Generation of an Oncolytic Herpes Simplex Viral Vector Completely Retargeted to the GDNF Receptor GFRα1 for Specific Infection of Breast Cancer Cells

Abstract

Oncolytic herpes simplex viruses (oHSV) are under development for the treatment of a variety of human cancers, including breast cancer, a leading cause of cancer mortality among women worldwide. Here we report the design of a fully retargeted oHSV for preferential infection of breast cancer cells through virus recognition of GFRα1, the cellular receptor for glial cell-derived neurotrophic factor (GDNF). GFRα1 displays a limited expression profile in normal adult tissue, but is upregulated in a subset of breast cancers. We generated a recombinant HSV expressing a completely retargeted glycoprotein D (gD), the viral attachment/entry protein, that incorporates pre-pro-GDNF in place of the signal peptide and HVEM binding domain of gD and contains a deletion of amino acid 38 to eliminate nectin-1 binding. We show that GFRα1 is necessary and sufficient for infection by the purified recombinant virus. Moreover, this virus enters and spreads in GFRα1-positive breast cancer cells in vitro and caused tumor regression upon intratumoral injection in vivo. Given the heterogeneity observed between and within individual breast cancers at the molecular level, these results expand our ability to deliver oHSV to specific tumors and suggest opportunities to enhance drug or viral treatments aimed at other receptors.

Keywords: breast cancer; herpes simplex virus; oncolytic.

Figure 1

Vector engineering. (A) Schematic depiction of the herpes simplex virus (HSV) genome. Terminal repeats (TR_L, TR_S) of the unique long (U_L) and unique short (U_S) segment, respectively; inverted internal repeats (IR_L, IR_S) of the unique long (U_L) and unique short (U_S) segment, respectively. The KNTc-ΔgD:GW genome represented underneath contains loxP-flanked bacterial artificial chromosome (BAC) sequences, a ubiquitin C (UbC)-promoter driven mCherry cassette (UbC-mCherry), a GW cassette in place of the glycoprotein D (gD) coding sequence (ΔgD:GW), and mutations in glycoprotein B that enhance retargeted virus entry (gB:NT) [36]. (B) Illustration of wild type gD, depicting the approximate size and location of the signal peptide (SP), transmembrane domain (TM), and cytoplasmic tail (CT), and the recombinant gD proteins gD:GDNF and gD:GDNFΔ38; Y38/Δ38 refers to the wild type (wt) gD residue 38 (Y) or the deleted (∆) residue 38. In both recombinant proteins the pre-pro-GDNF ligand was tethered to the C-terminus of gD at amino acid 25 by a flexible glycine-serine spacer (GS; (G₄S) × 3). (C) Western blot analysis of gD:wt and gD:GDNF proteins incorporated into purified virus particles. 1 × 10⁸ genome copies (gc) of purified virus particles were loaded per lane and blotted with either anti-gD or anti-gB antibodies. Lanes 1 and 2 represent virus stocks from two independent KNTc-gD:GDNF BAC isolates. The virus stock from lane 1 was used for the experiments pictured in Figure 2; the two isolates yielded comparable results.

Figure 2

Receptor dependence of KNTc-gD:GDNF virus entry. Infection of (A) J1.1–2 and J-GFRα1 cells or (B) B78H1 and B78-GFRα1 cells with KNTc-gD:GDNF virus at the indicated multiplicities of infection (MOIs) based on pfu (plaque-forming unit) titers on GFRα1-transduced U2OS cells (U2OS-ICP4-GFRα1, Figure S1) (Table 1). Virus entry into cells was visualized at 24 h post infection (hpi) as mCherry fluorescence.

Figure 3

Receptor specificity of KNTc-gD:GDNFΔ38 virus entry. Infection of (A) J1.1-2, J–C (nectin-1⁺) and J-GFRα1 cells or (B) B78H1, B78-C and B78-GFRα1 cells with KNTc-gD:GDNFΔ38 and KNTc-gD:wt viruses at 0.5 pfu/cell. Virus entry into cells was visualized as mCherry fluorescence at 24 hpi.

Figure 4

Virus infection of breast cancer cell lines. (A) MCF7, HCC1500, MDA-MB-231, and MDA-MB-453 cell lines were assessed for GFRα1 expression by Western blot analysis of whole cell lysates; β-actin detection was used as a loading control. (B) HCC1500, MCF7, MDA-MB-231, and MDA-MB-453 cells were infected with KNTc-gD:GDNFΔ38 or KNTc-gD:wt virus at 3 pfu/cell and flow cytometry for gB protein on the cell surface was performed at 24 hpi. For each cell line, uninfected cells are shown in blue compared to virus infected cells in red; KNTc-gD:wt infected cells (WT) and KNTc-gD:GDNFΔ38 infected cells (GDNFΔ38). (C) Quantification of gB⁺ cell populations as a percent of the total analyzed cells for each treatment group.

Figure 5

Virus infection of GFRα1 siRNA transduced cells. (A) MCF7 cells transfected with GFRα1-specific or non-specific (“Control”) siRNA pools were infected after 72 h with KNTc-gD:GDNFΔ38 (2 pfu/cell) or KNTc-gD:wt (0.2 pfu/cell) and stained at 6 hpi for ICP4 (upper rows) and DAPI (DAPI + ICP4, lower rows); representative images from triplicate infections are shown. (B) Quantification of the ICP4⁺ cell populations as a percentage of the total analyzed cells (DAPI) was performed for each treatment group using ImageJ software. Averages represent counts from 3–6 images ± SD; statistical differences were determined by two-way ANOVA (**** p < 0.0001). (C) Untreated and siRNA-treated MCF7 cells were assessed for GFRα1 protein expression by Western blot analysis of whole cell lysates; three biological replicates are shown for each condition. β-actin detection was used as loading control.

Figure 6

Virus-mediated cell death in vitro and tumor treatment. (A) MCF7 or (B) MDA-MB-453 cells were infected with KNTc-gD:GDNFΔ38 or KNTc-gD:wt virus at 3 pfu/cell and cell viability at 24, 48 and 72 hpi was measured by alamarBlue assay. Data are presented as the percentage of viable cells relative to uninfected cells at each time point. Averages presented at each time point represent 5–8 independent infections ± SEM. Statistics were determined by two-way ANOVA comparing virus infected cells to uninfected control cells at each time point. At 48 and 72 hpi, the viability of MCF7 cells infected with KNTc-gD:wt and KNTc-gD:GDNFΔ38 was significantly reduced compared to uninfected cells (KNTc-gD:wt, p < 0.0001 at 48 and 72 hpi, and KNTc-gD:GDNFΔ38, p = 0.0003 at 48 hpi and p < 0.0001 at 72 hpi). At 72 hpi, the viability of MDA-MB-453 cells infected with KNTc-gD:wt was significantly reduced compared to uninfected cells (p < 0.0001). The viability of MDA-MB-453 cells infected with KNTc-gD:GDNFΔ38 was not significantly different from that of uninfected cells at any time point tested. (C) MCF7 cells were implanted in the right hind flank in BALB/c athymic nude mice and tumors were injected with 1 × 10⁸ pfu of KNTc-gD:GDNFΔ38 or phosphate-buffered saline (PBS) when reaching a volume of approximately 70 mm³ (arrow, d22). Average tumor volumes in mm³ (mean ± SD of 3 animals/group) are presented over time. Statistical differences were determined by two-way ANOVA. KNTc-gD:GDNFΔ38 treated tumors were significantly reduced in volume compared to PBS-injected controls (d57, * p < 0.05; d64, ** p < 0.01; d72-d85, **** p < 0.0001).