Cytomegalovirus (CMV) has been shown to induce large populations of CD8 T-effector memory cells that unlike central memory persist in large quantities following infection, a phenomenon commonly termed “memory inflation”. Although murine models to date have shown very large and persistent CMV-specific T-cell expansions following infection, there is considerable variability in CMV-specific T-memory responses in humans. Historically such memory inflation in humans has been assumed a consequence of reactivation events during the life of the host. Because basic information about CMV infection/re-infection and reactivation in immune competent humans is not available, we used a murine model to test how primary infection, reinfection, and reactivation stimuli influence memory inflation. We show that low titer infections induce “partial” memory inflation of both mCMV specific CD8 T-cells and antibody. We show further that reinfection with different strains can boost partial memory inflation. Finally, we show preliminary results suggesting that a single strong reactivation stimulus does not stimulate memory inflation. Altogether, our results suggest that while high titer primary infections can induce memory inflation, reinfections during the life of a host may be more important than previously appreciated.

It is clear that latent CMV can reactivate in immunocompetent individuals, but the mechanism triggering such reactivations remains unclear. Recent clinical data suggest that reactivation can be subverted by CMV‐specific T‐memory. We therefore monitored CMV‐specific T cells in immunocompetent mice with latent mCMV after a known reactivation trigger (LPS). LPS induced transient systemic contraction of mCMV‐specific CD8 memory that was followed by transcriptional reactivation. Subsequent recovery of mCMV‐specific T cells coincided with resumption of latency. These data suggest that bacterial antigen encounters can induce transient T‐memory contraction, allowing viral recrudescence in hosts latently infected with herpes family viruses.

Abstract Both mice and humans can develop large “inflated” populations of CD8 effector memory T-cells after cytomegalovirus (CMV) infection, and such memory inflation has been associated with immune senescence. Not all humans develop CD8 memory inflation after CMV infection, so we hypothesized that memory inflation might be predicated by the initial viral burden. Because it is impossible to know an individuals CMV inoculum during natural infection, we infected mice with murine CMV (MCMV) titers ranging from 102 to 106 PFU. After infection, peripheral blood was evaluated monthly for MCMV-specific CD8 T-cells and antibody. Using MCMV-specific MHC class I tetramers we confirm that high titer infections induce inflationary T-memory responses. In contrast, low titer infections induce significantly less inflation over time. Serum ELISA for MCMV-antibody responses showed a similar pattern. DNA from lung tissues evaluated 16 weeks after infection confirmed successful infection in all mice. Quantitative PCR showed that latent viral load in lungs correlated well with eventual lung resident T-memory responses (R2=0.96). We therefore conclude that an individual’s initial infectious burden determines their latent viral load and eventual development of memory inflation after cytomegalovirus infection, possibly explaining the wide variability in T-cell responses to CMV observed in humans.