Discussion
In this study, we investigated lymphocyte numbers, percentage of proliferating CD4+ and CD8+ T cells identified by Ki-67 expression, and FoxP3+CD4+ T cells putatively representing Tregs, in BALF and PB in patients with sarcoidosis at diagnosis, and in PB repeatedly during the first year after diagnosis. No significant differences were observed over time for any of the examined parameters. However, at several time points, differences were observed in PB between patients with chronic progressive (worsening pulmonary manifestations, increasing inflammatory activity and/or in need of systemic treatment), compared with patients with chronic stable sarcoidosis (remaining pulmonary infiltrates without deterioration, no signs of inflammatory activity and no need for systemic treatment), but not significantly in BALF. In PB, chronic progressive patients disclosed lower concentration of lymphocytes, an increased frequency of Ki-67 expressing CD4+ and CD8+ T cells, and a tendency towards higher percentage FoxP3+CD4+ T cells compared with chronic stable patients. Furthermore, the lower PB lymphocyte concentration, the higher expression of Ki-67 in both CD4+ and CD8+ T cells. In PB, the percentage of Ki-67+CD4+ T cells correlated positively with the percentage FoxP3+CD4+ T cells.
An association between PB lymphopenia and less favourable prognosis was reported already in the 70s.14 15 In our investigation, we found a negative correlation between percentage of PB Ki-67+CD4+ and Ki-67+CD8+ T cells on one hand, and lymphopenia on the other, indicating a depletion of PB lymphocytes despite an increased proliferation, which is in line with the hypothesis of PB lymphocytes migrating to the lung and/or extrapulmonary organs.16 23 Speaking against that theory though is the fact that we did not detect a correlation between either PB and BALF lymphocytes, or PB lymphocytes and numbers of EPM. It is well established that the lung lymphocytosis in sarcoidosis patients is due to a CD4+T cell expansion but the PB lymphopenia seems attributed to a general lymphocyte decrease involving not only CD4+T cells, but also CD8+T cells and CD19+B cells,16 which may explain the lack of correlation in our study. Furthermore, PB and BALF lymphocytes were not determined the same day, which can have led to bias. In addition, we did not actively screen for all possible EPM, thus they may be under-rated.
However, it is also possible that the PB T cell depletion is due to an increased consumption or death. One study reported an association with PB lymphopenia in sarcoidosis and under expression of certain small noncoding microRNAs involved in apoptotic pathways and genes related to lymphopenia.24
We did not investigate healthy controls in our study. However, the frequency of PB Ki-67+CD4+ and Ki-67+CD8+ T cells we found in the included patients with sarcoidosis is higher than 0.8%, which is the frequency for both Ki-67+CD4+ and CD8+ T cells reported from healthy controls.25 Thus, our finding of the lower PB lymphocyte concentration, the higher expression of Ki-67 in both CD4+and CD8+ T cells, may speak for an increased turn-over and activation of PB T cells in sarcoidosis.
In tuberculosis, another disease associated with PB lymphopenia, macrophages infected with Mycobacterium tuberculosis could induce T cell apoptosis.26 Furthermore, replicating Tregs are found at higher frequency in patients with more severe compared with patients with milder tuberculosis, and antigenic challenge is believed to induce peripheral Tregs.27 28 Similarly, also in line with our results, Broos et al reported increased proportions of PB Tregs at diagnosis in sarcoidosis patients developing chronic disease compared with patients with spontaneous resolution and healthy controls.11 Taken together, it is tempting to speculate that the increased frequency of Ki-67+CD4+ and Ki-67+CD8+ T cells, as well as the tendency to increased proportion of FoxP3+CD4+ T cells we report here, reflects a compensatory mechanism for an increased T cell death due to a persistent stimulation from an unknown antigen, especially in patients with chronic active sarcoidosis. The positive correlation between Ki-67+CD4+ T cells and FoxP3+CD4+ T cells may reflect an increased proliferation of Tregs due to peripheral induction. It also needs to be stressed that the differences we observed between chronic progressive and chronic stable patients were not significant at every measurement. Whether this reflects timely events in the course of the sarcoid inflammation or rather is a consequence of the limited numbers of included patients, not enabling us to detect differences, is yet to be determined.
Treg expansion is stimulated by TNF-α, and patients with severe and progressive sarcoidosis have higher PB TNF-α levels than patients with milder and stable disease, providing a basis for escalating expansion of Treg.29 30 However, on binding to TNF-α, FoxP3 is dephosphorylated leading to suppression of Treg function31 32 but their antiproliferative function seem intact, which is hypothesised to explain the paradoxical PB lymphopenia despite dysfunctional Tregs.9 Indeed, this is quite in keeping with that anti TNF-α therapy in patients with sarcoidosis, but not corticosteroids and methotrexate, is reported to increase PB lymphocyte counts, and decrease both PB and BALF Tregs, suggesting TNF- α being of mechanistic importance.9 33–35
Subsequently, the findings from our investigation may indicate that measurement of PB lymphocytes, Tregs, Ki-67+CD4+ and Ki-67+CD8+ T cells, can help to early in the disease course select patients that will benefit from intervention with TNF-α inhibitors.
In contrast to PB, this study did not show any significant differences between chronic progressive and chronic stable sarcoidosis patients in BALF. But there was a trend towards higher median absolute numbers and percentage of lymphocytes, higher median percentage of Ki-67+CD4+ and FoxP3+CD4+ T cells in patients with chronic progressive than patients with chronic stable disease. Thus, our results point in the same direction as previous findings associating higher BALF lymphocytes with non-resolving disease and increased proportions of BALF Ki-67+CD4+ T cells with disease activity.2 3
Besides the already mentioned lack of screening for all possible EPM, this study has some other major limitations. The relatively small study sample, some patients not participating at every follow-up, an imbalanced sex distribution, and no patients that resolved completely, limit us to draw general conclusions from the study results. Furthermore, we cannot be sure that we properly identified the Treg population using FoxP3, as no definitive surface marker that uniquely isolates Treg cells from other T cell populations exists. However, FoxP3 is essential for the function and has been widely used for identification of Tregs.36 37
Major strengths include exploration of immune cells in two compartments, that is, lung and circulation, in treatment of naïve and phenotypically well-characterised patients, following them over a long time, enabling us to correlate immunological findings with clinical outcome.
To conclude, findings from this investigation indicate that the PB lymphopenia in patients with sarcoidosis is rather due to an increased consumption, and not sequestration in organs. Patients with PB lymphopenia should be carefully monitored as they are at risk of developing a chronic active disease, anticipating a need for treatment. Measurement of Tregs, Ki-67+CD4+ and Ki-67+CD8+ T cells may help to early distinguish patients that will benefit from pharmacological intervention.
We now continue to collect patient data, also including patients under treatment with immunosuppressant and resolving patients. Hopefully, this will increase our understanding of T cell responses associated with clinical phenotypes, and enable us to more precisely select patients at risk for chronic active disease.