Two studies have suggested a link between infection with the hepatitis C virus (HCV) and IPF. HCV is a small, enveloped, positive-sense single-stranded RNA virus in the Flavivirus family 32. Replication within hepatocytes causes a form of hepatitis, but the virus may be able to enter other cell types 33. Ueda et al. 34 were able to show a significant difference between the percentage of Japanese IPF patients who had serum antibodies to HCV (28.8%) and the corresponding percentage of control subjects (3.66%). A subsequent Italian study confirmed that 13% of patients with IPF were seropositive for HCV 35. In a control group of 4,614 blood donors, the prevalence of HCV antibodies was lower (0.3%). However, in a control group of 130 patients with non-interstitial lung disease, HCV antibody prevalence was 6.1%. While this was significantly higher than the blood donor group, the incidence of HCV in patients with non-interstitial lung disease and IPF were not statistically different. It should also be noted that a British study failed to find an association between IPF and HCV 36. One possible explanation for these findings is that there may be geographical differences in the prevalence of HCV infection, as the infection is more commonly seen in Japan and Mediterranean countries than it is in northern Europe 35. Given that HCV is not known to replicate in the lung, it is not clear whether these associations suggest that HCV is pathogenic in IPF, or rather if they indicate that IPF patients develop HCV cross-reactive antibodies. Thus, further research in this area will be needed.

Human adenoviruses have been suggested as etiological co-factors in the progression of interstitial lung disease 373839. Adenoviruses are medium-sized, non-enveloped, icosahedral, double-stranded DNA viruses. They are relatively resistant to chemical and physical agents, and as a result, they can remain infectious outside of the body for extended periods of time. Adenovirus infections typically cause respiratory symptoms and can be shed for long periods of time post-infection 37. Kuwano et al. 38 examined 19 patients with IPF, 10 patients with interstitial pneumonia associated with collagen vascular disease, and 20 patients with sarcoidosis using nested PCR and in situ hybridization for the adenovirus gene product E1A. E1A DNA was present in 3 out of 19 (16%) cases of IPF, in 5 of 10 (50%) cases of interstitial pneumonia associated with collagen vascular disease, and in 2 of 20 (10%) cases of sarcoidosis 38. While these data are not suggestive of a correlation between adenovirus infection and pulmonary fibrosis, Kuwano et al. found that the incidence of E1A DNA was considerably higher in patients who had been treated with corticosteroids (67%) compared to those patients left untreated (10%). This finding raises the interesting possibility that corticosteroids, a common therapy for IPF, may make patients more susceptible to adenovirus infection or reactivation from latency. However, studies investigating the titer of anti-adenoviral IgG in IPF patients have failed to demonstrate an elevation above normal 40. Despite this, it is important to note that E1A has been shown to upregulate production of the pro-fibrotic mediator, TGF-β, and to induce lung epithelial cells to express mesenchymal markers 41. This mechanism has been implicated in the architectural remodeling that occurs in chronic obstructive pulmonary disease. It is possible that a similar mechanism could contribute to extracellular matrix deposition and remodeling in those IPF patients who may also have adenovirus infections.

It is also conceivable that adenovirus infections could serve as exacerbating agents for patients with established lung fibrosis, although it is difficult to determine the frequency of this happening from published clinical literature. Furthermore, recent studies using an animal model of fluorescein isothiocyante (FITC)-induced fibrosis were unable to demonstrate significant exacerbation of FITC-fibrosis within the first 7 days post-mouse adenoviral infection 42. One note of caution in the interpretation of these experiments, however, is that human and mouse adenoviruses do show different tropisms, with human adenoviruses being predominantly respiratory pathogens. Interestingly, in studies using the same murine model, a gammaherpesvirus was able to augment FITC-induced fibrosis (see below). Whether the difference in the ability of mouse adenovirus and mouse gammaherpesvirus to exacerbate FITC-fibrosis represents differences in cell tropisms, inflammatory response or mediators released is unknown.

Human cytomegalovirus (HCMV), a betaherpesvirus, is a widespread opportunistic pathogen that persists in healthy individuals but normally only causes clinical manifestations in immune-compromised individuals 43. HCMV infects the respiratory tract, and it has been evaluated with regards to IPF. Dworniczak et al. 44 studied 16 patients, newly diagnosed with IPF and never treated, compared to 16 adult healthy volunteers. HCMV DNA copy number in broncho-alveolar lavage (BAL) cells, blood leukocytes, and serum was calculated by real-time PCR, and the prevalence of the HCMV DNA positive subjects in the patient group (75%) did not differ significantly from the prevalence of positive subjects in the control group (69%). IPF patients did show significantly higher DNA copy numbers in their blood compared to controls, however 44. Also, the viral copy number in the BAL cells of both IPF patients and healthy volunteers was elevated relative to respective viral copy numbers in blood leukocytes, suggesting an important role for the lungs (perhaps as a viral reservoir) in the pathobiology of HCMV. Consistent with this idea, a subsequent study by Tang et al. 45 did note higher levels of HCMV DNA in IPF lung tissue compared to control samples. Similarly, in a study by Yonemaru et al. 40, HCMV IgG and complement fixation titers were found to be elevated in the serum of patients with IPF when compared to several other disease-specific controls. In a retrospective study of lung transplant recipients, 102 patients were screened by urine test for evidence of HCMV infection on the day of transplant. Only five patients were found to be HMCV+ prior to transplant, and all five of the patients were IPF patients 46. Despite testing positive, none of these five patients exhibited symptoms of HCMV disease, suggesting that viral infections in this population can be occult. The increased incidence of HCMV infection prior to transplant correlated with an increased risk of HCMV infection post-transplant as well. In sum, there is evidence that suggests an association between the incidence of HCMV infection and the incidence of IPF, but a mechanism by which HCMV may affect fibrosis remains to be elucidated.

The virus that has been associated most strongly with IPF is EBV. EBV is a gammaherpesvirus that is present in all populations, infecting more than 95% of humans within the first decades of life 47. An association between EBV infection and IPF was first established when elevated levels of immunoglobulins A and G against EBV antigens were measured in a serological study of 13 patients with IPF 48. In contrast, 12 patients with interstitial lung disease of known cause had normal EBV serological profiles. This finding led to further research on EBV in the context of IPF. An immunohistochemical study indicated that EBV replicates within epithelial cells of the lower respiratory tract in IPF patients 49. Consequently, Stewart et al. 50 sought to confirm the presence of EBV DNA in the lung tissue of IPF patients using PCR. They found that EBV was present in the lungs of patients with IPF at a significantly higher percentage (48%) than in the lungs of control subjects (14%).

A couple of studies have associated the presence of active and latent EBV markers with IPF. Kelly et al. 51 investigated the occurrence of productive EBV replication by analyzing for the presence of an EBV gene rearrangement termed WZhet; 61% of EBV DNA-postive lung tissue biopsies from IPF patients were positive for WZhet. Buffy coat analysis for WZhet was positive in 16 of 27 IPF patients compared to none of 32 lung transplant recipients and 1 of 24 normal blood donors. Tsukamoto et al. 52 then determined that the presence of EBV latent membrane protein 1 (LMP1) is linked with more rapid disease progression. From a group of 29 patients, they found that patients positive for LMP1 died more quickly than patients who tested negative for EBV.

It should be noted that not all studies have found an association between EBV and IPF. In 1997, Wangoo et al. 53 published findings contrary to previous reports when they did not detect any EBV DNA in the lungs of IPF patients. Also, in 2005, an Italian study by Zamo et al. 54 failed to find evidence of either EBV or human herpesvirus (HHV)-8 DNA in their tissue banks of IPF samples. Whether these discrepancies reflect geographical distribution, technical sensitivities, or disease heterogeneity is still unclear.

Although EBV had been detected with more frequency in the lungs of IPF patients than in the lungs of control patients in most previous studies, many members of each IPF cohort analyzed did not test positive for EBV infection at all. Tang et al. 45 went on to test the hypothesis that at least one herpesvirus could be detected in the lungs of all IPF patients. They identified one or more of four herpesviruses – EBV, HCMV, HHV-7, and HHV-8 – in 32 of 33 patients with IPF and in 9 of 25 controls. They found two or more herpesviruses in 19 of 33 IPF patients and in 2 of 25 controls. These data strongly support the notion that at least one herpesviral infection accompanies the development of IPF.

Tang et al. drew other conclusions from their study that suggest susceptibility to viral infection and IPF depends on a genetic or acquired predisposition. Co-infection occurred more frequently in patients with the sporadic form of IPF compared to those with the familial form. Familial IPF is characterized by the incidence of IPF in two or more members of an immediate family 55. This led the authors to suggest that a patient with familial IPF may require less viral influence to trigger a progressive fibrotic response than a patient with sporadic IPF. In addition, Tang et al. note that the increased frequency of HHV-8 in the lungs of the IPF cohort is particularly interesting. In the United States, HHV-8 infection is predominantly found in patients with HIV infection and Kaposi's sarcoma 56, and all of the subjects tested negative for HIV in this study.

Collectively, the analyses of IPF lung tissue chronicled above create a rationale to study the association between viral infections and the occurrence of IPF, but do not provide evidence for a causal relationship between viruses and IPF. Demonstrating causation in humans requires detection of a virus in the lungs prior to clinical manifestations of IPF (which is clinically implausible) or evidence that an anti-viral therapy confers anti-fibrotic effects. The latter has been attempted with some success in a limited number of case studies, but no large trials have been conducted to date 4557. While at least four different viruses have been correlated with IPF, the most striking observations have linked EBV infection of lung tissue with the presence of IPF. However, we want to again stress that the pathogenesis of IPF is complex and multifactorial. In reality, IPF is likely a spectrum of diseases that result from a variety of genetic abnormalities and/or environmental factors. As mentioned above, the data regarding the association of viruses, even EBV, with IPF are controversial. What is particularly intriguing to us, however, is that these clinical observations are somewhat strengthened by emerging evidence in animal models that demonstrate that gammaherpesvirus infections can be linked to the development or the exacerbation of experimentally induced fibrosis. These animal studies will be highlighted below.

Pulmonary fibrosis has been reported to occur in both cats and horses 5859606162. Interestingly, recent studies of both horses and a single donkey have reported pulmonary interstitial fibrosis associated with herpesvirus-associated pneumonia 5963. It is interesting to note that infection with equine herpesvirus-5, a gammaherpesvirus, was detected by PCR in 19/24 (79.2%) horses affected with interstitial fibrosis whereas only 2/23 (8.7%) of control horses showed evidence of infection 59. While these data do not prove causality, the association between gammaherpesviruses and fibrosis in both horses and humans is striking.

Murine models of pulmonary fibrosis have enabled identification of pathogenic cells and mediators that are believed to be important in human fibrotic disease, and they have facilitated further exploration of a pathogenic role for viruses in humans 64. The human herpesviruses identified to be prevalent in IPF lung tissue have limited infection capability in mice, however. Thus, investigators have utilized a natural murine pathogen called murine gammaherpesvirus (MHV)-68 that has been characterized as genetically and biologically closely related to human gammaherpesviruses. The genome of MHV-68 is largely colinear with EBV and HHV-8, and there is evidence that both murine and human gammaherpesviruses infect the respiratory tract and can persist in B cells as well as lung epithelial cells 6566676869.

In 2002, Lok et al. 70 used MHV-68 to demonstrate that gammaherpesviruses could serve as a cofactor in the development of pulmonary fibrosis. BALB/c mice infected intranasally with MHV-68 one week prior to intratracheal administration with the fibrotic stimulant bleomycin later developed pulmonary fibrosis even though BALB/c mice are normally resistant to bleomycin. Mice infected with MHV-68 but not challenged with bleomycin did not develop fibrosis. The gammaherpesvirus infection alone was not sufficient to cause fibrosis, but Lok et al. proposed that a viral infection made the previously protected lungs susceptible to fibrotic disease upon the event of an exogenous injury. These results are enticing, but it should be noted that the bleomycin was delivered during the peak of lytic viral infection. It is difficult to infer from these studies whether chronic latent infection with MHV-68 might also predispose the lung to subsequent fibrotic responses. Also, the mechanism(s) for how MHV-68 infection augmented the subsequent fibrotic response to bleomycin were not defined. This is an area of active research in our laboratory. We have determined that MHV-68 infection is latent in the lung by day 14 post-infection 42. Mice given MHV-68 14 days prior to the administration of bleomycin and harvested 21 days post-bleomcyin show a different pattern of inflammation and fibrosis than is noted in mice mock infected prior to bleomycin administration (Figure 1). Of note, focal clusters of mononuclear inflammatory leukocytes are seen in the setting of viral infection. Thus, it is likely that latent viral infections alter the inflammatory response that occurs in response to a secondary fibrotic stimulus.

Ebrahimi et al. 71 showed that MHV-68 causes fibrosis in interferon (IFN)γR-/- mice. IFN-γ is a Th1 cytokine with anti-viral and anti-fibrogenic properties. It down-regulates the expression of both type I and type III collagens and fibronectin 72737475. The cytokine profile of IFNγR-/- mice is Th2-biased, resulting in a cytokine imbalance similar to that observed in the lungs of IPF patients 76. This study suggests that a herpesvirus infection delivered to lungs skewed towards a profibrotic cytokine environment is sufficient for fibrogenesis. In this case, the fibrogenesis was not limited to the lung only. The mice developed multi-organ fibrosis (liver, lung, spleen and lymph nodes). The development of multi-organ fibrosis correlated with an overproduction of pro-fibrotic mediators such as TNF-α, TNF-β, IL-1β, TGF-β1, lymphotactin, and macrophage inflammatory protein-1β (MIP-1β). These mediators were elevated on day 14 after infection whereas the anti-fibrotic chemokines CXCL9 (CXC chemokine ligand 9) and CXCL10 were significantly reduced. The authors noted that MHV-68 gene expression may have influenced the cytokine imbalance. These results are fascinating in light of the fact that MHV-68 infection in wild-type mice shows no signs of causing multi-organ fibrosis and suggest that the outcome of viral infection may be critically dependent on the cytokine milieu at the time of infection.

Mora et al. 57 extended studies in the IFNγR-/- model of herpesvirus-induced fibrosis by characterizing the disease seen in the lungs and examining possible pathogenic effects of the virus that contribute to fibrosis. They showed that MHV-68 induces epithelial damage and inflammatory responses leading to alveolar remodeling and ultimately to unresolving progressive interstitial fibrosis resembling human IPF. More recently, Mora et al. 77 published data implicating alveolar macrophages as integral profibrotic effectors in IFNγR-/- mice. These studies suggested that alveolar macrophages were chronically recruited to areas of epithelial hyperplasia and fibrosis and that these alveolar macrophages displayed signs of alternative activation, a process known to be driven by Th2 cytokines. The alternatively activated macrophages express arginase. Arginase metabolism of L-arginine to L-ornithine, L-proline, and polyamine promotes fibroblast proliferation, collagen production, and, ultimately, fibrosis. In addition, there is evidence that microvascular injury has a role in the pathogenesis of IPF 78, and Mora et al. 57 detected viral-induced vasculitis accompanied by red blood cell extravasation compatible with hemorrhage. With regards to viral-induced vasculitis, it is interesting that Magro et al. 78, who proposed microvascular injury as a pathogenic mechanism for IPF, also found evidence of CMV and parvovirus B19 infection in IPF patients. In fact, Magro et al. 78 speculated that endotheliotropic viral infections, such as CMV and B19, may be precursors for the microvascular injury that is noted in IPF 78. Similar mechanisms may be responsible for the MHV-68-induced damage noted in the studies by Mora et al. MHV-68 has previously been reported to cause vascular damage 79. Finally, Mora et al. 57 noted enhanced expression of the profibrotic cytokine TGF-β1 in the MHV-68-infected IFNγR-/- mice, and it is likely that this pro-fibrotic cytokine may make a significant contribution to the fibrosis in this model. In wild-type mice, however, IFN-γ signaling would be expected to inhibit transcription of the TGF-β gene 80.

The authors attribute some of the TGF-β1 dysregulation to epithelial cell damage. Interestingly, type II alveolar cells are a target of MHV-68, and Mora et al. 57 speculate that epithelial cell infection and injury may be triggering surfactant abnormalities as well as dysregulated epithelial cell repair. An association between the frequency of polymorphic variants of surfactants and IPF has been documented previously 81. Furthermore, another series of studies on IPF has found some evidence that a link between EBV and p53 expression leads to modifications in epithelial cell repair and apoptosis 8283. Flano et al. 84 have reported chronic low level reactivation of MHV-68 in the lung. It is possible that reactivation is a repetitive trigger in the setting of the IFNγR-/- mice that contributes to the progressive disruption of lung epithelial cells and unresolving fibrosis. In support of this, preventing chronic reactivation of MHV-68 to the lytic phase through the use of anti-viral drugs reduces fibrosis in IFN-γR-/- mice 8485. Table 1 summarizes potential mechanisms whereby preceding viral infection may predispose the host to the development of fibrosis.

While most IPF patients have a slow, progressive disease, some patients have an acute deterioration in function that carries a poor prognosis. In the placebo arm of a study of 32 patients who died from IPF-related causes, Martinez et al. 86 reported that 47% suffered an acute deterioration, and 27% of the acute deteriorations were associated with infection. We were recently able to model acute exacerbations of fibrosis in a murine model of FITC-induced pulmonary fibrosis. Wild-type mice infected with MHV-68 after the establishment of fibrosis (day 14) developed a significantly worse fibrotic response than fibrotic mice that were mock-infected with saline 42. This finding is especially interesting in light of the work done in Th2-biased mice discussed above because MHV-68 exacerbation of fibrosis in wild-type mice occurred despite a strong Th1-biased anti-viral immune response. In fact, MHV-68 was able to exacerbate FITC-induced fibrosis even in Th2-deficient (IL-4 and IL-13-/-) mice 42. In these experiments looking at exacerbation of established fibrosis, MHV-68 infection was lytic. The pro-fibrotic actions of MHV-68 may vary greatly depending on whether the infection precedes or follows the fibrotic stimulus, and the contribution of Th2 cytokines to the pathogenesis may vary as well depending on the timing of infection.

The recruitment of fibrocytes to the lung was also associated with viral exacerbation of FITC-induced fibrosis 42. Fibrocytes are bone marrow-derived cells that share the characteristics of both leukocytes and mesenchymal cells and are defined by the co-expression of CD45 and collagen 1 17. Fibrocytes migrate in response to chemokine signals and likely contribute to fibrogenesis both via differentiation into myofibroblasts as well as through the paracrine secretion of pro-fibrotic mediators 1819878889. If MHV-68 infection is able to recruit fibrocytes, then this mechanism may be able to explain the enhancement of fibrosis noted by Lok et al. 70 when the infection preceded the administration of bleomycin by 7 days. In fact, if MHV-68 infection in the lung is associated with prolonged recruitment of fibrocytes even after the virus has established latency, this could explain why herpesviral infections may predispose persons to an enhanced fibrotic response upon a second challenge. It may also explain, in part, the altered inflammatory responses noted when mice latently infected with MHV-68 are challenged with a fibrotic stimulus (Figure 1). We have gathered preliminary data in our lab that suggest that fibrocyte accumulation in the lung is elevated for at least 30 days post-MHV-68 infection (unpublished observation). Thus, this will be an area of active future exploration. It is likely that viral infections alter epithelial cell function, cytokine profiles and inflammatory cell accumulation to promote fibrosis. Table 2 summarizes potential mechanisms to explain the ability of a viral infection to exacerbate existing pulmonary fibrosis.