Cancer bio-immunotherapy XVI annual NIBIT-(Italian Network for Tumor Biotherapy) meeting, October 11–13 2018, Milan, Italy

Introduction

The XVI annual meeting of the Italian Network for Tumor Biotherapy (NIBIT) took place in Milan, Italy, on October 11–13, 2018. International leader scientists from academic institutions and pharmaceutical industries presented updates of the new research achievements on cancer bio-immunotherapy. The meeting aimed at building close connections between clinical and pre-clinical research. Many topics were discussed during the meeting, such as (a) Tumor microenvironment; (b) Adoptive cell therapy with CAR-T cells; (c) Commensal microorganisms, metabolism, immune cells and cancer; (d) Cancer vaccines, elicited T cells and their interaction with tumor cells; (e) Novel insights into the biology of anti-tumor T cells; (f) Prospective NIBIT collaborations. A brief summary of the topics discussed in the meeting is reported.

Session 1: tumor microenvironment, part I

The tumor microenvironment (TME) has a central role in restraining or unleashing tumor growth. Hence, a deep understanding of the dynamic interplay between cancer cells and their local environment is of paramount importance in designing the immunotherapy that best fits to each tumor. Wolf-Herman Fridman (Paris, France) proposed the TME as a dialog window between cancer cells and host cells. Growing evidence links the characteristics of the tumor immune infiltrate with the clinical outcome of the disease. TME is composed by several different types of cells, thus making investigation of its composition and function rather complex. A case was made for clear cell renal cell carcinoma (ccRCC), which is enriched in CD8⁺ TILs, which associate with poor clinical outcome, exhibits high expression of immune checkpoints and their ligands, abundant infiltration with dysfunctional dendritic cells (DC) and scarce tertiary lymphoid structures [1]. Fridman also illustrated the importance of transcriptomic analysis that allowed inter-sample comparisons together with the identification of biomarkers able to identify patients at high risk of disease recurrence and potentially sensitive to checkpoint blockade. Niels Halama (Heidelberg, Germany) discussed the need to characterize the immunosuppressive role of the TME in micro-satellite-stable (MSS) colorectal cancer (CRC) characterized by the resistance to all approved immunotherapy approaches. The immunohistochemistry analysis of solid tumor tissue slides is the gold standard technique to assess tumor immune infiltrate, since it allows for precise quantification of type, density and localization of immune cells [2]. Based on IHC studies, different types of immune phenotypes can be identified: “cold” tumors (or “immune desert” showing no immune cell infiltration), “immune-excluded” tumors (with immune cells aggregating at the tumor boundaries) and “hot” tumors (inflamed tumors). The cross-talk between cancer cells and immune cells may be cause of cancer immune evasion and progression. Different treatment approaches, such as PD-1/PD-L1 blockade or chemokine modulations, are able to successfully modify these interactions toward rejection or suppression of tumor cells. Given the large number of approved or candidate drugs able to control key aspects of tumor growth, the comparison of all immunotherapeutic agents in a clinical setting is not easy. In silico models of tumor growth can be designed to mimic different cellular interactions. In particular, Halama showed a spatial agent-based approach to represent tumor cells, fibrotic stroma, lymphocytes and macrophages in human metastatic colorectal cancer, able to investigate several combinations of treatments achieving optimal response rates in different tumors. This model represents a potent tool for a prospective validation. Alexandre Harari (Lausanne, Switzerland) defined a workflow for neoantigen identification and presented data about the spontaneous recognition of tumor neo-epitopes in immunotherapy-naive, chemotherapy-pretreated patients with recurrent and advanced epithelial ovarian cancer (EOC). Through peripheral blood analysis, Harari identified T-cell precursors recognizing neo-epitopes in 1/3 of the patients, while identification of these cells was more frequent using TILs. The data provided evidences that the repertoire of neo-epitope recognizing T cells can be largely discordant between blood and TILs. Since TILs exert a higher sensitivity to cognate neo-epitopes than their blood counterparts, this information has important implications to drive the choice of the source of T cells in the development of adoptive T-cell therapy [3]. Arianna Calcinotto (Bellinzona, Switzerland) showed how treatments that block IL-23 can oppose myeloid-derived suppressor cell (MDSC)-mediated resistance to castration in prostate cancer and synergize with standard therapies. This finding could have important clinical implications towards the understanding of the biological processes that underlie castration resistance [4]. Finally, Tiziana Triulzi (Milan, Italy) showed that the de-differentiation of adipocyte, the major breast component, caused by breast cancer (BC) cells, resulted in provision of pro-tumorigenic stimuli to BC cells and that the blockade of this program can hamper tumor progression.

Session 2: tumor microenvironment, part II

Session 2 was mainly focused on the role of myeloid cells in promoting tumor growth and invasion. Tumor growth induces massive myelopoiesis, that includes the expansion of myeloid-derived suppressor cells (MDSCs). As discussed by Vincenzo Bronte (Verona, Italy), the importance of depleting MDSCs in the TME could be crucial to restoring anti-tumor immunity. Cellular FLICE (FADD-like IL-1β-converting enzyme)-inhibitory protein (c-FLIP) confers to MDSCs chemotherapy resistance and the ability to induce tolerance towards cancer cells. However, the number of circulating CD11b⁺Ly6G⁻Ly6C^high cells can be diminished using low doses of diverse chemotherapeutic drugs that induce a lowering of c-FLIP expression mediated by the activation of the extrinsic apoptotic pathway, thereby enhancing the efficacy of adoptive cell transfer (ACT). Accordingly, engineering CD14⁺ human monocytes to overexpress c-FLIP renders them immunosuppressive and capable of controlling established graft versus host disease (GvHD). This effect is due to the ability of c-FLIP to activate a broad inflammatory pathway in myeloid cells and to trigger PD-L1 upregulation through NF-kB activation. Indeed, patients with pancreatic cancer showing a high number of infiltrating Ly6C^highFLIP⁺CD14⁺ monocytes expressing PD-L1/2 have a negative overall survival and disease-free survival prognosis [5]. Together with MDSCs, tumor-infiltrating regulatory T lymphocytes (Tregs) suppress effector T-cell functions against cancer cells. Massimiliano Pagani (Milan, Italy) showed that type-1 Tregs infiltrating non-small cell lung cancer (NSCLC) and colorectal carcinoma (CRC) have a unique transcriptome. Indeed, these cells significantly up-regulate the expression of several immune checkpoints (GITR, OX40, TIGIT, LAG-3, and TIM-3) and their ligands, and show high immunosuppressive features compared to circulating and tissue-infiltrating Tregs, as well as compared to Th1 and Th17 cells, suggesting a crucial role of environmental factors in driving different gene expression profiles. At the single cell level, Treg signature genes are co-expressed with FoxP3 and IL-2RA, and this transcription profile is characteristic of both primary and metastatic human tumors. One key gene related to a bad prognosis is chitinase 3 like 2 (CHI3L2), which is overexpressed in patients affected by NSCLC, CRC, and melanoma correlating with tumor progression and reduced overall survival. Combining RNA sequencing to multiplexed tissue imaging would be of great interest to gauge the interaction between Tregs and the tumor “neighborhood”, confirming transcriptomic data [6]. Chemotherapy generally provides clinical benefits to patients diagnosed with cancers, but in some cases, it has proven pro-metastatic effects. Supporting this evidence, Michele De Palma (Lausanne, Switzerland) demonstrated in two different mouse model of mammary tumor (MMTV-PyMT transgenic mice and Rag1^−/− mice injected with 4T1 cells) that the taxane paclitaxel (PTX) enhanced the incidence and size of pulmonary metastasis compared to vehicle-treated mice. This effect was in part mediated by the upregulation of CCL2 in the lungs of tumor-bearing mice, which promoted the recruitment of pro-metastatic Ly6C⁺ monocytes. Furthermore, chemotherapy with either PTX or the anthracycline doxorubicin (DOX) elicited tumor-derived extracellular vesicles (EVs), which are known to foster seeding and growth of metastatic cancer cells, thus facilitating metastasis. The proteomic analysis of 4T1-derived EVs revealed that their enhanced pro-metastatic potential after exposure to chemotherapy was dependent on the enhanced intravesicular accumulation of annexin A6 (ANXA6). Indeed, genetic Anxa6 depletion from 4T1 cells disrupted the ability of the isolated EVs to induce CCL2 upregulation and Ly6C⁺, monocyte increase in the lung, indicating that EV-associated ANXA6 supports a pro-metastatic environment through CCL2 and pro-metastatic myeloid cells. Finally, initial clinical data indicated that ANXA6 was detected and increased in circulating EVs of breast cancer patients undergoing neoadjuvant chemotherapy [7]. Another contribution to the establishment of an immunosuppressive TME is given by mast cells (MCs) that can promote adenocarcinoma development in the prostate microenvironment. Elena Jacchetti (Milan, Italy) demonstrated that in mice developing spontaneous prostate adenocarcinoma (TRAMP) and deficient for MCs, the incidence of carcinoma was significantly reduced, due to the restored ability to mount a tumor-specific cytotoxic T-cell response. This recovered T-cell-mediated immunity was associated with reduced activity of CD40-expressing polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) that was dependent on the impairment of the interaction with MCs through CD40L. Therefore, CD40–CD40L interaction fosters an immunosuppressive microenvironment promoting tumor onset and growth, suggesting immunotherapy targeting CD40 as an opportunity in prostate cancer [8]. Unfortunately, immune checkpoint inhibitors (ICIs) do not work in all patients. Indeed, Michele Sommariva (Milan, Italy) reported that a fraction of non-small cell lung cancer (NSCLC) patients show disease hyper-progression (HP) once treated with PD-1/PD-L1 mAbs, and their tumors are characterized by infiltrating CD162⁺CD33⁺PD-L1⁺ M2-like tumor-associated macrophages (TAMs). Also in mice, the boost in tumor growth of tumor-bearing mice treated with anti-PD-1 antibodies paralleled with F4/80⁺CD206⁺ TAMs accumulation. Surprisingly, using anti-F(ab)2 fragments, the increase in tumor growth was abrogated, outlining that FcR engagement by ICIs on TAMs may reprogram immune cell functions shaping a pro-tumorigenic phenotype and fostering HP onset [9].

Session 3: adoptive immunotherapy

Adoptive immunotherapy, especially with genetically modified T lymphocytes, represents one of the most exciting innovative approaches for cancer treatment. In the field of hematological malignancies, in particular pediatric B acute lymphoid leukemia (B-ALL), the most challenging results have been achieved through the use of the Chimeric Antigen Receptor (CAR)-T-cell therapy. Franco Locatelli (Rome, Italy) invested in the field with two trials: (1) CD19.CAR for B-ALL and (2) GD-2.CAR for neuroblastoma. Since the infusion of CAR-T cells can cause both on-target off-tumor toxicity and non-antigen-associated toxicities, in the vector construct used for these trials the inducible Caspase 9 (iCas9) have been included. Pre-clinical studies demonstrated the safeness of the inducible safety switch, supporting its translation to the clinic. In GD-2.CAR-T cells therapy, all the seven infused patients displayed a relevant expansion of CAR-T cells and metastatic clearance, demonstrating significant anti-tumor activity. Grade 1–2 cytokine-release syndrome (CRS) and two cases of reversible neuropathy, but no cases of central nervous system (CNS) toxicities, have been reported. In B-ALL patients enrolled in the phase I trial, eight out of nine patients infused with CD19.CAR-T cells reached complete remission and no toxicities have been reported. Although CD19.CAR-T-cell therapy is very successful, a significant number of patients fail because of CD19-neg relapse. Marco Ruella (Philadelphia, PA, USA) reported the intriguing case of a patient relapsed with CD19-neg leukemia expressing CAR. During the manufacturing process, CD19.CAR was integrated in a single leukemic cell that unexpectedly survived the CAR-T-cell production protocol and was reinfused into the patient. The CD19.CAR expressed on the leukemic cell bound the CD19, masking the CD19 epitope to CD19.CAR-T cells, allowing the leukemic cells to persist over time and cause the relapse [10]. To counteract this “masking” relapse mechanism, an anti-CAR19.CAR has been developed. Anti-CAR19.CAR-T cells can selectively kill CD19.CAR⁺ B-ALL but could be also used as CD19.CAR-T depleting agent in case of toxicities. In all CD19.CAR-T clinical trials, the construct encoding the CAR is virally transferred into patient lymphocytes, which are expanded ex vivo and infused back into the preconditioned patient. This complex and expensive manufacturing process limits the CAR-T immunotherapy to specialized center only. Chiara Magnani (Monza, Italy) described an alternative non-viral approach that consists of the use of the Sleeping Beauty (SB) transposon system to transfer by electroporation CD19.CAR construct into Cytokine-Induced Killer (CIK) cells, an allogeneic cell population with NK cell activity, highly cytotoxic, and with no risk of GvHD. SB-CAR-CIK-CD19 cell manufacturing process was effective, feasible and reproducible in GMP, and the cellular product was active and well tolerated in vivo. The approach can be applied to additional target, like BAFF receptor, which expression is restricted to B-ALL blasts and normal B cells. The treatment of solid tumors with CAR-T cells is limited by the lack of antigens on the surface of tumor cells that are not expressed by normal cells. In the context of glioblastoma (GBM), Gianpietro Dotti (Chapel Hill, NC, USA) identified CSPG4 as relevant target for CAR-T cells. Indeed, this proteoglycan has a restricted distribution in normal tissues, and it is not only expressed on GBM tumor cells, but also in tumor vessels. Although CSPG4 expression is heterogenous on GBM tumor cells, in vivo, it is up-regulated by the TNF- produced by the microglia. CSPG4.CAR- T cells efficiently controlled GBM-derived cells, grown as neurospheres, both in vitro and in vivo in xenograft model with long-term survival [11]. Anti-tumor effects by CD19.CAR-T cells are accompanied by three main toxicities: (1) B cell aplasia, (2) CNS toxicities and (3) CRS. Margherita Norelli (Milan, Italy) developed an SGM3 mouse model with human hematopoiesis that reconstituted human T cells, which can be CAR-engineered and infused into co-engrafted mice without causing GvHD. The results obtained thanks to this model indicated that monocytes, and not CAR-T cells, producing IL-1 and subsequently IL-6 are the main drivers of CRS and CNS toxicities [12]. Another limitation for cancer immunotherapy is the expression of inhibitory signals in tumor microenvironment that lead to T-cell exhaustion and to the failure of tumor cell killing. Beatrice Cianciotti (Milan, Italy) developed a strategy to disrupt by CRISPR/Cas9 system the inhibitory receptors LAG-3 and TIM-3 in long-living memory T cells gene-edited and redirected against tumor antigens by TCR gene transfer. The innovative tumor-specific cellular products resistant to inhibitory signals are effective in killing cancer cells in vitro and in vivo in xenograft model. Tumor-specific T cells for ACT, besides being generated by genetic engineering of polyclonal T-cell population, can be derived by antigen-specific expansion. Daniela Montagna (Pavia, Italy) exploited donor-derived CTLs directed toward patient leukemia blasts for the treatment of high-risk patients upon haploidentical stem cell transplantation. The anti-leukemia CTL infusions both control and prevent post-transplant recurrences, also leading to long-term remission, and without causing severe adverse reactions, included severe GvHD.

Session 4: microbiota and metabolism

In the last years, the role of microbiota to define the efficacy of anti-tumor immunotherapy has emerged. The number of papers that proved a connection between microbiota, metabolism and immune system in cancer development is growing. Gut microbiota has been demonstrated to affect not only digestion and pathogen infection, but also nutrient availability in the host tissues. As a consequence, microbiota impairs cell proliferation, tumor cell expansion as well as maturation and differentiation of cellular component of the immune system, thus affecting cancer onset, progression and response to therapeutics. Guido Kroemer (Paris, France) demonstrated that, according to the composition of gut microbiota, cancer patients can be stratified as responder and non-responder to immune checkpoint blockade. Analysis of gut microbiota in 100 patients affected by NSCLC, RCC or urothelial cancer revealed that significantly higher levels of Akkermansia muciniphila associated with better clinical outcome or progression-free survival higher than 3 months, after treatment with anti-PD1. Similarly, in melanoma patients, Fecalibacterium spp. or Bifidobacterium spp. has been demonstrated to be associated with better prognosis. These results suggest a specific microbiota fingerprint, related to the cancer histological type, connected with immunotherapy response [13]. Also, Luigi Nezi (Milan, Italy), analyzing gut and fecal samples from melanoma patients treated with anti-PD1, demonstrated a different microbiota composition between responder and non-responder patients. In particular, patients who responded to immunotherapy presented a higher concentration of Faecalibacterium spp. (Ruminococcaceae spp). In germ-free mice, the author proved that fecal micriobiota transplantation from responder to non-responder patients was associated with delay in tumor growth and superior response to immunotherapy. Immune profiling of the obtained tumors indicated a microbiota-induced enhancement of the anti-tumor immunity and higher expression of PD1 ligand (PD-L1) within tumor microenvironment [14]. Antonio Sica (Milan, Italy) demonstrated a connection between immunotherapy response and nutrient availability. In a murine model of fibrosarcoma (MN/MCA1), he proved that the efficacy of anti-PD1 treatment improved when the monoclonal antibody was administered in combination with inhibitors of nicotinamide phosphoribosyltransferase (NAMPT). NAMPT, the rate-limiting enzyme in the production of nicotinamide adenine dinucleotide (NAD), and often overexpressed in cancer, controls hematopoiesis in bone marrow and regulates myeloid-derived suppressor cells (MDSCs). In response to myeloid growth factors, NAMPT levels increased inducing the inactivation of CXCR4-dependent pathway and the mobilization of MDSCs that coordinate tumor immune suppression. Inhibition of NAMPT was associated with a reduction of tumor growth and metastatization, reduction of MDSC number and concomitant increase of CD4⁺ and CD8⁺ cell number, thus reactivating the T-cell-mediated anti-tumor immunity [15]. Nutrient involvement in cancer development has also been exploited by Antonio Moschetta (Bari, Italy), who described a gut-liver axis in hepatocarcinoma (HCC) onset, regulated by nuclear receptor FXR. FXR is a sensor of hepatocyte intracellular levels of Bile Acids (BAs) highly expressed not only in the liver, but also in the gut, where it controls the levels of CYP7a1 as well as other enzymes and proteins involved in the secretion and transportation of BAs. While BAs are key players in the elimination of lipid nutrients, dietary fats and steroids, they can also induce inflammation and DNA oxidative damages, impair cell proliferation and apoptosis resulting in neoplastic transformation of hepatocytes. Moschetta demonstrated that FXR-null mice, upon cholic acid-supplemented diet, presented a deregulation of CYP7a1, associated with increased BA hydrophilicity and inflammatory cytokine production and cell hyper-proliferation leading to spontaneous development of HCC. All together, these results suggest that microbiota, metabolism and cancer must be evaluated no longer as distinct factors but as a whole to predict cancer progression and response to novel anticancer therapies. Karolina Pilipow (Milan, Italy) showed a novel method to develop early differentiated memory stem-like T cells from naïve precursors (TN). The author demonstrated that treatment with the ROS scavenger N-acetylcysteine, during activation of TN cells, was associated with metabolic reprogramming of T cells and slow-down of their differentiation. Limiting ROS metabolism induced the switch from glycolysis to fatty acid oxidation and down-regulation of mTOR pathway, together with higher levels of expression of stem-associated genes (LEF1 and TCF7), enhanced proliferation and anti-tumor capability compared to untreated T cells.

Session 5: tumor and T-cell interactions

Sjoerd van der Burg (Leiden, The Netherlands) opened Session 5 showing that the clonality of TILs from melanoma lesions was patient specific, rather than shared among patients. Van der Burg group demonstrated that once harvested, expanded against the primary tumor and infused, CD8⁺ T cells lacked the reactivity against newly developed lesions and metastasis. Moreover, T cells expanded against secondary lesions did not react against epitopes from the primary tumor. Indeed, after recovery, they found changes in relative expansion of tumor-specific CD8⁺ T-cell clones. These data suggested that also in humans, tumor clones with lowered or absent expression of neoantigens can emerge under the pressure of immunoediting after adoptive T-cell transfer. Next, Mark S. Cragg (Southampton, UK) reported that the therapeutic activity and mechanism of action of antibodies targeting immunostimulatory receptors depend on the isotype. For instance, antibodies to CD40 or OX40 were generally effective as mIgG1 by engaging FcRIIB and mediating receptor clustering on the cell surface leading to activation, akin to their physiological ligands. In contrast, mIgG2a antibodies against OX40 depleted Tregs, facilitating the expansion of CD8⁺ T cells, whereas mIgG1 did not. Last invited speaker of the session was Dario Sangiolo (Turin, Italy). He showed that melanoma cells might express PD-1 molecules, which endowed them with stem properties. Contrary to PD-1⁻, PD-1⁺ cells were more efficient in propagating tumors in immunodeficient mice in an mTOR-dependent way and this ability was impaired by the administration of anti-PD-1 antibody. Stem-like PD-1⁺ melanoma cells resulted to be drivers of relapse after BRAF/MEK inhibitor therapy. Administration of anti-PD-1 mAb prolonged the anti-tumor response of BRAF/MEK inhibitors in melanoma xenograft-bearing mice [16]. Sara Bulfamante (Turin, Italy) demonstrated that chemotherapeutics (e.g. gemcitabine) were effective not only because directly killed tumor cells, but also because had immune modulatory activity. She demonstrated that chemotherapy augmented tumor-associated antigens (TAA) recognized by specific IgG in the serum of patients affected by pancreatic ductal adenocarcinoma (PDAC), favoring complement-dependent killing of tumor cells. In murine model of PDAC, gemcitabine better controlled tumor growth after vaccination with the TAA ENO1 compared with controls, thus, opening the possibility to employ chemotherapy to unleash new TAAs and boost the immune response against them. A very nice strategy to generate tumor antigens is the one reported by Alessia Melacarne (Milan, Italy). She exploited the ability of Salmonella to induce in infected tumor cells the expression of membrane hemichannels, that allow antigen transfer, to retrieve neoantigens for cancer vaccines. Melacarne found that some of these peptides were immunogenic, because prevented tumor progression in a murine model of melanoma. Similar peptides obtained from Salmonella-infected human melanoma cells were also able to activate CD8⁺ T cells against tumor cells in vitro. This strategy was tested in vivo, where it improved the survival of pet dogs affected by spontaneous osteosarcomas. Finally, Selene Ottonello (Genova, Italy) nicely correlated the baseline frequency of CD3⁺ cells in peripheral blood with the overall survival (OS) of non-small cell lung cancer patients who underwent anti-PD-1 therapy. Specifically, a higher baseline expression of PD-1 on CD3⁺ and CD8⁺ T cells correlated with shorter OS and progressive disease. Whereas patients with partial response or stable disease had lower exhausted CD8⁺ T cells both at baseline and during the treatment. Guido Kroemer’s lecture closed Session 5. He elegantly showed that immunogenic cell stress and death were prerequisite for a successful anticancer therapy. In this view, effective chemotherapy, radiotherapy and target therapy were de facto immunotherapy. Indeed, anticancer drugs had higher activity in immunocompetent hosts and patients. Kroemer showed that anticancer therapy supplied the immune system with tumor-derived antigens and tumor-derived adjuvants (e.g. Annexin, ATP, HMGB1 etc.). For instance, R-crizotinib, a tyrosine kinase receptor inhibitor, has a potent off-target immunostimulatory activity mediated by annexin 1 and HMGB1 released by dying tumor cells. As a consequence, blockade of annexin 1 and HMGB1 or depletion of T lymphocytes resulted in the loss of therapeutic activity of R-crizotinib in mice.

Session 6: T-cell biology

One of the main aims of cancer immunotherapy is the elicitation of effective anti-tumor immune responses that should distinguish mutated malignant cells from their normal counterparts, selectively kill them and provide long-lasting protection. Antigen-specificity and immunological memory are two distinctive properties of the adaptive arm of the immune system and especially of T cells, which are the major orchestrators and effectors of both cytotoxic and humoral adaptive responses. Although most human malignancies are T cell-infiltrated, the TME, comprising cancer cells, MDSCs, M2 macrophages and stromal cells, can use several mechanisms of immune evasion to render infiltrating T cells exhausted, anergic, regulatory, or even pro-tumor. Therefore, a deeper understanding of the biology of T cells (i.e., CD4⁺, CD8⁺ and innate NKT cells) is a prerequisite to elicit, boost and/or redirect effective immune responses to cancer. Keywords of model studies and clinical trials presented in the 2018 NIBIT section dedicated to T-cell biology were T-cell stemness, T-cell exhaustion, antigen specificity, and antigen adjuvanticity. Rahul Roychoudhuri (Cambridge, UK) opened the session by exploring tumor cell-intrinsic and extrinsic mechanisms of TME-mediated T-cell suppression. Extracellular potassium ([K +]e) released by necrotic tumor cells impairs T-cell function by inhibiting Akt-mTOR pathway. Such dysfunctional signaling can be reversed in vitro through phosphatase PP2A inhibition (either by gene silencing or by PPA2 inhibitor okadaic acid), or by overexpression of the voltage-gated potassium efflux channel Kv1.3. The gene-engineering approach could be exploited in the adoptive cell therapy of highly necrotic tumors, like colon cancer and melanoma, to render reinfused TILs less prone to local suppression. On the other hand, the transcriptional repressor BACH2, whose haploinsufficiency was described by Roychoudhuri and colleagues as the cause of BRIDA syndrome, was studied for its ability to limit terminal effector T-cell differentiation and favor Treg induction by acting on super-enhancer elements upstream genes controlling T-cell fate and senescence. The potential benefit of future BACH2-directed gene therapy was exploited in a pre-clinical tumor model, where the reduction of tumor growth in Bach2-deficient mice compared to wild-type mice was mirrored by an increase of rapidly proliferating and IFNγ-expressing T cells and by a decrease of Tregs at the tumor site. Enrico Lugli (Milan, Italy) addressed the heterogeneity of tumor-infiltrating CD8 + T cells, to define the most effective and durable anti-tumor identikit. Specifically, high-dimensional flow cytometry was performed on single CD8 + T cells sorted from the blood, normal lung tissues and tumor tissues of 53 human non-small-cell lung carcinoma (NSCLC) patients. With the combined reanalysis of public single RNAseq data, Lugli identified CXCR5⁺TIM-3^–CD8⁺ T cells as a cytotoxic population with a partial exhausted phenotype, yet retaining stem-like features of self-renewal, multipotency and polyfunctionality. These cells, featuring a PD-1^intTIM-3⁻TCF-1⁺EOMES⁺T-bet⁻ phenotype, showed enhanced ability to derive more differentiated effector cells, and were lost during NSCLC progression. To increase the intratumoral representation of such population by ACT, similar stem memory T cells can be easily derived in vitro from naïve precursors, by limiting ROS metabolism upon naïve CD8⁺ T-cell activation [17]. Sebastian Amigorena (Paris, France) examined the epigenetic control of stemness commitment in CD8⁺ T cells. SUV39h1 is a chromatin writer that silences stem/memory genes in terminally-differentiating CD8⁺ T-cell effectors. Interestingly, in a mouse model of tumor engraftment, OVA-B16 tumor growth was much better controlled by anti-PD-1 treatment in SUV39h1^−/− mice than in wild-type littermates. RNAseq analyses of CD8⁺ T cells sorted wild type and SUV39h1^−/− mice treated or not with anti-PD-1 mAb, outlined a sequential exhaustion program of tumor-infiltrating CD8⁺ T cells, starting from an early IFNγ⁺ memory-like stage (exh1), progressing to an early exhausted PD-1^int stage (exh2), then to a late exhausted PD-1^high stage (exh3), and finally to a senescent apoptosis and glycolysis-marked stage (exh4). In this view, anti-PD-1 treatment was more effective in SUV39h1^−/− mice, because the IFNγ⁺ memory-like fraction (exh1), endowed with both self-renewal and cytotoxic activity, was the more represented exhausted stage among infiltrating CD8⁺ T cells. ETP-69 is a SUV39h1 inhibitor that arrests the terminal T-cell differentiation, thus mimicking the anti-tumor activity observed in SUV39h1^−/− mice treated with anti-PD-1 mAb. By affecting SUV39h1-driven silencing, this drug could be exploited to render cells more prone to reprogramming and potentiate the efficacy of immune checkpoint inhibitors [18]. Massimo Massaia (Turin, Italy) discussed the main issues in the optimization of V9V2T-cell-based immunotherapy. These non-conventional T cells are activated by phosphoantigens (pAgs) generated by antigen-presenting cells (APCs) and tumor cells and best expanded with zoledronic acid (ZA). ZA inhibits the mevalonate pathway in APCs and tumor cells, inducing the intracellular accumulation and extracellular release of isopentenyl pyrophosphate (IPP), a potent V9V2T-cell activator. Tumor-infiltrating V9V2T cells are traditionally considered favorable indicators of local tumor containment, as they can exert several direct and indirect cytotoxic effects on cancer cells. Yet, TME can inhibit their activation and switch them from anti-tumor to pro-tumor functions. This is best exemplified in the multiple myeloma bone marrow (MM BM) where exhausted PD-1⁺ V9V2T cells are already found in asymptomatic patients and to persist in MM in clinical remission even after removal of myeloma cells by effective treatment. Single PD-1 blockade only partially recovered immune effector functions of V9V2T cells, because it induced the expression of additional immune checkpoint molecules (ICPs), such as TIM-3 and LAG-3, pushing BM-infiltrating V9V2T cells into a superanergic state. The best therapy to prevent ICP upregulation was combining multiple anti-ICP antibodies. A deeper analysis of the molecular interactions between different ICPs is, therefore, needed to design alternative strategies to prevent superanergy and boost the immune potency of ICP inhibitors [19]. Sara Bobisse (Lausanne, Switzerland) presented data on the feasibility of new mutanome-based personalized treatments in patients with high-grade epithelial ovarian cancer (EOC). EOC is a tumor with a low mutational load, but susceptible to immune recognition, as demonstrated by the presence of CD8⁺ TILs in late-stage patients with a better prognosis. Therefore, immunotherapy would be promising in EOC, provided that CD8⁺ T cells specific for private tumor neoantigens (neoAgs), derived from non-synonymous somatic mutations, could be expanded or primed to exert anti-tumor functions without unwanted cytotoxicity. In the presented study, peripheral blood and tumor samples from 19 patients were screened for the presence and prevalence of CD8⁺ T cells specific for predicted mutated antigens, generated by combining exome-sequencing data and HLA binding prediction analyses. Patients were analyzed before and after personalized immunotherapy consisting of vaccination with DCs loaded with autologous tumor cells and ACT with in vitro neoAg-expanded post-vaccine peripheral cells. Using an optimized TIL expansion protocol, neoAg-specific T cells were detected and expanded at an unsuspected high rate, even in the low mutational rate tumor EOC. Specific CD8⁺ T cells were detected in most immunotherapy-naïve patients, with discordant neoAg reactivities between circulating and tumor-infiltrating cells, the TCRs of the latter having higher predicted affinity and higher functional avidity. In summary, personalized immunotherapy can increase the frequency and the functional avidity of neoAg-specific T-cell responses (both pre-existing and de novo-generated). Cristina Corbetta (Milan, Italy) discussed how to refine immunotherapy for recurrent glioblastoma. After the first encouraging results obtained with DC-based vaccinations of primary glioblastoma patients, DC immunotherapy alone revealed scarcely effective and rarely curative in recurrent patients, hinting at the need to define an optimal combinatorial approach. The first attempt (the DENDR2 clinical study) was the combination with of the alkylating agent Temozolomide (TMZ), as a potential adjuvant, but this strategy failed to provide immune response activation and survival advantage. The next attempt (the V-DENDR2 clinical study) was to precondition the vaccine site with the recall antigen tetanus toxoid (TT): this positively associated with an increase in the number and activation of CD8⁺ and CD4⁺ T cells in long-term survivors. Ag-experienced CD8⁺ T cells from long-term survivors had a memory phenotype (i.e., IFN and KLRG1 expression). In conclusion, in V-DENDR2 clinical trial, the adjuvant change from TMZ to TT allowed: (i) CD8⁺ T-cell activation and memory formation, (ii) increased numbers of CD4⁺ T helper cells sustaining the anti-tumor CD8⁺ T-cell response and allowing their persistence during the treatment, and (iii) a better impact on patients’ survival. As the last talk of the section, Andrea Ponzetta (Milan, Italy) described a genetic approach to explore the conditioning of adaptive immunity by neutrophils in 3-methylcholanthrene (3-MCA)-induced sarcomagenesis. Neutrophils were shown to drive an IL-12/IFN-dependent type-1 polarization of innate-like CD4- and CD8-double negative (DN) T cells in the TME, providing resistance against primary carcinogenesis and controlling tumor progression in tumor-bearing mice. Multiparametric flow cytometry and single-cell RNAseq analyses revealed the heterogeneity of infiltrating DN T cells. The relevance of infiltrating neutrophils for sarcoma resistance in the mouse model was confirmed in a specific subtype of human sarcomas, namely undifferentiated pleomorphic sarcomas (UPS), both by interrogating the public Cancer Genome Atlas (TCGA) database and by performing immunohistochemistry analyses in a cohort of UPS patients. Of note, higher numbers of tumor-infiltrating innate cells expressing a neutrophil-specific gene signature correlated with local IFN-mediated responses in patients with a better clinical outcome (i.e., higher overall survival, recurrence-free survival and metastasis-free survival rates).

Session 7: clinical and research highlights around NIBIT

Growing evidence supports a considerable role for Tyrosine kinase inhibitors (TKi) in promoting anti-tumor immune responses. The co-administration of CD40 antagonists with BRAFi has been reported to interfere with the anti-tumor effect of the latter (PLX4720) in a melanoma mouse model. These findings suggest that the anti-tumor effect of BRAFi relies on dendritic cell maturation, which is induced by the biding of CD40 with its ligand. In this framework, Andrea Botticelli (Rome, Italy) showed that, in a randomized phase III study, metastatic renal cell carcinoma (mRCC) patients pretreated with the TKi pazopanib before nivolumab experienced a higher clinical benefit as compared to patients pretreated with Sunitinib. Notably, it was demonstrated that the ability of pazopanib to enhance nivolumab efficacy depended on pazopanib-mediated modification of DC functions. In fact, treatment with pazopanib compared to sunitinib, resulted in a higher expression of CD40 and primed DC activity through the down-regulation of the β-catenin pathway [11]. In the setting of cancer vaccination, Luigi Buonaguro (Naples, Italy) presented HepaVac: a European multi-center phase I/II clinical trial based on the administration of a vaccine multi-epitope, multi-target and multi-HLA allele that is currently ongoing for the treatment of patients affected by hepatocellular carcinoma (HCC). The HepaVac is based on an “off-the-shelf” vaccine, which comprises 18 newly identified MHC-I and -II tumor-associated peptides (TUMAPs) naturally processed and presented on primary tumor tissues of HCC patients. This treatment is complemented by an actively personalized vaccine (APVAC) approach, which can be used in a subgroup of patients for whom specific mutated, naturally processed and presented peptides have been identified. Giuseppe Procopio (Milan, Italy) further discussed recent advances of immunotherapy in the treatment of RCC. The randomized clinical trial CHECKMATE-025 showed a higher benefit of nivolumab (3 mg/kg intravenously every 2 weeks) compared to everolimus (10 mg per dose daily) in terms of OS (the median OS was 5.4 months longer with nivolumab than with everolimus) also in patients with bone and liver metastasis, regardless of the expression of PD-L1 [20]. Based on this evidence, 4 doses of nivolumab (3 mg/kg) has been administrated intravenously every 3 weeks in combination with ipilimumab (1 mg/kg) and tested versus sunitinib monotherapy (50 mg capsules by mouth once daily for 4 weeks), proving an increased efficacy of the combination in intermediate- and poor-risk patients with advanced or metastatic renal cell carcinoma (CHECKMATE-214). In terms of progression-free survival (PFS), patients with high expression of PD-L1 (PD-L1 ≥ 1%; n = 214) showed a significant improvement if treated with nivolumab and ipilimumab compared to sunitinib, though PFS was not different in patients with low PD-L1 (PD-L1 < 1%; n = 562) expression. Again, high PD-L1 expression was associated with increased PFS also in patients treated with the atezolizumab (1200 mg) and bevacizumab (15 mg/kg) combination (intravenously once every 3 weeks), but not with sunitinib alone (50 mg orally once daily for 4 weeks on, 2 weeks off), in a first-line randomized phase III study conducted in patients with advanced or metastatic RCC. These data suggest the need to stratify patients according to the PD-L1 status to select patients who could respond to first-line nivolumab + ipilimumab or atezolizumab + bevacizumab combinations (PD-L1 high), or who can be treated with TKI (Sunitinib) alone (PD-L1 low). Francesco Spagnolo (Genova, Italy) discussed the activity of target therapies in patients with metastatic melanoma, including those affected by brain metastases. The combination of target therapy with immunotherapeutic agents increased early benefit in terms of clinical activity and several trials are now ongoing to validate this effect on PFS and OS. For instance, in the IMspire170 randomized clinical trial, wild-type BRAF patients are treated with either cobimetinib (60 mg, orally daily on a 21 days on, 7 days off schedule) and atezolizumab (840 mg as IV infusion once every 2 weeks) or pembrolizumab (200 mg as IV infusion once every 3 weeks), while in the ongoing clinical trial TRICOTEL, cobimetinib and atezolizumab are combined with vemurafenib (960 mg, orally twice daily) only in patients bearing mutated BRAF and with central nervous system involvement. Two trials are now ongoing to investigate the sequential regimen of target therapy and immunotherapy (ECOG6134 and SECOMBIT trials). The session was closed by Anna Maria Di Giacomo (Siena, Italy), who reviewed the most recent clinical activities sponsored by the NIBIT Foundation in melanoma and mesothelioma patients. She presented a randomized phase III study (NIBIT-M2 study), which is supported by data showing that the combined administration of ipilimumab and nivolumab induced objective responses in 53% of melanoma patients with or without brain metastasis, with an 82% 1-year OS and an acceptable safety profile. NIBIT-M2 study will investigate the efficacy of the combination of ipilimumab and fotemustine or ipilimumab, and nivolumab versus fotemustine alone in patients with melanoma metastatic to the brain. In mesothelioma patients, a single arm, phase II clinical study provided preliminary evidence about safety and tolerability of anti-CTLA-4 tremelimumab combined with the anti-PD-L1 monoclonal antibody durvalumab in patients with unresectable malignant mesothelioma (NIBIT-MESO-1 study), independently of tumor cell PD-L1 expression [21]. Finally, the possibility of combining immunotherapy with epigenetic drugs, which are expected to increase tumor immunogenicity and immune recognition, was presented (NIBIT-M4 study).

Conclusions

The NIBIT meeting, in the awareness that the key to further success is fostering and strengthening scientific collaborations and exchanging of ideas, gave a great opportunity for young scientists and leaders in immuno-oncology to meet and discuss about emerging clinical and pre-clinical data arising in the field. The meeting provided the attendances with a broad overview of the most recent achievements in cancer bio-immunotherapy, highlighting some new fields of investigation, such as metabolism and the identification of favourable microbiota, as new possible areas for the clinical development of drugs able to synergize and potentiate the current strategies of immunotherapy. Excited by the recent goals achieved, and eager to learn new discoveries in the field of immunotherapy, together with the NIBIT Board of Directors, we look forward to seeing you in Verona on October 11–13, 2019 to enjoy the “XVII NIBIT Meeting’’.

Abbreviations

ACT

Adoptive cell therapy

APM

Antigen-presenting machinery

BAs

Bile acids

B-ALL

B acute lymphoid leukemia

BC

Breast cancer

BM

Bone marrow

CAR

Chimeric antigen receptors

CIK

Cytokine-induced killer

CRC

Colorectal cancer

CRS

Cytokine-release syndrome

CNS

Central nervous system

CSPG4

Chondroitin sulfate proteoglycan 4

CSC

Cancer stem cells

DC

Dendritic cells

EOC

Epithelial ovarian cancer

EV

Extracellular vesicles

FLIP

FLICE (FADD-like IL-1β-converting enzyme) inhibitory protein

FXR

Farnesoid X receptor

GBM

Glioblastoma multiforme

GITR

Glucocorticoid-induced tumor necrosis factor receptor

GvHD

Graft versus host disease

HCC

Hepatocellular carcinoma

HMGB1

High-mobility group box 1

ICI

Immune checkpoint inhibitors

MDSC

Myeloid-derived suppressor cells

MCA

Methylcholanthrene

MM

Multiple myeloma

mTOR

Mammalian target of rapamycin

NAMPT

Nicotinamide phosphoribosyltransferase

NSCLC

Non-small cell lung cancer

ORR

Objective response rate

OS

Overall survival

pCR

Pathologic complete response

PDAC

Pancreatic ductal adenocarcinoma

PFS

Progression-free survival

PMN-MDSC

Polymorphonuclear myeloid-derived suppressor cells

RCC

Renal cell carcinoma

ROS

Reactive oxygen species

TAAs

Tumor-associated antigens

TAM

Tumor-associated macrophages

TCF-1

Transgenic T-cell factor 1

TILs

Tumor-infiltrating lymphocytes

TME

Tumor microenvironment

Treg

Regulatory T cells

ZA

Zoledronic acid

Author contributions

All authors contributed to writing the manuscript. MB and VR collected and assembled contributions from all authors. All authors revised and approved the final version of the manuscript.

Funding

This meeting was supported in part by unrestricted grant from Roche, Bristol-Meyers Squibb, Novartis, Pierre Fabre, Amgen, MSD Oncology, Biotechne, Incyte, Euroclone and Labospace. It was organized under the auspices of the Italian Association for Cancer Research (AIRC), Ordine Provinciale dei Medici Chirurghi e degli Odontoiatri (OMCeO, Milano), Regione Lombardia, Società Italiana di Cancerologia, Ospedale San Raffaele ed Università Vita-Salute San Raffaele.

Compliance with ethical standards

Conflict of interest

No potential conflicts of interest were disclosed by the authors.