Impairment of CD4+ cytotoxic T cells predicts poor survival and high recurrence rates in patients with hepatocellular carcinoma

Authors


  • Potential conflict of interest: Nothing to report.

  • Supported in part by grants from the National Natural Science Foundation of China (30901328), the National Key Basic Research Program of China (2009CB522500, 2012CB519005), and the National Grand Program on Key Infectious Disease (2012ZX10002-007-002).

Abstract

The role of CD4+ cytotoxic T cells (CTLs) in hepatocellular carcinoma (HCC) remains obscure. This study characterized CD4+ CTLs in HCC patients and further elucidated the associations between CD4+ CTLs and HCC disease progression. In all, 547 HCC patients, 44 chronic hepatitis B (CHB) patients, 86 liver cirrhosis (LC) patients, and 88 healthy individuals were enrolled in the study. CD4+ CTLs were defined by flow cytometry, immunohistochemistry, and lytic granule exocytosis assays. A multivariate analysis of prognostic factors for overall survival was performed using the Cox proportional hazards model. Circulating and liver-infiltrating CD4+ CTLs were found to be significantly increased in HCC patients during early stage disease, but decreased in progressive stages of HCC. This loss of CD4+ CTLs was significantly correlated with high mortality rates and reduced survival time of HCC patients. In addition, the proliferation, degranulation, and production of granzyme A, granzyme B, and perforin of CD4+ CTLs were inhibited by the increased forkhead/winged helix transcription factor (FoxP3+) regulatory T cells in these HCC patients. Further analysis showed that both circulating and tumor-infiltrating CD4+ CTLs were independent predictors of disease-free survival and overall survival after the resection of the HCC. Conclusion: The progressive deficit in CD4+ CTLs induced by increased FoxP3+ regulatory T cells was correlated with poor survival and high recurrence rates in HCC patients. These data suggest that CD4+ CTLs may represent both a potential prognostic marker and a therapeutic target for the treatment of HCC. (HEPATOLOGY 2013)

See Editorial on Page 1

Hepatocellular carcinoma (HCC), one of the most common cancers in the world,1, 2 is characterized by a progressive development and poor prognosis, with 5-year survival rates of less than 5%. Although the effective CD8+ T-cell-mediated cytotoxicity plays a crucial role in controlling cancer development, CD4+ T cells are increasingly considered to contribute to antitumor immune responses through activating CD8+ T cells by way of their cytokine production. CD4+ T-cell cytotoxicity has long been regarded as an artifact, as these observations have been restricted to cell lines or CD4+ T-cell clones generated in vitro.3, 4 However, an increasing number of studies have focused on the role of peripheral CD4+ cytotoxic T cells (CTLs) in viral infections, autoimmune diseases, and malignancies.3-8

CD4+ CTLs are defined as a population of CD4+ T cells that constitutionally express granzyme (Gzm) and perforin and execute direct lytic activity through granular exocytosis.3-5, 9-13 Recent studies have identified peripheral CD4+ CTLs in patients with viral infections, such as human immunodeficiency virus (HIV), cytomegalovirus (CMV), hepatitis B virus (HBV), and hepatitis C virus (HCV).3, 4, 11, 14-16 These cells are also associated with autoimmune diseases, such as rheumatoid arthritis17 and ankylosing spondylitis,18 and circulatory tumors, such as B-cell chronic lymphocytic leukemia.19, 20 In contrast, few CD4+ CTLs can be detected in healthy individuals.3-5, 10 Recently, two groups have demonstrated that the transfer of naïve tumor-reactive CD4+ T cells that did not undergo in vitro manipulation into a mouse model of advanced melanoma significantly induced tumor regression.12, 13 In addition, this antitumor activity was dependent on the direct recognition of target cells through major histocompatibility complex (MHC) class II receptors and the degranulation of Gzm and perforin, but was independent of CD8+ T cells, B cells, natural killer (NK) cells, and NKT cells.12, 13 Similar findings were confirmed in a mouse HCC model.21 However, little information is available regarding either peripheral or intratumor CD4+ CTLs in HCC patients, as well as their associations with HCC progression and survival rates. The regulatory mechanisms that are responsible for the changes in CD4+ CTLs in HCC patients also need to be clarified.

The present study enrolled 547 HCC patients at various stages of disease progression with a homogeneous background of chronic HBV infection and characterized CD4+ CTLs from peripheral blood, tumor-, and nontumor-infiltrating lymphocytes in these HCC patients. We found that HCC patients exhibited an increase in CD4+ CTLs only at early stage disease, but their numbers and activities progressively decreased due to the increased forkhead/winged helix transcription factor (FoxP3+) regulatory T cells (Tregs). More important, the reduced incidence of CD4+ CTLs may represent a promising independent predictor for survival and recurrence in HCC patients. These findings also suggest that CD4+ CTLs may represent a therapeutic strategy for the treatment of HCC.

Patients and Methods

Abbreviations

ALT, alanine aminotransferase; CTLs, cytotoxic T cells; FoxP3, forkhead/winged helix transcription factor; Gzm, granzyme; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; LIL, liver-infiltrating lymphocytes; NC, normal controls; NIL, nontumor-infiltrating lymphocytes; PB, peripheral blood; PBMC, peripheral blood mononuclear cells; TIL, tumor-infiltrating lymphocytes; Treg, regulatory T cells.

Study Participants.

In all, 547 HBV-related HCC patients were enrolled in this study. Blood samples were obtained from 232 HCC patients, 86 liver cirrhosis (LC) patients, 44 chronic hepatitis B (CHB) patients, and 88 of age- and gender-matched healthy donors as normal controls (NC) between 2005 and 2012. Tumor-infiltrating lymphocytes (TILs), nontumor-infiltrating lymphocytes (NILs), and liver-infiltrating lymphocytes (LILs) were isolated from liver tissues of 30 HCC patients and 9 LC patients who had undergone surgical resection or liver transplantation. LILs from eight healthy donors whose livers were used for liver transplantation were also collected. Paraffin-embedded liver sections of resected tumor tissue from 315 HCC patients were used for immunohistochemical staining in our hospital between 2001 and 2004.

The diagnosis of HCC was based on the results of standard biopsies or imaging according to the American Association for the Study of Liver Diseases (AASLD) guidelines.22 A diagnosis of tumor recurrence after resection was based on imaging appearances. The stage of HCC disease was evaluated according to the criteria for the diagnosis and staging of primary liver cancer published by the Chinese Anti-Cancer Association in 2001.23 A comparison of the criteria between the Chinese classification system and the TNM system for the staging of primary HCC has been described.23, 24 The study protocol was approved by the Ethics Committee of the Beijing 302 Hospital, and written informed consent was obtained from each subject prior to blood and tumor sampling. The patients with concurrent HCV and HIV infections, and autoimmune or alcoholic liver disease, were excluded in the study.

Flow Cytometric Analysis, Intracellular Cytokine Detection, and Immunohistochemical Staining.

The material and methods for these techniques are basing on our previously reported protocols24, 25 and are shown in the Supporting Materials and Methods.

Cell Isolation.

CD4+ T cells were isolated from peripheral blood mononuclear cells (PBMCs) using a CD4-positive isolation kit (Miltenyi Biotech, Auburn, CA). CD27+, CD27, CD28+, and CD28 CD4+ T cells were sorted using FACSAria II (Becton Dickinson, San Jose, CA). CD4+CD25+ Treg cells were isolated from PBMCs by CD4-negative selection followed by CD25-positive selection using a CD4+CD25+ T cell isolation kit (Miltenyi Biotech) according to the manufacturer's instructions. The purity of CD4+CD25+ Tregs and CD4+ T cells was ≥90% and 98%, respectively. TILs and NILs were isolated separately based on our previously established method.24, 25

Analysis of CD4+ T Cell Degranulation.

Degranulation of CD4+ T cells was measured with a CD107a mobilization assay according to previous reports.26, 27 PBMCs, PBMC-Treg (Treg depleted), and PBMC-Treg+Treg (Treg added back to the PBMC-Treg population at different ratios) were incubated in RPMI medium containing 10% fetal calf serum (FCS), soluble anti-CD3 (1 μg/mL), and anti-CD28 (1 μg/mL) plus fluorescein isothiocyanate (FITC)-conjugated anti-CD107a. The cells were incubated for 1 hour (37°C, 5% CO2), followed by a 4-hour incubation with monensin (BD PharMingen, San Diego, CA). The cells were then washed, stained with peridin chlorophyll protein (PerCP)-conjugated anti-CD3 and phycorerythrin (PE)-conjugated anti-CD4, and analyzed by flow cytometry.

Detection of Gzm and Perforin Production and T Cell Proliferation.

Gzm and perforin production was measured as described.24 Briefly, CFSE-labeled or nonlabeled PBMCs, PBMC-Treg, and PBMC-Treg+Treg were incubated in RPMI containing 10% FCS plus soluble anti-CD3 (1 μg/mL) and anti-CD28 (1 μg/mL) for 48 hours or 96 hours. The cells were then stimulated with brefeldin A (10 μg/mL) for an additional 5 hours. The cells were then washed, stained for surface markers (CD3 and CD4) and intracellular GzmA, GzmB and perforin, and analyzed by flow cytometry.

Statistical Analyses.

Data were analyzed with SPSS v. 13.0 for Windows software (Chicago, IL) and expressed as mean ± standard deviation (SD) for percentages. The Mann-Whitney U test, Kruskal-Wallis H test, and Wilcoxon signed ranks test were used to compare groups. Actuarial survival rates were analyzed by the Kaplan-Meier method and survival was measured in weeks from diagnosis to death or the last review for the patients who did not receive any antitumor therapy from diagnosis to death. Disease-free survival (DFS) was measured in months from resection to tumor recurrence or the last observation. The overall survival (OS) was measured in months from resection to death or the last review. The log-rank test was applied for comparisons between groups. Multivariate analysis of prognostic factors for OS was made using Cox's proportional hazards model.24, 28 P < 0.05 was considered significant.

Results

Clinical Data of Enrolled HCC Patients.

The clinical data of the HCC patients are shown in Table 1. All of the HCC patients had a history of more than 20 years of chronic HBV infection and had been hospitalized or followed up in Beijing 302 Hospital, China. No patients had received anticancer therapy prior to sampling. The median survival duration was 10.5 weeks (range, 0.75 to 29.5 weeks) for HCC patients with stage III disease, 22 months (range, 1.8 to 63 months) for DFS in HCC patients with stage I and II when circulating CD4+ CTLs were used as an identification marker, and 55 months (range, 1.8 to 116 months) for both DFS and OS in HCC patients when intratumoral CD4+ CTLs were used as an identification marker.

Table 1. Clinical Data of Enrolled 547 of HBV-Related HCC Patients
VariableResults
  • HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; HBeAb, hepatitis B e antibody; CHB, chronic hepatitis B; LC, liver cirrhosis in chronic hepatitis B; ND, no data.

  • *

    HCC patient disease stage was evaluated according to the criterion for diagnosis and staging primary liver cancer constituted by the Chinese Anti-Cancer Association in 2001.

Mean age (year) mean ± SD52.2 ± 10.3
Gender (male/female)476/71
HBsAg (+/−)547/0
HBsAb (+/−)0/547
HBeAg (+/−)123/424
HBeAb (+/−)206/341
HBcAb (+/−)547/0
HBV DNA (+/−/ND)340/137/70
Background (CHB/LC)144/403
Child-Pugh classification (A/B/C)400/120/27
AFP (>400/400-200/≤200)286/85/176
Stage (I/II/III)*297/137/113

Progressive Reduction of CD4+ CTLs Associated With Disease Progression in HCC.

CD4+ CTLs are defined as a population of CD4+ T cells that express GzmA, GzmB, and perforin (Fig. 1A). It was found that the percentages of circulating CD4+ CTLs were significantly higher in HCC patients than in CHB and LC patients and NC subjects (Fig. 1A,B). There was no significant difference in the CD4+ CTL percentages among NC subjects and CHB and LC patients (Fig. 1B). Notably, we found that the proportion of CD4+ CTLs progressively decreased in HCC patients with advanced stage compared with those in early stages (Fig. 1C). These results indicate that a numerical decrease in CD4+ CTLs is associated with the progression of HCC.

Figure 1.

The progressive reduction of circulating CD4+ CTLs was associated with disease progression in HCC patients. (A) The representative prevalence of CD4+ CTLs from individual subjects in the four groups studied. The numbers in each quadrant represent the percentages of each cell population among CD4+ T cells. (B) Statistical analysis showed that the frequency of CD4+ CTLs in HCC patients was significantly higher than that in CHB and LC patients and normal controls. (C) The prevalence of CD4+ CTLs decreased with the progression of HCC. The boxplots in (C,D) show the 25th, 50th, and 75th percentiles. Vertical lines represent the 10th and 90th percentiles. NC, normal controls; CHB, chronic hepatitis B; LC, liver cirrhosis; HCC, hepatocellular carcinoma. **P < 0.01; *P < 0.05.

We then found that the percentage of CD4+ CTLs in TILs was lower than in NILs, and also gradually decreased in both TILs and NILs in HCC patients from stage I, stage II to stage III (Fig. 2A,B). In addition, the CD4+ CTL percentages in LILs and NILs from NC subjects and HCC patients, respectively, were significantly higher than those in peripheral blood, and the gaps of CD4+ CTL percentages between TILs and NILs became increasingly significant during the progression of HCC (Supporting Fig. 1). Immunohistochemical double-staining experiments further confirmed that CD4+GzmA+/GzmB+/perforin+ T cells were preferentially accumulated in nontumor regions rather than tumor regions or normal liver tissues. In contrast, CD4+GzmA/GzmB/perforin T cells were increasingly infiltrated inside the tumors relative to nontumor regions (Fig. 2C). Furthermore, we also found that there was a close correlation among the proportions of CD4+ CTLs in peripheral blood, NILs, and TILs (Supporting Table 1), which indicated that the proportion of peripheral CD4+ CTLs was indicative of the density of their counterparts within the tumor. These data suggest that CD4+ CTLs were redistributed in HCC patients compared with NC, CHB, and LC subjects, and were reduced within the tumors of HCC patients.

Figure 2.

Loss of CD4+ CTLs in TILs associated with disease progression of HCC. Tissue-infiltrating lymphocytes were isolated from liver tissues of 30 HCC patients, nine LC patients, and eight healthy donors. The cells were then collected and stained for flow cytometric analysis. (A) Representative prevalence of CD4+ CTLs from the tissue-infiltrating lymphocytes of an individual HCC patient, an LC patient, and a healthy control. The numbers indicated the percentages of GzA, GzB, and perforin-expressing CD4 T cells. (B) Statistical analysis showed that the frequencies of CD4+ CTLs were significantly higher in NILs than in TILs or LILs, but decreased with the progression of HCC. (C) Representative immunohistochemical double staining in tumor and nontumor specimens from an HCC patient and from the liver tissue of an LC patient and a healthy donor as a normal control. Double staining for GzmA, GzmB, and perforin (blue, in cytoplasm) and CD4 (red, on cell membrane) are shown. Positively stained cells are shown by arrows. NILs, nontumor-infiltrating lymphocytes; LILs, liver-infiltrating lymphocytes; TILs, tumor-infiltrating lymphocytes. **P < 0.01; *P < 0.05.

Decreased Release of Cytolytic Molecules by CD4+ CTLs in HCC Patients.

We further addressed whether the dysfunction of CD4+ CTLs was associated with HCC progression. We found that HCC patients showed a significant decrease in the expression of CD107a (a Gzm and perforin-release marker26, 27) by CD4+ T cells compared with NC subjects and CHB and LC patients (all P < 0.01); in particular, in HCC patients in the advanced disease stages, CD4 T cells displayed even lower levels of CD107a expression (Fig. 3A). Dynamic analysis showed that the CD107a expression was much lower in HCC patients than other groups at 3 hours and 5 hours poststimulation (P < 0.05) (Fig. 3B). Degranulation analysis further showed that the percentage of degranulation of CD4+ T cells in HCC patients was significantly lower than that measured in other groups (P < 0.05) (Fig. 3C,D).

Figure 3.

The release of cytolytic molecules by CD4+ CTLs was decreased in HCC patients. (A) HCC patients showed a significant decrease in CD107a expression by CD4+ T cells compared with normal controls, CHB, and LC patients after anti-CD3/CD28 stimulation for 5 hours. Patients with advanced stage disease showed a significant decrease of CD107a expression than patients with early stage. (B) Dynamic analysis showed that CD107a expression was lower in HCC patients than in normal controls and CHB and LC patients at 3 hours and 5 hours poststimulation. (C) Residual GzmA, GzmB, and perforin were examined after stimulation (5 hours) with anti-CD3/CD28. The percentage degranulation of CD4+ T cells was calculated using the following formula: 1 − (residual enzymes after 5-hour stimulation with anti-CD3/CD28/constitutive expression of enzymes before stimulation) ×100%. (D) Representative fluorescence-activated cell sorting (FACS) staining of percentage degranulation of CD4+ T cells from an individual subject in the four groups, respectively, before and after 5-hour stimulation with anti-CD3/CD28 (black line, before stimulation; red line, after stimulation). The stage of HCC patients studied in (B-D) as below: three patients with stage I disease, four patients with stage II disease, and one patient with stage III disease). **P < 0.01.

Cytolytic Activity of CD4+ CTLs Suppressed by Treg Cells in HCC Patients.

Treg cells have been demonstrated to be increased in HCC patients and associated with HCC progression in our previous study.24 We therefore analyzed how the increased number of Treg cells influenced the cytolytic activity of CD4+ CTLs. First, we found that the numbers of Treg cells were negatively correlated with the numbers of CD4+ CTLs both in the liver and peripheral blood by Spearman analysis (Supporting Fig. 2). Accordingly, immunohistochemical double-staining further demonstrated that there were more FoxP3+ lymphocytes and fewer enzyme+ lymphocytes within the tumor compared with nontumor regions (Supporting Fig. 3). Second, the depletion of Treg cells (PBMC-Treg) resulted in a significant restoration of CD107a expression at 3, 5, and 7 hours poststimulation. The addition of Treg cells into Treg-depleted PBMCs (PBMC-Treg+Treg) resulted in the suppression of CD107a expression in a dose-dependent manner. Additionally, this inhibition was further confirmed to be dependent on cell-to-cell contact in a transwell assay (Fig. 4A). Third, the production of GzmA, GzmB, and perforin by CD4+ T cells from HCC patients increased in PBMC-Treg cells compared with PBMCs after 48 hours of stimulation with anti-CD3/CD28, and were decreased in the PBMC-Treg+Treg group compared with the PBMC-Treg group, which was also dependent on cell-to-cell contact (Fig. 4B). Fourth, the production of GzmA, GzmB, and perforin by new CD4+ T cells from HCC patients was also enhanced following anti-CD3/CD28 stimulation for 4 days when Treg cells were depleted from PBMCs (Fig. 4C). These data strongly suggest that the cytolytic capability of CD4+ CTLs can be markedly suppressed by Treg cells by way of the inhibition of the release and self-renewal of cytolytic molecules, as well as by the prevention of a new generation of CD4+ CTLs.

Figure 4.

Treg cells suppress the cytolytic molecule release and production by autologous CD4+ CTLs in HCC patients. The expressions of CD107a (A) and GrA, GrB, and perforin (B) were detected in PBMCs, PBMCs-Treg cells, and PBMCs-Treg+Treg cells at different ratios for 3 hours, 5 hours, and 7 hours and for 48 hours, respectively, followed by anti-CD3/CD28 stimulation. Transwell experiments were also simultaneously performed. (C) Cell cycle analysis showed that Treg cells have the potential to inhibit CD4+ T-cell proliferation and Gzm/perforin generation by proliferative CD4+ T cells after anti-CD3/CD28 stimulation for 96 hours. Cells in the plots were gated on CD4+ T cells. Data were obtained from a representative HCC patient.

Decreased Numbers of CD4+ CTLs Predict Poor Survival in HCC Patients.

To investigate the association between CD4+ CTLs and HCC progression, 83 HCC patients with stage III disease were divided into two groups (the high CD4+ CTLs and low CD4+ CTLs groups), according to the median percentage of circulating CD4+ CTLs. The analysis showed that the low CD4+ CTL group patients had significantly poorer survival rates compared with the high CD4+ CTL group patients (P < 0.001) (Fig. 5A). In addition, we analyzed the association between peripheral CD4+ CTL percentages and HCC recurrence after resection in 100 HCC patients with stage I and II who underwent tumor resection and were followed until tumor recurrence. The data showed that the DFS rate in the high CD4+ CTL group patients was significantly higher than in the low CD4+ CTL group patients (P < 0.01, Fig. 5B). Cox's proportional hazards model analysis revealed that the GzmB+ and perforin+ CD4+ CTLs were independent prognostic factors for survival of HCC patients with stage III, and the hazard ratio (HR) was 0.391 (95% confidence interval [CI], 0.202-0.757; P = 0.005) and 0.373 (95% CI, 0.198-0.702; P = 0.002) for GzmB+ and perforin+ CD4+ CTLs, respectively (Table 2). Circulating GzmB+CD4+ CTLs were also independent prognostic factors for DFS in HCC patients with stage I and II (HR, 0.097; 95% CI, 0.021-0.438; P = 0.002), as well as disease stage (HR, 1.756; 95% CI, 1.032-2.772; P = 0.023) (Table 2). However, circulating GzmA+ and perforin+CD4+ T cells were not found to be independent prognostic factors for DFS in these HCC patients.

Figure 5.

The prevalence of CD4+ CTLs predicts the survival in HCC patients. (A) Eighty-three HCC patients with advanced stage were divided into two groups according to the median percentage of peripheral CD4+ CTLs. The low percentages of the CD4+ CTL group had significantly poorer survival rates compared to the high percentages of the CD4+ CTL group (P ≤ 0.001). (B) One hundred HCC patients with early stage (n = 53 for stage I, n = 47 for stage II) who had undergone tumor resection were divided into two groups according to the median percentage of circulating CD4+ CTLs. The DFS in the high percentages of CD4+ CTL group was significantly higher than that in the low percentages of CD4+ CTL group (P < 0.01). (C,D) Intratumoral GzmB-positive CD4+ CTLs were detected from 315 HCC patients by immunohistochemical double-staining for GzmB and CD4. All of these patients were divided into two groups according to the median number of GzmB+CD4+ T cells, termed the high (n = 151) and low (n = 164) percentages of CD4+ CTL groups. The low percentages of the CD4+ CTL group had significantly poorer DFS and OS rates compared to the high percentages of CD4+ CTL group (P < 0.001). Patients who failed to attend follow-up or died from known tumor-unrelated causes were excluded from the record. Actual OS rates were analyzed by the Kaplan-Meier method and survival was measured in weeks from diagnosis to death. DFS and OS were measured in months from tumor surgical resection to recurrence or death, respectively. The log-rank test was applied for comparisons between the groups. A multivariate analysis of prognostic factors for OS was made using Cox's proportional hazards model. DFS, disease-free survival; OS, overall survival.

Table 2. Multivariate Analysis of Factors Associated with Survival or Recurrence Using the Cox Proportional Hazard Regression Model
VariableCirculating Cytotoxic CD4+ T CellsIntratumoral Cytotoxic CD4+ T Cells
Survival for Stage IIIDFS for Stage I and IIDFSOS
Hazard Ratio95% CIPHazard Ratio95% CIPHazard Ratio95% CIPHazard Ratio95% CIP
  1. DFS, disease-free survival; OS, overall survival; CI, confidence interval; Gzm, granzyme; NA, not applicable.

Age, years (>50 vs. ≤50)1.0130.606–1.6920.9611.0600.543–2.0720.8640.8960.682–1.1780.4230.7440.549–1.0080.056
Sex (male vs. female)1.4390.639–3.2390.3790.7310.301–1.7750.4881.0220.644–1.6230.9261.1580.712–1.8840.555
AFP, ng/mL (>200 vs. ≤200)1.5700.895–2.7540.1161.7420.861–3.5260.1231.0890.834–1.4210.5310.9610.718–1.2880.791
LC (yes vs. no)0.5320.273–1.0380.0641.3750.698–2.7090.3581.2070.843–1.7290.3041.1110.757–1.6280.591
Tumor size, cm (>5 vs. ≤5)0.7610.393–1.4760.4202.9240.817–10.4730.0990.6760.395–1.1580.1540.9230.599–1.4210.715
Tumor number (multiple vs. single)1.4760.884–2.4650.1370.7360.273–1.9840.5441.2950.933–1.7950.1221.3070.911–1.8760.146
Tumor metastasis (yes vs. no)2.6631.375–5.1560.004NANANANANANANANANA
Child-Pugh score (B vs. A)0.9290.557–1.5490.7781.2630.345–4.6180.7241.6001.033–2.4770.0350.8470.448–1.5990.607
Stage (II vs. I)NANANA1.7561.032–2.7720.0232.1611.220–3.8280.0083.0711.689–5.583<0.001
GzmA+ (High vs. Low)0.7480.407–1.3750.3500.6910.245–1.9510.485NANANANANANA
GzmB+ (High vs. Low)0.3910.202–0.7570.0050.0970.021–0.4380.0020.6970.524–0.9260.0130.5970.443–0.8040.001
Perforin+ (High vs. Low)0.3730.198–0.7020.0020.5660.154–2.0880.393NANANANANANA

The association between intratumoral CD4+ CTLs and DFS or OS was further investigated by immunohistochemical double-staining in 315 HCC patients. The results showed that the low GzmB+CD4+ T cells group patients had significantly poorer DFS and OS in comparison to the high group of patients (P < 0.001) (Fig. 5C,D). Cox's proportional hazards model showed that GzmB+CD4+ T cells were independent prognostic factors for both DFS and OS (HR, 0.697; 95% CI, 0.524-0.926; P = 0.013 for DFS; HR, 0.597; 95% CI, 0.443-0.804; P = 0.001 for OS) (Table 2). It was also found that the disease stage was an independent prognostic factor for DFS and OS, whereas the Child-Pugh score was an independent prognostic factor for DFS in these HCC patients (Table 2). These results suggest that both circulating and intratumoral CD4+ CTLs are numerically associated with disease progression and therefore represent predictors of poor survival in HCC patients.

Discussion

CD4+ CTLs have been demonstrated to exert antitumor activity in mice through granular exocytosis,12, 13, 21 and their therapeutic potential in cancer was recently emphasized.29-32 However, limited information is available on the functional roles of these cells in human cancers. This study comprehensively characterized CD4+ CTLs in vivo in HCC patients and found that reduced numbers of CD4+ CTLs are associated with poor survival and a high recurrence of HCC.

The present study indicated that CD4+ CTLs were enriched in nontumor regions, and were significantly increased in early stage HCC patients. Furthermore, the loss of CD4+ CTLs was closely associated with HCC disease progression. We also found that CD4+ CTLs predominantly expressed interferon-gamma (IFN-γ) and tumor necrosis factor alpha (TNF-α) (Supporting Fig. 4C). These data suggest that these cells might participate in antitumor immunity. Most important, circulating and tumor-infiltrating CD4+ CTLs in HCC patients exhibited a strong prognostic value for survival times in naturally progressing HCC patients, and in terms of the DFS and OS rate in patients who had undergone surgical resection. The Cox's proportional hazards model showed that CD4+ CTLs are independent prognostic factors for naturally progressing survival, DFS, and OS rates. Taken together, these results strongly indicate that the number of CD4+ CTLs is a prognostic marker for human HCC progression.

Notably, we found that there were intrinsic qualitative defects in CD4+ CTLs through the detection of CD107a mobilization. CD107a is a lysosomal-associated membrane glycoprotein that surrounds the core of the lytic granules in cytotoxic T cells. Upon T cell receptor (TCR) engagement, CD107a is exposed on the cell membrane of cytotoxic T cells.26, 27 The surface mobilization of CD107a by CD4+ T cells is associated with the release of cytolytic granules.27 Our study indicated that CD4+ CTLs from HCC patients showed significantly lower levels of exocytosis of cytolytic molecules in response to TCR engagement compared with other groups of subjects. As such, the degranulation of CD4+ CTLs in HCC patients was functionally impaired and led to a small release of stored perforin and Gzm proteins from these cells.

This study also elucidates the possible mechanism that underlies the functional impairment of CD4+ CTLs in HCC patients. Our data support the notion that the increased numbers of Treg cells may potentially impair CD4+ CTL function. On the one hand, the number of CD4+ CTLs in TILs, NILs, and peripheral blood were negatively correlated with an increase in the number of Treg cells. Our previous studies indicated that FoxP3+ Treg cells in TILs, NILs, and peripheral blood were significantly increased with the progression of HCC.24 On the other hand, Treg cells have the potential to inhibit cytolytic molecule release in CD4+ CTLs and block the new generation of CD4+ CTLs in a dose- and cell-cell contact-dependent pattern. Furthermore, the high levels of CD95 and PD1 expression by CD4+ CTLs also implies that they have additional regulatory mechanisms (Supporting Fig. 4A). Future studies should determine which factors are responsible for the suppression of CD4+ CTLs in HCC patients. These data also suggest that the target of CD4+ CTLs in vivo could facilitate the boosting of the antitumor responses in HCC patients.

It is currently not known why CD4+ CTLs are increased in HCC patients with early stage disease. We found that the frequency of CD4+ CTLs in CHB and LC patients was much lower than in HCC patients, which was in accordance with the findings of a previous study15 that showed chronic HBV infection was not the principal reason for increased numbers of CD4+ CTLs in HBV-associated HCC patients. Three reasons may be involved in the increase in CD4+ CTL numbers in HCC patients: (1) the suppression of traditional cytotoxic immune cells might induce feedback compensation for the high incidence of CD4+ CTLs in HCC patients. For example, the cytolytic activity of CD8+ T lymphocytes and NK cells in HCC patients is significantly abrogated during tumorigenesis.24, 33, 34 Indeed, Williams and Engelhard35 found that CD4+ T cells develop perforin-dependent cytotoxicity only in the absence of activated CD8+ T cells; (2) Abnormal immune activation due to the chronic inflammatory microenvironment is thought to be another major driving factor that induces CD4+ CTLs differentiation. Numerous reports have demonstrated that the presence of increased numbers of CD4+ CTLs is associated with chronic inflammatory processes, such as chronic viral infection or autoimmune diseases.5, 15, 17, 18, 36 Additionally, inflammation is also involved in all stages of tumor development and correlates with poor survival rates in HCC.37-40 Consistent with this hypothesis, we found that CD4+ CTLs in HCC patients were highly activated (high levels of CD38 and HLA-DR expression) (Supporting Fig. 4A); and (3) the increased numbers of CD4+ CTLs in tumor and nontumor regions may also be due to their recruitment into the liver from the peripheral blood, which is a similar finding to the previously reported role of CD8+ T cells.41 Future studies are warranted to confirm these hypotheses.

In summary, this study demonstrated that a progressive decrease in the number of CD4+ CTLs was closely associated with HCC progression and poor survival rates in HCC patients. The intrinsic defects and extrinsic suppression by increased Treg cells may involve the impairment of CD4+ CTLs in HCC patients. These data highlight the novel role of CD4+ CTLs in the immunocompromised status of HCC patients, and also provide a potential therapeutic target for the treatment of HCC.

Ancillary

Advertisement