Temozolomide (TMZ) is an oral alkylating agent used for the treatment of glioblastoma and is now becoming a chemotherapeutic option in patients diagnosed with high-risk low-grade gliomas. The O-6-methylguanine-DNA methyltransferase (MGMT) is responsible for the direct repair of the main TMZ-induced toxic DNA adduct, the O6-Methylguanine lesion. MGMT promoter hypermethylation is currently the only known biomarker for TMZ response in glioblastoma patients. Here we show that a subset of recurrent gliomas carries MGMT genomic rearrangements that lead to MGMT overexpression, independently from changes in its promoter methylation. By leveraging the CRISPR/Cas9 technology we generated some of these MGMT rearrangements in glioma cells and demonstrated that the MGMT genomic rearrangements contribute to TMZ resistance both in vitro and in vivo. Lastly, we showed that such fusions can be detected in tumor-derived exosomes and could potentially represent an early detection marker of tumor recurrence in a subset of patients treated with TMZ.

Herpes simplex virus 1 (HSV-1) is a neurotropic pathogen that can infect many types of cells and establishes latent infections in the neurons of sensory ganglia. In some cases, the virus spreads into the central nervous system, causing encephalitis or meningitis. Cells infected with several different types of viruses may secrete microvesicles (MVs) containing viral proteins and RNAs. In some instances, extracellular microvesicles harboring infectious virus have been found. Here we describe the features of shedding microvesicles released by the human oligodendroglial HOG cell line infected with HSV-1 and their participation in the viral cycle. Using transmission electron microscopy, we detected for the first time microvesicles containing HSV-1 virions. Interestingly, the Chinese hamster ovary (CHO) cell line, which is resistant to infection by free HSV-1 virions, was susceptible to HSV-1 infection after being exposed to virus-containing microvesicles. Therefore, our results indicate for the first time that MVs released by infected cells contain virions, are endocytosed by naive cells, and lead to a productive infection. Furthermore, infection of CHO cells was not completely neutralized when virus-containing microvesicles were preincubated with neutralizing anti-HSV-1 antibodies. The lack of complete neutralization and the ability of MVs to infect nectin-1/HVEM-negative CHO-K1 cells suggest a novel way for HSV-1 to spread to and enter target cells. Taken together, our results suggest that HSV-1 could spread through microvesicles to expand its tropism and that microvesicles could shield the virus from neutralizing antibodies as a possible mechanism to escape the host immune response.IMPORTANCE Herpes simplex virus 1 (HSV-1) is a neurotropic pathogen that can infect many types of cells and establishes latent infections in neurons. Extracellular vesicles are a heterogeneous group of membrane vesicles secreted by most cell types. Microvesicles, which are extracellular vesicles which derive from the shedding of the plasma membrane, isolated from the supernatant of HSV-1-infected HOG cells were analyzed to find out whether they were involved in the viral cycle. The importance of our investigation lies in the detection, for the first time, of microvesicles containing HSV-1 virions. In addition, virus-containing microvesicles were endocytosed into CHO-K1 cells and were able to actively infect these otherwise nonpermissive cells. Finally, the infection of CHO cells with these virus-containing microvesicles was not completely neutralized by anti-HSV-1 antibodies, suggesting that these extracellular vesicles might shield the virus from neutralizing antibodies as a possible mechanism of immune evasion.

21Temozolomide (TMZ) is an oral alkylating agent used for the treatment of glioblastoma and is 22 now becoming a chemotherapeutic option in patients diagnosed with high-risk low-grade 23 gliomas. The O-6-methylguanine-DNA methyltransferase (MGMT) is responsible for the direct 24 repair of the main TMZ-induced toxic DNA adduct, the O6-Methylguanine lesion. MGMT 25 promoter hypermethylation is currently the only known biomarker for TMZ response in 26 glioblastoma patients. Here we show that a subset of recurrent gliomas carries MGMT genomic 27 rearrangements that lead to MGMT overexpression, independently from changes in its promoter 28 methylation. By leveraging the CRISPR/Cas9 technology we generated some of these MGMT 29 rearrangements in glioma cells and demonstrated that the MGMT genomic rearrangements 30 contribute to TMZ resistance both in vitro and in vivo. Lastly, we showed that such fusions can 31 be detected in tumor-derived exosomes and could potentially represent an early detection marker 32 of tumor recurrence in a subset of patients treated with TMZ. Introduction 34The therapeutic benefits of TMZ depend on its ability to methylate DNA, which takes place at 35 the N-7 and O-6 positions of guanine and N-3 position of adenine. Although the minor product 36 O6-Methylguanine (O6-meG) accounts for less than 10% of total alkylation, it exerts the greatest 37 potential for apoptosis induction 1 . O6-meG pairs with thymine as opposed to cytosine during 38 DNA replication. The O6-meG:thymine mismatch can be recognized by the post-replication 39Mismatch Repair (MMR) system and, according to the futile repair hypothesis, ultimately 40 induces DNA double-strand breaks, cell cycle arrest and cell death 2 . The O-6-methylguanine-41 DNA methyltransferase (MGMT) is responsible for the direct repair of O6-meG lesion by 42 transferring the alkyl group from guanine to a cysteine residue. Epigenetic silencing, due to 43 promoter methylation, of the MGMT gene prevents the synthesis of this enzyme, and as a 44 consequence increases the tumours sensitivity to the cytotoxic effects induced by TMZ and other 45 alkylating compounds 3,4 . As today, MGMT promoter hypermethylation is the only known 46 biomarker for TMZ response 4 . However, the discordance between promoter methylation and 47 Nevertheless, the desired genomic rearrangements were further validated using a break-apart 132 fluorescence in situ hybridization (FISH) assay ( Supplementary Fig. 4). 133 134