Drug-Induced Senescent Multiple Myeloma Cells Elicit NK Cell Proliferation by Direct or Exosome-Mediated IL15 Trans-Presentation

Natural killer (NK) cells are innate lymphoid cells capable of killing malignant or infected cells by an integrated interplay of activating and inhibitory receptor signaling and by secreting a wide array of cytokines and chemokines (1–3). Ample evidence indicates NK cell involvement in the protection against cancer in humans and experimental animals, and a number of NK cell–based immunotherapeutic approaches have been exploited (4, 5). NK cells, originally identified as “naturally active,” are endowed with cytotoxic activity when incubated in vitro with tumor target cells and can be increased by addition of stimulatory signals, such as cytokines (6). IL15 is central to the development, survival, and activation of NK cells (2, 7–9) and was first identified by its functional similarities to IL2 in promoting T and NK cell proliferation and signaling through the same beta- and gamma-receptor subunits (10, 11). These subunits are important for signal transduction, whereas the alpha-receptor subunit is responsible for the specificity and high-affinity binding of IL15 to the receptor (12).

IL15 mRNA expression can be detected in a broad range of tissues, including adherent mononuclear blood cells, activated macrophages, epithelial and fibroblast cells, placenta, and skeletal muscle, with an expression pattern similar to IL15Rα (IL15RA) (12). However, IL15 is rarely secreted but can be under pathologic conditions (11, 13). In vivo, IL15 predominantly exists associated with the plasma membrane, which explains the difficulty in its detection as soluble protein. IL15 trans-presentation is a unique mechanism that stimulates IL15 signaling on neighboring cells through cell–cell interactions, mediating different functions compared with conventional soluble cytokine delivery (14). IL15 trans-presentation is critical to support NK cell development, survival, and activation (15–17), suggesting that IL15 presented in trans may play a major role in augmenting NK cell–mediated immunosurveillance (17).

NK cells have a well-established role in the clearance of a number of hematological malignancies, including multiple myeloma (MM), a plasma cell (PC) tumor that mainly develops in the bone marrow (BM). We previously demonstrated that treatment with sublethal doses of genotoxic drugs increases the expression of NKG2D and DNAM-1 activating ligands preferentially on chemotherapy-treated MM senescent cells, which are preferentially killed by NK cells and trigger IFNγ production (18–20). In a mouse model of MM, the genotoxic drug melphalan (MEL) in vivo promotes the establishment of a senescent tumor cell population that displays an increased expression of NK cell–activating ligands and becomes more susceptible to NK cell killing (21).

Senescence is a complex cellular program induced by genotoxic, replicative, or oncogenic stress, in which cell-cycle–arrested cells remain metabolically active and secrete several soluble factors, also known as the “senescence-associated secretory phenotype” (SASP), which mediates a variety of cellular responses, including modulation of cancer immune surveillance. Studies have shed a new light on the role of exosomes in mediating the cell-to-cell transmission of senescence signals, suggesting that exosomes may act as new components of the SASP (22).

Herein, we observed that drug-induced senescent MM cells displayed a higher IL15/IL15RA complex on the plasma membrane, thus enhancing activation and proliferation of primary NK cells. MM cell–derived exosomes also expressed IL15RA and had the capability to stimulate NK cell proliferation in the presence of exogenous IL15. At present, the role of IL15 on myeloma cells refers only to the existence of an autocrine IL15 loop promoting their survival (23). Overall, our data demonstrated that beyond displaying activating ligands and increased susceptibility to NK killing, drug-induced senescent MM cells also promoted NK cell activation and proliferation via direct or exosome-mediated IL15 trans-presentation, thus further potentiating NK cell antitumor effector functions.

SKO-007(J3), ARK, and RPMI8226 MM cell lines were provided by Prof. P. Trivedi (Sapienza University of Rome, Italy). After thawing, cells were cultured for no longer than 4 weeks and tested for mycoplasma monthly. SKO-007(J3) and RPMI8226 cell lines were authenticated by IRCCS Azienda Ospedaliera Universitaria San Martino-IST, S.S. Banca Biologica e Cell factory by STR profile (Cell IDTM System, Promega). Detection of amplified fragments was obtained by ABI PRISM 3100 Genetic Analyzer. Data analysis was performed by GeneMapper software, version 4.0.

BM aspirates of 88 untreated MM patients were managed at the Department of Cellular Biotechnologies and Hematology (Sapienza University of Rome). Informed and written consent in accordance with the Declaration of Helsinki was obtained from all patients, and approval was obtained from the Ethics Committee of the Sapienza University of Rome. No exclusion criteria were used. Patient characteristics were described in Supplementary Tables S1 and S2. The BM aspirates were lysed to obtain bone marrow mononuclear cells (BMMC), and CD138 MicroBeads kit (cat. 130-097-616; Miltenyi Biotec) were used to isolate malignant PCs as previously described (18). Patients’ sera were obtained by centrifugation (2,000 rpm) of whole blood for 15 minutes at room temperature, collected, and stored at −80°C before the assay.

Sublethal doses of doxorubicin (DOX) and MEL, determined by an MTT assay as previously described (18), were used to treat SKO-007(J3), ARK, and RPMI8226 cell lines, and patients’ PCs. Cell lines and patient-derived PCs were cultured for 48 hours at a density of 3 × 10⁵ cells/mL and 5 × 10⁵ cells/mL, respectively, and then washed and left for further 24 hours without drugs.

The expression of IL15 and IL15RA on MM cells was evaluated with unconjugated anti-IL15 (MAB2471) and anti-IL15RA (ΜΑΒ1471), respectively (R&D Systems). Allophycocyanin (APC)-conjugated Goat affinity purified F(ab’)₂ fragment to Mouse IgG (GAM) was purchased from Jackson ImmunoResearch Laboratories. APC-conjugated anti-CD69 (FN50) was from Biolegend. To detect IL15 and ILRA expression on the MM cell surface, 1 × 10⁶ of untreated and MEL/DOX-treated cells were incubated for 20 minutes at 4°C with unconjugated IL15 and IL15RA monoclonal antibodies, then washed and incubated for further 20 minutes with secondary GAM-APC. For the intracellular staining of IL15, 1 × 10⁶ of untreated and MEL/DOX-treated cells were fixed with Flow Cytometry Fixation Buffer (cat. FC004) (R&D Systems), permeabilized with Flow Cytometry Permeabilization Buffer (cat. FC005; R&D Systems), and incubated with unconjugated IL15 (MAB2471) for 30 minutes. Cells were then washed and incubated for 20 minutes with secondary GAM-APC for final flow cytometric analysis. In some experiments, cells were stained with propidium iodide (PI; 50 μg/mL) to assess cell viability. PE-conjugated anti-CD38 (HIT2) and FITC-conjugated anti-CD138 (MI15) antibodies were purchased from BD Biosciences. NK cells were purified using human NK cell isolation kit (Miltenyi Biotec) from healthy donor PBMCs, isolated through Lymphoprep (Stemcell Technologies). Cell population was routinely more than 90% CD56⁺ CD3⁻, as assessed by immunofluorescence and flow cytometry analysis. In some experiments, NK cells were fixed and permeabilized with 30% methanol plus 0.4% paraformaldehyde (PFA) in PBS for 30 minutes, washed, and then incubated in 0.05% Tween plus 1% PFA in PBS and stained with anti-Ki67-FITC (clone MIB-1; Dako). Cells were analyzed with a FACS Canto II (BD Biosciences). Flow cytometric analysis was performed using the FlowJo software version 8.8.7 (TreeStar).

Total RNA was isolated from MM cells (SKO-007(J3), ARK, RPMI8226) using TRIzol reagent (Invitrogen) according to the manufacturer's protocol, and 1 μg was used for cDNA first-strand synthesis in a 25 μL reaction volume (M-MLV reverse transcriptase, cat. M170A; Promega). Resulting cDNA (1 μL) was used in a 25 μL PCR reaction. IL15 (Hs00174106_m1) and IL15RA (Hs00542604_m1) mRNA expression was analyzed by real-time PCR using specific TaqMan Gene Expression Assays (Applied Biosystems). Relative expression of each gene versus β-actin was calculated according to the ΔΔCt method (24).

Drug-treated MM cells were lysed for 20 minutes at 4°C in 1× RIPA lysis buffer (1% NP-40, 0.1% SDS,50 mmol/L Tris-HCl pH 7.4, 150 mmol/L NaCl, 0.5% sodium deoxycholate, 1 mmol/L EDTA) plus complete protease inhibitor mixture used at 1× (cat. P2714) and phosphatase inhibitors sodium orthovanadate and sodium fluoride (Sigma-Aldrich). The BioRad Protein Assay (BioRad Laboratories) was used to measure protein concentration. Total lysates (50 μg) were resolved by SDS–PAGE and transferred with transfer buffer (25 mmol/L Tris/HCl, 20 mmol/L glycine, and 10% (v/v) methanol; ref. 25) to nitrocellulose membranes (Whatman GmbH). After blocking with BSA, membranes were probed with the following specific antibodies (1 μg/mL): β-actin (AC15; Sigma-Aldrich), IL15 (MAB2471), IL15RA (ΜΑΒ1471), HSP70 (SC24; Santa Cruz Biotechnology), and calreticulin (PA3-900; Thermo Fisher Scientific). A horseradish peroxidase (HRP)-conjugated secondary antibody (cat. NA931V and NA934V; GE Healthcare Life Sciences) and an enhanced chemiluminescence kit (cat. RPN2106; GE Healthcare) were used to reveal immunoreactivity.

Freshly isolated peripheral blood NK cells were purified as previously described, adjusted to 1 × 10⁶ cells/mL, and cocultured overnight (o.n.) with untreated or MEL-treated SKO-007(J3) cells at E:T ratio of 1:1. The day after, MM cells were removed from the culture by CD138⁺ magnetic beads positive selection (Miltenyi Biotec), and NK cells were incubated for a further 5 days at 37°C and 5% CO₂. For blocking experiments, MM cells were pretreated 1 hour with 1 μg/mL of IL15-neutralizing monoclonal antibody (mAb) (MAB2471; R&D Systems), then washed, and left o.n. with primary NK cells (E:T = 1:1). Five days later, NK cell proliferation was evaluated by using the BrdUrd flow kits (cat. 552598) according to the instruction manual (BD Biosciences). In some experiments, 1 × 10⁶ NK cells were incubated with exosomes (5–10 μg/mL) derived from untreated or drug-treated MM cells (as described before) in the presence or absence of human recombinant IL15 (5–50 ng/mL; cat. 200-15) or IL2 (200 U/mL; cat. 200-02; PeproTech), and cell proliferation was measured after 5 days by CFSE (cat. 150347-59-4; Sigma-Aldrich) assay (26) or evaluating the Ki67 (MIB-1; Dako) proliferation marker.

Exosome-free medium was obtained as follows: fetal bovine serum (FBS; Thermo Fisher Scientific) was centrifuged at 100,000 × g for 3 hours in a Beckman ultracentrifuge (Beckman Coulter) to remove microvesicle-like exosomes. RPMI 1640 was supplemented with 10% FCS-exosome-free medium and antibiotics (penicillin, streptomycin, glutamine used at 1X; Thermo Fisher Scientific). ARK and SKO-007(J3) MM cell lines were cultured at 0.8 × 10⁶ to 1 × 10⁶ cells/mL in exosome-free medium for 48 hours. Exosome purification consists of different sequential centrifugations as previously reported (27). Briefly, cells were harvested by centrifugation at 300 × g for 10 minutes and supernatants were collected. Cell-free supernatants were then centrifuged at 2,000 × g for 20 minutes, followed by centrifugation at 10,000 × g for 30 minutes to remove cells and debris. Supernatants were filtered using a 0.22 μm filter and centrifuged at 100,000 × g for 70 minutes at 4°C in a Beckman ultracentrifuge to pellet exosomes. The resulting pellet was washed in a large volume of cold PBS and again centrifuged at 100,000 × g for 70 minutes at 4°C. Finally, exosomes were resuspended in PBS for further analyses and functional studies.

TEM was performed as previously described (28). Briefly, exosomes were fixed in 2% PFA and adsorbed on formvar-carbon-coated copper grids. The grids were then incubated in 1% glutaraldehyde for 5 minutes, washed with deionized water 8 times, and then negatively stained with 2% uranyl oxalate (pH 7) for 5 minutes and methyl cellulose/uranyl for 10 minutes at 4°C. Excess methyl cellulose/uranyl was blotted off, and the grids were air-dried and observed with a TEM (Philips Morgagni268D) at an accelerating voltage of 80 kV. Digital images were taken with Mega View imaging software.

Detection of IL15 in sera or plasma of MM patients was performed using specific ELISA kits from R&D Systems. Plates were developed using a peroxidase substrate system (DuoSet ELISA Development Kit, DY247-05; R&D Systems), and then read with the Victor3 multilabel plate reader (Model # 1420-033; Perkin Elmer) capable of measuring absorbance in 96-well plates using dual wavelengths of 450 to 540 nm. Results were expressed as picograms per milliliter (pg/mL) and referred to a standard curve obtained by plotting the mean absorbance for each standard on the y-axis against the concentration on the x-axis and drawing the best-fit curve through the points on the graph. Detection of IL15 in the MM-conditioned supernatants was performed with a Milliplex^MAP Human Cytokine/Chemokine Magnetic Bead Panel, Immunology Multiplex Assay, according to the manufacturer's instructions (Millipore). Plates were read with Bio-Plex MAGPIX Multiplex Reader (Bio-Rad) and analyzed by Bio-Plex Manager MP software.

Error bars represent standard deviation (SD) or standard error of the mean (SEM). Data were evaluated by paired Student t tests, with the exception of ELISA data, which were analyzed by Mann–Whitney test. For statistical analysis, GraphPad Prism software was used. Statistical significance is indicated with the P values < 0.05.

We previously demonstrated that sublethal doses of genotoxic drugs, such as doxorubicin (DOX) and melphalan (MEL), induce a SASP in primary malignant PCs and SKO-007(J3) MM cells (18, 29). We then asked whether drug-induced MM cells could also exhibit a SASP phenotype and release soluble mediators capable of affecting NK cell functions and focused our attention on IL15, a cytokine involved in NK cell activation and proliferation/maturation. We evaluated IL15 mRNA expression by real-time PCR analysis on SKO-007(J3), ARK, and RPMI8226 MM cell lines treated with DOX and MEL for 24, 48, and 72 hours (48 hours of treatment plus 24 hours without drug). Our findings showed that upregulation of IL15 mRNA with both drug treatments was evident by 24 hours, with a peak at 48 hours (Fig. 1A). All these cell lines underwent senescence in response to genotoxic drugs (Supplementary Fig. S1 A; ref. 19). Intracellular staining on DOX- or MEL-treated MM cells after overnight incubation with brefeldin A revealed an increase in IL15 intracellular protein expression (Fig. 1B) that was confirmed by Western blot analysis (Fig. 1C). However, augmented IL15 mRNA and intracellular protein expression was not accompanied by a parallel increase in cytokine release, as determined by Luminex technology (value <4 pg/mL, lower detection limit).

The absence of detectable IL15 in the culture supernatants of MM cell lines suggested that IL15 could be trapped by the surface-expressed IL15RA and could signal in trans. Thus, IL15 surface expression was determined by immunofluorescence and flow cytometry, and we found that all the MM cells analyzed exposed this cytokine on the plasma membrane upon drug treatment (Fig. 1D). We then investigated the ability of genotoxic drugs to regulate the expression of IL15RA. Upon 72 hours of drug treatment of SKO-007(J3), ARK, and RPMI8226 MM cells, we found increased IL15RA protein levels (Fig. 2A) and a significant augmentation of its cell-surface expression, as revealed by FACS analysis (Fig. 2B and C). Similar results were obtained on drug-treated, primary malignant PCs derived from the BM of patients at different states of disease, which displayed increased SA-βGal activity by FACS analysis (Supplementary Fig. S1B) and a clear perinuclear blue staining by microscopy (Supplementary Fig. S1C), both associated with a senescent phenotype. The upregulation of IL15 mRNA (Fig. 3A) was accompanied by a concomitant increase of both IL15 and IL15RA cell surface expression on MEL-treated primary PCs (Fig. 3B and C). We also investigated whether IL15/IL15RA expression could be associated with variation in patient clinical characteristics (Supplementary Table S1). Our data demonstrated that a more pronounced basal and drug-induced IL15/IL15RA expression is independent of the clinical stage, age, and/or the percentage of malignant PCs. Bortezomib and lenalidomide, two commonly used therapeutics that did not induce a senescence phenotype in our model (Supplementary Fig. S1D), did not upregulate IL15/IL15RA expression in MM patients (Supplementary Fig. S2A and S2B). Collectively, these data indicated that drug-induced senescent myeloma cells express increased IL15/IL15RA complex on the surface membrane.

To identify the functional role of increased expression of the IL15/IL15RA complex on drug-treated MM cells, we investigated whether IL15 trans-presentation could induce NK cell activation and proliferation. MEL-treated MM cells were incubated overnight to allow for trans-presentation of IL15 to freshly isolated peripheral blood NK cells, and then CD138⁺ cells were removed from the culture, and the expression of the CD69 activation marker on NK cells was evaluated by immunofluorescence and FACS analysis. As observed with soluble recombinant IL15 alone, used as control, primary NK cells in contact with MM cells showed an increased expression of CD69, which was higher for MEL-treated cells (Fig. 4A).

To examine the direct role of IL15 in NK cell activation induced by drug-treated MM cells, SKO-007(J3) cells were pretreated with an IL15-blocking monoclonal antibody (mAb) or with a nonreactive isotype control before coculture with the NK cells (Fig. 4B). We found that sublethal doses of MEL promoted IL15 trans-presentation by MM cells to NK cells, thus stimulating their activation. We also observed augmented proliferation of NK cells cocultured with drug-treated MM cells compared with untreated MM cells (Fig. 4C and D), which was associated with an increased proportion of cells in the S phase of the cell cycle (Supplementary Fig. S3). Blocking experiments with an anti-IL15 demonstrated a direct role of trans-presented IL15 in the enhanced NK cell proliferation (Fig. 4E). Similar results were obtained with the ARK MM cell line, where an increased CD69 expression (Supplementary Fig. S4A and S4B) and proliferation (Supplementary Fig. S4C and S4D) on primary NK cells was observed upon coculture with MEL-treated myeloma cells. Altogether, these data indicated that drug-treated MM cells, by displaying increased IL15RA, acquire the capability to induce NK cell activation and stimulate NK cell proliferation with a mechanism dependent on IL15 trans-presentation.

IL15RA has been shown to be associated with exosomes derived from DCs (30) or different IL15RA transfectants (31). Thus, we further investigated the expression of IL15/IL15RA complex on MM cell–derived exosomes. To this purpose, exosomes were isolated from the conditioned media of SKO-007(J3) cells as previously reported (28) and characterized by transmission electron microscopy and Western blot. Ultrastructural analysis showed that exosomal preparations contained typical nano-sized cup-shaped vesicles (Fig. 5A). Heat-shock protein 70 (HSP70), a canonical exosomal marker, was detected on exosomes, and the absence of calreticulin, which is exclusively associated with the endoplasmic reticulum (ER), confirmed the purity of the nanovesicle preparations (Fig. 5B). Both IL15RA and IL15 were associated with MM-derived exosomes, although at lower expression compared with the cell lysates (Fig. 5B). No differences were found on IL15RA expression on exosomes derived from drug-treated SKO-007(J3) and ARK MM cells (Supplementary Fig. S5).

As a next step, we endeavored to address whether MM-derived exosomes have the capability to stimulate NK cell proliferation. Our findings revealed a slight increase of cell proliferation in the presence of exosomes alone that was further stimulated by IL15 (Fig. 5C). To further investigate, freshly isolated human NK cells were labeled with CFSE and cultured with exosomes in the presence of exogenous IL15, and we observed a significant increase of IL15-induced NK cell proliferation (Fig. 5D and E).

Because IL2 and trans-presentation of IL15 by IL15RA is required for NK cell proliferation (32), we asked whether exosome-mediated NK cell proliferation was dependent only on the presence of exogenous IL15 or if it could also be mediated by IL2. The augmentation of NK cell proliferation, measured by the expression of the Ki67 marker, was observed only with IL15 and not with IL2, strongly suggesting that this effect could be mediated by IL15 trans-presentation (Fig. 5F). We also asked whether IL15 could be found in the sera from MM patients with different clinical characteristics (Supplementary Table S2). We observed that about 21% of patients with active MM exhibited variable serum IL15 (Fig. 6A), and IL15 was also found in the plasma from BM aspirates isolated from different MM patients (Fig. 6B), thus indicating the presence of this cytokine in the BM tumor microenvironment. Serum IL15 was detectable in MM patients with a less favorable disease progression (Table 1). Altogether, these data highlight that IL15RA harbored by MM-cell derived exosomes is functional, leading to NK cell proliferation when associated with IL15.

Senescent cells show numerous changes in gene expression, leading to an increased expression of many proteins that can promote or repress tumor progression (33, 34). This secretory program is a hallmark of senescence and is referred to as the SASP (33). Herein, we provide evidence that drug-induced multiple myeloma senescent cells overexpress IL15, a cytokine involved in NK cell activation and proliferation/maturation. Upregulation of IL15 expression was accompanied by an increased expression of the IL15/IL15RA complex on the membrane of senescent myeloma cells, which also released exosomes carrying both IL15 and IL15RA, leading to increased activation and proliferation of NK cells. Unsurprisingly, no increased release of the cytokine was observed. Accordingly, with the current theory of trans-presentation proposing that IL15 and IL15RA are pre-associated within presenting cells prior to be shuttled to the cell surface (14), our data showed increase of IL15 both at mRNA and protein levels upon MEL treatment, whereas IL15RA upregulation was only attributable to an increase of the cell surface protein.

The IL15 trans-presentation by drug-treated MM cells leads to NK cell activation and proliferation, as demonstrated by the enhanced expression of the CD69 activation marker and the increased incorporation of BrdUrd by freshly isolated NK cells upon coculture with MEL-treated MM cells. Our findings also demonstrated that both IL15RA and IL15 were present on the exosomes released by SKO-007(J3) MM cells at steady-state conditions, as well as upon MEL treatment. MM-derived exosomes had the capability to promote NK cell proliferation induced by exogenous IL15 but not by IL2, strongly suggesting a key functional role of IL15RA.

In accordance with the expression of IL15RA, no differences were observed between exosomes derived from untreated and MEL-treated cells in terms of stimulatory effects on NK cell proliferation. However, based on our previous findings demonstrating that MEL-treated MM cells can release a higher number of exosomes with respect to untreated cells (28), we propose that low-dose chemotherapy promotes stronger IL15-dependent NK cell responses as a consequence of the enhanced exosome release by drug-treated senescent tumor cells. Our results also provided evidence of soluble IL15 presence in about 21% of patients affected by active myeloma, and we cannot rule out that it is complexed to IL15RA on the exosome surface, as previously reported (31). IL15 infusion in cancer patients has been shown to affect mainly the expansion of the CD56^high NK cell subset and stimulate cytotoxic activity (35).

Complexed IL15, as compared with IL15 alone, has been demonstrated to more efficiently reduce tumor burden (36). It has been also shown that in vivo delivery of IL15/IL15RA complexes triggers rapid and significant regression of established solid tumors in two murine models (37). The physiologic trans-presentation of IL15 has been mimicked by generating a construct of IL15 linked to the extended sushi domain of IL15RA, and it has been established a more effective stimulation of NK and T cells, in terms of proliferation and cytotoxicity, than the targeted presentation of IL15 alone (38). In this respect, a clinical trial of ALT-803, an IL15 super-agonist, in patients with relapsed or refractory MM is currently ongoing (ClinicalTrials.gov NCT02099539). However, according to previous data (39), we found a correlation between IL15 release in the sera of MM patients and disease progression. To this regard, it is important to underline that the induction of senescence by genotoxic agents, on the one hand, inhibits tumor cell growth, and on the other, enhances the IL15-driven effect on NK cell antitumor activity. The findings that bortezomib and lenalidomide, two drugs with different mechanisms and unable to induce senescence in our model, did not upregulate IL15/IL15RA expression in MM patients, further strengthening the hypothesis that increased expression of the IL15/IL15RA complex is driven by senescence.

Induction of cellular senescence by chemotherapeutic agents has emerged as an appealing option to arrest cancer cell proliferation. Immunosurveillance of senescent cells in cancer and other pathological processes by the innate immune system has been highlighted. Our group contributed to report the importance of NK cells in senescent cell recognition and elimination (19, 21). Using a model of oncogene-induced senescence (OIS), Iannello and colleagues demonstrated that p53-restored senescent tumor cells recruit NK cells by secreting CCL2 (40). Our findings indicated that in addition to these immune modulatory activities, it can be envisaged that the success of senescence-based anticancer therapies might be also rely on the ability of senescent cells to trigger an IL15-dependent antitumor innate immune response.

No potential conflicts of interest were disclosed.

Conception and design: C. Borrelli, A. Soriani

Development of methodology: C. Borrelli, B. Ricci, E. Vulpis, C. Fionda, A. Peri, A. Zingoni

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Borrelli, B. Ricci, M.R. Ricciardi, M.T. Petrucci, L. Masuelli

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Borrelli, M. Cippitelli, A. Zingoni

Writing, review, and/or revision of the manuscript: M. Cippitelli, A. Santoni, A. Soriani

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Soriani

Study supervision: A. Santoni, A. Soriani

This work was supported by the Italian Association for Cancer Research (AIRC 5 × 1000 cod. 9962), the Sapienza University of Rome (Progetto di ricerca 2016, cod. RM116154C8F24748). E. Vulpis is supported by a fellowship from AIRC.

The expert technical assistance of Bernardina Milana is gratefully acknowledged. We also thank Lucilla Simonelli for technical assistance in electron microscopy procedures.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.