Development of recombinant vesicular stomatitis viruses that exploit defects in host defense to augment specific oncolytic activity

Abstract

Vesicular stomatitis virus (VSV) is a negative-stranded RNA virus normally sensitive to the antiviral actions of alpha/beta interferon (IFN-alpha/beta). Recently, we reported that VSV replicates to high levels in many transformed cells due, in part, to susceptible cells harboring defects in the IFN system. These observations were exploited to demonstrate that VSV can be used as a viral oncolytic agent to eradicate malignant cells in vivo while leaving normal tissue relatively unaffected. To attempt to improve the specificity and efficacy of this system as a potential tool in gene therapy and against malignant disease, we have genetically engineered VSV that expresses the murine IFN-beta gene. The resultant virus (VSV-IFNbeta) was successfully propagated in cells not receptive to murine IFN-alpha/beta and expressed high levels of functional heterologous IFN-beta. In normal murine embryonic fibroblasts (MEFs), the growth of VSV-IFNbeta was greatly reduced and diminished cytopathic effect was observed due to the production of recombinant IFN-beta, which by functioning in a manner involving autocrine and paracrine mechanisms induced an antiviral effect, preventing virus growth. However, VSV-IFNbeta grew to high levels and induced the rapid apoptosis of transformed cells due to defective IFN pathways being prevalent and thus unable to initiate proficient IFN-mediated host defense. Importantly, VSV expressing the human IFN-beta gene (VSV-hIFNbeta) behaved comparably and, while nonlytic to normal human cells, readily killed their malignant counterparts. Similar to our in vitro observations, following intravenous and intranasal inoculation in mice, recombinant VSV (rVSV)-IFNbeta was also significantly attenuated compared to wild-type VSV or rVSV expressing green fluorescent protein. However, VSV-IFNbeta retained propitious oncolytic activity against metastatic lung disease in immunocompetent animals and was able to generate robust antitumor T-cell responses. Our data indicate that rVSV designed to exploit defects in mechanisms of host defense can provide the basis for new generations of effective, specific, and safer viral vectors for the treatment of malignant and other disease.

FIG. 1.

(A) Construction of VSV-IFNβ cDNA. Mouse IFN-β cDNA was inserted between the glycoprotein (G) and polymerase protein (L) genes of pVSV-XN2, a full-length cDNA clone of the Indiana serotype, as described in Materials and Methods. N, nucleocapsid protein gene; P, phosphoprotein gene; M, matrix protein gene. (B) One-step growth curves of recombinant viruses. BHK-21 cells were infected at an MOI of 10 PFU per cell with VSV-IFNβ, VSV-GFP, or rVSV not containing a foreign gene. The supernatants of infected cells were harvested at the indicated times, and viral titers were determined by a standard plaque assay. (C) Expression of IFN-β in VSV-IFNβ-infected cells. BHK-21 cells were infected with VSV-IFNβ or VSV-GFP at an MOI of 10 PFU per cell for 24 h. Supernatants (25 μl from 1 ml of 10⁶ infected cells) were analyzed for IFN-β expression by immunoblotting using an anti-mouse IFN-β polyclonal antibody. Purified recombinant mouse IFN-β (UnitedStates Biological) (120 ng) was used as a positive control. (D) Biological assay of IFN-β expressed by VSV-IFN. BHK-21 cells were mock infected or infected with VSV-GFP, VSV-IFNβ, or rVSV at an MOI of 10 PFU per cell for 24 h. Supernatants (500 μl) were HI to remove residual virus and used to incubate B16(F10) cells for 24 h. Treated cells were then infected with wild-type VSV (Indiana strain) at an MOI of 0.1 for 24 h, and CPE was assessed under microscopy. (Di) B16(F10) cells treated with supernatants from VSV-GFP-infected BHK-21 cells; (Dii) B16(F10) cells treated with supernatants from rVSV-infected BHK-21 cells; (Diii) B16(F10) cells treated with supernatants from VSV-IFNβ-infected BHK-21 cells; (Div) B16(F10) cells treated with purified recombinant murine IFN-α/β (1,000 U/24 h); (Dv) B16(F10) cells treated with medium from uninfected BHK-21 cells; (Dvi) uninfected B16(F10) cells. (E) Quantitation of IFN-β expressed by VSV-IFNβ. The supernatant of BHK-21 cells infected with VSV-IFNβ or VSV-GFP at an MOI of 10 PFU per cell for 24 h was serially diluted and used to incubate C57BL/6 primary cells for 24 h. The cells were infected with wild-type VSV (Indiana strain) at an MOI of 0.01 PFU per cell for 24 h, and CPE was assessed by microscopy. Biological activity data are presented as units per milliliter and represent the means from two experiments.

FIG. 2.

In vitro oncolytic activity of VSV-IFNβ. TS/A (A), DA-3 (B), or B16(F10) (C) cells were mock infected or infected with VSV-IFNβ or VSV-GFP at an MOI of 0.1 (open bars) or 5 (solid bars) PFU per cell for 24 h. After this period, cell viability was assessed by trypan blue exclusion assay (A, B, and C) or visually by microscopy (magnification, ×20) (D).

FIG. 3.

In vitro replication of VSV-IFNβ in primary or transformed cells. Primary or transformed MEFs derived from BALB/c (A and B) or C57BL/6 (C to E) cells were mock infected or infected with VSV-IFNβ or VSV-GFP at an MOI of 1 (shaded bars), 0.1 (open bars), or 0.01 (solid bars) PFU per cell for 24 h (A and D) or infected at an MOI of 0.01 PFU per cell for 24 (shaded bars in panels B and C and panels in panel E), 48 (open bars), or 72 (solid bars) h (B, C, and E). Cell viability was assessed by trypan blue exclusion assay (A to D) or visually by microscopy (magnification, ×20) (E).

FIG.4.

In vitro replication of rVSV expressing the hIFNβ gene in human normal and tumor cell lines. (A) Expression of hIFNβ in VSV-hIFNβ-infected cells. BHK-21 cells were infected with VSV-hIFNβ or VSV-GFP at an MOI of 5 PFU per cell for 24 h. Supernatants (25 μl of 1 ml from 10⁶ infected cells) were analyzed for hIFNβ expression by immunoblotting using an anti-hIFNβ polyclonal antibody. Purified recombinant hIFNβ (5,000 U) was used as a positive control. (B) In vitro replication of VSV-hIFNβ in human normal cells. Human foreskin fibroblast cells (hTERT) were mock infected or infected with VSV-hIFNβ or VSV-GFP at an MOI of 1 (shaded bars), 0.1 (open bars), or 0.01 (solid bars) PFU per cell for 48 h. Cell viability was assessed by trypan blue exclusion assay. (C) In vitro replication of VSV-hIFNβ in human tumor cells. HeLa cells were mock infected or infected with VSV-hIFNβ or VSV-GFP at an MOI of 1 (shaded bars), 0.1 (open bars), or 0.01 (solid bars) PFU per cell for 48 h. Cell viability was assessed by trypan blue exclusion assay. (D) CPE of VSV-hIFNβ-infected human cells. hTERT or HeLa cells were mock infected or infected with VSV-hIFNβ or VSV-GFP at an MOI of 0.01 PFU per cell for 48 h. After this period, cell viability was assessed visually by microscopy (magnification, ×20). (E) In vitro oncolytic activity of VSV-hIFNβ. Human tumor cell lines (MCF-7, K562, and PC-3) or normal cell lines (HMVECs and HMECs) were mock infected or infected with VSV-hIFNβ or VSV-GFP at an MOI of 0.1 PFU per cell for 48 h. Cell viability was assessed by trypan blue exclusion assay.

FIG. 5.

In vitro replication of VSV-IFNβ in Stat1^+/+ or Stat1^−/− cells. Immortalized embryonic fibroblasts derived from Stat1^+/+ (A and B) or Stat1^−/− (C and D) C57BL/6 mice were treated with (solid bars) or without (open bars) murine IFN-α/β (1,000 U/ml) for 24 h and were mock infected or infected with VSV-IFNβ or VSV-GFP at an MOI of 0.01 PFU per cell for 24 h. Cell viability was assessed by trypan blue exclusion assay (A and C) or visually by microscopy (magnification, ×20) (B and D).

FIG. 6.

VSV-IFNβ is attenuated in vivo. (A) Various amounts of VSV-IFNβ, VSV-GFP, or rVSV without an insert or wild-type (WT) VSV (Indiana strain) were used for i.v. inoculation of BALB/c mice (6 to 8 weeks old). Mortality of the animals was monitored. (B) Survival times of animals receiving i.v. inoculations of 10⁸ PFU of VSV-IFNβ, VSV-GFP, or wild-type VSV (Indiana strain). (C) Weights of BALB/c animals that received i.v. inoculations of phosphate-buffered saline or recombinant viruses at the various doses indicated and that were monitored for 12 weeks postinfection. No residual or persistent virus was detected in a variety of organs analyzed at 20 weeks postinfection.

FIG. 7.

rVSV-IFNβ inhibits the growth of syngeneic renal cell carcinoma. (A) BALB/c mice (6 to 8 weeks old) were implanted s.c. with 7 × 10⁵ Renca cells. After palpable tumors had formed, animals were treated intratumorally with 1.5 × 10⁸ PFU of VSV-IFNβ, VSV-GFP, or HI-VSV. Three injections of the virus were given as indicated by the arrows. Tumor volumes were calculated, and values are means ± standard errors of the means (n = 6). (B) Tumor volumes (in cubic millimeters) were measured at day 19 postimplantation, and animals that showed no tumor development at 1 month after treatment were identified. (C) Antitumor IFN-γ-secreting T cells were measured in animals (n = 2) receiving VSV-GFP, VSV-IFNβ, or HI-VSV at 10 days posttreatment as described in Results.

FIG. 8.

VSV-IFNβ increases the survival of immunocompetent animals harboring metastatic tumors. BALB/c animals (6 to 8 weeks; n = 8) were injected via tail vein with 5 × 10⁴ TS/A cells, followed 2 days later by a single i.v. injection (5 × 10⁷ PFU) of VSV-GFP, VSV-IFNβ, or HI-VSV. Following injection, the survival rates of the animals were monitored. A total of 50% of animals receiving VSV-IFNβ continued to remain disease free after 120 days.