GBM-Targeted oHSV Armed with Matrix Metalloproteinase 9 Enhances Anti-tumor Activity and Animal Survival

Abstract

The use of mutant strains of oncolytic herpes simplex virus (oHSV) in early-phase human clinical trials for the treatment of glioblastoma multiforme (GBM) has proven safe, but limited efficacy suggests that more potent vector designs are required for effective GBM therapy. Inadequate vector performance may derive from poor intratumoral vector replication and limited spread to uninfected cells. Vector replication may be impaired by mutagenesis strategies to achieve vector safety, and intratumoral virus spread may be hampered by vector entrapment in the tumor-specific extracellular matrix (ECM) that in GBM is composed primarily of type IV collagen. In this report, we armed our previously described epidermal growth factor receptor (EGFR)vIII-targeted, neuronal microRNA-sensitive oHSV with a matrix metalloproteinase (MMP9) to improve intratumoral vector distribution. We show that vector-expressed MMP9 enhanced therapeutic efficacy and long-term animal survival in a GBM xenograft model.

Keywords: Glioblastoma multiforme = GBM; OVs derived from herpes simplex virus = oHSV; Oncolytic vectors = OVs; extracellular matrix = ECM; matrix metalloproteinase 9 = MMP9.

Figure 1

KMMP9 Mediates Expression of Enzymatically Active MMP9 Compared to the KGW Control oHSV (A) The oHSV vectors were created from the KOS-37 BAC full-length genomic clone of the HSV-1 KOS strain. They are also deleted for the internal repeat (joint) region and carry two gain-of-function mutations in the external domain of glycoprotein B (gB) that improve virus entry (D285N/A549T, gB:N/T). The GFP (EGFP) gene was linked downstream of the glycoprotein C (gC) gene-coding sequence using a T2A skipping sequence in order to track virus replication and spread using fluorescence. The vectors were retargeted to tumor cells by modifying glycoprotein D (gD) using a single-chain variable fragment (scFv) antibody that recognizes both EGFR and its constitutively active mutant form EGFRvIII. Recognition of the natural gD receptors was ablated by deletion of the N-terminal amino acids 2–24 of gD and deletion of the residue at position 38 (Δ38). A double Red recombination technique was used to insert the Gateway cassette (GW) and the bovine growth hormone polyadenylation sequence (bGHpA) between UL3 and UL4 in order to produce KGW^BAC as an MMP9-deficient control vector. The oncolytic KMMP9^BAC was created by Gateway recombination, replacing the Gateway cassette with the MMP9 gene driven by the highly active CAG promoter. (B) Cell supernatants from U2OS mock-infected cells, U2OS cells transfected with a plasmid expressing MMP9 (+ Control), and U2OS cells infected with KGW or KMMP9 (MOI of 100 gc/cell) were collected 48 hpi and analyzed via a gelatin zymography assay followed by Coomassie blue staining. White bands are indicative of gelatinase activity (top panel), signifying enzymatically active MMP9. MMP9 expression in the media was verified by western blot analysis using a monoclonal anti-MMP9 antibody (middle panel). Monoclonal anti-tubulin antibody (bottom panel) was used to detect cellular tubulin gene expression as a loading control.

Figure 2

EGFR-Retargeted oHSV Expressing MMP9 Displays Entry Selectivity for EGFR-Bearing Cells and Enhanced Tumor Cell Killing In Vitro (A) KMMP9, KGW MMP9 negative control oHSV, and the KG4:T124 backbone viruses were employed to confirm EGFR-dependent entry of the EGFR-retargeted KMMP9 and KGW viruses into J cells expressing the EGFR (J/EGFR) compared to J cells expressing either of the two HSV gD cognate receptors nectin-1/HveC (J/C) or TNFRSF14/HVEM/HveA (J/A). Cells were infected at a range of multiplicities (MOI of 0.01, 0.1, and 1.0), and at 6 hpi, infected cell monolayers were immunostained for HSV ICP4 to detect virus entry/infection of the J cells bearing the different cell surface receptors. (B) U87, SNB19, and GBM30 glioma cell lines were infected with KMMP9 and the KGW control virus at an MOI of 10 genome copies (gc)/cell or 100 gc/cell for 3 (left panel) or 7 days (right panel), and the percent cell viability determined by a 3-(4,5-dimethylthiazol-2-yl)-2,5-dimethyltetrazolium bromide (MTT) assay was compared for each sample to uninfected cells. Bars represent the mean ± SD of six replicates. Significant differences between the KMMP9 and KGW control viruses (p < 0.05, unpaired Student’s t test) are indicated by an asterisk (*).

Figure 3

MMP9 Improves oHSV Spread in KMMP9-Infected GBM30 Spheroids (A) GBM30 cells were grown in suspension and infected with either KMMP9 or KGW (MOI of 0.01 or 0.05), both expressing the EGFP fluorescent reporter that is only expressed when the virus is actively replicating within the tumor cells. Viral replication and spread were monitored daily by visualization of EGFP expression in whole-mount spheroids. (B) GBM30 spheroids were infected with either KMMP9 or KGW (MOI of 0.01) and imaged by confocal microscopy at 72 hpi: (Ba) representative images of two KMMP9- and two KGW-infected spheroids using 3D reconstruction from 0 to 150 μm; (Bb) z sections of two KMMP9- and two KGW-infected spheroids at z = 100 μm. (Bc) Each spheroid was divided into five segments along the z axis. Relative signal intensity of each segment of the spheroid was calculated by averaging EGFP signal (green) divided by DAPI signal (blue) (n = 7; *p < 0.05).

Figure 4

KMMP9 Treatment of Nude Mice Bearing GBM30 Intracranial Tumors Increases Animal Survival Kaplan-Meier survival plot for nude mice receiving intracranial implant of GBM30 cells and 5 days later injected at the same coordinates with PBS or 10⁴ PFU of KMMP9 or KGW. Animals treated with KMMP9 or KGW survived significantly longer than did those treated with PBS (KMMP9, p < 0.001; KGW, p < 0.01). KMMP9 also showed significantly improved efficacy compared to KGW (p < 0.01).

Figure 5

MMP9 Improves Oncolysis Even in Well-Established Brain Tumors (A–C) T2-weighted brain MRI images of the three animals from each treatment group (A, PBS; B, KGW; and C, KMMP) on day 9 after tumor cell implantation and 1 day before virus injection (day 10) and on days 13, 16, 19, and 23 after tumor cell implantation (4, 7, 9, and 13 days after virus injection). (D) The overall tumor volume calculated based according to MRI imaging is plotted at different time points (n = 3; ****p < 0.0001).