CD8+CXCR5+T cells infiltrating hepatocellular carcinomas are activated and predictive of a better prognosis

High infiltration of CD8+CXCR5+ T cells into HCC tumors predicts a better prognosis

To assess the presence of CD8+CXCR5+ T cells within HCC tumors, we used flow cytometry to analyze the CD8+CXCR5+ T cell content in 20 healthy blood samples and 40 HCC specimens (Table 1), each of which included blood, peritumoral liver, and tumor tissue samples the same patient. Regardless of whether blood or tumor was assessed, the majority of CD8+ T cells were CXCR5-, though a clear CD8+CXCR5+ T cell population was detected. In HCC patients, the percentage of CD8+CXCR5+ T cells was obviously larger in tumor tissue than in the peritumoral liver or blood. There was no significant difference in the numbers CXCR5+ cells among circulating CD8+ T cells in HCC patients and healthy donors (Figure 1A and 1B). Importantly, the percentage of tumor-infiltrating CD8+CXCR5+ T cells correlated negatively with microvascular invasion (Figure 1C) and early recurrence (Figure 1D and Table 2). Moreover, the levels of PD-1 and ICOS markers expressed by CD8+CXCR5+ T cells were significantly higher in tumor tissue than in the matched blood or peritumoral liver tissues or in healthy blood samples (Figure 1E and 1F and Supplementary Figure 1). Together, these data demonstrate that CD8+CXCR5+ T cells strongly infiltrated HCC tumors and predicted a better prognosis.

Figure 1. Strong infiltration of CD8+CXCR5+ T cells into HCC tumors predicts a better prognosis. Fresh samples were stained with anti-CD8, anti-CXCR5, anti-PD-1 and anti-ICOS antibodies. (A–B) Following gating of CD8+ T cells, the frequencies of CD8+CXCR5+ T cells from healthy PBMCs (n=20), and matched HCC tumor tissue, peritumoral liver tissue and PBMC samples (n=40) were analyzed. (A) One representative experiment is shown. (C) Association of tumor-infiltrating CD8+CXCR5+ T cells with microvascular invasion (n=25 for positive, n=15 for negative) is shown. (B–C) The data indicate the median with the interquartile range. (D) Patients were divided into two groups (Low/High) based on the median of the tumor-infiltrating CD8+CXCR5+ T cell percentages. The early recurrence rate was compared between the two groups using the log-rank test. (E–F) PD-1 and ICOS expression by CD8+CXCR5+ T cells differed among tumor tissue and matched peritumoral tissues and peripheral blood (n=19). The data indicate the median with the interquartile range. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 determined using the Mann-Whitney U test (B, C, E and F).

Strong infiltration of IL-21-producing CD8+CXCR5+ T cells in HCC correlates with disease stage

Numbers of IL-21-producing CD8⁺CXCR5⁺ T cells are significantly increased in chronic hepatitis B (CHB) [14].

We investigated whether IL-21 was produced by CD8+CXCR5+ T cells in HCC. We found that numbers of IL-21-producing CD8+CXCR5+ T cells were significantly higher within HCC tumor tissue than in the matched blood and peritumoral liver tissue or in healthy blood samples (Figure 2A and 2B). We subsequently found that strong tumor infiltration by IL-21-producing CD8⁺CXCR5⁺ T cells led to the accumulation of IL-21+ cells within the peritumoral stroma and that this accumulation correlated negatively with the TNM staging of the patients (Figure 2C–2E).

Figure 2. Strong infiltration of IL-21-producing CD8+CXCR5+ T cells in HCC correlates with disease stage. (A–B) Flow cytometric analysis of IL-21 production by CD8+CXCR5+ T cells (n=9). The cells were characterized using FACS with sequential gating of lymphocyte cells, CD45+ cells and then CD8+CXCR5+ cells. (A) One representative experiment is shown. (B) The data indicate the median with the interquartile range. (C–E) Immunohistochemical staining of IL-21+ cells in paraffin-embedded HCC tissue (n=96). The distribution of IL-21+ cells is shown in (C and D). Micrographs at higher magnification show the stained peritumoral liver (1), peritumoral stromal region (2), and cancer nest (3). The association of the density of tumor-infiltrating IL-21+ cells with the TNM staging of patients is shown in (E). *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 determined using Mann-Whitney U test (B and D) or student’s t test (E).

Migration and induction of CD8+CXCR5+ T cells by HCC-SN

Having established the presence of an antitumor CD8+CXCR5+ T cell subset in HCC, we next performed experiments to investigate whether environmental factors facilitated the migration and induction of CD8+CXCR5+ T cells. Culture supernatant from primary HCC tumor (HCC-SN) cells, but not from normal liver cells, showed a great ability to induce chemotaxis in CD8+CXCR5+ cells from healthy blood (Figure 3A and 3B). Moreover, the addition of HCC-SN cells to cultures of CD8+ T cells from healthy blood induced a considerable proportion of CD8+CXCR5+ T cells (Figure 3C), and upregulated IL-21 production by CD8+CXCR5+ T cells (Figure 3D). By contrast, culture with supernatant from normal liver cells failed to augment CXCR5 or IL-21 expression in CD8+ T cells.

Figure 3. Differentiation and chemotaxis of CD8+CXCR5+ T cells is induced by HCC-SN cells. (A–B) Culture supernatants from primary HCC cell cultures (HCC-SN1 and HCC-SN2 cells), but not normal liver cultures (Hem-liver-SN cells), induce chemotaxis of healthy blood CD8+CXCR5+ T cells sorted by FACS (n=3). (A) One representative experiment is shown. (B) The data indicate the mean±SD. (C–E) HCC-SN1 and HCC-SN2 cells, but not Hem-liver-SN cells, are able to induce the CD8+CXCR5+ or IL-21+ CD8+CXCR5+ phenotype in healthy blood CD8+ T cells (n=3). The results shown represent four separate experiments. The data indicate the mean±SD. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 determined using student’s t test (B, D and F).

Tumor-infiltrating CD8+CXCR5+ T cells from HCC patients were potent inducers of plasmablasts in vitro

We demonstrated that large numbers of IL-21-producing CD8+CXCR5+ T cells accumulate within HCC tumor tissue. Because IL-21 plays a critical role in regulating immunoglobulin production [15, 16], we investigated the possible effects of tumor-infiltrating CD8+CXCR5+ T cells on B cells. Confocal microscopy showed that IL-21+ T cells and CD19+ B cells were located in the same region (Figure 4A). In HCC patients, moreover, there was a correlation between the frequencies of CD19+ B cells and CD8+CXCR5+ T cells (Figure 4B). These data demonstrate that immunoglobulin produced by B cells recruited by CD8+CXCR5+ T cells may play an important role in antitumor activity.

Figure 4. Tumor-infiltrating CD8+CXCR5+ T cells from HCC patients are potent inducers of plasmablasts differentiation in vitro. (A) Representative immunofluorescence images of CD19 (green), IL-21 (red) and nuclear staining with DAPI (blue) in HCC tissue. (B) Associations between tumor-infiltrating CD8+CXCR5+ T cells and tumor-infiltrating CD19⁺ B cells (n=19). (C–D) Coculture of autologous CD19+ B cells with healthy blood CD8+CXCR5+ T cells or tumor-infiltrating CD8+CXCR5+ T cells. On day 5, the supernatants were harvested. IgG (C) and IgM levels (D) were determined using an ELISA (n=3). (E–F) Immunohistochemical staining of CD138+ B cells in paraffin-embedded HCC tissue (n=96). (G–H) Patients were divided into two groups (Low/High) based on the median of the tumor-infiltrating CD138+ B cell percentages. The DFS and OS curves between the two patient groups were compared using the log-rank test. **P<0.01.

In addition, coculture of CD8+CXCR5+ T cells with CD19+ B cells led to B cell class switching, and IgM and IgG levels in the supernatant differed between HCC patients and healthy donors (Figure 4C and 4D). These results indicate that CD8+CXCR5+ T cells in HCC patients can provide helper B cells and induce immunoglobulin-producing plasmablasts. We subsequently examined the relationship between CD138+ expression and HCC patient survival (Figure 4E–4H). HCC patients who underwent curative resection, and for whom follow-up data were available, were divided into two groups based on the median CD138+ plasma cell density (n=96, Table 3). This revealed obvious positive associations between CD138+ plasma cell density and both disease-free survival (DFS) (n=96, p=0.025, Figure 4G) and overall survival (OS) (n=96, p=0.014, Figure 4H). Univariate and multivariate analyses revealed that tumoral CD138 expression was an independent prognostic factor for DFS and OS (Table 4).

Patients and specimens

Tissue-infiltrating leukocytes were isolated from fresh tumor tissue, peritumoral liver tissue, and peripheral blood samples obtained from patients who underwent curative resection at the Third Affiliated Hospital of Sun Yat-sen University between September 2016 and June 2018 (Table 1). Tissues for immunohistochemistry (IHC) or immunofluorescence (IF) were also obtained from patients receiving surgical resection at the same hospital between September 2011 and December 2012 (Table 3). All of these patients had HBV infection. None had previously received any anticancer therapy. Patients with HIV infection or other autoimmune disease or cancer were excluded. The clinical stages of the tumors were classified according to the guidelines of the International Union Against Cancer. Written informed consent was obtained from all patients, and the protocols were approved by the Review Board of the Third Affiliated Hospital of Sun Yat-sen University.

Isolation of immune cells from peripheral blood and tissue

Peripheral blood mononuclear cells (PBMCs) were isolated using a standard Ficoll procedure. Tissue-infiltrating leukocytes were dissociated from tissue specimens and collected as described previously [27]. Briefly, specimens were cut into small pieces and digested in RPMI 1640 (Gibco) supplemented with 100 μg/ml Liberase TL (Roche), 100 μg/ml DNase I (Sigma-Aldrich) and 20% FBS (Gibco) for 30 min at 37°C. The homogenates were filtered through a 150-mm cell strainer and separated by Ficoll density gradient centrifugation (Axis-Shield). The isolated PBMCs were washed with Hanks’ balanced salt solution and resuspended in RPMI 1640 supplemented with 10% FBS.

Flow cytometric analysis of cell surface molecule expression

Cells were stained with anti-CD4-APC-CY7, anti-CD8-FITC, anti-CXCR5-APC, anti-PD-1-PE, and anti-ICOS-PECF594 (all from Biolegend), after which they were washed and marker expression was detected using flow cytometry (BD LSR2).

Flow cytometric analysis of intracellular cytokine expression

PBMCs from healthy individuals and HCC patients were cultured in a 96-well plate (1×10⁶ cells/well) in 500 μl of culture medium. The cells were stimulated with PMA (50 ng/ml, Sigma-Aldrich) and ionomycin (400 ng/ml, Sigma-Aldrich) for 1 h at 37°C. Brefeldin A (10 lg/ml, Sigma-Aldrich) was then added, and the cells were incubated for another 4 h. Thereafter, the samples were fixed, permeabilized and stained with anti-CD4-APC-CY7, anti-CD8-FITC, anti-CXCR5-APC, anti-PD-1-PE, anti-ICOS-PECF594 (all from Biolegend), and anti-IL-21-BV421 (BD Biosciences). Cytokine-producing cells were detected using flow cytometry (BD LSR2).

Preparation of culture supernatant conditioned by primary HCC tumor cells

Tumor specimens of HCC patients or healthy liver specimens from hemangioma patients (normal liver tissue far away from the hemangioma) without concurrent autoimmune disease or HBV, HCV, HIV, or syphilis infection were completely digested and then washed with medium containing polymyxin B (20 g/mL; Sigma-Aldrich) to exclude endotoxin contamination. A total of 1×10⁷ dissociated cells were resuspended in 10 mL of complete medium and cultured in 100-mm dishes. After 2 days, the supernatants were harvested, centrifuged, and stored at -80°C.

Cell migration assay

Migration of CD8+CXCR5+ T cells was assayed using a 24-well Transwell plate with a 6.5-μm pore polycarbonate membrane insert (Corning, USA) after the cells were purified using a FACS ARIA cell sorter (BD). The inserts were loaded with CD8+CXCR5+ T cells (5×10⁵ cells in 100 μl serum-free medium) and placed into the 24-well plates containing 600 μl of HCC-SN, Hem-liver-SN or serum-free medium. After incubating for 3 h at 37 °C, the migrated cells were fixed with 100% methanol for 30 min and then stained with 0.1% crystal violet (Leagene, China) for 15 min. To evaluate chemotaxis, the number of migrated cells were counted under a light microscope (Leica, Germany).

Separation of human CD19+ B or CD8+ T cells from peripheral blood

PBMCs were isolated from peripheral blood samples from healthy donors by Ficoll centrifugation (Axis-Shield). CD19+ B or CD8+ T cells (collected populations were more than 95% CD19+ or CD8+) were sorted from among the PBMCs using a MACS CD19+ B or CD8+ T cell Isolation Kit (Miltenyi Biotech).

TFH-B cell culture

PBMCs and TILs were labeled with anti-CD8-FITC and anti-CXCR5-APC (both from Biolegend), after which CD8+CXCR5+ T cells were purified using a FACS ARIA cell sorter (BD). The purity was higher than 98%. A total of 2×10⁴ T cells were cocultured with the same number of CD19+ B cells in a total volume of 200 μl of Gibco 1640. The supernatant was harvested on day 5 and stored at -20 °C. IgG and IgM levels in the supernatant were measured using an ELISA according to the manufacturer’s instructions (R&D).

Statistical analysis

Results are expressed as the mean±SD or median with interquartile range. Statistical comparisons were made as indicated in the figure legends using two-sided tests. Group data were analyzed using Student’s t test or log-rank test for normally distributed variables, and the Mann-Whitney U test was used for non-parametric comparisons. Correlations between two parameters were assessed using Pearson correlation analysis. Multivariate analysis of the prognostic factors for OS and DFS was performed using the Cox proportional hazards model and log-rank test. Cumulative survival time was assessed using the Kaplan-Meier method. Values of P<0.05 were considered significant.

Ethics approval

The biopsy specimens were obtained under protocols approved by the ethics committees of The Third Affiliated Hospital of Sun Yat-sen University and informed consent was obtained from all patients.