Detection of Viruses Using Biophysical Properties of Infected Cells for Biologics and Vaccines Testing

Virus detection assays are critical in support of the production of biologics and vaccines to ensure safe, effective drugs. Current methods capable of broad range detection of virus such as the in vitro virus assay are slow requiring weeks to complete and therefore typically done at the end of biologic production. Here, we present new approaches for broad detection of viruses in biologic production using label-free, biophysical cell properties. These properties change with viral infection enabling a simple, in process test for virus and quantitative detection of virally infected cells. 

Laser force cytology (LFC) probes infected cells for optical properties using measurements of optical force imposed by a laser as well as imaging capabilities. We demonstrate the determination of an important set of cellular features from LFC that can distinguish a wide variety of viruses using principal component analysis. Using velocity, size, and shape factor of the cells, the sensitivity of LFC was determined to be at least equal and more sensitive with some viruses than the in vitro assay. Further, we demonstrated the ability of LFC to monitor live viral vaccine potency during bioreactor production. The titer of a measles virus vector propagating over time in a bioreactor was shown to correlate well with the percentage of microcarrier and supernatant cells with a high optical force index. 

Mechanical property change of cells with infection was investigated using micropipette aspiration. We observed cells infected with reovirus 3 had nearly doubled increased stiffness as measured by elastic modulus by 3 days post infection. The stiffness of the infected cells did not continue to increase at later time points post infection as infection progressed. Alterations to internal actin filaments as well as actin protrusions at the cell periphery were observed with infection and likely the cause of increased stiffness. These protrusions are thought to be related to viral processing and escape as disruption of actin filaments decreased the amount of virus released into the supernatant. 

A high-throughput microfluidic device using constrictions to deform cells was developed. Using the time required by cells to transit the constrictions, elastic modulus and cell fluidity mechanical parameters were determined using power law rheology. Similar results to micropipette aspiration with an increased stiffness of reovirus 3 infected cells were measured with a microfluidic constriction device. With a device capable of measuring cell mechanics in a high-throughput manner, how other various viruses affect the mechanical property of cells can be studied. 

Finally, histograms of LFC descriptors of measles infected samples were decomposed into underlying infected and uninfected distributions using the fractions of infected and uninfected cells from a kinetic model for measles propagation in bioreactors. The decomposed infected cell distribution remained consistent 72 hours post infection and beyond but failed at earlier time points due to lack of infected cell sampling. Probability analysis for the expected number of infected cells based on a sample size further confirmed the under sampling of the infected population at early time points with only 300 cells. It appears likely a decomposition analysis would increase detection sensitivity down to 48 hours post infection with an increased sample of at least 2500 cells.