The concept that extracellular vesicles (EVs) from the diet can be absorbed by the intestinal tract of the consuming organism, be bioavailable in various organs, and in-turn exert phenotypic changes is highly debatable. Here, we isolate EVs from both raw and commercial bovine milk and characterize them by electron microscopy, nanoparticle tracking analysis, western blotting, quantitative proteomics and small RNA sequencing analysis. Orally administered bovine milk-derived EVs survive the harsh degrading conditions of the gut, in mice, and is subsequently detected in multiple organs. Milk-derived EVs orally administered to mice implanted with colorectal and breast cancer cells reduce the primary tumor burden. Intriguingly, despite the reduction in primary tumor growth, milk-derived EVs accelerate metastasis in breast and pancreatic cancer mouse models. Proteomic and biochemical analysis reveal the induction of senescence and epithelial-to-mesenchymal transition in cancer cells upon treatment with milk-derived EVs. Timing of EV administration is critical as oral administration after resection of the primary tumor reverses the pro-metastatic effects of milk-derived EVs in breast cancer models. Taken together, our study provides context-based and opposing roles of milk-derived EVs as metastasis inducers and suppressors.

Osteosarcoma, a common malignancy in large dog breeds, typically metastasises from long bones to lungs and is usually fatal within 1 to 2 years of diagnosis. Better therapies are needed for canine patients and their human counterparts, a third of whom die within 5 years of diagnosis. We compared the in vitro sensitivity of canine osteosarcoma cells derived from 4 tumours to the currently used chemotherapy drugs doxorubicin and carboplatin, and 4 new anti-cancer drugs. Agents targeting histone deacetylases or PARP were ineffective. Two of the 4 cell lines were somewhat sensitive to the BH3-mimetic navitoclax. The proteasome inhibitor bortezomib potently induced caspase-dependent apoptosis, at concentrations substantially lower than levels detected in the bones and lungs of treated rodents. Co-treatment with bortezomib and either doxorubicin or carboplatin was more toxic to canine osteosarcoma cells than each agent alone. Newer proteasome inhibitors carfilzomib, ixazomib, oprozomib and delanzomib manifested similar activities to bortezomib. Human osteosarcoma cells were as sensitive to bortezomib as the canine cells, but slightly less sensitive to the newer drugs. Human osteoblasts were less sensitive to proteasome inhibition than osteosarcoma cells, but physiologically relevant concentrations were toxic. Such toxicity, if replicated in vivo, may impair bone growth and strength in adolescent human osteosarcoma patients, but may be tolerated by canine patients, which are usually diagnosed later in life. Proteasome inhibitors such as bortezomib may be useful for treating canine osteosarcoma, and ultimately may improve outcomes for human patients if their osteoblasts survive exposure in vivo, or if osteoblast toxicity can be managed.

Outcomes for patients diagnosed with the bone cancer osteosarcoma have not improved significantly in the last four decades. Only around 60% of patients and about a quarter of those with metastatic disease survive for more than five years. Although DNA-damaging chemotherapy drugs can be effective, they can provoke serious or fatal adverse effects including cardiotoxicity and therapy-related cancers. Better and safer treatments are therefore needed. We investigated the anti-osteosarcoma activity of IAP antagonists (also known as Smac mimetics) using cells from primary and metastatic osteosarcomas that arose spontaneously in mice engineered to lack p53 and Rb expression in osteoblast-derived cells. The IAP antagonists SM-164, GDC-0152 and LCL161, which efficiently target XIAP and cIAPs, sensitized cells from most osteosarcomas to killing by low levels of TNFα but not TRAIL. RIPK1 expression levels and activity correlated with sensitivity. RIPK3 levels varied considerably between tumors and RIPK3 was not required for IAP antagonism to sensitize osteosarcoma cells to TNFα. IAP antagonists, including SM-164, lacked mutagenic activity. These data suggest that drugs targeting XIAP and cIAP1/2 may be effective for osteosarcoma patients whose tumors express abundant RIPK1 and contain high levels of TNFα, and would be unlikely to provoke therapy-induced cancers in osteosarcoma survivors.

Osteosarcoma is the most common form of primary bone cancer. Over 20% of osteosarcoma patients present with pulmonary metastases at diagnosis, and nearly 70% of these patients fail to respond to treatment. Previous work revealed that human and canine osteosarcoma cell lines are extremely sensitive to the therapeutic proteasome inhibitor bortezomib in vitro. However, bortezomib has proven disappointingly ineffective against solid tumors including sarcomas in animal experiments and clinical trials. Poor tumor penetration has been speculated to account for the inconsistency between in vitro and in vivo responses of solid tumors to bortezomib. Here we show that the second-generation proteasome inhibitor ixazomib, which reportedly has enhanced solid tumor penetration compared to bortezomib, is toxic to human and canine osteosarcoma cells in vitro. We used experimental osteosarcoma metastasis models to compare the efficacies of ixazomib and bortezomib against primary tumors and metastases derived from luciferase-expressing KRIB or 143B human osteosarcoma cell lines in athymic mice. Neither proteasome inhibitor reduced the growth of primary intramuscular KRIB tumors, however both drugs inhibited the growth of established pulmonary metastases created via intravenous inoculation with KRIB cells, which were significantly better vascularized than the primary tumors. Only ixazomib slowed metastases from KRIB primary tumors and inhibited the growth of 143B pulmonary and abdominal metastases, significantly enhancing the survival of mice intravenously injected with 143B cells. Taken together, these results suggest ixazomib exerts better single agent activity against osteosarcoma metastases than bortezomib. These data provide hope that incorporation of ixazomib, or other proteasome inhibitors that penetrate efficiently into solid tumors, into current regimens may improve outcomes for patients diagnosed with metastatic osteosarcoma.

Development of drugs targeting Bcl-2 relatives and caspases, for treating diseases including cancer and inflammatory disorders, often involves measuring interactions with recombinant target molecules, and/or monitoring cancer cell killing in vitro. Here, we present yeast-based methods for evaluating drug-mediated inhibition of Bcl-2 relatives or caspases. Active Bax and caspases kill Saccharomyces cerevisiae, and pro-survival Bcl-2 proteins can inhibit Bax-induced yeast death. By measuring the growth or adenosine triphosphate content of transformants co-expressing Bax with pro-survival Bcl-2 relatives, we found that the Bcl-2 antagonist drugs ABT-737 or ABT-263 abolished Bcl-2 or Bcl-xL function and reduced Bcl-w activity, but failed to inhibit Mcl-1, A1 or the poxvirus orthologs DPV022 and SPPV14. Using this technique, we also demonstrated that adenoviral E1B19K was resistant to these agents. The caspase inhibitor Q-VD-OPh suppressed yeast death induced by caspases 1 and 3. Yeast engineered to express human apoptotic regulators enable simple, automatable assessment of the activity and specificity of candidate drugs targeting Bcl-2 relatives or caspases.