Pancreatic cancer incidence is rising both in Europe (Partyka et al., 2023) and the USA (Siegel et al., 2023), with the lowest 5-year relative survival rate (12 %) among all cancer types. In particular, pancreatic ductal adenocarcinoma (PDAC), the prevalent type of pancreatic cancer, is one of the deadliest cancers, with 90 % of PDAC patients dying within a year of diagnosis, as reported by the World Health Organization (WHO) (Li et al., 2020). The disease is asymptomatic during its early stages; patients are often diagnosed with metastasized stage IV with no option for surgery (Lev et al., 2017).

Gemcitabine hydrochloride (GEM), approved by the Food and Drug Administration (FDA) in 1996, was historically used as monotherapy for unresectable and metastatic PDAC despite a modest improvement in survival compared to 5-fluorouracil (FU) (Frezza et al., 2008). Nowadays, current standard-of-care treatments, which mainly consist of GEM plus nab-paclitaxel or FOLFIRINOX [a combination of leucovorin calcium (folinic acid), FU, irinotecan, and oxaliplatin], can profoundly improve the prognosis of advanced PDAC (Conroy et al., 2022; Klein-Brill et al., 2022).

However, intrinsic or acquired chemoresistance combined with a stiff desmoplastic pancreatic tumor stroma, responsible for the drug's poor penetration, contributes to poor clinical efficacy and a high risk of side effects (Zeng et al., 2019).

Most of the studies on PDAC chemoresistance have focused on GEM resistance, and several mechanisms have been described (Amrutkar and Gladhaug, 2017; Adamska et al., 2018, Zeng et al., 2019). They can affect GEM cellular accumulation or metabolism, ranging from downregulation of the GEM transporter to upregulation of drug efflux pumps or impairment of enzymatic functions involved in the metabolism of the drug, such as upregulation of the cytidine deaminase activity able to metabolize GEM very quickly into inactive compounds (Baker et al., 2013).

In addition to drug metabolization studies, other mechanisms can be involved in GEM chemoresistance. The stiffness of the desmoplastic pancreatic tumor stroma can induce epithelial-mesenchymal transition (EMT) (Rice et al., 2017), and a sustained EMT via upregulation of the mesenchymal transcriptional factors is critical not only for local invasion but also for imparting GEM resistance (Adamska et al., 2018). Overexpression of mutated KRAS oncogene, registered in 90 % of pancreatic cancer, also correlated with activation of Nuclear factor erythroid 2-related factor 2 (Nrf2), the master regulator of the antioxidant responses (DeNicola et al., 2011; Barrera et al., 2021). Nrf2, highly expressed in GEM-resistant pancreatic cells, can also crosstalk with YAP, a critical Hippo pathway effector involved in cancer progression and drug resistance (Grattarola et al., 2021). Another major player in the GEM chemoresistance mechanisms is the tumor-related glycoprotein HAb18G/CD147, whose expression can be induced by GEM, eliciting increased invasion and metastatic cell phenotype (Xu et al., 2016).

Several strategies can overcome chemoresistance, including those targeting impaired signal pathways (Yang et al., 2018, Grattarola et al., 2021, Hosein et al., 2022, Zhang et al., 2022). However, in the last decades, nanosized drug delivery systems have been proposed as a solution to overcome poor efficacy and resistance issues, allowing site-specific delivery and improvement of drug bioavailability of encapsulated medicines (Shi et al., 2017). With this aim, GEM has been loaded in liposomes (Tamam et al., 2019), polymeric nanoparticles, or solid lipid nanoparticles (Cai et al., 2021), always showing enhanced antitumoral effects compared to free GEM. Furthermore, nanomedicines represent a valid technology for combination therapy, allowing for a highly efficient treatment that can counteract different mechanisms of cancer progression and resistance (Alshememry et al., 2022, Boggio et al., 2023).

Pancreatic cancer is resistant not only to chemotherapy but also to immunotherapy. Indeed, pancreatic cancer presents an immunosuppressive microenvironment (Falcomatà et al., 2023), and less than 1 % of PDAC patients are sensitive to immune-checkpoint blockade via the PD-1/PD-L1 pathway (Le et al., 2015). The mechanisms of immune suppression in PDAC have different origins, such as the desmoplastic stroma that produces a hypoxic environment that inhibits the CD8 + T-cell and the infiltration of immune suppressor cells (mainly myeloid-derived suppressor cells and tumor-promoting M2-polarized Tumor-associated Macrophages) (Ullman et al., 2022). To date, several immunotherapeutic approaches have been tested in pre-clinical and clinical studies, including immune checkpoint modulators, chimeric antigen receptor CAR-T cells, and cancer vaccines (Yoon et al., 2021). Nevertheless, all the efforts in developing novel immunotherapies did not lead to a breakthrough for PDAC patients and left the need to investigate alternative strategies open.

Recently, the inducible T-cell co-stimulator (ICOS)/ICOS ligand (ICOSL) signaling has emerged as an interesting immunomodulatory pathway with anti-tumoral activity (Zhang et al., 2016; Clemente et al., 2020, Raineri et al., 2020). ICOS is a T-cell costimulatory molecule detected initially in activated T cells (Redoglia et al., 1996, Buonfiglio et al., 2000). ICOS binds to the CD28 family member ICOSL (also named B7H), which is expressed in several cells of the immune system, such as dendritic cells (DCs), macrophages, B cells, but also endothelial cells (ECs), epithelial cells, and fibroblasts (Sharpe and Freeman, 2002). Besides the regulation of immune responses (Nurieva, 2005), ICOS/ICOSL is also involved in regulating the tumor immune microenvironment, as demonstrated in colorectal cancer (Zhang et al., 2016), multiple myeloma (Boggio et al., 2021; ), melanoma (Fu and He, 2011), and pancreatic cancer (Blair et al., 2022). Moreover, ICOSL triggering by a bivalent soluble form of ICOS, consisting of the Fc portion of IgG1 and two molecules of the extracellular portions of ICOS (ICOS-Fc), inhibited adhesiveness and migration of DCs, ECs, and several cancer cells (Dianzani et al., 2010; Occhipinti et al., 2013; Dianzani et al., 2014), as well as HUVEC tubulogenesis (Clemente et al., 2020). Interestingly, in vivo treatment with ICOS-Fc inhibited lung metastatization of several types of tumor cell lines in mice, but it failed to inhibit the growth of primary tumors. Conversely, ICOS-Fc loaded in nanoparticles inhibited the growth of established subcutaneous B16-F10 melanoma (Clemente et al., 2020).

One of the key issues in developing novel therapies, including nanomedicines for PDAC treatment, is the lack of clinically relevant animal models of the disease. PDAC is a type of tumor characterized by high inter- and intra-tumoral heterogeneity, high stromal component, and infiltration of immunosuppressive cells. These features are very difficult to represent by preclinical models consisting of human or murine cell lines, patient-derived cells, or genetically engineered mouse models (GEMMs) (Yu et al., 2021). Substantial differences between 2D cultures and actual tumors lie in their different interactions with the microenvironment, changes in cell polarity or spatial gradients of nutrients; these are well-known parameters able to regulate several crucial steps of cancer development, such as tissue organization, cell motility, proliferation, and metastatic diffusion (Lo et al., 2000, Cavo et al., 2018). Organoids and spheroids have been shown to resemble many physiological aspects better than cells grown in monolayers; moreover, 3D scaffold-based cancer models can further improve the in vivo mimicking since they resemble the extracellular Matrix (ECM), which is an important component of the tumor microenvironment (TME), playing crucial roles in cancer progression and invasion (Dutta and Dutta, 2009; Marrella et al., 2021). Finally, to better recapitulate the whole in vivo scenario, fluid-dynamic conditions were added to 3D cultures (Marrella, 2021). These innovative organ-on-chip technologies can reproduce the physiological blood flow element, which can affect cell survival and metastatic potential, alongside a reliable drug distribution (Hoarau-Véchot et al., 2018, Trujillo-de Santiago et al., 2019). These models offer the opportunity to ameliorate the evaluation of tumor cell behavior and to perform investigations such as dynamic skin or gastrointestinal drugs absorption assay, anti-proliferative assay, 3D cancer model immune cells infiltration, as well as dynamic migration assay (Cavo et al., 2018, Marrella et al., 2020; Marrella et al., 2021; Pulsoni et al., 2022).

The aim of this work was to evaluate β-cyclodextrin-based nanosponges (NS) loaded with GEM and decorated with ICOS-Fc as an innovative combination therapy, potentially suitable to overcome drug resistance in pancreatic cancer. NS-GEM effects on viability, proliferation, and invasion were assessed in the 2D and 3D standard cultures of pancreatic cancer cells. Moreover, in order to increase the anti-invasion properties, NS-GEM were functionalized with the immunomodulant agent ICOS-Fc. The MULTI-ORGAN single flow device (MIVO®) technology was exploited to confirm the 2D anti-invasion effect, raising the complexity of the system and achieving evidence closer to the in vivo scenario.