The SK-MEL-147 human melanoma cell line was authenticated by analysis of short tandem repeats using GenePrint 10 (Promega, Internal Standard-ILS 600, performed by the Rede Premium Core Facility, FMUSP) and tested negative for mycoplasma by a PCR assay using conditioned medium as template and amplification using the following oligonucleotides:

This cell line as well as HEK293 were cultured in DMEM with 10% fetal calf serum, supplemented with antibiotic-antimycotic (Thermo Fisher Scientific, Waltham, MA, USA) and maintained at 37°C and 5% CO₂ atmosphere.

The strategy for constructing the adenoviral vectors has been described previously (21).

For the generation of the recombinant adenovirus we first constructed the “pEntr-PG” vector containing transgenes of interest: i) Luc2, used as control, ii) Luc2-p14ARF, and iii) Luc2-hIFNβ ( Figure S1 ). Next, site directed recombination was performed with the “destiny” vector encoding the Ad5 backbone (non-replicating, E1/E3 deleted, RGD modified fiber) utilizing Gateway L/R Clonase II Enzyme (Life Technologies, Carlsbad, CA, USA) as previously described (21, 34), giving rise to AdRGD-PG-Luc2, AdRGD-PG-Luc2-p14ARF, and AdRGD-PG-Luc2-hIFNβ. Following viral amplification, purification was performed using an iodixanol gradient followed by desalting, as described by Peng et al. (35) and as per our previous studies (21, 36). For the determination of biological titer, we used the Adeno-X Rapid Titer Kit (Clontech, Mountain View, CA, USA) which is based on immunodetection of the adenoviral hexon protein in transduced cells. The biological titer (transducing units per milliliter, TU/ml) was used for the calculation of the multiplicity of infection (MOI) indicated in each experiment.

SK-MEL-147 cells were seeded in medium containing 2% FBS together with the corresponding vectors at a final MOI of 50: i) AdRGD-PG-Luc2, ii) AdRGD-PG-Luc2-p14ARF, and iii) AdRGD-PG-Luc2-hIFNβ and iv) combination of AdRGD-PG-Luc2-p14ARF and AdRGD-PG-Luc2-hIFNβ (p14ARF+IFNβ, MOI 25 each). After 4 h incubation at 37°C, an equal volume of DMEM/10% FBS was added and the cells incubated for 24 and 48 h, unless otherwise noted. The described strategy was employed to guarantee the consistent number of viral particles for each treatment condition.

For validation of luciferase transgene activity, we used the Luciferase Assay System kit as per the manufacturer’s protocol (Promega, Madison, WI, USA). Sample luminescence was read on a 1420 Multilabel Counter VICTOR3 (Perkin Elmer, Waltham, MA, USA). The luminescence results were normalized by protein concentration. Interferon-β detection was performed using the Verikine human IFN beta ELISA kit (PBL Assay Science, Piscataway, NJ, USA) and conditioned medium derived from transduced cell cultures. For p14ARF, we use immunofluorescence and Western blot as previously described (37–39), using anti-p14ARF antibody (Santa Cruz Biotechnology, Dallas, TX, USA). Immunofluorescence detection was obtained using EVOS FL Cell Imaging System (Thermo Fisher Scientific).

Cells were transduced and incubated as described before harvest of both detached and adherent cells, fixation in 70% ethanol, RNAse treatment and staining with PI (21). Alternatively, fresh cells were treated with the LIVE/DEAD ™ Fixable Green Dead Cell Stain Kit, for 488 nm excitation (Thermo Fisher Scientific, #L23101) following the manufacturer’s recommendations. For either assay, cells were subjected to flow cytometry and analysis using the manufacturer’s software (Attune™, Thermo Fisher Scientific). For measurement of caspase activity, transduced cells were collected and treated with Cell Event Caspase 3/7 Green Detection reagent (Thermo Fisher Scientific, Cat. No. C10423) diluted and incubated according to the recommendations of the manufacturer. The proportion of fluorescent cells resulting from activation of caspases 3 and 7 was assessed by flow cytometry (Attune™, Thermo Fisher Scientific).

All animal experimentation and the protocols described here were approved by the Ethics in Animal Use Committee (FMUSP) and performed at the Centro de Medicina Nuclear (CMN), FMUSP, São Paulo, Brazil. Assays were performed using 8–12 weeks old female Nod-Scid mice obtained from the Bioterio Central, FMUSP. After local trichotomy, SK-MEL-147 cells (1 × 10⁶) were implanted subcutaneously in the left flank and animals were observed until tumors reached 60 mm³ and treatment was initiated. Intratumoral injection was performed on four occasions at 48-h intervals (days 1, 3, 5, and 7), applying 2 × 10⁸ TU in 50 µl of 1× PBS as excipient. Tumors were treated with PBS, AdRGD-PG-Luc2 or AdRGD-PG-Luc2-IFNβ, AdRGD-PG-Luc2-p14ARF, and combination of p14ARF+IFNβ. The number of animals/group is indicated in the figure legends. Tumors were measured with calipers and the volume calculated according to the formula: ½ × (length) × (width)² (39, 40). We consider the tumor volume of 1000 mm³ as experimental end point when the animals were anesthetized in a chamber with 4% isoflurane and euthanized by CO₂ inhalation. Otherwise, animals that did not reach the end point were monitored until 60 days. Distribution of animals/cage was maintained according to the pre-defined experimental groups at the beginning of the treatment. Alternatively, on experimental day 9 (48 h after the last treatment), some of these animals were euthanized and samples collected for histologic analyses.

The ICD markers were analyzed following the protocols previously described (21). For each assay, 1 × 10⁵ cells were treated with the vectors for 48 h, and then, cells and supernatant were collected. For the evaluation of calreticulin exposure through flow cytometry, cells were probed with rabbit anti-calreticulin antibody (Novus Biologicals, Littleton, CO, USA), followed by the Alexa488-conjugated anti-rabbit secondary antibody (Thermo Fisher Scientific). The cells were then analyzed using the Attune cytometer (Thermo Fisher Scientific). The ATP secreted in the media was evaluated using the ENLITEN ATP Assay System (Promega), following the manufacturer protocol, and the luminescence obtained with tube reader Sirius L Tube Luminometer (Titertek Berthold, Germany).

All experiments were performed after the approval of the institutional Committee for Ethics in Research (Fundação Pró-Sangue, CEP#03, FMUSP). Healthy donors’ peripheral blood was obtained from leukoreduction chambers (41) after signed written informed consents. PBMCs were isolated by density gradient centrifugation over Ficoll-Paque (GE Healthcare, Uppsala, Sweden). For the generation of monocyte-derived DCs, PBMCs were either plated for 2 h for adherence of monocytes to the plastic and subsequent removal of non-adherent cells (42) or by positive magnetic selection (Milteny Biotec, Bergisch Gladbach, Germany) for the isolation of CD14⁺ monocytes. The monocytes were cultured at 37°C and 5% CO₂ in RPMI-1640 medium supplemented with 10% FBS, antibiotic-antimycotic (Thermo Fisher Scientific) and 50 ng/ml of IL-4 plus 50 ng/ml of GM-CSF (both from PeproTech, Rocky Hill, NJ, USA) for 5 days to obtain immature monocyte-derived dendritic cells (iDCs) (43). At day 5, iDCs cells were harvested, washed, counted, and activated for 24 h with previously transduced SK-MEL-147 cells at a 1:1 and 1:10 ratios. Cells were harvested and analyzed by flow cytometry using CD209 (DCN46, #551545, BD), HLA-DR (G46-6, #556643, BD), CD80 (L307.4, #340294, BD), CD83 (HB15e, #561132, BD), CD86 (2331, #561124, BD) or used for the priming of autologous T cells.

T cells were enriched by either recovery of non-adherent cells after 2 h of PBMCs plastic adherence or by negative magnetic selection with the Pan T cell isolation kit (Milteny Biotec). Aliquots of T cells were cryopreserved until used in subsequent experiments. For T cell priming, activated Mo-DCs were harvest, washed, counted, and co-cultured for 7 days with autologous T cells at a 1:10 ratio in 96-well U-bottom plates (Corning, Tewksbury, MA, USA), with human IL-7 (50ng/ml), IL-2 (250 U/ml), and IL-15 (5 ng/ml) (all from Peprotech) with media replacement with fresh cytokines every 3 days. T cell proliferation was determined by carboxyfluorescein succinimidyl ester (CellTrace™ CFSE Cell Proliferation Kit, #C34554, Thermo Fisher Scientific) dilution, where T cells were previously stained with 5 μM CFSE. After 7 days, T cells were harvested and stained for surface markers CD3 (SK7, #340542, BD), PD1 (EH12.1, #561273, BD), TIM3 (F38-2E2, #25-3109-42, Thermo Fisher Scientific), LAG3 (3DS223H, #12-2239-41, Thermo Fisher Scientific), and CFSE dilution was determined by flow cytometry.

Primed T cells were seeded in 96-well U-bottom plates together with fresh non-transduced SK-MEL-147 in a 10: 1 ratio. After 48 h of incubation, SK-MEL-147 viability was assessed through LIVE/DEAD staining (Thermo Fisher Scientific) by flow cytometry. SK-MEL-147 cells were gated based on size (FSC^hi) and absence of CD3 staining.

To determine the cytokine profiles of stimulated T cells, we use co-cultures with allogeneic Mo-DCs primed with adenoviral-transduced SK-MEL-147 cells. Supernatants were collected after 3 days of cultures and analyzed by CBA. Analysis of IL-2, IL-4, IL-6, IL-10, IL-17A, TNF-α, and IFN-γ levels was performed using the Cytokine Bead Array Th1/Th2/Th17 kit (BD Biosciences, San Jose, CA, USA). Samples were acquired with a FACS II LRS Fortessa (Becton, Dickinson and Company, USA) according to the manufacturer’s instructions and analyzed using the FCAP Array Software 3.0 (BD Biosciences, San Jose, CA, USA).

Statistical analyses were performed using GraphPad Prism software version 5.0 (GraphPad Software, San Diego, CA, USA). Statistical significance was calculated using one-way or two-way ANOVA when appropriate, followed by Bonferroni post-test. Differences were considered significant when p < 0.05. Number of replicates and sample sizes for all experiments are detailed in the figure legends.

Adenoviral vectors encoding Luc2, p14ARF or IFNβ were constructed ( Figure S1 ) and transgene expression was validated ( Figures S2 and S3 ). In this work, we focus on the SK-MEL-147 human melanoma cell line since it harbors wild-type p53. Subsequently, we measured hypodiploid cell populations after transduction, revealing that the combination of adenoviruses p14ARF+IFNβwas superior to individual gene transfer for the induction of hypodiploidy ( Figure 1A ). Figure 1B shows the increase hypodiploidy over time, reaching 45% of cells in 48 h.

Next, we assessed the alteration of cell membrane permeability in a live/dead assay (L/D) and also measured caspase 3/7 activity 48 h after transduction. Figure 1C shows that the number of live cells (Caspase 3/7^neg, L/D^neg) was significantly reduced in the groups that received IFNβ, p14ARF or the combination as compared to the Mock or Luc control groups. Representative dot-blot graphs are shown in Figure S4 . Apoptotic cells, those positive for caspase 3/7 activity (Casp3/7⁺) and with cell membrane integrity (L/D^neg), were more frequent upon treatment with p14ARF ( Figure 1D ). In contrast, combined gene transfer was associated with necrosis (Casp3/7^neg, L/D⁺), indicating loss of cell membrane integrity even without caspase 3/7 activity ( Figure 1E ). Caspase activity with concomitant loss of cell membrane integrity (Casp 3/7⁺, L/D⁺) was associated with IFNβ treatment alone or in combination with p14ARF ( Figure 1F ). These data indicate that treatment with p14ARF+IFNβ may induce cell death by a mechanism that is independent of caspase 3/7 activity.

After seeing that adenovirus carrying the p14ARF and IFNβ transgenes triggered cell death in human melanoma cells in vitro, we questioned whether this potential could also be observed in vivo. For this, tumors were established by injection (s.c.) of 1 × 10⁶ cells in the right flank of Nod-Scid mice. Once the tumors had reached a volume of 60 mm³, we initiated the treatment regimen consisting of four doses of adenovirus (2 × 10⁸ TU per dose) administered by intratumoral injections with 48-h intervals between doses. The first day of treatment is considered as day 1 (thus virus was administered on days 1, 3, 5, and 7). The experimental endpoint in the PBS and Luc controls was reached around the tenth day of treatment, while IFNβ alone or in combination with p14ARF show significantly reduced tumor growth ( Figure 2A ) and prolonged survival ( Figure 2B ). We noticed an intermediate effect when p14ARF was used alone. While IFNβ was sufficient to inhibit tumor progression, we postulate that the benefit of p14ARF may lie in the mechanism of cell killing and the impact of oncolysis on the host response.

Having verified the potential to generate oncolysis, we investigated whether treatment with the adenoviral vectors would trigger ICD, measured by the emission of DAMPs and activation of cellular mediators of the immune response. For this, SK-MEL-147 cells were transduced with the vectors encoding p14ARF, IFNβ or the combination and then incubated for 48 h before observing the exposure of calreticulin in the cell membrane and the secretion of ATP. Figure 3A shows that the increase in calreticulin exposure in the cell membrane of transduced cells was greatest in response to the IFNβ and p14ARF combination. This condition also resulted in increased granularity, indicating cellular stress. Figure 3B shows that secretion of ATP is most intensely induced by the p14ARF+IFNβ combination. The production of IFN-β, another DAMP, is also supported by our gene transfer approach ( Figure 3C ). Thus, combined gene transfer of p14ARF and IFNβ resulted in the emission of three critical DAMPs expected to contribute to immune cell activation.

To further investigate whether the induction of cell death and immunogenic factors upon transduction with the p14ARF+IFNβ adenoviral vectors could elicit an immune response, we exposed immature monocyte-derived dendritic cells (iDCs) to previously transduced SK-MEL-147 cells ( Figure 4A ). Combined p14ARF+IFNβ gene transfer was shown to be especially efficient for the release of ICD markers, thus was used in these assays. First, we verified that the viability of iDCs was not lost when co-cultured at different ratios (1:1 and 1:10) with SK-MEL-147 cells previously transduced at MOI 10, 50, or 100 ( Figure S5 ). By gating on the live CD209⁺ iDCS, we are able to demonstrate that the frequency and expression level of HLA-DR, CD80, and CD86 increased in iDCs exposed to p14ARF+IFNβ-transduced SK-MEL-147 cells compared to iDCs exposed to mock-transduced SK-MEL-147 cells (representative histograms presented in Figure 4B ). Statistically, iDCs behaved the same when treated at a ratio of 1:1 (black dots) or 5:1 (gray dots) iDC:SK-MEL-147 transduced cells, thus these results were pooled. When co-cultured in direct contact with iDC, p14ARF+IFNβ-transduced SK-MEL-147 cells induced an increase in expression of HLA-DR, CD80, and CD86 in iDCs, while co-culture in a transwell chamber revealed that soluble factors induced the upregulation of only HLA-DR in iDCs ( Figure 4C ). This suggests that membrane-associated factors presented by the transduced SK-MEL-147 cells are responsible for most of their dendritic cell activation effect.

Next, we explored whether iDCs exposed to transduced SK-MEL-147 cells were able to prime T cells and polarize cytokine production ( Figure 5A ). Using an autologous system supported by IL-15, IL-7, and IL-2, we found that iDCs activated with Luc and p14ARF+IFNβ-transduced SK-MEL-147 cells were able to switch the CD4/CD8 ratio, favoring the proliferation of CD8⁺ T cells rather than CD4⁺ ( Figure 5B ). Compared to unstimulated CD8⁺ T cells, iDCs activated with mock-transduced, Luc-transduced and p14ARF+IFNβ-transduced SK-MEL-147 cells favored a non-significant upregulation of the checkpoint molecules LAG-3, PD-1, and TIM-3 among the CD8⁺ T cell population ( Figure 5C ). To characterize the cytokine milieu induced during antigen presentation and co-stimulation, we co-cultured activated DCs with allogeneic T cells to avoid the interference of external cytokines needed in the autologous system. We found that iDCs activated with p14ARF+IFNβ-transduced SK-MEL-147 cells potently induced the production of IFN-γ by T cells, compared to iDCs activated with Luc-transduced SK-MEL-147 cells; TNF-α and IL-10 were also significantly increased, however at lower concentrations than IFN-γ. In turn, IL-6 was upregulated in T cells co-cultures with both, iDCs activated with Luc and p14ARF+IFNβ-transduced SK-MEL-147 cells. When iDCs were activated with p14ARF+IFNβ-transduced SK-MEL-147 cells separated through a transwell membrane, they lost the ability to induce cytokine production by T cells. No significant differences were found when iDCs were activated at a 1:1 (black) or 5:1 (gray) ratio with p14ARF+IFNβ-transduced SK-MEL-147 cells ( Figure 5D ).

Finally, to explore whether iDCs activated with adenovirus-transduced SK-MEL-147 cells were able to induce a specific cytotoxic population, autologous primed T cells from two healthy donors were challenged with fresh non-transduced SK-MEL-147 cells. After 48 h, we assessed the abundance of T cells in samples and the viability of CD3^neg cells within the SK-MEL-147 gate (higher FSC). We observed that within the lymphocyte gate, more than 90% of the cells were CD3⁺ ( Figure 6A ). For both donors, we found higher viability loss among fresh SK-MEL-147 cells when challenged with T cells primed by iDCs activated with p14ARF+IFNβ-transduced SK-MEL-147 cells, compared to all controls ( Figure 6B ). Based on these findings, we propose in Figure 6C the mechanism of immune activation in response to treatment of SK-MEL-147 cells with the combination of p14ARF+IFNβ gene transfer.