Hello, my name is Shree Gopakumar. I'm a medical degree candidate at Baylor College of Medicine in Houston, Texas. Thank you to the Scientific Programming Committee for this opportunity. I will be discussing oncolytic viral delivery into the surgical resection cavity for treatment of glioblastoma in a murine model. This research was conducted at MD Anderson Cancer Center under the guidance of Professor and Chairman Dr. Frederick Lange. Dr. Lange is a patent holder on the oncolytic virus I'll be discussing today. We are familiar with the terrible prognosis of GBM, but specifically with regards to this project, it is important to highlight that surgical resection is part of the current standard of care, and this includes an opportunity for therapeutic agents to be delivered at the time of initial resection. One new treatment category that has emerged over the past decade for GBM is oncolytic virotherapy, in which viruses are genetically engineered to selectively replicate and kill cancer cells, but not normal surrounding tissue. The oncolytic virus DNX2401 used in this study was originally developed at the MD Anderson Brain Tumor Center and is a tumor-selective, replication-competent adenovirus. Tumor selectivity arises from a 24 base pair deletion in the E1A gene that prevents replication in normal cells. Additionally, an RGD motif increases integrin-mediated attachment to GBM cells. A Phase 1 clinical trial published in 2018 demonstrated that DNX2401 was capable of infecting, replicating in, and killing human glioma cells after direct intratumoral injection into recurrent human GBM, ultimately achieving complete and durable responses in 12% of patients, with 20% of patients surviving over 4 years after treatment. To increase treatment efficacy, our group hoped to combine upfront surgical resection with oncolytic viral therapy, however the optimal method for delivering oncolytic virus into the surgical resection cavity has not been determined. Our group turned to these bone marrow-derived mesenchymal stem cells, or MSCs, as cellular vehicles for viral delivery, because MSCs have been previously shown to be effective carriers of oncolytic virus, with intrinsic tumortropic properties, seen here, that allow them to migrate selectively towards gliomas. But one major concern of using MSCs was how to best increase retention of these stem cells and virus within the hostile, necrotic, and hypoxic nature of the resection cavity and prolong exposure of the stem cells releasing virus to glioma cells. To this end, work in other labs has shown that using fibrin as a scaffold for MSCs is highly effective in retaining the cells within the surgical bed so that they are not lost with intraoperative surgical irrigation or postoperative CSF flow. We hypothesized that delivering MSCs loaded with DNx2401 into the tumor resection cavity using a fibrin scaffold will retain the stem cells within the surgical bed and eradicate residual tumor cells, thereby decreasing tumor recurrence and extending overall survival. This is a schematic of the overall therapeutic model, with tumor resection, transplantation of virus-loaded stem cells in fibrinogen, and the addition of thrombin to form the fibrin scaffold within the resection cavity that delivers virus with the goal of eradicating residual disease. To test this hypothesis and concept, we first performed in vitro transwell studies to investigate if fibrin would interfere with stem cell viral delivery. To do so, we wanted to confirm the ability of virus-loaded MSCs secured in fibrin to deliver oncolytic virus through a 0.4 micrometer transwell pore to U87 tumor cells plated below. Transwell results showed that treatment with MSCs loaded with DNx virus were just as effective in releasing virus with cytotoxic effect on tumor cells whether they were seated in fibrin or not, indicating that fibrin does not hinder viral release from stem cells. For in vivo studies, we used a murine-glioma model of tumor resection and recurrence. Half a million U87 tumor cells were implanted into the left parietal lobes of athomic mice to a depth of 1 millimeter. Fluorescence-guided surgical resection was performed after 10 days of tumor xenograft growth, and the treatment was delivered in the resection cavity. Weakly luciferase bioluminescence imaging was used to monitor and track tumor recurrence thereafter. For this murine model of glioma resection and recurrence, the U87 glioma cell line we implanted was transduced with both mCherry protein, which fluoresces red, to aid with surgical resection, as well as the luciferase protein to track tumor recurrence with BLI imaging. Shown here are key elements of the surgical procedure, with a 2x2 mm craniectomy performed, exposure and visualization of the tumor under normal light, followed by surgical debulking using fluorescent microscopy under 8x magnification, with minimal residual disease left behind. At this point, MSCs loaded with DNx and fibrin were delivered into the resection cavity. In these images, the MSCs are labeled with GFP to confirm the presence of stem cells within the cavity intraoperatively. Fluorescence microscopy of collected brains 4 days after surgery and sectioned through the resection cavity confirmed the intracavitary presence of GFP-labeled MSCs within the tumor bed. Complete tumor eradication was seen within the brains of 5 out of 10, or 50% of animals treated with MSCs loaded with DNx and fibrin, with complete responses seen at varying time points. Selected H&E stained slides comparing control brains with large tumor recurrence in successfully treated animals with no evidence of residual GBM at the resection site were observed. Kaplan-Meier survival analysis, specifically comparing the green treatment curve and the yellow control curve, demonstrates a significant survival benefit with treatment using MSCs and DNx2401 in fibrin, with a 40% overall survival and extended median overall survival to 100 days. In addition, comparing the green and black curves, the use of fibrin specifically enhanced treatment effectiveness when compared to delivery of virus-loaded stem cells without fibrin, suggesting the importance of fibrin in holding the MSCs within the resection site. Finally, no difference was observed between the blue and green curves, suggesting that the value of MSC tumor tropism remains unclear in this U87 glioma cell line model, as U87 xenografts tend to grow as uniform spheres without invasive foci. Overall, this preclinical translational study validates delivery of oncolytic biotherapy using MSCs and fibrin as a treatment strategy for attempting to prevent GBM tumor recurrence after surgical resection. Future studies utilizing a more invasive tumor cell line may better elucidate the value of intrinsic MSC tumor tropism for eliminating invading GBM foci. Ultimately, however, we hope this project is a step forward to the eventual clinical translation and application of this approach in human patients undergoing surgical resection of primary GBM. I would like to express my gratitude for all members of the MD Anderson Brain Tumor Center for their assistance with this project, as well as recognize my funding sources. Thank you.

oncolytic viral delivery

surgical resection cavity

glioblastoma

MSCs as carriers

DNX2401