Lung CD4+ Vα2.3+ T-cells in sarcoidosis cohorts with Löfgren’s syndrome

An accumulation of CD4+ T-lymphocytes in bronchoalveolar lavage fluid (BALF), resulting in an increased CD4/CD8 ratio (> 3.5), has been shown to associate with sarcoidosis [1]. Further, an expansion in BALF of CD4+ T-cells expressing T-cell receptor (TCR) Vα2.3 gene segment has been identified in HLA-DRB1*03+ sarcoidosis patients, usually with Löfgren’s syndrome (LS) i.e. with an acute disease onset with bilateral ankle arthritis and/or erythema nodosum (EN), bilateral hilar lymphadenopathy (BHL), sometimes combined with parenchymal infiltrates and usually fever [2]. Dr. Sven Löfgren, described in his study from 1946 different causes of EN and where joint symptoms were reported in 149 of 185 patients, distributed fairly evenly over all etiological groups [3]. Within the cohort he found a group of patients with BHL, subsequently named to have “the bilateral hilar lymphoma syndrome”, today Löfgren’s syndrom [4]. Later, EN has been reported to be significantly more common in women with sarcoidosis [5], whereas joint symptoms without EN is seen preferentially in men [6].

In sarcoidosis today, bronchoalveolar lavage (BAL) is commonly used as a diagnostic method and in this study we aimed at investigating the usefulness of measuring the proportion of CD4+ Vα2.3+ T-cells, as an additional diagnostic test to the CD4/CD8-ratio, by evaluating the sensitivity and specificity for sarcoidosis of such BALF expansions.

Out of 749 patients with sarcoidosis, 268 subjects were identified with BALF CD4+ Vα2.3+ T-cells expansions. Out of these, 193 had LS. In contrast, all except three of the 108 controls had normal proportions of CD4+ Vα2.3+ T-cells in BALF, see Table 1. One of the individuals with a high expression in BALF was a healthy control without signs of sarcoidosis or any other pulmonary disease on chest X-ray, but with a CD4/CD8-ratio of 4.4. The other two were patients investigated because of suspicion of sarcoidosis with radiological NSIP, one with unknown cause and one with dermatomyositis, both with low CD4/CD8-ratio of 0.4 and 0.2.

A proportion of CD4+ Vα2.3+ T-cells in BALF > 10.5% was found to be highly specific for sarcoidosis, with a specificity of 97% but with a sensitivity of 36% (p < 0.0001), calculated on all 749 sarcoidosis patients, including LS as well as non-LS patients, and 108 controls. ROC curves show that testing for CD4+ Vα2.3+ T-cells in BALF was a more useable test in individuals with LS [area under the curve (AUC) 0.82, p < 0.0001] compared to the whole patient group (AUC 0.64, p < 0.0001).

For comparison, an elevated CD4/CD8-ratio was seen in 203 of the patients with LS, in 332 with non-LS and in 9 controls (all of them healthy volunteers). The specificity for CD4/CD8-ratio > 3.5 was 92% and the sensitivity 73% (p < 0.0001). ROC curves for the CD4/CD8-ratio showed a high diagnostic capacity in both the complete patient group with sarcoidosis (AUC 0.89, p < 0.0001) and in patients with LS (AUC 0.92, p < 0.0001). In patients with LS there was a positive correlation between the CD4/CD8-ratio and the percentage of CD4+ Vα2.3+ T-cells in BALF, see Fig. 2. The proportion of non-LS patients who had expansion of CD4+ Vα2.3+ T-cells in BALF (> 10.5%) or elevated CD4/CD8-ratio was distributed in a similar manner in different radiographic stages, see Fig. 3.

We set out to evaluate if an expansion of CD4+ Vα2.3+ T-cells per se in BALF could be used as an additional diagnostic tool for sarcoidosis. As shown, healthy controls and patients with other pulmonary diseases which constitute differential diagnosis to sarcoidosis only rarely seem to have such expansions. The analysis of this particular T-cell subset add value especially in sarcoidosis cohorts where Löfgren’s syndrome is frequently occurring and/or where HLA-DRB1*03 is common e.g. in Scandinavian countries. However, there were also patients with non-LS who had elevated levels of this specific subset of cells although less frequently. We suggest therefore that analyzing the proportion of CD4+ Vα2.3+ BAL T-cells in patients with sarcoidosis could be valuable as a complement to the CD4/CD8-ratio. A CD4/CD8-ratio > 3.5 is a more useful test as ROC-analysis show. However, increased CD4+ Vα2.3+ T-cells in BALF (> 10.5%) is highly specific for sarcoidosis (specificity 97%). As a result, in cases with normal or only marginally increased CD4/CD8-ratio but elevated CD4+ Vα2.3+ T-cells in BALF, further invasive procedures can be avoided.

A lung accumulation of CD4+ Vα2.3+ T-cells is well known and characteristic for sarcoidosis. The resulting increase in the CD4/CD8 ratio, in particular when > 3.5, has been suggested to be of help in diagnosing sarcoidosis [1]. The association between a CD4/CD8-ratio > 3.5 and sarcoidosis was also confirmed in this study, which to our knowledge is the largest patient cohort showing this. The results are consistent with a metanalysis of 999 patients and 886 controls [11]. A number of serological tests have also been proposed to reflect the activity and support the diagnosis of sarcoidosis. One of these is angiotensin-converting enzyme (ACE) which has been extensively investigated since it was first reported to be associated with sarcoidosis in 1975 by Lieberman [12]. The sensitivity was in a study including 101 patients diagnosed with sarcoidosis 62% and the specificity 76% [13]. However, ACE may be normal in early stages of the disease. Also, neopterin was found to be elevated in serum and to correlate with ACE [14] but for both substances the most important value seems to be when sequential analyses are made. According to a study on 232 sarcoidosis patients by Bargagli et al. [15] the enzyme chitotriosidase is also correlating to ACE and showed for sarcoidosis a high sensitivity (88.6%) and specificity (92.8%). In a study by Popevic et al. [16] the corresponding values for sensitivity were 82.5% and for specificity 70%. Serum-ACE in the latter study had 66% sensitivity and 54% specificity. Soluble IL-2 receptor has been shown by Bargagli et al. [17] to correlate to chitotriosidase and to be a useful marker of disease progression. Eurelings et al. [13] reported a sensitivity of 88 and 85% specificity in a total of 101 patients diagnosed with sarcoidosis. Several other attempts have been made to identify biomarkers. One of them is KL-6 that showed a sensitivity of 78% and seems to reflect the more severe disease stages well [18]. Another is serum amyloid A which is known to be involved in inflammatory reactions and immune disorders [19]. It has been reported to be increased in sarcoidosis and suggested to be valuable especially at follow-up. Taken together, no serological biomarker has so far proven ideal and in this setting we believe that CD4+ Vα2.3+ T-cells could provide additional valuable information in diagnosing and in some cases monitoring sarcoidosis. The Va2.3+ cells could perhaps be a component in the future of suggested multi-assays as proposed by Ramos-Casals et al. [20]. A weakness of the present study is the composition of the controls. We did not include patients with e.g. tuberculosis and only one with lymphoma since we seldom see patients with these diagnoses at the lung clinic. Studying the CD4+ T-cell subgroup that express Vα2.3, we found particularly in LS patients a strong association with the diagnose and a capacity to separate sarcoidosis from controls. In non-LS this finding was also existing but weaker.

The accumulation of CD4+ Vα2.3+ T-cells in the lung indicates the presence of specific lung-antigens, that are recognized by CD4+ Vα2.3+ T-cells, leading to a specific immune response in the lung. The function of these lung accumulated T-cells is unknown, although recent data show a preferential pairing with TCR Vb22, and a high degree of clonality, implicating these cells to interact with an as yet unknown antigen [21]. This is also in line with the finding that the percentage of CD4+ Vα2.3+ T-cells in BALF normalizes as the disease resolves [22]. The hypothesis has been further strengthened by the discovery of T-cells clones that are identical between different individual patients, as well as by a strong association with HLA-DRB1*03, which may be considered to present specific antigens for these T-cells [2]. The good prognosis in patients with CD4+ Va2.3+ T cells may indicate that precisely these T-cells for unknown reasons are particularly effective in their immune response to and elimination of such a potential antigen. Thus, the percentage of CD4+ Vα2.3+ T-cells in BALF may also serve as a prognostic marker as the proportion of patients with resolving disease increases gradually with increased proportions of CD4+ Vα2.3+ T-cells in BALF [23]. LS patients have the highest levels of CD4+ Vα2.3+ T-cells as well as the best prognosis and are in addition found to have very few extrapulmonary manifestations (EN and ankle arthritis excluded) [24].

How a presumed antigen is presented to the T-cells in sarcoidosis seems to play a role for the disease course. There is a correlation with HLA-DRB1*03 and subtypes of *13 and elevated levels of CD4+ Vα2.3+ T-cells in BALF [23]. Furthermore, the majority of patients with LS have either HLA-DRB1*03 and/or DRB3*01, in turn associated with good prognosis [25]. This phenomenon may explain why CD4+ Vα2.3+ T-cells is of best diagnostic use in LS patients. The other T-cell subsets that cause elevated CD4/CD8-ratio in patients with other HLA-types need to be investigated in further studies.

In conclusion the results in this study indicate that Vα2.3 can be an additive diagnostic marker to the CD4/CD8-ratio for sarcoidosis. A low value does not exclude sarcoidosis, but an increased percentage of > 10.5% strengthens the diagnosis. In comparison, the CD4/CD8-ratio > 3.5 had a lower specificity for sarcoidosis but a higher sensitivity. Whether Vα2.3 may be of clinical value also in other populations than Scandinavian needs to be analyzed in further studies, as well as the significance in other granulomatous disorders.