Four Distinct Patterns of Memory CD8 T Cell Responses to Chronic Murine Cytomegalovirus Infection

Abstract

CMVs are β herpesviruses that establish lifelong latent infection of their hosts. Acute infection of C57BL/6 mice with murine CMV elicits a very broad CD8 T cell response, comprising at least 24 epitopes from 18 viral proteins. In contrast, we show here that the CD8 T cell response in chronically infected mice was dominated by only five epitopes. Altogether, four distinct CD8 T cell kinetic patterns were evident. Responses to some epitopes, including M45, which dominates the acute response, contracted sharply after day 7 and developed into stable long-term memory. The response to m139 underwent rapid expansion and contraction, followed by a phase of memory inflation, whereas the response to an M38 epitope did not display any contraction phase. Finally, responses against two epitopes encoded by the immediate early gene IE3 were readily detectable in chronically infected mice but near the limit of detection during acute infection. CD8 T cells specific for the noninflationary M45 epitope displayed a classic central memory phenotype, re-expressing the lymph node homing receptor CD62L and homeostatic cytokine receptors for IL-7 and IL-15, and produced low levels of IL-2. Responses to two inflationary epitopes, m139 and IE3, retained an effector memory surface phenotype (CD62Llow, IL-7Rα−, IL-15Rβ−) and were unable to produce IL-2. We suggest that immunological choices are superimposed on altered viral gene expression profiles to determine immunodominance during chronic murine CMV infection.

The (gradual) rise of memory inflation

“…The third feature observed was maintained effector functions such as cytokine release—in contrast to exhausted CD8 + T cells 8. There are certainly some differences in the level of cytokine production comparing inflationary populations and classical non‐inflationary memory cells in the same model, but over time there does not appear to be clear attrition of such functionality (and the lack of transcriptional and phenotypic markers of exhaustion is consistent with this) 3, 4, 16, 17. The findings were given some further relevance since they are quite parallel to those seen in human responses to HCMV—in other words very large responses identifiable by tetramer, with a phenotype described as “effector‐memory” (also CCR7, CD62L, CD28, CD27, CD127 low) and with maintained function.…”

“…At a similar time to this, Ann Hill's group performed some very detailed mapping experiments using vectors and peptide libraries to map the dominant CMV‐specific CD8 + T‐cell responses in the C57BL/6 mouse. Thus in different mouse strains, an “inflationary” and a classical response were seen to evolve in parallel against distinct peptide epitopes even within the same animal 3, 4…”

“…The first point is that it depends on the viral context, since epitopes which show inflation in the adenovirus or MCMV models do not induce inflation if presented in other recombinant settings such as vaccinia viruses—in other words it is not an inherent property of the peptide (for example related to some aspect of cross‐reactivity) 16, 22. Within a given model where inflation can occur, the answer does not appear to be related to initial immunodominance, since even very subdominant responses can inflate at later time‐points 2, 3, 4, 71. Similarly, it is not dependent on the affinity of the epitope itself for MHC, since responses to the same epitope can be converted from a non‐inflationary to an inflationary type by modulating the epitope context 2, 3, 71…”

“…This is quite relevant in the MCMV model, where specific antigens are generated at different points in the replication cycle, and where in the BALB/c model, the first dominant inflationary response seen was to IE1, which is in the Immediate Early locus 1, 13. However, inflationary responses are seen against a range of epitopes from different proteins, and there are examples of proteins which include an inflationary and non‐inflationary response 4. Furthermore, in the adenovirus model, where there is 1 promotor and 1 protein transcribed, there is still the capacity to produce inflationary and non‐inflationary responses in parallel 16, 25.…”

“…It is a striking phenomenon that was made quite evident by the use of MHC‐peptide tetramers to track responses in the infected animals over time 2, 3, 4. The term is now in reasonably common use5, 6 and the actual features of the responses induced which can be called “inflationary” are quite robustly reproducible between different laboratories using different virus stocks, mouse lines, and infection protocols.…”