New laboratory criterion for the differential diagnosis of sarcoidosis and pulmonary tuberculosis

In some cases there is a problem of differential diagnosis of sarcoidosis (SD) and tuberculosis (TB) because of the similarities in clinical, X-ray and laboratory features. The aim of this study was to search for new differential diagnostic criteria for sarcoidosis and tuberculosis by calculating the index, based on the level of autoantibodies to modied citrullinated vimentin (anti-MCV) and the level of B-cell subpopulations. These parameters were measured in patients with sarcoidosis (n = 93), tuberculosis (n = 28) and healthy donors (n = 40) using the ELISA and cytometry. The absence of a statistically signicant difference when comparing the level of anti-MCV, the number of B-cells in SD and TB suggests that these changes may be characteristic of granulomatous diseases. The use of the formula Ds=([B-naïve%]\[B-memory%])*([B-CD38%]+[B-CD5%])/[anti-MCV] might allow to differentiate SD with an increase in the calculated index of more than 5 units with a sensitivity of 80.00% and specicity of 93.10% (AUC = 0.926).


Take-home Message
This study provides a new diagnostic index based on the antibodies to mutated citrullinated vimentin concentration and the level of different B subsets. Determination of index over 5 Units proved to be characteristic of sarcoidosis, what might help to differentiate SD from TB.

Background
Sarcoidosis and tuberculosis are granulomatous diseases and have many similarities in clinical symptoms, which makes differential diagnosis di cult and leads to the prescription of improper therapy. Tuberculosis is an infectious granulomatous disease caused by Mycobacterium tuberculosis and is characterized by the formation of caseous granulomas in the lungs, as well as in other internal organs and bones. Sarcoidosis is a granulomatous disease of unknown etiology, in which noncaseating granulomas in various organs and tissues are detected, most often in the lungs, mediastinal lymph nodes, and skin (1). At the moment, differential diagnosis is based on the detection of M. tuberculosis in the patient's biomaterials and on the data of histological examination, if the mycobacterium was not isolated by laboratory diagnostic methods (2). However, only in 40-60% of cases the infectious agent is diagnosed (3), and according to the results of histological examination, diagnostic errors are 40% (4).
Because of the similarity of clinical and radiological signs of both diseases, making decisions about the correct diagnosis is signi cantly complicated.
It is assumed that the presence of common features in these diseases may be due to some similarities in their pathogenesis. Despite the fact that the etiology of sarcoidosis is not yet fully understood, an autoimmune theory is being actively discussed (5). Infectious agents, for example, Mycobacterium tuberculosis, Propionibacterium acnes, Chlamydophila pneumoniae, viruses (herpes viruses HHV6 and HHV8, cytomegalovirus, retroviruses), mold (6), and environmental hazards are considered to play an important role in the development of this disease (7)(8)(9)(10). The greatest attention is given to the role of M. tuberculosis, which relationship with the disease was de ned when bacteria contents and their nucleic acids were detected in sarcoid granulomas (11,12). The immunological studies shown the presence of antimycobacterial antibodies in patients' serum (13-15) and described Mycobacteria-speci c cytotoxic T cells in patients with sarcoidosis (16,17). The possible association of Mycobacteria infection and sarcoidosis development was revealed with animal models of sarcoidosis (18).
It is important to note that patients with tuberculosis often have autoimmune complications, such as granulomatosis polyangitis, arthritis, and uveitis (19,20). The possibility of autoimmune reactions development during M.tuberculosis infections has been described in various experimental studies, showing that mycobacteria promoted an autospeci c T-cell immune response (21)(22)(23). The presence of autoantibodies in patients with tuberculosis is widely described in the literature. The high titers of antibodies to cardiolipids, β2-glycoprotein, prothrombin, proteinase-3, neutrophil cytoplasm (24), and cyclic citrulline peptides (25) were found in patients, which allude the ability of mycobacteria to induce autoimmune processes.
Vimentin is considered to be one of the possible autoantigens for sarcoidosis and tuberculosis.

Wahlström et al. identi ed vimentin autoantibodies and vimentin-speci c T cells in the bronchoalveolar
uid of patients with sarcoidosis and also described the relationship of the HLA-DRB1*03 genotype with the production of these antibodies (26). According to a serological study elevated titers of autoantibodies to modi ed citrullinated vimentin were also found in blood serum of tuberculosis patients (27).
The development of autoimmune in ammation due to M. tuberculosis could be explained with the molecular mimicry and similarity of immunogenic epitopes of bacterial and autoantigens, resulting in a cross-reaction and activation of T-lymphocytes (28)(29)(30). Despite the list of possible mycobacterial antigens involved in the cross-reaction (heat shock proteins Mtb-HsP60, Mtb-HsP65, protein p36, ESAT-6 protein, catalase (mKatG) (31), the bacterial antigen with molecular similarity to vimentin is not described yet. Perhaps this reaction may be caused by vimentin expressed by macrophages infected with mycobacteria. It was shown that due to the oxidative stress and in the presence of pro-in ammatory factors, upregulation of vimentin expression activates the NKp46 receptor of natural killers, which leads, in turn, to the lysis of infected macrophages (32,33).
In addition to the autoantibodies detection in patients with sarcoidosis, the typical changes in B-cell populations were observed (34,35). An increase in the number of naive B cells and a decrease in the number of memory B cells can be described in granulomas (36,37), blood, and bronchoalveolar uid (34,35). Similar changes have been revealed in some autoimmune diseases, such as rheumatoid arthritis and granulomatous polyangiitis (38,39). One of the con rmations of the active role of B cells in the pathogenesis of sarcoidosis is the effectiveness of rituximab in the treatment of different clinical forms of the disease (40,41).
In tuberculosis, signs of the B cells involvement in the pathogenesis of the disease were also found (42,43). According to Willem J. du Plessis, a similar picture is observed in such patients, with a reduced level of memory B cells and an increased level of naive B cells in comparison with healthy individuals. During the treatment of tuberculosis, not only the normalization of the ratio of naive B cells and memory cells is observed, but also the changing of prevailing memory B cells population from a switched class to a nonswitched class (44).
Thus, in addition to clinical and radiological similarities in sarcoidosis and tuberculosis, similar pathogenetic mechanisms can be described in these diseases, as the appearance of autoantibodies to modi ed citrullinated vimentin and changes in B-cell subpopulations. Immunological similarities leave even more questions regarding the pathogenesis of both diseases, which requires a more detailed study of the characteristics of the immune response in sarcoidosis and tuberculosis.
The aim of this study was to reveal new differential diagnostic criteria for sarcoidosis and tuberculosis by to development the new immunologic index, based on the level of autoantibodies to modi ed citrullinated vimentin and the level of B-cell subpopulations. Patients with lung sarcoidosis (n = 93), lung tuberculosis (n = 28) and healthy subjects (n = 40) were included. Healthy subjects did not have any chronic diseases, contacts with tuberculosis and positive immunologic test (ELISPOT). All study participants signed informed consent. The clinical characteristics of patients are presented in Table 1. The groups were comparable by gender and age. Exclusion criteria were: more than 2 years after the diagnosis evaluation, immunosuppressive and anti-tuberculosis therapy, plasmapheresis less than 2 months from the date of inclusion, HIV infection, syphilis, tumor diseases, decompensated diabetes mellitus.

Materials Of The Study
Methods of the study.
All patients underwent a complex of examinations, including a clinical examination, multislice computed tomography (MSCT) of the chest, laboratory blood tests, a standard set of tests for tuberculosis, histological veri cation of the lungs lesions and intrathoracic lymph nodes (obtained with transbronchial and video thoracoscopic biopsy).
The diagnosis of pulmonary sarcoidosis was made according to standard criteria of the American Thoracic Society (ATS), European Respiratory Society (ERS) and the World Association of Sarcoidosis and Other Granulomatous Diseases (WASOG). Criteria included typical radiological changes (mediastinal lymphadenopathy, disseminated foci in lung tissue); histological veri cation of lung lesions or intrathoracic lymph nodes (detection of granulomas of epithelioid cells without caseous necrosis and acid-resistant mycobacteria); the exclusion of other causes of granulomatous changes, especially tuberculosis.
The diagnosis of pulmonary tuberculosis was made using typical radiological changes (mediastinal lymphadenopathy, focal and in ltrative changes with or without destruction); positive results of a tuberculosis tests (detection of M. tuberculosis (MBT) and/or MTB DNA in sputum with molecular genetic and bacteriological methods).
Determination of the antibodies level.
In the serum of patients with sarcoidosis (n = 93), tuberculosis (n = 28), and the control group (n = 40), the level of antibodies to the modi ed citrullinated vimentin (anti-MCV) was determined using ELISA (ORGENTEC, Germany). All measurements were performed using a BIO-TEK ELx800 ELISA spectrophotometer. A positive result is the detection of the level of these antibodies was considered to be more than 19.5 U/ml. Immunotyping.

Statistical processing of results
Statistical analysis was performed using GraphPad Prism 6 (Graph Pad Software, USA), Statistica 10 (Statsoft, USA) using Mann-Whitney criterion, Spearman statistical analysis. ROC analysis was used to determine the diagnostic signi cance of the results. The differences were considered statistically signi cant at a p level of less than 0.05, diagnostically signi cant at AUC> 0.80.

Results
Testing the level of antibodies to modi ed citrullinated vimentin The level of anti-MCV was analyzed in 93 patients with lung sarcoidosis, 28 patients with lung tuberculosis and 40 healthy donors, the results are presented in Table 2. Immunophenotyping results.
An analysis of B-cell populations in the studied groups was performed and the most signi cant B cell subsets were revealed in comparison with the healthy control group. According to the classi cation of B cells by the presence of IgD and CD27 expression, B cells were divided into "naive" and memory B cells.
It was found that in patients with sarcoidosis and tuberculosis, the number of "naive" B cells were higher than in healthy donors. The results of "naive" B cells measurement are presented on Figure 2.
According to the Mann-Whitney test, the level of "naïve" B cells in patients with sarcoidosis was signi cantly higher when compared with healthy donors (p <0.0001) and patients with tuberculosis (p = 0.015). The level of "naïve" B cells of patients with tuberculosis is also signi cantly higher than the number of "naïve" B cells in control group (p = 0.007), Figure 2.
According to the results of the ROC analysis, in comparison with the control group, diagnostically signi cant results were obtained only for patients with sarcoidosis (AUC = 0.819, p<0.0001). For this group, the reference value for "naïve" B cells number was 70% within total CD19+ cells (diagnostic sensitivity was 76%, while speci city -70%). There was no diagnostically signi cant difference between the sarcoidosis and tuberculosis groups.

IgD-CD27+ memory B cells
The levels of memory B cells in patients with sarcoidosis and tuberculosis were lower than in the control group. The results of memory B cells analysis are presented on Figure 3.
According to Mann-Whitney U-test the levels of memory B cells in sarcoidosis and tuberculosis groups were signi cantly lower then in healthy control (p <0.0001 and p = 0.007, respectively). Furthermore, the signi cant difference was also found when sarcoidosis and tuberculosis groups were compared (p = 0.005), gure 3.
Using the ROC analysis, diagnostically signi cant results were determined only for patients with sarcoidosis when compared with the control group (AUC = 0.819, p <0.0001). In patients with sarcoidosis, the level of memory B cells was below 30% with a diagnostic sensitivity of 76% and a speci city of 70%. But signi cant difference between the sarcoidosis and tuberculosis groups was not shown.
The level of CD24 +++ CD38 +++ B cells was elevated in patients with sarcoidosis and tuberculosis in comparison with the control group. The results of measuring the level of CD24 +++ CD38 +++ B-cells are presented in Figure 4.

Mann-Whitney U-test showed a signi cant increase of CD24+++CD38+++ B cells patients with
sarcoidosis and tuberculosis when compared with healthy donors (p<0.001 in both cares). But, no signi cant differences between the groups with sarcoidosis and tuberculosis were found, gure 4.
According to the ROC analysis, a diagnostically signi cant increase in CD24 +++ CD38 +++ B cells relative to healthy individuals was found only in patients with sarcoidosis (AUC = 0.904, p <0.0001), the reference value was 6.52% with a diagnostic sensitivity of 91%, speci city of 88%. There was no signi cant differenceы between the groups of sarcoidosis and tuberculosis.

CD5+CD27-B cells
An increase in CD5+CD27-B cells levels of relative to healthy individuals was found in the sarcoidosis and tuberculosis groups. The results of measuring the level of CD5 + CD27-B cells are presented in Figure  5.
According to the Mann-Whitney test, statistical signi cant increase of the CD5+CD27-B cells level was found only in patients with sarcoidosis (in comparison with healthy donors (p <0.0001), with tuberculosis patients (p = 0.001)), gure 5.
The ROC analysis revealed a diagnostically signi cant increase in CD5+CD27-B cells relative to the control group of more than 12.45% for patients with sarcoidosis (AUC = 0.795, p <0.0001) with a diagnostic sensitivity of 76%, speci city of 80%. No diagnostically signi cant difference was found between the sarcoidosis and tuberculosis groups.
Thus, according to the ROC analysis of a signi cant difference in the level of anti-MCV, the number of "naive" IgD + CD27-B cells, IgD-CD27 + B-cells memory, CD24 +++ CD38 +++ B-cells, CD5 + CD27-B cells were not found. In this connection, a comprehensive analysis of the data was performed in order to develop a diagnostic index.
A search for a comparative index showed that when using the formula (1)

Discussion And Conclusions
An increased level of autoantibodies to citrullinated modi ed vimentin is shown for both patients with sarcoidosis and tuberculosis (p <0.0001), which re ects the presence of autoimmune in ammation in both diseases.
One explanation for the presence of high concentrations of autoantibodies to MCV may be the high immunogenicity of this antigen. Vimentin is a protein of connective tissue and is present in many cells (47). The development of autoimmune processes against citrullinated vimentin was described in rheumatoid arthritis, systemic lupus erythematosus, and other autoimmune diseases (47). It is known that the processes of citrullination are necessary for the arginine-containing proteins elimination (vimentin, brin, and others) by macrophages (48,49). However, in chronic in ammation, permanent tissue damage occurs, which leads to an increase in calcium concentration and hyperactivation of peptidyl-arginine deaminase, an enzyme involved in protein citrullination (50). For the carriers of the HLA-DRB1 genotypes, this process leads to the activation of autoimmune reactions against citrullinated peptides (51). A similar process has been described in the pathogenesis of rheumatoid arthritis and chronic lung diseases, in which patients nd high titers of antibodies to citrulline proteins (48).
According to our previous work, high anti-MCV titers were not correlating with alveolitis and polyangiitis (27). Given that there was no diagnostically signi cant difference in the anti-MCV levels in sarcoidosis and tuberculosis, and that the reference level of anti-MCV for both groups was less than the recommended value for the diagnosis of rheumatoid arthritis (20 Me/ml), it can be assumed that in tuberculosis autoimmune in ammation against vimentin could exist, which, under certain conditions, can become systemic and lead to the formation of sarcoid granulomas. Determination of anti-MCV levels over 12 units/ml may indicate granulomatous in ammation in the lungs, but, the determination of this biomarker can`t be used for differentiation between tuberculosis and sarcoidosis.
Immunophenotyping showed that patients with tuberculosis and sarcoidosis had reduced level of memory B cells, an increased number of "naïve" B cells and CD24 +++ CD38 +++ B cells in comparison with healthy controls. Moreover, sarcoidosis patients had signi cantly lower frequencies of memory B cell, while their peripheral blood "naive" B cells were signi cantly up regulated when compared with tuberculosis group numbers. It is important to note that in patients with sarcoidosis, the levels of CD5+CD27-B cells were statistically higher then in blood samples from healthy individuals as well as patients with tuberculosis.
Alterations in B cell subsets may be closely related to the in ammatory reactions, that are the essential part of foreign antigen elimination by immune system. However, it remains unclear is there reference meanings of the analyzed cells re ecting the presence of an autoimmune process.
According to our data, sarcoidosis and tuberculosis patients showed an increase in "naive" B cells levels and a decrease in memory B cells numbers, though, the imbalance of memory B cells and "naive" in sarcoidosis patients was more pronounced, and it was con rmed by statistical analysis. Similar results were obtained in the study of chronic sarcoidosis (35) and active forms of tuberculosis (44). According to the results of the Willem J. du Plessis, the described changes in B cell subsets were shown only for patients with tuberculosis in comparison with other lung diseases (viral or bacterial pneumonia, bronchiectasis, asthma, COPD), which did not statistically differ from the group of healthy individuals (44). According to O'Shea et al., during active tuberculosis, a decrease in memory B cells was detected, while the frequency of "naive" B cells was unaltered (52). Impaired memory and "naïve" B cells distribution have been described in some autoimmune diseases (38,39).
Given the absence of changes in the ratio of memory B cells and "naive" in other pulmonary diseases, it can be assumed that such changes characterize diseases in which granulomatous in ammation plays an important role, which is shown in some infectious and autoimmune diseases.
Next, we analyzed the frequency of CD24+++CD38+++ B cells and revealed their increased levels patients with sarcoidosis or tuberculosis versus healthy controls, but no signi cant differences were found between those two patients groups. It is known, that CD24+++CD38+++ B cells represent a population of immature transitional B cells that perform various regulatory functions (53). One of the main features is the ability to synthesize the anti-in ammatory cytokine IL-10 (54). Many studies have shown that B cells producing IL-10 could inhibit the activity of self-speci c CD4+ T cells (55). A decrease of CD24+++CD38+++ B cell levels were found in patients with rheumatoid arthritis, which allowed researchers to con rm the role of CD24+++CD38+++ B cells in preventing autoimmune reactions (56).
However, an increase in the level of these cells was noted in primary Sjogren's syndrome and systemic lupus erythematosus (57), as well as in patients in the active phase of sarcoidosis (34). In a Blair study in patients with systemic lupus erythematosus while the level of B cells with the CD24+++CD38+++ was increased, lower levels of CD38intCD24int B cells and CD24+++CD38-B cells were found (54). Perhaps an impairment of the distribution of those B cells indicates the activation of a compensatory antiin ammatory response in infectious and autoimmune diseases. In autoimmune processes at some point an activation of immune system might be replaced by decompensation, characterized by a decrease in the level of these cells.
Interesting results were obtained when comparing CD5-expressing B cells. CD5-expressing B cells also belong to regulatory B cells capable of synthesizing IL-10 and can be found in various human tissues. Moreover, CD5+ B cells are capable of producing autoantibodies (including rheumatoid factor and antibodies against ssDNA), and the number of CD5+ B cells increases in autoimmune diseases such as rheumatoid arthritis and Sjogren's syndrome (58,59). The ability of cells to produce autoantibodies was shown in mouse models; it was shown that CD5-expressing B cells belonging to a subpopulation of B1a, which is usually localized in the abdominal cavity, produce low-a nity IgM antibodies with autoreactive speci city (60). In addition, IL-10 synthesized by CD5+ B cells takes part in the control of autoimmune reactions in experimental encephalomyelitis in mice (61). In humans, CD5 is found on the cell membrane of transient CD24+++CD38++ T1 B cells (62), but according to recent studies, these cells produce a relatively low level of IL-10 compared to other transient B cell subsets (57). There is evidence that CD5 can be considered as a marker of activation of B cells in humans, and CD5 negative human B cells can be activated in vitro by incubation with phorbol or thymoma EL4 cells, followed by the appearance of CD5 molecules on their membrane (63). According to Zhang et al., tuberculosis patients can also detect elevated levels of CD5-expressing B cells. At the same time, the ability of these cells to inhibit Th17 activity was shown (64).
Our study revealed a statistically signi cant increase in the number of B cells with the CD5+CD27phenotype only in patients with sarcoidosis. At the moment, it is di cult to explain the absence of differences in the levels of CD5+CD27-В cells in patients with tuberculosis and healthy individuals, perhaps the results are associated with a small sample size.
According to the ROC analysis of immunophenotyping results, there were no diagnostically signi cant results for the differentiation of sarcoidosis and tuberculosis, however, a tendency to more pronounced changes in the level of B cells in patients with sarcoidosis is visible.
Despite the proposed laboratory parameters of the immune response (anti-MCV level, the number of "naive" IgD+CD27-B cells, IgD-CD27+ memory B cells, CD24+++CD38+++ B cells, CD5+CD27-B cells) have no diagnostic signi cance in the differential diagnosis of sarcoidosis and tuberculosis, some markers (increase in anti-MCV more than 14 U/ml, increase in "naive" B cells and decrease in memory B cells, increase in CD24+++CD38+++ B cells) may indicate the presence of granulomatous in ammation in the lungs.
At the moment, in clinical practice, the determination of the immune response is used in the diagnosis of tuberculosis. Immunological tests detect an enhanced T-cell immune response against M. tuberculosis antigens (65). The obtained diagnostic criterion allows us to differentiate sarcoidosis and tuberculosis and may indicate a predominance of autoimmune changes, characterizing the humoral immune response.

Conclusions.
To date, there is no diagnostic criterion for the differential diagnosis of sarcoidosis and tuberculosis.
However, the use of a formula we developed with a sensitivity of 80.00% and a speci city of 93.10% suggests the presence of sarcoidosis with an increase in the calculated index of more than 5 units (AUC = 0.926).
This indicator may present that changes in B cell subsets, that typical for autoimmune diseases are more characteristic of sarcoidosis.
To con rm the diagnostic signi cance of the described criteria, studies are needed on the application of the Ds formula, determining an increase in the anti-MCV level of more than 14 U ml, changing the ratio of memory B cells and "naive", increasing the level of CD24+++CD38+++ B cells in groups of patients with various autoimmune and granulomatous diseases, which will allow us to identify more accurate and speci c values for differential diagnosis.