Maintenance of large numbers of virus genomes in human cytomegalovirus-infected T98G glioblastoma cells.

After infection, human cytomegalovirus (HCMV) persists for life. Primary infections and reactivation of latent virus can both result in congenital infection, a leading cause of central nervous system birth defects. We previously reported long-term HCMV infection in the T98G glioblastoma cell line (1). HCMV infection has been further characterized in T98Gs, emphasizing the presence of HCMV DNA over an extended time frame. T98Gs were infected with either HCMV Towne or AD169-IE2-enhanced green fluorescent protein (eGFP) strains. Towne infections yielded mixed IE1 antigen-positive and -negative (Ag(+)/Ag(-)) populations. AD169-IE2-eGFP infections also yielded mixed populations, which were sorted to obtain an IE2(-) (Ag(-)) population. Viral gene expression over the course of infection was determined by immunofluorescent analysis (IFA) and reverse transcription-PCR (RT-PCR). The presence of HCMV genomes was determined by PCR, nested PCR (n-PCR), and fluorescence in situ hybridization (FISH). Compared to the HCMV latency model, THP-1, Towne-infected T98Gs expressed IE1 and latency-associated transcripts for longer periods, contained many more HCMV genomes during early passages, and carried genomes for a greatly extended period of passaging. Large numbers of HCMV genomes were also found in purified Ag(-) AD169-infected cells for the first several passages. Interestingly, latency transcripts were observed from very early times in the Towne-infected cells, even when IE1 was expressed at low levels. Although AD169-infected Ag(-) cells expressed no detectable levels of either IE1 or latency transcripts, they also maintained large numbers of genomes within the cell nuclei for several passages. These results identify HCMV-infected T98Gs as an attractive new model in the study of the long-term maintenance of virus genomes in the context of neural cell types. Our previous work showed that T98G glioblastoma cells were semipermissive to HCMV infection; virus trafficked to the nucleus, and yet only a proportion of cells stained positive for viral antigens, thus allowing continual subculturing and passaging. The cells eventually transitioned to a state where viral genomes were maintained without viral antigen expression or virion production. Here we report that during long-term T98G infection, large numbers of genomes were maintained within all of the cells' nuclei for the first several passages (through passage 4 [P4]), even in the presence of continual cellular division. Surprisingly, genomes were maintained, albeit at a lower level, through day 41. This is decidedly longer than in any other latency model system that has been described to date. We believe that this system offers a useful model to aid in unraveling the cellular components involved in viral genome maintenance (and presumably replication) in cells carrying long-term latent genomes in a neural context.