We have previously shown that a gammaherpesvirus infection can exacerbate established FITC-induced fibrosis in a murine model . In this study, we found that TLR-9 signaling plays a role in limiting the profibrotic exacerbations of gammaherpesviral infection. There are many potential mechanisms by which TLR-9 signaling might influence the fibrotic environment during viral infection. One possible mechanism is simply by limiting viral replication. Guggemoos et al. showed that TLR-9 signaling is important in control of γHV68 infection when given intraperitoneally ; however, they were not able to show that TLR-9 signaling in the lung was important in controlling an intranasal infection with γHV68. Our data are consistent with this finding. Although replication of the virus was greater in mice treated with bleomycin compared with non-fibrotic mice, the absence of TLR-9 had little effect on viral gene expression in either group. Although it is not entirely clear why viral replication is enhanced in fibrotic mice, it is likely that proinflammatory and profibrotic factors enhance viral gene transcription. In support of this hypothesis, prostaglandin E2 has been shown to promote viral replication . Thus, our results confirm that TLR-9 signaling is not required for control of lytic γHV68 infection in the lung. The reasons why TLR-9 would differentially regulate lung versus peritoneal infection probably reflect the differences in the cell populations that are initially infected in each site.
Within the lung, AECs are one of the primary cell types infected with γHV68, as confirmed by our immunohistochemical findings in this study (Figure 4A) and our previous studies [31, 32]. Although it is known that γHV68 expresses proteins that can prevent apoptosis and enable establishment of latent infection [37–39], previous studies in human cells and tissues have suggested a role for gammaherpesviruses in the induction of AEC apoptosis [10, 23]. Furthermore, TLR-9 stimulation has been shown to inhibit macrophage apoptosis . Thus, we were surprised to discover that levels of apoptosis in AECs isolated from mice treated with bleomycin plusγHV68 were similar between Balb/c and TLR-9-/- genotypes. This was true regardless of whether apoptosis was assessed by M30 staining or caspase activation.
One important aspect of AEC function is to limit fibroproliferation. Profibrotic stimuli (for example, chemokine (C-C motif) ligand 2 (CCL2)) are known to alter the ability of isolated AECs to limit fibroproliferation . Thus, we sought to determine whether there were basal differences in the ability of the AECs from Balb/c or TLR-9-/- mice to control fibroblast proliferation. However, AECs from both genotypes of mice were equivalent in their ability to limit fibroblast proliferation from both strains. Taken together, we could not find substantial evidence of altered AEC apoptosis or function to explain the exaggerated viral-induced fibrotic response noted in TLR-9-/- mice.
We next investigated the inflammatory response between Balb/c and TLR-9-/- mice treated with bleomycin plus γHV68. After viral exacerbation, both genotypes of mice had an increase in total numbers of inflammatory cells, but there were few differences noted in the particular cell types recruited. There were no significant differences in fibrocyte accumulation, or in the percentages of neutrophils or CD4, natural killer or B cells between genotypes. The TLR-9-/- mice did have a higher percentage of CD8+ cells, and it is possible that if activated, these CD8+ cells could contribute to tissue damage, which might exaggerate the fibrotic response in the TLR-9-/- mice. Both genotypes had a loss of B cells after infection, and we believe this represents the migration of B cells to the spleen, the known major reservoir for latent viral infection .
Because TLR-9 signaling leads to NFκB activation and the production of Th1 cytokines , it seemed likely that TLR-9-/- mice would have a more Th2-biased cytokine environment. Because Th2-biased mice are known to be prone to the development of fibrosis in response to infection with γHV68 , it was possible that a cytokine imbalance could explain the exacerbation of fibrosis in TLR-9-/- mice. However, there was no evidence of Th2 skewing or defective Th1 signaling in the TLR-9-/- mice during the acute response to bleomycin and infection. In fact, IL-4 and IL-13 levels were reduced on day 21 after bleomycin plus γHV68 administration in TLR-9-/- mice. We were surprised that IFN-γ was not diminished in the TLR-9-/- mice, and we believe that this may reflect the fact that in AECs at least, activation of other TLRs may help to stimulate NFκB activation for IFN-γ production. We confirmed that lung AECs from TLR-9-/- mice express normal levels of TLR7 and TLR8, for instance (data not shown). Additionally, differences in production of TGF-β were not noted between genotypes. The observation that fibrosis is exaggerated despite strong induction of IFN-γ confirms our previous findings of viral exacerbation in the C57Bl/6 background . Thus, we conclude that neither reduced production of Th1 cytokines nor increased production of Th2 cytokines in response to infection can explain the enhanced exacerbation of fibrosis noted in TLR-9-/- mice.
Not surprisingly, the one cytokine whose production was different between Balb/c and TLR-9-/- mice was IFN-β. TLR-9 stimulation is known to induce type I interferons . Levels of IFN-β were reduced after bleomycin plus virus in whole-lung homogenates and in infected fibroblasts. It is interesting that after bleomycin administration alone, levels of IFN-β were reduced (Figure 6A). The fact that this happened in both Balb/c and TLR-9-/- mice suggest a TLR-9-indepdent mechanism for the reduction in IFN-β after a fibrotic insult. This should not necessarily be interpreted to mean that bleomycin causes a reduction in TLR-9 expression, as we have no evidence of that, at least in fibroblasts isolated from fibrotic mice (data not shown). As IFN-β is known to be a potent inhibitor of lung fibroblast proliferation , a fact we have confirmed (Figure 7), it is reasonable to conclude that early viral infection of fibroblasts would result in the stimulation of TLR-9 by viral CpG DNA sequences and a concomitant decrease in fibroblast proliferation. In fact, we found that fibroblasts from Balb/c mice potently upregulated IFN-β secretion in response to even low-level infection. Furthermore, exogenous IFN-β could limit proliferation of fibroblast from both strains, suggesting that the defect in TLR-9-/- mice is not sensitivity to, but rather production of, IFN-β. Taken together, we conclude that although viral infection after fibrotic challenge increases the fibrotic response of both strains, the more extensive exacerbation of fibrosis noted in the TLR-9-/- mice is most likely due to deficiencies in IFN-β production, which in turn allow for increased fibroproliferation. When cell number and viral dose were equivalent, fibroproliferation was inhibited in Balb/c, but not TLR-9-/- mice (Figure 8). Because there may be some TLR-9-independent signaling that leads to IFN-β production in the lung by day 7 after infection (Figure 6A), it is possible that at later time points, sufficient IFN-β may be available to limit fibroproliferation even in TLR-9-/- mice. It is important to remember that secretion of IFN-β in the lung would probably inhibit the proliferation of other resident fibroblasts, not just those that happened to be infected.
TLR-9-/- mice did not respond differently from wild-type mice to bleomycin alone or FITC alone. These results suggest that fibrotic insults alone may not generate endogenous ligands for TLR-9 stimulation. Thus, we next wanted to determine whether therapeutic stimulation of TLR-9 with CpG ODN could protect against fibrotic challenge. We initially tried to perform these experiments in the Balb/c background, but had poor responses to bleomycin in two separate experiments. Although levels of collagen were somewhat lower in CpG ODN-treated mice than in control mice, the overall levels of fibrosis were so low that meaningful interpretations of the data were not possible. Given the expense of these experiments, we chose to perform them in the C57Bl/6 background, which has a more reproducible fibrotic response to bleomycin or Blenoxane . We found that mice treated intranasally with CpG ODN were significantly protected from the development of bleomycin-induced fibrosis (Figure 9). Although these results in murine studies were very exciting, recent studies in human cells have made it clear that it is unlikely that these results can be extrapolated to humans.
Our results are not the first to describe an enhanced fibrotic response in TLR-9-/- mice. Earlier studies have shown that granulomas that form in response to Schistosoma mansoni eggs in TLR9-/- mice are larger and have more collagen deposition within the granuloma than in wild-type mice . However, in that study, the results were associated with diminished Th1 and augmented Th2 responses, probably reflecting the strong Th2-skewing nature of the S. mansoni egg antigens, whereas we used a Th1-inducing viral infection. Thus, there are now at least two examples of TLR-9 signaling showing a protective effect in lung-fibrosis models in mice. By contrast, Meneghin et al. found that TLR-9 expression in human lung fibroblasts promotes myofibroblast differentiation , and was anticipated to worsen fibrotic disease. This human study also found that TLR-9 was expressed at low levels in surgical lung biopsies of normal subjects, but was dramatically upregulated in the biopsies from patients with fibrotic lung diseases. These data suggest that TLR-9 expression might be low in normal human lung fibroblasts, whereas our studies suggest that murine lung fibroblasts express this receptor constitutively. In addition, TLR-9 expression in human lung epithelium is modest compared with that at other sites in the body . Studies in both human lung fibroblasts and human hepatic stellate cells have shown the ability of TLR-9 stimulation via CpG ODN to induce myofibroblast differentiation [28, 46]. Furthermore, expression of TLR-9 on fibroblasts obtained from initial surgical lung biopsies of patients with IPF can distinguish patients with rapid disease progression (TLR-9-positive) from patients with a slower disease course (TLR-9-negative) . Thus, it seems that in humans, TLR-9 stimulation may promote fibrotic reactions, whereas in mice, it may protect. Another important difference that may explain the improved outcomes in TLR-9-expressing mice is that mice have a larger repertoire of CpG-responsive hematopoietic cells. In mice, all DC subsets, B cells and macrophages respond to CpG ODN, whereas in humans, only plasmacytoid DCs and B cells respond . Thus, it is likely that production of IFN-β in response to TLR-9 stimulation is greater in the murine system than it would be in humans. Collectively, these studies identify important differences between mice and men, and thus suggest that CpG therapy, although beneficial in rodent models of fibrosis, is unlikely to be beneficial for human treatment of IPF.