Churg–Strauss syndrome: still a clinical challenge

Figure 1. Antigen-driven clonal expansion of CD4⁺ and CD8⁺ T cells in Churg–Strauss syndrome.

T cells produce both Th2 cytokines, which promote eosinophilia, and IFN-g, which is involved in granuloma formation. Vascular damage may depend on T-cells, ANCA- or eosinophil-mediated cytotoxicity. ANCA: Antineutrophil cytoplasmic antibodies; DC: Dendritic cell; HMGB: High mobility group-1 protein; MPO: Myeloperoxidase; TEFF: T-cell terminally differentiated; TEM: T cell effector memory; TRAIL: Tumor necrosis factor-related apoptosis-inducing ligand.

History

The original description of Churg–Strauss syndrome (CSS) was based on an autopsy series of 13 patients who died from severe asthma [1]. All patients had fever, hypereosinophilia and a spectrum of terminal events, including cardiac failure, pericarditis, peripheral neuropathy, renal failure, as well as cutaneous and gastrointestinal involvement. Churg and Strauss proposed the syndrome of ‘allergic granulomatosis and angiitis’ as distinct from the traditional periarteritis nodosa (PN). They considered the specificity of the syndrome on the basis of clinical (asthma, eosinophilia, fever and vasculitis involving many organs, particularly heart, peripheral nerves, skin, the lungs and kidneys) and, mainly, pathological features. Characteristic histopathologic changes were infiltration of extravascular tissues with eosinophils, a characteristic granuloma known as the ‘allergic granuloma’, and vasculitis involving arteries and possibly veins. Vasculitis had some changes similar to the changes observed in PN, that is necrotizing lesions and fibrosis, but here eosinophils could be seen in and around vessels and, specifically, a granulomatous response could be observed within vessels in the majority of cases.

It is clear that the patients described by Churg and Strauss represent the worst evolution of the syndrome, which actually had a poor prognosis, before corticosteroid therapy was available. It is also clear that the Churg and Strauss criteria, based largely on a post-mortem series, could not serve today as a guide for diagnosing the syndrome in clinical practice. To this aim, Lanham et al. [2] proposed that diagnosis could be obtained on clinical basis, pointing to the recognition of signs of vasculitis in two or more organs in patients at high risk for the syndrome, that is asthmatics with rhinosinusitis and eosinophilia. Lanham et al. [2] pointed out in their paper that the three histological components, previously described by Churg and Strauss, often do not coexist in space (i.e., in the same organ) or in time (i.e., at the time of biopsy). Each organ is affected differently by the vasculitis. As an example, granulomas are infrequently found in the respiratory tract. In the following years, CSS has been defined by American College of Rheumatology (ACR) [3], Chapel Hill Consensus Conference (CHCC) [4] and, more recently, by a panel of experts [5].

All these definitions were not intended for diagnostic purposes but just to distinguish individual forms of vasculitis from each other. The use of these criteria for diagnostic purposes has some potential problems because the capacity of any diagnostic criteria set to identify a patient with a disease is better when disease prevalence is high than when it is low [6]. The positive predictive value of the ACR criteria improves as the prevalence of specific types of vasculitis increases. However, if ACR criteria were applied in a setting in which the prevalence of vasculitis is very low, such as a primary care, or even in a population of asthmatic patients, the likelihood of identifying individuals who meet ACR vasculitis criteria but do not have vasculitis would be high. For example, a patient with asthma, eosinophilia and sinusitis with extravascular eosinophil infiltration in a nasal biopsy specimen will be classified as having CSS according to ACR criteria (four positive criteria out of six).

In 1994, the CHCC produced definitions for vasculitis but, again, they were not intended as diagnostic criteria [4]. CHCC recognized that histological data would not be available for all patients, especially when the clinical condition of the patient might preclude obtaining appropriate biopsy samples or when the sample might not be representative and might miss salient histological features. Therefore, the concept of surrogate markers of vasculitis was introduced, but a list of markers was not provided.

Very recently, a stepwise algorithm, based on Lanham’s, ACR, CHCC criteria and surrogate markers, was developed by consensus between experts, with the aim of using these criteria and definitions for epidemiological studies [5]. As far as CSS is concerned, experts suggest to use Lanham’s or ACR criteria in order to separate CSS from the other vasculitides. In a few cases, patients clinically diagnosed with other vasculitis might be classified as CSS. Once again, even this new algorithm may not help clinicians to decide whether their patients with the eosinophilic phenotype of asthma associated with rhinosinusitis are going to develop CSS when the same patients showed, for example, pulmonary infiltrates. It would be quite possible that, applying ACR criteria, an asthmatic patient with allergic bronchopulmonary aspergillosis or with associated eosinophilic pneumonia received the diagnosis of CSS, whenever extravascular eosinophil infiltration should be present in transbronchial biopsy specimen. In both cases, four ACR criteria would be fulfilled: asthma, rhinosinusitis, eosinophilia and pulmonary infiltrates, which actually were due to either mucoid impaction and atelectasis or alveolar filling by eosinophils and not to vasculitis.

Diagnostic challenge

We think the true clinical problem with CSS is not to differentiate this syndrome from other vasculitis, but to diagnose the onset of the syndrome in those asthmatics with rhinosinusitis and eosinophilia, who report symptoms and/or signs suspect of vasculitis. The finding of antineutrophil cytoplasmic antibodies (ANCA) may help to define different phenotypes of the disease. ANCA have been reported in approximately 40% of CSS patients and their presence has been associated with a higher frequency of necrotizing crescentic glomerulonephritis, constitutional symptoms, purpura, alveolar hemorrhage and moneuritis multiplex [7]. Vasculitis was documented less frequently in histologic specimens of ANCA-negative patients, in whom perivascular and tissue eosinophil infiltration is more prominent.

According to Lanham et al. [2], the natural history of the disease can be divided in three main, and usually successive, phases: the prodromic phase, consisting of the eosinophilic phenotype of asthma associated with rhinosinusitis; the eosinophilic phase, resulting from eosinophil infiltration into tissues, particularly the lung, myocardium and gastrointestinal tract; and the systemic phase, with the development of necrotizing vasculitis and also affecting the skin, peripheral nerves and kidneys.

The eosinophilic phase of the disease is the most challenging phase from the clinical diagnostic point of view. In this phase, patients may report symptoms suggestive of vasculitis, such as abdominal pain, for which an obvious explanation is not found. Endoscopy is often near-normal and bioptic samples may reveal histopathological pattern consistent with gastrointestinal eosinophilic disease, associated with a diffuse interstitial eosinophilic infiltration. We know that gastrointestinal eosinophilic disease may cause important symptoms, depending on the site of eosinophilic infiltration, such as heartburn, emesis, abdominal pain and diarrhea [8]. What we do not know is whether or not these patients are going to develop catastrophic abdominal complications, such as bowel perforation, due to the development of vasculitis. Therefore, what we need in these cases would be surrogate markers of vasculitis, which would help clinicians to distinguish patients with an eosinophilic phenotype of asthma associated with rhinosinusitis and eosinophilic digestive disease from those with CSS. The same diagnostic challenge arises when asthmatic patients report limb paresthesias [9] or present chest x-ray pulmonary infiltrates with eosinophilia in bronchoalveolar lavage and/or eosinophilic tissue infiltration in transbronchial bioptic samples, but without the other two pathologic markers of disease, granulomas and vasculitis. An important clue to the diagnosis of vasculitis may be the neurophysiologic study showing mononeuritis multiplex [10], as well as a low value of exhaled nitric oxide in an asthmatic patient presenting pulmonary infiltrates due to alveolar hemorrhage [11].

The definite answers to the important and challenging diagnostic questions may come from research on specific markers, as well as on the pathogenesis of the syndrome.

Promising markers of the disease are those referring to survival of eosinophils, such as the expression of tumor necrosis factor-related apoptosis-inducing ligand receptor 3, a decoy receptor that acts as an antiapoptic receptor, which is significantly higher on eosinophils from patients with CSS compared with asthmatics and patients with parasitic infection [12]. Other markers that have been shown to discriminate CSS patients from asthmatic patients are the serum levels of VEGF [13] and high mobility group-1 protein (HMGB1) [14], which has potent inflammatory and proangiogenic properties.

Pathogenetic hypothesis: the role of T lymphocytes

We think that an improvement in diagnostic assessment of patients with suspected CSS could come from a better understanding of its pathogenesis. Eosinophils and autoantibodies (ANCA) certainly play a role that, we suggest, may be orchestrated by T cells. Increases in concentration in serum of eosinophilic cationic protein (suggesting eosinophil activation) and soluble thrombomodulin (indicating endothelial cell damage) have been shown to correlate with soluble IL-2 receptor, suggesting T-cell activation [15]. Impairment of CD95 ligand-mediated apoptosis of lymphocytes and eosinophils has been observed to be associated with oligoclonal T-cell expansion [16].

A central pathogenic role for T lymphocytes in CSS is also suggested by the following observations. First, T cells are commonly present in biopsies of active lesions, both vasculitic and granulomatous [17]. A study of the clinicopathological features of neuropathy associated with the syndrome showed that CD8⁺ and CD4⁺ T cells outnumber eosinophils in neuronal biopsy specimens, suggesting a T-cell-mediated damage of epineural vessels [18]. Second, T-cell lines from CSS patients, generated in vitro by polyclonal stimulation, are polarized towards a Th2 phenotype (IL-4 and IL-13), which may explain the eosinophilia, the hallmark of the disease [19]. The same cells have also been shown to release significant amounts of IFN-γ, which may be related to the granuloma formation. Third, immunosuppressive therapy targeting lymphocytes is useful in CSS patients, strengthening the concept of CSS as a T-cell-driven disease. Moreover, T cells from perinuclear ANCA-positive patients have been shown to proliferate in response to myeloperoxidase, suggesting their helper function in the production of autoantibodies [20]. Inhaled allergens, vaccination, drugs, desensitization and/or infections have been implicated in triggering some cases of CSS [21,22], resulting in lymphocytes activation via antigen or superantigen (SAg) stimulation.

Lymphocyte activation by SAgs has been suggested to occur in Kawasaki disease, PN and microscopic polyangitis [23]. SAg stimulation of T cells bearing specific Vβ T-cell receptors (TCRs) results in the polyclonal expansion of selected Vβ families in CD4⁺ and CD8⁺ cells, followed by T-cell-restricted deletion. Conversely, antigen stimulation may cause oligoclonal T-cell expansions, which have been found in CSS patients by immunoscope technique [16]. We found one or more expanded Vβ families in eight CSS patients [24]. These expanded Vβ families were found mostly in CD8⁺ lymphocytes with a late memory phenotype. Each CD8⁺/Vβ⁺ expanded family is the result of the proliferation of a clonal T cell, bearing the same TCR, as detected by analysis of TCR rearrangement. These data are more consistent with a persistent antigenic stimulation that has been reported to occur in some cases of CSS and other autoimmune diseases [25,26], compared with a superantigenic stimulus, as reported in other vasculitides [23].

We found that the CD8⁺/Vβ⁺ expanded populations had an effector memory (TEM) or a terminally differentiated (TEFF) phenotype, as demonstrated by flow cytometric analysis (CD45RA^-/CD62L^- and CD45RA⁺/CD62L^-, respectively). While central memory T cells are increased after acute viral infection, a chronic antigen exposure drives a selective shift towards TEM or TEFF cells [27], which have the propensity to populate peripheral tissues as well as inflammatory sites. We can speculate that in CSS, circulating CD8⁺/Vβ⁺ TEM or TEFF are antigen-driven cells, which may infiltrate target organs and participate in the inflammatory cascade of both vasculitic and granulomatous lesions (Figure 1). In Wegener’s granulomatosis, for example, the driving force of both granuloma formation and autoimmunity has been identified in TEM cells [28]. In agreement with this hypothesis, we found that a sinus bioptic sample from one patient with CSS was infiltrated by T cells bearing the same clonal TCR rearrangement that was previously detected in peripheral CD8⁺/Vβ⁺ TEM lymphocytes [24]. A more detailed analysis of T cells infiltrating damaged tissues in CSS patients is clearly needed to confirm this hypothesis.

In conclusion, the diagnosis of CSS may be difficult because the syndrome arise as a common association between asthma, rhinosinusitis and eosinophilia. Asthma itself may be associated with occasional pulmonary infiltrates (i.e., mucus plugging and atelectasis) and corticosteroids dependency, so that a clear diagnosis of CSS might not be made until evidence of vasculitis in one or more organs develops. Unfortunately, there are no validated markers of the disease. We suggest that the detection of clonal-expanded peripheral late-memory T-cell populations (assessed by flow cytometry combined with analysis of TCR gene rearrangement, two methods routinely used in immunopathology laboratories) might be useful in the early diagnosis of CSS, as suggested by the negative findings in asthmatic control patients.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.