Dulbecco’s Modified Eagle’s Medium (DMEM), Roswell Park Memorial Institute 1640 medium (RPMI 1640), fetal bovine serum (FBS), penicillin–streptomycin, and glutamine were purchased from Gibco (Thermo Fisher, Germany). Trypsin-EDTA solution, adenosine, inosine, AMP, ADP, and ATP, the A2b (alloxazine) and A3 adenosine receptors (VUF5574), and caffeine, a non-selective adenosine antagonist, were purchased from Sigma-Aldrich (Germany). The antagonists for A1 (DPCPX) and A2a (SCH-58261) adenosine receptors were purchased from Tocris, UK.

The C57Bl/6 mouse-derived melanoma cell line B16F10 was purchased from ATCC (#CRL-6475) and propagated in DMEM supplemented with 10% FBS and penicillin/streptomycin (D10) at 37°C in a 5% CO₂ atmosphere. NIH/3T3 fibroblast cell line was purchased from ATCC (#CRL-1648) and cultured in RPMI 1640 supplemented with 10% FBS, streptomycin/penicillin, and glutamine (R10). Human synovial tissue samples were obtained from knee joints of patients with rheumatoid arthritis from the orthopedic rheumatology unit of the Waldkrankenhaus St. Marienin Erlangen. An informed consent was obtained from patients, and their use was approved by the local ethics committee (Permit # 52_14B_3). Human fibroblast-like synoviocytes (FLS) were dissected by cutting off the villi of the synovial membrane. The tissue was digested using Collagenase IV solution (Sigma) in a shaking thermomixer at 37°C in two steps for 45 min. The samples were vortexed vigorously to release the cells. The collected cells were allowed to adhere to culture flasks for 2 days, with addition of fresh medium every day. Then, complete medium was removed together with non-attached cells, and cells were washed rigorously. Adherent cells were a mixture of two major cell subtypes: type A macrophage-like and type B FLS. Short trypsinization steps of about 2 min at each passage allow detachment of only fibroblastic-like cells and thereby removal of the monocytic cells from the cell mixture. The terminally differentiated macrophages have a limited life span in vitro, which contributes to the formation of a relatively clean pool of FLS. The isolated FLS were expanded in culture flasks for three passages by splitting them in a 1:2 ratio at confluency. The cells were used for proliferation assays in passages 3–6 in R10 medium supplemented with 1% fungizone (Gibco). Viable cells were seeded in 24-well plates at very low densities (200–1,000 cells/well) in the presence of dead and dying cells or their SNs. Control wells containing only 200 B16F10 cells in D10, 1,000 FLS in R10, or in indicated SN were included in the same plate during one experiment. Agonists, antagonists, and vehicle were added as indicated. Transwell experiments were performed with ThinCerts™ cell culture inserts from Greiner Bio-One employing a 0.4 µm pore size membrane on 24-well plates (Greiner Bio-One, Germany).

Cell death was induced by irradiation with ultraviolet light type B (UVB) at 1.5 mJ/cm²/s or by heat shock (56°C for 30 min). The phenotype of cells after cell death induction was confirmed by a morphophysiological classification by flow cytometry (25) as previously shown [see Figures 2E,F in Ref. (13)]. At the indicated time points, SNs were collected, centrifuged at 6,000 g for 10 min, stored at −70°C until further use, and thawed only once. When necessary, the SNs were boiled on water bath for 15 min for deproteinization. A further protein-free fraction was obtained by filtration through Amicon^® (Millipore, Germany) filters with a 3 kDa size cutoff membrane.

High-performance liquid chromatography and size exclusion chromatography (HPLC–SEC) were performed with a Perkin Elmer Series 200 HPLC system using Strong Cation Exchange (SCX) column purchased from Shiseido CAPCELL PAK SCX UG 5 µm 150 mm × 1.5 mm. The protocol for detection of adenosine was developed based on manufacturer’s recommendations. The data was recorded for 10 pps and at 254 nm on column data with a 50 mM potassium dihydrogen orthophosphate (KH₂PO₄)–dipotassium hydrogen orthophosphate (K₂HPO₄) buffer (pH = 2.6) as a mobile phase and a flow rate of 0.5 ml/min (~860 psi) detection (26). The quantification of the concentration of purinergic metabolites was performed by liquid chromatography coupled with mass spectrometry.

Cell cultures were harvested at the indicated time points by collecting the medium containing dead and spontaneously detached cell together with adherent cells after treatment with trypsin-EDTA solution for 20 min. Cell growth was quantified at different time points by flow cytometry employing a Gallios flow cytometer (Beckman-Coulter, Miami, FL, USA). Only viable cells excluding propidium iodide were recorded. Since the harvesting procedure required prolonged incubation and pipetting steps, an alternative colorimetric method for the quantification of growth was established for experiment requiring multiple simultaneous harvesting (27). Briefly, dead and non-adherent cells were washed out with warm PBS; adherent cells were fixed for 30 min in a solution of glutaraldehyde (1% in PBS), washed with PBS (pH 7.4), and subsequently stained with a 0.01% crystal violet solution (in 20% methanol). After removing excess of dye, the crystal violet-stained cells were dissolved in 1 ml of a 10% sodium dodecyl sulfate solution, and the optical density of the extracted dye was read in a spectrophotometer at 590 nm (27).

Wild-type C57Bl/6 mice were bred at the Institute of Cell Biology, Lviv, Ukraine and kept on a standard diet with drinking water available ad libitum. A total of 5 × 10⁵ viable cells in Ringer’s solution, with adenosine or with SNs of dying cells (four mice per group), were implanted subcutaneously on the back of each mouse. Tumor volume was measured using a caliper from the day of appearance until the end of the experiment. Animal studies were conducted according to European principles and local guidelines for the care and use of laboratory animals. Experiments were performed on mice matched for age and sex and evaluated with blinded identity.

Data are reported as mean ± standard error of replicate mean values. The software package GraphPad Prism 5.0 was used for graphics and statistical tests. For comparisons between control and experimental groups, Mann–Whitney U test or two-way ANOVA tests were employed as appropriate. Statistical significance was assumed if p < 0.05.

In order to simulate the conditions during cancer treatment where the vast majority of tumor cells are killed by irradiation and only a few survive, we evaluated the growth of few viable B16F10 cells (200 cells/well) in the presence of higher quantities of dying cells killed by UVB irradiation (apoptosis induction). Cultures were kept without adding fresh medium during the whole indicated culture period. Dying cells alone and viable cells alone were cultured simultaneously in separate wells as controls. We observed that cocultures of viable cells with dying cells by apoptosis supported rapid growth of viable cells seeded at very low densities. Seeding cells at these densities in D10 medium only showed no proliferation during the same period of time (Figure 1A). This effect was more pronounced when lower numbers of dying cells reach a maximum at the viable/dead cell ratio of 1:100 (Figure 1B). When the number of dead cells was further increased, the promotion of proliferation was completely abolished (Figure 1B). In order to explore whether physical contact between viable and dying cells is necessary to trigger cell growth enhancement, viable B16F10 cells were cultured in transwell plates (ThinCert Greiner, Germany), which allow physical separation of cells by a 0.4 µm pore membrane. Viable cells were plated in the upper side of the transwell, and 100 times more dying cells were placed in the lower compartment. The promotion of proliferation was also achieved when dying cells were separated from the viable ones (Figure 1C), indicating that soluble stimulating factors released by dead and dying cells concur to promote melanoma cell proliferation.

In order to elucidate whether the promotion of proliferation was specific for the melanoma cell line, we also employed NIH/3T3 mouse fibroblasts cultured with SNs from apoptotic homologous cells. Employing the staining method for adherent cells we quantified proliferation more efficiently as with the flow cytometric method and observed that the conditioned medium from irradiated fibroblast significantly stimulated the proliferation compared to fibroblasts cultured in R10 medium alone (Figure 1D). Furthermore, the source of the stimulatory potential and the proliferating target cell were not limited to immortalized cells. Primary FLS from patients with rheumatoid arthritis were also stimulated to proliferate by SNs of apoptotic neutrophils (Figures 1E,F).

The analysis of SNs collected at different time points after the induction of cell death showed that cells dying by apoptosis after irradiation released the stimulating factors after 12 h of culture and lasted for 36 h more (Figure 2A). In contrast, heat shock treatment resulted in immediate necrosis and release of proliferation-stimulating factors (Figure 2B). In order to investigate the nature of the released factors, we boiled SN collected after 24 h of irradiation and after 8 h of heat shock treatment and compared the stimulating capacity with that of untreated SNs and fresh D10 medium. We found that boiling only partially reduced the stimulating capacity of the SNs from both apoptotic (Figure 2C) and necrotic cells (Figure 2D). These results suggest that proteins released after cell death are unlikely to be responsible for the stimulation of proliferation in melanoma cells. We confirmed this hypothesis when we prepared the protein-free and -rich fractions by filtration through a 3 kDa membrane of SNs and medium. We observed that the proliferation-stimulating activity was present only in the protein-free fraction of dead cell SNs (Figures 2E,F). Interestingly, cells proliferated even in the absence of high molecular weight nutrients, which rather represents a starving condition.

One of the most important non-proteic metabolites released is the nucleotide ATP. Therefore, we set up a HPLC–SEC analysis of the composition of protein-free fractions of SNs from dying cells contrasted with the standards ATP, adenosine, and inosine. The runs for two samples of SNs and fresh medium are shown in Figure 3A, upper panel. The purified standards are shown in the lower panel (Figure 3A). The chromatographic separation of the metabolites allows speculating about ATP and inosine being present in the SNs of dying cells. The same is not true for adenosine. The mass spectrometric analysis confirmed the presence of inosine (10 nM) or its metabolite hypoxanthine (10 nM) in SNs of apoptotic and necrotic cells, respectively (Figure 3B). Adenosine was detectable but at much lower concentration (<0.2 μM; Figure 3B). Adenosine in the body is produced by the action of ectonucleotidases from nucleotides like ATP, ADP, and AMP. Adenosine is either immediately transported to the intracellular space or is degraded by the action of adenosine deaminase into inosine (28). The phosphorylation of the ribose from the inosine results in the generation of hypoxanthine and ribose-1-PO₄ (29). We confirmed that all these steps of the purinergic metabolism take place spontaneously in cell-free SNs. The addition of 500 µM purified adenosine to cell-free SNs from dying cells resulted in the clearance of adenosine and the production of inosine and hypoxanthine (Figure 3C). These results confirm that nucleotides are released by dying cells and are immediately catabolized by ectonucleotidases present in the culture into adenosine and inosine.

After confirming the presence of inosine and ATP as major purinergic metabolites in the SN of both apoptotic and necrotic cells, we investigated the relative potency of the purified nucleotides ATP, ADP, and AMP (Figure 4A) and nucleosides adenosine and inosine (Figure 4B) after 9 days of culture of melanoma cells without feeding. Nucleotides begin to exert stimulating activity at 10 µM concentrations (Figure 4A), whereas nucleosides stimulate proliferation at 50 µM concentration (Figure 4B). Both types of metabolites exhibited a bell-shaped dose–response curve similar to that observed for dead/viable cells cocultures (Figure 1B).

In order to investigate whether adenosine receptors participate in the stimulation of proliferation induced by purinergic metabolites, we added the pan-adenosine receptor antagonist caffeine and the specific adenosine receptor antagonists alloxazine (A2bR) and VUF (A3R) to adenosine-stimulated cultures and observed that the stimulation of proliferation most likely mediated by inosine was abolished in the presence of the antagonists for the adenosine receptors A2b and A3 and not by the antagonists of the receptors A1 and A2a (Figure 4C). Additionally, we confirmed the expression of the adenosine receptors A1, A2b, and A3 on the surface of viable B16F10 melanoma cells (Figure 4D).

The implantation of few (50,000) B16F10 cells in the subcutaneous space of the back of C5Bl/6 mice led to the late development of solid tumors. The first measurable tumor appeared at day 24 after implantation (Figure 5, black line). Coimplantation with either SNs from apoptotic cells or with 200 µM adenosine caused the development of larger tumors (Figure 5, colored lines).