Functionally impaired follicular helper T cells induce regulatory B cells and CD14⁺ human leukocyte antigen‐DR ⁻ cell differentiation in non‐small cell lung cancer

2.1. Patients

From June 2015 to March 2017, 57 newly diagnosed NSCLC patients (34 males and 23 females; median age 60.48 years; range 38 to 81 years) at Zhejiang Provincial People's Hospital before any systemic or local treatment was given were enrolled in the present study. Of all 57 NSCLC patients, 14, 12, 15, and 16 were classified into stages I, II, III and IV, respectively, based on criteria described in the seventh edition International Association for the Study of Lung Cancer (IASLC) TNM staging system. Detailed clinical characteristics of NSCLC patients are shown in Table 1. We also recruited 20 age‐matched healthy subjects (HS) from the physical examination department. The study was approved by the Zhejiang Provincial People's Hospital Review Board. All methods were carried out in accordance with the 1975 Declaration of Helsinki. All tumor specimens in the pathological analysis were obtained with informed consent with approval by the ethics committee of our hospital.

2.2. Cell isolation

Peripheral blood mononuclear cells were isolated from blood samples using a Ficoll‐Hypaque density gradient centrifugation method. Serum samples were obtained by centrifugation at 1500 g for 10 minutes and were immediately stored at −80°C. Serum IL‐21 was assessed using ELISA (eBioscience, San Diego, CA, USA).

CD4⁺CXCR5⁺ICOS⁺PD‐1⁺ Tfh cells, CD19⁺IgD⁺ naive B cells, and CD14⁺HLA‐DR⁻ cells from six HS and six NSCLC patients were purified using a FACSAria III Aria cell sorter (Becton Dickinson, Sparks, MD, USA) based on the expression of CD4, CXCR5, ICOS, and PD‐1 or CD19 and IgD or CD14 and HLA‐DR. Cell purity was confirmed to be >95% by flow cytometry.

2.3. Flow cytometry analysis

The following cell surface antibodies were used: PerCP‐CD3 (clone SK7; BD Biosciences, San Diego, CA, USA), PC7‐CD4 (clone 13B8.2; Beckman Coulter, Marseille cedex, France), Alexa Fluor 488‐CXCR5 (clone RF8B2; BD Biosciences), APC‐ICOS (clone ISA‐3; BD Biosciences), PerCP‐cy7‐PD‐1 (clone H12.1; BD Biosciences), APC‐CXCR3 (clone IC6; BD Biosciences), PerCP‐cy5.5‐CCR6 (clone 11A9; BD Biosciences), FITC‐CD19 (clone J4.119; Beckman Coulter) and PE‐CD14 (clone RMO52; Beckman Coulter). After cells were incubated with cell surface antibodies for 30 minutes at 4°C in the dark, they were washed with PBS and then analyzed by flow cytometer. CD4⁺CXCR5⁺ICOS⁺PD‐1⁺ Tfh cells were identified based on ICOS and PD‐1 expression after cells were gated on CD3⁺CD4⁺CXCR5⁺ (Figure S1). Tfh subtypes were determined according to CXCR3 and CCR6 expression after cells were gated on CD3⁺CD4⁺CXCR5⁺ (Figure S1) and the PD‐1 expression of the three subtypes was further analyzed.

For the detection of intracellular cytokines following cell surface staining, cells were fixed and permeabilized using a Cytofix/Cytoperm kit (BD Biosciences) and then stained using PE‐IL‐10 (clone JES5‐19F1; BD Biosciences) and PE‐cy7‐TGF‐β (clone TW4‐9E7; BD Biosciences) or Alexa Fluor 488‐TNF‐α (clone MAb11; BD Biosciences). Stained cells were then analyzed using a FACS Canto II flow cytometer and Diva software (Becton Dickinson). All staining was carried out according to the manufacturer's protocol.

2.4. Function analyses of Tfh cells

Isolated Tfh cells (1.5 × 10⁴) were cultured either alone or 1:1 with purified CD19⁺IgD⁺ cells (1.5 × 10⁴) in complete RPMI 1640 containing l‐glutamine, NaHCO₃, 10% FCS and penicillin/streptomycin (100 U/mL) in 96‐well U‐bottom plates in the presence of 2 μg/mL Staphylococcal Exterotoxin B (SEB) for 72 hours, with PIB (phorbol‐12‐myristate‐13‐acetate + ionomycin + brefeldin A) added in the last 5 hours, as described elsewhere. Cells were then stained with FITC‐CD19, permeabilized, stained intracellularly with PE‐IL‐10 and PE‐cy7‐transforming growth factor beta (TGF‐β) and analyzed by flow cytometry. The supernatant was harvested for IL‐10 and TGF‐β detection.

Isolated Tfh cells (1.5 × 10⁴) were cultured either alone or 1:1 with purified CD14⁺HLA‐DR⁻ cells (1.5 × 10⁴) in complete RPMI 1640 containing l‐glutamine, NaHCO₃, 10% FCS and penicillin/streptomycin (100 U/mL) in 96‐well U‐bottom plates for 72 hours, with PIB added in the last 5 hours as described elsewhere. Cells were then stained with PE‐CD14, permeabilized, stained intracellularly with Alexa Fluor 488‐tumor necrosis factor (TNF)‐α and analyzed by flow cytometry. Supernatant TNF‐α levels were examined by ELISA.

2.5. Enzyme‐linked immunosorbent assay

Human IL‐21, IL‐10, TGF‐β and TNF‐α ELISA Ready‐Set‐Go Kits (eBioscience) were used to examine cytokine levels following instructions provided by the manufacturer.

2.6. Immunohistochemistry for PD‐L1

All measurements for PD‐L1 were obtained according to the immunohistochemistry (IHC) protocols provided by the manufacturers. All IHC results were checked independently by two pathologists. The cutoff for PD‐L1 expression on tumor cells (Dako, 22C3, Copenhagen, Denmark; approved by the FDA) was equal to or more than 50% staining.

2.7. Statistical analysis

Statistical analysis was carried out with GraphPad Prism 5.01 software (GraphPad Software Inc., San Diego, CA, USA). The statistical tests used for data analysis included the Mann‐Whitney U test and the Pearson test for correlation analysis. Quantitative data are presented as the mean values ± standard deviations (SD). Differences were considered to be statistically significant at values of P < .05 and P < .01.

3.1. Elevated numbers of Tfh cells and skewing to PD‐1⁺‐Tfh2 and PD‐1⁺‐Tfh17 subtypes in NSCLC patients

Compared to HS, a significantly higher frequency and number of circulating CD4⁺PD‐1⁺, Tfh cells were observed in NSCLC patients (Figure 1A,B), but there was no significant difference in the frequency and number of CD4⁺CXCR5⁺ cells between the two groups (Figure 1A,B). Because Tfh cells can be divided into Tfh1 (CXCR3⁺CCR6⁻), Tfh2 (CXCR3⁻CCR6⁻) and Tfh17 (CXCR3⁻CCR6⁺) subtypes (Figure 1C,D), we further compared the prevalence of these three subtypes in NSCLC patients. Tfh2 and Tfh17 subtypes were significantly elevated in NSCLC patients, whereas Tfh1 subtypes were decreased in NSCLC patients (Figure 1C,D).

We also investigated PD‐1 expression of three Tfh subtypes in HS and NSCLC patients and found higher mean fluorescence intensity (MFI) of PD‐1 in Tfh1 than in Tfh2 and Tfh17, and higher MFI of PD‐1 in Tfh1 in NSCLC patients than in HS (Figure 1E,F). However, we further compared PD‐1 expression of three Tfh subtypes in NSCLC patients with those in HS and found that the frequency and the number of PD‐1⁺‐Tfh2 and PD‐1⁺‐Tfh17 in NSCLC patients were higher than those in HS (Figure 1G). These data suggest that circulating Tfh cells, especially PD‐1⁺‐Tfh2 and PD‐1⁺‐Tfh17 subtypes, expand in NSCLC and indicate a potential involvement of Tfh cells in NSCLC.

Non‐small cell lung cancer patients can be classified into stage I, stage II, stage III and stage IV according to the TNM classification system, with severity increasing from stage I to stage IV. Thus, we further analyzed the numbers of circulating Tfh cells in NSCLC patients with different stages. Our data showed lower numbers of CD4⁺PD‐1⁺ (Figure 1I), Tfh cells (Figure 1J), Tfh2 (Figure 1L) and Tfh17 subtypes (Figure 1M) but a higher number of Tfh1 subtypes (Figure 1K) in NSCLC patients with early stage (I and II) relative to those with advanced stage (III and IV); however, there was no significant difference in the number of CD4⁺CXCR5⁺ cells in NSCLC with different stages (Figure 1H).

3.2. Decreased serum IL‐21 levels and its negative association with circulating Tfh cells in NSCLC patients

Interleukin‐21 is produced mainly by Tfh cells. Although the number of Tfh cells was elevated in NSCLC patients, in the present study, we observed decreased serum IL‐21 levels (Figure 2A), especially in those with advanced stage (III and IV; Figure 2B). Moreover, IL‐21 levels were significantly and negatively correlated with the number of circulating Tfh cells in NSCLC patients (Figure 2C). We further analyzed IL‐21 levels relative to Tfh subtypes and found that the number of Tfh17 subtypes was negatively correlated with IL‐21 levels (Figure 2F), whereas the number of Tfh1 and Tfh2 subtypes were not correlated with IL‐21 levels in NSCLC patients (Figure 2D,E). These results indicated that the function of Tfh cells to produce IL‐21 could be impaired in NSCLC patients.

3.3. Follicular helper T cells from NSCLC patients induce differentiation of regulatory B cells and CD14⁺HLA‐DR⁻ cells

We know that interaction of PD‐1 and PD‐L1 inhibits immune responses in tumors. Our previous study found high numbers of IL‐10⁺ regulatory B cells and CD14⁺HLA‐DR^low/− myeloid‐derived suppressor cells (MDSC) in NSCLC patients22 and, thus, in the present study, we investigated the functional effect of Tfh cells on CD19⁺IgD⁺ naïve B cells and CD14⁺HLA‐DR⁻ cells. Tfh cells were sorted from six HS and six NSCLC patients and were cocultured 1:1 with autologous CD19⁺IgD⁺ naïve B cells in the presence of SEB. After 72 hours, we observed a slightly higher frequency of IL‐10‐ and TGF‐β‐producing CD19⁺ B cells (Figure 3A‐C) and supernatant IL‐10 and TGF‐β levels (Figure 3D,E) when cocultured with Tfh cells compared to CD19⁺IgD⁺ cells alone from NSCLC patients.

Similarly, after coculture with Tfh cells from NSCLC patients, a higher, statistically significant frequency of TNF‐α‐producing CD14⁺HLA‐DR⁻ cells (Figure 3F,G) and supernatant TNF‐α levels (Figure 3H) was observed when compared to CD14⁺HLA‐DR⁻ cells alone. However, coculture with Tfh cells from HS had no significant effect on the cytokine production of CD19⁺IgD⁺ naïve B cells and CD14⁺HLA‐DR⁻ cells (Figure 3).

3.4. High expression of PD‐L1 on tumor cells is significantly correlated with high numbers of circulating Tfh cells in NSCLC patients

Non‐small cell lung cancer is known to upregulate PD‐L1 expression, and high PD‐L1 expression is an independent predictor of poor prognosis.23, 24 In the present study, we found that 32 NSCLC samples (56.1%) stained negative for PD‐L1, and 25 NSCLC samples (43.9%) stained positive for PD‐L1 (Figure 4A,B). PD‐L1‐positive NSCLC tumor cells were significantly correlated with high numbers of CD4⁺PD‐1⁺ (Figure 4D) and Tfh cells (Figure 4E), but there was no significant correlation with CD4⁺CXCR5⁺ (Figure 4C) and clinicopathological factors (data not shown). As previously reported, we also found that NSCLC patients with PD‐L1‐negative tumor cells had longer disease‐free survival after tumor resection (Figure 4F).

3.5. Negative correlation between Tfh1 subtypes with disease‐free survival

Thus, we determined whether Tfh cells with high PD‐1 expression were also predictive of survival time in NSCLC. We analyzed the correlation between Tfh cells and disease‐free survival after tumor resection. Interestingly, the number of Tfh1 subtypes was positively correlated with disease‐free survival of NSCLC patients (Figure 5D). However, no significant association was found between CD4⁺CXCR5⁺, CD4⁺PD‐1⁺, Tfh cells, Tfh2, or Tf17 subtypes and disease‐free survival of NSCLC patients (Figure 5A,B,C,E,F).

3.6. Circulating Tfh cells in NSCLC patients with different histological subtypes

Non‐small cell lung cancer comprises several subtypes based on histological examination in which adenocarcinoma and squamous cell carcinoma are the two major subtypes. We investigated the levels of circulating CD4⁺CXCR5⁺, CD4⁺PD‐1⁺ and Tfh cells in these NSCLC histological subtypes. Our data showed that the frequency and the number of circulating CD4⁺CXCR5⁺ and Tfh cells were higher in adenocarcinoma than in squamous cell carcinoma or in the other types, but the level of CD4⁺PD‐1⁺ was no different in the three histological subtypes (Figure 6). However, the level of circulating CD4⁺CXCR5⁺, CD4⁺PD‐1⁺ and Tfh cells was not significantly different between squamous cell carcinoma and other types.