The heterogeneity of the different entities that constitute diffuse interstitial lung disease (ILD) means that a sequential and systematized study with a multidisciplinary diagnostic approach1–3 is required to achieve a definitive diagnosis. Several factors should be considered in the initial presentation of a patient with diffuse ILD, as these entities may share similar clinical, functional or radiological characteristics.2 In fact, in previously undiagnosed ILD, an underlying cause, such as systemic autoimmune diseases (SAD) must always be ruled out.

For this reason, the application of tools that add useful information to the ILD diagnostic algorithm is essential. Autoantibodies are primary biomarkers in the diagnosis of SAD-ILD.4 These antibodies may support a suspected diagnosis and, in some cases, help predict disease progression.5 Autoantibodies are usually studied in serum, but they can also be isolated in other biological matrices and appear to synthesize in situ in multiple tissues.6,7 Starting from this premise and taking into account the description of germinal centers in the pulmonary epithelium of patients with SAD,8 we hypothesized that autoantibodies might also be detected in bronchoalveolar lavage (BAL) in patients with ILD. The main objective of the study was to identify autoantibodies to extractable nuclear antigens (ENA) in the BAL of patients with suspected ILD. A secondary objective was to associate the presence of these autoantibodies in BAL and serum with patients’ clinical, radiological and respiratory parameters.

This study shows the presence of autoantibodies in the BAL of patients with ILD during the diagnostic process. As far as we know, this is a new finding not previously described. Positivities were detected for 7 different ENA specificities, anti-Ro52 being the most frequent. BAL studies revealed high baseline levels of autoantibodies in all patients, regardless of the ANA titer in serum.

ENA is detected in BAL using an inexpensive, widely available technique. Our results show that these determinations, combined with the serum study, can provide additional information for screening ILD of autoimmune origin. Detection using different techniques (immunoblot, IIF) confirms that this is a reproducible strategy that could be incorporated into routine diagnostic procedures.

BAL analysis is widely used in the diagnostic algorithm of ILDs,2,25 and can reflect the humoral and immunobiological mechanisms that occur in the lung in a more reliable and predictable manner than peripheral blood.26 However, information on biomarkers of ILD in BAL is still limited. Most of the molecules studied are associated with the aberrant activation of epithelial cells or alveolar macrophages and, to a lesser extent, a modified immune response.8,27,28

The appearance of autoantibodies in serum can occur years before the first clinical manifestations.1,12 Identifying the organ in which autoimmunity begins could open the door to new diagnostic or therapeutic strategies. In this series, all patients were studied at the onset of pulmonary manifestations (clinical or radiological), so we propose that the appearance of autoantibodies in BAL could serve as an early marker of ILD development in patients with SAD. In contrast with idiopathic interstitial pneumonias, the study of autoantibodies in SAD-ILD is essential, since their presence is directly related to pathogenesis and tissue damage.28,29

Three patients showed autoantibodies in BAL that were undetectable in serum, while 3 others had serum autoantibodies that were not detectable in BAL. These disparities may reflect differences in local response. Autoantibodies detected in BAL could only derive directly from the lung, while those found only in serum could be synthesized in other tissues. These findings support the need for studying both matrices in parallel.

The fact that no positive results were obtained in another group of common ILDs, including idiopathic pulmonary fibrosis or hypersensitivity pneumonitis, may also indicate that the detection of autoantibodies in BAL is highly indicative of underlying SAD. In this regard, it should be noted that 1 patient with silicosis in the group with positive BAL also had positive serum. Although this result may seem contradictory, exposure to silica dust has been described as a risk factor for the development of several types of SAD.30,31 It should also be noted that autoantibodies were not detected in the control group or in patients with diseases that could be confused clinically or radiologically with onset ILD (congestive heart failure, infection in immunosuppressed patients, and others).2

When the group of patients with positive BAL was compared with those diagnosed with SAD or IPAF and negative BAL, we found no significant differences in clinical or functional parameters. Nevertheless, we believe that it might be interesting to analyze the progress of these patients to assess the role of autoantibodies in the prognosis of these diseases.

The ENA most frequently detected in the BAL of these patients was anti-Ro52. Anti-Ro52 is found in several SADs, including Sjögren’s syndrome, systemic lupus erythematosus, scleroderma, and others. Its presence in serum has been associated with a higher prevalence of ILD32 and with a worse prognosis when it coexists with myositis-specific autoantibodies.32–34 In our series, patients with positive anti-Ro52 in BAL were mostly women and presented different diagnoses and widely varying clinical and radiological manifestations, ranging from asymptomatic cases that were investigated due to a radiological pattern consistent with ILD to patients whose initial manifestations included acute respiratory failure, alveolar hemorrhage, or signs of acute interstitial pneumonia on high-resolution computed tomography. The appearance of ILD is of interest in patients with systemic lupus erythematosus, in whom the prevalence of this clinical finding is generally lower than in other SAD.13 However, when it occurs, it can present as an acute complication, as occurred in our cases. These results are in line with studies suggesting that patients with positive anti-Ro52 in serum may have a highly heterogeneous clinical phenotype.35

In the group of patients with positive BAL, autoantibodies with high specificity and clinical relevance were also detected, particularly Jo-1, SM/RNP and Scl-70. These autoantibodies are included in the classification criteria for various SADs, such as antisynthase syndrome, mixed connective tissue disease and diffuse sclerosis, respectively. These markers can also provide prognostic information.29,35,36

No standardized protocols are currently available for immunological studies in ILD screening. Many laboratories study only ANA by IIF, and if these are negative, ENA specificities are not investigated. In our study, the isolation of ENAs in both BAL and serum also supports the recommendations made by other authors to expand serum autoantibody testing to other matrices. In addition to myositis markers, anti-Ro52, that is highly prevalent and lacks a characteristic ANA pattern in serum (it may be negative), should be routinely studied in these patients.4,29,35,36

Another unexpected finding is the higher prevalence of IgG antibodies over IgA antibodies. IgA antibodies are dominant in mucous membranes and the concentration of IgG in the lung is very low. However, we have found patients who are positive for IgG antibodies and negative for IgA antibodies. This apparent incongruity could be explained by a lack of sensitivity of the polymeric IgA conjugate. It could also be because the polymeric IgA antibodies are strongly bound to the mucous layer and do not penetrate the liquid phase that is collected in the BAL. In peripheral blood, IgA isotype immunoglobulins have a half-life of a few days, so they are less prevalent.37 For this reason, it is highly unlikely that the IgA antibodies detected in our patients are a result of contamination with peripheral blood. Likewise, despite being scant, their mere presence demonstrates synthesis in situ. This would support the hypothesis posited in other studies that, in at least some patients, the lung is an organ involved in autoimmunity mechanisms.26,38 Our results, that open the door to future research, would need to be replicated in functional studies.

This study has some limitations. First, it has no validation cohort to determine the predictive value of the test. Since the primary objective of the study was to detect the presence of autoantibodies in BAL, our findings answer the research question. Second, the patients in the control group were not healthy volunteers and some had a history of smoking. Although BAL is a minimally invasive technique, it can be an unpleasant experience for some patients, and, as such, is difficult to justify in a healthy population. Finally, although the study includes patients from 2 centers, the group of patients with SAD is too small for adequate statistical analysis. However, we assume that it reflects real clinical practice in ILD units in our setting, where a high percentage of patients are referred for follow-up of a previous diagnosis of SAD.

In summary, this study shows that autoantibodies can be detected in BAL in patients with early-onset ILD using conventional techniques. The parallel analysis of autoantibodies in serum adds valuable information to the diagnostic algorithm for these diseases. This approach may be useful for the early identification of interstitial involvement in SAD patients in the different possible scenarios.

This study has been funded by the following grants: research grant from the Spanish Society of Pulmonology and Thoracic Surgery (SEPAR): project 020, EPID Futuro 2014 grant, SEPAR international research fellowship 2018 (Karina Portillo), FIS P16/0216 P16/0216 (Marian García-Nuñez), and SEPAR‐Boehringer EPID 2018 Grant for Young Researchers (Laura Rodríguez Pons).

The authors state that they have no conflict of interest.