Despite the short-term success of lung transplantation, long-term outcomes remain among the worst of all routinely transplanted solid organs. The lung, perhaps because of direct environmental interactions and innate host defenses, appears uniquely vulnerable to alloimmune injury despite modern immunosuppression. Over 50% of lung transplant recipients experience acute rejection within the first year (1). Furthermore, a majority of patients develop bronchiolitis obliterans syndrome (BOS), a condition of progressive allograft dysfunction, within five years after transplant (1).
Although the precise etiology of BOS remains elusive, it was often thought a manifestation of chronic alloimmune injury. This was consistent with the observation that prior acute rejection predicted subsequent BOS. However, a more complex view is emerging, in which non-alloimmune factors like primary graft dysfunction, infection or gastroesophageal reflux also appear to contribute to the risk of BOS (2). Given the lack of reliable treatment, prevention of BOS becomes paramount to improving outcomes after lung transplantation. The identification of risk factors for the onset of BOS has therefore been the focus of intense research in the lung transplant community.
Most reports on the role of infections in BOS have focused on viruses, such as cytomegalovirus or community-acquired respiratory viruses like influenza and parainfluenza. Viral infection can induce cellular immune responses which cross-react with the allograft or increase tissue inflammation and upregulation of histocompatibility antigens in lung epithelia promoting alloimmune responses. However, the exact mechanism by which infection might contribute to BOS and the connection of other infectious agents to lung allograft rejection has remained controversial.
In this issue of AJT, Weigt et al. add another piece to the associative puzzle of infection and lung rejection (3). The authors retrospectively analyzed 171 lung transplant recipients with sufficient PFT data transplanted between 2000 and 2006. They identified 48 patients with Aspergillus colonization prior to the onset of BOS, and compared outcomes to the 122 non-colonized patients. Using appropriate time-dependent Cox models, the authors show a clear association of Aspergillus colonization with the subsequent onset of BOS, with colonization being an independent predictor of morbidity and mortality due to BOS regardless of acute rejection scores.
There are several limitations to this study, which make the results intriguing rather than confirmatory. First, the analysis did not attempt to determine if multiple colonization episodes or increased fungal burden increased the risk for BOS; if present, such a dose response relationship would have strengthened the overall hypothesis. Also, due to the retrospective study design, patients with Aspergillus colonization underwent more bronchoscopies than their counterparts. Thus, we are ultimately not certain whether Aspergillus is a true risk factor for BOS, or whether increased sampling due to another underlying problem contributed to differences in colonization rates. Finally, because of variation in treatment and infrequent followup bronchoscopies the authors could not assess the effects of Aspergillus eradication upon BOS.
Nevertheless, this novel finding has significant implications. Placed in the context of previous work, the study by Weigt et al. suggests that the relationship between lung infection and BOS may not be limited to specific pathogens, but rather depend on the host response to a broad spectrum of pathogen-associated motifs. The paradigm best fitting this response may involve innate immune pattern recognition through toll-like receptors (TLRs). For example, Aspergillus is recognized through TLR2 and TLR4; in addition, other specific TLRs recognize viral or bacterial pathogen associated molecular patterns. We have previously shown that pulmonary immune activation by lipopolysaccharide, a TLR4 ligand, promotes alloimmune lung injury in a murine model (4) and that hypo-responsive TLR4 genotypes protect lung transplant recipients from the onset of BOS (5). If innate immune activation is sufficient for the induction of BOS, this would explain the observation by Weigt et al. that Aspergillus colonization, even in the absence of frank infection, was enough to induce BOS. Placed in an even broader context, non-infectious injury to the allograft could also fit within this paradigm. For example, in primary graft dysfunction tissue injury leading to the release of endogenous innate immune ligands like short-fragment hyaluronan could contribute to the development of BOS even in the absence of infectious triggers (6).
If Aspergillus colonization is confirmed in further studies as a risk for BOS, these results have intriguing implications for clinical practice. Should all lung transplant recipients receive prophylactic azole antifungals? Such an approach would have to be weighed carefully against the cost, risk for hepatotoxicity, and potential to select for highly resistant fungal infections. More importantly and generally, if innate immune activation can lead to BOS, is then modulation of innate immunity a viable candidate for the prevention or treatment of BOS? The present paper by Weigt et al. does not answer these questions. Rather, it offers intriguing food for thought and further impetus to rethink our mechanistic understanding of BOS.