Tumour-Infiltrating FoxP3+ and IL-17-Producing T Cells Affect the Progression and Prognosis of Gallbladder Carcinoma after Surgery



Tumour-infiltrating lymphocytes (TILs) have been found to play crucial roles in a series of cancers. However, the impact of these cells on gallbladder carcinoma (GBC) remains poorly understood. In this study, we examined infiltrating FoxP3+, IL-17+, CD4+ and CD8+ cells by immunohistochemistry in specimens of 104 patients with GBC and evaluated the association of these cells with clinicopathological features and prognosis. The number of FoxP3+ cells was increased in a stepwise manner from CC to GA and GBC (GA versus CC,= 0.036; early GBC versus GA,= 0.032; advanced versus early GBC,= 0.025). Both intratumoral FoxP3+ and IL-17+ cells correlated with nodal metastasis and TNM stage. Additionally, there were more infiltrating FoxP3+ cells in specimens with distant metastasis (= 0.014). The group with high FoxP3+ cells showed poor overall survival (OS,< 0.001) and disease-free survival (DFS,< 0.001), and high infiltration of IL-17-producing cells was also a predictor of poor OS (= 0.024). Multivariate analysis revealed that the presence of intratumoral FoxP3+ cells was an independent prognostic indicator for poor DFS (< 0.01). In summary, these findings indicate that FoxP3+ and IL-17+ cells cooperatively facilitate pathogenesis and progression of GBC and show prognostic significance for OS or DFS.


Gallbladder carcinoma (GBC) is the most common and lethal malignancy of biliary tract, although its incidence is lower than gastrointestinal cancers. Despite remarkable advances in imaging modalities, GBC is usually diagnosed at an advanced stage due to lack of specific manifestations. A radical resection is currently the most effective and potentially the only curative therapy, but unfortunately most patients undergo recurrences after surgery, and the overall survival remains poor, ranging from 23.8% to 36% in different reports from Asian countries [1-3]. Thus, unravelling the nature and precise mechanism of gallbladder carcinogenesis to develop novel diagnostic or therapeutic strategies is urgently recommended.

Nowadays, microenvironment has been increasingly suggested to play an important role in the initiation and progression of many cancers. Stroma-infiltrating lymphocytes, including CD4+ and CD8+ T cells, constitute at least a part of the tumour environment and represent host antitumour immunity. CD8+ T cells, defined as cytotoxic T cells, possess the capability to kill target cells by releasing perforins and granzymes in acquired immune responses, thereby playing a critical role in antitumour immunity.

CD4+ T cells, which involve in orchestrating host immunity by releasing distinct cytokines, can be differentiated into four types: Th1, Th2, Th17 and regulatory T cells (Tregs) based on the local cytokine profiles [4-6]. Tregs usually function by maintaining self-tolerance and regulating host immune responses against infection and tumour [7, 8]. Transcriptional factor forkhead box P3 (FoxP3) is critical for the development and function of Tregs and thus regarded as the specific marker for Tregs. Accumulating evidence showed an increased level of FoxP3+ cells in peripheral blood or resident tissues in a wide range of cancers [9-15]. Additionally, in mice models, depletion or attenuation of these cells inhibited tumour growth and invasion [16]. Therefore, intratumoral FoxP3+ cells have a pivotal impact on carcinogenesis and progression. As another part of CD4+ T lymphocytes, T helper 17 (Th17) cells are characterized of producing IL-17 and expressing transcription factor RORγ. Although IL-17-producing cells are closely related to inflammatory and autoimmune diseases [17, 18], the role of these cells in tumorigenesis remains elusive and controversial in different cancers [19-23].

More recently, the balance status between FoxP3+, IL-17+ and CD8+ cells is thought to be a key regulator in host antitumour immune responses, and the skewed balance among these immune cells has been observed in several cancers [24-27]. However, little is known of the distribution, value and balance status of these immune cells in GBC. In our series, we sought to elucidate the exact role of FoxP3+, IL-17+ and CD8+ cells in the pathogenesis, progression and prognosis of GBC.

Materials and methods

Patients and specimens

Paraffin-embedded specimens from 104 patients with GBC who underwent surgical resection in Tianjin Nankai hospital between October 2006 and October 2011 are enrolled in the study, whereas patients with concurrence of autoimmune diseases or without complete clinicopathological data were excluded. It has been generally accepted that the unresolved inflammation may result in the atypical hyperplasia, which may further promote carcinogenesis, so 28 patients with CC and 28 patients with GA a well-established precursor lesion of GBC were both taken as controls. The characteristics of patients with GBC are summarized in Table 1. The clinical TNM stage of GBC is on the basis of the 7th edition of AJCC cancer staging manual. The study protocol was approved by the Ethics Committee of Tianjin Nankai Hospital, and written informed consent was obtained from each patient.

Table 1. Clinicopathologic characteristics of patients with GBC
AgeMean ± SD66.13 ± 11.88
Degrade of differentiationWell/Moderate/Poor30/40/34
Depth of invasionT1/T2/T3/T45/37/58/4
Nodal metastasisNegative/Positive88/16
Distant metastasisNegative/Positive92/12
Perineural invasionNegative/Positive83/21
Vascular or lymphatic invasionNegative/Positive78/26
TNM stageI/II/III/IV7/37/48/12


Follow-up data were available in all cases, ranging from 2 to 76 months (mean, 39 months). The last follow-up was completed on 20 February 2013. All patients with GBC were monitored prospectively for CA19-9 (A tumour-associated antigen), abdominal ultrasonography and chest X-ray every 1–6 months according to the postoperative time. For patients with clinical manifestations or test results indicative of relapse, computed tomography (CT) or magnetic resonance imaging (MRI) was applied to verify whether it was true. A diagnosis of relapse was based on the representative CT or MRI images. Overall survival (OS) and disease-free survival (DFS) were defined as the interval from surgery to death and that from surgery to the first relapse, respectively. In addition, the interval between surgery and the last observation when patients are still alive or do not relapse is also viewed as OS or DFS.

Immunohistochemical assay

Paraffin-embedded, formalin-fixed samples were consecutively cut into 5-μm-thick sections and processed for immunohistochemical staining as previously described [9, 10]. Briefly, sections were first deparaffined and hydrated. After antigen retrieval in a pressure cooker with citrate buffer for 3 min, tissue sections were treated with 3% hydrogen peroxide for 10 min to block the endogenous peroxidase activity. Monoclonal antibodies against FoxP3 (1:100; Abcam, Cambridge, UK), IL-17 (AF-317-NA; R&D systems, Abingdon, UK), CD8 (1:50; Abcam, Cambridge, UK) and CD4 (1:50; Zhongshan Goldenbridge Biotech, Beijing, China) were used for incubating sections at 4 °C overnight. After secondary antibody staining, diaminobenzidine (DAB) was used as the chromogen for 3 min and then the nuclei were counterstained with haematoxylin.

The distribution of FoxP3+, IL-17+, CD4+ and CD8+ T lymphocytes was observed using the optical microscope (BX51; Olympus, Tokyo, Japan) by two senior independent pathologists without prior knowledge of the clinical data. The number of these cells was respectively counted using image-pro plus 5.1 software (Media Cybernetics Inc, Silver Spring, MD, USA) in five randomized high-power fields (400×), and average number of each lymphocyte was calculated.

Statistical analysis

Statistical analyses were performed with spss 17.0 software (SPSS, Chicago, IL, USA). Comparisons between groups were examined by Mann–Whitney U-test, Student's t-test or χ2 test. Survival curves were plotted using the Kaplan–Meier method and analysed by the log-rank test. Multivariate analyses of prognostic factors for OS and DFS were conducted using the Cox proportional hazards regression model. Results were considered significant when < 0.05.


The distribution of tumour-infiltrating T lymphocytes

In this study, we examined the absolute number of FoxP3+, IL-17+, CD4+ and CD8+ cells by immunohistochemistry. The representative images of each lymphocyte in GBC, GA and CC tissues were shown in Fig. 1.

Figure 1.

Detection of infiltrating FoxP3+, IL-17+, CD4+ and CD8+ T cells in paraffin-embedded tissues. B–E, G–J and L–O refer to FoxP3+, IL-17+, CD4+ and CD8+ cells in GBC, GA and CC tissues, respectively. A, F and K refer to HE staining in GBC, GA and CC tissues, respectively (Original magnification, 400×).

Intratumoral FoxP3+ cells were synchronically increased according to disease progression (GA versus CC, = 0.037; early GBC versus GA, = 0.032; advanced versus early GBC, = 0.025, Fig. 2A). Likewise, intratumoral IL-17-producing cells were slightly increased from CC to GA and GBC, but there was no statistical difference. Most importantly, intratumoral IL-17-producing cells were strikingly elevated from early to advanced stage. The infiltration of these cells at advanced stage was higher than that in CC, GA or early-stage group (versus CC, = 0.036; versus GA, = 0.040; versus early GBC, = 0.045, Fig. 2B). Besides, intratumoral CD8+ cells in GBC and CC were both significantly higher than that in GA (versus CC, = 0.043; versus GBC, = 0.038, Fig. 2D). However, there was no remarkable difference concerning the number of CD4+ cells among specimens from patients with CC, GA and GBC (data not shown).

Figure 2.

The counts of inline image and inline image and the ratio of inline image/inline image and CD8+ T in sample tissues. A–B refers to the absolute number of FoxP3+ and IL-17+ cells in CC, GA, early and advanced GBC, respectively; C refers to the ratio of inline image/inline image in CC, GA, early and advanced GBC; D refers to the absolute number of CD8+ cells in CC, GA and GBC (*< 0.05, **< 0.01, ***< 0.001).

Although the increasing trend occurred in both FoxP3+ and IL-17+ cells, the ratio of inline image/inline image was gradually decreased following the process from CC to GA and GBC (GA versus CC, = 0.040; early GBC versus advanced GBC, = 0.034, Fig. 2C). Nevertheless, intratumoral FoxP3+ cells did not correlate with IL-17-producing cells. Additionally, the ratio of inline image/inline image was found not to be associated with carcinogenesis of cancer (data not shown).

Correlation between FoxP3+, IL-17-producing cells and clinicopathological features

To address whether TILs are associated with the progression of GBC, we analysed the correlation between intratumoral TILs and clinicopathological features. Notably, the infiltration of FoxP3+ cells was positively associated with nodal, distant metastasis and TNM stage (= 0.030, = 0.014, = 0.017, respectively, Table 2). Besides, we found that in specimens with nodal metastasis or advanced stage (III-IV), intratumoral IL-17-producing cells were significantly higher (= 0.045, = 0.020, respectively, Table 2).

Table 2. Correlation between tumor-infiltrating FoxP3+ and IL-17 producing cells and clinicopathologic features of 104 patients with GBC.
Parameters inline image inline image
Low (= 52)High (= 52) P Low (= 50)High (= 54) P
  1. The number in bold refers to statistical significance.

Gender (male/female)19/3322/300.54721/2930/340.605
Age, Mean ± sd, years65.1 ± 12.867.0 ± 11.10.41665.8 ± 12.766.4 ± 11.10.784
CA19-9, Median, U/ml46.4340.460.73640.7547.170.794
Depthof invasion (T1–T2/T3–T4)25/2717/350.11024/2618/360.128
Differentiation (WD/MD/PD)17/23/113/17/220.11217/18/1513/22/190.535
Nodal metastasis (–/+)48/440/12 0.030 46/442/12 0.045
Distant metastasis (–/+)50/242/10 0.014 45/547/70.637
Perineural invasion (–/+)45/738/140.08743/740/140.130
Vascular lymphatic invasion (–/+)42/1036/160.17440/1038/160.257
TNM stage (I–II/III–IV)28/2416/36 0.017 26/2416/38 0.020

Prognostic value of FoxP3+ and IL-17-producing cells in GBC

OS and DFS were evaluated in patients with GBC, and the 5-year survival and disease-free survival rate were 22.6% and 13.5%, respectively. Remarkably, we identified the presence of high FoxP3+ cells as a prognostic factor for poor OS and DFS (< 0.001, respectively, Fig. 3A,B). Moreover, patients with high intratumoral IL-17 cells showed an unfavourable DFS (= 0.024, Fig. 3C).

Figure 3.

Survival curves for univariate analyses. A–B refers to OS and DFS of low versus high FoxP3+ cells; C–D refers to OS and DFS of low versus high IL-17-producing cells; E–F refers to OS and DFS of low versus high ratio of inline image/inline image.

As FoxP3+ and IL-17-producing cells are reciprocally antagonistic and contrarily regulated in gallbladder carcinogenesis and progression, we also analysed the correlation of the ratio of inline image/inline image with OS or DFS. As a result, it was found to be prognostic for OS. Patients with higher ratio of inline image/inline image showed longer OS (= 0.033, Fig. 3E). In addition, we did not observe a prognostic significance of the ratio of inline image/inline image for OS or DFS (data not shown).

To explore independent prognostic factors affecting OS and DFS, multivariate analysis was performed using the Cox proportional hazards model, and univariate variables that correlated with prognosis were also adopted as covariates. The results indicated that the high infiltration of FoxP3+ cells was an independent prognostic factor for poor DFS (= 0.003, Table 3. Note: NA, not adopted as a covariate for multivariate analysis).

Table 3. Univariate analyses of prognostic factors associated with overall survival and disease free survival
VariablesOverall survivalDisease free survival
Age, years (>67.5/≤67.5)1.460.148 NA1.350.183 NA
Sex (Male/Female)0.940.941 NA0.950.829 NA
CA199, U/ml (>44.25/≤44.26)1.390.212 NA1.360.183 NA
Differentiation(PD/MD/WD)1.170.345 NA1.100.514 NA
Vascular lymphatic invasion (+/–)1.060.841 NA1.070.791 NA
Perineural metastasis (+/–)0.920.820 NA1.080.789 NA
Depth of invasion (T1/T2/T3/T4/)1.910.0031.050.9161.490.0320.730.480
Nodal metastasis (+/–)2.390.0051.780.1082.200.0062.000.030
Distant metastasis (+/–)3.260.0013.200.0272.790.0012.760.025
TNM stage (I/II/III/IV)1.890.0010.920.8541.590.0041.230.620
inline image (High/Low)2.580.0011.990.1182.50<0.0012.170.003
inline image (High/Low)1.810.0281.800.0661.500.074 NA
Ratio of inline image/inline image (High/Low)0.580.0380.860.7010.690.104 NA


Stroma lymphocytes, a constitutive part of microenvironment, have been identified as a manifestation of host immune responses against pathogens or cancers. It is clear that lymphocytes are functionally heterogeneous and consist of various immune cell subgroups, which may suppress or promote carcinogenesis. So far, FoxP3+ cells have been suggested to suppress protective antitumour immunity through direct cell–cell contact or cytokine-mediated mechanism [28]; thus, it is not surprising that intratumoral FoxP3+ cells were associated with a poor outcome in a wide range of cancers [9, 10, 26, 27]. Consistently, we demonstrated for the first time that FoxP3+ cells were enriched in GBC and had an impact on pathogenesis and progression of GBC. Interestingly, in other cancers, infiltrating FoxP3+ cells were believed to be a favourable predictor [29] or did not affect any prognosis [30]. The discrepancy could arise from the differences in methodologies or biological properties of cancers. The FoxP3 was originally known as a nuclear marker of CD4+ CD25+ regulatory cells, but compelling data have showed that it can also be expressed in the effector or other helper T cells [31, 32], indicating that apart from Tregs, FoxP3+ cells incorporate several other subgroups, albeit to a minority, some of which may express cytotoxic molecules such as granzymes and perforins. Indeed, Ronald J reviewed 58 studies including 16 different malignant tumours and found that the prognostic value of FoxP3 varies widely even in the same cancer type. Nevertheless, in four of these studies in which double labelling of FoxP3 combined with CD4, CD25 or CD8 was applied, the double-positive T cells were generally associated with a poor prognosis except colorectal cancer [33]. Additionally, these different clinical outcomes suggested that FoxP3+ cells might play a dual role in carcinogenesis. In our opinion, intratumoral FoxP3+ cells extensively suppress host immune responses including antitumour and anti-inflammation immunity via inactivating effector T cells. It appears that FoxP3+ cells may be harmful by suppressing the antitumour immunity and can also benefit the host by modulating inflammation, which is linked to the angiogenesis of cancer. Consequently, the ultimate effect of FoxP3+ cells attributes to which role of suppressing antitumour or anti-inflammation prevails on earth during the carcinogenesis.

IL-17-producing cells also seem to play a dual role in carcinogenesis. On the one hand, IFNγ and IL-17 secreted by Th17 cells synergize to increase the production of CXCL9 and CXCL10, which in turn trigger the recruitment of Th1, NK and CD8+ cells into the tumour microenvironment, and these effector cells play an active role in anticancer immunity [34]. Thus, the presence of intratumoral IL-17-producing cells was correlated with prolonged survival in oesophageal and gastric carcinoma [35, 36]. On the other hand, Th17-cell-derived cytokines such as IL-6 nourish tumour cells and foster the pathogenesis and progression of cancer via IL-6–STAT3 or VEGF signalling pathway [37, 38]. Finally, infiltration of high IL-17-producing cells is viewed as a poor prognostic factor in other cancers [19, 20]. In the present study, patients with advanced GBC revealed higher infiltration of IL-17-producing cells compared with the early stage, and the high infiltration of these cells was a significant predictor of poor OS. In our opinion, IL-17-producing cells are a double-edged sword, and whether they play a potent or negative role in carcinogenesis appears to correlate with the property of the tumour cell itself. In malignancies such as hepatocellular carcinoma and GBC, which are inclined to metastasize at early stage via lymphatics or vessels, IL-17-producing cells are more likely to facilitate pathogenesis and progression of cancers through increasing angiogenesis or lymphangiogenesis, whereas in cancers with less invasiveness, they are responsible for suppressing carcinogenesis by secreting tumour inhibitors.

Interestingly, we found intratumoral CD8+ cells downregulated from CC to GA, but then elevated at the stage of GBC. This is because tumorigenesis is a complex, multifactorial and multistep process. As a component of immune effector T cells, CD8+ cells play an antitumour role, thus the infiltration of these cells diminishing in the precancerous lesion of GA. However, with more IL-17+ cells storing up, which may recruit CD8+ cells in tumour [36], intratumoral CD8+ cells elevated in GBC, exerting a positive role against the immunosuppression induced by FoxP3+ cells.

Previous studies indicated that FoxP3+ and IL-17+ cells inversely correlated with each other and might be contrarily regulated during differentiation [39]. By contrast, we did not find a correlation between FoxP3+ and IL-17+ cells (data not shown). Nonetheless, the ratio of inline image/inline image gradually decreased following the process from CC to GA and GBC, in line with a previous study concerning cervical cancer [27]. As described above, FoxP3+ and IL-17+ cells play antagonistic roles in the process of inflammation, although they both promote tumorigenesis by different mechanisms. In the stage of CC, inflammation prevails, so the balance is tipped towards IL-17-producing cells, but then the balance is tipped in favour of FoxP3+ cells following the progression despite the similar increasing trend, indicating that the microenvironment has been transformed from the state of inflammation to cancer. Furthermore, our study identified the lower ratio of inline image/inline image as a predictor of poor OS. Not all consistent with our result, recent reports in gastric and breast cancer [24, 25] showed that IL-17+ and FoxP3+ cells accumulated in early stage and then the infiltration of IL-17-producing cells gradually decreased according to the progression. From our perspective, the dynamic distribution of these immune cells varies with the type of malignancy due to different properties of tumour biology.


Our study demonstrated for the first time that intratumoral FoxP3+ and IL-17-producing cells cooperate and facilitate the pathogenesis and progression of GBC. Most importantly, the presence of high FoxP3+ cells was an independent prognostic factor for poor DFS. Thus, the development of strategies to deplete or attenuate FoxP3+ or IL-17-producing cells may be a novel therapy for GBC.


The authors thank research staff in the laboratories of Nankai hospital for their contributions to the investigation, meanwhile appreciate the department of pathology for providing specimens and technical assistance.