The frequency of HSV-1 reactivation from latency in ex vivo TG cultures is inversely proportional to the size of the ganglionic CD8+ T cell population . Based on these findings some have advocated that all future HSV-1 vaccines be evaluated based on their capacity to enhance the HSV-specific CD8+ T cell population in latently infected sensory ganglia. Such an approach would only be feasible if the HSV-specific CD8+ T cells expanded in peripheral lymphoid organs through immunization have access to latently infected ganglia.
Evidence obtained in a model system in which latently infected DRG are transplanted under the kidney capsule of recipient mice suggest that HSV-specific CD8+ T cells do not infiltrate latently infected DRG even when the resident CD8+ T cell population is substantially disrupted . However, a number of factors that may be unique to the transplanted tissue could have influenced CD8+ T cell migration in that model. The authors suggested that the loss of CD8+ T cells from the DRG following transplantation was likely due to death resulting from the trauma of transplantation. Such trauma could also induce other changes within the microenvironment of the latently infected ganglion such as alterations in the chemokine and cytokine milieu that could influence CD8+ T cell infiltration. Moreover, revascularization of the transplanted tissue could influence T cell migration.
Based on these concerns we sought a more physiological model in which to examine the HSV-specific CD8+ T cell infiltration into latently infected TG. We employed a model in which HSV-1 reactivation from latency is induced by exposure to psychological stress rather than trauma to the TG, depletion of the ganglion-resident CD8+ T cell population is accomplished through exposure to corticosterone, and CD8+ T cell migration into infected TG is examined at the orthotopic site.
Our findings strongly support the notion that the unperturbed CD8+ T cell population in latently infected TG is maintained without detectable replenishment from the peripheral blood. We provide two types of evidence in support of this theory. We first identified chemokine receptors that are used to direct CD8+ T cells into acutely infected TG and asked if blocking these receptors during latency would result in diminution of the CD8+ T cell population in latently infected TG. The chemokine receptors CCR5 and CXCR3 are highly expressed on activated T cells and are expressed in HSV- acutely and latently infected mouse and human TG as are the corresponding chemokines CCL5 and CXCL10 [24, 29, 41–43]. However, a direct influence of CCR5 and CXCR3 on CD8+ T cell infiltration into acutely or latently infected TG has not been established. Here we demonstrate that systemic treatment with TAK-779, a chemical non-peptide inhibitor of both CXCR3 and CCR5  significantly reduced the infiltration of CD8+ effector T cells into acutely infected TG, but did not influence the size of the CD8+ memory T cell population within latently infected TG. These findings suggested that if the CD8+ T cell population in HSV-1 latently infected TG is maintained through infiltration of CD8+ T cells from the peripheral blood, the CCR5 and CXCR3 chemokine receptors do not appear to play an essential role in directing their migration despite the presence of their ligands in the TG.
More direct evidence that HSV-specific CD8 T cells do not migrate into latently infected TG came from the observation that adoptively transferred gB-CD8 that are retained in the blood, spleen, lymph nodes, and lungs of recipient mice over an extended period of at least 4.5 weeks are not detectable in the TG during the same period (Figure 3). The above findings strongly suggest that TG containing an established CD8+ memory T cell population and latent HSV-1 are not permissive to CD8+ T cell infiltration. We next asked if diminution of the resident CD8+ T cell population through elevated serum corticosterone levels and reactivation of HSV-1 from latency through exposure of mice to restraint stress would change the microenvironment of the TG, rendering it permissive to CD8+ T cell infiltration. However, adoptively transferred gB-CD8 T cells did not enter the TG from the blood even when reactivation and diminution of the TG-resident CD8+ T cell population was induced by exposure of latently infected mice to stress and corticosterone (Figure 4).
A caveat to these findings is that the adoptively transferred CD8+ T cells obtained from gBT1.1 mice may not accurately reflect the migratory capability of the endogenous gB-specific CD8+ T cell memory population in the lymphoid organs and blood of the host. This seems unlikely given that neither effector cells generated through in vitro stimulation for 3 days with the gB498-505 peptide, nor memory cells obtained by incubating the effector cells for an additional 10 days in medium containing IL-15 were able to enter the TG. The transferred memory population expressed high levels of the LFA-1 and VLA-4 adhesion molecules typically used by CD8+ T cells to enter infected tissue. Moreover, CD8+ T cells that were isolated from spleens of HSV-1 latently infected mice (30 dpi) also failed to enter the TG during recovery of the CD8+ T cell population following stress and corticosterone treatment (data not shown). These findings combined with observations made with transplanted HSV-1 latently infected DRG, and with vesicular stomatitis virus infected brains strongly suggest that once a Trm CD8+ T cell population is established in nervous tissue, further infiltration of cells from the blood is effectively blocked [33, 44].
Since infiltration of CD8+ T cells from the blood did not seem to account for recovery of the CD8+ T cell population in TG of stress and corticosterone treated mice, we concluded that recovery likely resulted from proliferation of the remaining cells. However, proliferation (based on BrdU incorporation into CD8+ T cells) was the same or slightly lower in the recovered population when compared to the homeostatic proliferation of the CD8+ T cell population in control mice that were not exposed to stress and corticosterone. While it is possible that a burst of proliferation was missed, the low level of gB-specific CD8+ T cell proliferation measured from 24-36 hours after terminating stress and corticosterone treatment would not seem to account for the rapid recovery of the CD8+ T cell population.
Interesting differences between our findings and those obtained with transplanted latently infected DRG deserve discussion. In our study, the kinetics of CD8+ T cell recovery, the composition of CD8+ T cell population (frequency of gB-CD8 T cells), and the size of the recovered population were all independent of CD4+ T cell help (Figure 6). This is in contrast to the transplantation model where donor CD4+ T cells were required for recovery of the donor CD8+ T cell population in the DRG. Also differing in the two models is the kinetics of recovery of the CD8+ T cell population. In our model recovery was complete 4 days after terminating stress and corticosterone treatment. In contrast, recovery of the CD8+ T cell population in the transplanted DRG was negligible until 6 days after transplant, and peaked at 9 days after transplant. Finally the size of the recovered CD8+ T cell population in our model closely approximated that in unperturbed control TG, while the recovered gB-specific CD8+ T cell population in the transplanted DRG was approximately 17-fold higher than that observed prior to excision.
We believe these differences can be explained by differences in the two models and may be informative. The loss of CD8+ T cells from the DRG following transplantation may reflect death of the CD8+ T cells due to the trauma of DRG excision and transplantation as suggested by the authors of that report. Thus, recovery of the population would require proliferation of the surviving cells that were at very low levels. This could explain the delayed kinetics of recovery and the requirement for DC and CD4+ T cell help. In contrast, corticosterone causes T cells to emigrate from tissues . In our model, depletion of the CD8+ T cell population from the latently infected TG may have resulted from migration out of the TG, followed by immigration back into the TG over the ensuing 4 days. This would explain the lack of proliferation of the recovered cells as well as the lack of requirement for CD4+ T cell help. It would also explain the identical frequency of gB-specific CD8+ T cells in the recovered and pre-treatment population. However, the failure of adoptively transferred cells to enter the TG during recovery would suggest selective re-entry of the original TG-resident CD8+ T cells. This would suggest acquisition by the TG resident CD8+ T cells of a homing receptor that permits selective re-entry into the latently infected TG. This possibility is under investigation.
Our findings suggest that at some point after establishment of an HSV-specific CD8+ T cell population the infected TG becomes resistant to further T cell infiltration. We have shown that the gB-CD8 T cells exhibit a progressively higher functional avidity (ability to detect a low epitope density) over time in latently infected TG, whereas their counterparts in the spleen and lungs showed decreased functional avidity over the same period . The ability to detect very low levels of MHC/peptide complexes on latently infected neurons would likely enhance the ability of CD8+ T cells to provide immunesurveillance of latently infected ganglia, but be of lesser importance in the periphery. Therefore, there might be a selective advantage to the host to restricting infiltration of HSV-specific CD8+ T cells into the TG. However, restricting entrance of CD8+ T cells into the latently infected TG will complicate the development of therapeutic vaccines designed to bolster the TG-resident CD8+ T cell population.