Th17, synchronically increased with T_regs and B_regs, promoted by tumour cells via cell‐contact in primary hepatic carcinoma

Summary

Documented reports about T helper type 17 (Th17) cells have revealed that Th17 plays a critical role in inflammation and autoimmunity diseases. However, the role of Th17 in cancer remains contradictory. The interplay between Th17 and tumour cells in the tumour microenvironment of primary hepatic carcinoma (PHC) needs to be explored further and the relationship between Th17, regulatory T cells (T_regs) and regulatory B cells (B_regs) has not been defined completely. In this study, numerous experiments were undertaken to elucidate the interaction of Th17 and T_reg/B_reg cells involved in PHC. Our work demonstrated that an increased Th17 was detected in the peripheral circulation and in tumour tissues in PHC patients. In addition, increases in peripheral blood Th17 corresponded with tumour–node–metastasis (TNM) stage progression. Also, further studies indicated that Th17 cells were promoted by tumour cells in the PHC tumour microenvironment through both contact‐dependent and ‐independent mechanisms, but cell‐contact played the major important role in promoting the production and proliferation of Th17. When isolated CD4⁺CD25⁺CD127^low T_regs and CD4⁺CD25^–CD127⁺ non‐T_regs were cultured with autologous tumour cells, it implied that the phenotype of Th17 and T_regs was modified by tumour cells in the tumour microenvironment. As well as this, Th17 cells were also found to correlate positively with CD4⁺forkhead box protein 3⁺ T_regs and CD19⁺CD5⁺CD1d^hi B_regs in PHC. Notably, Th17 increased synchronically with T_regs and B_regs in PHC. These findings may provide new clues to reveal the mechanisms of immune escape in PHC.

Introduction

Primary hepatic carcinoma (PHC) is one of the most prevalent and lethal malignancies in the world 1, 2. Although it is the fifth most common cancer worldwide and the third most common cause of cancer‐related death 3, the relative mechanisms by which the immune system is modulated in patients with PHC remains largely unclear.

T helper type 17 (Th17) cells and their effector cytokines are being recognized increasingly as key players in inflammation and autoimmune diseases 4, 5. Although investigations into Th17 cells have been made for several years, the roles of Th17 cells in tumour immunity are still inconsistent. As to the function of Th17 cells in tumours, it has been shown that Th17 cells can promote tumour growth or might potentially mediate an anti‐tumour effect 6, 7, 8, 9, 10, so the precise expression of Th17 cells in tumour progression remains unclear and the exact role of Th17 in tumour immunity remains undefined and needs to be explored further.

Regulatory cells include regulatory T cells (T_regs) and regulatory B cells (B_regs). As we know, T_reg cells play important roles in immune regulations 11, 12. The immune suppressive roles of T_regs in the development and progression of cancer have also been studied extensively 13, 14.

In addition to T_regs, a newly confirmed subset of B cells, described as CD19⁺CD1d^hiCD5⁺ 15 or CD19⁺CD24^hiCD38^hi 16, has recently been at the forefront of investigation and plays an important regulatory role in autoimmune diseases and tumours 15, 16, 17, 18.

In this study we quantified the percentage of Th17 cells in PHC, and investigated the potential relationship and mechanisms behind the interplay between the levels of Th17 cells and the development of PHC in humans. In addition, the relationship between Th17 and T_regs (or B_regs) will be investigated in patients with PHC. To our knowledge, this study is the first to elaborate the relationship between Th17 and B_regs in human PHC.

Materials and methods

Patients and specimens

Patients with PHC (n = 93, 13 females and 80 males, mean age = 49·02 years, age range from 26 to 75 years) without previous treatments were enrolled into the study upon giving informed consent. According to the tumour–node–metastasis (TNM) classification for primary hepatic carcinoma (Union for International Cancer Control, UICC 2010, 7th edition), the patients were divided into four stages: stage I (n = 11, 11·83%), stage II (n = 28, 30·11%), stage III (n = 40, 43·01%) and stage IV (n = 14, 15·05%). Closely age‐matched, healthy volunteers (n = 51, 12 males and 39 females, mean age= 46·00 years, age range from 25 to 77 years) were recruited as healthy controls. Tumour tissues, peritumoral tissues (at least 2 cm distant from the tumour site) and tumour‐free tissues (at least 4 cm distant from the tumour site) were obtained intra‐operatively from 15 patients. This study was approved by the ethical committees of Changhai Hospital and the Eastern Hepatobiliary Hospital of the Second Military Medical University.

Cell preparation

Peripheral blood mononuclear cells (PBMC), CD4⁺ T cells, tumour infiltrating lymphocytes (TIL), non‐tumour infiltrating lymphocytes (NIL), CD14⁺ monocyte cells, tumour cells and non‐tumour cells were isolated from peripheral blood or tumour tissues. As monocytes are indicated to enhance Th17 cell levels 19, CD14⁺ cells were depleted from cells obtained from tumour and non‐tumour tissues to avoid the effect of monocytes/macrophages. Detailed isolation is described in the Supporting information, Table S1.

Flow cytometry analysis and cell sorting

Detailed activation and incubation of the cells mentioned previously are described in Supporting information, Table S1. The following anti‐human antibodies were used: CD4 phycoerythrin‐cyanin 5·5 (PE‐Cy5·5), interleukin (IL)‐17A PE, forkhead box protein 3 (FoxP3) fluorescein isothiocyanate (FITC), CD86 PE, CD1d allophycocyanin (APC), FoxP3 PE, CD127 PE‐Cy5·5 (all from eBioscience, San Diego, CA, USA), CD80 FITC, inducible T cell co‐stimulator ligand (ICOSL) (anti‐CD275) PE (both from BioLegend, San Diego, CA, USA), CD19 PE‐Cy7, CD5 FITC, CD4 FITC (all from Beckman Coulter, USA) and CD25 APC (BD, USA). Stained cells were analysed by flow cytometry using a fluorescence activated cell sorter (FACS)Canto II (BD Biosciences, San Jose, CA, USA). Data were analysed using FlowJo software (Tree Star, Ashland, OR, USA).

After staining with surface markers, CD19⁺B cells, CD4⁺ CD25⁺CD127^low T_regs, CD4⁺CD25^–CD127⁺ non‐T_regs were sorted from healthy controls and PHC patients using FASC Aria II (BD Biosciences) according to the manufacturer's instructions.

Th17 cell isolation and cell proliferation assay

Th17 cells were isolated by using stimulation and staining reagents from the IL‐17 secretion assay detection kit (Miltenyi Biotec, Bergisch Gladbach, Germany). In brief, isolated CD4⁺ T cells were stimulated with CytoStim for 4 h; cells were first labelled with IL‐17 catch reagent followed by another 45‐min incubation period at 37°C. Finally, cells were stained with an IL‐17 detection antibody. Isolated Th17 cells were stained first with carboxyfluorescein diacetate succinimidyl ester (CFSE; Invitrogen, Carlsbad, CA, USA); cells were then co‐cultured with or without autologous tumour cells from tumour tissues or non‐tumour cells from peritumoral tissues at 10 : 1 in the presence of anti‐CD3/CD28 (10 μg/ml) and IL‐2 (100 U/ml) for 7 days. After 7 days of incubation, cells were stained for further analysis on flow cytometry (FACSCanto II; BD Biosciences). Analysis of cell division was performed using ModFit software version Mac3·2 (Verity Software House Inc., Topsham, ME, USA).

Isolated Th17 cells were co‐cultured with sorted CD19⁺ B cells (at a density of 2 × 10⁶ cells/ml) at 1 : 1 followed by stimulation with 100 nM cytosine–phosphatase–guanine (CpG) ODN2006 (Invivogen, Toulouse, France) for 48 h. CD19⁺CD5⁺CD1d^hi B_regs were then analysed with flow cytometry.

RNA extraction and real‐time polymerase chain reaction (PCR)

The sequences of primers are listed in Table 1 and Supporting information, Table S1. Total RNA was extracted from PBMC with TRIzol (Invitrogen), according to the manufacturer's instructions. The RNA was quantified using a NanoDrop‐2000 spectrophotometer (Thermo Scientific, Fremont, CA, USA), and 500 ng total RNA was reverse‐transcribed to cDNA using SYBR PrimeScript RT reagent Kit (Takara, Shiga, Japan), according to the manufacturer's instructions. The real‐time PCR reactions were performed with the Rotor‐Gene 6000 real‐time PCR system using SYBR Premix Ex Taq from Takara following the manufacturer's recommendations. Relative expression was determined using the 2^–ΔΔCt method with averaged relative levels of β‐actin used for normalization.

Table 1.

Nuclear sequences of primes used in real‐time polymerase chain reaction

Gene	Sequence	Product (base pairs)	GeneBank Accession number
β‐actin	5′‐TGGCACCCAGCACAATGAA‐3′ 5′‐CTAAGTCATAGTCCGCCTAGAAGCA‐3′	186	NM_001101
IL‐1R	5′‐CCTGTGATTGTGAGCCCAGCTA‐3′ 5′‐ ACTCAACTGGCCGGTGACATTAC‐3′	90	NM_000877
IL‐6	5′‐AAGCCAGAGCTGTGCAGATGAGTA‐3′ 5′‐TGTCCTGCAGCCACTGGTTC‐3′	150	NM_000600
IL‐17A	5′‐CTGAACATCCATAACCGGAATACCA‐3′ 5′‐AGCGTTGATGCAGCCCAAG‐3′	162	NM_002190
IL‐17F	5′‐TAACATCGAGAGCCGCTCCAC‐3′ 5′‐ CATTGATGCAGCCCAAGTTCC‐3′	113	NM_052872
IL‐21	5′‐GCCACATGATTAGAATGCGTCAAC‐3′ 5′‐TGGAGCTGGCAGAAATTCAGG‐3′	92	NM_021803
IL‐22	5′‐CGCACCTTCATGCTGGCTAA‐3′ 5′‐ AAGTTCAGCACCTGCTTCATCAGA‐3′	131	NM_020525
IL‐23p19	5′‐CAGCTTTCACAGAAGCTCTGCAC‐3′ 5′‐TGACTGTTGTCCCTGAGTCCTTG‐3′	158	NM_016584
TGF‐β	5′‐AGCGACTCGCCAGAGTGGTTA‐3′ 5′‐GCAGTGTGTTATCCCTGCTGTCA‐3′	130	NM_000660
RORc	5′‐ACCTCACCGAGGCCATTCAG‐3′ 5′‐TAGGCCCGGCACATCCTAAC‐3′	139	NM_001001523
FoxP3	5′‐GTTCACACGCATGTTTGCCTTC‐3′ 5′‐CACAAAGCACTTGTGCAGACTCAG‐3′	91	NM_014009

IL = interleukin; TGF = transforming growth factor; RORc = RAR‐related orphan receptor C; FoxP3 = forkhead box protein 3.

Immunohistochemistry and confocal microscopy

PHC tissues were fixed in 4% paraformaldehyde and embedded in paraffin. Paraffin‐embedded samples were cut into 3‐μm sections and resected specimens were dewaxed in xylene, rehydrated, washed in distilled water and then stained with haematoxylin and eosin (H&E). Immunohistochemical staining was performed as described previously 17, 20. Briefly, sections were deparaffinated in xylene and rehydrated. Antigen retrieval was achieved by autoclaving in 1 mM ethylenediamine tetraacetic acid (EDTA) (pH 8·0) at 121°C for 10 min. Endogenous peroxidase was blocked with 3% H₂O₂ for 5 min and non‐specific staining was prevented by soaking in 5% bovine serum albumin (BSA; Nacalai Tesque, Kyoto, Japan). Sections were stained overnight at 4°C with mouse anti‐human CD4 (at a 1 : 40 dilution; Novocastra, Newcastle upon Tyne, UK) and rabbit anti‐human IL‐17 (at a 1 : 100 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA). After washing in PBS, the samples were incubated with secondary antibodies: FITC‐conjugated goat anti‐mouse immunoglobulin (Ig)G at a 1 : 200 dilution and PE‐conjugated goat anti‐rabbit IgG at a 1 : 100 dilution (all from Sigma, St Louis, MO, USA) for 30 min at room temperature before washing with PBS. Images were viewed and assessed using a scanning confocal microscopy (LSM510 META; Zeiss, Jena, Germany) and analysed by LSM510 META software. The mean fluorescence intensity (MFI) of CD4 and IL‐17 were analysed with Image J (National Institutes of Health, Bethesda, MD, USA).

Transwell and cell co‐culture systems

A contact‐independent co‐culture system was set up using Transwell inserts in which the two chambers were separated by a semipermeable membrane. The tumour cells were seeded at a density of 4 × 10⁴ cells/well in 24‐well plates and cultured overnight. A Transwell insert membrane with a 0·4 μm pore‐size (Millipore, Burlington, MA, USA) was set in the well before autologous 4 × 10⁵ cells/well CD4⁺ T cells were added to the upper chamber. CD4⁺ T cells were also cultured with tumour cells at a ratio of 10 : 1 in the cell‐contact co‐culture system. All the cells were co‐cultured for 4 h in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin, after which the CD4⁺ T cells were incubated with CD4 PE‐Cy5·5, IL‐17A PE for Th17 detection and analysed by flow cytometry, as described previously.

Neutralization assay

Neutralization antibodies were diluted in PBS according to the manufacturer's instructions. To block IL‐23, IL‐6 and transforming growth factor (TGF)‐β, tumour cells were incubated with isotypes or anti‐human IL‐23p19 neutralizing monoclonal antibodies (mAb) (R&D Systems, Minneapolis, MN, USA; final concentration of 5 μg/ml), anti‐IL‐6 (Millipore; final concentration of 10 μg/ml) and anti‐TGF‐β mAb (Millipore; final concentration of 10 μg/ml) singly or in combination for 1 h in the Transwell co‐culture system before autologous CD4⁺ T cells were seeded. Mouse IgG or anti‐CD80, anti‐CD86 and anti‐ICOSL (all from Ancell, Stillwater, MN, USA; final concentration of 10 μg/ml) were incubated with tumour cells for 1 h singly or in combination in the direct‐cell‐contact co‐culture system. CD4⁺ T cells were then added to both culture systems. All the cells were co‐cultured for 48 h in DMEM containing 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin, after which the CD4⁺ T cells were incubated with CD4 PE‐Cy5·5, IL‐17A PE for Th17 detection.

Statistical analysis

The results are expressed as means ± standard deviation (s.d.). Statistical analyses were performed using the GraphPad Prism version 5·0 software package (GraphPad Software, San Diego, CA, USA). Differences between groups were analysed using linear regression, t‐test (Mann–Whitney U‐test or paired test, two groups) or one‐way analysis of variance (anova) (Kruskal–Wallis test and Dunn's multiple comparison test, more than two groups). All tests were two‐sided and significance was determined when P < 0·05 (*P < 0·05, **P < 0·01 and ***P < 0·001).

Results

Significantly increased Th17 cells in peripheral blood correlate with PHC TNM stage

PBMC in patients with PHC (n = 93) and in healthy volunteers (n = 51) were examined for the prevalence of Th17 cells. The Th17 cell population as a percentage of total CD4⁺ T cells was evaluated by flow cytometry. Representative plots showed that the Th17 cell populations in the gated CD4⁺ cells were increased in the patients with PHC in comparison to normal controls (Fig. 1a). Summarized data from all individuals indicated that the proportion of Th17 cells in the peripheral blood of PHC patients was significantly higher than that in healthy donors (4·621 ± 0·856% versus 3·148 ± 0·716%, P < 0·001, Fig. 1a). In PBMC, compared with normal controls (NC), PHC expressed extremely high MFI levels of IL‐17 (P < 0·001, Fig. 1b).

Figure 1.

Significantly increased T helper type 17 (Th17) cells in peripheral blood mononuclear cells (PBMC) were detected in primary hepatic carcinoma (PHC) patients compared with those in normal controls (NC). (a) CD4⁺interleukin (IL)‐17⁺ Th17 cells in peripheral blood were analysed by flow cytometry; the percentage of Th17 cells in PHC was higher than that in NC (representative fluorescence‐activated cell sorter (FACS) data showing cell percentages). Th17 cells in the PHC group (n = 93) were higher than those in the NC group (n = 51), the differences between NC and PHC were significant (***P < 0·001). (b) Increased mean fluorescence intensity (MFI) level of IL‐17 was detected in PHC compared to NC (***P < 0·001). (c) Th17 cells were enhanced with PHC tumour–node–metastasis (TNM) stage progression; the percentage of Th17 in advanced stages (stages III–IV, n = 54) was higher than that in early stages (stages I–II, n = 39) (**P < 0·01). (d) Compared with NC (n = 51), significantly increased mRNA expression of Th17‐related factors in PBMC were detected in PHC (n = 93) with real‐time polymerase chain reaction (PCR). The mRNA expression of IL‐1R, IL‐6, IL‐17A, IL‐17F, IL‐21, IL‐22, IL‐23p19, RAR‐related orphan receptor C (RORc) and transforming growth factor (TGF)‐β were higher than that in NC (*P < 0·05; **P < 0·01; ***P < 0·001, respectively). [Colour figure can be viewed at http://wileyonlinelibrary.com]

The clinical characteristics of the study subjects are summarized in Table 2. The potential relationship between increased peripheral blood Th17 cells and the clinical characteristics of the 93 patients, including gender, age, the diameter and number of the tumours, portal vein tumour thrombosis (PVTT), metastasis, the preoperative concentration of AFP and TNM stages were analysed. The percentage of Th17 cells in PHC patients correlated positively with tumour size, PVTT and especially the TNM stages (P < 0·05 for each group). The average percentage of Th17 cells in PHC patients was increased with TNM stage progression (I < II < III < IV, P < 0·001), where Th17 cells in patients with advanced stages of PHC (III–IV) were significantly higher than those in the early stages (I–II) (P < 0·01, Fig. 1c). These observations indicated that the patients with advanced PHC had an increased population of Th17 cells in their PBMC. The mRNA expression levels of Th17‐related factors in PBMC from PHC patients and normal controls were investigated using real‐time PCR. Increased expression of IL‐1R, IL‐6, IL‐17A, IL‐17F, IL‐21, IL‐22, IL‐23p19, retinoic acid‐related orphan receptor C (RORc) and TGF‐β mRNA was observed in PHC (P < 0·05 for each group, Fig. 1d). All these results suggested that Th17 cells were enriched in the peripheral blood of PHC patients, and increased Th17 population was correlated positively with PHC TNM stage.

Table 2.

The relationship between T helper type 17 cells and clinical characteristics of the primary hepatic carcinoma (PHC) patients

	Number	Th17 ( $\overline{\mathrm{x}}$ ± s) (%)	P‐values
Gender			0·8898
Male	80	4·60 ± 0·81
Female	13	4·78 ± 1·15
Age (years)			0·6034
≥ 50	41	4·66 ± 1·00
< 50	52	4·58 ± 0·75
Tumour size (cm)			0·0074 **
≥ 5	57	4·77 ± 0·86
< 5	36	4·25 ± 0·79
Tumour number			0·9387
≥ 2	22	4·61 ± 0·79
1	71	4·62 ± 0·88
Portal vein thrombosis			0·0363 *
Yes	20	4·94 ± 0·80
No	73	4·46 ± 0·70
Lymph intumesce			0·1632
Yes	9	4·96 ± 0·76
No	84	4·59 ± 0·88
Metastasis			0·3768
Yes	8	4·95 ± 1·03
No	85	4·59 ± 0·84
Preoperative AFP (ug/l)			0·1234
≥ 20	60	4·57 ± 0·91
< 20	33	4·71 ± 0·82
TNM stages			<0·0001 ***
I	11	4·07 ± 0·75
II	28	4·29 ± 0·70
III	40	4·87 ± 0·84
IV	14	5·02 ± 0·87

*P < 0·05; **P < 0·01; ***P < 0·001. AFP = alpha‐fetoprotein; TNM = tumour–node–metastasis.

CD4⁺IL‐17⁺Th17 cells accumulated in situ in PHC tumours

To examine the distribution of Th17 cells in different liver sites, H&E staining, fluorescence labelling and confocal microscopy were used. We demonstrated that compared with peritumoral tissues, more lymphocytes infiltrated into tumour tissues were detected in the intrahepatic region (Fig. 2a,b). To visualize the distribution of liver‐infiltrating Th17 cells, we used double immunostaining: CD4 and IL‐17. Greatly infiltrated CD4⁺IL‐17⁺Th17 cells were found in the lobular and portal areas of the livers in PHC tumour tissues (Fig. 2c), whereas the liver tissues from autologous peritumoral tissue controls had few Th17 cells (Fig. 2d). The MFI of CD4 and IL‐17 of Fig 2c,d were quantified as shown in Supporting information, Fig. S1. These data indicate that CD4⁺IL‐17⁺ cells were accumulated in tumour tissues of PHC patients.

Figure 2.

T helper type 17 (Th17) cells infiltrated into primary hepatic carcinoma (PHC) tumour tissues. Tumour tissues and peritumoral tissues in patients (n = 10) with PHC were stained with haematoxylin and eosin (H&E) or fluorescent antibodies and analysed by confocal laser scanning microscopy; green represents anti‐CD4 fluorescein isothiocyanate (FITC), red shows anti‐interleukin (IL)‐17 phycoerythrin (PE), yellow stands for merged FITC and PE. (a) In non‐tumour tissues, liver structure was intact and rare inflammatory cell infiltration was observed, (b) while in cancer tissues, the integral liver structures were severely destroyed and there was frequent inflammatory cell infiltration. (c) A large number of CD4⁺IL‐17⁺Th17 cells can be detected in tumour tissues, (d) while rare Th17 cells can be seen in non‐tumour tissues. Scale bar = 150 μm (all photomicrographs are from the same patient visualized under ×200 magnification). [Colour figure can be viewed at http://wileyonlinelibrary.com]

Tumour cells promote the proliferation of Th17 cells via cell‐contact in PHC

In addition, isolated Th17 cells stained with CFSE were co‐cultured with or without autologous tumour cells from tumour tissues or non‐tumour cells from peritumoral tissues for 7 days. Significant proliferation was found in Th17 cells co‐cultured with tumour cells (Fig. 3a). As we showed that tumour cells play important roles in promoting Th17 cells, to investigate this issue further purified CD4⁺ T cells were left untreated or cultured with autologous tumour cells for 48 h in cell‐contact‐dependent or cell‐contact‐independent conditions. Here, we observed that the tumour cells enhanced the IL‐17 production in both systems, but the proportion of IL‐17 in the direct‐contact co‐culture system was much higher than that in the Transwell co‐culture system (Fig. 3b). Our findings clearly support the emerging concept that tumour environmental factors drive the generation and expansion of Th17 cells. We therefore investigated which factors were involved in elevating Th17 cells in tumours.

Figure 3.

Factors involved in elevating T helper type 17 (Th17) in the tumour microenvironment. (a) Isolated interleukin (IL)‐17⁺cells stained with carboxyfluorescein diacetate succinimidyl ester (CFSE) were left untreated or co‐cultured with non‐tumour cells or tumour cells at 10 : 1 in the presence of anti‐CD3/CD28(10 μg/ml) and IL‐2 (100 U/ml) for 7 days. Tumour cells promoted autologous Th17 cell proliferation (n = 3). (b) CD4⁺T cells were left untreated or were cultured in Transwell or in direct contact with autologous tumour cells at a ratio of 10 : 1 for 48 h. Th17 cells were enhanced by tumour cells both in cell‐contact‐independent and cell‐contact‐dependent systems, with cell‐contact playing the major role in promoting Th17 cells (*P < 0·05, **P < 0·01, n = 5). (c) Th17 cells were decreased by blocking IL‐23p19, IL‐6 and transforming growth factor (TGF)‐β singly or in combination (n = 3, *P < 0·05). Neutralization of CD80, CD86 and inducible T cell co‐stimulator ligand (ICOSL) (CD275) in tumour cells singly or in combination decreased Th17 activation (n = 3, *P < 0·05). (d) Tumour cells irradiated with 150 Gy γ‐ray also promoted Th17 cell development (n = 3, **P < 0·01). [Colour figure can be viewed at http://wileyonlinelibrary.com]

Given that cell‐contact‐independent enhancement of Th17 cells in the tumour microenvironment is possible, we hypothesized that tumour cells and the tumour microenvironment create an optimal cytokine milieu to facilitate the generation of Th17 cells. Taking into account the role of IL‐23p19, IL‐6 and TGF‐β in the differentiation and proliferation of Th17, we incubated tumour cells with neutralizing antibodies IL‐23p19, IL‐6 and TGF‐β singly or in combination for 1 h before CD4⁺ T cells were added to the contact‐independent co‐culture system for 48 h. The proportion of Th17 cells was then analysed by flow cytometry. As shown in Fig. 3c, the proportion of Th17 cells was decreased by the blocking antibody. Therefore, IL‐23p19, IL‐6 and TGF‐β promoted the differentiation and proliferation of Th17 cells in the tumour microenvironment.

Because cell‐contact also enhances Th17 cells in the tumour microenvironment, it is possible that Th17 cells may be activated by tumour cells in a cell–cell‐contact manner. To address this issue, we used flow cytometry to detect the expression of co‐stimulatory molecules required for human CD4⁺T cell activation on tumour cells derived from PHC tumour tissues required for human CD4⁺ T cell activation. As shown in Supporting information, Fig. S2, we observed that co‐stimulatory molecules such as CD80 and CD275 (ICOSL) were expressed highly by tumour cells. To remove the effect of ICOSL, CD80 and CD86, tumour cells were incubated with neutralizing antibodies to ICOSL, CD80 and CD86 singly or in combination for 1 h before CD4⁺ T cells were added to the cell direct‐contact co‐culture system. Two days later, Th17 cells were analysed by flow cytometry. Intriguingly, the proportion of Th17 cells was decreased by the blocking antibody (Fig. 3c). Thus, elevated Th17 cells in the tumour microenvironment may be regulated by membrane co‐stimulatory molecules (CD80, CD86 and ICOSL) expressed on tumour cells.

To avoid the effect of cytokines in the cell‐contact system and to evaluate the signal provided by binding of membrane co‐stimulatory molecules on tumour cells, freshly isolated tumour cells were irradiated with 150 Gy γ‐ray at a 15 Gy/min dose rate, then autologous CD4⁺ T cells were incubated with them at a ratio of 10 : 1 for 48 h. This also led to an increase in Th17 cells (Fig. 3d, 4·11–10·69%). Therefore, factors involved in antigen presentation such as CD80, CD86 and ICOSL also played important roles in promoting Th17 cells. All these data indicate that CD4⁺IL‐17⁺ cells are promoted markedly by tumour cells of PHC patients via cell‐contact.

The phenotypical features of T cells are altered by the tumour microenvironment

Given that Th17 cells accumulate in tumours in situ, it is possible that tumour cells and the tumour microenvironment play important roles in promoting Th17 cell development. However, are there Th17 cells that migrate into tumour microenvironments? To address this issue, chemokine expression of tumour cells was detected by real‐time PCR. We found that tumour cells expressed higher levels of CCL20 but lower CCL5 and CCL19 (Fig. 4a, P < 0·01).

Figure 4.

The phenotypical features of T cells are altered by the tumour microenvironment. (a) Compared with normal controls (NC) (n = 51), significantly increased mRNA expression of T helper type 17 (Th17)‐related chemokines in peripheral blood mononuclear cells (PBMC) were detected in primary hepatic carcinoma (PHC) (n = 93) with real‐time polymerase chain reaction (PCR). The mRNA expression of CCL20 was higher than that in NC (**P < 0·01), while decreased CCL5 and CCL19 were detected in PHC (***P < 0·001 and **P < 0·01, respectively). (b) Compared with uncultured CD4⁺CD25^–CD127⁺ non‐regulatory T cells (T_regs), increased CD4⁺forkhead box protein 3 (Foxp3)⁺ cells were detected in non‐T_regs cultured with autologous tumour cells (n =5, *P < 0·05). (c) Compared with uncultured CD4⁺CD25⁺CD127^low T_regs, increased CD4⁺IL‐17⁺ cells were detected in T_regs cultured with autologous tumour cells (n = 5, *P < 0·05). [Colour figure can be viewed at http://wileyonlinelibrary.com]

To assess further the relationship between Th17 and tumour cells, we identified the phenotypical features of Th17 and T_regs in the tumour microenvironment. The expression of FoxP3 varied with CD25 expression. The majority of FoxP3⁺ cells fell into the cell subset of CD4⁺CD25^high and CD4⁺CD25⁺ cells; in particular, the CD4⁺CD25⁺CD127^low cells correlated positively with FoxP3 in CD4⁺ T cell (Supporting information, Fig. S3). Therefore, PBMC stained with CD4, CD25 and CD127, and subsequently CD4⁺CD25⁺CD127^low T_regs and CD4⁺CD25^–CD127⁺ non‐T_regs, were isolated according to the strategy shown in Supporting information, Fig. S4.

Isolated CD4⁺CD25^–CD127⁺ non‐T_regs and CD4⁺CD25⁺CD127^low T_regs were incubated with autologous tumour cells for 48 h before being analysed by flow cytometry. Uncultured CD4⁺CD25^–CD127⁺ non‐ T_regs and CD4⁺CD25⁺CD127^low T_regs were used as biological controls. Interestingly, increased CD4⁺FoxP3⁺ T_regs were detected in CD4⁺CD25^–CD127⁺ non‐T_regs cultured with tumour cells (Fig. 4b). Similarly, increased CD4⁺IL‐17⁺ Th17 cells were detected in the CD4⁺CD25⁺CD127^low T_regs cultured with tumour cells (Fig. 4c). These results suggest that the phenotypical features of T cells can be altered by tumour cells, and some Th17 cells and T_regs may convert into each other in vitro.

Increased Th17 cells correlated positively with both T_regs and B_regs in PHC

To determine the relationship between Th17 and T_regs in patients with PHC, we analysed the percentage of T_regs and the ratio of T_regs to Th17. The data show that the T_regs were increased in PHC patients and increased with TNM stage progression (Supporting information, Fig. S5). We found the ratio in PHC was higher than that in normal controls, as shown in Fig. 5a (P < 0·001). We also analysed the linear relationship between Th17 and T_regs. As shown in Fig. 5b, Th17 and T_regs showed a positive linear correlation and r = 0·52, P < 0·001. To assess further the relationship between Th17 and B_regs in PHC, we detected the percentage of CD19⁺CD5⁺CD1d^hi among CD19⁺ B cells in PHC, increased percentage of B_regs were detected in PHC and increased B_regs were correlated positively with increased Th17 cells in PHC (r = 0·65, P < 0·05, Fig. 5c). These data emphasize that increased Th17 cells correlated positively with both T_regs and B_regs in PHC.

Figure 5.

T helper type 17 (Th17) were increased concurrently with regulatory T cells (T_regs) and regulatory B cells (B_regs) in patients with primary hepatic carcinoma (PHC). (a) Significantly elevated Th17 : T_regs cell ratios in the peripheral blood mononuclear cells (PBMC) of PHC patients (n = 93) compared to the normal controls (NC, n = 51) (***P < 0·001). (b) The graph indicates a significant positive correlation between Th17 and T_regs in PHC (r = 0·52, P < 0·001, n = 93). (c) Representative flow cytometry plot show the frequency of CD19⁺CD5⁺CD1d^hi B_regs among CD19⁺ B cells from PHC. Th17 and B_regs were increased in a positive linear correlation (r = 0·60, P < 0·01, n = 20). [Colour figure can be viewed at http://wileyonlinelibrary.com]

We therefore hypothesized that there might be an interaction between Th17 cells and CD19⁺CD5⁺CD1d^hi B_regs. To assess the effect of Th17 on B_regs, sorted CD19⁺ B cells from normal controls were cultured with autologous isolated Th17 cells for 48 h and CD19⁺CD5⁺CD1d^hi B_regs were examined with flow cytometry. Compared with controls, Th17 induced the up‐regulation of CD19⁺CD5⁺CD1d^hi B_regs (P < 0·05, Supporting information, Fig. S6). Th17 cells may promote the generation of CD19⁺CD5⁺CD1d^hi B_regs.

Discussion

In this study, we detected that significantly increased Th17 cells in PHC peripheral blood correlated with TNM stages and accumulated in in‐situ tumours of PHC. We showed that elevated expression of IL‐17 production in the PHC cell‐contact co‐culture system and Th17 cells proliferation were promoted by tumour cells via cell‐contact in PHC. As well, we were the first to show that increased Th17 cells correlated positively with both T_regs and B_regs in PHC. These results provide important information and new insights into Th17, T_regs and B_regs and other immune cells in the PHC tumour microenvironment.

Although investigations into Th17 cells have been made for several years, the relationship between Th17 cells and cancer and their roles in anti‐tumour immunity need to be explored further. Previous studies revealed that Th17 cells were elevated in patients with gastric cancer 21, ovarian cancer 22, cervical cancer 23, colon cancer 24, hepatocellular carcinoma 25, 26 and myeloma 27. Inconsistent with these observations is that in some carcinomas, such as breast cancer, Th17 cells were lower than in healthy controls 28. To determine the prevalence of Th17 cells in patients with PHC, we evaluated the Th17 cell population in the circulation or tumour micro‐environment. Our study showed that the frequency of Th17 cells was increased predominantly in PBMC, and Th17 cells were associated with the PHC TNM stage progression. Th17 cells in TIL were higher than those in NIL and peripheral blood. In the present study we also observed that CD4⁺IL‐17⁺ Th17 cells were enriched predominantly in the tumour microenvironment. All these results suggest that Th17 cells might be involved in PHC progression.

We show that tumour cells play important roles in promoting Th17 cells in the tumour microenvironment. We found that IL‐6, IL‐23 and TGF‐β levels were increased in PHC and CD80 and CD86, ICOSL expression was increased in tumour cells. On the basis of these studies, we demonstrate that the increased proportion of Th17 cells could be attributed to the observed increased secretion of IL‐6, IL‐23 and TGF‐β and the increased CD28–CD80/CD86, ICOS–ICOSL contact. That is to say, the cytokines in the tumour microenvironment and the contact between tumour cells and CD4⁺ T cells through co‐stimulatory molecules were all involved in promoting Th17 cells. Although some data indicated that the production of IL‐17A and IL‐17F by T lymphocytes was regulated by IL‐23 produced by dendritic cells, independently of cell–cell‐contact or traditional T cell receptor (TCR) activation 29, our study demonstrated that Th17 cells were increased via both cell‐contact‐independent and ‐dependent mechanisms in the tumour microenvironment, especially for cell‐contact‐dependent mechanisms, including co‐stimulatory molecules. Our findings clearly support the emerging concept that tumour environmental factors drive the generation and expansion of Th17 cells through both cytokines and cell‐contact.

Our results show that tumour cells secrete several key cytokines and chemokines present an array of cell‐contact signals, and form a microenvironment that regulates and prompts the proliferation of Th17 cells. The PHC microenvironment may convert CD4⁺CD25^–CD127⁺ non‐T_regs to CD4⁺FoxP3⁺ T_regs and increased CD4⁺CD25⁺CD127^low T_regs may be converted into IL‐17‐procucing cells in the tumour microenvironment, which is consistent with a previous study on ovarian cancer 30. That is to say, the sources of Th17 cells in tumour tissues may include the trafficking of circulating Th17 cells to tumours and locally induced Th17 cells.

T_regs have been accepted widely as an important suppressive regulation in human tumours, contributing to tumour progression 31. As well as the traditional T_regs, a new discrete subset of B cells, described and confirmed as B_regs 32, has been brought into focus. Numerous researches 33, 34 have been conducted concerning B_regs as negative mediators, both in mice and human autoimmune diseases, but this is still rare in human tumour. In our study, we detected an increased percentage of CD19⁺CD5⁺CD1d^hi B_regs in PHC. As we have shown, enhanced Th17 cells have positive correlations with both T_regs and B_regs; that is to say, Th17 cells increased synchronically with T_regs and B_regs in PHC. The evidence shows that IL‐17 family members play an active role in inflammatory diseases and cancer 29. Th17 cells promote inflammation, T_regs and B_regs suppress the anti‐tumour reaction, and all these cells may promote the development and progression of PHC together. To our knowledge, we are the first to reveal the relationship between Th17 cells and B_regs in human PHC. Our results may provide a new clue to the study of immune regulation in PHC.

Collating our results (as shown in Fig. 6), it is hypothesized that in the tumour microenvironment, tumour cells recruit Th17 cells, prompting the expansion and proliferation of Th17 cells (step 1) and some Th17 cells, and T_regs may convert into each other (step 2); Th17 cells, which can also facilitate the generation of B_regs (step 3), all of which impair the anti‐tumour response, collaborate with tumour growth and promote immune escape in the tumour microenvironment (step 4). Of course, the precise involvement of Th17 cells, T_regs and B_regs in anti‐tumour immunity needs to be elucidated further. These results may provide the basis and new clues for further investigation of the function and mechanisms of Th17, T_regs and B_regs in PHC and, possibly, other human tumours.

Figure 6.

A hypothesis model for the function of T helper type 17 (Th17) cells in primary hepatic carcinoma (PHC) tumour microenviroment. There is a hypothesis model for the interaction of tumour cells, Th17 cells, regulatory T cells (T_regs) and regulatory B cells (B_regs) in the PHC tumour microenvironment. Tumour cells may recruit Th17 cells and promote the expansion and proliferation of Th17 cells in the tumour microenvironment. Th17 cells are crucial for PHC due to Th17 cells and T_regs may convert into each other, facilitating the generation of B_regs, all of which may contribute to tumour progression in the tumour microenvironment. [Colour figure can be viewed at http://wileyonlinelibrary.com]

Disclosure

The authors declare no conflicts of interest.

Supporting information

Additional Supporting information may be found in the online version of this article at the publisher's web‐site:

Fig. S1. The mean fluorescence intensity (MFI) of CD4 and interleukin (IL)‐17 in tumour tissues and peritumoral tissues. MFI was quantified with Image J. (a,b) Representative plots showed that the fluorescence of CD4 and IL‐17 in Fig. 2c was quantified with Image J. (c,d) The MFI of CD4 and IL‐17 was increased in the tumour tissues with primary hepatic carcinoma (PHC) in comparison to peritumoral tissue controls (**P <0·01).

Fig. S2. Membrane co‐stimulatory molecules CD80, CD86 and inducible T cell co‐stimulator ligand (ICOSL) (CD275) in primary hepatic carcinoma (PHC) were evaluated with flow cytometry. CD80 (a), ICOSL (c) expressed in tumour cells derived from tumour tissues were higher than those in non‐tumour cells derived from non‐tumour tissues, while CD86‐positive cells (b) in tumour appeared to be lower than that in non‐tumour.

Fig. S3. Forkhead box protein 3 (FoxP3) expression varied with CD25 and correlated positively with CD4⁺CD25⁺CD127^low cells in normal controls. The majority of FoxP3⁺ cells fell into the cell subset of CD4⁺CD25^high and CD4⁺CD25⁺ cells, especially the CD4⁺CD25⁺CD127^low cells correlated positively with FoxP3 in CD4⁺ T cells of normal controls (r ² = 0·5662, *P < 0·05, n = 10).

Fig. S4. The sorting strategy of CD4⁺CD25⁺CD127^low regulatory T cells (T_regs) and CD4⁺CD25^–CD127⁺ non‐T_regs. Cells stained with CD4 fluorescein isothiocyanate (FITC), CD25 allophycocyanin (APC) and CD127 phycoerythrin‐cyanin 5·5 (PE‐Cy5·5), T_regs and non‐T_regs were gated as shown in plots.

Fig. S5. Significantly increased regulatory T cells (T_regs) in peripheral blood correlate with primary hepatic carcinoma (PHC) tumour–node–metastasis (TNM) stage. (a) CD4⁺forkhead box protein 3 (FoxP3)⁺ expression T_regs in peripheral blood were analysed by flow cytometry. Representative plots showed that the T_reg populations in the gated CD4⁺ cells were increased in the patients with PHC in comparison to normal controls. (b) T_regs in the PHC group (n = 93) were higher than those in the NC group (n = 51), the differences between NC and PHC were significant (***P < 0·0001). (c) T_regs were enhanced with PHC TNM stage progression, the percentage of T_regs in advanced stages (stages III–IV, n = 54) was higher than that in early stages (stages I–II, n = 39) (**P < 0·01).

Fig. S6. T helper type 17 (Th17) cells promoted the generation of CD19⁺CD5⁺CD1d^hi regulatory B cells (B_regs). (a) Representative plots showed that increased CD19⁺CD5⁺CD1d^hi B_regs were analysed in B cells co‐cultured with isolated Th17 cells in the presence of 100 nM cytosine–phosphatase–guanine (CpG) ODN2006 for 48 h. (b) Compared with unco‐cultured CD19⁺B cell controls, increased percentage of CD19⁺CD5⁺CD1d^hi B_regs were detected in B cells cultured with autologous Th17 (n = 4, *P < 0·05).

Table S1. Nuclear sequences of chemokine primes used in real‐time polymerase chain reaction (PCR).