Formulating a TMEM176B blocker in chitosan nanoparticles uncouples its paradoxical roles in innate and adaptive antitumoral immunity

Abstract

The immunoregulatory cation channel TMEM176B plays a dual role in tumor immunity. On the one hand, TMEM176B promotes antigen cross-presentation to CD8⁺ T cells by regulating phagosomal pH in dendritic cells (DCs). On the other hand, it inhibits NLRP3 inflammasome activation through ionic mechanisms in DCs, monocytes and macrophages. We speculated that formulating BayK8644 in PEGylated chitosan nanoparticles (NP-PEG-BayK8644) should slowly release the compound and by that mean avoid cross-presentation inhibition (which happens with a fast 30 minutes kinetics) while still triggering inflammasome activation.

Chitosan nanocarriers were successfully obtained, exhibiting a particle size within the range of 200 nm; they had a high positive surface charge and a 99 % encapsulation efficiency. In in vitro studies, NP-PEG-BayK8644 did not inhibit antigen cross-presentation by DCs, unlike the free compound. The NP-PEG-BayK8644 activated the inflammasome in a Tmem176b-dependent manner in DCs. We administered either empty (eNP-PEG) or NP-PEG-BayK8644 to mice with established tumors. NP-PEG-BayK8644 significantly controlled tumor growth and improved mice survival compared to both eNP-PEG and free BayK8644 in melanoma and lymphoma models. This effect was associated with enhanced inflammasome activation by DCs in the tumor-draining lymph node and infiltration of the tumor by CD8⁺ T cells. Thus, encapsulation of BayK8644 in chitosan NPs improves the anti-tumoral properties of the compound by avoiding inhibition of antigen cross-presentation.

GRAPHICAL ABSTRACT

INTRODUCTION

Immunotherapeutic approaches have revolutionized cancer therapy [1]. However, only a small proportion of patients with advanced disease experience clinical benefits [2–4]. Thus, it is crucial to identify novel immunoregulatory molecules to trigger effective anti-tumoral immune responses. In contrast to immune checkpoints whose physiological role is to block effector immune responses, other immunoregulatory molecules such as IL-2 and interferons play dual functions, promoting but also inhibiting immunity [5,6]. This complexity implies a huge challenge to therapeutically manipulate those pathways.

TMEM16B has been proposed as an emergent player in immunoregulation [7–9]. It is an intracellular acid-sensitive non-specific cation channel which belongs to the MS4A family [10–12. It is largely expressed in dendritic cells (DCs) from mice, rats, and humans [10,12–16], as well as other leukocyte subsets. High TMEM176B expression in the tumor has been associated with reduced overall survival in patients with colon [17], glioblastoma [18] and gastric cancer [19], making it a potential pharmacological target in oncology. Accordingly, deletion of TMEM176B was associated with strong anti-tumoral immune responses in mouse cancer models, dependent on host inflammasome activation and CD8⁺ T cells [17]. Moreover, pharmacological inhibition of TMEM176B using BayK8644 has shown to control tumor growth in murine cancer models in preventive but not in therapeutic protocols [17]. We speculated that the limited therapeutic efficacy of TMEM176B blockers might be related to its potential capacity of inhibiting antigen cross-presentation. In fact, CD8⁺ T cells recognize tumoral antigens in the context of MHC I molecules through the cross-presentation pathway [20, 21], and it was shown that TMEM176B promotes antigen cross-presentation by controlling phagosomal acidification in DCs [10].

In cross-presentation, endo-phagosomal antigen processing occurs through a fast kinetics, with a proposed time limit of approximately 25 minutes [22]. Thus, exogenous antigens that are not processed before that time, would not be presented on MHC class I molecules [22]. It is therefore anticipated that developing a formulation of the TMEM176B blocker, BayK8644, that allows slow release in endosomes could prevent the inhibition of cross-presentation while still promoting inflammasome activation.

The working hypothesis is that nanoencapsulation of BayK8644 in chitosan nanocarriers could prove to be an effective strategy to achieve site-specific drug delivery and sustained release. Chitosan (CS) is an excellent candidate for developing these nanovehicles due to its natural origin and properties, such as low toxicity, low immunogenicity, biocompatibility, biodegradability, pH-dependent behaviour, and low production cost. Chitosan-based nanoformulations can improve the dispersion of loaded hydrophobic biomolecules in aqueous environments, increasing their uptake by target cells and thus potentially improving their therapeutic efficacy via various administration routes. CS nanoparticles have been already reported in a wide range of applications, including controlled drug delivery systems [23–27] and cancer immunotherapy [28–31].

Moreover, previous work has shown that CS nanocapsules allow high encapsulation efficiency of a highly hydrophobic drug due to the presence of the chitosan-coated oily core in which the drug is contained [32].

Here we proved that formulating BayK8644 in slow-releasing chitosan nanoparticles (NPs) triggered inflammasome activation while preventing the inhibition of antigen cross-presentation, and enhances the compound’s anti-tumoral efficacy.

2. MATERIALS AND METHODS

2.1. Materials

Ultrapure chitosan hydrochloride salt (Protasan UP CL 113, with a deacetylation degree (DD %) of 85.3 % by ¹H NMR, (supplementary figure S2) was purchased from Novamatrix. Its viscosity-average molecular weight was determined by intrinsic viscosity measurements in 0.3 M acetic acid/0.2 M sodium acetate at 25 °C. The value of molecular weight was calculated using the Mark–Houwink–Sakurada equation (K= 0.074 ml/g and a= 0.76) and was equal to 33 KDa (See supplementary material, figure S1). BayK8644 of 99% of purity was purchased from Sundia. Crodamol GTCC, a mixed ester consisting of caprylic and capric acids derived from coconut or palm oil, was purchased from Croda. Biotin-PEG-COOH (Mw 3073Da) and MeO-PEG-COOH (Mw 5056Da) were obtained from Iris Biotech. Poloxamer 188, N-hidroxysuccinimide (NHS), N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) and Avidin from egg white (≥ 10 units/mg protein) were purchased from Sigma Aldrich. Phospholipon^® 90 NG (purified phosphatidylcholine from soybean lecithin) was purchased from Phospholipid GmbH. All other reagents used were of analytical grade.

2.2. Synthesis of CS-PEG copolymer

100 mg of chitosan-HCl (Protasan UP CL113), 27.6 mg of MeO-PEG-COOH, 2.4 mg Biotin-PEG-CO₂H and 3.4 mg of NHS were dissolved in 14 mL of water. Then, 69 mg of EDC was added in four equal portions every 30 minutes. The resulting solution was stirred for 24 h at room temperature, and then was ultrafiltered (Amicon 10 KDa) with water and lyophilized to give chitosan-PEG as a white foam.

2.3. Preparation of CS nanoparticles

Synthesis of blank chitosan nanoparticles (Nps) were prepared by using the solvent displacement technique, described by Torrecilla et al. [32] based on the solvent displacement method, with some modifications. Briefly, 40 mg of lecithin were dissolved in 0.5 mL of ethanol before adding 125 μL of Crodamol and 9.5 mL of acetone. This organic phase was directly poured over 20 mL of aqueous phase containing 10 mg of CS and 50 mg of Poloxamer 188. The mixture turned milky immediately as a result of the formation of NPs. Finally, the solvents were evaporated under vacuum with stirring at 150 rpm until reaching a final volume of 10 mL. BayK8644 and fluorescent probe (6 coumarin) loaded NPs formulations were prepared as described above by adding 0.5 mL of ethanolic BayK8644 solution (20 mg/mL) and 1 mL of ethanolic Coumarin solution (1 mg/mL) to the organic phase respectively. The final BayK8644 concentration obtained in NPs was 1 mg/mL.

2.4. Particle size and zeta potential

Hydrodynamic diameter and zeta potential of samples were obtained using a Nano-ZetaSizer (Malvern Instruments Ltd., UK). Analyses were performed in aqueous medium at 25 °C, viscosity of 0.8872 cP and RI of 1.33. Particle size was read in terms of z-average and zeta potential was calculated from measured electrophoretic mobility using the Helmoltz-Smoluchowski equation. Samples were analyzed in triplicate.

For short-term stability testing, particle sizes of NP-PEG-BayK8644 formulation were determined after 6 months of storage at 4 °C.

2.5. Surface morphology

The morphology of the nanocarriers was examined by transmission electron microscopy (TEM), JEOL JEM-1010.Samples were stained with 2 % (w/v) phosphotungstic acid solution and placed on copper grids (200-mesh) for viewing by TEM.

2.6. Raman spectroscopy

Formation of NPs was analyzed by Confocal Raman spectroscopy carried out using a WITec Alpha 300-RA 202 Raman-Confocal Microscope (WITec GmbH, Ulm, Germany). All samples (Lecithin, CS, eNP-PEG, NP-PEG-BayK8644 and BayK8644) were placed on a silicon wafer utilized as a substrate and were air-dried for 24 h.

2.7. Encapsulation Efficiency

The encapsulation efficiency of BayK in the NPs was indirectly determined by the difference between the total amount of BayK in the formulation and the free drug found in the supernatant of the formulation. Accordingly, the total amount of drug was estimated by dissolving an aliquot of no isolated NP-PEG-BayK8644 with acetonitrile and measured with a high performance liquid chromatography (HPLC) system. The no encapsulated drug was determined by the same method following separation of the CS structures from the aqueous medium by ultrafiltration (Amicon 10KDa). The loading capacity was determined by the amount of BayK8644 loaded per unit weight of the nanoparticle. The HPLC system consisted of a Thermo Scientific Ultimate 3000 equipped with a UV detector set at 274 nm and a reverse phase Zorbax 300SB-C18 column (4.6 × 250 mm i.d., pore size 5 μm Agilent, U.S.A.). The mobile phase consisted of a 50/50 acetonitrile/buffer (0.02 % ammonium acetate solution at pH 5) mixture at a temperature of 30 °C. The standard calibration curves of BayK were linear (r² > 0.9998) in the range of concentrations between 0.1–1.4 μg/mL.

2.8. In vitro drug release study

NP-PEG-BayK8644 solution was dissolved in RPMI culture media at different pHs (pH 5.0, pH 6.0 and pH 7.4). The amount of free BayK8644 released was determined by HPLC at different time points.

2.9. Animals

Six-to-ten weeks old male or female C57BL/6 TMEM176B (WT) and Tmem176b^−/− mice were used (Jackson Lab; Bar Harbor, ME) and bred for up to 20 generations in the Laboratory Animals Biotechnology Unit (UBAL)of the Pasteur Institute in Montevideo in conditions free of specific pathogens (SPF). These were kept in a controlled environment, with a temperature between 19 and 21°C, and cycles of 14 hours of light and 10 hours of darkness. Mice received water and sterile ration administered under ad libitum conditions. Tmem176b^−/− mice were generated in the 129/SvJ strain and heterozygous mice were backcrossed for 10 generations onto the C57BL/6 background (Janvier, Saint Berthevin, France). ARRIVE guidelines were followed for experimental research using mice. All animal protocols were approved by the institutional animal care and use committee (CEUA) of Institut Pasteur Montevideo (020–22).

2.10. Cell line culture

The E.G7-OVA (expressing OVA antigen) cell line (ATCC^® CRL-2113^™, (Manassas, VA), thymic lymphoma of murine origin) was cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS), 10 mM HEPES, 1 mM sodium pyruvate, MEM-1% non-essential amino acids, 0.05 mM β-mercaptoethanol, penicillin/streptomycin 100 units/mL and 0.4 mg/mL geneticin. It was kept at 37 °C in the presence of 5 % CO₂.

2.11. Primary culture of dendritic cells derived from bone marrow (BMDCs)

Bone marrow-derived Dendritic Cells (BMDCs) were obtained as previously described in [33]. Briefly, bone marrow cells were cultured in RPMI supplemented medium (10 % fetal bovine serum, 10 mM HEPES, 1 mM sodium pyruvate, 1 % non-essential MEM-amino acids, 2 mM L-glutamine, 0.05 mM β-mercaptoethanol, penicillin/streptomycin 100 units/mL). BMDCs were differentiated by culturing bone marrow cells for 8 days in the presence of 0.4 ng/mL GM-CSF. During the in vitro differentiation of bone marrow cells into DCs, culture medium was replaced at days 3 and 6 of culture. At day 8, adherent cells were > 95% CD11c⁺CD11b⁺MHC II^int.

2.12. Nanoparticles uptake by Bone Marrow Dentritic Cells (BMDCs).

At the 8th day, BMDCs were detached with a buffer solution containing FBS and EDTA. For confocal microscopy internalization studies, 1.0 ×10⁵ cells were seeded in a 12-well plate, glass cover slips were previously placed in each well. Cells were allowed to adhere for 1.5 hours and stimulated for 30 minutes with NP-PEG-6cou at 37 °C to allow endocytosis process or 0 °C (as a negative control), washed and then chased for 15 minutes at the same conditions. All images were taken using a Zeiss LSM 800^™ semi-spectral confocal laser scanning microscope (Beckman Coulter). For flow cytometry internalization studies, the same protocol was followed as mentioned above. BMDCS were collected and samples were acquired using Cyan ADP Flow Cytometer equipped with a 488 nm laser and a band filter (530/40 nm). For data analysis FlowJo v.X software (BD, Ashland, OR, USA) was used.

2.13. Colocalization Studies by Confocal Microscopy

Tmem176b^+/+ (WT) and Tmem176b^−/− C57BL/6 BMDCs were plated in 12-well plate (1 ×10⁵ cells/well) containing glass cover slips and cultured for 1.5 hours. Then cell were treated with NP-PEG-6-cou during 30 minutes, washed and chased for 15 minutes. Cells were fixed in 2 % (w/v) paraformaldehyde (PFA) for 15 min, permeabilized and blocked using TritonX100 0.1 %, 3 % BSA in PBS. Primary antibody, rabbit anti-mTMEM176B Polyclonal antibody (Proteintech Group) 5 μg/mL, or a control IgG antibody, was diluted in BSA-PBS-Triton solution and incubated overnight at 4 °C. Three washing steps were performed using PBS solution. Biotin-donkey anti-Rabbit IgG (H+L) secondary antibody diluted 1:500 in BSA-PBS-Triton was added and incubated for 1,5 hours. After washing, cells were incubated with Streptavidin Alexa Fluor 568 conjugate (Molecular Probes) diluted 1:1000 in BSA-PBS-Triton. Finally, nuclei were stained with DAPI diluted 1:5000 (Life Technologies) and incubated for 5 minutes. Sections were mounted in ProLong antifade mounting medium (Molecular Probes). Fluorescence was observed by confocal microscopy Zeiss LSM 800^™ semi-spectral.

2.14. In Vitro Inflammasome Activation

IL-1β was quantified in BMDCs culture supernatant as a readout of the NLRP3 inflammasome activation. As mentioned above, on the 8th day, 5.0 ×10⁴ BMDCs were seeded in a 96-well plate and allowed to adhere for 1.5 hours. Then, cells were stimulated for 3 hs with 0.25 μg/mL LPS, washed and treated for 2 hours with the indicated doses of free BayK8644, NP-PEG-BayK8644 and eNP-PEG. The presence of IL-1β was determined using an ELISA kit following the manufacturer’s protocol. (Biolegend, 432603). To determine Caspase-1 activation by Western blot, culture supernatants from BMDCs stimulated in the absence of FBS were precipitated with 20% (v/v) TCA and washed with acetone. Cell lysates were generated with RIPA buffer in the presence of a protease inhibitor cocktail. Cell lysates and precipitates from culture supernatants were electrophoresed blotted and probed with anti-Caspase-1 (Adipogen, AG-20B-0042) antibodies. Quantification was carried out using ImageJv.1.5i software and for statistical analysis.

Caspase-1 activation was quantified in TMEM176B BMDCs culture as a readout of the NLRP3 inflammasome activation. As mentioned above, on the 8th day, BMDCs were seeded (5 ×10⁵cells) in 35 mm culture dishes and allowed to adhere for 1.5 hours. Cells were stimulated for 2 h with 20 μM of free BayK8644, NP-PEG-BayK8644 and eNP-PEG and then washed. BMDCs were stained with FLICA-1 (FLICA 660 Caspase-1 Assay, ImmunoChemistry) for 30 minutes, and analyzed by flow cytometry using Cyan ADP Flow Cytometer equipped with a 488 nm laser and a band filter (530/40 nm). For data analysis FlowJo v.X software (BD, Ashland, OR, USA) was used.

2.15. Antigen cross-presentation assays in mouse splenic dendritic cells.

To obtain splenic DCs, spleen of C57Bl6 mice was extracted, cut into small pieces and treated with collagenase-D (Roche, Merck). After 30 minutes of incubation at 37 °C, the enzymatic reaction was stopped with a 10 mM EDTA solution. Lysis of red blood cells was performed with an ammonium chloride solution. Finally, cells were washed with PBS/FBS1%. For the isolation of CD11c⁺ cells, murine anti-CD11c monoclonal antibodies coupled to magnetic microspheres supplied in the commercial kit “CD11c MicroBeads UltraPure, mouse” (Miltenyi Biotec Inc, Auburn, CA, USA) were used according to the protocol suggested by the manufacturer. The purity of the CD11c⁺ fraction was confirmed by flow cytometry.

CD11c⁺ spleen cells were incubated with 3 mg/mL of soluble OVA, or OVA plus 20 μM of free BayK8644, 20 μM of BayK8644 encapsulated into chitosan nanoparticles (NP-PEG-BayK8644) or empty counterpart (eNP-PEG) for 4 hours in supplemented-RPMI medium. Also, as a control, CD11c⁺ spleen cells were incubated with SIINFEKL peptide at indicated concentrations or SIINFEKL peptide plus BayK8644 for 4 hours in supplemented-RPMI medium. Cells were then washed with PBS/BSA 1 % and fixed with 0.008 % glutaraldehyde for 3 minutes on ice and quenched with 0.2 M glycine solution. After one final wash with PBS, CD11c⁺ spleen cells were incubated with the B3Z cell line (T cell hybridoma cell line) for 18 hours at 37 °C and 5% CO₂. B3Z cell line specific for SIINFEKL allowed us to study the T cells activation (Membrane Fusion Laboratory, IHEM -CONICET; Faculty of Medical Sciences, UNCuyo). The B3Z cell line is a β-galactosidase (lacZ) inducible T cell hybridoma. They are stably transfected with the Dectin-1-CD3ζ chimeric construct (Dectin-1 extracellular region fused to the CD3ζ cytoplasmic region). Furthermore, these cells are transiently transfected with an NFAT-LacZ reporter gene (activated T-cell nuclear factor NFAT, fused to the E. coli lacZ gene); since CD3ζ triggers an intracellular signal that leads to the activation of the transcriptional factor NFAT. Previous studies demonstrated that the heterologous lacZ gene, under the control of the complete enhancer region of IL-2 or NFAT alone, is specifically induced in transfected and activated T cells. Therefore, activation of transfected T cells results in the synthesis of the products of the IL-2 and lacZ genes. The activation of B3Z cell line was measured detecting β-galactosidase activity by optical density (absorbance at 595–655 nm) using chlorophenol red-β-D-galactopyranoside (CPRG; Roche) as substrate for the reaction.

2.16. Ex vivo isolation and flow cytometry analysis of CD8+ T cells

C57BL/6 mice were injected subcutaneously (s.c) with 0.5 × 10⁶ EG7 thymic lymphoma cells. When tumors were palpable (tumor size of 5 mm × 5 mm), 70 nmoles of BayK8644 encapsulated in NPs was intratumorally (i.t.) injected. NPs injections were repeated at the +2, +4 days after the first injection. After the third NP-PEG-BayK8644 injection, tumors were resected.

Animals treated with NP-PEG-BayK8644 were classified according to tumor growth kinetics. Those animals that control tumor growth over time (tumor size at the resection time were <100 mm²) were named Responders. Inversely, those animals did not control the tumor size were called Progressors. T cells were isolated by dissociating tumor tissue in the presence of collagenase D for 10 min at 37 °C. Reaction was stopped with EDTA 10 mM, and the suspension was filter through a 50 μm strainer. After red blood cell lysis, flow cytometry analysis was performed. For flow Cytometry analysis, single cell suspensions were stained with antibodies against surface molecules CD8α-APC-Cy7, TCRβ-PE, CD19-PercPCy5.5, OVA-pentamer-APC and TCRvβ12-FITC to discriminate EG.7-OVA tumor cells. All data were collected on a FACS Aria Fusion (BD Biosciences) and analyzed with FlowJo software (FlowJo v.X software (BD, Ashland, OR, USA).

2.17. Tumor Models and Treatments

Tmem176b^+/+ (WT) and Tmem176b^−/− C57BL/6 mice were injected s.c with 0.5 ×10⁶ E.G7-OVA thymic lymphoma cells or 0.5 ×10⁵ 5555 melanoma cells. At the indicated experiment, i.t injection was performed alternating one WT and one Tmem176b^−/− mouse until completing both groups. In treated animals, alternation was done between BayK8644- and vehicle-treated animals or NP-PEG-BayK8644 and eNP-PEG-treated animals. Tumor growth was caliper measured manually every 2–3 days, and the two major diameters were taken. 70 nmoles of free BayK8644 or BayK8644 encapsulated in chitosan nanoparticles, or the same amount of empty NPs injections were repeated at the +2, +4, +6, +8 days after the first injection. Mice were sacrificed when one of the major tumor diameters reached 2 cm.

2.18. Statistical analysis

Statistical analysis were performed by GraphPad Prism 7 (GraphPad Software, Inc., La Jolla, CA). Survival analysis were conducted with the Log-rank (Mantel-Cox) test. The comparison of the two experimental conditions was done with unpaired Student’s t test. Comparison among multiple conditions were performed using one-way ANOVA tests. Differences in gene expression between responder and progressors groups were assessed using the unpaired t-test when normality assumption was met.

3. RESULTS AND DISCUSSION

3.1. Encapsulation of BayK8644 in chitosan nanoparticles.

To begin with, it was first studied whether BayK8644 inhibits antigen cross-presentation in vitro or not. Enriched splenic CD11c⁺ DCs from WT mice were pulsed with OVA protein and co-cultured with the OVA-specific CD8⁺ B3Z T cell hybridoma in which the β-Gal enzyme is under the control of the IL-2 promoter [34]. OVA induced a dose-dependent activation of B3Z cells (Supplementary figure S3).

It could be observed that treating splenic DCs with BayK8644 significantly decreased their capacity to activate B3Z cells (Figure 1a). Conversely, BayK8644 did not inhibit activation of B3Z cells induced by the minimal OVA peptide SIINFEKL (Figure 1b), which does not require processing to be loaded onto MHC I molecules. These results suggest that Bayk8644 interferes with the processing of OVA protein through the cross-presentation pathway in splenic DCs as expected for a TMEM176B inhibitor [10]. This observation may account for the limited therapeutic efficacy of the compound [17]. Conversely, BayK8644 did not impair antigen presentation to CD4+ T cells (Supplementary figure S4).

Figure 1. Free BayK8644 inhibits antigen cross-presentation by DCs in vitro.

a) Antigen cross-presentation assays in mouse splenic dendritic cells. A pool of three independent experiments is shown. ***p < 0.001, ****p < 0.0001. One-way ANOVA test. b) Splenic DCs were loaded for 45 minutes with the OVA peptide SIINFEKL, which does not need further processing in intracellular compartments to be presented on H-2K^b MHC-I molecules ± 20μM BayK8644. DCs were then washed and fixed and co-cultured with B3Z T cells. A pool of three independent experiments is shown. ***p < 0.001. Unpaired Student’s t test.

It was then hypothesized that formulating BayK8644 through a slow releasing mechanism in the endosomal compartment may avoid inhibition of cross-presentation by the compound. Nanoparticles (NPs) are known as drug-encapsulation systems that can release compounds with slow kinetics [26, 35]. Chitosan NPs, which are semi-crystalline linear polysaccharides composed of N-acetyl D-glucosamine and D-glucosamine units linked by β(1–4) glucosidic bonds, were synthesized and coated with polyethylene glycol (PEG) to form a protective hydrophilic layer around the nanoparticles [32]. There were three types of NPs synthetized: empty PEGylated NPs (eNP-PEG), PEGylated NPs containing BayK8644 (+)-BayK8644 (NP-PEG-BayK8644), and PEGylated NPs containing the fluorescent probe 6-coumarin (NP-PEG-6cou). The average size, the polydispersity index (PI) and the Z potential of the NPs, as well as the efficiency of encapsulation of the formulation containing BayK8644, are detailed in Table 1. Particle size measurements using a Nano Zetasizer indicated an average size close to 200 nm, which is optimal for enhancing biological fluid circulation time and cellular internalization [36]. The reduction in nanoparticle size upon the addition of PEG can be attributed to the increased solubility provided by the amphiphilic copolymer, which reduces interfacial tension and results in smaller particles. The PEG coating incorporated to the nanoparticle surface is intended, not only to enhance the physicochemical properties of the nanocarriers but also to facilitate the future conjugation of specific antibodies, further enhancing the antitumor effect. The colloidal solutions remained stable for at least 6 months at 4°C, with the size remaining nearly constant (203.2 ± 2.95 nm, PDI: 0.147 ± 0.022). This stability may be attributed to the high positive surface charge (zeta potential close to 30 mV), which provides substantial stability to the nanosystems. The zeta potential, primarily influenced by the chitosan outer layer, is positive due to the amino groups in chitosan, especially at slightly acidic pH levels (below 6.5). The encapsulation efficiency of BayK8644 in all syntheses was consistently greater than 99%.

Table 1.

Physico-chemical properties of the nanosystems obtained (mean ± s.d.; n=3). Nps, eNP-PEG, NP-PEG-6cou do not contain BayK

Type of nanosystem	Size ± SD (nm)	PDI ± SD	Zeta potencial ± SD (mV)	BayK8644 encapsulation efficiency (%)	BayK8644 loading capacity (%)
Nps	180.7 ± 15.0	0.249 ± 0.075	27.8 ± 1.3	-	-
eNP-PEG	173.4 ± 4.1	0.218 ± 0.020	25.9 ± 5.7	-	-
NP-PEG-BayK8644	209.9 ± 11.9	0.198 ± 0.032	24.6 ± 2.2	99.8 ± 0.1	5.40 ± 0.8
NP-PEG-6cou	199.1 ± 2.65	0.228 ± 0.006	24.2 ± 1.04	-	-

The TEM images of eNP-PEG and NP-PEG-BayK8644 revealed their spherical shape and a homogeneous size distribution (Figure 2a and b). The sizes obtained by TEM are smaller than those measured by Nano Zetasizer, as shown in Table 1. This difference in size measurement arises because, in solution, the ions of opposite charge will be attracted to the surface of the nanoparticles generating ionic layers that move along with it. The Nano Zetasizer measures the hydrodynamic diameter of the nanoparticles in solution, including these ionic layers, while the TEM reports the actual diameter of the NPs without the ionic layer [37].

Figure 2. Encapsulation of BayK8644 in chitosan nanoparticles.

a and b) Transmission electron micrograph (TEM) images and their respective particle size histograms obtained from TEM images. c) The Raman spectrum of NP-PEG-BayK (in red), free BayK8644 (blue) and eNP-PEG (black). d) Cumulative drug release profiles of NP-PEG-BayK8644 at different pH conditions.

The confocal Raman microscopy study revealed significant differences in the spectra among pure BayK8644, eNPs-PEG and NPs-PEG-BayK8644. The spectrum obtained from NPs-PEG-BayK8644 differed evidently from what would be expected by simply combining the individual spectra (Figure 2c).

These differences were most pronounced in specific regions: between 1400 and 1480 cm⁻¹ (mainly from C-H bending vibrations), between 2700 and 3000 cm⁻¹ (predominantly C-H stretching vibrations), and at 1739 cm⁻¹ (indicative of C=O stretching in ester bonds) [38, 39]. Shifts and changes in intensity ratios were observed in these bands when comparing BayK NPs to empty nanoparticles. Furthermore, it can be observed a significant decrease or even absence of the 939 cm⁻¹ band in Np-PEG-BayK8644, which is present in eNP-PEG and corresponds to vibrational stretching modes of the rings [38]. These findings highlight possible chemical interactions between NPs constituents and the BayK8644 molecule and underscore the importance of these interactions in achieving optimal release of the BayK8644 molecule [38].

The in vitro release kinetics of BayK8644 from the nanoparticles exhibited a slow, pH-dependent release profile. As shown in Figure 2d, an acidic pH environment increased the increased the release of BayK8644 after 24 hours of incubation with the buffer.

3.2. NP-PEG-BayK8644 do not inhibit antigen cross-presentation by DCs but still trigger inflammasome activation.

To further investigate whether encapsulation of BayK8644 in chitosan NPs could enhance the compoundś anti-tumoral properties, it was studied whether and how NP-PEG-BayK8644 modulate DCs biology. Initially, NP internalization by DCs was assessed. Flow cytometry studies showed that bone marrow-derived DCs (BMDCs) efficiently internalized NP-PEG-6cou, as evidenced by an increase in both the percentage of 6-coumarin⁺ cells (Figure 3a, b) and the mean 6-coumarin fluorescence intensity (MFI) (Figure 3a, c) compared to untreated cells and BMDCs incubated at 0 °C (as a control to minimize active internalization). Confocal microscopy experiments showed that at 0 °C, NPs were mostly associated to the cell surface (arrows), while at 37 °C, they were fully internalized by BMDCs (arrowheads; Figure 3d).

Figure 3. Chitosan nanoparticles are internalized by DCs, do not inhibit antigen cross-presentation and trigger inflammasome activation.

a) BMDCs were treated in vitro with NP-PEG-6cou for 30 minutes, at 37 °C or 0 °C as a negative endocytosis control. A representative dot plot is shown with the percentage of 6-coumarin positive cells of BMDCs treated with NP-PEG-6cou at 37 °C (red dots), compared with BMDCs treated with NP-PEG-6cou at 0 °C (cyan dots). Grey dots represent untreated BMDCs. One experiment representative of three is shown. b) Percentage of 6-coumarin-positive cells. A pool of three independent experiments is shown. ** p < 0.01, One-way ANOVA test. c) Mean Fluorescence Intensity (MFI) of 6-coumarin dye determined by flow cytometry. A pool of three independent experiments is shown. *** p < 0.001; ****p < 0.0001. One-way ANOVA test. d) Confocal microscopy analysis of BMDCs treated with NP-PEG-6cou (green fluorescence) for 30 minutes at the indicated temperatures (0 °C and 37 °C). Nuclei were stained with DAPI (blue fluorescence). One experiment representative of three is shown. At least 50 cells were analyzed/experiment. e) Confocal microscopy study showing the detection of NP-PEG-6cou and the endogenous TMEM176B protein. BMDCs from WT mice were treated with NP-PEG-6cou for 30 minutes, then immunostaining was performed with anti-TMEM176B (red) antibody. Nuclei were stained with DAPI (blue). The graph shows the intensity profiles of anti-TMEM176B staining (red) and NP-PEG-6cou (green) obtained using ImageJ software inside the white square. The fluorescence intensities are plotted along the y-axis and the image position along the x-axis. One experiment representative of three is shown. At least 50 cells were analyzed/experiment. f) The graph shows the intensity profiles of anti-Tmem176b staining (red) and NP-PEG-6cou (green) obtained using ImageJ software inside the white square. The fluorescence intensities are plotted along the y-axis and the image position along the x-axis. One experiment representative of three is shown. At least 50 cells were analysed/experiment. g) Confocal microscopy analysis of BMDCs treated with NP-PEG-6cou or free 6-cou (green fluorescence) 15 minutes and then chased for 30 or 120 minutes. Nuclei were stained with DAPI (blue fluorescence). One experiment representative of two is shown. h) CD11c⁺ splenic DCs were left untreated or loaded with 3mg/mL soluble OVA for 4 hr ± 20 μM BayK8644, eNP-PEG or NP-PEG-BayK8644. B3Z T cell activation was assessed through a colorimetric assay. A pool of three independent experiments is shown. *p<0.05, ***p < 0.001, ****p < 0.0001. One-way ANOVA test. i) WT and TMEM176B^−/− BMDCs were treated with LPS (0.25 μg/mL) for 3 hr, washed and exposed for 2 hr to 50 μM BayK8644 encapsulated in chitosan Nps, or an equivalent amount of eNP-PEG. IL-1β was measured in the culture supernatant by ELISA. One experiment representative of three is shown. ****p < 0.0001. Two-way ANOVA test. j) WT BMDCs were treated with LPS (0.25 μg/mL) for 3 hr, washed and exposed for 2 hr to 50 μM BayK8644 encapsulated in NP-PEG-BayK8644, or an equivalent amount of eNP-PEG or 50 μM free BayK8644. Inflammasome activation was studied by assessing Caspase-1 activation by flow cytometry using the FLICA-1 reagent. Representative dot plots are shown for the FLICA quantification by flow cytometry. One experiment representative of five is shown. k) A pool of the five experiments commented in H is shown. *p<0.05, **p<0.01, ****p<0.0001. One-way ANOVA test.

Additionally, NP-PEG-6-cou were localized within Tmem176b⁺ compartments (Figure 3e), likely endosomes [10]. The intensity profiles of anti-TMEM176B staining (red) and NP-PEG-6cou (green), obtained using ImageJ software, are shown inside the white square in Figure 3f. This suggests that BayK8644 released by NPs within endosomes may be in close proximity to TMEM176B. It was then studied whether encapsulating a hydrophobic compound in NPs would delay its release to the cytosol after internalization by the DC. The fluorescent probe 6-coumarin, which mimics BayK8644 in molecular weight and hydrophobicity, was used in this study. At early time points after NP internalization (30 minutes), 6-coumarin showed a dot profile, compatible with endosomal localization (Figure 3g). However, at 120 minutes after NP internalization, 6-coumarin dot pattern was substituted by a diffuse profile, strongly suggesting cytosolic localization of the fluorescent probe. In contrast, free 6-coumarin rapidly entered the cytosolic compartment. These results demonstrate that small hydrophobic compounds encapsulated in NP are released through a slow kinetics within DCs. Thus, encapsulating BayK8644 in NP would block its ability to inhibit antigen cross-presentation (which happens during the first 30 minutes after antigen internalization) while still promoting inflammasome activation. It was further explored whether BayK8644 encapsulated in NPs could inhibit antigen cross-presentation by DCs. Splenic DCs were loaded with OVA protein ± BayK8644 or ± NP-PEG-BayK8644 and co-cultured with B3Z CD8⁺ T cells. It was observed that in contrast to free BayK8644, NP-PEG-BayK8644 failed to block activation of B3Z cells (Figure 3h). Moreover, treatment of DCs with NP-PEG-BayK8644 induced higher levels of IL-1β secretion compared to eNP-PEG in WT, but not in Tmem176b^−/− DCs (Figure 3i). NP-PEG-BayK8664 also proved were more effective than eNP-PEG and BayK8644 in inducing Caspase-1 activation in DCs (Figures 3j–k and Supplementary figure S5). Thus, NP-PEG-BayK8644 trigger inflammasome activation in DCs while preventing inhibition of antigen cross-presentation by BayK8644.

3.3. Therapeutic injection of BayK8644 encapsulated in chitosan NPs increases survival and retards tumor growth in a TMEM176B-dependent manner.

Having shown that NP-PEG-BayK8644 do not inhibit antigen cross-presentation (Figure 3g) while still triggering inflammasome activation (Figures 3h–k and Supplementary figure S5), it was then investigated whether those NPs may control tumor growth in vivo in a therapeutic setting. It was previously shown that BayK8644, when administered intraperitoneally before the tumors establishment (preventive protocol), it effectively controlled its growth. However, it failed to control tumor growth in animals with established tumors (therapeutic protocol) [17].

To evaluate the therapeutic efficacy of NP-PEG-BayK8644, in vivo studies were carried out with WT mice carrying established 5555 (melanoma) and EG7 (lymphoma) tumors (Figure 4a and b, and Supplementary figure S6). It was observed that intratumoral (i.t.) injection of NP-PEG-BayK8644 significantly improved mouse survival in melanoma and lymphoma (Figure 4a and b) models, compared to animals injected with eNPs-PEG. In contrast, no difference in survival was observed in Tmem176b^−/− mice treated with either NP-PEG-BayK8644 or eNP-PEG (Figure 4c), indicating that NP-PEG-BayK8644 control tumor growth through a Tmem176b-dependent mechanism. Additionally, i.t. injection of free BayK8644 also failed to control tumor growth and to improve mouse survival in comparison to vehicle-treated animals (no significant difference between free BayK8644 and vehicle) (Figure 4d). To further understand how NP-PEG-BayK8644 controls tumor growth, inflammasome activation in the tumor-draining lymph nodes (TDLN) of animals treated with NP-PEG-BayK8644 or eNP-PEG was assessed. Fluorescent NPs injected i.t. were found in the TDLN 24 hours post-administration, primarily in resident but not in migratory conventional (co)DCs (Supplementary Figure S7). We did not observe different percentages of active Caspase-1 (FLICA-1+) cells within resident coDCs in animals injected with eNP-PEG versus NP-PEG-BayK8644 (Figure 4e). Moreover, there were no differences in Caspase-1 activation within resident coDCs between progressor and responder mice treated with NP-PEG-BayK8644 (Figure 4f). Thus, the control of tumor growth in mice treated with NP-PEG-BayK8644 versus eNP-PEG does not appear to be explained by differential capacity to trigger inflammasome activation.

Figure 4. Therapeutic treatment with NP-PEG-BayK8644 increases mice survival and reduces tumor growth compared with eNP-PEG.

a) Mouse survival of WT mice bearing 5555 melanoma and injected with NP-PEG-BayK8644 or eNP-PEG. *p< 0.05. Log-rank (Mantel-Cox) Test. b) Mouse survival of WT mice bearing EG7 lymphoma and injected with NP-PEG-BayK8644 or eNP-PEG. *p< 0.05. Log-rank (Mantel-Cox) Test. c) Mouse survival of Tmem176b^−/− mice bearing EG7 lymphomas and injected with NP-PEG-BayK8644 or eNP-PEG. ns= not significant. Log-rank (Mantel-Cox) Test. d) Mouse survival of WT mice injected with vehicle control or free BayK8644. ns= not significant. Log-rank (Mantel-Cox) Test. e) Percentage of FLICA-1 (Active Caspase-1) positive cells within resident coDCs in the TDLN 3 days after the third NP injection in EG tumor-bearing WT animals. ns= not significant. Student’s t test. f) Percentage of FLICA-1 (Active Caspase-1) positive cells within resident coDCs in the TDLN 3 days after the third NP-PEG-BayK8644 injection in EG tumor-bearing WT animals. Comparison of progressors and responder animals. ns= not significant. Student’s t test.

3.4. Increased tumor infiltration by total and tumor-specific CD8⁺ T cells is associated with tumor control in NP-PEG-BayK8644-treated animals.

It was then speculated that preventing antigen cross-presentation inhibition by BayK8644 might enhance the anti-tumoral efficacy of NP-PEG-BayK8644. Anti-tumoral CD8⁺ T cells are primed in the TDLN through the cross-presentation pathway [21]. Therefore, CD8⁺ T cells were analysed by flow cytometry in the TDLN form EG7 tumors-bearing animals treated with eNP-PEG, NP-PEG-BayK8644 or free BayK8644. As hypothesized, the absolute numbers of both total and tumor-specific CD8⁺ T cells were increased in the TDLN from NP-PEG-BayK8644 mice in comparison with free BayK8644 or eNP-PEG-treated animals, although these differences did not reach statistical significance (Figure 5a–b). A similar trend was observed in the tumor microenvironment for absolute numbers of total and tumor-specific CD8⁺ T cells (Figure 5c–d), when comparing the three therapeutic groups. We also found a significant negative correlation between tumor size and the absolute numbers of total (Figure 5e) and tumor-specific (Figure 5f) CD8⁺ T cells within the NP-PEG-BayK8644-treated group in the tumor microenvironment. Accordingly, there was a significant increase in the percentage and absolute number of intratumoral total CD8⁺ T cells in animals with controlled tumor growth (responders) compared to progressors within the NP-PEG-BayK8644-treated group (Figure 5g–i). Similar results were found when analysing intratumoral tumor-specific CD8⁺ T cells (Figure 5j–l). These findings indicate an association between clinical response and tumor infiltration by CD8⁺ T cells, supporting the role of preserved antigen cross-presentation in the anti-tumoral properties of NP-PEG-BayK8644.

Figure 5. Tumor control in NP-PEG-BayK8644-treated mice is associated with tumor cell infiltration by CD8⁺ T cells.

WT mice were subcutaneously injected with 5.0×10⁵ cell line E.G7-OVA. BayK8644, eNP-PEG or NP-PEG-BayK8644 were i.t injected every 2 days when tumors size reached 25 mm².

a) Absolute numbers of total CD8⁺ T cells in tumor-draining lymph nodes (TDLN). b) Absolute numbers of OVA-specific CD8⁺ T cells in TDLN. c) Absolute numbers of total CD8⁺ T cells in tumors. d) Absolute numbers of OVA-specific CD8⁺ T cells in tumors. e) Correlation study between the tumor-infiltrating total CD8⁺ T cell absolute numbers and tumor size (mm²). Correlation test. f) Correlation study between the tumor-infiltrating OVA-specific CD8⁺ T cell absolute numbers and tumor size (mm²). Correlation test. g) Relative frequency of total CD8⁺ T cells among Vβ12⁻ cells in tumors of NP-PEG-BayK8644-treated mice comparing progressor and responder animals. *** p < 0.001, Student’s t test. h) Absolute numbers of total CD8⁺ T cells among Vβ12⁻ cells in tumors of NP-PEG-BayK8644-treated mice comparing progressor and responder animals. **** p < 0.0001, Student’s t test. i) Representative flow cytometry dot plots of total CD8⁺ T cells in tumors of NP-PEG-BayK8644-treated mice comparing progressor and responder animals. j) Relative frequency of OVA-specific CD8⁺ T cells among Vβ12⁻ cells in tumors of NP-PEG-BayK8644-treated mice comparing progressor and responder animals. * p < 0.05, Student’s t test. k) Absolute numbers of OVA-specific CD8⁺ T cells among Vβ12⁻ cells in tumors of NP-PEG-BayK8644-treated mice comparing progressor and responder animals. * p < 0.05, Student’s t test. l) Representative flow cytometry dot plots of OVA-specific CD8⁺ T cells in tumors of NP-PEG-BayK8644-treated mice comparing progressor and responder animals.

4. CONCLUSIONS

High expression of TMEM176B/A has been linked to reduced overall survival in various human tumors, suggesting its potential as a pharmacological target for controlling cancer progression [17–19].

In this work, we developed PEGylated oil-core chitosan nanosystem that efficiently encapsulates the TMEM176B inhibitor BayK8644 in a very efficient manner and with suitable characteristics for targeting CD8+ T cells.

Results suggest that a slow release of the BayK8644 from the nanosystem prevents inhibition of antigen cross-presentation. More specifically, it was observed that 30 minutes post-endocytosis, the nanoparticles are localized within Tmem176b+ endosomes. Furthermore, after a 2-hour incubation time, NP-PEG-BayK8644, but not eNP-PEG, trigger IL-1β secretion in a TMEM176B-dependent manner in DCs. Thus, these results can be explained if the NP-PEG-BayK8644 are internalized in the cell through endocytosis, leading to TMEM176B inhibition once the compound is released between 30 and 120 minutes after internalization.

Additionally, NP-PEG-BayK8644 significantly controlled tumor growth and mice survival in melanoma and lymphoma models compared to eNP-PEG and free BayK8644, in a TMEM176B-dependent manner.

In conclusion, this study provides preclinical proof-of-concept of a pegylated chitosan nanoparticles-based approach that improves the anti-tumoral therapeutic efficacy of BayK8644 by uncoupling the dual role of TMEM176B on innate and adaptive immunity.