Single-cell analyses of viral infections reveal heterogeneity that is not detected by traditional population-level studies. This study applies drop-based microfluidics to investigate the dynamics of HSV-1 infection of neurons at the single-cell level. We used micron-scale Matrigel beads, termed microgels, to culture individual murine Superior Cervical ganglia (SCG) neurons or epithelial cells. Microgel-cultured cells are subsequently encapsulated in individual media-in-oil droplets with a dual fluorescent-reporter HSV-1, enabling real-time observation of viral gene expression and replication. Infection within drops revealed that the kinetics of initial viral gene expression and replication were dependent on the inoculating dose. Notably, increasing inoculating doses led to earlier onset of viral gene expression and more frequent productive viral replication. These observations provide crucial insights into the complexity of HSV-1 infection in neurons and emphasize the importance of studying single-cell outcomes of viral infection. The innovative techniques presented here for cell culture and infection in drops provide a foundation for future virology and neurobiology investigations.

The microfluidic device containing hydrogel microgels with infected mature neurons were placed in a microscope stage top incubation chamber at 37°C. Images in Phase Contrast/YFP/RFP were taken every 15 min for 16 h. 

Percentage of YFP Positive Cells. To quantify the percentage of YFP positive cells, end-point images were analyzed using Fiji ImageJ. Cells were found in brightfield. The background intensity of the YFP channel was subtracted from the image, and the percentage of YFP positive cells was quantified. Experiments were repeated in triplicate. Cell counts for Vero cells embedded in-microgels ranged from 79 -280 cells/replicate, for Vero cells grown on-microgels from 165 - 390 cells/replicate, for Vero cells in-suspension from 228 - 456 cells/replicate, and for SCG neurons from 84 - 209 cells/replicate. Multiple comparison tests and one-way ANOVAs were performed in Matlab.

The Timing of YFP and RFP Detection. To quantify the timing of YFP and RFP detection, time-lapse images were analyzed using Fiji ImageJ. Cells were difficult to identify in brightfield, so cells were identified by finding YFP positive cells at 16 hpi. Once located, a 30 μm diameter circular region of interest (ROI) was drawn around each individual cell. The maximum pixel intensity in YFP and RFP for the ROI was then measured for each time-frame. The frame 1 maximum pixel intensity of each ROI was subtracted from the subsequent frames for that ROI. The noise threshold was found to be 50 arbitrary units (a.u.) for YFP and 75 a.u. for RFP. The timing of YFP and RFP detection were defined as the first time-frame the YFP and RFP pixel intensities in each ROI were greater than the noise threshold for two consecutive frames. For SCG time-lapse studies, experiments were repeated 3-4 times with a total of 174 cells analyzed. For Vero time-lapse studies, experiments were repeated 3 times with a total of 669 cells analyzed. One-way ANOVAs and linear regressions were performed in Matlab.