to the survival and functions of effector T cells, causing their G1 cycle arrest and apoptosis [11,12]. Meanwhile, Kyn, the metabolite, can promote the differentiation and activation of regulatory T cells (Tregs) [13], facilitating the recruitment of myeloid-derived suppressor cells (MDSCs), culminating in a tolerogenic tumor microenvironment (TME) [11,13,14]. Multiple molecule IDO inhibitors have been developed and tested in pre-clinical and clinical studies [15]. PDT may induce the secretion of interferon γ (IFN-γ) and tumor necrosis factor α (TNF-α), leading to increased infiltration of Tregs and expansion of MDSCs with elevated IDO expression. Combining with IDO inhibition may promote PDTinduced antitumor immunity. The concept, however, has only recently been tested [2].
Herein, we report a composite, core/satellite nanoparticle that allows for co-delivery of photosensitizer and IDO inhibitor for combination PDT and immune checkpoint blockade (ICB) therapy. Photosensitizers and IDO inhibitors belong to two categories of therapeutics that have different physiochemical and pharmacokinetic properties. While it is possible to load both types of therapeutics onto one nanoplatform, the drugs may interfere with each other and negatively affect loading and release. To solve the issue, we construct a composite nanostructure where zinc hexadecafluoro-phthalocyanine (ZnF 16 Pc) and NLG919 are encapsulated into separate nanocompartments. More specifically, ZnF 16 Pc, a near-infrared photosensitizer (ex: ~ 670 nm) is encapsulated into ferritin (FRT) protein cage, whereas NLG919, a potent enzymatic inhibitor of IDO, is encapsulated into poly(lactide-coglycolic)-block-poly(ethylene glycol) (PEG-PLGA) nanoparticles. The two particles are covalently conjugated to form a composite nanoparticle (Fig. 1a) [15]. While ZnF 16 Pc is minimally released from ferritin thus enabling steadfast production of reactive oxygen species (ROS) under photo-irradiation [16,17], NLG919 is released from PEG-PLGA nanoparticles in a controlled fashion to allow for sustained inhibition of IDO post PDT. We administered the resulting, PEG-PLGA and FRT composite nanoparticles (hereafter referred to as PPF NPs), into B16F10-tumor-bearing mice to examine their tumor suppression efficacy and antitumor immunity.
To construct PPF core/satellite nanoparticles, the surface carboxyl groups of NLG919@PLGA NPs were activated by EDC/NHS chemistry; the resulting nanoparticles were mixed with ZnF 16 Pc@FRT to form a covalent linkage. The ratio between the two nanocompartments were adjusted so that in the final conjugate (i.e. PPF NPs), the NLG919 to ZnF 16 Pc ratio was ~ 7:1; this ratio was chosen based on the effective doses of the two therapeutics according to previous studies [17,20]. Transmission electron microscopy (TEM) confirmed the coupling (Additional file 1: Figure S3). Dynamic light scattering (DLS) found an increased hydrodynamic diameter from 112.9 ± 0.7 to 144.0 ± 1.3 nm (Fig. 1b). This was accompanied by a surface charge increase from − 33.30 ± 1.03 to − 20.23 ± 1.72 mV (Fig. 1c), also confirming the conjugation.
We also studied drug release using UV-vis spectrophotometry. The release of NLG919 from PPF NPs was relatively slow at pH 7.4 (77.5% at 24 h, Fig. 2a), and accelerated under acidic conditions (98.2% at 24 h, Fig. 2a). Meanwhile, ZnF 16 Pc release was minimal at both neutral and acidic pH (Fig. 2b), which is consistent with our previous observations [19]. Notably, photo-irradiation did not significantly accelerate the release of either drug molecules (Fig. 2a, b).
We then evaluated the cytotoxicity of PPF NPs in B16F10 cells, which are a murine melanoma cell line. Briefly, B16F10 cells were incubated with PPF NPs (25 µg/mL, ZnF 16 Pc concentration) in the dark and irradiated at 4 h (671 nm, 0.1 W/cm 2 , 200 s). PPF NPs in the absence of radiation showed low toxicity (Additional file 1: Figure S4). Under radiation, there was a significant viability drop, which is attributed to PDT. We also evaluated the IDO inhibition efficacy of PPF NPs. Because B16F16 cells have a relatively low IDO expression, this was examined in HeLa cells, which express a high level of IDO and are commonly used for in vitro IDO inhibition assessment [20][21][22]. Briefly, HeLa cells were treated with IFN-γ to induce IDO expression, followed by incubation with PPF NPs for 48 h; the amounts of Kyn in the culturing medium were quantified using a colorimetric assay (Fig. 2c) [20]. PPF NPs showed efficient and concentration-dependent inhibition of Kyn production. The EC 50 is 1.25 µg/mL, which is comparable to free NLG919 (Fig. 2c).
We next evaluated the therapeutic effects of PPF NPs in vivo in B16F10 tumor bearing C57BL/6 mice. We intratumorally (i.t.) injected PPF NPs (2.5 mg/kg NLG919, equivalent to 0.375 mg/kg ZnF 16 Pc) into the animals (n = 10). This was followed by photo-irradiation with a 671-nm laser (0.3 W/cm 2 , 15 min) applied to tumors at 4 h (Day 0, the treatment group was denoted as PPF + IR). Two more treatments were applied on Day 3 and 6. For comparison, NLG919@PLGA NPs in the absence of irradiation (NLG919@PLGA, n = 10) and ZnF 16 Pc@FRT plus irradiation (ZnF 16 Pc@FRT + IR, n = 10) were also tested. PBS only in the absence of irradiation was tested as a control (PBS, n = 8).
ZnF 16 Pc@FRT + IR caused a modest tumor suppression, inhibiting tumor growth by 51.27% relative to the PBS control group on Day 12 (Fig. 3a). However, this was followed by a rapid tumor rebound (Fig. 3a, c), with 40% of the animals reaching a humane endpoint point by Day Fig. 1 Preparation of characterizations of PPF NPs. a Schematic illustration showing the mechanism behind PPF NPs enabled immunotherapy. The PPF conjugate contains two nanocompartments, which are FRT and PLGA NPs. These two compartments are separately encapsulated with ZnF 16 Pc and NLG919, and covalently linked. Under photo-irradiation, ZnF 16 Pc is activated, leading to cancer cell death and release of tumor associated antigens. Meanwhile, NLG919 causes the suppression of IDO and restored the Trp/Kyn balance; this results in reduced tumor infiltration of Tregs and MDSCs and increased cytotoxic T cell activity, leading to enhanced antitumor immunity. b Hydrodynamic sizes of FRT, NLG919@PLGA and PPF NPs, measured by DLS. c Zeta potential of FRT, NLG919@PLGA and PPF NPs 15. NLG919@PLGA failed to suppress tumor growth but rather promoted it in the short-term, although the difference was insignificant between the NLG919@PLGA and PBS groups (p = 0.81, Fig. 3a, c). All animals in the NLG919@PLGA group either died or had to be euthanized by Day 12. As a comparison, PPF + IR resulted in remarkable tumor suppression (84.19% on Day 12, Fig. 3a, c). 30% of the mice in the PPF + IR group showed complete tumor eradication and remained alive after 6 months. For dead or euthanized animals, we harvested their tumors and major organ tissues for histology analysis. H&E and Ki-67 staining confirmed efficient cancer cell killing and tumor growth suppression in the PPF + IR group (Fig. 3d).
For animals showing complete tumor elimination in the PPF + IR group, a re-challenge study was performed to assess whether the treatment promoted anticancer immunity. Briefly, we injected live B16F10 cells into the opposite flank of these animals on Day 69 and monitored the tumor growth. All of the mice successfully rejected the second inoculation, and remained healthy through the remainder of the experiment (Additional file 1: Figures S5, S6). As a comparison, the only surviving animal from the ZnF 16 Pc@FRT + IR group succumbed to the second inoculation on Day 84. We euthanized animals from the PPF + IR group on Day 159, and analyzed memory T cells by flow cytometry. We found that both central (CD3 + CD8 + CD62L + CD44 + ) and effector memory T cell (CD3 + CD8 + CD62L − CD44 + ) populations were significantly increased relative to naïve mice (Additional file 1: Figure S7a, b). These results suggested that animals treated with PPF + IR developed a strong antitumor immunity.
To better understand treatment-induced immune responses, in a separate study, we treated animals with the same regimens (PPF + IR, PP, ZnF 16 Pc@FRT + IR, and PBS, single dose) and euthanized them 1 day after the treatment. We collected blood samples and analyzed serum concentrations of Trp and Kyn by LC-MS. We found that Trp/Kyn ratio was slightly reduced in the ZnF 16 Pc@FRT + IR group relative to the PBS control (Fig. 4a, p = 0.28), which is attributable to PDTinduced IDO upregulation. In both NLG919@PLGA and PPF + IR groups, Trp/Kyn ratios were significantly increased (Fig. 4a), which is owing to NLG919-based IDO inhibition.
We then repeated the therapy study but euthanized animals 7 days after the third treatment and collected tumor and spleen tissues to examine lymphocyte profiles. Compared to the PBS and ZnF 16 Pc@FRT groups, there was a significant decrease of MDSCs (Gr-1 + CD11b + ) population in tumors treated with NLG919@PLGA (Fig. 4b). PPF + IR led to an even greater reduction in the MDSCs population. More remarkable reduction was observed with Tregs (CD3 + CD4 + FoxP3 + ), whose abundance was decreased by 74.03 and 68.15%, respectively, in the NLG919@PLGA and ZnF 16 Pc@FRT + IR groups, and by 95. 97% in the PPF + IR group (Fig. 4b). Similar trends were observed in spleen tissue samples (Fig. 4c). Meanwhile, PPF + IR led to a significant increase in the dendritic cell (CD11c + ) and cytotoxic T cell (CD3 + CD8 + ) populations in both tumor and spleen samples (Fig. 4b,  c). CD8/Treg ratio was dramatically increased from 54.52% in the ZnF 16 Pc@FRT + IR group to 848.43% in the PPF + IR group in tumor (Fig. 4d). Overall, these studies confirm that combination PDT and ICB with PPF NPs plus photo-irradiation induced a strong anticancer immunity.
Meanwhile, the treatment was well tolerated by the animals. There were no signs of actuate toxicity after PDT. No body weight loss was observed throughout the studies (Fig. 5a). We also harvested tissues from normal organs, such as the heart, kidney, liver, lung, and spleen, and H&E staining found no signs of toxicity in the PPF + IR group (Fig. 5b).

Conclusions
PDT can cause ICD, promoting recruitment and activation of innate immune cells such as dendritic cells [10]. This in turn may enhance antigen cross-presentation to T cells, stimulating specific antitumor immune responses. However, tumors may hijack immunoregulatory mechanisms to subvert PDT-induced immunity. Taking the current study for instance, ZnF 16 Pc@FRT + IR only caused a modest benefit in tumor suppression and animal survival, and failed to induce an abscopal effect. As a comparison, PPF NPs enable combination PDT and ICB therapy that significantly improves tumor management, resulting in complete tumor eradication in 30% of the treated animals and protecting them from a subsequent live cell challenge. Our unique composite nanoparticle is considered an important factor for effective treatment, for it enables efficient PDT and controlled release of IDO inhibitors. In current studies, nanoparticles were topically applied, which is suitable for melanoma therapy. It is postulated that this strategy can be used to treat stage III and recurrent melanoma, eliciting an antitumor immunity that not only eradicates local or residual cancer cells after surgery but also prevents metastasis. Notably, melanoma is relatively resistant to PDT due to the absorbance and anti-oxidant effects of melanin [23]. It is possible that switching to a photosensitizer with even longer excitation wavelength may improve the treatment. PPF nanoparticles may be systemically administered for treatment of other cancer types such as prostate and head and neck cancer. For these applications, the pharmacokinetics of the nanoparticles need to be re-assessed. PPF treatment may also synergize with anti-PD-1/PD-L1 antibodies that block other immunoregulatory pathways. These possibilities will be explored in future studies.

PLGA-b-PEG-COOH synthesis
PLGA-PEG-COOH polymer was synthesized using EDC/ NHS chemistry. Briefly, 500 mg PLGA-COOH (0.071 mmol) and 109 mg EDC (0.7 mmol) were dissolved in 7 mL DCM, under magnetic stirring at room temperature for 10 min. Then, 160 mg NHS (1.4 mmol) was dissolved in 7 mL DCM, followed by dropwise addition into the above solution. The reaction was conducted overnight at room temperature. PLGA-NHS was washed with cold 1:1 methanol/diethyl ether mixture for three times, collected by centrifugation, and lyophilized. Then, 200 mg PLGA-NHS (0.028 mmol) was dissolved in 5 mL DCM under magnetic stirring at room temperature for 10 min. 30 mg NH 2 -PEG-COOH (0.03 mmol) and 60 µL DIEA (See figure on next page.) Fig. 3 In vivo treatment efficacy, tested in C57BL/6 mice bearing B16F10 tumors. The animals were i.t. injected with PPF NPs and irradiated with a 671-nm laser at 4 h (PPF + IR). A total of three doses of treatment were given 2 days apart. PBS, NLG919@PLGA NPs in the absence of irradiation (NLG919@PLGA), and ZnF 16 Pc@FRT plus irradiation (ZnF 16 Pc@FRT + IR) were tested for comparison (n = 8-10). a Relative tumor volume change. Values are presented as means ± SEM. (0.28 mmol) were dissolved in 2 mL DCM, and dropwise added into the PLGA-NHS solution. The reaction was conducted overnight. The conjugated PLGA-b-PEG-COOH polymer was washed with cold, 1:1 v/v methanol/ diethyl ether mixture for three times, collected by centrifugation, and lyophilized. The product was stored at -20 °C. Chemical structure was confirmed by 1 H-NMR on a Varian Mercury Plus 400 system.

NLG919@PLGA nanoparticle synthesis
NLG919@PLGA NPs were synthesized through a nanoprecipitation method. Briefly, PLGA-b-PEG-COOH and NLG919 were first dissolved in acetonitrile; the final concentrations were 5 mg/mL for PLGA-b-PEG-COOH and 1.5 mg/mL for NLG919. The mixture was dropwise added into sterilized nanopure water with constant stirring for 2 h in fume hood. The resulting NLG919@PLGA NPs were collected on an amicon ultracentrifugation unit (MWCO 100 kDa) and washed 3-4 times with sterilized nanopure water. Finally, the purified NLG919@PLGA NPs were resuspended in sterilized nanopure water and kept at 4 °C.

Preparation of ZnF 16 Pc@FRT
Protocols for ferritin (FRT) expression and purification were published before [17]. ZnF 16 Pc loading was achieved by following a published protocol with modification. Briefly, FRT was dispersed in PBS (pH 7.4), and the solution pH was reduced to 2.0 by adding 0.2 M HCl. ZnF 16 Pc (5 mg/mL in DMSO) was added into the above solution to reach a final FRT/ZnF 16 Pc w/w ratio of 5:1.
After gently shaking at room temperature for 30 min, the pH of the mixture was slowly adjusted back to 7.4 by adding 1 M NaOH. NAP-10 column was used to remove unloaded ZnF 16 Pc.

Preparation of PPF NPs
Briefly, NLG919@PLGA NPs (10 mg/mL) were resuspended in 3 mL sterilized nanopure water with constant stirring. Then, 720 µL EDC (10 mg/mL) was dropwise added, followed by the addition of 440 µL NHS ( [20]. Polymeric nanoparticles are reported to accumulate in tumors at a rate of 3-5%. Based on our computation, a ZnF 16 Pc-to-NLG919 ratio of 1:7 is reasonable for combination therapy.

Drug release
For drug release studies, 100 µL nanoparticle solution was loaded onto a dialysis unit, with or without irradiation (671 nm, 0.

Cell culturing
B16F10 and HeLa cells were obtained from American Type Culture Collection (ATCC) and cultured in DMEM or EMEM medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. The cells were incubated humidly at 37 °C with 5% CO 2 .

Cytotoxicity assays
B16F10 cells were seeded onto a 96-well plate at 8000 cells per well density and allowed to grow overnight. Nanoparticles of varied concentrations were added to the incubation medium. After 4 h, the cells were exposed to a 671-nm laser (0.1 W/cm 2 for 200 s). Cells not exposed to photo-irradiation were tested for comparison. After another 12 h, cell viability was measured by MTT assay.

IDO activity
Briefly, HeLa cells were seeded onto a 96-well plate at a density of 8000 cells per well and allowed to grow overnight. Recombinant human IFN-γ was added to each well to reach a final concentration of 50 ng/mL. Meantime, a gradient of NLG919@PLGA NPs or PPF NPs (NLG919 concentration 0, 0.4, 0.8, 1.5, 3, 6 µg/mL) were added to the incubation medium. After 48 h, 150 µL of the supernatants per well was transferred to a separate 96-well plate. 75 µL 30% trichloroacetic acid was added into each well and incubated with cells at 50 °C for 30 min, converting N-formylkynurenine to kynurenine. The supernatant was transferred to a new 96-well plate, mixed with equal volume of Ehrlich reagent (2% p-dimethylamino-benzaldehyde w/v in glacial acetic acid), and incubated for 10 min at room temperature. Reaction product was quantified by measuring 490-nm absorbance using a plate reader.

Animal model
C57BL/6 mice were purchased from Envigo laboratories. The animal model was established by subcutaneously injecting 2 × 105 B16F10 cells into the right hind limb of each mouse. All of the experimental procedures were conducted following a protocol approved by the University of Georgia Institutional Animal Care and Use Committee.

Serum tryptophan and kynurenine analysis
Trp/Kyn ratios in serum samples were examined by LC/MS on a Bruker Daltonics Impact II system. C57BL/6 mice bearing B16F10 tumor at ~ 50 mm 3 were treated with single-dose of PBS, NLG919@PLGA NPs, ZnF 16 Pc@FRT + IR, or PPF + IR (NLG919 2.5 mg/kg, ZnF 16 Pc 0.375 mg/kg). One day after the treatment, the plasma samples were collected and mixed with methanol (plasma:methanol, 1:2.5, v/v) and centrifuged at 13,500 rpm for 15 min. Supernatants were collected and purified with TopTip for LC-MS quantification of kynurenine and tryptophan.

Lymphocyte profiling studies
Seven days after the third treatment, animals were euthanized; spleen and tumors were harvested, and immune cell populations in the tissues was examined by flow cytometry. Briefly, single-cell suspensions from tissue samples were filtered and red blood cells were lysed. For extracellular staining, cells were incubated with the indicated combinations of antibodies (e.g. anti-CD11c, -Gr-1, -CD3, -CD8α, -CD4, and -CD45). For intracellular staining, cells were fixed and permeabilized immediately after cell surface staining according to manufacturer's protocol (eBioscience). Antibodies (e.g. anti-Foxp3) were added into the permeabilization buffer and incubated with cells. Animals surviving from re-challenge studies were euthanized after 30 days, and spleen tissues were harvested for analysis of memory T cell populations. Immune cell populations were examined on a Beckman Coulter Cyto-FLEX system. Data analyzed was performed using FlowJo software.

Statistical analysis
Comparison of multiple assays was performed using a one-way ANOVA test. Comparisons of only two groups was performed using a paired t-test. Significance was set at p < 0.05. All experiments were performed with at least three replicates unless specified otherwise. All the data is represented as mean ± S.D.