Combined action of suicide gene exosomes from pancreatic cancer-associated fibroblasts and from mesenchymal stem cells as a pancreatic ductal adenocarcinoma treatment approach Authors: Jana Jakubechova, Ursula Altanerova, Michal Andrezal, Dajana Vanova, Tatiana Zeleznikova, Peter Bartek, Marina Cihova, Bozena Smolkova, Verona Buocikova, Peter Makovicky, Kristina Jakic, Monika Burikova, Maria Urbanova, Veronika Repaska, Benjamin Spanik, Miroslav Tomas, Peter Dubovan, Georgina Kolnikova, Zbynek Zdrahal, Vaclav Pustka, David Potesil, Cestmir Altaner Published: 2025 ABSTRACT BackgroundThe pancreatic cancer-associated fibroblasts (pCAFs) are among the most active components of the pancreatic ductal adenocarcinoma (PDAC). The pCAFs being of mesenchymal stem/stromal cell origin, interact directly with tumor stromal elements, modulate tumor development, and are involved in the formation of pre-metastatic niches that result in unsatisfactory PDAC treatment outcomes. This study aimed to develop an innovative approach for the treatment of desmoplastic pancreatic carcinoma via intracellularly targeted exosomes derived from pCAFs and from mesenchymal stem cells (MSCs) transduced with the suicide gene – yeast cytosine deaminase::uracil phosphoribosyl transferase (yCD::UPRT).MethodspCAFs were isolated from four PDAC tumor specimens and MSCs from various tissues. Their transduction with yCD::UPRT gene produce homogenous gene transduced cell populations capable of secreting suicide gene exosomes. Both gene- transduced and naive cells and their exosomes underwent characterization by biophysical, biochemical, microscopic, and LC–MS/MS proteomic methods. Tumor cell-killing functionality was assessed using three pancreatic cancer cell lines. The killing efficacy of combined suicide gene exosomes of MSCs and pCAFs was measured in a mixture of pCAFs and MIA PaCa-2 cells as a simulated desmoplastic pancreatic tumor in vitro.ResultsMSCs and pCAFs suicide gene exosomes act as cancer cell-targeted drugs, effectively killing pancreatic carcinoma cells. Exosomes intracellular convert the non-toxic prodrug 5-fluorocytosine into cytotoxic 5-fluorouracil and its metabolites in a dose-dependent manner. In experiments simulating the desmoplastic microenvironment of PDAC, we have found that the suicide gene exosomes from both cells conjugated with prodrug effectively target and inhibit the growth of simulated PDAC.ConclusionExosomes containing the yCD::UPRT gene from pCAFs and MSCs function as “Trojan horse” therapies, efficiently and dose-dependently eliminating pancreatic cancer cells. PDAC environment-targeted yCD::UPRT-gene exosomes from MSCs and pCAFs show promise for a novel PDAC treatment.Graphical Abstract Background Pancreatic ductal adenocarcinoma (PDAC) is one of the deadliest cancers, with an increasing incidence worldwide. The current treatment modalities, which include surgical resection combined with radiation and chemotherapy regimens that have a well-known toxicity profile and resistance pattern, provide a median survival of only 6 months [1]. Several different factors contribute to the poor prognosis of pancreatic cancer patients. Due to the lack of early stage-specific symptoms, most patients are diagnosed at an advanced stage. PDAC tumors have an excessive desmoplastic stroma, accounting for up to 90% of the total tumor volume, causing limited intra-tumor treatment options [2]. Pancreatic carcinoma comprises various stromal cells, such as inflammatory, immune, and cancer-associated fibroblasts (CAFs). CAFs are key and active components of the prominent PDAC stroma. They originate from mesenchymal stem/stromal cells and the pancreatic tumor cells modify them [3–5]. To address this, we named the CAFs in PDAC pancreatic cancer associated fibroblasts (pCAFs). The pCAFs and their generated extracellular matrix function as a physical barrier, preventing efficient drug delivery. pCAFs secrete extracellular vesicles - exosomes that mediate cross-talk between tumor cells and other stromal cells, can initiate PDAC pre-metastatic niche formation in the liver and lungs, and contribute to the induction of gemcitabine resistance [6]. Although many studies consider pCAFs as a potential therapeutic target, their high heterogeneity limits the development of an effective anti-PDAC drug [7]. Efforts to exploit the tumor microenvironment (TME) for therapy through deconstructing the surrounding desmoplastic stroma have largely failed. Thus far, this chemotherapeutic treatment approach has been largely unsuccessful. Exosomes are ideal drug delivery vehicles capable of efficiently targeting cancer cells, and exhibit high biocompatibility and low immunogenicity, thus minimizing off-target damage and side effects [8, 9]. Exosomes derived from mesenchymal stem cells are entering the clinic as effective anti-inflammatory agents for tissue regeneration or as effective immunomodulatory drugs [10]. Any innovative PDAC therapy that is potentially curative should be targeted intracellular to the PDAC tumor desmoplastic compartment and disable the metastatic process, thus preventing metastases. We hypothesize that Gene-Directed Enzyme Prodrug Therapy (GDEPT) - suicide gene therapy mediated by yeast cytosine deaminase::uracilphosphoribosyl transferase realized through yCD::UPRT-exosomes from mesenchymal stem cells (MSCs) and exosomes from pCAFs transduced with the yCD::UPRT gene may be an innovative PDAC treatment. The working hypothesis is supported by data showing tumor cell targeting behavior of yCD::UPRT-MSCs, secretion of exosomes containing the mRNA of the yCD::UPRT gene from yCD::UPRT-MSCs, and tumor-homing character of suicide gene exosomes secreted from tumor cells transduced with suicide gene [11–13]. Previously, the yCD::UPRT-MSCs exosomes were proven therapeutically active in vivo in rats bearing intracerebral glioblastoma cells. A notable number of rats were cured when the conditioned medium (CM) containing yCD::UPRT-MSC-exosomes conjugated with the prodrug 5-FC was repeatedly injected systemic or intranasal [14]. Recently, we demonstrated the therapeutic potential of yCD::UPRT suicide gene exosomes present in concentrated CM from yCD::UPRT gene transduced placenta-derived MSCs (Plac-MSCs) in preclinical PDAC models [15]. Materials and methods Cell cultures The human adipose tissue-derived MSCs (AT-MSCs) were isolated from lipoaspirates obtained from healthy individuals undergoing elective lipoaspiration using a collagenase type VII digestion and plastic adherence technique described previously [16]. Human bone marrow-derived MSCs (BM-MSC) were isolated from bone marrow aspirates taken from the iliac crest of normal donors by density gradient centrifugation. BM-MSCs adhered to plastic dishes were grown in a complete culture medium DMEM low glucose (1 g/l) supplemented with 5% human platelet lysate (PL) and incubated at 37 °C in a humidified atmosphere with 5% CO2. Dental pulp stem cells (DP-MSCs) were isolated from dental pulp tissue fragments adhered to plastic tissue culture dishes as described previously [17]. The same technique was used to obtain placenta-derived MSCs from the chorionic plate tissue and mesenchymal stem cells from umbilical cord tissue (UC-MSCs). The tissue was cut into small fragments, washed with PBS, adhered to a Petri culture dish, and surrounded with a complete culture medium (Gibco low glucose DMEM) supplemented with 5% PL. Three days later UC-MSCs began to migrate from the adhered umbilical cord tissue explants. To expand all types of MSCs, the cells were seeded at 4000 cells/cm2 to plastic dishes and grown with medium exchange every 2–3 days. For cryo-preservation, the cells were then harvested with trypsin/EDTA, suspended at 1 × 106 cells/ml in 10% dimethyl sulfoxide and 30% human serum albumin, and frozen in 1 ml aliquots in liquid nitrogen. Human pancreatic cancer cell lines MIA PaCa-2 (RRID: CVCL_0428), BxPC-3 (RRID: CVCL_0186), and SU.86.86 (RRID: CVCL_3881) were cultured in DMEM (Gibco) supplemented with 5% of Fetal Bovine Serum and antibiotics (100 unit/ml penicillin and 100 µg/ml streptomycin). Creation of MIA PaCa-2 cell spheroids. MIA PaCa-2 cells were cultured in ultra-low attachment plates to form spheroids. Their growth was evaluated by the Incucyte® Live-Cell Analysis System. All experiments were performed with mycoplasma-free cells. All human cell lines have been authenticated using STR profiling. Isolation of pCAFs and omentum fibroblasts From the resected tumor tissue, a specimen for histopathological examination was taken, and the rest of the tissue was fragmented into small pieces, suspended in a freezing medium, and stored in a liquid nitrogen biobank. To isolate the pCAFs, tumor tissue fragments were adhered to a plastic tissue culture dish and surrounded with a complete culture medium (Gibco low glucose DMEM) supplemented with 5% PL. The next day, the fragments were overlaid with medium and cultivated at 37 °C in a humidified atmosphere with 5% CO2. Two to three days later, the pCAFs began to migrate from the adhered tumor tissue explants. When the dish of pCAFs was near confluence, the fragments were transferred to a new dish. Up to 20 transfers of tumor stroma specimens were carried out with continued pCAFs production. The same technique was used to isolate fibroblasts from human omentum tissue. Platelet lysate preparations Platelets lysate (PL) was prepared from expired bags of human platelets. Platelets (5 × 1011/l) in bags were lysed by ping-pong freezing at −80o C and thawing at 37o C three times. The platelet bodies were eliminated by centrifugation, and the supernatant was filtered through a 0.22 μm GP Millipore Express Plus Membrane Stericup. PL was negative for bacterial, fungal, and mycoplasma contaminations. Transduction of pCAFs with yCD::UPRT gene To prepare pCAFs expressing the yCD::UPRT gene from retrovirus integrated into cell DNA, the sub-confluent cells were infected with replication-defective retrovirus with mixed amphotropic and ecotropic envelope glycoproteins. The recombinant retrovirus possessing the yCD::UPRT gene was prepared as described in detail previously [18]. The multiplicity of infection of MSCs and pCAFs was 10 infectious units per cell. Virus-infected pCAFs were cultivated in selection media containing a pretested concentration of G418 (0.5–1.2 mg/ml of G418) to obtain a homogenous population of yCD::UPRT gene-transduced cells. The addition of 5-FC to yCD::UPRT gene-transduced cells induced apoptotic death of all cells as a control of their homogeneity. Cell growth assessment using an IncuCyte® live cell monitoring system Cell growth was monitored by “real-time in vitro micro-imaging” using the IncuCyte® ZOOM system (Essen BioScience, UK). Most cell viability data obtained from the IncuCyte® system were in good accordance with data obtained by subjecting the plates to the MTT Proliferation Assay (Sigma-Aldrich, St. Louis, MO). Animal studies The toxicity of 24-hour conjugates of yCD::UPRT-pCAFs, yCD::UPRT-UC-MSCs, yCD::UPRT-DP-MSCs, was tested in a group of SCID mice. For each conjugate and for the control group six mice were used. The mice were injected intraperitoneal every second day with a 0.5 ml dose for 5 weeks. Their body weights were continuously measured. All experiments were conducted in accordance with EU directive 2010/63 (https://eur-lex.europa.eu/eli/dir/2010/63/oj), with approvals from the institutional ethics committee and by the national competent authority of the State Veterinary and Food Administration of the Slovak Republic Ro-290–3/2020–220 and 5993/2024–220 in the authorized animal facility under license No. SK UCH 02022. Isolation of exosomes from conditioned medium The source of exosomes was a 24-hour CM free of growth supplements collected from 80% confluent cell cultures of 98% cell viability as assessed by Trypan-Blue exclusion (Sigma-Aldrich, St. Louis, MO). CM was centrifuged at 800xg for 5 min, and filtered through a 0.22 μm syringe filter to remove debris. Exosomes from CM were concentrated via various techniques, such as TFF-Easy cartridge, membrane affinity spin columns, using the exoEasy Maxi Kit (Qiagen, Hilden, Germany), or by size-exclusion chromatography. Purification and concentration of yCD::UPRT-exosomes by tangential flow filtration The exosomes were purified and concentrated by a TFF-Easy tangential flow filtration-based device with 5 nm cut-off holes (HansaBioMed Life Sciences Ltd. Tallinn, Estonia). Separation of the exosomes out of CM by size-exclusion chromatography and the quantification of exosomes According to manufacturer’s instructions, the CM from yCD::UPRT-transduced pCAFs was separated on the IzonQev10/35nm column. Each fraction was sequentially collected using the peristaltic pump P-1 (Pharmacia Lkb, Uppsala, Sweden). The concentration of proteins was estimated using the NanoDrop (Thermo Scientific™, Wilmington, DE, USA), or by the Micro BCA protein Assay kit (Thermo Fisher Scientific). Nanoparticle tracking analysis (NTA) NTA was carried out using a NanoSight Pro equipped with a 405 nm laser, a CCD camera, and a syringe pump system (Malvern Instruments, Ltd., UK). Before exosome analysis, instrument performance was assessed using 100 nm latex beads (Malvern Instruments). Samples were diluted using nanoparticle-free water to ensure the particle concentration fell within the linear range of the instrument. Samples were administered and recorded under controlled flow from a syringe pump (set to a speed 2.5 µl/second). Five replicate videos of 30 s were recorded. Cryo electron microscopy of exosomes Four µl of the exosome sample was applied to freshly glow-discharged holey carbon grids (Quantifoil, Cu, 300mesh, R2/1) and vitrified by plunging into liquid ethane using Thermo Scientific Vitrobot Mark IV (4 °C, 100% rel. humidity, 60 s waiting time, 6 s blotting time). Grids were subsequently mounted into Autogrid cartridges and loaded into Glacios (Thermo Scientific) transmission electron microscope for imaging. The microscope was operated at 200 kV. Exosome cryo-TEM images were collected on Falcon4i direct electron detector at 150000x nominal magnification with underfocus in the range 1.5–3 μm and overall dose ~ 20 e/Å2. Detection of interleukins in CM of pCAFs and yCD::UPRT-pCAFs According to the manufacturer’s instructions, cytokines IL-6, IL-8, VEGF, and MCP-1 were measured using the multiplex bead-based immunoassay BD™ CBA Human Flex test (Beckton Dickinson). Briefly, samples (50 µl) were incubated for 1 h at room temperature, protected from light, with a 50 µl working solution of Capture Beads (1 µl) of beads for each appropriate cytokine. Then, to the working solution, the PE detection reagent was added (including 1 µl detection reagent for each appropriate cytokine). Test tubes were incubated for 2 h (in the dark at room temperature). After incubation, 500 µl of Wash Buffer was added to each assay tube and centrifuged at 700xg for 5 min. The supernatant was aspirated and discarded. The bead pellet in each assay tube was suspended in 200 µl of Wash Buffer. The fluorescence intensity was measured using a flow cytometer BD FACS Canto II (reported channel PE, beads channels APC and APC-Cy7™). The final concentration of cytokines was determined from calibration curves (prepared from lyophilized Standard components included in the BD™ CBA Human Flex test). Expression of pluripotent in pCAFs and yCD::UPRT-pCAFs and their exosomes Relative expression levels of 15 pluripotent stem cell genes at the protein level was measured by Proteome Profiler Human Pluripotent Stem Cell Array (R&D Systems Minneapolis, MN 55413). The same amount of protein was used for incubation of all arrays, and the procedure was performed as directed by the manufacturer. Membranes were exposed to Western HRP Substrate (LuminataTM Forte), covered with plastic wrap, and exposed to X-ray film. Pluripotent stem cell array data on developed X-ray film were quantified by scanning the film using a transmission mode scanner. ImageJ software (NIH) was used for the quantitative evaluation. Immunocytochemical staining of cells pCAFs and corresponding yCD::UPRT gene transduced pCAFs, MSCs, and yCD::UPRT-MSCs were cultured in a dish containing glass slides. Cultured cells adhered to slides were washed with PBS and fixed with 4% paraformaldehyde/PBS for 15 min at RT. pCAFs were permeabilized with 0.05% Triton/PBS for 15 min and blocked with 3% BSA/PBS. Immunofluorescence staining was performed by incubating the cells adhered to glass slides for 1 h at 37 °C. The following primary antibodies were used: cytokeratin 19 (CK19), carcinoembryonic antigen (CEA), vimentin (VIM), Ki67, and CD9. After being washed in PBS, the cells were incubated with Alexa Fluor 488 secondary antibody for 1 h at 37 °C. After that, cells were incubated with Phalloidin Alexa Fluor 546 antibody for 1 h at 37 °C and counterstained using DAPI in PBS for 15 min. Glass slides were mounted, and the edges were sealed with clear acrylic nail polish, and viewed under fluorescence microscopy. Haematoxylin and Eosin staining The cells adhered to glass slides were washed with PBS and fixed with 4% paraformaldehyde/PBS for 15 min at RT. Cells were permeabilized with 0.05% Triton/PBS for 15 min, and hematoxylin was added to stain the nuclei for 5 min, after which the sections were rinsed with tap water. Eosin was added for 1 min, after which the sections were rinsed with ethanol. Then, the samples were dehydrated, cleared, and sealed. The cells were observed and photographed under an Olympus BX46 microscope with an Olympus U-CMAD3 camera. Proteomic analysis Proteins extracted from exosome samples were subjected to filter-aided sample preparation as described elsewhere [19]. The resulting peptides were analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) performed using UltiMate 3000 RSLCnano system (Thermo Fisher Scientific) online coupled with a timsTOF Pro mass spectrometer (Bruker). Methodical details have been described previously [20]. Statistical analysis Unless noted otherwise, all experiments were repeated at least three times to enable statistical analysis and all results were similar between replicates. Error bars represent the standard error of the mean from independent samples within each experiment. All statistical analyses were performed using inferential statistics with the Student’s t-test. Results Isolation of pCAFs from primary PDAC tumor specimens The pCAFs were isolated from pieces of tumor tissue obtained during surgical therapy of PDAC patients without previous chemotherapy. The designation of cells, patients, tumor classification, and histological analysis of the tumor from which the pCAFs originated are shown in Table 1. Table 1Origin of isolated pCAFsPatientCulture nameSexYear of birthTumor siteGradeStageSizeHistology27pCAFs27F1943TailG3HA; pT3pN0 pM0Pn + V0 L072 × 38 × 32 mmInvasive PDAC31pCAFs31F1954Head & processus uncinatusG2III; pT3pN2pM034 × 27 × 32 mmInvasive PDAC78pCAFs78F1965HeadG1IIB; T2pN1pMxd. 28 mmInvasive PDAC79pCAFs79M1948Processus uncinatusG1-2IIB; pT2pN1pMxd. 25 mmInvasive PDAC Table 1Origin of isolated pCAFsPatientCulture nameSexYear of birthTumor siteGradeStageSizeHistology27pCAFs27F1943TailG3HA; pT3pN0 pM0Pn + V0 L072 × 38 × 32 mmInvasive PDAC31pCAFs31F1954Head & processus uncinatusG2III; pT3pN2pM034 × 27 × 32 mmInvasive PDAC78pCAFs78F1965HeadG1IIB; T2pN1pMxd. 28 mmInvasive PDAC79pCAFs79M1948Processus uncinatusG1-2IIB; pT2pN1pMxd. 25 mmInvasive PDAC Origin of isolated pCAFs Pictures of tumor tissue fragments with migrating pCAFs and their histological examination of more than 10 transfers are shown in Fig. 1A. KRAS mutations are commonly found in pancreatic cancers. The analysis of pCAFs and yCD::UPRT-PCAFs for KRAS mutations by Droplet Digital PCR was negative, confirming pure pCAFs population. Principle of Gene-Directed enzyme prodrug therapy (GDEPT) The fused yeast cytosine deaminase::uracilphosphoribosyl transferase gene transduced into MSCs via retrovirus is transcribed and translated, leading to the production of an enzyme. The enzyme is included into exosomes, which are excreted from cells into CM. When such exosome enters the recipient cell in the presence of nontoxic prodrug 5-FC, the fused enzyme converts the prodrug into potent chemotherapeutic agent 5-FU and metabolites. Nuclear Magnetic Resonance spectroscopy measurements of yCD::UPRT-PlacMSC-cCM demonstrated that 10% of non-toxic 5-FC was effectively and uniformly metabolized into 5-FU in recipient cells [15]. Fig. 1Isolation and characterization of pCAFs. (A) Pictures of explants during their transfers ending with necrotic tissue material. (B) Growth curves of homogenous yCD::UPRT-pCAFs in the presence/absence of prodrug 5-fluorocytosine. (C) Phenotypical characterization of isolated pCAFs and yCD::UPRT-pCAFs by flow cytometry. (D) Differentiation induction assay of isolated pCAFs, yCD::UPRT-pCAFs, AT-MSCs, and BM-MSCs. (E) Immunocytochemical characterization of the isolated pCAFs, yCD::UPRT-pCAFs, and MSCs of different tissue-origin. (F) Secretion of interleukins into the conditioned medium of pCAFs and yCD::UPRT-pCAFs compared to MSCs and yCD::UPRT-MSCs. (G) Comparison of expression of pluripotent genes at the level of proteins in pCAFs, yCD::UPRT-pCAFs Isolation and characterization of pCAFs. (A) Pictures of explants during their transfers ending with necrotic tissue material. (B) Growth curves of homogenous yCD::UPRT-pCAFs in the presence/absence of prodrug 5-fluorocytosine. (C) Phenotypical characterization of isolated pCAFs and yCD::UPRT-pCAFs by flow cytometry. (D) Differentiation induction assay of isolated pCAFs, yCD::UPRT-pCAFs, AT-MSCs, and BM-MSCs. (E) Immunocytochemical characterization of the isolated pCAFs, yCD::UPRT-pCAFs, and MSCs of different tissue-origin. (F) Secretion of interleukins into the conditioned medium of pCAFs and yCD::UPRT-pCAFs compared to MSCs and yCD::UPRT-MSCs. (G) Comparison of expression of pluripotent genes at the level of proteins in pCAFs, yCD::UPRT-pCAFs Transduction of pCAFs by retrovirus with the yCD::UPRT suicide gene Growing evidence suggests that signals from stromal cells are involved in the metastatic progression of PDAC. In concordance with our working hypothesis, and with our experience with the transduction of MSCs by the yCD::UPRT suicide gene, which has led to the secretion of tumor-killing exosomes, we transduced isolated pCAFs. The mode of retrovirus replication through integrating its genome into cell DNA as a DNA provirus can be used for foreign gene insertion into the cell. Such a gene is expressed due to potent retrovirus promoters and the message is translated to the corresponding proteins. Using replication-defective mixed eco- and amphotropic envelope retrovirus containing yCD::UPRT suicide gene linked by IRES to neo gene we obtained pCAFs transduced with the suicide gene. Co-translation of yeast cytosine deaminase and neo gene facilitates the selection of a homogenous population of yCD::UPRT-pCAFs using a pretested concentration of antibiotic G418. The homogeneity of cells with integrated yCD::UPRT retrovirus can be easily checked by the addition of nontoxic prodrug 5-FC causing apoptotic death of all cells. The prepared yCD::UPRT gene transduced pCAFs differed neither in growth behavior nor cell morphology from naive pCAFs. They were selected by G418 and found to be the homogenous population as evidenced by the addition of prodrug 5-FC which led to the death of all cells (Fig. 1B). Phenotypic characterization of isolated pCAFs and corresponding yCD::UPRT gene transduced ones Phenotypic characterization of pCAFs by flow cytometry (Fig. 1C) revealed cells of mesenchymal lineage origin, together with markers of CAFs. All isolated pCAFs were positive for the tetraspanins CD9, CD81, and CD63, the MSC markers CD73, CD105, CD44, and CD90, and negative for the hematopoietic markers CD45 and CD3. The membrane glycoprotein CD36 was negative in all pCAFs except pCAFs79 but became negative in gene transduced cells. All pCAFs were positive for the platelet-derived growth factor receptor- β CD140b. The obtained data are summarized in Table 2. Table 2Phenotypical characterization of pCAFs and yCD::UPRT-pCAFs by flow cytometry expressed as a percentage of cells binding to a monoclonal antibodyCD34CD63CD90CD9CD140bCD44CD73CD105CD45CD81CD36CD3pCAFs270.1499.991.1191.9598.9399.9898.8999.90.5299.870.511.36yCD::UPRT-pCAFs272.0972.7299.5888.7753.7493.0583.4984.12.2199.11.689.65pCAFs310.1999.2897.9975.9999.1490.9887.8988.390.0690.761.270.06yCD::UPRT-pCAFs311.1498.9410099.7489.0599.8199.0698.140.6499.9710.580.24pCAFs780.799.4458.0186.2880.0599.0395.497.312.7198.258.0212.02yCD::UPRT-pCAFs780.1399.5899.7857.9967.1485.8993.8298.981.396.140.856pCAFs790.1499.0994.5281.2799.8699.9395.7199.040.0799.7275.540.15yCD::UPRT-pCAFs792.2295.199.9999.8193.8999.8496.6398.062.2699.10.430.89 Table 2Phenotypical characterization of pCAFs and yCD::UPRT-pCAFs by flow cytometry expressed as a percentage of cells binding to a monoclonal antibodyCD34CD63CD90CD9CD140bCD44CD73CD105CD45CD81CD36CD3pCAFs270.1499.991.1191.9598.9399.9898.8999.90.5299.870.511.36yCD::UPRT-pCAFs272.0972.7299.5888.7753.7493.0583.4984.12.2199.11.689.65pCAFs310.1999.2897.9975.9999.1490.9887.8988.390.0690.761.270.06yCD::UPRT-pCAFs311.1498.9410099.7489.0599.8199.0698.140.6499.9710.580.24pCAFs780.799.4458.0186.2880.0599.0395.497.312.7198.258.0212.02yCD::UPRT-pCAFs780.1399.5899.7857.9967.1485.8993.8298.981.396.140.856pCAFs790.1499.0994.5281.2799.8699.9395.7199.040.0799.7275.540.15yCD::UPRT-pCAFs792.2295.199.9999.8193.8999.8496.6398.062.2699.10.430.89 Phenotypical characterization of pCAFs and yCD::UPRT-pCAFs by flow cytometry expressed as a percentage of cells binding to a monoclonal antibody The ability of pCAFs and yCD::UPRT-pCAFs to be induced to differentiation in vitro Mesenchymal stem/stromal cells can be induced to differentiate into osteoblasts, chondrocytes, and adipocytes in vitro. The yCD::UPRT gene transduced MSCs did not lose their ability to be induced to differentiate. We subjected isolated pCAFs and yCD::UPRT-pCAFs to a differentiation assay and found that they differed. Naive pCAFs27, pCAFs31, and corresponding yCD::UPRT-gene transduced cells did not differentiate into adipocytes, while pCAFs78 and yCD::UPRT-pCAFs78, pCAFs79 did. All naive pCAFs differentiated into osteoblasts, whereas yCD::UPRT-gene transduction of pCAFs27, and pCAFs31 led to the loss of osteo-differentiation. In contrast, the osteo-differentiation of pCAFs78 and pCAFs79 was reduced (Fig. 1D). The differentiation induction assay of isolated pCAFs confirmed their individual features and tumor origin. Histopathological examination of isolated pCAFs and matching yCD::UPRT-pCAFs compared to similar isolates of MSCs of different tissue origin The obtained data are presented as pictures in Fig. 1E. The tetraspanin CD9 was found to be abundant on the surface of all isolated pCAFs from the patient’s PDAC correlating with the high tumor stage mentioned in Table 1. Secretion of interleukins into the conditioned medium of PCAFs and yCD::UPRT-PCAFs compared to MSCs and yCD::UPRT-MSCs PCAFs, besides many other functions, coordinate cytokine-mediated inflammatory cross-talk within the TME that may dictate the response to therapeutic agents. Further, characterizing isolated pCAFs and yCD::UPRT-gene transduced pCAFs we measured the secretion of IL-6, IL-8, VEGF, and MCP-1 (Fig. 1F). The interleukin secretion profile of pCAFs confirmed the individual features of isolated pCAFs. The transduction of pCAFs with the yCD::UPRT gene retroviral vector led to small quantitative changes in the interleukin profile. The interleukin secretion profile of AT-MSCs, UC-MSCs, Plac-MSCs, and DP-MSCs, and corresponding yCD::UPRT-MSCs was found to be type specific, and gene transduction did not change the profile. Generally, the level of interleukins from MSCs was lowered compared to pCAFs. The secretion of VEGF was low in both stromal cell types, whether naive or gene transduced. The increased concentration of IL8 in gene transduction pCAFs was influenced by pro-inflammatory cytokines (e.g., TNF-a, IL1-beta, IL-6, IFN-γ) (Fig. 1F). Individual pCAFs differ in the protein expression of pluripotent genes Tissue cultures of pCAFs and yCD::UPRT-pCAFs derived from early passages were examined for relative expression of 15 genes by the Proteome Profiler Human Pluripotent Stem Cell Array. The data, quantified by ImageJ software, are presented in Fig. 1G. The comparison of relative expression of pluripotent genes at the protein level of pCAFs confirms the individual features of each patient’s PDAC. The transduction of pCAFs with the yCD::UPRT gene retrovirus vector and subsequent selection for the homogenous gene transduced cell population, led to small quantitative changes. Relative protein expression of all tested genes in yCD::UPRT-pCAFs did not differ significantly from pCAFs. The individual character of yCD::UPRT-pCAFs was sustained. pCAFs and stably transduced pCAFs with yCD::UPRT gene secrete exosomes The size distribution analysis of exosomes present in CM of pCAFs and yCD::UPRT-pCAFs, using nanoparticle tracking analysis (NTA) (Fig. 2A) revealed that they fell within the 50 to 100 nm size range. The modal average particle size for pCAFs-exosomes was 74.5 ± 8.5 nm, with an average concentration of 1.75 × 1011/ml. For yCD::UPRT-pCAFs-exosomes, the average modal particle size was 72.5 ± 2.5 nm with an average concentration of 4.62 × 1011/ml. The similar size distribution analysis of exosomes secreted into CM of four types of MSCs and corresponding yCD::UPRT-gene transduced cells using NTA by NanoSight™ Pro (Malvern) (Fig. 2B) revealed that they fell within the size range of 50 to 130 nm. The size range of the yCD::UPRT-gene stably transduced MSCs did not differ from that of naive cells. The mean concentration of particles AT-MSCs, UC-MSCs, and the corresponding gene transduced was 2.0 × 1011/ml, for Plac-MSCs and DP-MSCs (both types), the particle concentration was 1.8 × 1011/ml. Fig. 2Characterization of exosomes secreted by pCAFs, yCD::UPRT-pCAFs, and yCD::UPRT-MSCs. (A) Nanoparticle tracking analysis of exosomes produced by pCAFs and stably yCD::UPRT gene transduced pCAFs. (B) Nanoparticle tracking analysis of exosomes produced by four types of MSCs and stably yCD::UPRT gene transduced MSCs. (C) Proteomic comparison of exosome-specific antigens – tetraspanins present in exosomes of pCAFs and yCD::UPRT-pCAFs. (D) Scatterplots heat map of normalized data comparing pCAFs- exosomes and yCD::UPRT-pCAFs-exosomes. (E) Heat map of top 100 protein groups in exosomes of naive individual pCAFs. (F) Comparison of proteomes of exosomes isolated from naive pCAFs and yCD::UPRT gene transduced pCAFs showing presence of yCD::UPRT gene products in transduced cells (FUR1 – Uracil phosphoribosyltransferase, FCY1 – Cytosine deaminase). (G) Size exclusion chromatography of CM from yCD::UPRT-pCAFs on an IzonQev10/35nm column. (H) Comparison of exosomes produced by yCD::UPRT-DP-MSC CM-5-FC conjugate and yCD::UPRT-DP-MSCs isolated from the dental pulp of various tissue donors Characterization of exosomes secreted by pCAFs, yCD::UPRT-pCAFs, and yCD::UPRT-MSCs. (A) Nanoparticle tracking analysis of exosomes produced by pCAFs and stably yCD::UPRT gene transduced pCAFs. (B) Nanoparticle tracking analysis of exosomes produced by four types of MSCs and stably yCD::UPRT gene transduced MSCs. (C) Proteomic comparison of exosome-specific antigens – tetraspanins present in exosomes of pCAFs and yCD::UPRT-pCAFs. (D) Scatterplots heat map of normalized data comparing pCAFs- exosomes and yCD::UPRT-pCAFs-exosomes. (E) Heat map of top 100 protein groups in exosomes of naive individual pCAFs. (F) Comparison of proteomes of exosomes isolated from naive pCAFs and yCD::UPRT gene transduced pCAFs showing presence of yCD::UPRT gene products in transduced cells (FUR1 – Uracil phosphoribosyltransferase, FCY1 – Cytosine deaminase). (G) Size exclusion chromatography of CM from yCD::UPRT-pCAFs on an IzonQev10/35nm column. (H) Comparison of exosomes produced by yCD::UPRT-DP-MSC CM-5-FC conjugate and yCD::UPRT-DP-MSCs isolated from the dental pulp of various tissue donors Proteomic analysis of exosomes secreted from pCAFs and stably transduced yCD::UPRT-pCAFs LC-MS/MS analysis revealed slight differences in the levels of exosome-specific antigens when comparing exosomes from individual pCAFs and the corresponding yCD::UPRT-exosomes (Fig. 2C). Incorporating a gene into the cell DNA does not interfere with the mechanisms of EV creation. Because the gene is of yeast origin, its expression, followed by translation is probably recognized as garbage and included in exosomes to be excreted. Data analysis at the whole proteome level revealed no clear differentiation between naive pCAFs and yCD::UPRT-pCAFs (Fig. 2D). It means that the transduction of the yCD::UPRT gene into the genome of pCAFs did not influence the protein cargo of exosomes. Some observed differences might be related to the presence of contaminants from the CM that serve for the isolation of exosomes. The abundance of the top 100 protein groups in exosomes produced by individual pCAFs (Fig. 2E) indicates that the pCAFs isolated from individual patients are different, based on different protein cargo of exosomes. In agreement with other features of the exosomes mentioned above, in the gene transduced pCAFs, the secreted exosomes are enriched with the products of the yCD::UPRT gene on the protein level (Fig. 2F). Proteomic analysis of exosomes secreted from yCD::UPRT-DP-MSCs compared to corresponding yCD::UPRT-DP-MSCs-CM-5-FC conjugate isolated from the dental pulp of unrelated donors revealed great similarity, suggesting their biological activities may be preserved (Fig. 2H). Exosomes secreted from yCD::UPRT-pCAFs into CM are responsible for the tumor cell-growth-killing activity of pancreatic tumor cell lines in the presence of prodrug 5-FC The secretome, represented by CM from yCD::UPRT-pCAFs, was composed of biologically active factors, including many low molecular weight proteins, such as cytokines, chemokines, growth factors, and exosomes. To determine the extent to which the exosomes secreted from these cells are responsible for their tumor cell growth/inhibiting activity, we separated the CM using the size-exclusion chromatography on an IzonQev10/35nm column. The obtained fractions were tested for their tumor-killing activity in the presence/absence of 5-FC on MIA PaCa-2 cells. The assay arrangement allowed for the detection of growth inhibition and stimulation simultaneously. Figure 2G shows that most tumor growth inhibition activity was localized to the nanoparticle fractions. Addition of prodrug 5-FC to CM of yCD::UPRT-pCAFs forms 5-FC exosome conjugate We previously reported that the suicide gene yCD::UPRT inserted in the cell DNA of MSCs via a retrovirus is transcribed into mRNA and translated to proteins. Furthermore, the MSCs engineered to express the yCD::UPRT suicide gene secreted exosomes into CM, possessing mRNA of the suicide gene in their cargo. The yCD::UPRT-MSCs-exosomes in the presence of prodrug 5-FC were shown to induce the death of various kinds of tumor cells in a dose-dependent manner. We discovered that adding 5-FC to the CM from several types of yCD::UPRT-MSCs led to the formation of a mixture of nanoparticles where 5-FC was conjugated. We were interested in whether exosomes in the CM of yCD::UPRT gene transduced pCAFs behaved similarly. Furthermore, we co-cultivated MIA PaCa-2 cells with the individual pCAFs in a ratio of one to four, testing various yCD::UPRT-pCAFs-CM-5-FC conjugates to find the extent of tumor cell growth inhibition with the plain conjugate or with an additional dose of 2 µg/ml of gemcitabine (GCB) using the MTT test. Data obtained after 5 days of co-cultivation compared to the growth of cells in a medium without conjugates fluctuated in a range of 80 to 55% depending on the type of pCAFs used (Fig. 3A). Fig. 3Tumor cell- killing activity of suicide gene exosomes. (A) Assessment of the killing effect of various yCD::UPRT-MSCs-CM-5-FC, yCD::UPRT-pCAFs-CM-5-FC conjugates, and a combination of both tested in simulated pancreatic carcinoma composed of a mixture of MIA PaCa-2 cells and individual pCAFs. (B) Growth assessment of human omentum fibroblasts treated with CM of yCD::UPRT-pCAFs and yCD::UPRT-MSCs, and with corresponding exosomes. (C) Assessment of growth-inhibiting activity of CM and exosomes from four isolates of naive and yCD::UPRT gene transduced pCAFs and two types of MSCs in the presence/absence of prodrug 5-FC in three pancreatic cancer cell lines. (D) Assessment of tumor cell growth-inhibiting activity of yCD::UPRT-MSCs-exosomes tested on cells of PDAC cell lines and several pCAFs. (E) Toxicity test of the 5-FC conjugates yCD::UPRT-pCAFs, yCD::UPRT-UC-MSCs, yCD::UPRT-DP-MSCs, and a control in SCID mice Tumor cell- killing activity of suicide gene exosomes. (A) Assessment of the killing effect of various yCD::UPRT-MSCs-CM-5-FC, yCD::UPRT-pCAFs-CM-5-FC conjugates, and a combination of both tested in simulated pancreatic carcinoma composed of a mixture of MIA PaCa-2 cells and individual pCAFs. (B) Growth assessment of human omentum fibroblasts treated with CM of yCD::UPRT-pCAFs and yCD::UPRT-MSCs, and with corresponding exosomes. (C) Assessment of growth-inhibiting activity of CM and exosomes from four isolates of naive and yCD::UPRT gene transduced pCAFs and two types of MSCs in the presence/absence of prodrug 5-FC in three pancreatic cancer cell lines. (D) Assessment of tumor cell growth-inhibiting activity of yCD::UPRT-MSCs-exosomes tested on cells of PDAC cell lines and several pCAFs. (E) Toxicity test of the 5-FC conjugates yCD::UPRT-pCAFs, yCD::UPRT-UC-MSCs, yCD::UPRT-DP-MSCs, and a control in SCID mice Assessment of growth-inhibiting activity of CM and exosomes from four isolates of Naive and yCD::UPRT gene transduced pCAFs and two types of MSCs in the presence/absence of prodrug 5-FC in three pancreatic cancer cell lines All cells secrete exosomes to reiterate the original homing feature of their mother cells. It is expected that yCD::UPRT-pCAFs-exosomes should be oriented to the PDAC tumor stroma and cooperate more specifically in the growth inhibition of PDAC tumor cells. pCAFs were isolated from histologically different types of PDAC. To determine, how specific each yCD::UPRT-pCAFs-exosomes is in the growth inhibition of various pancreatic tumor cell lines (SU.86.86, MIA PaCa-2, and BxPC-3) we compared CM and isolated exosomes for their tumor cell growth inhibiting activity in the presence of prodrug 5-FC (Fig. 3C). In addition, the biological activity of exosomes secreted from pCAFs was assessed. Both types of exosomes were isolated and concentrated by tangential flow filtration (TFF) with 5 nm holes that separated exosomes (retentate) from small biologicals (permeate). The tumor cell-killing activity was assessed using exosomes secreted for 24 h into a medium free of PL. The isolation of exosomes via TFF led to their concentration, therefore for the assays, the number of exosomes was adjusted to be equal to the original. The CM and exosomes from all naive pCAFs did not affect the tested pancreatic cell lines. However, the CM and exosomes of the yCD::UPRT-pCAFs27 and yCD::UPRT-pCAFs31 in the presence of 5-FC inhibited the growth of epithelial-like cells SU.86.86 by over 80%. pCAFs31 and pCAFs27 were less effective in growth inhibiting, resulting in growth inhibition of Su.86.86 cells in the 65% to 75% range. We also tested the CM and exosomes from DP-MSCs, AT-MSCs, and cells transduced with the yCD::UPRT gene. The CM and exosomes from naive cells did not affect the growth of SU.86.86 cells. However, the CM and exosomes from yCD::UPRT gene transduced MSCs in the presence of 5-FC, inhibited the tumor cell growth to a similar extent as yCD::UPRT-pCAFs78 and yCD::UPRT-pCAFs79. The growth of cells MIA PaCa-2 cell line, capable of inducing metastases, reacted differently. The tumor cell growth inhibition with CM from yCD::UPRT-pCAFs27 and yCD::UPRT-pCAFs31 was around 70%. The CM and exosomes of naive pCAFs, as well as from yCD::UPRT gene transduced cells, did not stimulate the growth of MIA PaCa-2 cells. On the other hand, CM, exosomes, and permeate of naive pCAFs78 and yCD::UPRT-pCAFs78 stimulated the growth of MIA PaCa-2 cells, indicating that the permeate contains small growth-stimulating factors. The CM of DP-MSCs and the permeate of yCD::UPRT-DP-MSCs significantly stimulated the growth of MIA PaCa-2 cells. Consequently, following this growth stimulation, the CM and exosomes from the yCD::UPRT gene transduced inhibited the growth of MIA PaCa-2 cells by about 30%. In contrast, CM and exosomes from yCD::UPRT-AT-MSCs inhibited the growth of MIA PaCa-2 cells by about 65%. Cells of pancreatic cancer cell line BxPC-3 lack a KRAS mutation, though it is commonly found in pancreatic cancers. BxPC-3 cells have high expression of cancer stem cell markers, and are tumorigenic in animals, despite being a KRAS-wild type cell line. The CM, exosomes, and permeate from all naive pCAFs and yCD::UPRT transduced ones slightly stimulated the growth of BxPC-3 cells. The CM and exosomes of all tested yCD::UPRT-pCAFs in the presence of 5-FC inhibited the growth in the range of 60 to 40%. The CM and exosomes from yCD::UPRT-AT-MSCs and yCD::UPRT-DP-MSCs inhibited the growth of BxPC-3 cells very effectively. The yCD::UPRT-MSCs-exosomes eradicated 85% of BxPC-3 cells. The naive pCAFs-exosomes in the form of TFF permeate did not inhibit the growth of tumor cell lines regardless of the presence of 5-FC. The overview of growth-inhibiting and stimulating activities of CM and isolated exosomes is presented in Fig. 3C. In general, the data revealed the individual character of isolated pCAFs reflected in the biological activity of secreted exosomes. All suicide gene exosomes from transduced pCAFs and MSCs in the presence of prodrug killed pancreatic tumor cell lines to various extents. The individuality of pCAFs was also mirrored in cancer cell growth support by some pCAFs and MSCs, but this was always overcome by the suicide gene exosomes conjugated with 5-FC. Pancreatic tumor cell growth-inhibiting activity of yCD::UPRT-MSCs-exosomes The biological activity of 24 h CM and exosomes isolated from it by TFF technology, adjusted to match the original in terms of the number of exosomes, was assayed on several tumor and normal cells. The obtained data are presented in a heat map (Fig. 3D). The highest killing effect of yCD::UPRT-MSCs-exosomes was observed in BxPC-3 and epithelial-like cells SU.86.86 cells, where the resulting growth inhibition of these cells exceeded 80%. Similar results were obtained with metastasis-inducing cancer cell line MIA PaCa-2, where the tumor growth inhibition was around 70%. pCAFs were found to be quite resistant to yCD::UPRT-MSCs-exosomes. Exosomes secreted from naive AT-MSCs and DP-MSCs were compared with exosomes secreted from corresponding yCD::UPRT gene transduced MSCs. Cryo-EM demonstrated particles of the expected size for exosomes for both naive and yCD::UPRT-gene transduced MSCs (Fig. 4A). Similar observation was found for the exosomes from all MSCs and yCD::UPRT-MSCs- exosomes used in this study. Tumor targeting specificity of yCD::UPRT-pCAFs-exosomes and yCD::UPRT-MSCs-exosomes in vitro and in vivo To see how specific the yCD::UPRT-pCAFs-exosomes are in the cell-targeting, we tested their biological activity in human primary omentum fibroblasts. The CM and exosomes from all tested yCD::UPRT-pCAFs slightly stimulated the growth of omentum fibroblasts in the absence of 5-FC, while CM and exosomes from yCD::UPRT-AT-MSCs and DP-MSCs did not (Fig. 3B). Both CM and exosomes slightly inhibited growth in the presence of the prodrug. To find whether the toxicity data from in vitro assays correspond to conditions in vivo, we tested the toxicity of the 24-hour conjugates of yCD::UPRT-pCAFs, yCD::UPRT-UC-MSCs, yCD::UPRT-DP-MSCs, and control in a group of 6 old SCID mice. The animals of mixed sex were injected with a 0.5 ml dose intraperitoneal every second day for 5 weeks. There was no weight decrease in animals during the experiment, with at least a 1 g increase in all mice groups at the end (Fig. 3E). Size-exclusion chromatography of yCD::UPRT-pCAFs-exosomes and yCD::UPRT-pCAFs 5-FC conjugated exosomes Adding prodrug 5-FC to CM from all yCD-UPRT-pCAFs led to the conjugation of 5-FC to exosomes. The assessment of the biological activity of fractions from size-exclusion chromatography on an IzonQev10/35nm column in the presence/absence of 5-FC on MIA PaCa-2 cells revealed the existence of nanosize particles (fraction No. 9, 10, and 11) possessing high activity in stimulating the growth of tumor cells, but inhibiting their growth when additional 5-FC was supplemented (Fig. 4B). This finding provides evidence for the presence of fused yeast cytosine deaminase phosphoribosyl transferase enzyme in these fractions. The proteomic data confirmed the presence of peptides identifying yeast cytosine deaminase and phosphoribosyl transferase, which is compatible with the character of exomeres [24, 25]. Fig. 4(A) Cryo-EM images of AT-MSCs-exosomes, yCD::UPRT-AT-MSCs-exosomes, DP-MSCs-exosomes, and yCD::UPRT-DP-MSCs-exosomes. (B) Size-exclusion chromatography comparison of yCD::UPRT-pCAFs-CM-exosomes and yCD::UPRT-pCAFs-exosomes 5-FC conjugated. Detection of the growth-inhibiting activity of fractions on MIA PaCa-2 cells. (C)The killing effect of simulated pancreatic carcinoma composed of a mixture of MIA PaCa-2 + pCAFs27 and a mixture of MIA PaCa-2 + pCAFs31 in a ratio of one to five, treated with yCD::UPRT-UC-MSCs-CM-5-FC-conjugate, and yCD::UPRT-pCAFs-5-FC pooled conjugate in the presence or absence of 5-FC. The treatment was repeated. (D)The endpoint of the killing effect of simulated pancreatic carcinoma composed of a mixture of MIA PaCa-2 + pCAFs27 and a mixture of MIA PaCa-2 + pCAFs31 in a ratio of one to five, treated with yCD::UPRT-UC-MSCs-CM-5-FC-conjugate, and yCD::UPRT-pCAFs-5-FC pooled conjugate, evaluated by MTT assay. (E) The endpoint of the killing effect of simulated pancreatic carcinoma composed from a mixture of MIA PaCa-2 + pCAFs78 and a mixture of MIA PaCa-2 + pCAFs79 in a ratio of one to five treated with yCD::UPRT-UC-MSCs-CM-5-FC conjugate, and yCD::UPRT-pCAFs-5-FC pooled conjugate evaluated by MTT assay. The statistical significance: P < 0.05 - *; P < 0.01 - **; P < 0.001 - ***; P < 0.0001 - **** (A) Cryo-EM images of AT-MSCs-exosomes, yCD::UPRT-AT-MSCs-exosomes, DP-MSCs-exosomes, and yCD::UPRT-DP-MSCs-exosomes. (B) Size-exclusion chromatography comparison of yCD::UPRT-pCAFs-CM-exosomes and yCD::UPRT-pCAFs-exosomes 5-FC conjugated. Detection of the growth-inhibiting activity of fractions on MIA PaCa-2 cells. (C)The killing effect of simulated pancreatic carcinoma composed of a mixture of MIA PaCa-2 + pCAFs27 and a mixture of MIA PaCa-2 + pCAFs31 in a ratio of one to five, treated with yCD::UPRT-UC-MSCs-CM-5-FC-conjugate, and yCD::UPRT-pCAFs-5-FC pooled conjugate in the presence or absence of 5-FC. The treatment was repeated. (D)The endpoint of the killing effect of simulated pancreatic carcinoma composed of a mixture of MIA PaCa-2 + pCAFs27 and a mixture of MIA PaCa-2 + pCAFs31 in a ratio of one to five, treated with yCD::UPRT-UC-MSCs-CM-5-FC-conjugate, and yCD::UPRT-pCAFs-5-FC pooled conjugate, evaluated by MTT assay. (E) The endpoint of the killing effect of simulated pancreatic carcinoma composed from a mixture of MIA PaCa-2 + pCAFs78 and a mixture of MIA PaCa-2 + pCAFs79 in a ratio of one to five treated with yCD::UPRT-UC-MSCs-CM-5-FC conjugate, and yCD::UPRT-pCAFs-5-FC pooled conjugate evaluated by MTT assay. The statistical significance: P < 0.05 - *; P < 0.01 - **; P < 0.001 - ***; P < 0.0001 - **** Killing efficacy of various yCD::UPRT-MSCs-CM-5-FC and yCD::UPRT-pCAFs-CM-5-FC conjugates tested in a mixture of pCAFs and MIA PaCa-2 cells as a simulation of pancreatic carcinoma’s feature in vitro To simulate the PDAC, composed of tumor cells and an overwhelming number of stromal cells, we mixed spheroids of MIA PaCa-2 cells with naive pCAFs27, pCAFs31, pCAFs78, and pCAFs79 in a ratio of one to five, independently. The simulated pancreatic tumors of 7 × 103 cells were plated into the 96-well clear round bottom cell adherent microplate and independently into the ultra-low attachment 96-well microplate, thus attempting to simulate the cell neighborhoods communication that was found to be important during chemotherapeutic treatment [21]. The mixture of the cells in the well of 96-well microplates was treated with yCD::UPRT-UC-MSCs-CM-5-FC and yCD::UPRT-pCAFs-CM-5-FC conjugates. The course of the treatment was evaluated by an IncuCyte® live cell monitoring system (Fig. 4C), and at the end of the treatment by MTT assay (Fig. 4D and E). To approximate more closely the usual therapeutic approach consisting of repeating drug applications, we subjected the simulated pancreatic carcinoma constructed from MIA PaCa-2 cells and pCAFs to a new second dose of treatment with appropriate agents in half of the treatment course. The obtained data revealed efficient killing of the simulated pancreatic carcinoma in all tested PCAFs-combinations with yCD::UPRT-UC-MSCs-CM-5-FC conjugate or yCD::UPRT-pCAFs-CM-5-FC conjugate versus controls (P < 0.0001). Repeated treatment with the same agents progressed the killing effect. The high killing efficiency of yCD::UPRT-UC-MSCs-CM-5-FC conjugate and yCD::UPRT-pCAFs-CM-5-FC conjugate did not differ substantially (P < 0.05). A subsequent addition of the prodrug 5-FC did not alter effect (P < 0.05). The different end-point killing efficiencies of simulated PDAC tumors reflect the distinct features of individual pCAFs. Discussion Our study demonstrates promising therapeutic potential of suicide gene exosomes derived from MSCs and from pCAFs when used in a combined action in preclinical simulated desmoplastic model of the PDAC. The therapeutic benefits of MSCs are largely dependent on paracrine factors, mainly their secreted exosomes. MSCs isolated from various tissues and their exosomes possess an inherent tumor homing capability. It was shown that the proteomes of MSCs derived from the bone marrow, placenta, and umbilical cord differ, which was reflected in their different migration abilities [22] and depends on the type of target cells to which they are directed [23]. Tissue origin of MSCs and secreted exosomes differ in the reparatory and therapeutic capabilities, with those of younger embryonic origin appearing more effective. For instance, DP-MSCs differ from adipose and bone marrow derived MSCs in the protein expression of some pluripotent stem cells genes. DP-MSCs being developed from ecto-mesenchymal elements and containing neural crest-derived cells display plasticity and multi-potential capabilities [20, 24]. Differences in MSCs from different tissue origin are reflected in the therapeutic abilities of their secreted exosomes. Exosomes secreted from evolutionary young MSCs, like plac-MSCs exosomes exhibit high therapeutic activities [25]. Plac MSC-exosomes play a role in pro-angiogenic activity, suggesting that they may be a novel therapeutic approach for treatment of diseases associated with ischemia, including cancer [15, 26]. Transduction of MSCs from various origin with the suicide gene yCD::UPRT has been shown to maintain their tissue-specific characteristics together with sustained tumor cell therapeutic capability. Recently, we demonstrated the therapeutic potential of yCD::UPRT-Plac-MSCs-exosomes in preclinical PDAC models [15]. Exosomes secreted from pCAFs execute not only the cross-talk between tumor cells and other stromal cells of pancreatic TME, but also are involved in PDAC pre-metastatic niche formation in the liver and lungs. Tumor-derived exosomes taken up by organ-specific cells prepare the pre-metastatic niche [27]. We proved that PCAFs transduced with yCD::UPRT secrete gene suicide gene exosomes, similarly as MSCs. They represent an attractive option for a novel platform to treatment of PDAC, similar to smart exosomes [28]. In this study, we have tested whether the combined action of suicide gene exosomes derived from yCD::UPRT gene transduced pCAFs and MSCs is capable of penetrating the desmoplastic nature of pancreatic carcinomas. We have shown through in vitro experiments using simulated desmoplastic pancreatic tumors that the combined action of these targeted exosomes represents a potential strategy for treating pancreatic ductal adenocarcinoma. The obtained results are in an agreement with our translational studies with yCD::UPRT-MSCs in nude mice. Furthermore, the expression of the suicide gene in MSCs driven by strong retrovirus promoters and its translation results in the secretion of exosomes containing gene’s mRNA [13]. Suicide gene exosomes act like a Trojan horse sneaking into cancer cells and killing them from within, by converting the non-toxic prodrug 5-fluorocytosine into the cytotoxic 5-fluorouracil. pCAFs isolated from PDAC specimens and transduced with yCD::UPRT suicide gene behaved similarly, releasing targeted suicide gene exosomes. The addition of prodrug 5-FC to the CM of yCD::UPRT-MSCs or yCD::UPRT-pCAFs leads to the conjugation of 5-FC to the nanoparticles in the CM. Size-exclusion chromatography revealed the existence of nanoparticles resembling described exomeres [29, 30]. We believe that intravenous application of suicide gene exosomes from pCAFs after surgical resection of PDAC could damage the pre-metastatic niche and thus prevent or inhibit metastasis [31]. All these suicide gene nanoparticles in the presence of prodrug 5-FC caused apoptotic cell death intracellularly dose-dependently, thereby mimicking the effects of conventional drugs. The therapeutic potential of yCD::UPRT-MSCs-exosomes can be potentiated by chemotherapeutic drugs. We successfully incorporated gemcitabine (GCB) into yCD::UPRT-DP-MSCs-exosomes, which led to efficient growth inhibition of the pancreatic cell lines PANC-1 and MIA Paca-2 in vitro [32]. The pancreatic carcinoma microenvironment is rather complex; therefore, it might be of interest to test a mixture of suicide gene exosomes from different tissue specific MSCs, expecting further improvement in their therapeutic activity. The emergence of chemotherapy-resistant tumor cells poses a significant obstacle to the effectiveness of cancer chemotherapy. It was reported that the incorporation of the yCD::UPRT transgene into 5-FU resistant colorectal cells, which are derived from colorectal adenocarcinoma, has successfully overcome this chemoresistance [33]. Exosomes containing suicide genes from MSCs or pCAFs, delivered as a 5-FC conjugate, may represent a promising treatment option for resistant carcinomas. The research of exosomes as drug delivery vehicles targeted to pancreatic tumor cells or tumor adjacent fibroblasts has intensified [34, 35]. Several reviews on the therapeutic potential of modified exosomes or other components of the MSCs secretome has appeared recently [36, 37]. Conclusions and future considerations CAFs have long been considered an attractive therapeutic target by drugs, but a poor understanding of these heterogeneous cell populations has limited the development of effective treatments. CAFs of pancreatic cancer (pCAFs) despite many reports of combined chemotherapies, including targeted therapies, still fail to improve PDAC therapy. Our study provides experimental evidence that suicide gene exosomes homing in pCAFs can kill them by disrupting the desmoplastic pancreatic carcinoma and allowing MSCs-derived suicide gene exosomes to cause cell death to tumor cells by intracellular conversion of the prodrug to an effective anti-cancer drug. The current study had several limitations. The pCAFs heterogeneity, their changes during tumor development, and their patient-specific character may limit their use in proposed PDAC treatment. However, all isolated pCAFs from different histopathologic tumors were capable of being transduced with the suicide gene. All secreted suicide gene exosomes inhibited the growth of various pCAFs creating a hope that they may act more broadly. Therefore, an individualized tumor/patient-specific treatment protocol may not be required. The proposed PDAC innovative treatment approach was tested on metastatic pancreatic cancer cell lines. We realize the necessity of choosing the right standardized methodology for suicide gene exosome isolation, purification, concentration, large-scale production quantification, and characterization, as next steps. Our long-lasting laboratory working experience with exosomes will help us achieve this. We are considering to expand the translational studies in PDAC preclinical models, such as patient-derived xenografts (PDX), based on the experience we obtained with yCD::UPRT-Plac-MSC-cCM [15]. It will include testing of primary pancreatic cancer cells from surgical resections in experiments in vitro and in vivo. PDX can answer the presumed therapeutic ability of the combined action of suicide gene exosomes derived from MSCs and pCAFs. The orthotopic application of metastatic pancreatic cancer cells MIA PaCa-2 is leading to liver metastases in SCID mice. Animal model opens a possibility of finding whether intravenous injections of pCAFs-derived suicide gene exosomes can damage the pre-metastatic niche and thus inhibit or prevent metastasis.