Hantaviruses appear to infect their rodent reservoirs with little or no discernible pathology. For SNV and deer mice, and SEOV and rats, the immune responses at persistence strongly resemble that of a regulatory T cell response with prominent expression of TGF-β and lack of inflammatory signatures (9
The work here provides the most detailed examination of the kinetics of infection and host response in deer mice infected with SNV. By day 2 postinfection, low levels of viral RNA were detected in the lungs of some animals. At day 5 and beyond, viral RNA was detected in lungs of all animals, with some animals exhibiting high viral loads (>1,000 copies/mg) while others had low viral loads (<100 copies/mg). The high degree of variation occurred both between groups and within groups. This difference is striking and suggests that polymorphisms within the deer mice may contribute to control of viral gene expression and replication levels. Additionally, some deer mice may be “super spreaders,” animals that, for unknown reasons, produce substantially more virus than other members of the species (36
). Deer mouse 7086 may represent one such animal, although no studies have examined naturally infected deer mice for this characteristic. Excluding deer mouse 7086, viral RNA levels peaked on day 15 postinfection, compared to levels in rats infected with SEOV, which peaked on day 30 (9
). Another limitation of the present work is the small sample size, with 5 deer mice per group. Larger groups may have reduced the variation observed in the present work.
IgG specific for nucleocapsid was first detected on day 10 at low titers in two of the five deer mice, and all five mice from this group had detectable viral RNA in the lungs, including deer mouse 7086, which had the highest copy number of viral RNA in the experiment. The other seropositive deer mouse, number 7040, had 594 copies/mg, which was lower than that of the other deer mice in the group that were seronegative. We previously determined that some experimentally infected deer mice can remain seronegative to nucleocapsid for up to 45 days (29
). Surprisingly, neutralizing antibodies were detected only in the first 7 days postinfection and likely represent low-titer IgM. Presumably, some of these IgM-secreting B cells should undergo class switching to IgG (and presumably other isotypes) and affinity maturation; however, the data presented suggest that this does not occur within 20 days of infection. Because high-titer IgG to the nucleocapsid was produced, class-switching and affinity maturation do not appear to be impaired. The dramatic difference in antibody responses is likely due to multiple factors, including a greater abundance of nucleocapsid antigen for the immune response and the increased sensitivity of ELISA compared to plaque reduction assays. The lack of neutralizing antibody and the presence of viremia suggest that deer mice may secrete infectious virus for several weeks after infection, which could improve the likelihood of transmission to other deer mice, an important component of virus ecology.
Antigen load in tissues appears to be low to moderate, as has been previously reported in deer mice (3
). In contrast, levels of antigen appear to be higher in Syrian golden hamsters infected with the human-pathogenic Andes, Maporal (MAPV), or Choclo (CHOV) hantaviruses (11
), although only ANDV and MAPV are pathogenic in hamsters. SNV causes an apathogenic infection in hamsters and elicits a strong neutralizing antibody response by 28 dpi (18
); thus, it may be informative to examine the kinetics of neutralizing antibodies in this model. The low levels of antigen in deer mice likely reduce antigenic stimulus and could be partially responsible for the lack of neutralizing antibodies by day 20. Cultivation of SNV nucleocapsid-specific CD4+
T cells requires 5 μg/ml, a relatively high concentration; thus, low tissue levels of antigen may not be sufficient for eliciting an aggressive helper T cell response that would be required for affinity maturation and class switching in B cells.
In terms of host immune modulation, several genes were expressed at severalfold increases in the spleens of infected deer mice. As with viral RNA copy number, the increases within infected groups varied substantially, with some animals exhibiting more than 4-fold increases, while others in the same group had expression levels similar to those of the sham-inoculated control groups. In contrast, the 95% confidence intervals of the sham-inoculated groups were small for all time points and all genes, suggesting that the variation observed in infected groups reflects diverse response kinetics in the outbred deer mice. Because many immune genes are expressed only for hours, it may be that expression of these genes in some animals is missed by collecting the RNA, which requires euthanasia of the deer mice, before or after gene expression. This could account for the substantial variation observed since samples were collected at 2- to 5-day intervals.
Splenic expression levels of Ccl2, Ccl3, and Ccl5 were elevated at different points. Ccl3 and Ccl5 are thought to play roles in stimulating inflammatory responses by recruiting neutrophils and T cells to sites of infection (8
). Ccl2 can direct the development of Th2 cells (7
), and IL-4, but not IL-5, expression was elevated in some deer mice, which also has been attributed to Ccl2 expression (20
). In Long Evans rats infected with SEOV, splenic expression of Ccl2 and Ccl5 also occurs (9
The expressions of IL-12 p35 (surrogate for IL-12), IL-21, and IL-23 p19 (surrogate for IL-23) (1
) inflammatory cytokine genes were elevated in spleens of some infected deer mice on days 7 and 10. IL-21 contributes to B cell germinal center formation (21
), and its peak expression correlated with the earliest appearance of IgG to nucleocapsid on day 10. TGF-β expression was elevated on day 5 but declined until day 15 and persisted at day 20. Its biphasic expression has been noted in other infections in which dendritic cells are thought to express it early, followed later by regulatory T cells (19
), a pattern that appears to occur in deer mice infected with SNV. In Long Evans rats infected with SEOV, TGF-β expression is detected on days 3 and 15 postinfection in the spleens and then subsides; however, it is unknown if its expression is reduced between days 3 and 15 (9
). In contrast, in deer mice, pulmonary TGF-β is elevated only on day 20 in only some animals but is elevated in all rats on days 3, 15, and 30 (9
). It is noteworthy that while Long Evans rats are outbred, their genetic diversity is limited compared to that of the deer mice used in these studies since the strain is derived from several Wistar Institute female rats and a single male wild rat. This may account for the more homogeneous response observed in these rats infected with SEOV.
Splenic CD4, CD8α, and TCRβ expressions were elevated on day 5, but CD4 expression declined to background levels on day 10, suggesting that T cell responses were limited. The modulation of these levels could represent an increased expression within given cells, clonal expansion, or a combination of both, and we interpret this modulation as an indicator of T cell activation. Without CD4+
T cells, sustained CD8+
CTL response are impaired and result in persistent infection of laboratory mice (Mus musculus
) with lymphocytic choriomeningitis virus (23
). The lack of a strong helper T cell response in deer mice may contribute to persistence of SNV. IgM expression in the spleens was repressed on days 7 and 10 but elevated in some deer mice on day 15 before declining on day 20. Its detection represents expansion of B cells and/or increased synthesis of IgM heavy chain mRNA. Its decline on day 20 may signify class-switching events to other isotypes since on that day seropositive animals have high IgG titers to the nucleocapsid. TRAF2, a signal transduction protein involved in several pathways, including the NF-κB, MAPK, and caspase 8 pathways, was slightly elevated on day 5 in the lungs of some deer mice; however, we did not detect increased levels of TNF expression, so it is unclear what its elevation means in this context. TRAF2 is a frequent evasion target of viruses (17
) and may be manipulated by SNV in deer mice.
Cluster analysis revealed that several genes involved in T cell mobilization were coordinately expressed (), including those for selection and activation of effector T cells (Ccl2, IL-12 p35, IL-21, IL-23, CD4, CD8, and TCRβ genes) and inflammation (Ccl3 and Ccl5 genes), but most subsided by day 20. These data suggest that in the spleen, a subtle to modest inflammatory T cell immune response is initiated but is not sustained. Correlating with this decline is the expression of the TGF-β gene, which can have anti-inflammatory activities and is prominently expressed by T cells from persistently infected deer mice and rats (10
). Despite the inflammatory gene expression signature, no histological evidence of inflammation was noted in any of the deer mouse tissues in this study. It may be that translational regulation prevents inflammation or that insufficient amounts of inflammatory proteins are made.
In the lungs, chemokine gene expression was not detected until 20 dpi, when Ccl2 and Ccl5 were elevated. Despite the expression of these genes, we did not have evidence that recruitment or activation of T cells occurred in the lungs. This is in agreement with previous work in which pulmonary tissue histology failed to show marked leukocytic infiltrates in the lungs of infected deer mice (3
). At all other time points in the lungs, cytokine and chemokine gene expressions were at baseline, except for expressions of Ccl2, Ccl5, and TGF-β genes, which were elevated on day 20 in some animals. The expression of these three genes is seemingly paradoxical considering one is involved in Th2 responses (Ccl2 gene), one in inflammatory responses (Ccl5 gene), and one in anti-inflammatory responses (TGF-β gene). Evaluation of deer mice beyond 20 dpi may clarify the roles of these genes.
The expression of transcription factors was significantly repressed in both lungs and spleens on day 2 postinfection. While the increase of expression of most of these genes is low, transcription factors are enzyme-like in their abilities to initiate transcription of genes. The pattern of expression suggests differentiation of both Th1 and Th2 cells and correlates with initial expansion of T cells on day 5 postinfection. We did not observe elevation in Fox-p3 expression in either the lungs or spleens of infected deer mice. Our previous work demonstrated that SNV-specific CD4+ T cell lines established from most deer mice 42 days postinfection expressed Fox-p3. It is possible that the low sensitivity (i.e., signal-to-noise ratio) of using whole tissues, in which only a small percentage of the T cells are likely virus specific, versus homogeneous antigen-specific T cell lines, or that the fact that in the present work the T cells are from days 20 and before, may account for this difference.
In our previous work with SNV-specific CD4+
T cell lines (34
), we detected severalfold increases of IFN-γ, IL-4, IL-5, and IL-10. Only increased expression of IL-4 was detected in spleens of some deer mice in the present study. Levels of TNF and LT, both of which are found in autopsy specimens of HCPS patients (30
), generally were not elevated at any time point in the experiment, suggesting that the roles for these cytokines are minimal during infection. The low-level expression also suggests that in hantavirus reservoir models, immune responses are subdued relative to pathogenic models of viral infections (9
). With T cell lines, this problem is likely obviated since all the T cells in the culture are virus specific and are harvested after 2 days of cultivation with antigen. Unfortunately, a better comparison would be pathogenic hantavirus models in Syrian golden hamsters; however, immunologic reagents and molecular and cellular methodologies are not available for such an assessment in this species.
The dramatic repression of gene expression 2 dpi suggests antiviral effector mechanisms are active in deer mice infected with SNV. Considering that many of these proteins are highly conserved, it should be possible to use antibodies specific to these proteins in other species to dissect the events that occur during cellular infection since SNV may have transcriptional activity (27
). Propagation of deer mouse cells susceptible to infection should be a high priority so that further scrutiny of the early events of infection can be performed. Additional effort should also focus on developing antibodies to detect the cytokines that are expressed now that evidence exists that they are involved in the host response to SNV.