Synthesis of PEG-C′ dots and αMSH-PEG-C′ dots
C′ dots were synthesized in aqueous solution as previously described21,22,23. Both plain (PEGylated) C′ dots and αMSH-PEG-C′ dots encapsulate the fluorescent organic dye, Cyanine 5 (Cy5). αMSH peptides with an N-terminal acetylated cysteine thiol enabled reaction with a maleimido-terminated heterobifunctional PEG-silane, the latter obtained by reacting maleimido-NHS ester heterobifunctional PEG (Sigma) with aminopropyl-triethoxysilane (Sigma). The resulting αMSH-PEG-silane was conjugated to the particle surface during the PEGylation step as previously reported27. Once synthesized, particles were purified using gel permeation chromatography (Bio-Rad) and up-concentrated using Vivaspin ultrafiltration spin columns (Cytiva, molecular weight cut-off 30,000). Absorbance measurements of both plain PEG-C′ dots and αMSH-PEG-C′ dots were obtained using a Varian Cary 5000 spectrophotometer, and the number of αMSH peptides per particle was determined via the absorbance of the peptide and its calculated extinction coefficient. Particles were further characterized by fluorescence correlation spectroscopy (FCS) to determine their hydrodynamic radius, concentration and number of encapsulated dyes per particle21.
Cell culture and cell line authentication
B16F10 and B16-GMCSF melanoma cell lines were maintained in RPMI medium supplemented with 10% fetal calf serum (FCS) and 1% penicillin with streptomycin and Plasmocin (InvivoGen catalogue number ant-mpp). B16-Blue IFNα/β (InvivoGen catalogue number bb-ifnt1) and B16-Blue IFNγ cells (InvivoGen catalogue number bb-ifng) cells were maintained in RPMI1640, 10% heat-inactivated FBS, 50 U ml−1 penicillin, 50 μg ml−1 streptomycin, 100 μg ml−1 Normocin and 100 μg ml−1 Zeocin. Cell lines were routinely mycoplasma tested.
Generation of STING-KO B16-GM cells
STING1 KO and scrambled control (scr) B16-GM cells were generated by lentiviral transduction with pLentiCRISPRv2-mSTING-gRNA_1 (GCTGGATGCAGGTTGGAGTA) or pLentiCRISPRv2-mSTING-scrambled_gRNA_1 (CTTTGCGGAGATTGGAGGGA), respectively (Addgene), using standard protocols89. Cells were selected in fresh puromycin (1 μg ml−1; InvivoGen) for 10 days, with media replenished every 2 days. Knockout efficiency was confirmed by western blot. B16-GM, B16-GM STING-KO and B16-GM STING scr cells (5 × 106) were lysed in 1 ml Pierce RIPA buffer (ThermoFisher) on ice for 15 min. Lysates were centrifuged at 15,000 × g for 20 min at 4 °C, and 20 ml supernatant was mixed with 2× Laemmli sample buffer (Bio-Rad), boiled at 95 °C for 10 min and resolved on Mini-PROTEAN 4–20% pre-cast gels (Bio-Rad). Proteins were transferred to nitrocellulose membranes (Bio-Rad), washed once in TBS and blocked with EveryBlot Blocking Buffer (Bio-Rad) for 15 min. Membranes were incubated overnight at 4 °C with rabbit anti-STING (D2P2F, Cell Signaling Technology) or rabbit anti-β-actin (Poly6221, BioLegend) primary antibodies in blocking buffer, followed by HRP-conjugated secondary antibody for 1 h at room temperature (RT). Bands were visualized using Clarity Western ECL substrate (Bio-Rad) and imaged on a ChemiDoc system (Bio-Rad).
Gene expression analysis
Cells were collected at indicated time points and RNA extracted using the Purelink RNA minikit (ThermoFisher catalogue number 12183020) per the manufacturer’s instructions. DNase removal was performed on a column per the manufacturer’s instructions (ThermoFisher catalogue number 12185-010). RNA was extracted from FFPE sections using the Quick RNA-FFPE kit (Zymo Research catalogue number R1008) when tissue samples were limited. Reverse transcription was carried out using the High-Capacity RNA-to-cDNA Kit (ThermoFisher catalogue number 4387406) using between 100 ng and 1 μg of RNA per the manufacturer’s instructions. Quantitative polymerase chain reaction (qPCR) was performed for analysing 2 μl cDNA in a 20 μl reaction using TaqMan assays (FAM-labelled, ABI), TaqMan Fast Advanced Master Mix (ThermoFisher catalogue number 4444556) and a QuantStudio 3 Real-Time PCR Instrument for analysis using the following conditions: 40 cycles, 50 °C (2 min), 95 °C (2 min), PCR cycling 95 °C (1 s) and 60 °C (20 s). mRNA levels were normalized to a reference gene, GAPDH (VIC dye-labelled) using ΔCt = Ct (gene of interest) – Ct (Gapdh), where Ct is the threshold cycle. Results were reported as the relative fold change in mRNA expression over baseline (that is, untreated cells), or ΔΔCt = 2−(ΔCt (sample) − ΔCT (control)) and normalized using a log2 transformation, with untreated cells as the baseline. TaqMan assays are listed in Supplementary Table 1.
Bulk RNA-seq and analysis
Following RNA isolation, FFPE total RNA integrity was checked using a 2100 Bioanalyzer with a DV200 reading as 47–61% for all 6 samples (Agilent Technologies). RNA concentrations were measured using the NanoDrop system (Thermo Fisher Scientific). We used 200 ng of FFPE total RNA sample as the input for fragmentation by treating with metal ions present in the 2X Priming Buffer at elevated temperature (Agilent SureSelect XT HS2 RNA Library Preparation and Target Enrichment Protocol number G9989-90000, Agilent Technologies). Fragmented RNA was used to perform the first- and the second-strand cDNA synthesis. After AmpPure XP Beads purification, cDNA samples proceeded to end repair, A-tailing and adapter ligation following the manufacturer’s instructions. The adaptor-ligated cDNA library was then PCR amplified and hybridized by SureSelect XT Mouse All Exon probes (Agilent Technologies, catalogue number 5190-4641). Captured and target-enriched cDNA libraries were amplified and purified. Final libraries were pooled and subjected to sequencing on the NovaSeq6000 sequencer (Illumina) at PE 2 × 100 cycles. The raw sequencing reads in BCL format underwent processing through bcl2fastq v.2.20 (Illumina) for FASTQ conversion and demultiplexing. Following adaptor trimming with Cutadapt v.1.18 (ref. 90), RNA reads were aligned and mapped to the GRCm39 mouse reference genome using STAR v.2.5.2 (ref. 91). Read counts per gene were extracted using HTSeq-count v.0.11.2 (ref. 92). Gene expression profiles were constructed for differential expression, clustering and principal component analyses using the DESeq2 package v.1.30.1 (ref. 93). For differential expression analysis, pairwise comparisons between two or more groups utilized parametric tests where read counts followed a negative binomial distribution with a gene-specific dispersion parameter. Corrected P values were calculated based on the Benjamini–Hochberg method to adjust for multiple testing.
The volcano plot was generated using the R EnhancedVolcano package v.1.16.0 (ref. 94). GSEA was conducted using the R clusterProfiler package v.4.6.2 (ref. 95), and the enrichment results were visualized using the R enrichplot package v.1.18.3.
Type I/II IFN assays
B16-Blue IFNα/β (InvivoGen catalogue number bb-ifnt1) and B16-Blue IFNγ cells (InvivoGen catalogue number bb-ifng) were plated in 96-multi-well plates for 24 h (4.2 × 105 cells per ml), followed by addition of media (180 μl) and supernatant/sera (20 μl), and incubated (37 °C and 5% CO2 for 24 h). To determine secreted embryonic alkaline phosphatase production, supernatants from treated and untreated B16-Blue (50 μl) cells were added to new 96-well plates and used immediately or snap frozen at −80 °C. QuantiBlue assay reagent (150 μl; InvivoGen catalogue number rep-qbs) was added to samples to a final volume of 200 μl, and the mixture was incubated at 37 °C and 5% CO2 for 1–6 h and absorbance (l = 620–655 nm) was quantitatively measured using a SpectraMax iD3 Multi-Mode Microplate Reader (Molecular Devices).
Secreted cytokine/chemokine determination
Cytokine/chemokine secretion from treated specimens was detected using a Proteome profiler array kit (R&D Systems, mouse, catalogue number ARY006), according to the manufacturer’s instructions. Arbitrary values of abundance were calculated as integrated densities (ImageJ software, v1.51w) of each dot plot normalized to a reference spot.
Total cellular and mitochondrial ROS quantification
After plating treated and untreated cells in 6-well dishes, adherent cells were stained with CellROX Green Reagent (ThermoFisher catalogue number C10444) or MitoSOX Red Mitochondrial Superoxide Indicator (Thermo Fisher catalogue number M36008) per the manufacturer’s instructions. ROS generation was assessed by an LSRFortessa flow cytometer (BD Biosciences).
In vitro analysis of BMDMs
Using established protocols, BMDMs were derived from tibias and fibulas collected from B16-GM tumour-bearing mice (WT, STING-KO). In brief, bone marrow cell suspensions were cultured for 7 days in complete medium (RPMI 1640 supplemented with L-glutamine, 10% FBS, 1% penicillin and streptomycin, 1 mM sodium pyruvate and 1% non-essential amino acids) in the presence of 10 ng ml−1 recombinant mouse macrophage colony-stimulating factor (M-CSF; PeproTech catalogue number 315-02), with media replaced every 2–3 days. One week later, attached cells were collected and used as part of stimulation assays. For phenotypic analysis, gene expression and IFN/cGAMP analysis, BMDMs were cultured with or without C′ dots for 48 h in complete media. Phenotypic analysis was performed using antibodies listed in Supplementary Table 3 and the Aurora Spectral Cytometer (Aurora Cytek); gene expression analysis used primers listed in Supplementary Table 1. Phagocytosis assays were performed by labelling tumour cells with carboxyfluorescein succinimidyl ester (eBioscience CFSE, ThermoFisher catalogue number 65-0850-84) per the manufacturer’s instructions after being treated with 15 μM C′ dots or vehicle control for 72 h. This was followed by addition of BMDMs at an effector-to-target cell ratio of 1:1. Cells were collected 24 h later, stained for CD11b and analysed on an LSRFortessa flow cytometer; CD11b+ cells were examined for CFSE positivity.
In vitro analysis of T cells
Spleens from pmel-1 TCR transgenic mice were isolated and ground through 100 μm filters to obtain T cells. After red blood cell lysis, T cells were activated by culturing splenocytes with Dynabeads Mouse T-Activator CD3/CD28 (ThermoFisher catalogue number 11456D) in RPMI 1640 medium containing 1% penicillin/streptomycin (MediaTech), 50 mM β-mercaptoethanol (Gibco, Invitrogen), 1% L-glutamine (Gibco) and 100 U ml−1 IL-2 (PeproTech). Cells were labelled with eBioscience CFSE (Thermo Fisher catalogue number 65-0850-84). Following treatment, cytotoxicity was quantified at a T cell-to-tumour cell ratio of 1:1 using the CyQUANT LDH Cytotoxicity Assay (Thermo Fisher catalogue number C20300); supernatants were analysed for type I/II IFNs. Antigen-specific T cells were quantified via H-2Db gp100 Tetramer-EGSRNQDWL-PE (MBL catalogue number TS-M546-1) analysis. Proliferation, tetramer and surface marker analysis were performed 72 h post-stimulation using the LSRFortessa flow cytometer.
DNA damage, immune-related cell death and inflammasome analysis
All antibodies/assays were used per the manufacturer’s instructions. To quantify the degree of cellular DNA damage, PE anti-H2A.X Phospho (Ser139) Antibody (BioLegend catalogue number 613411) was incubated with particle- and vehicle-treated cells and analysed by flow cytometry (BD Biosciences LSR II). For analysis of immune-related cell death, cell surface calreticulin was assayed using PE-conjugated calreticulin (D3E6) XP rabbit mAb (Cell Signaling Technology catalogue number 19780) and analysed with a BD Biosciences LSR II flow cytometer. Extracellular ATP was quantified using the RealTime-Glo Extracellular ATP Assay (Promega catalogue number GA5010), and HMGB1 was quantified using the Lumit HMGB1 Human/Mouse Immunoassay (Promega catalogue number W6110). To assess caspase-1, the Caspase-Glo 1 Inflammasome Assay (Promega catalogue number G9951) was used. Luminescence values were determined by subtracting background luminescence levels (that is, no cells).
Analysis of senescence induction, cell cycle and autophagy
Senescence was quantified using two independent methods: (i) bright-field analysis using the Senescence β-Galactosidase Staining Kit (Cell Signaling Technology, catalogue number 9860) for particle- and vehicle-treated cells (6-well plate, 3 wells per condition) (100 cells at a minimum were counted for each condition), and (ii) flow cytometric analysis (LSRFortessa, BD Biosciences) using the Cellular Senescence Detection Kit SPiDER-βGal (catalogue number SG03-10; Molecular Technologies) for cells treated in the same manner as (i) (10,000 cells per condition were analysed with data shown representing the mean of 2 independent experiments).
For cell-cycle distribution studies, cells were initially synchronized using the cyclin-dependent kinase 1 (CDK1) inhibitor RO-3306 (Selleck, catalogue number S7747), followed by C′ dot or vehicle treatment. Cells were then fixed with 70% ethanol overnight, washed in triplicate with phosphate-buffered saline (PBS), stained with Alexa Fluor 488 anti-Histone H3 Phospho (Ser10) antibody (BioLegend catalogue number 650803) for 1 h, washed again in triplicate with PBS and resuspended in 200 ml propidium iodide/RNase Staining Solution (Cell Signaling Technology catalogue number 4085). Cells were analysed by flow cytometry (LSRFortessa); gates for each phase of the cell cycle were manually determined. At least 10,000 cells per condition were analysed for 2 independent experiments.
Autophagy was assessed by measuring autophagolysomal formation with DALGreen (catalogue number D675-10; Dojindo Molecular Technologies). In brief, DALGreen (1 mmol l−1), after reconstituting (per the manufacturer’s instructions), was added to 5 × 105 B16-GM cells in RPMI 1640 media (30 min, at 37 °C), washed twice with PBS, followed by addition of fresh media, with or without C′ dots. Cells were collected 72 h post-treatment and analysed by flow cytometry (BD Biosciences LSR II) with gates manually determined. At least 10,000 cells per condition were analysed for 2 independent experiments.
Evaluation of STING pathway activation by C′ dots and a reference agonist
WT B16-GM cells were incubated with 15 μM C′ dots for 72 h, washed and stained intracellularly for p-STING (Supplementary Table 2). ROS levels were assessed in parallel. For comparison, cells were incubated over the same time interval with the molecular STING agonist, ADU-S100 (2′3′-c-di-AM(PS)2 (Rp, Rp); InvivoGen) and stained for pSTING. Flow cytometry was used to quantify ROS production and STING activation.
cGAMP analysis via ELISA
B16-GM cells (5 × 105) were resuspended in 250 μl RIPA lysis buffer (Pierce catalogue number 89901) supplemented with cOmplete Protease Inhibitor Cocktail (Sigma catalogue number 11697498001) and lysed with a 28½ gauge needle. Lysates were incubated on ice (30 min), centrifuged (16,000 × g) and cooled at 4 °C (10 min). 2′3′-cGAMP levels in lysates and supernatants from collected tumour cells or BMDMs were quantified using the 2′3′cGAMP ELISA kit (Arbor Assays catalogue number K067-H1W) and then normalized to total numbers of cells per condition.
Subcellular fractionation and quantification of mitochondrial and nuclear DNA
For analysis and quantification of cytosolic DNA, cell fractionation assays were performed using NE-PER nuclear and cytoplasmic extraction reagents (ThermoScientific, catalogue number 78833) per the manufacturer’s instructions48. Primers specific for mitochondrial DNA (mt-Dloop, forward sequence AATCTACCATCCTCCGTGAAACC and reverse sequence TCAGTTTAGCTACCCCCAAGTTTAA) and genomic (TERT, forward sequence CTAGCTCATGTGTCAAGACCCTCTT and reverse sequence GCCAGCACGTTTCTCTCGTT) were used for analysis normalized to whole cell extract values. qPCR was performed using the PowerUp SYBR Green Master Mix (Applied Biosystems catalogue number A25742) with melt curve analysis on a QuantStudio 3 Real-Time PCR Instrument.
In vivo murine models
All animal studies were performed in accordance with protocols approved by the Weill Cornell Medical College Institutional Animal Care and Use Committee (IACUC) and conformed to NIH guidelines for animal welfare. All mice were obtained from the Jackson Laboratory. Six- to eight-week-old female C57BL/6J, NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) and C57BL/6J-Sting1gt/J mice were implanted intradermally with either B16-GM (1.5 × 104) or B16F10 (1 × 105) cells. When tumours reached a size of 150–200 mm3, mice were i.v. injected with a multi-dose regimen (n = 3 doses, every 3 days) of either particles (36 nmol) or saline vehicle. Pmel-1 mice with a gp100-reactive transgenic T cell receptor were used for in vitro studies to assess antigen-specific tumour killing and expansion. For ferroptosis studies, mice were also i.p. injected with liproxstatin-1 (Selleckchem), as previously described37. Tumour volumes were measured daily via calipers, and volumes were calculated using the following formula: volume = (long axis × short axis2)/2. No tumours were allowed to exceed the maximal size of 1,500 mm3, as permitted by our institutional IACUC.
Antibody depletion and survival/treatment studies
Five cohorts of female C57BL/6J mice (6–8 weeks, Jackson Laboratory) were inoculated with 1.5 × 104 B16-GM cells for performing antibody depletion studies 7 days post-implantation. Mice were i.v.-injected with vehicle, C′ dots or C′ dots administered with one of the following i.p. injected mAbs (200 mg each): (i) anti-CD4 (clone GK1.5), (ii) anti-NK1.1 (clone PK136), (iii) anti-CD8 (clone 53-5.8) or (iv) anti-CSF1R (clone AFS98) (BioXCell). mAbs were given on day 1 of treatment, and then biweekly for 2 weeks until study termination. Depletion was also confirmed by flow cytometry using collected tumour tissue and blood samples. For treatment studies, mice received anti-mouse PD-L1 rat IgG2a (clone RMP1-14) or anti-mouse IL-6 (clone MP5-20F3) (BioXCell) antibody by i.p. injection (200 µg per dose, as indicated in treatment schemas). Tumour volumes were measured every 2–3 days until tumours reached 1,500 mm3. Kaplan–Meier survival curves were generated, and the log-rank test was used to investigate the significance of various treatments for survival.
Cellular IF
Cells were grown in 8-well chamber slides (ibidi USA, catalogue number 80841) and then fixed in 4% paraformaldehyde for 20 min at RT. Cells were permeabilized (0.1% Tween 20, 0.01% Triton X-100 in PBS) for 10 min, blocked with 1% BSA in PBS (30 min) and then incubated with 1:500 dilution dsDNA antibody in blocking buffer (clone HYB331-01; Abcam catalogue number ab27156) overnight in a humidified chamber at 4 °C. Cells were triply washed with PBS and then incubated with Alexa Fluor 594 anti-mouse antibody in 0.3 ml blocking buffer (1 h, RT). Cells were again triply washed with PBS, counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and fixed with Invitrogen ProLong Gold Antifade Mountant (ThermoFisher catalogue number P36941) for curing overnight. PicoGreen staining was performed with Quant-iT PicoGreen dsDNA Reagent and Kits (ThermoFisher, catalogue number P7589). For confocal microscopy, cells were incubated with PicoGreen (37 °C, 1 h) (ThermoFisher, catalogue number P11495), which was diluted in cell culture medium (3 ml ml−1). Cells were counterstained with DAPI (Thermo Fisher catalogue number P36941), washed and fixed for microscopy (Invitrogen ProLong Gold Antifade Mountant).
Flow cytometric analysis of collected tumours
Mouse tumour samples were minced with scissors and scalpels before incubation with 1.67 U ml−1 Liberase (Roche) and 0.2 mg ml−1 DNase (Roche) in RPMI for 30 min at 37 °C. Tumour samples were homogenized by repeated pipetting and filtered through a 100 mm nylon filter (BD Biosciences) in RPMI supplemented with 5% FCS to generate single-cell suspensions. Cell suspensions were washed once with PBS and resuspended in FACS buffer (PBS/1% BSA). Cells from mouse spleens were isolated by grinding spleens through 40 mm filters. After RBC lysis (ACK Lysing Buffer, Lonza), all samples were washed and resuspended in FACS buffer (PBS/1% BSA) or RPMI. Cells isolated from mouse tumours and spleens were pre-incubated (15 min, 4 °C) with anti-CD16/32 mAb (Fc block, clone 2.4G, BD Biosciences catalogue number 553142) to block non-specific binding and stained (30 min, 4 °C) with combinations of fluorochrome-conjugated antibodies at a ratio of 1:50 (Supplementary Table 3). Fixation and permeabilization (15 min fixation, RT) were carried out with the Foxp3/Transcription Factor Staining Buffer set (eBioscience catalogue number 00-5523-00), followed by intracellular staining in permeabilization buffer for 30 min at 4 °C. Dead cells were excluded using a Zombie NIR Live/Dead fixable dye (1:2,500 in PBS; BioLegend catalogue number 423105). Flow cytometry was performed (Aurora Spectral Cytometer, Aurora Cytek) and data analysed with FlowJo (v.10) software (FlowJo LLC). A representative gating strategy is shown in Supplementary Fig. 17.
Assessment of ROS and STING activation in collected tumour specimens
Tumours were collected from vehicle- and C′ dot-treated B16-GM-bearing mice (n = 4 per group) 3 days after the final dose, dissociated into single-cell suspensions and subjected to RBC lysis and 70 µm filtration. Cells were stained with anti-CD45 and -TYRP, anti-F4/80 and anti-CD3 mAbs for evaluating tumour and immune cell subsets. ROS was measured with CellROX Green, and STING activation assessed by intracellular staining for p-STING after fixation and permeabilization. Data were acquired by flow cytometry and analysed in FlowJo, with gating on tumour cells (CD45−/TYRP1+), macrophages (CD45+/F4/80+) and T cells (CD45+/CD3+).
IHC and IF for collected tissue specimens
Tissues were collected and incubated for 24 h in 10% buffered formalin and then transferred to 70% ethanol for 24–48 h. Tissues were embedded in paraffin embedding blocks, and 5 μM sections were prepared on slides. The pathology slides were deparaffinized and rehydrated by washing sequentially in xylene 3 times for 3 min each, absolute (100%) ethanol 3 times for 3 min each, 95% ethanol 3 times for 3 min and the final wash with water 2 times for 5 min each. Antigen retrieval was performed by autoclaving the slides in citric acid solution at 121 °C for 20 min in a pH 9.0 buffer. These samples were cooled to RT overnight. After antigen retrieval, slides were washed twice in water for 5 min, PBS for 5 min one time and PBS with 0.1% Tween 20 once for 5 min. Slides were blocked for 1–4 h with 10% normal goat serum in 1% BSA and PBS at RT. IHC for Ki-67 was performed on paraffin tumour sections (3 sections per tumour) using a Leica Bond RX automated stainer. After heat-induced epitope retrieval, the primary antibody, rabbit monoclonal antibody clone D3B5 (Cell Signaling number 12202), was applied at a concentration of 1:500, followed by a polymer detection system, according to the manufacturer’s instructions (DS9800, Novocastra Bond Polymer Refine Detection, Leica Biosystems). The chromogen used was 3,3-diaminobenzidine tetrachloride (DAB), and sections were counterstained with haematoxylin. IF staining was performed for γH2AX, STING and PD-L1 on paraffin tumour sections as follows. After heat-induced epitope retrieval in a pH 9.0 buffer, primary antibodies (anti-γH2AX 0.2 mg ml−1; Abcam, number 11174, anti-STING 0.1 mg ml−1; Abcam, number 178847 or anti-PD-L1 0.5 mg ml−1; Cell Signaling, number 64693) were applied. Secondary biotinylated goat anti-rabbit IgG secondary antibody was subsequently added (5.75 mg ml−1; Vector Labs, number PK6101). Application of streptavidin-HRP D (DAB Map kit, Ventana Medical Systems) was followed by incubation with Tyramide Alexa Fluor 488 (Invitrogen, number T20922), prepared at a 1:150 dilution. Slides were counterstained with DAPI (5 mg ml−1, Sigma-Aldrich, number D9542) and mounted with Mowiol. Quantification was performed using QuPath software (v.0.6.0), and a minimum of 10 fields of view were analysed.
Multiplex IF
Staining
Multiplex IHC was conducted using a tyramide signal amplification approach96,97. Validation and condition optimization for the primary antibody was performed using chromogenic IHC. Slides were heated to 60 °C for 30 min, followed by processing in a Bond Rx Research Stainer (Leica Biosystems), where slides were deparaffinized using the ‘dewax’ option and submitted for antigen retrieval at pH 9 for 25 min at 98 °C (BOND Epitope Retrieval Solution 2, catalogue number AR9640, Leica Biosystems). After several washes with immunostainer buffer 1× (Wash Solution 10× Concentrate, catalogue number AR9590, Leica Biosystems), slides were incubated with the primary antibody anti-AQP1 (30 min, RT) at a specified dilution using an antibody diluent (catalogue number ab64211, Abcam) and then treated with washing buffer. Slides were then incubated with rabbit anti-mouse IgG (Bond Polymer Refine Detection kit, catalogue number DS9800, Leica Biosystems) for 8 min at RT, treated with washing buffer and incubated with polymer anti-rabbit Poly-HRP-IgG. After further washes with buffer and deionized water, slides were incubated with peroxide block for 5 min and 3,3′-diaminobenzidine (10 min, RT). Slides were then counterstained with haematoxylin for 5 min at RT (Bond Polymer Refine Detection). The slide was then dehydrated manually using a series of increasing concentrations of ethanol and Citrisolv (catalogue number 89426-270, Decon Labs). Finally, slides were mounted with a coverslip and Canada balsam.
Multiplex IF staining for six markers was also performed on a Bond Rx Research Stainer, using the Opal 6-plex detection kit (catalogue number NEL821001KT, Akoya Biosciences) for the detection of protein markers, using DAPI as a counterstain for nuclear detection (Opal 6-plex detection kit). All washes were done with BOND wash solution (catalogue number AR9590, Leica Biosystems). After baking, dewaxing and epitope retrieval steps were performed before the first cycle of staining was started. Slides were incubated with the blocking buffer 1× (Opal Antibody Diluent/Block, catalogue number ARD1001EA, Akoya Biosciences) for 6 min at RT, followed by the primary antibody for 30 min at RT, which had been diluted in antibody diluent (catalogue number ab64211, Abcam) to the dilution previously established. After washing, sections were incubated with SuperBoost Goat anti-mouse Poly-HRP (ready-to-use solution, catalogue number B40961, Thermo Scientific) or with anti-rabbit Poly-HRP-IgG (reagent from Bond Polymer Refine Detection, catalogue number DS9800, Leica Biosystems), and then treated again with wash solution. After incubating with Opal 480 reagent (10 min, RT), slides were treated with wash solution and incubated with BOND Epitope Retrieval Solution 1 pH 6.0 (catalogue number AR9961, Leica Biosystems) for 20 min at 98 °C. Following treatment with washing reagent, slides were again incubated with blocking buffer 1×, and another staining cycle started for a total of 6 cycles. Slides were then taken out of the immunostainer and mounted with ProLong Diamond Antifade Mountant (catalogue number P36961, ThermoFisher Scientific).
Image acquisition, processing and analysis
Double IF-stained slides were left to dry and then acquired using the Vectra Polaris instrument (v.1.0, Akoya Biosciences). In brief, slides were mounted on a tray and pre-scanned for the acquisition. After finding the focal plane on the DAPI channel and optimizing the exposure time to avoid saturation for all the channels (DAPI, 6 opal dyes and autofluorescence channel), slides were acquired and then exported into a drive for post-acquisition processing. Spectral unmixing was performed using InForm software (v.2.6, Akoya Biosciences), and tiles produced were stitched and visualized using the Halo software (v.3.6, Indica Labs).
For analysis purposes, 3 tissue areas were considered and defined through annotation: central tumour area, periphery of the tumour (defined as 400 µm of the most external portion of the tumour) and the non-neoplastic peritumoural area. To separate the different tissue compartments, a DenseNet classifier (HALO AI add-on, v.3.6, Indica Labs) was trained to recognize glass background, stroma, melanoma cells, necrotic areas and artefact areas (out-of-focus regions, foldings, foreign bodies and so on). The positivity for each marker was determined manually by an experienced pathologist (F.S.), adjusting the threshold value for each single slide. The analysis module on Halo was used to extract the cytoplasmic positivity for Melan-A, CD8, CD4, PD-L1, F4/80 and nuclear positivity of FOXP3 on all samples, stratifying the results on the basis of areas (that is, central tumour, tumour periphery and non-neoplastic/peritumoural). Double positivity for PD-L1-Melan-A, PD-L1-F4/80 and CD4-FOXP3 was also measured by the module. Results were expressed as the percentage of single- and double-positive cells for each marker in the stroma.
Histopathological analysis of treated hepatic and splenic specimens
Hepatic and splenic specimens were collected from particle-treated and control mice, fixed in 10% neutral-buffered formalin and embedded in paraffin. Sections (5 mm) were cut, mounted on glass slides and stained with haematoxylin and eosin following standard protocols. Stained sections were examined by light microscopy, and histopathologic evaluation was performed in a blinded manner by an experienced pathologist (S.M.) to assess tissue architecture and treatment-related changes.
Statistical analysis
Statistical comparisons for double-arm studies in which animals received either the particle treatment or vehicle control were analysed using a two-sided unpaired t-test. Group means and s.e.m. were calculated for time- and/or concentration-dependent changes in cellular uptake, tumour and immune cell markers, functional assays and tumour volumes. Statistical comparisons across multiple experimental groups used a non-parametric one-way analysis of variance (ANOVA) followed by post hoc Tukey’s test. A two-way analysis of variance (ANOVA) followed by Sidak’s multiple comparison test was used for analysing growth inhibition studies. Statistical significance was set at *P < 0.05, **P < 0.01, ***P < 0.005 and ****P < 0.001 for all studies, unless otherwise indicated. All graphs were constructed and analysed using GraphPad Prism 7 (GraphPad Software).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
