Animal studies were approved by the Ethics Committee of Animal Experimentation (CEEA) of the CNB and of the CIMA with compliance with national, institutional and EU guidelines. Six- to eight-week-old female BALB/c and C57Bl/6 were purchased from Envigo (www.envigo.com). C57Bl/6 Batf3 ^tm1Kmm/J (Batf3^−/−) [33] and IFN-a/bR ^o/o (IFNARKO) [34] were kindly provided, by Dr. Kenneth M. Murphy, Washington University, St. Louis, MO and by Matthew Albert (Institut Pasteur, Paris) respectively, and were bred at CIMA in specific pathogen-free conditions. Mice were housed in the Animal Facility of Centro Nacional de Biotecnologia (CNB-CSIC, Madrid, Spain) and Centro de Investigacion Medica Aplicada (CIMA, Pamplona, Spain).

B16-F10 mouse melanoma cells and 4 T1 mouse breast carcinoma were purchased from the ATCC, and B16-OVA melanoma cells and MC38 colon carcinoma cells were a kind gift from Dr. Lieping Chen (Yale University, New Heaven, CT) and Dr. Karl E. Hellström (University of Washington, Seattle, WA) respectively. Tumor cells were cultured in RPMI 1640 (Gibco) containing 10% fetal bovine serum (FBS, Sigma-Aldrich), 2 mM glutamine (Gln, Gibco), 100 U/ml penicillin and 100 μg/ml streptomycin (100 U/ml), and 50 μM 2-mercaptoethanol (Gibco). The B16-OVA cell line was supplemented with 400 μg/mL Geneticin (Gibco). Cell lines were routinely tested for mycoplasma contamination (MycoAlert Mycoplasma Detection Kit, Lonza).

UMBY and ICNI human melanoma were derived from primary surgical samples of metastatic lesions of patients at the Department of Dermatology, University Hospital Erlangen and grown in DMEM (Gibco) containing 10% FBS, 4 mM Gln and 1% P/S. HT-29 and HCT 116 colon cancer from the ATCC were cultured in RPMI, 2 mM Gln, 10% FBS and 1% P/S. SK-BR-3 and BT-474 breast cancer cell lines were a kind gift from Dr. López-Botet, IMIM, Barcelona and were grown in DMEM/F12 (1:1) (Invitrogen), containing 2.5 mM Gln, 10% STF and 1% P/S.

BO-112 was developed and provided by Bioncotech Therapeutics (Madrid, Spain). All experiments were performed with the same batch.

The in vitro cytotoxicity of BO-112 in mouse and human cell lines was continuously assessed by measuring electric impedance in an xCELLigence machine (ACEA). Tumor cells (1.5-2 × 10⁵) were seeded on specific 8-well plates to measure electric impedance. After 4-5 h, BO-112 or Poly I:C (Sigma) was added to culture media at identical concentrations in a final volume of 200 μL per well. PEI (Polyplus-transfection®) was added to culture media at the same concentrations as it is present in BO-112 formulation. Electric impedance was measured every five minutes for 48 h. In vitro BO-112 cytotoxicity was also assessed by the CellTiter AQueous One Solution Cell Proliferation Assay (MTS, Promega). Briefly, tumor cells (5 × 10³ cells/well; 96 flat-well plates, 8 replicates per condition) were cultured for 48 h, alone or with BO-112 (0.25, 0.5 and 1 μg/mL) and absorbance (OD 492 nm) measured in an ELISA plate reader. Three independent MTS assays were performed. Cell viability is referred to untreated cells (100%).

For RNA expression analyses, B16-OVA cell lines were cultured 24 h with BO-112 at 0.5 μg/mL or in the presence of equivalent volumes of BO-112 vehicle.

HMGB1 detection in culture supernatants was performed with HMGB1 ELISA detection kit following the manufacturer’s instructions (IBL International ST51011).

B16-F10 and B16-OVA melanoma, MC38 colon carcinoma or 4 T1 breast carcinoma cells were injected subcutaneously (5 × 10⁵–10⁶) into the right flank of 8- to 10-week-old female C57BL/6 or BALB/c (6–11 mice/group) on day 0. Tumors were measured twice per week with calipers and the volume calculated (length x width²/2). When tumors reached a volume of 80–100 mm³ (day 0) mice were randomized into different groups of treatment according to the experiment. Poly I:C or BO-112 formulation (2.5 mg/Kg, 100 μl), was administered by intratumoral injection twice per week for three weeks (six doses in total). The control group received intratumoral injections of 5% glucose (BO-112 vehicle, identical volume) or PEI (identical amount per dose as present in each dose of BO-112). Survival was monitored daily, and tumors were measured twice per week until the animals died or the tumor volume reached the maximum allowed size.

To evaluate the systemic antitumor effects, 5 × 10⁵ (injected/treated tumor) and 1.5 × 10⁵ (contralateral tumor) B16-OVA cells were injected into each flank respectively. For evaluation of cDC1 in BO-112-meditated antitumor response, identical experiments were performed in Batf3^−/− mice (in parallel with WT mouse groups). For evaluation of intratumoral BO-112 in combination with systemic immunostimulatory monoclonal antibodies, mice were intratumorally injected with BO-112 or vehicle. The intratumoral treatment schedule was the same as that described for single tumor models. Starting at the second BO-112 administration, mice received concomitant intraperitoneal administration (150 μg/dose) of either InVivoPlus anti-PD-L1 (10F.9G2), InVivo anti-CD137 (3H3) or InVivoMAb rat IgG from BioXCell.

For flow cytometry, IHC and RNA extraction experiments, mice received two intratumoral administrations and were euthanized 48 h post-second dose.

Fresh tumors were excised from mice, weighed and mechanically dissected and enzymatically digested for 15–30 min at 37 °C with appropriate medium: DMEM F-12 (Gibco), 1 mg/mL Collagenase 1A (Roche), 2.5 U/mL Dispase (Roche), 20 U/mL DNAse-I (Roche), 20 mM HEPES (Lonza) and antibiotics. The enzymatic reaction was stopped with 5% FBS in phosphate-buffered saline (PBS). After hypotonic lysis, cell aggregates were removed by filtering the cell suspension with a 70-μm cell strainer (BD Falcon, BD Bioscience) and counted. Lymph nodes were excised from mice and mechanically disrupted and passed through a 70-μm cell strainer. Perfect count microspheres (Cytognos) were used as an internal standard according to the manufacturer’s instructions to calculate absolute cell counts in cell suspensions.

Single cell suspensions from tumors and lymph nodes were previously treated with FcR-Block (anti-CD16/32 clone 2.4G2; BD Biosciences Pharmingen) and were then surface stained at 4 °C with fluorochrome-labeled antibody cocktails defined for each staining. Tetramer staining was performed according to the manufacturer’s protocol. Flow cytometry antibodies, tetramers, cell death stainings and isotype controls are listed in Additional file 1. Table S1. For intracellular FOXP3 staining, cells were fixed and permeabilized using the True-Nuclear™ Transcription Factor Buffer Set (Biolegend).

Samples were acquired in a Gallios Cytometer (Beckman Coulter), a FACSCanto II (BD Biosciences) and a CytoFLEX Flow cytometer (BD Biosciences). Kaluza Flow Analysis Software (Beckman Coulter) and FlowJo (Treestar) software were used for data analysis.

100–300 μg/dose of anti-CD4 (GK1.5), anti-CD8 (2.43), anti-NK1.1 (PK136), anti-Gr1 (RB6-8C5) mAbs or Rat IgG2b (LTF-2) from BioXCell, were injected one day before therapy, concurrently with the first intratumoral injection and at days 3, 7, and 10 after the beginning of therapy. Cell depletion was validated in blood samples by flow cytometry analysis, showing a specific reduction of more than 95% of each respective cell subset. Gr1 depletion was confirmed in the tumor microenvironment on day 7 (the day of the first BO-112 injection). For IFNγ neutralization, mice were treated with 250 μg of anti-IFNγ (XMG1.2) or rat IgG the day prior to each BO-112 treatment. Then, mice were injected weekly with for depletion maintenance (100 μg/dose).

Formaldehyde-fixed and paraffin-embedded tissue sections (3 μm thick) were cut, dewaxed and hydrated. Heat induced antigen retrieval was applied for 30 min at 95 °C in 0,01 M Tris-1 mM EDTA buffer (pH = 9) in a Pascal pressure chamber (S2800, Dako). Sections were incubated overnight at 4 °C with anti-CD4 (1:1000; Abcam, ab183685) or anti-CD8 (1:300; Cell Signaling, 98,941) Visualisation was performed using MACH 2 rabbit AP-polymer (Biocare Medical, RALP525) with StayRed (Abcam, ab103741) as chromogen according to the manufacturer’s instructions.

Total RNA was isolated in two steps using TRIzol (Life technologies) and Rneasy Mini-Kit (Quiagen) purification, following the manufacturer’s RNA cleanup protocol. The assessment of RNA integrity was performed with the Agilent 2100 bioanalyzer (Agilent Technologies) and high-quality RNA was hybridized to Affymetrix Clariom S Mouse Affymetrix microarrays following the manufacturer’s protocol.

The transcriptome experiment with Clariom S Mouse Affymetrix microarrays was normalized using the Robust Multichip Average (RMA) algorithm [35]. After quality assessment, a filtering process was performed to eliminate low expression probe sets. Applying the criterion of an expression value greater than 4 in at least 3 samples of one of the experimental conditions (BO-112 or control samples), 21,731 probe sets were selected for statistical analysis in the in vivo experiment. Regarding the differential expression analysis upon BO-112 incubation of B16-OVA in vitro, 18,412 probe sets were selected for statistical analysis after applying the criterion of an expression value greater than 4 in at least 2 samples of one of the experimental conditions (BO-112 or control samples).

Linear Models for Microarray Data (LIMMA) [36] was used to identify the probe sets that showed significantly differential expression between experimental conditions. Genes were selected as significant using a B-statistic cut-off (B > 0). R and Bioconductor were used for preprocessing and statistical analysis [37].

The functional enrichment analysis was performed using Ingenuity Pathway Analysis (Ingenuity Systems, www.ingenuity.com), whose database includes manually curated and fully traceable data derived from literature sources. In addition, enrichment analyses of some gene sets of interest extracted from different publications [38, 39] were performed using the hypergeometric distribution in R [37]. Microarray expression data can be downloaded from Gene Expression Omnibus (GEO) under the Series accession number GSE116078.

Statistical analyses were performed using Prism software (GraphPad Software, Inc.). A two-tailed Student’s t-test or Mann–Whitney tests were used to analyze statistical differences between groups. The Mantel-Cox test was used for survival analysis. For tumor growth data analyses, mean volumes of tumors over time were fitted using the formula y = A x e (t-t0) / (1 + e(t-t0)/B), where t represents time, A the maximum size reached by the tumor and B its growth rate. Treatments were compared using the extra sum-of-squares F test. Values of p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) were considered significant.

BO-112 is a GMP-grade pharmaceutical composition of nanoplexed Poly I:C (300–5000 mer) coupled to polyethylenimine characterized by their monomodal diameter distribution (at least 90% of particles mono-modal diameter distribution bellow 300 nm) and z-average diameter (less or equal 150 nm) (PCTEP2016078078, Additional file 2: Figure S1). Previous forms of Poly IC-PEI nanocomplexes termed BO-110 have been intravenously delivered in mouse models giving rise to therapeutic effects against subcutaneous transplanted melanomas in a fashion related to its ability to induce tumor cell apoptosis in the context of intense autophagy elicited via MDA5 stimulation in malignant cells [31, 32].

In keeping with those findings, Fig. 1a shows that BO-112 induces death in a dose-dependent manner in cultured cell lines able to engraft in syngeneic mice representing melanoma (B16-F10, B16-OVA), colon cancer (MC38) and triple negative breast cancer (4 T1). Tumors derived from such cell lines are described as poorly immunogenic and difficult to treat with immunotherapy [40]. BO-112-induced cytotoxicity was also observed in a panel of human cell lines (Additional file 3: Figure S2A), including melanoma, colon cancer and breast cancer. Of particular importance, identical concentrations of Poly I:C were not cytotoxic in either mouse or human tumor cell lines (Fig. 1a and Additional file 3: Figure S2A). These results were further confirmed by measuring cytotoxicity in MTS assays in mouse and human cell lines (Additional file 3: Figure S2B).Fig1Fig. 1

Local injection of BO-112 exerts antitumor effects. a. Cell viability (in terms of electric impedance) of cultured tumor cell lines was measured in xCELLigence plates over time in the presence of different concentrations of BO-112 or Poly I:C as indicated, to study effects on cell viability. b. Tumor volume follow-up of in vivo engrafted syngeneic B16F10 tumors treated intratumorally with control vehicle, Poly I:C or BO-112 as indicated in the diagram. Representative photographs of mice treated with BO-112, Poly I:C or control vehicle are included as an inset. c. Individual follow-up of tumor volume means ± SD (in graphs on the right) of MC38 and 4 T1-bearing mice treated with BO-112 or control vehicle as indicated. Experiments are representative of two similarly performed. ***P < 0.001

Since direct intratumoral injection of BO-112 therapeutically controlled tumor progression, we tested whether similar results could be achieved by subcutaneous delivery elsewhere in the mouse. Such experiments comparing intratumoral versus subcutaneous delivery at the same doses were performed in MC38 and B16-OVA tumor-bearing mice (Additional file 5: Figure S4). As shown in Additional file 5: Figure S4, while intratumoral delivery controlled tumor progression, this was not the case for subcutaneous administrations which failed to control tumor progression. Accordingly, intratumoral release is more efficacious than subcutaneous release under identical experimental conditions.

Intratumoral BO-112 therapeutic effects could be the result of direct tumor cell death or the result of enhanced antitumor immune responses or a combination of both factors. To start addressing this question, we studied the composition of the lymphoid infiltrates following BO-112 intratumoral delivery in B16-OVA-derived tumors (Fig. 3a and Additional file 6: Figure S5). Interestingly, we observed changes in the leukocyte tumor microenvironment in favor of CD8⁺ cells in proportion to CD4⁺ T cells and regulatory T cells (Tregs) (Fig. 3b). In fact, higher numbers of CD8⁺ cells could be observed per gram of tumor tissue, whereas FOXP3⁻CD4⁺ and FOXP3⁺CD25⁺CD4⁺ Tregs cells decreased (Fig. 3c). These infiltrates were evenly distributed in the tumors as representative microphotographs in Fig. 3d show.Fig3Fig. 3

BO-112 intratumoral injection enhances T lymphocyte infiltrates. a. Schematic representation of the experiments to surgically harvest tumors following treatment to generate cell suspensions that were analyzed by flow cytometry. b. CD8/CD4 and CD8/Treg ratios in cell suspensions. c. Percentage of CD8⁺, CD4⁺ and CD25⁺FOXP3⁺ over total intratumoral CD45⁺ leukocytes and absolute numbers per gram of tumor tissue. d. Representative microphotographs of CD4 and CD8 immunohistochemistry analyses of sections derived from B16-OVA tumors treated as indicated. Scale bar of the main microphotograph: 100 μm. Scale bar of the inset: 60 μm. Positive cells are stained in magenta. *P < 0.05, **P < 0.01***P < 0.001

The intratumoral route in all detectable tumors might be possible in oligometastatic patients, but impossible in most advanced cancer patients. Furthermore, microscopic metastatic disease will not be amenable to intratumoral treatment. Therefore, we performed experiments in B16-OVA tumor-bearing mice in which only one of the subcutaneously engrafted tumors was treated (Fig. 4a). These mice were intraperitoneally co-treated with control antibody or immunomodulatory mAbs agonistic for CD137 [43] or antagonistic for PD-L1 [44]. As seen in Fig. 4b and c, BO-112 exerted clear local control of the disease as compared to vehicle. Such local control was enhanced to some extent by both anti-CD137 and anti-PD-L1 mAbs, which did not show any meaningful therapeutic activity by themselves, as shown in their combination with intratumoral control vehicle.Fig4Fig. 4

Immunotherapeutic effects of combinations of intratumoral BO-112 with systemic anti-CD137 or anti-PD-L1 monoclonal antibodies. a. Schematic representation of experiments in mice bearing two B16-OVA-derived tumors engrafted on opposite flanks and intratumorally treated with BO-112 only in the right lesion and with intraperitoneal administrations of immunomodulatory monoclonal antibodies as indicated. b. Tumor volume follow-up of the injected and distant tumors in the different groups of treatment. c. Mean ± SD summary indicating statistical significance of the listed comparisons. *P < 0.05, **P < 0.01***P < 0.001

Intratumoral release not only reaches the tumor microenvironment itself, but also drains to lymph nodes. In our hands, two intratumoral injections of BO-112 at therapeutic doses (Fig. 5a) in B16-OVA-tumor bearing mice resulted in prominent draining lymph node enlargements (Fig. 5b), as a result of an enhanced contents of CD45⁺ leukocytes. Again, CD8⁺/ CD4⁺ and CD8⁺/Treg ratios were markedly increased by treatment in most tumors, while non-draining lymph nodes remained normal (data not shown). Of interest, while numbers of effector CD8⁺ and CD4⁺ T cells rose, Tregs remained stable (Fig. 5c).Fig5Fig. 5

BO-112 intratumoral injection induces tumor-draining lymph node enlargement and increases CD8 T cells recognizing specific antigens. a. Scheme of experimental treatment showing representative size of TDLN and their total leukocyte content in the graph comparing mice treated intratumorally with BO-112 or control vehicle. b and c.: Analysis by flow cytometry of individual TDLN cell suspensions. b. CD8 to CD4 ratios and CD8/Treg ratios. c. represents the absolute number of the indicated T-cell subsets in TDLNs. d. Class I MHC tetramer stainings to identify T cells recognizing OVA-specific epitope and TRP-2 among CD8 T cells per gram of malignant tissue in mice bearing B16-OVA tumors. e Class I MHC tetramer stainings to identify the numbers OVA- and TRP2-specific CD8+ T cells in TDLN. Absolute numbers are provided for antigen-specific CD8 T cells. *P < 0.05, **P < 0.01***P < 0.001

Results from tumor infiltrates suggested that an important component in the therapy was mediated by the immune system. To determine which cell subsets were involved in the BO-112 tumor growth delay, depletion experiments were performed in MC38 and B16-OVA tumors. Antitumor effects on MC38-derived tumors completely disappeared when CD8_β ⁺ T cells were depleted in vivo, while CD4+ T cells were dispensable (Additional file 9: Figure S8A and B). By contrast, single CD8 subset depletion induced only a partial loss of efficacy and single CD4+ and NK+ subset depletion had not impact in BO-112-mediated antitumor response in B16-OVA mouse models. (Additional file 10: Figure S9 A and B). Interestingly, triple depletion of NK1.1+, CD4+ and CD8+ had a major impact on BO-112 therapeutic effects (Additional file 10: Figure S9 B), although such loss of efficacy was not complete, indicating that other mechanisms are involved in BO-112-induced antitumor response. In addition, myeloid Gr-1+ cells were dispensable (Additional file 10: Figure S11 A and B). Depletions were verified in peripheral blood by flow cytometry (Additional file 9: Figure S8C and Additional file 10: Figure S9 C and D), and Gr-1 depletion was additionally verified in both blood and tumor (Additional file 11: Figure S10 C and D).

The immunotherapeutic effects mediated by CD8⁺ T cells are usually dependent on IFNγ and the critical contribution of this cytokine to intratumoral BO-112 efficacy was revealed in BO-112-treated B16-F10 tumors when IFNγ was systemically neutralized with a specific mAb (Additional file 11: Figure S10E).

Various tumor specificities of CD8⁺ T lymphocytes in TDLN and in the tumor microenvironment can be monitored with H-2k^b tetramers refolded with well-studied immunodominant CTL epitopes such as Ovalbumin (OVA) and TRP2 in B16-OVA, and gp70 in the MC38 model. In this regard, injections of BO-112 into B16-OVA tumors resulted in increased contents CD8⁺ T cells recognizing TRP-2 and OVA as a surrogate tumor antigen in tumors (Fig. 5d) and in TDLN (Fig. 5e).

In line with these findings in B16-OVA models, a remarkable increase of gp70-specific intratumor CD8+ T cells was found in MC38-tumor bearing mice treated intratumorally with BO-112 (Additional file 12 Figure 11).

Previous reports have linked Poly I:C delivery to IFNα/β release [45, 46]. We genome-wide analyzed the mRNAs expressed in B16-OVA tumors 48 h following the second intratumoral BO-112 administration in comparison with control Vehicle (Fig. 6a and Additional file 13: Figure S12) and in B16-OVA cell cultures following 24 h incubation with BO-112 of Vehicle (Additional file 14: Figure S13).Fig6Fig. 6

Intratumoral BO-112 induces potent type-I IFN-related transcriptomic changes. a. Mice bearing B16-OVA tumors were treated with intratumoral BO-112 or vehicle (n = 5 per group) and total RNA was extracted as indicated to be genome-wide analyzed by gene expression microarrays. Differentially expressed transcripts were obtained by Linear Models for Microarray Data (LIMMA) analysis (b). Hierarchical clustering of differentially expressed genes between both experimental conditions. Most relevant genes for immune functions are indicated as upregulated by BO-112. c. Top canonical pathways upregulated by BO-112 treatment as defined by Ingenuity Pathway Analysis of the differentially expressed transcripts. d. Heat map representing enrichment analyses of key previously described signatures for IFNα and IFNγ stimulation, for tumor cell infiltration and activation of TILs as well as T-cell effector-related transcripts