Friederich Wegener’s original paper “On generalised septic vessel disease” suggests he thought it likely there was an infectious cause for the condition which now bears his eponym.1 The characteristic pathological features of Wegener granulomatosis (WG) are: a necrotising granulomatous inflammation of the respiratory tract with vasculitis affecting both arteries and veins; focal necrotising glomerulonephritis; and a varying degree of systemic vasculitis—the so-called “Wegener’s triad”.2 The granulomatous inflammation is conspicuous for the absence of any obvious microorganism, although granulomatous infections can sometimes be misdiagnosed as WG.3 During the 70 years since Wegener’s description there have been some remarkable advances in both the diagnosis and treatment of this condition and in our understanding of its pathogenesis. However, the precise nature of the initiating factor(s) remains elusive.

What we do know is that there is a strong association between WG and the human leucocyte antigen (HLA)-DPB1*0401 allele, suggesting that there is an inherited predisposition for the condition.4 Interestingly, there is also an association with α-1 antitrypsin deficiency.5 We also know that virtually all patients who subsequently develop the systemic disease have circulating antineutrophil cytoplasmic antibodies (ANCAs) and these are mainly directed against proteinase-3 (PR3), the so-called “Wegener’s autoantigen”—a serine protease which regulates cell proliferation, differentiation and death.6 This antibody has proved to be a useful biomarker for the diagnosis of WG,7 8 although less so for the likelihood of relapse.9 Furthermore, our increasing knowledge of its biological properties has provided new insights into the pathogenesis of WG. Specifically, whilst PR3 is the target antigen for ANCAs, the observation that elevated levels of PR3 at sites of granulomatous inflammation correlate with increased tumour necrosis factor α (TNFα) has led to the hypothesis that PR3 is directly involved in the modulation of cytokines associated with an aberrant immune response (see Csernok et al10 for a review).

Following tissue injury, increased levels of PR3, released from activated or dying neutrophils, might act as a danger/alarm signal interacting with proteinase-activated receptor-2 (PAR-2), the so-called “gateway receptors” on dendritic cells. This would result in their maturation and instruction to initiate an innate T helper 1 (Th1) immune response. However, PR3 also has further effects on inflammation mediated by two distinct actions on interleukin 32 (IL32), a recently identified cytokine which induces production of TNFα, IL1β, IL6 and 2CXC chemokines by immune cells.10 It is known that PR3 activates IL32 by enzymatic cleavage, thereby enhancing its proinflammatory effects.11 However, PR3 is also known to be a specific binding protein for IL32α, independently of its enzymatic activity.12 It thus has the ability both to activate and to inhibit IL32, and the latter could represent an important negative feedback mechanism designed to limit adverse effects of prolonged inflammation. It therefore follows that inhibition of this feedback, for example by ANCAs, could result in an IL32-mediated, uncontrolled inflammatory reaction which may be relevant to the vasculitis of WG.

Further evidence that PR3–ANCA interactions may be directly related to the pathogenesis of WG comes from a recent in vitro study of human oral, lung and kidney epithelial cells. These cells do not normally secrete proinflammatory cytokines when stimulated with pathogen-associated molecular patterns (PAMPs). However, in the presence of either anti-PR3 antibodies or serum from patients with WG, such cells release IL6, IL8, monocyte chemotactic protein-1 (MCP-1) and TNFα.13 ANCA also causes dysregulated neutrophil apoptosis resulting in increased inflammatory mediator release.14

Another, but not yet final, piece of this rather elaborate jigsaw relates to speculation on what might trigger the switch from localised granulomatous inflammation in the upper respiratory tract to the life-threatening systemic vasculitis and, further, what might cause relapse of such vasculitis in patients whose condition has been stabilised with anti-inflammatory treatment.

The report by Richter et al in this issue of Thorax (see page 692) is the first study to investigate the extent to which the lower respiratory tract of patients with WG is colonised by microorganisms.15 They compare these patients with both normal healthy volunteers and a group of patients with idiopathic pulmonary fibrosis (IPF). Interestingly, they demonstrate that the lower respiratory tract of patients with WG in relapse or remission (but not at first presentation) are frequently colonised by Staphylococcus aureus and to a lesser extent other organisms. Their findings are consistent with previous studies which have related relapse rates of WG to nasal carriage of this organism.16 17 Furthermore, they show by careful, serial bronchoalveolar lavage that this colonisation is transient rather than persistent. It seems unlikely that such colonisation merely relates to structural damage in the lungs since they found no apparent relationship between the presence of S aureus and radiological evidence of bronchiectasis. The authors’ preferred explanation comes from their further observation that the lung epithelial lining fluid of patients with WG is rich in IL1 receptor antagonist (IL1Ra), probably generated by IL32 via mechanisms described above. This molecule is known to be permissive of the growth of S aureus as has been noted in patients with acute respiratory distress syndrome (ARDS),18 19 supporting a role for a direct relationship between an inflammatory alveolar microenvironment and bacterial growth.

This report therefore provides evidence to support the hypothesis that bacterial colonisation of the lower respiratory tract may trigger a relapse of WG. It is also consistent with the seminal observation by DeRemee that treatment with the antimicrobial trimethoprim/sulfamethoxazole (co-trimoxazole) combination has a beneficial effect on the clinical course of WG, particularly in its localised form.20 21 Finally, it supports the notion that bacteria in the lower respiratory tract may play a wider role as initiators or progression factors in other pulmonary diseases of unknown cause. The inclusion of a group of patients with IPF is of particular interest in this regard.

As its name suggests, IPF remains a disease of unknown aetiology. However, there is accumulating evidence that there may be a link between infections, particularly Epstein–Barr gammaherpesvirus, and at least some cases of lung fibrosis.22 The idea that infectious agents contribute to the initiation and/or progression of IPF is not new; indeed considerable effort has been put into the identification of viral DNA in the lungs of patients with IPF, and increasing evidence supports the hypothesis that viral infection may be an important factor.23 24

Much less is known about the role of bacteria in this context. By using a group of patients with IPF as a “disease control”, the study by Richter et al also provides new evidence for transient bacterial colonisation of the lungs in patients with IPF. It is therefore interesting to speculate whether this may be one, neglected factor, responsible for the alveolar epithelial injury observed in IPF. Certainly, recurrent “microinfections” would be consistent with the observed temporal and spatial heterogeneity of lung fibrosis seen in usual interstitial pneumonia. The observation that a patient with IPF given empirical treatment with co-trimoxazole (for presumed Pneumocystis jiroveci) showed considerable clinical improvement led to a recent, small, pilot study of this antibiotic in patients with pulmonary fibrosis.25 The results are encouraging and a bigger trial is currently underway in the UK to examine this observation further.

It is tempting to speculate that in both WG and IPF, co-trimoxazole treatment may exert a beneficial effect by reducing bacterial colonisation and, thereby, halting the, downstream cellular and molecular interactions which ensue. Whatever agent(s) initiate and/or perpetuate these conditions remains elusive. The study by Richter et al is a timely reminder that the complex relationship between microorganisms and the respiratory tract has yet to yield all its secrets.