(A-1) Computational modeling of replication efficiency and late protein synthesis variabilities in HCMV-infected ARPE19 cells

Introduction:: Introduction: Human cytomegalovirus (HCMV) is a highly prevalent betaherpesvirus leading to symptomatology ranging from mononucleosis to sepsis in immunocompromised individuals (PMID: 27501258). A major, yet often overlooked, consequence of HCMV infection and immunosuppression is CMV retinitis (PMID: 18052600). HCMV lytic replication is supported by many cell types with different kinetics and efficiencies (PMID: 27501258). It is thought that infection of the cell type comprising the retinal pigment epithelium is a mechanism of HCMV entry into the neural layer of the retina (PMID: 12706995). Thus, we sought to investigate the differences in late lytic replication cycle kinetics by applying the experimental and computational methodologies proposed in Monti et al. (PMID: 35994667) to adult retinal epithelial (ARPE19) cells infected with HCMV produced from MRC5 fibroblasts and ARPE19 cells.


Materials and Methods:: Materials and Methods: Viral stocks were generated using HCMV TB40/E-BAC4 containing IE2-T2A-eGFP and pp28-mCherry and propagated on MRC5 (TB40/Fib) or ARPE19 cells (TB40/Epi). We infected ARPE19 cells using two multiplicities of infection (MOIs, 0.1 and 0.5) and measured cell-associated viral genomes at 2 hpi (vDNA_in,0) and changes over time in viral DNA (vDNA) using qPCR, late protein levels (pUL44 and pp28) using immunoblot, and cell-free viral titers using infectious unit assays. An empiric model of vDNA synthesis was fit to the qPCR data, which was then used to drive a mechanistic model of late lytic replication and protein synthesis proposed in Monti et al. (PMID: 35994667). This model was parameterized using the pseudo-Monte Carlo method applied in Monti et al. for each viral source-target cell combination using immunoblot and titering data. The late protein synthesis kinetic data were corroborated post-hoc using an IncuCyte live imaging platform and confocal microscopy.


Results, Conclusions, and Discussions:: Results: By fitting the empirical vDNA model to the qPCR data, we found a consistently larger initial input vDNA (vDNA_in,0) in the TB40/Epi virus but a larger maximal vDNA replication (vDNA_rep,max­) in the TB40/Fib virus, similar to TB40/Fib infecting fibroblast cells. By fitting the mechanistic model of the late protein synthesis proposed in Monti et al. to the immunoblot data, we found a larger synthesis constant for the viral pp28 protein in TB40/Epi than in TB40/Fib. We corroborated this finding experimentally using live cell imaging. Interestingly, we found using TB40/Epi and confocal microscopy that increased pp28 correlated with larger cytoplasmic assembly compartments likely due to syncytia formation. TB40/Epi produced approximately 1-log higher viral titer compared to TB40/Fib. By comparing vDNA_in,0-matched simulations of infection between TB40/Epi and TB40/Fb, we showed that increased vDNA_in,0 for TB40/Epi compensated for a smaller vDNA_rep,max allowing for increased output titer, likely related to improved entry.



Discussion and Conclusion: We demonstrated that vDNA_rep,max, pp28 synthesis rate, and syncytia formation are more dependent upon viral source and less dependent on target cell type. In contrast, output titers are dependent upon vDNA_in,0 which is dependent upon a match between viral source and target cell type. Application of the framework devised in Monti et al. allowed for the analysis of infection of ARPE19 cells by HCMV derived from two different cellular sources. This work provides proof-of-concept for application of this framework to the investigation of other aspects of the late lytic replication cycle. Furthermore, this work highlights the need for a vDNA_in,0, rather than MOI, based method of experimental infection for use in computational studies and to accurately compare between experimental conditions. Future directions may include investigation of other cell types and expansion to include other parts of the lytic replication cycle.




Acknowledgements (Optional): : Funding: NIH R21-AI149039


References (Optional): :