Fusion peptide is superior to co-expressing subunits for arming oncolytic herpes virus with interleukin 12

Abstract

Background: G47∆ is a triple-mutated oncolytic herpes simplex virus type 1 (HSV-1) recently approved as a new drug for malignant glioma in Japan. As the next-generation, we develop armed oncolytic HSV-1 using G47∆ as the backbone. Because oncolytic HSV-1 elicits specific antitumor immunity, interleukin 12 (IL-12) can function as an effective payload to enhance the efficacy.

Methods: We evaluate the optimal methods for expressing IL-12 as a payload for G47∆-based oncolytic HSV-1. Two new armed viruses are generated for evaluation by employing different methods to express IL-12: T-mfIL12 expresses murine IL-12 as a fusion peptide, with the genes of two subunits (p35 and p40) linked by bovine elastin motifs, and T-mIL12-IRES co-expresses the subunits, with the two genes separated by an internal ribosome entry site (IRES) sequence.

Results: T-mfIL12 is significantly more efficient in producing IL-12 than T-mIL12-IRES in all cell lines tested, whereas the expression methods do not affect the replication capabilities and cytopathic effects. In two syngeneic mouse subcutaneous tumor models of Neuro2a and TRAMP-C2, T-mfIL12 exhibits a significantly higher efficacy than T-mIL12-IRES when inoculated intratumorally. Furthermore, T-mfIL12 shows a significantly higher intratumoral expression of functional IL-12, causing stronger stimulation of specific antitumor immune responses than T-mIL12-IRES.

Conclusions: The results implicate that a fusion-type expression of IL-12 is a method superior to co-expression of separate subunits, due to higher production of functional IL-12 molecules. This study led to the creation of triple-mutated oncolytic HSV-1 armed with human IL-12 currently used in phase 1/2 trial for malignant melanoma.

Fig. 1. The structures of T-mfIL12 and T-mIL12-IRES.

The boxes (top line) represent inverted repeat sequences flanking the long (U_L) and short (U_S) unique sequences of HSV-1 DNA. Oncolytic HSV-1 uses G47∆ as the backbone and contains 1.0 kb deletions in both copies of the γ34.5 gene, a 0.3 kb deletion in the α47 gene, and a 0.9 kb deletion in the ICP6 gene. The lacZ gene and the cytomegalovirus (CMV) promoter-driven interleukin-12 (IL-12) gene are oriented in opposite directions, and inserted into the deleted ICP6 locus. T-mfIL12 contains a cassette carrying the p35 and p40 genes of murine IL-12 linked by two bovine elastin motifs. Murine IL-12 is expressed as a single fusion peptide and folds to become active. T-mIL12-IRES contains a cassette in which the p40 and p35 genes of murine IL-12 are separated by an internal ribosome entry site (IRES) sequence from the encephalomyocarditis virus (EMCV). The p40 and p35 subunits are efficiently co-expressed to form active murine IL-12. T-01 contains a cassette with no transgene.

Fig. 2. In vitro replication capabilities and murine IL-12 expressions of T-mfIL12 and T-mIL12-IRES.

a In vitro replication assay. Vero cells were infected with G47∆, T-01, T-mfIL12 candidate (4 clones) or T-mIL12-IRES candidate (4 clones) at a multiplicity of infection (MOI) of 0.01, the progeny virus was recovered 48 h after infection, and the number determined by plaque assay. T-mfIL12 and T-mfIL12-IRES showed no significant difference in replication capability (p = 0.736, t-test). b In vitro murine IL-12 expression. Vero cells were infected with G47∆, T-01, T-mfIL12 candidate (4 clones) or T-mIL12-IRES candidate (4 clones) at an MOI of 1, and the amount of murine interleukin-12 (IL-12) (p70) secreted was determined by enzyme-linked immune-sorbent assay (ELISA). T-mfIL12 expressed a significantly higher amount of p70 IL-12 than T-mIL12-IRES (p < 0.001, t-test). c The time course of viral yields. Vero cells were infected with G47∆, T-01, T-mfIL12 or T-mIL12-IRES in duplicate at an MOI of 0.01, the progeny virus was recovered 0 h, 6 h, 24 h and 48 h after infection, and titrated by plaque assay. The time course for the viral yields of T-mfIL12 was comparable to that of T-mIL12-IRES. d In vitro murine IL-12 expression in murine tumor cell lines. Neuro2a, Pr14-2 or TRAMP-C2 cells were infected with T-mfIL12 or T-mIL12-IRES at an MOI of 1, and the amount of murine IL-12 (p70) secreted was determined by ELISA. T-mfIL12 expressed a higher amount of p70 IL-12 than T-mIL12-IRES in all three cell lines. All assays were performed in duplicate. ***, p < 0.001; NS, not significant.

Fig. 3. In vitro cytopathic effects of T-mfIL12 and T-mIL12-IRES.

In all three murine cancer cell lines, both T-mfIL12 and T-mIL12-IRES showed cytopathic effects comparable to the control virus T-01 at both multiplicity of infection (MOI) = 0.1 (dotted lines) and MOI = 1 (solid lines). The number of surviving cells was counted daily and expressed as a percentage of mock-infected controls. The results are the means of triplicate wells.

Fig. 4. In vivo efficacy of T-mfIL12 and T-mIL12-IRES.

a, b and c Assessment by subcutaneous tumor growth inhibition. a Male C57BL/6 mice harboring unilateral subcutaneous TRAMP-C2 tumors (prostate cancer) were treated with intratumoral inoculations with T-01, T-mfIL12 or T-mIL12-IRES (5 × 10⁶ pfu) or mock on days 0 and 3 (n = 6). T-mfIL12 showed the highest efficacy. All three viruses were significantly more efficacious than mock (p < 0.001 for all) (Significance not indicated in the figure). b Female A/J mice harboring bilateral subcutaneous Neuro2a tumors (neuroblastoma) were treated with intratumoral inoculations with T-01, T-mfIL12 or T-mIL12-IRES (5 × 10⁴ pfu) or mock into the left tumors only on days 0 and 3 (n = 9-12). T-mfIL12 exhibited the highest efficacy in both treated and untreated sides. All three viruses were significantly more efficacious than mock in the treated side (p = 0.001 for T-01 and p < 0.001 for both T-mfIL12 and T-mIL12-IRES) and T-mfIL12 and T-mIL12-IRES than mock in the untreated side (p < 0.001 for both viruses) (Significance not indicated in the figure). c NOG mice harboring bilateral subcutaneous Neuro2a tumors were treated with intratumoral inoculations with T-01, T-mfIL12 or T-mIL12-IRES (1 × 10⁶ pfu) or mock into the left tumors only on days 0 and 3 (n = 10). T-01, T-mfIL12 and T-mIL12-IRES showed no difference in efficacy in the treated side. The effect of IL-12 expression was completely abolished for both T-mfIL12 and T-mIL12-IRES in both the treated side and the untreated side. Tumor volume = length × width × height × 0.52. Results represent the mean. Bars, standard error of the mean (SEM). *, p < 0.05; **, p < 0.01; ***, p < 0.001; Two-way ANOVA with Bonferroni’s multiple comparisons test. d and e Assessment by survival. Female A/J mice harboring bilateral subcutaneous Neuro2a tumors were treated with intratumoral inoculations with T-01, T-mfIL12 or T-mIL12-IRES, either 5 × 10⁵ pfu (d) or 5 × 10⁴ pfu (e), or mock into the left tumors on days 0 and 3. Animals were sacrificed when the maximum diameter of one of the bilateral tumors reached 22 mm. Both T-mfIL12 and T-mIL12-IRES significantly prolonged the survival of tumor-bearing mice compared with T-01 at both doses (log-rank test). One or two T-mfIL12-treated mice showed a long-term survival for both doses, but there was no significant difference between T-mfIL12 and T-mIL12-IRES.

Fig. 5. Immune responses by T-mfIL12 and T-mIL12-IRES.

a Immunohistochemistry. Bilateral subcutaneous Neuro2a tumors were generated in A/J mice, left tumors only were inoculated with T-mfIL12, T-mIL12-IRES, T-01 (2 × 10⁵ pfu) or mock on days 0 and 3, and the tumors were harvested on day 6 (n = 3 per group). An increased infiltration of CD4⁺ and CD8⁺ lymphocytes were observed in the tumor for all three viruses, both in the treated and the untreated side, most prominently with T-mfIL12 (Fig. 5a). HSV-1 positive cells were observed in the tumor with all viruses in the treated side, but not in the untreated side. HE, hematoxylin and eosin. Scale bars, 100 μm. b In vivo levels of interleukin-12 (IL-12) and Interferon γ (IFNγ). In the unilateral subcutaneous Neuro2a model, established tumors were inoculated with T-01, T-mfIL12, T-mIL12-IRES (2 × 10⁶ pfu) or mock, sera and tumor samples were collected on days 1, 3 and 6 (n = 3 per group), and the levels of mouse IL-12 and IFNγ were measured by ELISA. The intratumoral IL-12 levels for T-mfIL12 were significantly higher than those for T-mIL12-IRES at all time points (p = 0.018, p = 0.016 and p = 0.046 for days 1, 3 and 6, respectively). The levels of IL-12 detected from the serum were remarkably lower than those in the tumor. Correlating with the intratumoral IL-12, the serum IL-12 level for T-mfIL12 was higher than that for T-mIL12-IRES on day 1 (p = 0.047). The IFNγ levels of T-mfIL12 were significantly higher than those of T-mIL12-IRES both in the tumor and serum on day 1 (p = 0.014 and p = 0.027, tumor and serum, respectively). c Immune responses specific to Neuro2a cells. In the unilateral subcutaneous Neuro2a model, established tumors were inoculated with T-01, T-mfIL12, T-mIL12-IRES (5 × 10⁴ pfu) or mock on days 0 and 3, and the spleen was harvested on day 6. By ELISpot assay, splenocytes from T-mfIL12-treated mice showed a significantly higher number of IFNγ release stimulated by Neuro2a cells than those from T-01- and T-mIL12-IRES- treated ones (p = 0.005 and p = 0.004 vs T-01 and T-mIL12-IRES, respectively). No significant difference in number of IL-4 releasing splenocytes was observed among the three virus-treated groups. The IFNγ or IL-4 releases specific to Neuro2a cells were calculated by subtracting the numbers of SaI/N responding spots from those of Neuro2a responding spots. For b and c, Graphs show the means. Dots represent individual data. Bars, SD. *, p < 0.05; **, p < 0.01; NS, not significant; one-way ANOVA with Tukey’s multiple comparisons.

Fig. 6. Comparison of in vivo efficacy of T-mfIL12 and direct intratumoral injection with recombinant interleukin-12 (rIL-12).

In the bilateral subcutaneous Neuro2a model, rIL-12, T-01 (5 × 10⁴ pfu) without or with rIL-12, T-mfIL12 (5 × 10⁴ pfu) or mock was inoculated into the left tumors only on days 0 and 4 (n = 10 per group). Three different doses were used for rIL-12; 500 ng (a), 50 ng (b) and 1 ng (c). The dose 50 ng represents the intratumoal IL-12 level treated with T-mfIL12 at 2 × 10⁶ pfu and 1 ng represents that at 5 × 10⁴ pfu, the T-mfIL12 dose used in these experiments. a When the dose of 500 ng was used for rIL-12, rIL-12, T-01, T-01+rIL-12 and T-mfIL12 were all significantly more efficacious than mock in the treated side (p < 0.001 vs mock for all). In the untreated side, rIL-12 alone showed no significant antitumor effect compared with mock, whereas T-01+rIL-12 and T-mfIL12 showed a significantly higher efficacy than mock (p = 0.039 and p < 0.001 vs mock, respectively). Further, in the untreated side, T-mfIL12, but not T-01+rIL-12, was significantly more efficacious than rIL-12 alone (p = 0.003). b Results similar to a were obtained when the dose of 50 ng was used for rIL-12. c When the dose of 1 ng was used for rIL-12, rIL-12 alone showed no significant efficacy in both treated and untreated sides, and T-mfIL12, but not T-01+rIL-12, was significantly more efficacious than rIL-12 alone in the treated side. Tumor volume = length × width × height × 0.52. Results represent the mean. Bars, standard error of the mean (SEM). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant; Two-way ANOVA with Bonferroni’s multiple comparisons test.