Poliomyelitis is a highly infectious disease caused by the poliovirus and transmitted through person-to-person contact (fecal-oral route). Most people experience asymptomatic infections and are “silent” spreaders. Others present symptoms including fever, fatigue, headache, etc. Approximately one in 200 infections leads to permanent paralysis.
According to the Global Polio Eradication Initiative, the global incidence of poliomyelitis has decreased by 99.9 percent—from 350,000 estimated cases in more than 125 endemic countries to 29 reported cases in 2018—since the partnership’s foundation in 1988. Only one of the three strains of wild poliovirus is still reported today, circulating in Afghanistan and Pakistan. Vaccine-derived poliovirus has also emerged in areas of low vaccination; the most recent cases concern the Democratic Republic of the Congo, Nigeria, and Somalia.
Failure to eradicate polio from these last remaining areas could result in as many as 200,000 new cases every year within 10 years worldwide. In an attempt to help eliminate polio, I joined the IBM Almaden Research Center in 2017 to work on a Defense Advanced Research Projects Agency (DARPA)-funded project in collaboration with the University of California, San Francisco; Stanford University; and the University of Haifa.
Our research focuses on the creation of a set of mathematical models to direct the design of defective interfering particles (DIPs) as therapeutic antiviral agents for poliovirus. DIPs are deletion mutants of a wild-type (WT) virus that lack genes necessary to complete their reproduction cycle. Their survival depends on the presence of the WT virus, from which they can steal their lacking elements when co-infecting the same cell. By doing so, DIPs act as parasites of the WT virus by hindering growth and potentially preventing further spread of the disease. When DIPs succeed in helping an infected organism clear a WT infection, they eventually die out because of their inability to carry out their reproduction cycle alone. Naturally-occurring DIPs appear in a number of viral species, including dengue , Ebola , Influenza A , and very occasionally poliovirus [4-5]. However, their role in the evolution and control of viral infections is currently unknown and thus a hot topic of investigation. Natural poliovirus DIPs feature an in-frame deletion in the region that encodes capsid proteins . This constitutes an essential envelope of protection for the viral genome to survive outside of the cells and be able to enter and infect new cells. The starting point for our investigation is a synthetic defective construct with such an in-frame deletion, so that the DIP must rely on the WT virus for capsid proteins (see Figure 1).
Figure 1. Schematic of the within-cell reproduction cycle of wild-type (WT) poliovirus (blue) and defective interfering particles (DIPs) (red). DIPs have been engineered with two features: faster replication of its genome (as it is shorter), and stealing of capsid proteins produced by the WT virus for its own encapsidation to produce new virions. DIPs engineered in this way can decrease the WT viral load due to this competition process.
Our goal is to optimize the design of a therapeutic DIP against poliovirus infection. For this purpose, we are building experimentally-guided mathematical models to understand and predict the factors shaping the competition between DIPs and WT poliovirus. Predictions ultimately provide guidelines to improve the construction of efficient therapeutic DIPs. We adopt a bottom-up multiscale modeling approach, starting from the within-cell level dynamic before moving to cell-to-cell, within-host, and finally host population level (see Figure 2).
Rousseau and her collaborators employ a bottom-up multiscale modeling approach to describe the competition between wild-type poliovirus and defective interfering particles, starting from the within-cell level then moving to the cell-to-cell, within-host, and finally host population level. This allows them to infer the antiviral design principles, which they then use to engineer the defective particles in the lab. Poliovirus image courtesy of Manuel Almagro Rivas
and cell image courtesy of Servier Medical Art
We created a mechanistic model at the within-cell level describing the replication and encapsidation processes — the process of enveloping the viral genome in a capsid protein of the WT and DIP genomes . We parameterized the model with experimental competition data to understand the DIP’s main interference mechanisms. Interestingly, DIPs interfere with WT viruses by competing for limiting resources necessary for genome replication and capsid proteins produced by the WT; they thus act as a parasite of the parent virus. To assess the biological validity of our modeling strategy and results, we tested the model’s predictive power by generating simulations corresponding to easily-testable initial experimental conditions. We found that our model does indeed describe the essential features of the competition process. Our project’s objective is to establish design principles for the antiviral particles. To this end, we relied on an extensive sensitivity analysis of model parameters on an output quantity of interest, namely the proportion of WT virions produced per cell. This quantity is biologically important because it is a proxy of the relative contribution of WT and DIP to the viral load from co-infected cells. Interestingly, we found that the DIP-relative replication rate—which is in part an indication of the difference in genome length between the two species—and encapsidation rate—which represents the faster encapsidation of DIPs compared to WT—are the most relevant parameters. This set of observations is paramount in driving the design of an optimal antiviral, a task currently undertaken by our experimental partners.
Next we built a cell-to-cell competition model. The input for this level of investigation is the lower level within-cell model for WT and DIP production per cell. Computational exploration of the parameter space revealed that the most sensitive parameters for the proportion of WT virions in the viral load (i.e., virions circulating outside of cells) come from the within-cell level: the relative WT and DIP production per cell, which are proxies of the WT and DIP virions produced in co-infected cells. Interestingly, the proportion of WT virions in the viral load is minimized over 10 weeks of simulation when WT- and DIP-relative production per cell is maximized. This rather counterintuitive result emerges because increasing WT- and DIP-relative production per cell increases WT and DIP viral loads, which in turn increases the probability of a cell’s co-infection by WT and DIP, a necessary condition for DIP to interfere with the WT. Looking back at the within-cell model predictions, we could infer that one could achieve such maximization of WT- and DIP-relative production per cell by increasing the relative DIP encapsidation rate rather than its replication rate. These major results provide guidelines for the optimization of our DIP’s design.
The intra-host competition model adds a layer of complexity by having several organs—which represent different environments that impact WT- and DIP-specific characteristics—that are connected to one another and in which WT and DIP can spread. The goal is to identify the conditions required to prevent WT virus’s spread into a given organ (like the brain, in the case of poliovirus). This allows us to study biological conditions that are more closely related to the antiviral’s therapeutic potential, such as the effect of difference in tissues, doses, times of infection and treatment, etc. Finally, the host-to-host epidemiological model features transmissible DIPs. The goal is to study the impact of these transmissible therapeutic DIPs on WT poliovirus’s global epidemiology. While researchers have explored the spread of preventative vaccines in a limited number of papers [8-10], no one seems to have conducted any work on the spread of therapeutic agents. This original analysis will thus be the first of its kind and provide important guidelines for the control of epidemics with therapeutic DIPs.
Our bottom-up approach and the guidance of competition experiments allowed us to build mathematical models with a solid biological background. Our method should be applicable to other viruses, as one may engineer therapeutic DIPs for any virus, with case-specific interference mechanisms that can be fine-tuned. For example, unlike poliovirus DIPs, other naturally-occurring DIPs (such as in Influenza A ) are replication-defective and require the helper WT virus for this step of the reproduction cycle.
Ultimately, DIP use as a therapeutic antiviral seems like a promising tool. If a DIP cannot fully suppress a WT infection, it can at least reduce the WT viral load to relieve a patient before the immune system gets triggered and becomes operational to clear the infection. The potential for transmission of DIPs between hosts adds another layer of efficiency for the control of WT virus epidemics.
The author presented this research during a minisyposium talk at the 2018 SIAM Conference on the Life Sciences, which took place in Minneapolis, Minn. last August.
Acknowledgments: The author would like to thank Raul Andino’s lab for a successful collaboration and for monitoring essential experiments. This material is based upon work supported by DARPA’s INTERfering and Co-Evolving Prevention and Therapy (INTERCEPT) Program, under contract number HR00-17-2-0027. Any opinions, findings, and conclusions, or recommendations expressed in this material are those of the author and do not necessarily reflect the views of DARPA.
 Li, D., Lott, W.B., Lowry, K., Jones, A., Thu, H.M., & Aaskov, J. (2011). Defective interfering viral particles in acute dengue infections. PloS One, 6(4), e19447.
 Calain, P., Monroe, M.C., & Nichol, S.T. (1999). Ebola virus defective interfering particles and persistent infection. Virol., 262(1), 114-128.
 Saira, K., Lin, X., DePasse, J.V., Halpin, R., Twaddle, A., Stockwell, T.,..., Ghedin, E. (2013). Sequence analysis of in vivo defective-interfering (DI)-like RNA of influenza A H1N1 pandemic virus. J. Virol., 87(4), 8064-74.
 Cole, C.N., Smoler, D., Wimmer, E., & Baltimore, D. (1971). Defective interfering particles of poliovirus I. Isolation and physical properties. J. Virol., 7(4), 478-485.
 Wimmer, E., Hellen, C.U., & Cao, X. (1993). Genetics of poliovirus. Ann. Rev. Gen., 27(1), 353-436.
 Lundquist, R.E., Sullivan, M., & Maizel Jr, J.V. (1979). Characterization of a new isolate of poliovirus defective interfering particles. Cell, 18(3), 759-769.
 Shirogane, Y., Rousseau, E., Voznica, J., Rouzine, I., Bianco, S., & Andino, R. (2019). Experimental and mathematical insights on the competition between poliovirus and a defective interfering genome. BioRxiv, 519751. https://doi.org/10.1101/519751
 Fine, P.E., & Carneiro, I.A. (1999). Transmissibility and persistence of oral polio vaccine viruses: implications for the global poliomyelitis eradication initiative. Am. J. Epidem., 150(10), 1001-1021.
 Nuismer, S.L., Althouse, B.M., May, R., Bull, J.J., Stromberg, S.P., & Antia, R. (2016). Eradicating infectious disease using weakly transmissible vaccines. Proc. Roy. Soc. B: Bio. Sci., 283(1841), 20161903.
 Bull, J.J., Smithson, M.W., & Nuismer, S.L. (2018). Transmissible viral vaccines. Trends Microbio., 26(1), 6-15.
 Magnus, P.V. (1947). Studies on interference in experimental influenza. Arkiv För Kemi, Mineralogi, Och Geologi, 24(7), 1-6.
|| Elsa Rousseau earned her Ph.D. in computational biology at the National Institute of Research in Informatics and Automatics in 2016. Her research interests are in the evolution and epidemiology of pathogens. She has been a postdoctoral researcher at IBM Almaden Research Center since 2017, where she works on the design of defective interfering particles as therapeutic agents against disease-causing viruses.