Glioblastoma is the most prevalent and malignant brain tumor identified in adults. Surgical resection followed by radiotherapy and chemotherapy, mainly with temozolomide (TMZ), is the chosen treatment for this type of tumor. However, the average survival of patients is around 15 months. Novel approaches to glioblastoma treatment are greatly needed. Here, we aimed to investigate the anti-glioblastoma effect of the combination of matteucinol (Mat) (dihydroxyflavanone derived from Miconia chamissois Naudin) with the chemotherapeutic TMZ in vitro using tumor (U-251MG) and normal astrocyte (NHA) cell lines and in vivo using the chick embryo chorioallantoic membrane (CAM) assay. The combination was cytotoxic and selective for tumor cells (28 μg/mL Mat and 9.71 μg/mL TMZ). Additionally, the combination did not alter cell adhesion but caused morphological changes characteristic of apoptosis in vitro. Notably, the combination was also able to reduce tumor growth in the chick embryo model (CAM assay). The docking results showed that Mat was the best ligand to the cell death membrane receptor TNFR1 and to TNFR1/TMZ complex, suggesting that these two molecules may be working together increasing their potential. In conclusion, Mat-TMZ can be a good candidate for pharmacokinetic studies in view of clinical use for the treatment of glioblastoma.

Cancer remains one of the leading causes of death worldwide and continues to pose a serious public health challenge. The limited success of many current treatments—often due to toxicity, poor selectivity, and the development of drug resistance—highlights the need for new and more effective therapeutic options. Phytochemicals have emerged as a valuable source of anticancer agents, offering rich structural diversity and a wide range of biological activities. However, identifying promising compounds from the vast chemical space of natural products remains difficult using conventional screening methods, which are typically slow, costly, and inefficient. In recent years, artificial intelligence (AI) has begun to transform phytochemical-based drug discovery. Machine learning and deep learning approaches are now used to support key steps in the discovery process, including metabolite identification, virtual screening, target prediction, and toxicity assessment. By integrating chemical, biological, and multi-omics data, AI enables a more systematic and data-driven exploration of natural product diversity. Despite these advances, challenges persist, particularly the scarcity of high-quality experimental data, the structural complexity of phytochemicals, and their limited representation in public databases. This review critically examines current AI applications in phytochemical-based anticancer drug discovery and discusses emerging strategies aimed at overcoming these limitations. Overall, AI-driven phytochemical screening represents a promising path toward accelerating the development of next-generation cancer therapies. Graphical Abstract

Glioblastoma is one of the most aggressive and widely recognized types of brain tumors, characterized by significant cellular and molecular diversity and an inherently aggressive nature. The treatment remains highly challenging, with limited effectiveness and persistently low survival rates. For this reason, researchers are continuously expanding the chemical space of anticancer agents by exploring complex sources such as animal venoms and incorporating in silico tools to accelerate discovery. Indeed, venoms serve as libraries of proteins and peptides, providing a rich source of novel chemical structures for glioblastoma therapy. Some review articles have examined the mechanisms by which venom‐derived peptides target glioblastoma cell lines; nevertheless, key structural insights and computational analyses remain underexplored. In this era of artificial intelligence (AI) and advancements in in silico approaches, our review documented the antiglioblastoma properties of venom peptides and underscores the value of computational methods in peptide‐based drug development. To this end, a comprehensive search was conducted in PubMed, Elsevier, Springer, Lilacs, Google Scholar, and SciELO databases. Furthermore, in silico analyses were conducted to evaluate the anticancer potential, hemolytic activity, toxicity, and blood–brain barrier (BBB) penetrating properties of venom‐derived peptides. In total, 26 unique sequences were identified, with their structural properties and mechanisms of cell death comprehensively characterized. The development of peptide‐based anticancer drugs remains in its early stages, with minimal advancement toward preclinical evaluation using in vivo models. The advancement of AI models offers opportunities to accelerate peptide discovery. However, our case study revealed divergences among AI‐based predictions, as well as discrepancies between computational and experimental findings, underscoring the need for further model refinement and validation through experimental data integration. In summary, venoms remain promising peptide libraries that offer valuable natural molecular templates. These peptides require chemical optimization to enhance their stability and BBB permeability. Such advances could enable selective targeting within the glioblastoma niche and support the development of more effective therapies.

The discovery of novel cytotoxic drugs is of paramount importance in contemporary medical research, particularly in the search for treatments with fewer side effects and higher specificity. Antimicrobial peptides are an interesting class of molecules for this endeavor. In this context, the LyeTx III, a new peptide extracted from the venom of the Lycosa erythrognatha spider, stands out. The peptide exhibits typical antimicrobial traits: a positive net charge and amphipathic α -helix structure in lipid-like environments. Its unique sequence (GKAMKAIAKFLGR-NH₂), identified via mass spectrometry and Edman degradation, shows limited similarity to existing peptides. Significantly, when liposome-encapsulated, LyeTx III demonstrates selective activity against tumor cells in culture. Our MTT results showed that the cytotoxicity of the peptide increased against HN13 cells when administered as liposomes, with their viability in HN13 cells alone being 98%, compared to 38% in liposome-encapsulated form. This finding underscores that the LyeTx III peptide may be a good candidate for the development of new drugs against cancer. Its activity when encapsulated is promising, as it can increase its half-life in the body and can also be targeted to specific tumors.

Cancer, a longstanding global challenge, remains a leading cause of death, prompting an urgent search for effective treatments. Conventional therapies, while prevalent, often cause adverse effects due to their lack of specificity. This review explores an innovative approach, focusing on animal toxins as a rich source of bioactive compounds which have demonstrated efficacy against cancer cells. Peptides from various species, including scorpions, snakes, bees, spiders, and frogs, show promising antiproliferative and cytotoxic effects. Emphasizing the most prevalent types of cancer, this review outlines the discovery and development stages of potential anticancer drugs derived from toxin peptides. The comprehensive overview includes in vitro and in vivo assessments of their anticancer activity and toxicity. This pioneering exploration extends to different tumors, offering valuable insights into mechanisms of action and potential therapeutic applications. The findings highlight a paradigm shift in cancer research, showcasing the potential of toxin-derived compounds for developing targeted and efficient cancer therapies with reduced side effects.