Targeted therapy for multiple myeloma: an overview on CD138-based strategies.

Multiple myeloma (MM) is an incurable hematological disease characterized by the uncontrolled growth of plasma cells primarily in the bone marrow. Although its treatment consists of the administration of combined therapy regimens mainly based on immunomodulators and proteosome inhibitors, MM remains incurable, and most patients suffer from relapsed/refractory disease with poor prognosis and survival. The robust results achieved by immunotherapy targeting MM-associated antigens CD38 and CD319 (also known as SLAMF7) have drawn attention to the development of new immune-based strategies and different innovative compounds in the treatment of MM, including new monoclonal antibodies, antibody-drug conjugates, recombinant proteins, synthetic peptides, and adaptive cellular therapies. In this context, Syndecan1 (CD138 or SDC1), a transmembrane heparan sulfate proteoglycan that is upregulated in malignant plasma cells, has gained increasing attention in the panorama of MM target antigens, since its key role in MM tumorigenesis, progression and aggressiveness has been largely reported. Here, our aim is to provide an overview of the most important aspects of MM disease and to investigate the molecular functions of CD138 in physiologic and malignant cell states. In addition, we will shed light on the CD138-based therapeutic approaches currently being tested in preclinical and/or clinical phases in MM and discuss their properties, mechanisms of action and clinical applications.

A novel complement-fixing IgM antibody targeting GPC1 as a useful immunotherapeutic strategy for the treatment of pancreatic ductal adenocarcinoma.

BACKGROUND: Pancreatic ductal adenocarcinoma (PDAC) is one of the most aggressive cancers with a very low survival rate at 5 years. The use of chemotherapeutic agents results in only modest prolongation of survival and is generally associated with the occurrence of toxicity effects. Antibody-based immunotherapy has been proposed for the treatment of PDAC, but its efficacy has so far proved limited. The proteoglycan glypican-1 (GPC1) may be a useful immunotherapeutic target because it is highly expressed on the surface of PDAC cells, whereas it is not expressed or is expressed at very low levels in benign neoplastic lesions, chronic pancreatitis, and normal adult tissues. Here, we developed and characterized a specific mouse IgM antibody (AT101) targeting GPC1. METHODS: We developed a mouse monoclonal antibody of the IgM class directed against an epitope of GPC1 in close proximity to the cell membrane. For this purpose, a 46 amino acid long peptide of the C-terminal region was used to immunize mice by an in-vivo electroporation protocol followed by serum titer and hybridoma formation. RESULTS: The ability of AT101 to bind the GPC1 protein was demonstrated by ELISA, and by flow cytometry and immunofluorescence analysis in the GPC1-expressing "PDAC-like" BXPC3 cell line. In-vivo experiments in the BXPC3 xenograft model showed that AT101 was able to bind GPC1 on the cell surface and accumulate in the BXPC3 tumor masses. Ex-vivo analyses of BXPC3 tumor masses showed that AT101 was able to recruit immunological effectors (complement system components, NK cells, macrophages) to the tumor site and damage PDAC tumor tissue. In-vivo treatment with AT101 reduced tumor growth and prolonged survival of mice with BXPC3 tumor (p < 0.0001). CONCLUSIONS: These results indicate that AT101, an IgM specific for an epitope of GPC1 close to PDAC cell surface, is a promising immunotherapeutic agent for GPC1-expressing PDAC, being able to selectively activate the complement system and recruit effector cells in the tumor microenvironment, thus allowing to reduce tumor mass growth and improve survival in treated mice.

A comprehensive overview on antibody-drug conjugates: from the conceptualization to cancer therapy.

Antibody-Drug Conjugates (ADCs) represent an innovative class of potent anti-cancer compounds that are widely used in the treatment of hematologic malignancies and solid tumors. Unlike conventional chemotherapeutic drug-based therapies, that are mainly associated with modest specificity and therapeutic benefit, the three key components that form an ADC (a monoclonal antibody bound to a cytotoxic drug via a chemical linker moiety) achieve remarkable improvement in terms of targeted killing of cancer cells and, while sparing healthy tissues, a reduction in systemic side effects caused by off-tumor toxicity. Based on their beneficial mechanism of action, 15 ADCs have been approved to date by the market approval by the Food and Drug Administration (FDA), the European Medicines Agency (EMA) and/or other international governmental agencies for use in clinical oncology, and hundreds are undergoing evaluation in the preclinical and clinical phases. Here, our aim is to provide a comprehensive overview of the key features revolving around ADC therapeutic strategy including their structural and targeting properties, mechanism of action, the role of the tumor microenvironment and review the approved ADCs in clinical oncology, providing discussion regarding their toxicity profile, clinical manifestations and use in novel combination therapies. Finally, we briefly review ADCs in other pathological contexts and provide key information regarding ADC manufacturing and analytical characterization.

Surface antibody changes protein corona both in human and mouse serum but not final opsonization and elimination of targeted polymeric nanoparticles.

BACKGROUND: Nanoparticles represent one of the most important innovations in the medical field. Among nanocarriers, polymeric nanoparticles (PNPs) attracted much attention due to their biodegradability, biocompatibility, and capacity to increase efficacy and safety of encapsulated drugs. Another important improvement in the use of nanoparticles as delivery systems is the conjugation of a targeting agent that enables the nanoparticles to accumulate in a specific tissue. Despite these advantages, the clinical translation of therapeutic approaches based on nanoparticles is prevented by their interactions with blood proteins. In fact, the so-formed protein corona (PC) drastically alters the biological identity of the particles. Adsorbed activated proteins of the complement cascade play a pivotal role in the clearance of nanoparticles, making them more easily recognized by macrophages, leading to their rapid elimination from the bloodstream and limiting their efficacy. Since the mouse is the most used preclinical model for human disease, this work compared human and mouse PC formed on untargeted PNPs (uPNPs) and targeted PNPs (tPNPs), paying particular attention to complement activation. RESULTS: Mouse and human serum proteins adsorbed differently to PNPs. The differences in the binding of mouse complement proteins are minimal, whereas human complement components strongly distinguish the two particles. This is probably due to the human origin of the Fc portion of the antibody used as targeting agent on tPNPs. tPNPs and uPNPs mainly activate complement via the classical and alternative pathways, respectively, but this pattern did not affect their binding and internalization in macrophages and only a limited consumption of the activity of the human complement system was documented. CONCLUSIONS: The results clearly indicate the presence of complement proteins on PNPs surface but partially derived from an unspecific deposition rather than an effective complement activation. The presence of a targeting antibody favors the activation of the classical pathway, but its absence allows an increased activation of the alternative pathway. This results in similar opsonization of both PNPs and similar phagocytosis by macrophages, without an impairment of the activity of circulating complement system and, consequently, not enhancing the susceptibility to infection.

Machine learning to improve interpretability of clinical, radiological and panel-based genomic data of glioma grade 4 patients undergoing surgical resection.

BACKGROUND: Glioma grade 4 (GG4) tumors, including astrocytoma IDH-mutant grade 4 and the astrocytoma IDH wt are the most common and aggressive primary tumors of the central nervous system. Surgery followed by Stupp protocol still remains the first-line treatment in GG4 tumors. Although Stupp combination can prolong survival, prognosis of treated adult patients with GG4 still remains unfavorable. The introduction of innovative multi-parametric prognostic models may allow refinement of prognosis of these patients. Here, Machine Learning (ML) was applied to investigate the contribution in predicting overall survival (OS) of different available data (e.g. clinical data, radiological data, or panel-based sequencing data such as presence of somatic mutations and amplification) in a mono-institutional GG4 cohort. METHODS: By next-generation sequencing, using a panel of 523 genes, we performed analysis of copy number variations and of types and distribution of nonsynonymous mutations in 102 cases including 39 carmustine wafer (CW) treated cases. We also calculated tumor mutational burden (TMB). ML was applied using eXtreme Gradient Boosting for survival (XGBoost-Surv) to integrate clinical and radiological information with genomic data. RESULTS: By ML modeling (concordance (c)- index = 0.682 for the best model), the role of predicting OS of radiological parameters including extent of resection, preoperative volume and residual volume was confirmed. An association between CW application and longer OS was also showed. Regarding gene mutations, a role in predicting OS was defined for mutations of BRAF and of other genes involved in the PI3K-AKT-mTOR signaling pathway. Moreover, an association between high TMB and shorter OS was suggested. Consistently, when a cutoff of 1.7 mutations/megabase was applied, cases with higher TMB showed significantly shorter OS than cases with lower TMB. CONCLUSIONS: The contribution of tumor volumetric data, somatic gene mutations and TBM in predicting OS of GG4 patients was defined by ML modeling.

Targeted chitosan nanobubbles as a strategy to down-regulate microRNA-17 into B-cell lymphoma models.

Introduction: MicroRNAs represent interesting targets for new therapies because their altered expression influences tumor development and progression. miR-17 is a prototype of onco-miRNA, known to be overexpressed in B-cell non-Hodgkin lymphoma (B-NHL) with peculiar clinic-biological features. AntagomiR molecules have been largely studied to repress the regulatory functions of up-regulated onco-miRNAs, but their clinical use is mainly limited by their rapid degradation, kidney elimination and poor cellular uptake when injected as naked oligonucleotides. Methods: To overcome these problems, we exploited CD20 targeted chitosan nanobubbles (NBs) for a preferential and safe delivery of antagomiR17 to B-NHL cells. Results: Positively charged 400 nm-sized nanobubbles (NBs) represent a stable and effective nanoplatform for antagomiR encapsulation and specific release into B-NHL cells. NBs rapidly accumulated in tumor microenvironment, but only those conjugated with a targeting system (antiCD20 antibodies) were internalized into B-NHL cells, releasing antagomiR17 in the cytoplasm, both in vitro and in vivo. The result is the down-regulation of miR-17 level and the reduction in tumor burden in a human-mouse B-NHL model, without any documented side effects. Discussion: Anti-CD20 targeted NBs investigated in this study showed physico-chemical and stability properties suitable for antagomiR17 delivery in vivo and represent a useful nanoplatform to address B-cell malignancies or other cancers through the modification of their surface with specific targeting antibodies.

Machine Learning Application Identifies Germline Markers of Hypertension in Patients With Ovarian Cancer Treated With Carboplatin, Taxane, and Bevacizumab.

Pharmacogenomics studies how genes influence a person's response to treatment. When complex phenotypes are influenced by multiple genetic variations with little effect, a single piece of genetic information is often insufficient to explain this variability. The application of machine learning (ML) in pharmacogenomics holds great potential - namely, it can be used to unravel complicated genetic relationships that could explain response to therapy. In this study, ML techniques were used to investigate the relationship between genetic variations affecting more than 60 candidate genes and carboplatin-induced, taxane-induced, and bevacizumab-induced toxicities in 171 patients with ovarian cancer enrolled in the MITO-16A/MaNGO-OV2A trial. Single-nucleotide variation (SNV, formerly SNP) profiles were examined using ML to find and prioritize those associated with drug-induced toxicities, specifically hypertension, hematological toxicity, nonhematological toxicity, and proteinuria. The Boruta algorithm was used in cross-validation to determine the significance of SNVs in predicting toxicities. Important SNVs were then used to train eXtreme gradient boosting models. During cross-validation, the models achieved reliable performance with a Matthews correlation coefficient ranging from 0.375 to 0.410. A total of 43 SNVs critical for predicting toxicity were identified. For each toxicity, key SNVs were used to create a polygenic toxicity risk score that effectively divided individuals into high-risk and low-risk categories. In particular, compared with low-risk individuals, high-risk patients were 28-fold more likely to develop hypertension. The proposed method provided insightful data to improve precision medicine for patients with ovarian cancer, which may be useful for reducing toxicities and improving toxicity management.

Machine learning in onco-pharmacogenomics: a path to precision medicine with many challenges.

Over the past two decades, Next-Generation Sequencing (NGS) has revolutionized the approach to cancer research. Applications of NGS include the identification of tumor specific alterations that can influence tumor pathobiology and also impact diagnosis, prognosis and therapeutic options. Pharmacogenomics (PGx) studies the role of inheritance of individual genetic patterns in drug response and has taken advantage of NGS technology as it provides access to high-throughput data that can, however, be difficult to manage. Machine learning (ML) has recently been used in the life sciences to discover hidden patterns from complex NGS data and to solve various PGx problems. In this review, we provide a comprehensive overview of the NGS approaches that can be employed and the different PGx studies implicating the use of NGS data. We also provide an excursus of the ML algorithms that can exert a role as fundamental strategies in the PGx field to improve personalized medicine in cancer.

Biological Features of Nanoparticles: Protein Corona Formation and Interaction with the Immune System.

Nanoparticles (NPs) are versatile candidates for nanomedical applications due to their unique physicochemical properties. However, their clinical applicability is hindered by their undesirable recognition by the immune system and the consequent immunotoxicity, as well as their rapid clearance in vivo. After injection, NPs are usually covered with layers of proteins, called protein coronas (PCs), which alter their identity, biodistribution, half-life, and efficacy. Therefore, the characterization of the PC is for in predicting the fate of NPs in vivo. The aim of this review was to summarize the state of the art regarding the intrinsic factors closely related to the NP structure, and extrinsic factors that govern PC formation in vitro. In addition, well-known opsonins, including complement, immunoglobulins, fibrinogen, and dysopsonins, such as histidine-rich glycoprotein, apolipoproteins, and albumin, are described in relation to their role in NP detection by immune cells. Particular emphasis is placed on their role in mediating the interaction of NPs with innate and adaptive immune cells. Finally, strategies to reduce PC formation are discussed in detail.

Novel Nanotechnology Approaches to Overcome Drug Resistance in the Treatment of Hepatocellular Carcinoma: Glypican 3 as a Useful Target for Innovative Therapies.

Hepatocellular carcinoma (HCC) is the second most lethal tumor, with a 5-year survival rate of 18%. Early stage HCC is potentially treatable by therapies with curative intent, whereas chemoembolization/radioembolization and systemic therapies are the only therapeutic options for intermediate or advanced HCC. Drug resistance is a critical obstacle in the treatment of HCC that could be overcome by the use of targeted nanoparticle-based therapies directed towards specific tumor-associated antigens (TAAs) to improve drug delivery. Glypican 3 (GPC3) is a member of the glypican family, heparan sulfate proteoglycans bound to the cell surface via a glycosylphosphatidylinositol anchor. The high levels of GPC3 detected in HCC and the absence or very low levels in normal and non-malignant liver make GPC3 a promising TAA candidate for targeted nanoparticle-based therapies. The use of nanoparticles conjugated with anti-GPC3 agents may improve drug delivery, leading to a reduction in severe side effects caused by chemotherapy and increased drug release at the tumor site. In this review, we describe the main clinical features of HCC and the common treatment approaches. We propose the proteoglycan GPC3 as a useful TAA for targeted therapies. Finally, we describe nanotechnology approaches for anti-GPC3 drug delivery systems based on NPs for HCC treatment.

Zebrafish: A Useful Animal Model for the Characterization of Drug-Loaded Polymeric NPs.

The use of zebrafish (ZF) embryos as an in vivo model is increasingly attractive thanks to different features that include easy handling, transparency, and the absence of adaptive immunity until 4-6 weeks. These factors allow the development of xenografts that can be easily analyzed through fluorescence techniques. In this work, ZF were exploited to characterize the efficiency of drug-loaded polymeric NPs as a therapeutical approach for B-cell malignancies. Fluorescent probes, fluorescent transgenic lines of ZF, or their combination allowed to deeply examine biodistribution, elimination, and therapeutic efficacy. In particular, the fluorescent signal of nanoparticles (NPs) was exploited to investigate the in vivo distribution, while the colocalization between the fluorescence in macrophages and NPs allows following the elimination pathway of these polymeric NPs. Xenotransplanted human B-cells (Nalm-6) developed a reproducible model useful for demonstrating drug delivery by polymeric NPs loaded with doxorubicin and, as a consequence, the arrest of tumor growth and the reduction in tumor burden. ZF proved to be a versatile model, able to rapidly provide answers in the development of animal models and in the characterization of the activity and the efficacy of drug delivery systems.

Nanocarriers as a Delivery Platform for Anticancer Treatment: Biological Limits and Perspectives in B-Cell Malignancies.

Nanoparticle-based therapies have been proposed in oncology research using various delivery methods to increase selectivity toward tumor tissues. Enhanced drug delivery through nanoparticle-based therapies could improve anti-tumor efficacy and also prevent drug resistance. However, there are still problems to overcome, such as the main biological interactions of nanocarriers. Among the various nanostructures for drug delivery, drug delivery based on polymeric nanoparticles has numerous advantages for controlling the release of biological factors, such as the ability to add a selective targeting mechanism, controlled release, protection of administered drugs, and prolonging the circulation time in the body. In addition, the functionalization of nanoparticles helps to achieve the best possible outcome. One of the most promising applications for nanoparticle-based drug delivery is in the field of onco-hematology, where there are many already approved targeted therapies, such as immunotherapies with monoclonal antibodies targeting specific tumor-associated antigens; however, several patients have experienced relapsed or refractory disease. This review describes the major nanocarriers proposed as new treatments for hematologic cancer, describing the main biological interactions of these nanocarriers and the related limitations of their use as drug delivery strategies.

The Proteoglycan Glypican-1 as a Possible Candidate for Innovative Targeted Therapeutic Strategies for Pancreatic Ductal Adenocarcinoma.

Pancreatic ductal adenocarcinoma (PDAC) accounts for 90% of all pancreatic cancers, with a 5-year survival rate of 7% and 80% of patients diagnosed with advanced or metastatic malignancies. Despite recent advances in diagnostic testing, surgical techniques, and systemic therapies, there remain limited options for the effective treatment of PDAC. There is an urgent need to develop targeted therapies that are able to differentiate between cancerous and non-cancerous cells to reduce side effects and better inhibit tumor growth. Antibody-targeted strategies are a potentially effective option for introducing innovative therapies. Antibody-based immunotherapies and antibody-conjugated nanoparticle-based targeted therapies with antibodies targeting specific tumor-associated antigens (TAA) can be proposed. In this context, glypican-1 (GPC1), which is highly expressed in PDAC and not expressed or expressed at very low levels in non-malignant lesions and healthy pancreatic tissues, is a useful TAA that can be achieved by a specific antibody-based immunotherapy and antibody-conjugated nanoparticle-based targeted therapy. In this review, we describe the main clinical features of PDAC. We propose the proteoglycan GPC1 as a useful TAA for PDAC-targeted therapies. We also provide a digression on the main developed approaches of antibody-based immunotherapy and antibody-conjugated nanoparticle-based targeted therapy, which can be used to target GPC1.

Machine Learning Application in a Phase I Clinical Trial Allows for the Identification of Clinical-Biomolecular Markers Significantly Associated With Toxicity.

Machine learning (ML) algorithms have been used to forecast clinical outcomes or drug adverse effects by analyzing different data sets such as electronic health records, diagnostic data, and molecular data. However, ML implementation in phase I clinical trial is still an unexplored strategy that implies challenges such as the selection of the best development strategy when dealing with limited sample size. In the attempt to better define prechemotherapy baseline clinical and biomolecular predictors of drug toxicity, we trained and compared five ML algorithms starting from clinical, blood biochemistry, and genotype data derived from a previous phase Ib study aimed to define the maximum tolerated dose of irinotecan (FOLFIRI (folinic acid, fluorouracil, and irinotecan) plus bevacizumab regimen) in patients with metastatic colorectal cancer. During cross-validation the Random Forest algorithm achieved the best performance with a mean Matthews correlation coefficient of 0.549 and a mean accuracy of 80.4%; the best predictors of dose-limiting toxicity at baseline were hemoglobin, serum glutamic oxaloacetic transaminase (SGOT), and albumin. The feasibility of a prediction model prototype was in principle assessed using the two distinct dose escalation cohorts, where in the validation cohort the model scored a Matthews correlation coefficient of 0.59 and an accuracy of 82.0%. Moreover, we found a strong relationship between SGOT and irinotecan pharmacokinetics, suggesting its role as surrogates' estimators of the irinotecan metabolism equilibrium. In conclusion, the potential application of ML techniques to phase I study could provide clinicians with early prediction tools useful both to ameliorate the management of clinical trials and to make more adequate treatment decisions.

Chitosan-Based Biocompatible Copolymers for Thermoresponsive Drug Delivery Systems: On the Development of a Standardization System.

Chitosan is a natural polysaccharide that is considered to be biocompatible, biodegradable and non-toxic. The polymer has been used in drug delivery applications for its positive charge, which allows for adhesion with and recognition of biological tissues via non-covalent interactions. In recent times, chitosan has been used for the preparation of graft copolymers with thermoresponsive polymers such as poly-N-vinylcaprolactam (PNVCL) and poly-N-isopropylamide (PNIPAM), allowing the combination of the biodegradability of the natural polymer with the ability to respond to changes in temperature. Due to the growing interest in the utilization of thermoresponsive polymers in the biological context, it is necessary to increase the knowledge of the key principles of thermoresponsivity in order to obtain comparable results between different studies or applications. In the present review, we provide an overview of the basic principles of thermoresponsivity, as well as a description of the main polysaccharides and thermoresponsive materials, with a special focus on chitosan and poly-N-Vinyl caprolactam (PNVCL) and their biomedical applications.

Thermoresponsive Chitosan-Grafted-Poly(N-vinylcaprolactam) Microgels via Ionotropic Gelation for Oncological Applications.

Microgels can be considered soft, porous and deformable particles with an internal gel structure swollen by a solvent and an average size between 100 and 1000 nm. Due to their biocompatibility, colloidal stability, their unique dynamicity and the permeability of their architecture, they are emerging as important candidates for drug delivery systems, sensing and biocatalysis. In clinical applications, the research on responsive microgels is aimed at the development of "smart" delivery systems that undergo a critical change in conformation and size in reaction to a change in environmental conditions (temperature, magnetic fields, pH, concentration gradient). Recent achievements in biodegradable polymer fabrication have resulted in new appealing strategies, including the combination of synthetic and natural-origin polymers with inorganic nanoparticles, as well as the possibility of controlling drug release remotely. In this review, we provide a literature review on the use of dual and multi-responsive chitosan-grafted-poly-(N-vinylcaprolactam) (CP) microgels in drug delivery and oncological applications.

Role of Virus-Related Chronic Inflammation and Mechanisms of Cancer Immune-Suppression in Pathogenesis and Progression of Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) can be classified as a prototypical inflammation-driven cancer that generally arises from a background of liver cirrhosis, but that in the presence of nonalcoholic steatohepatitis (NASH), could develop in the absence of fibrosis or cirrhosis. Tumor-promoting inflammation characterizes HCC pathogenesis, with an epidemiology of the chronic liver disease frequently encompassing hepatitis virus B (HBV) or C (HCV). HCC tumor onset and progression is a serial and heterogeneous process in which intrinsic factors, such as genetic mutations and chromosomal instability, are closely associated with an immunosuppressive tumor microenvironment (TME), which may have features associated with the etiopathogenesis and expression of the viral antigens, which favor the evasion of tumor neoantigens to immune surveillance. With the introduction of direct-acting antiviral (DAA) therapies for HCV infection, sustained virological response (SVR) has become very high, although occurrence of HCC and reactivation of HBV in patients with co-infection, who achieved SVR in short term, have been observed in a significant proportion of treated cases. In this review, we discuss the main molecular and TME features that are responsible for HCC pathogenesis and progression. Peculiar functional aspects that could be related to the presence and treatment of HCV/HBV viral infections are also dealt with.

Characterization of Thermoresponsive Poly-N-Vinylcaprolactam Polymers for Biological Applications.

Poly-N-Vinylcaprolactam (PNVCL) is a thermoresponsive polymer that exhibits lower critical solution temperature (LCST) between 25 and 50 °C. Due to its alleged biocompatibility, this polymer is becoming popular for biomedical and environmental applications. PNVCL with carboxyl terminations has been widely used for the preparation of thermoresponsive copolymers, micro- and nanogels for drug delivery and oncological therapies. However, the fabrication of such specific targeting devices needs standardized and reproducible preparation methods. This requires a deep understanding of how the miscibility behavior of the polymer is affected by its structural properties and the solution environment. In this work, PNVCL-COOH polymers were prepared via free radical polymerization (FRP) in order to exhibit LCST between 33 and 42 °C. The structural properties were investigated with NMR, FT-IR and conductimetric titration and the LCST was calculated via UV-VIS and DLS. The LCST is influenced by the molecular mass, as shown by both DLS and viscosimetric values. Finally, the behavior of the polymer was described as function of its concentration and in presence of different biologically relevant environments, such as aqueous buffers, NaCl solutions and human plasma.

A New Epigenetic Model to Stratify Glioma Patients According to Their Immunosuppressive State.

Gliomas are the most common primary neoplasm of the central nervous system. A promising frontier in the definition of glioma prognosis and treatment is represented by epigenetics. Furthermore, in this study, we developed a machine learning classification model based on epigenetic data (CpG probes) to separate patients according to their state of immunosuppression. We considered 573 cases of low-grade glioma (LGG) and glioblastoma (GBM) from The Cancer Genome Atlas (TCGA). First, from gene expression data, we derived a novel binary indicator to flag patients with a favorable immune state. Then, based on previous studies, we selected the genes related to the immune state of tumor microenvironment. After, we improved the selection with a data-driven procedure, based on Boruta. Finally, we tuned, trained, and evaluated both random forest and neural network classifiers on the resulting dataset. We found that a multi-layer perceptron network fed by the 338 probes selected by applying both expert choice and Boruta results in the best performance, achieving an out-of-sample accuracy of 82.8%, a Matthews correlation coefficient of 0.657, and an area under the ROC curve of 0.9. Based on the proposed model, we provided a method to stratify glioma patients according to their epigenomic state.

A Novel Epigenetic Machine Learning Model to Define Risk of Progression for Hepatocellular Carcinoma Patients.

Although extensive advancements have been made in treatment against hepatocellular carcinoma (HCC), the prognosis of HCC patients remains unsatisfied. It is now clearly established that extensive epigenetic changes act as a driver in human tumors. This study exploits HCC epigenetic deregulation to define a novel prognostic model for monitoring the progression of HCC. We analyzed the genome-wide DNA methylation profile of 374 primary tumor specimens using the Illumina 450 K array data from The Cancer Genome Atlas. We initially used a novel combination of Machine Learning algorithms (Recursive Features Selection, Boruta) to capture early tumor progression features. The subsets of probes obtained were used to train and validate Random Forest models to predict a Progression Free Survival greater or less than 6 months. The model based on 34 epigenetic probes showed the best performance, scoring 0.80 accuracy and 0.51 Matthews Correlation Coefficient on testset. Then, we generated and validated a progression signature based on 4 methylation probes capable of stratifying HCC patients at high and low risk of progression. Survival analysis showed that high risk patients are characterized by a poorer progression free survival compared to low risk patients. Moreover, decision curve analysis confirmed the strength of this predictive tool over conventional clinical parameters. Functional enrichment analysis highlighted that high risk patients differentiated themselves by the upregulation of proliferative pathways. Ultimately, we propose the oncogenic MCM2 gene as a methylation-driven gene of which the representative epigenetic markers could serve both as predictive and prognostic markers. Briefly, our work provides several potential HCC progression epigenetic biomarkers as well as a new signature that may enhance patients surveillance and advances in personalized treatment.