Harnessing endogenous miRNAs to control virus tissue tropism as a strategy for developing attenuated virus vaccines

Abstract

Live attenuated vaccines remain the safest, most cost-effective intervention against viral infections. Because live vaccine strains are generated empirically and the basis for attenuation is usually ill defined, many important viruses lack an efficient live vaccine. Here, we present a general strategy for the rational design of safe and effective live vaccines that harnesses the microRNA-based gene-silencing machinery to control viral replication. Using poliovirus as a model, we demonstrate that insertion of small miRNA homology sequences into a viral genome can restrict its tissue tropism, thereby preventing pathogenicity and yielding an attenuated viral strain. Poliovirus strains engineered to become targets of neuronal-specific miRNAs lost their ability to replicate in the central nervous system, leading to significant attenuation of neurovirulence in infected animals. Importantly, these viruses retained the ability to replicate in nonneuronal tissues. As a result, these engineered miRNA-regulated viruses elicited strong protective immunity in mice without producing disease.

Figure 1. Engineering viruses with restricted tissue tropism

(A) Target sequences complementary to two distinct miRNAs (let-7a and miR124a) were inserted into two locations within the poliovirus genome. The 5’ site is located in the “variable” segment between the poliovirus internal ribosome entry site (IRES) and the start codon. The 3’ site is located between the structural and nonstructural genes. (B) Perfect sequence complementarity between the endogenous miRNA and the target sequence inserted into the poliovirus genome is illustrated. Silence mutations were engineered into the target sequences to create virus controls that disrupt base pairing with the endogenous miRNA, while conserving the wild type encoded amino acid sequence. (C) Northern Blot analysis of let-7a and miR-124a miRNA expression levels observed in cell lines (HeLa and NTERA-2, lanes 1 and 2) and mouse tissues (lanes 3–7).

Figure 2. Engineered PV-L7 virus exhibits attenuated replication kinetics due to endogenous let-7a-mediated repression

(A) Replication kinetics of wild type poliovirus (squares) and engineered viruses, PV-L7 (circles) and control PV-L7M (triangles), in the permissive (NTERA-2, white symbols) and non-permissive (HeLa, filled symbols) cell lines by virtue of endogenous let-7a expression. PV-L7 viral replication is attenuated in HeLa cells by six orders of magnitude, whereas the PV-L7M control virus replicates with wild type kinetics. Viral titer values represent the mean ± SD of three independent experiments. Error bars, SD. (B) RT-PCR analysis of total RNA extracted from HeLa cells infected with wildtype, PV-L7, or PV-L7M virus at time zero (lanes 4–6) and six hours post infection (lanes 7–9). We amplified by RT-PCR a 269nt-long fragment (amplicon), spanning the 5’ target insertion site, from 200ng total RNA. The molecular marker is shown in lane 1. Lane 2 controls for the reverse transcription reaction, and lane 3 corresponds to RNA isolated from uninfected HeLa cells used as template for the RT-PCR reaction. (C) Inhibition of endogenous let-7a using anti-mir technology rescues PV-L7 viral replication in the non-permissive HeLa cells. HeLa cells were transfected with small oligonucleotides corresponding to the let-7a complementary sequence (anti-mir let-7a, black symbols) or control nonspecific oligonucleotide (ctrl anti-mir, white symbols). HeLa cells were then infected with wild type poliovirus (squares) or engineered PV-L7 virus (circles). Viral titer values represent the mean ± SD of three independent experiments. Error bars, SD. (D) Knocking-down core components of the RNAi machinery rescues PV-L7 replication. siRNAs targeting Ago2 (circles) and drosha (triangles) partially rescues replication of PV-L7 (white symbols) in the non-permissive HeLa cells, while a siRNA targeting GAPDH has no effect (squares). Transfecting siRNAs has not effect on wild type poliovirus replication (WT, black symbols). Viral titer values represent the mean ± SD of three independent experiments. Error bars, SD.

Figure 3. Engineered viruses, PV-L7 and PV-124, show attenuated neuropathogenicity in transgenic mice

Percentage of (A) cPVR mice and (B, C) IFNAR mice surviving intramuscular injections of different doses (10⁸ - 10³ PFU); n=20 mice per group. (A) cPVR mice infected with 10⁸ PFU PV-L7 (filled squares) and PV-124 (filled squares) viruses are unaffected, whereas mice infected with control viruses, PV-L7M (filled triangles) and PV-124M (filled circles), and wild type (white squares) poliovirus do not survive past day six. B) IFNAR mice infected with 10⁷ PFU of PV-L7 (filled squares) virus all survive, whereas most mice infected with 10³ PFU wild type poliovirus (white squares) and the PV-L7M control virus (filled circles) were paralyzed by day 11. C) PV-124 virus is attenuated in IFNAR mice, albeit less dramatically than PV-L7 infected IFNAR mice. The white and black squares indicate wild type and PV-124, respectively.

Figure 4. Tissue tropism is controlled by the miRNA machinery

A set of five cPVR mice was infected intravenously with each engineered virus. Virus isolated from the brain (A), spinal cord (B), and spleen (C) was analyzed by a standard plaque assay six days post infection. Viral titers (PFU/ml), normalized by tissue mass, observed in the three tissues are plotted for five mice infected with each virus. Black squares designate wild type, white squares PV-L7, black circles PV-L7M, white circles PV-124, and black triangles PV-124M. In the brain and spinal cord, PV-L7 and PV-124 infected mice exhibit viral titers that are four to six orders of magnitude lower than viral titers observed in mice infected with wild type poliovirus and the control viruses, PV-L7M and PV-124M. PV-L7 viral replication is measurable in the spleen, but is still an order of magnitude lower than viral titers observed in mice infected with the other viruses. In the spleen, wild type levels are observed for PV-124 virus.

Figure 5. Infection with PV-L7 or PV-124 virus induces high levels of neutralizing antibody and confers immunity against a lethal challenge with wild type poliovirus

(A) Neutralizing antibody titers (reciprocal of the serum dilution able to neutralize 100 TCID₅₀ of wild-type poliovirus) were determined for serum collected from immunized mice. Data for individual mice (black squares specify UV-inactivated wild type virus, white squares PV-L7, black circles PV-124, white circles G64S, and black triangles for PBS) and the group average (bar) are shown. Titers for PBS-immunized mice were below the detection level of the assay, indicated by the dashed line. (B) cPVR and IFNAR mice were immunized with either phosphate buffer saline (PBS) or 1×10⁷ PFU (0.1LD₅₀) of wild type irradiated virus, PV-L7 virus, or PV-124 virus. Four weeks after immunization, serum was collected. Mice were subsequently challenged with either 10LD₅₀ or 10⁴LD₅₀ of wild-type poliovirus via intramuscular injection. The numbers of mice surviving are indicated. PV-124 immunized cPVR mice are protected against a lethal challenge of wild-type poliovirus, whereas cPVR mice immunized with PV-L7 and UV-inactivated wild-type virus are much less protected. PV-L7 immunized IFNAR mice were protected against a lethal challenge with 10LD₅₀ of wild-type poliovirus, and seventy percent of the immunized IFNAR mice even survived a challenge dose of 10⁴LD₅₀.