Seneca Valley Virus 3Cpro Substrate Optimization Yields Efficient Substrates for Use in Peptide-Prodrug Therapy

Abstract

The oncolytic picornavirus Seneca Valley Virus (SVV-001) demonstrates anti-tumor activity in models of small cell lung cancer (SCLC), but may ultimately need to be combined with cytotoxic therapies to improve responses observed in patients. Combining SVV-001 virotherapy with a peptide prodrug activated by the viral protease 3Cpro is a novel strategy that may increase the therapeutic potential of SVV-001. Using recombinant SVV-001 3Cpro, we measured cleavage kinetics of predicted SVV-001 3Cpro substrates. An efficient substrate, L/VP4 (kcat/KM = 1932 ± 183 M(-1)s(-1)), was further optimized by a P2' N→P substitution yielding L/VP4.1 (kcat/KM = 17446 ± 2203 M(-1)s(-1)). We also determined essential substrate amino acids by sequential N-terminal deletion and substitution of amino acids found in other picornavirus genera. A peptide corresponding to the L/VP4.1 substrate was selectively cleaved by SVV-001 3Cpro in vitro and was stable in human plasma. These data define an optimized peptide substrate for SVV-001 3Cpro, with direct implications for anti-cancer therapeutic development.

Fig 1. Conceptual schematic of use of an SVV 3C^pro activated peptide prodrug in combination with SVV virotherapy as a novel form of VDEPT.

SVV infects a fraction of tumor cells (1), producing 3C^pro during the viral life cycle. Upon cell lysis, new SVV virions and 3C^pro are released into nearby tissue (2). Administered peptide prodrug would be excluded from cells by the presence of the attached peptide (3), sparing normal tissues, which are non-permissive and therefore cannot express 3C^pro. The 3C^pro present at high concentration exclusively within the tumor microenvironment cleaves this peptide sequence (4), allowing the cytotoxic moiety to enter both infected and adjacent uninfected cells within the tumor, resulting in a powerful local bystander effect (5).

Fig 2. SVV-001 polyprotein and identification of efficient 3C protease substrates A.

The SVV-001 polyprotein is hypothesized to include twelve mature proteins based on sequence alignment with closely related viruses. Protein regions (drawn to scale) are presented with proposed 3C^pro cleavage sites (arrowheads) depicted. The 2A ribosome skipping sequence (RS; diamondhead arrow) is also shown. B. Sequence alignment of proposed 3C^pro cleavage sites. Sites are named based on the two mature proteins flanking these sites. The presumed scissile bond is depicted with an arrowhead with the consensus Q↓GP cleavage sequence amino acids highlighted. C. Schematic of FRET substrate construction and kinetic assays. Substrates are cloned between two fluorescent proteins, CyPET and YPET. These molecules exhibit FRET when in close proximity; therefore, there will be a high level of emission at 527 nm (YPET) and lower emission at 475 nm (CyPET). As the 3C protease cleaves and releases YPET from proximity to CyPET, the amount of FRET will decrease observed as an increase in CyPET emission (475 nm) and a decrease in YPET emission (527 nm). D. Conversion of FRET substrates by purified HRV or SVV-001 3C^pro over time. Data points represent the average of three replicates at each time point. The HRV substrate was used as a positive control. L/VP4 and 2B/2C are endogenous substrates. L/VP4.1 (P2’ N→P substitution) is a further optimized version of the endogenous L/VP4 substrate. Lines of the same color correspond to curve fits from GraphPad. Error bars on data points were removed for figure clarity.

Fig 3. Optimization of the L/VP4 Substrate A.

Alignment of SVV L/VP4 cleavage site (shown in italics) with L/VP4 junction sequences of closely related cardioviruses. The amino acids in each sequence that diverge from the consensus sequence (shown in bold) are highlighted. B. Conversion of modified L/VP4.1 FRET substrates by purified SVV-001 3C^pro over time. Data points represent the average of three replicates at each time point. The endogenous L/VP4 substrate and optimized L/VP4.1 substrate are shown for reference. L/VP4.3 and L/VP4.4 are P4 Y→F substitution and P4 Y→M substitution substrates, respectively. Lines of the same color correspond to curve fits from GraphPad. Error bars on data points were removed for figure clarity. C. Conversion of truncated L/VP4.1 FRET substrates by purified SVV-001 3C^pro over time. Data points represent the average of three replicates at each time point. The endogenous L/VP4 substrate and optimized L/VP4.1 substrate are shown for comparison. L/VP4.5 and L/VP4.6 are P6 and P5/P6 truncations of L/VP4.1, respectively. Lines of the same color correspond to curve fits from GraphPad. Error bars on data points were removed for figure clarity.

Fig 4. SVV-001 3C^pro substrates are cleaved in the context of a cellular infection.

A. Conversion of FRET substrates by SVV-001 3C^pro produced by a cellular SVV infection of permissive SCLC line, NCI-H446. Data points represent the average of three replicates at each time point. The L/VP4.1 FP substrate incubated with uninfected and infected cells, using similar incubations of NHL FP substrate as a negative control. Lines of the same color correspond to curve fits from GraphPad. Error bars on data points were removed for figure clarity. B. Initial reaction rates of L/VP4.1 peptide cleavage by recombinant SVV-001 3C^pro. Data points represent the initial rate of reaction at each concentration of L/VP4.1 peptide calculated from three replicate experiments. Data points were fit to a Michaelis-Menten nonlinear regression from GraphPad and the kinetic constants determined by the curve fit were reported. Standard deviation values of the kinetics constants were calculated by the GraphPad software and propagated through second order rate constant calculations. C. Proteolysis of CPQ2/5-FAM peptides by native SVV-001 3C^pro in a cellular assay with NCI-H446. Data points represent the average relative fluorescence units (RFUs) increase of three replicates at each time point relative to fluorescence at time zero. The L/VP4.1 FQ peptide was incubated with uninfected and infected cells, using similar incubations of NHL FQ peptide as a negative control. Lines of the same color correspond to connecting line between points.

Fig 5. Stability of L/VP4.1 fluorescently quenched peptide in human plasma.

Data points represent the average relative fluorescence units (RFUs) increase of four replicates at each time point relative to fluorescence at time zero for L/VP4.1 FQ peptide or NHL FQ peptide incubated with human plasma. The peptides were also incubated with human plasma supplemented with SVV 3C^pro or SVV C160A 3C^pro as positive and negative controls, respectively. Uncertainty is expressed by standard deviation.