Flow cytometric analysis of lymphocyte profiles in mediastinal lymphadenopathy of sarcoidosis

Lymphocyte profiles in mediastinal lymph nodes may reflect the immune status of patients with sarcoidosis. Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is useful for the diagnosis of diseases with mediastinal lymphadenopathy including sarcoidosis. The purpose of this study was to determine lymphocyte profiles of lymph nodes in sarcoidosis by analyzing EBUS-TBNA samples. We prepared single cell suspensions from EBUS-TBNA samples of mediastinal lymph nodes from patients with sarcoidosis or lung cancer and analyzed surface markers (CD3, CD4, CD8, CD19, CD25) and FoxP3 expression in the resultant lymphocytes using flow cytometry. We studied 26 patients with sarcoidosis and 16 with lung cancer with mediastinal lymph node metastases. In sarcoidosis, the CD4/CD8 ratio was significantly more elevated in lymph nodes than in bronchoalveolar lavage fluid (P<0.001), although both were strongly correlated. The CD4/CD8 ratio was significantly higher in stage I than in stage II both in the BAL fluid and lymph nodes. When compared with lung cancer lymph node metastasis, the CD4/CD8 ratio was significantly higher in sarcoidosis, whereas the CD3/CD19 ratio was significantly higher in lung cancer. The proportion of regulatory T cells (CD4+, CD25+, FoxP3 high) did not differ between sarcoidosis and lung cancer samples. Lymphocyte profiles in mediastinal lymphadenopathy can be analyzed by flow cytometry of EBUS-TBNA samples. These findings might help elucidate the immunopathology of sarcoidosis.


Introduction
Although CD4-positive T-lymphocytes are thought to play an important role in sarcoidosis, its pathogenesis has not been fully elucidated, although the excessive activation of CD4 positive T lymphocytes is a hallmark of sarcoidosis [1]. An elevated ratio of CD4/CD8 positive T-lymphocytes in bronchoalveolar lavage (BAL) fluid is suggestive for sarcoidosis diagnosis [2]. However, little is known about lymphocyte dynamics in the affected lymph nodes in a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 sarcoidosis. Although previous reports reported the lymphocyte profile in sarcoidosis lymph nodes including the CD4/CD8 ratio, the results were inconsistent among these reports [3,4]. With recent advances in respiratory endoscopic technologies including endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), the diagnostic accuracy for various mediastinal or hilar lymphadenopathies has been markedly improved. EBUS-TBNA samples are available for histological diagnosis and for molecular diagnosis or immunohistochemical analyses [5,6,7]. Recent studies showed gastrointestinal endoscopic ultrasound guided fine needle aspiration (FNA) biopsy is an established procedure for the diagnosis of lymphoma [8,9,10]. FNA-biopsy samples can be used for flow cytometric analysis, which is essential for the diagnosis of lymphoma. The purpose of this study was to analyze lymphocyte profiles in sarcoidosis lymph nodes by analyzing EBUS-TBNA samples using flow cytometry. We investigated the ratio of T-lymphocytes/B-lymphocytes, CD4/CD8 T-lymphocytes and regulatory T cells (Treg)/CD4 T-lymphocytes in affected lymph nodes of sarcoidosis patients. Comparative analyses of lymphocyte profiles in BAL fluid and lymph nodes were also performed. To elucidate the specific lymphocyte profile of sarcoidosis, we also compared the lymphocyte profile between sarcoidosis lymph nodes and lung cancer metastatic lymph nodes.

Subjects
From November 2017 to March 2018, 52 patients underwent EBUS-TBNA for the diagnosis of mediastinal lymphadenopathy in Fujita Health University Hospital. Among these patients, written informed consent could not be received from four patients and one patient refused to participate in the study. Thus, 47 patients were included in this study. Five patients with a final diagnosis that was not sarcoidosis or lung cancer metastases were excluded. Finally, 26 cases of sarcoidosis and 16 cases of lung cancer with mediastinal lymph node metastases were analyzed. The patient characteristics are shown in Table 1.
The ethics committee of Fujita Health University approved this study protocol, which was conducted in accordance with the tenets of the Declaration of Helsinki (approval number FH-14-035). The registration number of this study is UMIN 000021850. Written informed consent was obtained from all participants in this study.

Bronchoscopy
For EBUS-TBNA, we used a convex probe endobronchial ultrasound (Olympus UC260F-W-OL8, Olympus Co Ltd, Tokyo, Japan) and dedicated fine needles (NA-201SX-4022, 22 gauge, Olympus, Tokyo, Japan). We selected the target lymph node station according to the chest CT or positron emission tomography (PET)-CT findings. More specifically, among #2 (upper paratracheal), #4 (lower tracheal), #7 (subcarinal), and #10 (hilar) lymph nodes, an enlarged lymph node (short axis > 1 cm) with high uptake of 18 F fluorodeoxyglucose (FDG) was selected according to the physician's choice [11]. Under a real-time echoic view, TBNA was performed five to six times for each patient. We did not use rapid onsite evaluation. When we performed bronchoscopy on patients suspected of sarcoidosis, we performed bronchial alveolar lavage (BAL) immediately after EBUS-TBNA. BAL procedures and the interpretation of specimens were carried out according to the standard method of the American Thoracic Society [12]. In brief, after a bronchoscope (Olympus 260 or Olympus 1T-260, Olympus Ltd. Tokyo, Japan) was wedged into the target segment bronchus, 50 ml of saline was administered to the target segment (mostly middle lobe or lingular lobe) via the working channel three times. Instillation and retrieval of saline were performed manually using a 50 ml syringe. Retrieved BAL fluid was filtered and collected.

Analyses of lymphocyte profile by flow cytometry
Two to three pieces of EBUS-TBNA specimens were collected and used for flow cytometry. TBNA samples were pushed out from the puncture needle into high glucose Dulbecco's modified Eagle medium (DMEM) containing 5% fetal bovine serum (FBS). Phosphate buffered saline (PBS) was added to the samples, mixed by inversion, and centrifuged. The cell pellet was washed and centrifuged twice. Then the pellet was minced on a petri dish with a scalpel blade. The minced samples were dispersed into a single cell suspension with repeated pipetting. For hemolysis treatment, (EasyLyse Agilent Technologies, Inc., Santa Clara, CA, USA) was added to samples and stirred, followed by incubation at room temperature for 15 minutes according to the manufacturer's instruction. After washing again with PBS, samples were stirred and filtered. The cell count of each sample was calculated by the Trypan Blue exclusion method. We prepared a single cell suspension sample that contained 1-5×10 7 cells/ml from each sarcoidosis sample, and 5-10×10 5 cells/ml from each lung cancer sample. Each dispensed sample was labeled with anti-CD3-PC5 + anti-CD4-FITC + anti-CD8-PE antibody (Beckman Coulter Inc., CA, diluted 1:10), anti-CD3-FITC + anti-CD19-PE antibody (Beckman Coulter Inc., diluted 1:10), or anti-CD4-FITC + anti-CD25-PE + anti-FoxP3-APC antibody (Miltenyi Biotec Inc., Bergisch, Gladbach, Germany, diluted 1:10). For FoxP3 labeling, cell fixation and permeabilization treatments were performed. For the analyses of BAL specimens, a mixture of monoclonal anti-CD3-PC5, anti-CD4-FITC, and anti-CD8-PE antibodies (BD Multitest, BD Biosciences, San Jose, CA) were used for flow cytometry. We checked cell counts and fractionation on each BAL sample. For each analysis using different antibody combinations, 1-3×10 4 cells were counted. All flow cytometric analyses were performed using FACS Calibur (BD Biosciences).

Statistical analysis
For the analysis of data from two related samples, the P value was calculated by Wilcoxon signed-rank test. The Mann-Whitney U-test was used to compare two independent groups of data. Regarding the correlation between two groups of corresponding data, the P value was calculated by Spearman's rank correlation coefficient. Linear regression analysis was also used to calculate r 2 . All statistical analyses were performed using JMP (SAS Institute, Tokyo, Japan).  Table 2. Flow cytometry clearly differentiated between CD3+ and CD19+ cells (R2 and R3 respectively, Fig 1). The analysis also clearly identified CD4 T-lymphocytes (R4, Fig 1) and CD8 T-lymphocytes (R5, Fig 1) in this sample. We also identified CD4+/CD25+ lymphocytes and FoxP3 highly positive cells in this population. We determined the CD4/CD25 positive and FOXP3 high lymphocyte population as Treg (Fig 1, S1 Fig). In sarcoidosis, lymphocytes in lymph nodes consisted of 56.6% CD4+ cells (SD: 13.8%), 9.60% CD8+ cells (SD: 5.1%), and 29.5% CD19+ cells (SD 15.0%) on average.

Lymphocyte profile in sarcoidosis lymphadenopathy
The CD4/CD8 ratio was significantly higher in lymph nodes than in the BAL fluid (p < 0.0002, Wilcoxon signed-rank test). CD4/CD8 ratios in the lymph nodes and BAL fluid were significantly correlated (p < 0.0001, Spearman's rank correlation coefficient) (Fig 2). Tanriverdi et al. reported that the optimal cut-off value of CD4/CD8 to distinguish from other interstitial lung diseases was 1.34 and that a cut-off value of CD4/CD8 > 3.5 has high specificity (more than 95%) [2]. When we applied both values (1.34 and 3.5) to our BAL CD4/CD8 ratio in sarcoidosis patients, 92.3% and 61.5% of patients fell above the cut-off value, respectively. Interestingly, the CD4/CD8 ratio in lymph node samples was significantly higher in Stage I than in Stage II (p = 0.014, Mann-Whitney U-test) (Fig 3). The CD4/CD8 ratio in the BAL fluid of Stage I was also higher than in Stage II but the difference between the two stages was smaller compared to the difference observed in lymph nodes. As shown in Table 2, all sarcoidosis patients (except one with a history of autoimmune pancreatitis) had not received systemic steroid therapy. Thus, the CD4/CD8 ratio difference between Stage I and II in our study was independent of steroid therapy. Furthermore, the CD4/CD8 ratio in patients with or without extrapulmonary lesions was 6.43 ± 2.51 and 7.96 ± 4.33, respectively (mean ± SD). There was no significant difference between the two groups (p = 0.089, Mann-Whitney U-test).

Comparative analysis of the lymphocyte profile of lymphadenopathy between sarcoidosis and lung cancer samples
Finally, we compared the lymphocyte profile in sarcoidosis lymph nodes with those of lung cancer lymph nodes (five cases with small cell carcinoma, nine with adenocarcinoma, one with squamous cell carcinoma and one with "not otherwise specified" (NOS)). The detailed clinical information of lung cancer patients is shown in Table 3. As shown in Fig 4, the CD3/CD19 (Tlymphocytes /B-lymphocytes) ratio was markedly different between the two diseases. In sarcoidosis, the relative B-lymphocytes proportion was significantly higher than that of lung cancer metastatic lymph nodes. In addition, the CD4/CD8 ratio was significantly higher in sarcoidosis (p < 0.001, Mann-Whitney U-test) (Fig 4, Table 4). In contrast, no significant difference was observed in the ratio of Treg /CD4 T-lymphocytes between the two diseases (Fig 4, Table 5). As a larger number of younger patients was included in the sarcoidosis group, we assessed the influence of patient age to the T-lymphocyte profile. We compared the CD3/ CD19, CD4/CD8, and Treg/CD4 ratios in lymph nodes from age-matched patients (60-79 years old) from two groups (sarcoidosis and lung cancer). As shown in S2 Fig, the results were the same as those from the analyses of all patients. The CD4/CD8 ratio was still higher in sarcoidosis than lung cancer even when we compared patients whose ages were 60-80 years (sarcoidosis 9.31

Discussion
With the use of EBUS-TBNA, lymph node biopsy by EBUS-TBNA has become a routine examination for hilar and mediastinal lymphadenopathy in the field of respiratory medicine [13]. Although EBUS-TBNA has some complications, needle biopsy can be performed safely and less invasively compared with surgical procedures such as mediastinoscopy [14]. One of the advantages of EBUS-TBNA is that one can easily obtain fresh samples directly from pathogenic lymph nodes at most stations in the mediastinum [15]. To analyze lymphocyte profiles, it is clearly better to use freshly obtained samples than archived formalin fixed samples.
Many studies have shown that chest radiographic stage is an indicator of sarcoidosis prognosis [16]. Our study showed that the CD4/CD8 ratio in lymph node samples was significantly different by chest radiographic stage, suggesting that the lymphocyte profile is related to the pathology and prognosis of sarcoidosis. In addition, when sarcoidosis was compared with lung cancer, the CD4/CD8 ratio was much higher in sarcoidosis, and the CD3/CD19 ratio was higher in lung cancer indicating a marked difference in the immune system in both diseases. These data suggest samples obtained directly from the lymph nodes are very informative and are expected to help understand the pathological condition of lymphadenopathy.
Regulatory T cells play an important role in many immune-pathologic diseases including various inflammatory diseases and malignancies [17]. Intriguingly, we observed no difference in the percentage of Treg between sarcoidosis and lung cancer. Previous studies have shown sarcoidosis was associated with Treg subset amplification, which had a reduced activity [18,19]. Other studies proposed Treg cells in sarcoidosis have a regulatory role to control disease activity or establishment anergy [20]. Our data and that of other studies support the idea that Treg cells have an important role in the pathogenesis of sarcoidosis [21]. Recently, more advanced analysis of mediastinal lymph nodes using flow cytometry showed that Th17 cells have an important role in sarcoidosis pathogenesis [22,23,24] To elucidate the pathogenic role of the Th1/Th17/Treg balance in our cohort should be the next step of our study [24].
Previous studies of the lymphocyte profile in lymphadenopathy of sarcoidosis showed markedly different results for the CD4/CD8 ratio. Oda et al. reported that the CD4/CD8 ratio in sarcoidosis lymph nodes was lower than that in the BAL fluid using immunohistochemical staining [4]. In addition, Darlington et al. reported that the CD4/CD8 ratio in EUS-FNA samples from sarcoidosis lymph nodes was significantly lower than in BAL fluid samples [25]. They used flow cytometric analyses for their needle aspiration sample, which was the same method used in the current study. In contrast, our study showed the CD4/CD8 ratio was significantly higher in TBNA samples (lymph nodes) than in the BAL fluid. In 84% of sarcoidosis patients in our study, the CD4/CD8 ratio was higher in lymph nodes than the BAL. In another report, the CD4/CD8 ratio in EBUS-TBNA samples was highly variable without any specific tendency [3]. In that study, the CD4/CD8 ratio in lymph nodes was higher than in the BAL fluid in 48% of the studied patients. Although we could not elucidate the reason for these discrepancies, we hypothesize that it might be related to the different methods of analyses (flow cytometric analysis versus immunohistochemical analysis) or different ethnic populations (Caucasian or Japanese populations) [26]. However, even though there is a marked difference in the CD4/CD8 ratio in lymph nodes among these studies, most of them including our study are consistent with the idea that the CD4/CD8 ratio in the BAL fluid and lymph nodes have a strong correlation. This suggests the CD4/CD8 ratio might reflect the pathology of sarcoidosis as a systemic inflammatory and immunologic disorder.
There were some limitations in our study. First, the sample size was relatively small, leading to the possibility of selection bias. Especially, lung cancer metastatic lymph nodes may exhibit various immune-cell profiles according to the histology of lung cancer or patients' condition. Thus, more patients are required for the detailed analysis of lung cancer metastatic lymph nodes. Second, because we focused only on sarcoidosis and lung cancer metastases, the lymphocyte profile in other benign diseases such as infectious lymphadenitis are unknown. Whether lymphocyte profiles can be used to distinguish sarcoidosis from other benign lymphadenopathies may be an interesting clinical issue. Third, this study was performed at a single academic center in Japan. Although we used the generalized and well-accepted procedures of EBUS-TBNA and flow cytometric analyses, it may be difficult to exclude potential institutional bias. Furthermore, ethnic bias might have a large influence on inflammation during

Conclusion
Lymphocyte phenotypic analysis of lymphadenopathy was performed using flow cytometry. A high CD4/CD8 ratio was more notable in lymph nodes than in BAL fluid. The lymphocyte profile in sarcoidosis lymph nodes was distinctive from that in lung cancer metastasis. Flow cytometry analysis of lymph nodes might help to clarify the pathophysiology of diseases accompanying lymphadenopathy.  (60-79 years). We compared the CD3/CD19, CD4/CD8, and Treg/CD4 ratios in lymph nodes from age-matched patients (60-79 years) from two groups (sarcoidosis and lung cancer). The results were the same as those from the analyses of all patients (Fig 4). The CD4/CD8 ratio was still higher in sarcoidosis than in lung cancer even when we compared patients aged 60-80 years old (sarcoidosis 9.31 ± 4.15, lung cancer 1.70 ± 1.34, mean ± SD) (p < 0.03, Mann-Whitney U-test).