Lung in chronic autoimmune diseases and hypersensitivity – how to separate these from idiopathic pulmonary fibrosis


 Background

Lung involvement in autoimmune disease (AID) is not uncommon, and may precede other organ manifestations. The histologic pattern is variable and difficult to interpret. Another problem was recently encountered: chronicity presenting with fibrosing pneumonia. This resulted that patients with chronic AID and usual interstitial pneumonia (UIP) were excluded from antifibrotic therapy. Creating a new category of UIP with autoimmune features (IPAF) for patients with AID enabled treatment for these patients. This however raised the opinion, that a pathological investigation for the underlying disease is not necessary.
Results

We retrospectively evaluated 66 cases of AID, 31 cases of hypersensitivity pneumonias (HP), and 10 clinically confirmed usual interstitial pneumonia/idiopathic pulmonary fibrosis (UIP/IPF) cases. 12 additional cases could not be assigned into any of these categories. Acute AID presented with lymphocytic interstitial pneumonia, immune complex deposition, cytotoxicity for specific cell compartments, and alveolar hemorrhage. Subacute patterns most often presented with organizing pneumonia, whereas chronic pattern most often presented as UIP. Granulomas were present in some patients. UIP pattern was the most common presentation in chronic AID and HP, whereas NSIP was rather rare.
Conclusion

The most important, statistically significant feature differentiating chronic AID or HP, from UIP/IPF are lymphocytic infiltrations into myofibroblastic foci. Other features such as granulomas, Langhans giant cells, and protein deposits were significantly associated with AID and HP, but were not encountered in all cases. This study demonstrated that the morphological analysis can uncover the possible etiology in many cases, and a more specific diagnosis can be provided to the clinicians.


Background
Autoimmune diseases (AID) are heterogenous, some of them affecting joints, others involving blood vessels, and some with quite unspeci c clinical symptoms [1][2][3]. They all have in common a deregulation of the immune system resulting in auto-aggression against normal tissues [4][5][6][7][8]. In some diseases, a clinical diagnosis is usually straight-forward as they present with a typical clinical picture: e.g. of skin affection in lupus erythematosus (LE), joint in ammation in rheumatoid arthritis (RA), or glomerular disease in systemic LE or granulomatosis with polyangiitis/Wegener's granulomatosis (GPA). Most of these AIDs can affect the lung. In a minority of cases lung involvement can precede the classical symptoms delaying or impeding the diagnosis [9]. This can result in an undetected and untreated pulmonary disease, which might progress into chronic AID. Hypersensitivity pneumonia (HP) in the acute stage is easily diagnosed, clinically by an exposure anamnesis and typical undulating symptoms associated with exposure to the allergen, and morphologically by loose peripherally concentrated epitheloid cell granulomas combined with lymphocytic interstitial pneumonia (LIP) dominated by CD8 + Tcells. Due to increased sensitivity of CT scan and awareness of brosing pneumonia more biopsies are performed, which results in an increasing detection of chronic AID and HP presenting with lung brosis.
Subacute and chronic AID and HP in lung can present with three classic patterns: usual interstitial pneumonia (UIP), non-speci c interstitial pneumonia (NSIP), and organizing pneumonia (OP). Rarely unspeci c brosis of the lung is seen in this setting [1,[10][11][12][13][14][15]. UIP is the typical pattern of idiopathic pulmonary brosis (IPF), a deadly brosing pneumonia with incompletely understood etiology. Some underlying mechanisms have been reported, which might be correlated to morphologic features: continuous toxic injury by tobacco smoke toxins inducing apoptosis of pneumocytes,, repair by myo broblasts, and regeneration, during which some epithelial cells differentiate into senescent cells [16][17][18][19]. Factors associated with aging are mutations of telomere maintenance genes [20][21][22] (hTERT, telomerase), affecting the regeneratory capacity, regulators of homeostasis and in ammation [23][24][25] (surfactant apoprotein genes), contributing to prolonged in ammation, and MUC5B [26,27]. Mutation in the promoter of MUC5B might be responsible for a reduced mucociliary clearance and probably also disrupts regeneration in the peripheral lung parenchyma [28]. Senescent cells release in ammatory cytokines, stimulate the proliferation of myo broblasts, and prolong the repair process, nally resulting in brosis. What patterns in IPF might be expected from these underlying mechanisms? Toxic injury caused by tobacco smoke products will cause apoptosis of pneumocytes, a denuded basal lamina, and a proliferation of myo broblasts -in ammatory cells likely will be absent, and only be seen in cystic remodeled areas, due to bacterial colonization. The proliferation of the myo broblasts is driven by the senescent cells [19].
What about UIP in AID and HP? An UIP pattern has been reported in chronic AID and chronic HP. These chronic stages often clinically cannot be attributed to a speci c AID, and a new term, interstitial pneumonia with autoimmune features (IPAF) has been created [29]. In chronic HP the causing allergen can often no longer be identi ed. Senescence seems to play also a role in these immune diseases [19]. If UIP in IPF can be differentiated from UIP in chronic AID and HP has not been evaluated in larger studies, but it might be of therapeutic and prognostic relevance [1,12,14,[30][31][32]. There are divergent reports about survival of chronic AID or HP with UIP pattern versus IPF [14,[31][32][33][34][35]. In both AID and HP immune cells play a major role by inducing injury to the peripheral lung. As both diseases are, similar to IPF, stepwise progressing, additional features enabling a more detailed and etiology-based diagnosis should be reported. Here we aimed to analyze the different patterns of chronic AID and HP in a retrospective series, and compare this to IPF, to identify features which might allow an etiology-based diagnosis. We believe, a more detailed pathological report over just an UIP pattern report will help in a better strati cation of patients and concise discussion within the multidisciplinary team [36].

Methods
Out of 103 cases with autoimmune diseases and 148 cases of chronic HP, 78 cases of AID and 31 cases of HP were retrieved from lung tissue archive of the Institute of Pathology, Medical University of Graz. A total of 119 tissues samples from patients with immune diseases and IPF were selected, for which a clinical diagnosis and clinical response was received either before or after the pathological report was issued. These cases were predominantly consultation cases submitted to one of the authors. All 10 IPF cases came from one institution and were clinically veri ed (see Suppl. Table 1 for patient characteristics and clinical data). De nition of patterns UIP was de ned by the presence of myo broblastic foci, brosis, cystic remodeling of lung lobules, uninvolved normal lung parenchyma, and peripheral accentuation (temporal and spatial heterogeneity). NSIP pattern was de ned by uniform widening of alveolar septa by a mixed lymphocytic and histiocytic in ltration (cellular NSIP) without remodeling, or brosis of alveolar septa with residual in ammatory in ltrates ( brosing NSIP) again without remodeling. Organizing pneumonia (OP) was de ned by intraalveolar granulation tissue with newly formed capillaries and in ammatory in ltrates predominantly macrophages, but sometimes with residual granulocytes. LIP, a dominant pattern in immune disorders, is characterized by a diffuse monomorphic in ltration of lymphocytes and plasma cells, with a few scattered immunoblasts, with or without follicular hyperplasia of BALT (bronchus associated lymphoid tissue). In contrast to cellular NSIP, cells of the innate immune system are scarce or absent in LIP. Other features of AID and HP are histiocytic or epitheloid granulomas, amyloid or immune complex deposition with/without complement activation, isolated BALT hyperplasia, lymphocytic bronchitis/bronchiolitis, follicular bronchiolitis, and hemorrhage, often mixed in different proportions. Amyloid was de ned by a positive Congored stain with green birefringency, and in addition by immunohistochemistry for amyloid A or P. Immune complex deposition was veri ed by a positive immunohistochemical reactions for IGG and activation of complement components 1q, 3c, and 5-9 complex.
The tissues were obtained either by video-assisted thoracoscopic or cryobiopsies. Hematoxylin and Eosin (H&E) stained slides were available in all cases. In several cases immunohistochemistry (IHC) for lymphocyte subtypes was performed, using antibodies for CD3, CD4, CD8, CD20, and rarely for FOXP3 (regulatory T cells). All cases were re-evaluated by HP and ES. The presence or absence of the following patterns were recorded for each case: myo broblastic foci, cystic remodeling of the peripheral lung lobules, spatial and temporal heterogeneity, brosis, lymphocytic in ltrates (including LIP), lymphocytes in ltrating the myo broblastic foci, hyperplasia of bronchus associated lymphoid tissue (BALT), histiocytic and/or epitheloid granulomas, isolated Langhans giant cells, amyloid deposition, immune complex deposition, complement activation, vasculitis, hemorrhage, vasculopathy (myxoid changes of the intima with few scattered lymphocytes), vascular sclerosis, neutrophils, eosinophils. Additional features not generally associated with brosing pneumonias were also recorded (e.g., bronchial and/or bronchiolar in ammation and pleuritis).
For statistical analysis cases were classi ed as AID, HP, and IPF. The program Kaleidagraph (Synergy Software, v4.5) was used. A signi cance was stated, if the p-value was ≤ 0.05.

Study cohort
There were altogether 66 cases with acute, subacute, and chronic AID, 31 cases of subacute or chronic HP, and 10 con rmed UIP/IPF cases. 12 cases could not be assigned into any of these categories (

Histological Findings in autoimmune disease
In 44 cases an UIP pattern was seen, in two cases combined with OP; in 16 cases UIP was combined with LIP pattern. In 12 cases LIP without UIP was found, in six cases combined with OP (within this group there were acute AIDs); OP alone was seen in 5 cases, in two cases unspeci c interstitial brosis was seen. In four cases NSIP, one brosing, and three cellular were seen. In one case only BALT hyperplasia was found ( Table 1).
Out of these 66 cases of AID a more speci c diagnosis could be established: RA in12, systemic sclerosis SSc in 11, SLE in5, Sjogren's disease in4 cases, and one case each for Behcet disease, Raynaud's syndrome (secondary form), dermatomyositis, and Good Pasture syndrome. In 26 cases AID was clinically con rmed, but without a more speci c diagnosis. In one of the cases a combination of SSc and SLE was discussed and in another case either Sjogren's disease or RA was discussed as a possible diagnosis. There were two cases with an antiphospholipid antibody syndrome. In both a diagnosis of an AID was suggested, and as both presented with alveolar hemorrhage, systemic vasculitis or LE was suggested.

Findings in hypersensitivity pneumonia
In 22 of 31 cases of HP an UIP pattern was seen, which was combined with LIP in 12 cases. Four cases presented with LIP only (acute or recurrent HP), in 5 cases LIP was combined with OP; in 9 cases a differential diagnosis of either AID or HP was rendered, three of them presented with OP, airway centered interstitial brosis (ACIF), or unspeci c interstitial brosis, respectively ( Table 1).

Differences of AID versus HP by Immunohistochemistry
In cases of HP and AID, where a su cient number of lymphocytes could be evaluated, a predominance of CD8 + lymphocytes were seen in HP, whereas in AID CD20 + Bcells, CD4 + , and CD8 + Tcells were present; whereas scattered FOXP3 + regulatory Tcells were seen in HP, these were scarce in AID (Table 2). Findings in IPF All ten cases of IPF showed the typical UIP pattern without pronounced in ammation (Table 2). In 3 additional cases a differential diagnosis of either AID or IPF was rendered -these were cryobiopsy derived tissue samples.
Different patterns in AID, HP, and IPF ( see Tables 2 and 3)  The presence of lymphocytes in ltrating myo broblastic foci in AID (Fig. 1) was statistically signi cant different from IPF (p < 0.0001, Table 2), whereas no signi cant difference was seen when AID was compared to HP (p = 0.92). Next, we compared the presence of histiocytic/epitheloid cell granulomas and/or Langhans giant cells in AID versus IPF (Fig. 2), which also showed a signi cant difference (p = 0.028), as granulomas were absent in all IPF cases (Table 2); however, granulomas were present in a minority of AIDs. Granulomas were more often found in HP compared to AID, therefore the presence of Langhans cells or granulomas (Fig. 3) favor HP over AID (p < 0.001). Isolated histiocytic giant cells were more common in HP compared to AID. The presence of hyperplasia of BALT (Fig. 4) was in favor of AID or HP (p = 0.001 and p = 0.023, respectively) when compared to IPF, whereas no signi cance was seen when AID and HP were compared (p = 0.09; Table 2). There was no deposition of amyloid or autoimmune complexes (Fig. 5) in IPF. When the presence or absence of deposits was compared between AID and HP, this was signi cant (p = 0.004), because deposits were not seen in HP (Table 2). However, only a small number of AID cases presented with deposits, which explains the p-value. Other features such as chronic bronchitis/bronchiolitis, vasculitis, hemorrhage, and degenerative changes of the intima (vasculopathy) were seen in few cases, therefore due to the small numbers, a statistical analysis was not performed.
In more than half of the cases a more speci c diagnosis of AID could be suggested to the clinicians (Table 3AB). Examples were cases of RA presenting with combinations of UIP, LIP, granulomas, and amyloid deposits (Fig. 5, 6). A combination of arterial thrombosis, immune complex deposits, and hemorrhage, sometimes combined with pleuritis pointed to SLE (Fig. 7). In acute juvenile SLE an LE phenomenon could be seen in pleuritis (Fig. 7). An LIP pattern combined with OP and lymphoepithelial lesions with predominant CD8 + Tcells favored Sjøgren's disease (Fig. 8). Finally, UIP combined with LIP, hyperplasia of BALT, and myxoid changes of the intima of pulmonary arteries was suggestive of SSc ( Fig. 9). Good Pasture disease could be suggested due to alveolar hemorrhage, brosis, and the proof of linear deposits of immunoglobulin at the basal membrane of alveolar septa and capillaries combined with complement activation. A pathological diagnosis of anti-phospholipid syndrome and secondary Raynaud's syndrome was not possible; in Behcet disease a rheumatoid disorder was suggested because of large brotic areas, few ill-formed epitheloid cell granulomas, amyloid, and scattered lymphocytic in ltration (Fig. 10).

Discussion
Sorting brosing pneumonias into UIP, NSIP, and OP are usually performed in pathological laboratories. The pattern report is further discussed in a multidisciplinary team composed of clinicians, radiologists, and pathologists, proposed by international guidelines [37,38]. These recommendations suggested not to perform biopsies, when clinical and radiological features were in favor of UIP/IPF. But, new studies showed that radiologic features of UIP are not speci c enough to make a diagnosis of IPF. Wright et al evaluated 23 cases, and found that features of peribronchiolar metaplasia and giant cells or granulomas were in favor of HP and excluded IPF -features which can only be evaluated in biopsies [39]. One third of their cases could even not be solved by a multidisciplinary team discussion. Similarly Churg pointed to other features favoring HP over IPF, such as upper-lobe predominance, giant cells or granulomas, and peribronchiolar metaplasia [40]. Not much reports have focused on AID /connective tissue diseases, although patterns seen in chronic AID are very similar to IPF [40]. Even genetic alterations of surfactant and telomerase genes have been reported in AID similar to IPF [41,42]. Whereas an NSIP pattern almost exclusively is associated with AID or HP, an UIP pattern has a wider range of differentials. Positive antinuclear antibodies (ANA) are not always present in chronic AID, and CT scans might also not always discriminate IPF from AID [43]. If chronic AID with UIP pattern has a better overall survival remains uncertain.
In this study we demonstrated, that a pathological analysis of patterns in addition to UIP, NSIP, or OP can add more information about the underlying etiology. Lymphocytes in ltrating myo broblastic foci are indicators of an immune disease, either AID or HP. Other combinations of UIP with granulomas and/or histiocytic giant cells, hyperplasia of BALT, and protein deposits are also characteristic for immune disorders. Based on these ndings the pathological report can exclude IPF without knowing radiologic and clinical data. The issued report can further strengthen the discussion in a multidisciplinary team to narrow down the differential diagnosis, by including distribution patterns, clinical presentation and laboratory data.
Even more, by including speci c patterns in some of the AIDs the pathologic report can provide suggestions for one of the AIDs, as we have shown in roughly 40% of these cases (Table 3). RA with lung involvement can present with LIP in the acute form, but with UIP or NSIP in the chronic and subacute form, often combined with LIP in the former. Granulomas are seen especially in seropositive forms of RA.
In chronic RA deposition of immune complexes as well as amyloid is common. Very large idiotypicantiidiotypic immune complexes with granulomatous reaction are most often encountered in RA less in SLE [6,44,45]. Therefore, if a combination of these patterns is seen in a biopsy RA can be suggested. Subacute or chronic SLE in the authors experience rarely present with UIP, but often a combination of OP, unspeci c brosis, thrombosis, and deposition of immune complexes is seen. In SLE these antigenantibody complexes usually activate complement, which can be proven by IHC. SSc is known for its high numbers of autoantibodies and circulating immune complexes. In our cases an UIP pattern with hyperplasia of BALT, and vasculopathy of the pulmonary arteries, e.g. a concentric deposition of myxoid material and scattered lymphocytes within the intima of pulmonary arteries was commonly seen and did allow the suggestion of SSc. Sjøgren's disease is characterized by an aggressive lymphocytic in ltration of the mucosa of salivary and lacrimal glands, and also a similar in ltration of the mucosa of bronchi and bronchioles, mimicking lymphoepithelial lesions as seen in MALT lymphomas. A combination of OP and LIP with a CD8 + Tcell in ltration (in our cases) and lymphoepithelial lesions are suggestive for Sjøgren's disease.
Chronic HP can in some instances be differentiated from AID. Protein deposits either immune complexes or amyloid help to rule out HP, whereas scattered giant cells and ill-formed granulomas favor HP. In contrast to other reports, we could not see any signi cant difference between AID and HP with respect to BALT hyperplasia [46]. NSIP and OP patterns were relatively rare in our cases. NSIP and an isolated OP pattern in all cases were associated with AIDs, similar to previous reports [47]. However, OP pattern was combined with UIP in some cases, and more important LIP was present in several cases with an immunebased etiology.
Since several years cryobiopsies are preferred over open lung biopsies in diagnosing interstitial lung diseases, including brosing pneumonias. These biopsies are useful in all cases, where a clinical and radiological diagnosis already favors IPF, AID or HP. However, in those cases with unusual clinical and radiological pattern videothorascopic biopsy is superior and should be considered if the cryobiopsy is not diagnostic. The distribution of patterns as described above can be very focal, involving different lung segments or even lobes. This is also seen on CT scans, where honeycombing might be seen focally, while ground glass opacities are present in other foci. A larger piece of tissue will increase the likelihood of sampling all these different patterns.

Conclusion
Here we have shown that a pathological analysis can provide more than just diagnosing IPAF, which does not even include chronic HP. In our diagnostic workup we should try to be as speci c as possible, and dig for the underlying etiology. In some cases, we might be able to even discern chronic HP from AID. This might provide additional clues for treatment. Even in those cases, where the pathological report can only exclude IPF this would enable clinicians to investigate additional treatment options.
A multidisciplinary team discussion is recommended by international guidelines [48]. In the view of the authors a biopsy should be performed not only in cases, where CT scan ndings are inconclusive [38]. Lynch recommended a re-review of IPF cases on a regular basis, as the diagnosis might change. However, a pathological analysis can provide much more information and might early on help clinicians to better stratify their patients. In our consultation praxis we have seen cases, initially diagnosed as IPF by CT scan and clinical presentation, which later on turned into chronic AID.

Declarations
Page 13/16 Ethics approval and consent to participate: A consent was not required as the patient data were anonymized. The study was approved by the Ethical Committee of the Medical University of Graz (EK 24-135 ex 11/12).

Helmut Popper can submit the manuscript in behalf of all authors
All data are provided in the method and result section.