Classical IL-22–producing T helper cells (Th22 cells) mediate inflammatory responses independently of IFN-γ and IL-17; however, nonclassical Th22 cells have been recently identified and coexpress IFN-γ and/or IL-17 along with IL-22. Little is known about how classical and nonclassical Th22 subsets in human diseases are regulated. Here, we used samples of human blood, normal and peritumoral liver, and hepatocellular carcinoma (HCC) to delineate the phenotype, distribution, generation, and functional relevance of various Th22 subsets. Three nonclassical Th22 subsets constituted the majority of all Th22 cells in human liver and HCC tissues, although the classical Th22 subset was predominant in blood. Monocytes activated by TLR2 and TLR4 agonists served as the antigen-presenting cells (APCs) that most efficiently triggered the expansion of nonclassical Th22 subsets from memory T cells and classical Th22 subsets from naive T cells. Moreover, B7-H1–expressing monocytes skewed Th22 polarization away from IFN-γ and toward IL-17 through interaction with programmed death 1 (PD-1), an effect that can create favorable conditions for in vivo aggressive cancer growth and angiogenesis. Our results provide insight into the selective modulation of Th22 subsets and suggest that strategies to influence functional activities of inflammatory cells may benefit anticancer therapy.

IL-22 is a member of the IL-10 family that induces the production of chemokines and antimicrobial peptides by cells in tissues (1–3). Besides possible involvement in host defense against microbes, this cytokine has been linked to the pathogenesis of several inflammatory and autoimmune diseases (4–6). In animal models, it has been shown that IL-22 plays an important role in regulating autoimmune encephalitis, collagen-induced arthritis, colitis, and psoriasis (2, 7, 8). In humans, increased levels of IL-22–producing cells have been observed in epidermis from individuals with inflammatory skin disorders, and the IL-22⁺ T cell clone derived from psoriasis patients can enhance wound healing and tissue remodeling, an effect that depends exclusively on IL-22 (2, 9).

Although IL-22 was originally assumed to be a T helper type 1–associated (Th1-associated) cytokine (1, 2, 10), it has subsequently been linked to expression of IL-17 in Th17 cells (11–13). More recently, a novel IL-22–producing Th cell subset designated classical Th22 was described in humans and found to be unrelated to IFN-γ and IL-17 (14, 15). Nevertheless, it has been emphasized that a considerable proportion of IL-22⁺ Th cells from inflamed human tissues can coproduce IFN-γ and/or IL-17, and these can be called nonclassical Th22 cells (9). Thus, data on the composition and immune function of Th22 subsets in local environments are essential for understanding the potential mechanisms of these cells in immunopathogenesis.

The precise mechanisms of Th22 polarization are not yet clear, and this applies in particular to the nonclassical Th22 subsets in humans. A related issue that must be addressed is the nature of the antigen-presenting cells (APCs) that are capable of inducing Th22 responses. Here, we found that nonclassical Th22 cells represented about 80% of the entire Th22 population in tissues of normal human liver and tumor. Monocytes activated by LPS were the APCs that most efficiently triggered the expansion of nonclassical Th22 cells from memory T cells and classical Th22 cells from naive T cells. Selective repression of IFN-γ–producing Th22 subsets occurred in hepatocellular carcinoma (HCC) tissue, and blockade of B7-H1 on tumor monocytes directed Th22 cell differentiation away from IL-17 and toward IFN-γ. Moreover, we found evidence that an increase in IL-17–expressing Th22 cells in HCC promoted aggressive cancer growth and angiogenesis.

Th22 subset composition in human blood and tissue. We used FACS to analyze the composition of IL-22–producing Th subsets in 12 normal blood samples, 4 normal liver samples (tissue distal to a liver hemangioma), and 26 HCC tumor samples paired with peritumoral liver and blood samples (Supplemental Table 1; supplemental material available online with this article; doi:10.1172/JCI74381DS1). We classified 8 Th22 subsets based on their abilities to secrete IFN-γ, IL-4, and/or IL-17 (Figure 1A). In all the samples we analyzed, the four IL-4–producing Th22 subsets represented only a very small proportion (0.05% to ~2.7%), whereas 4 other subsets, including the classical Th22 (IFN-γ^–IL-17^–), Th22/Th1 (IFN-γ⁺IL-17^–), Th22/Th17 (IFN-γ^–IL-17⁺), and Th22/Th17/Th1 (IFN-γ⁺IL-17⁺) cells, constituted 96.5% ± 2.4% (Figure 1B). Classical Th22 cells constituted the predominant subset (~50% of total Th22 cells) in blood from both healthy donors (n = 12) and HCC patients (n = 26) (Figure 1, B and C). We found increased proportions of IFN-γ–associated Th22/Th1 and Th22/Th17/Th1 cells (44.2 ± 8.3% and 21.8 ± 4.7%, respectively) in both normal and peritumoral liver tissues. However, compared with the peritumoral liver tissue, the tumor tissue contained a significant proportion of Th22/Th17 cells (40.8 ± 6.9%; n = 26, P < 0.001 compared with peritumoral liver Th22/Th17) accompanied by an attenuated amount of Th22/Th1 cells (22.8 ± 4.8%) (Figure 1, B and C). Also, the HCC patients who were infected with HBV or HCV and those without such infection exhibited a similar Th22 subset composition in blood and peritumoral liver and tumor tissues (Figure 1, B and C).

Figure 1

Th22 subset composition in blood, liver, and HCC. (A and B) FACS analysis of IL-22, IFN-γ, IL-4, and IL-17 expression in T cells from normal human blood (A) and in Th22 cells isolated from blood, normal/peritumoral liver tissue, and HCC tissue (B). (C) FACS analysis of IFN-γ and IL-17 expression in IL-4^– Th22 cells from blood, normal/peritumoral liver tissue, and HCC tissue. Data shown in A–C are from 30 independent experiments. Numbers adjacent to the outlined areas (A) or in quadrants (A and C) indicate the percentage of cells in each quadrant. Cells used in the analyses came from samples collected as follows: normal blood from 12 healthy individuals (A–C); normal liver tissue from 4 liver hemangioma patients (B and C); paired blood, peritumoral liver tissue, and tumor tissue from 20 patients with HBV-related HCC, 3 patients with HCV-related HCC, and 3 patients with HCC but no HBV or HCV infection (B and C).

Phenotypic characteristics and transcription factor profiles of human Th22 subsets. Inasmuch as the classical Th22, Th22/Th1, Th22/Th17, and Th22/Th17/Th1 subsets constituted nearly the entire Th22 cell population, we next used FACS (Supplemental Figure 1) to investigate the phenotypic and transcriptional features of these subsets isolated from paired blood, peritumoral liver tissue, and HCC tumor tissue (Supplemental Table 1).

All the Th22 subsets, together with subsets Th1, Th17, and Th1/Th17, had a CD45RO⁺CD62L^–CCR7^– effector memory phenotype exhibiting substantial expression of CCR4 and CCR6, but low or negligible expression of other chemokine receptors (Figure 2). A remarkably large portion of all the Th22 subsets expressed the proinflammatory cytokines IL-2 and TNF-α, but not the antiinflammatory IL-4, IL-9, and IL-10, which illustrates the proinflammatory properties of IL-22⁺ cells (Figure 2). Analysis of immune modulatory molecules on these subsets isolated from HCC tumor tissue revealed that both Th22/Th17/Th1 and Th22/Th17 cell subsets expressed higher levels of the activation markers HLA-DR, CD25, and CD69, as well as the regulatory molecules CTLA-4 and GITR, than did their subset counterparts Th1/Th17 and Th17, respectively (Figure 2, right panel). These results suggest that IL-22–related Th17 cells represent highly activated T cell subsets that express more inhibitory molecules to attenuate T cell responses in tumors. In contrast, the classical Th22, Th1, and Th22/Th1 subsets displayed only weak expression of the above-mentioned molecules (Figure 2).

Figure 2

Phenotypic characteristics and transcription factor profiles of Th22 subsets. FACS analysis was performed to determine the phenotypic characteristics of Th0 (IFN-γ^–IL-17^–IL-22^–), Th1 (IFN-γ⁺IL-17^–IL-22^–), Th17 (IFN-γ^–IL-17⁺IL-22^–), Th17/Th1 (IFN-γ⁺IL-17⁺IL-22^–), Th22 (IFN-γ^–IL-17^–IL-22⁺), Th22/Th1 (IFN-γ⁺IL-17^–IL-22⁺), Th22/Th17 (IFN-γ^–IL-17⁺IL-22⁺), and Th22/Th17/Th1 (IFN-γ⁺IL-17⁺IL-22⁺) subsets isolated from blood, peritumoral liver tissue, and tumor tissue of HCC patients. Data for each of the indicated markers represent at least 4 patients. Data shown are from 30 independent experiments.

We subsequently analyzed the transcription factors involved in the development of nonclassical Th22 subsets. The Th22/Th1 and Th22/Th17/Th1 subsets expressed T-bet at a high level, comparable to that exhibited by the Th1 and Th17/Th1 subsets; in parallel, RORγt was widely expressed by Th17 subsets. Consistently, the classical Th22 subsets expressed neither T-bet nor RORγt in any parts. We found that none of the analyzed Th cell subsets expressed the Treg-specific transcription factor FOXP3 (Figure 2). These results indicate that induction of IL-22 in human Th cells may be regulated by an unknown transcription factor different from that regulating Th1, Th17, and Tregs.

Activated APCs induce Th22 subset expansion. APCs are essential for initiating and maintaining the responses of Th cells (16–18). Having established the Th22 subset composition in human blood, liver, and HCC, we performed experiments to elucidate the nature of the APCs and innate stimuli that trigger the expansion and differentiation of Th22 subsets. We used freshly isolated monocytes from normal blood as well as cultured monocyte–derived macrophages (Mϕ) (19) and DCs (20). APC activation was induced only by Pam3CysSK4 (a TLR2 agonist) and LPS (a TLR4 agonist), that is, not by CpG-DNA ODN (CpG, a TLR9 agonist) or poly (I:C) (PIC, a TLR3 agonist), as assessed by determining the upregulation of CD80 and CD86 (data not shown). The cytokine production in response to Pam3CysSK4 and LPS depended on the type of APCs present. In particular, monocytes produced larger amounts of TNF-α, IL-1β, IL-6, IL-23, and TGF-β, and, notably, this was not associated with production of IL-12p70 (Figure 3A).

Figure 3

Activated APCs induce Th22 subset expansion. (A) Monocytes (MOs), Mϕ, and DCs were left untreated or stimulated with Pam3CysSK4, LPS, poly (I:C), or CpG oligodeoxynucleotides (ODNs) for 18 hours. Cytokine production was determined by ELISA. (B–D) Monocytes, Mϕ, or DCs were left untreated or stimulated with LPS for 5 hours and then cultured for 8 days with autologous T cells, naive T cells, or memory T cells as described in Methods. Proliferation (Ki67) and expression of IFN-γ and IL-17 in Th22 cells were detected by FACS. Numbers in quadrants indicate the percentage of cells in each quadrant. (E) Blockade of TNF-α, IL-1β, IL-6, IL-23, and/or TGF-β changed the subset composition of Th22 cells expanded by LPS-stimulated monocytes. (F and G) T cells were left untreated or cultured with conditioned media from LPS- or Pam3CysSK4-stimulated monocytes, Mϕ, or DCs for 7 days (F) or for the indicated times (G), as described in Methods. Activation of STAT proteins was detected by immunoblotting. IL-4–treated T cells served as a positive control. (H) Inhibiting STAT1 and STAT3 activation changed the subset composition of Th22 cells expanded by LPS-stimulated monocytes. Th22 subset composition was determined by FACS. Th22 subset composition is shown in C–E and H: classical Th22, white bar; Th22/Th1, dark gray bar; Th22/Th17, light gray bar; Th22/Th17/Th1, black bar. Data represent the mean ± SEM (A, C–E, and H). Results in all panels represent 5 separate experiments (n = 5–7). *P < 0.05 and **P < 0.01, compared with untreated cells (A) or the indicated groups.

In subsequent experiments, autologous T cells were incubated with monocytes, monocyte-derived Mϕ or DCs, or such cells stimulated with LPS, and this was done in the absence or presence of CD3–cross-linking Abs. We measured T cell proliferation and production of IL-17, IFN-γ, and IL-22 on day 8 after incubation with IL-2. In general, all APCs induced T cell proliferation, although unstimulated APCs were much less potent than LPS-stimulated APCs in this context (Supplemental Figure 2A). LPS-activated monocytes were the APCs that most efficiently triggered expansion of Th22 cells (Figure 3B). Notably, most of these monocyte-expanded Th22 subsets were nonclassical (i.e., Th22/Th1, Th22/Th17, and Th22/Th17/Th1) (Figure 3C). In support of the view that IL-12p70 maintains Th1 responses, activated DCs mainly expanded the Th22/Th1 cell subset. We further purified naive and memory T cells and then cultured those cells with LPS-stimulated monocytes. The activated monocytes effectively promoted the expansion of Th22 cells from memory CD4⁺ T cells, and most of the expanded cells were positive for IFN-γ and/or IL-17 (Figure 3D). By comparison, monocyte-primed naive T cells gave rise to only a moderate amount of Th22 cells, but over 80% of them showed the classical Th22 phenotype, which suggests that nonclassical Th22 cells are expanded from memory T cells and that classical Th22 cells are polarized from naive T cells.

To probe the mechanisms involved in the induction of Th22 cells by APCs, particularly activated monocytes, we used neutralizing Abs that effectively abolished the individual roles of the above-mentioned cytokines in our monocyte-expanded Th22 cell systems. Expression of IL-22 in CD4⁺ T cells was inhibited 30%–60% by blocking IL-1β, IL-6, IL-23, or TGF-β, but was not reduced by blocking TNF-α alone (Supplemental Figure 2B). Precisely, blocking IL-1β mainly decreased the proportions of the IFN-γ–expressing Th22 cells, but an anti–TGF-β Ab primarily reduced the proportions of IL-17–producing Th22 cells (Figure 3E). In agreement with this, recombinant IL-1β effectively increased the prevalence of IFN-γ⁺ Th22, and TGF-β selectively induced expansion of IL-17⁺ Th22 cells (Supplemental Figure 2C). In addition, STAT proteins are well known for their essential roles in specifying Th cell differentiation. We found that conditioned media from APCs stimulated by LPS or Pam3CysSK4 could elicit marked activation of both STAT1 and STAT3, but not STAT5 or STAT6, in T cells in a time-dependent manner (Figure 3, F and G). Abolishing STAT1 activation by fludarabine effectively suppressed the LPS-activated mononcyte-mediated IFN-γ⁺ Th22 expansion; in parallel, AG490, a specific inhibitor of STAT3, could effectively impair the generation of IL-17–related Th22 cells (Figure 3H).

Monocyte B7-H1 regulates the balance between IFN-γ⁺ and IL-17⁺ Th22 cells. After finding that Th22/Th17 cells selectively accumulated in HCC tissue, we evaluated the influence of HCC-infiltrating monocytes on the development of Th22 subsets. Compared with corresponding cells in blood, monocytes isolated from both peritumoral liver and tumor tissues expressed at least 10-fold higher levels of TNFA, IL1B, IL6, and TGFB, but not IL12A (Figure 4A). Furthermore, analysis of immune modulatory molecules on monocytes revealed that only such cells isolated from tumor tissue showed markedly elevated expression of coinhibitory B7-H1, although monocytes from both peritumoral liver and tumor tissues exhibited a CD80^hiCD86^hiHLA-DR^hiB7-H4^– phenotype (Figure 4B), suggesting that B7-H1 regulates Th22/Th17 cells in tumors. To address this, we purified monocytes from peritumoral liver and paired tumor tissues and then cultured those cells together with autologous peripheral T cells. The monocytes from both sources markedly expanded Th22 cells, albeit with a subset bias: the peritumoral monocytes promoted expansion of IFN-γ–producing Th22 cells more strongly, whereas tumor monocytes associated selectively with the IL-17–expressing Th22 cell subset (Figure 4C). In support of our hypothesis, blocking B7-H1 by preincubating tumor monocytes with the mAb MIH1 led to a marked increase in Th22 cells that were related to IFN-γ (Figure 4C). Notably, such treatment also elicited robust production of IL-22 and IFN-γ by T cells (Supplemental Figure 3).

Figure 4

Monocyte B7-H1 regulates the balance between IFN-γ– and IL-17–related Th22 cells. (A) Real-time PCR determination of relative fold changes in the indicated cytokine mRNAs in monocytes from paired samples of blood and peritumoral liver and tumor tissues from 11 HCC patients. (B) FACS analysis of expression of the indicated markers on monocytes from paired samples of blood and peritumoral liver and tumor tissues from 4 HCC patients. (C and D) Purified T cells were cultured for 8 days with autologous peritumoral liver or tumor monocytes (C) or LPS-stimulated peripheral monocytes (D) in the presence of an anti–B7-H1 or a control Ab, as described in Methods. Expression of IL-22 in Th cells and expression of IFN-γ and IL-17 in Th22 are shown. Numbers in quadrants indicate the percentage of cells in each quadrant. (E and F) T cells were left untreated or were cultured with mock, B7-H1, or B7-DC transfectant in the presence of IL-1β, IL-6, IL-23, and TGF-β for 9 days as described in Methods. In parallel, B7-H1–expressing transfectants were preincubated for 1 hour with an anti–B7-H1 or a control Ab. STAT activation on day 1 (E) and Th22 subset composition on day 8 (F) in T cells were determined by immunoblotting and FACS, respectively. Th22 subset composition is shown in F: classical Th22, white bar; Th22/Th1, dark gray bar; Th22/Th17, light gray bar; Th22/Th17/Th1, black bar. Data represent the mean ± SEM (B and F). Results represent 4 separate experiments (n = 4–6; C–F). *P < 0.05; **P < 0.01.

Considering that B7-H1 was also expressed by the LPS-activated monocytes (Supplemental Figure 4A), we subsequently analyzed the involvement of this protein in LPS-stimulated monocyte-mediated Th22 subset bias. We found that blocking B7-H1 markedly and consistently increased the proportion of IFN-γ–producing Th22 cells expanded by exposure to LPS-activated monocytes (Figure 4D). Furthermore, we transduced HEK293T cells with pBABE-Puro retroviral vector encoding human B7-H1 or B7-DC (Supplemental Figure 4B) and then cultured those cells with T cells in a cytokine-expanded Th22 system. B7-H1–expressing HEK293T cells could effectively suppress the cytokine-mediated early STAT1 activation and IFN-γ production in T cells and later Th1-related Th22 polarization; the effects could be reversed by the B7-H1–blocking mAb MIH1 (Figure 4, E and F, and Supplemental Figure 5). These results suggest that the B7-H1/PD-1 inhibitory pathway negatively regulates STAT1 by limiting IFN-γ production, which in turn suppresses Th1-related subset development. We found that the B7-DC transfectant did not affect such processes (Figure 4, E and F and Supplemental Figure 5).

Repression of IFN-γ–producing Th22 subsets in human HCC. IFN-γ–producing Th22 cells were repressed by the monocyte B7-H1 ex vivo, which prompted us to analyze the association between the proportions of Th22 and PD-1⁺ Th cells in HCC. Upregulation of all Th22 subsets was much more pronounced in tumor and peritumoral liver tissues than in blood from HCC patients (n = 26) (Figure 5A). Due to declines in IFN-γ–related Th22 cell subsets, the prevalence of all Th22 cells was lower in the tumor samples than in the paired peritumoral liver samples, although we found significantly more IL-17–producing Th22 cells in tumor tissue than in peritumoral liver tissue (Figure 5B) (2.1 ± 0.3% and 0.7 ± 0.1% in CD4⁺ T cells, respectively; n = 26 for both). Contrary to the infiltration pattern of the entire Th22 cell population, the proportion of PD-1⁺ Th cells was markedly increased in tumor tissue (60 ± 13%, n = 26; P < 0.001 compared with proportions in blood and peritumoral liver tissue; Figure 5C). As expected, in the tissues, the proportions of PD-1⁺ Th cells were inversely correlated with both the size of the entire Th22 cell population and the proportions of the IFN-γ–producing Th22 subsets (Figure 5, D and E), whereas the levels of PD-1⁺ Th cells differed from the proportions of IL-17–expressing Th22 cells (Figure 5F). We observed analogous changes in the proportions of Th22 and PD-1⁺ Th cells in tissues that were or were not infected with hepatitis B virus (HBV) or hepatitis C virus (HCV) (Figure 5, A–C). Furthermore, both confocal microscopy and FACS confirmed that IL-22 protein was expressed mainly in PD-1^– Th cells (n = 8) and only weakly in PD-1⁺ cells (Figure 5, G and H).

Figure 5

Repression of IFN-γ–producing Th22 subsets in human HCC. (A) FACS analysis of IL-22 expression in fresh Th cells isolated from paired blood and peritumoral liver and tumor tissues. (B) FACS analysis of Th22 subset composition in Th cells isolated from paired blood and peritumoral liver and tumor tissues: classical Th22, white bar; Th22/Th1, dark gray bar; Th22/Th17, light gray bar; Th22/Th17/Th1, black bar. Data represent the mean ± SEM. (C) FACS analysis of PD-1 expression on Th cells isolated from paired blood and peritumoral liver and tumor tissues. (D–F) Correlations between the levels of PD-1⁺ Th cells and the following cells isolated from samples of peritumoral liver and tumor tissues from HCC patients (n = 26): total Th22 cells (D), IFN-γ⁺ Th22 cells (E), and IL-17⁺ Th22 cells (F). (G) Expression of IL-22 and PD-1 in Th cells isolated from paired samples of blood and peritumoral liver and tumor tissues from 8 HCC patients. Numbers in quadrants indicate the percentage of cells in each quadrant. (H) Expression of IL-22, PD-1, and CD3 in HCC tumor tissue. One of 5 representative micrographs is shown. Scale bars: 40 μm. Cells analyzed in A–C were isolated from samples collected from 20 patients with HBV-related HCC, 3 patients with HCV-related HCC, and 3 patients with HCC but no HBV or HCV infection. *P < 0.05; **P < 0.01; ***P < 0.001.

Protumorigenic role of Th22 and Th17 in HCC. Previous studies have shown that IL-22 exhibits protumorigenic activity in several types of human cancers (21–26). To evaluate the potential role of IL-22 and its related subsets in HCC pathogenesis, we divided the 26 HCC patients described above into 2 groups according to their complete clinical and pathological characteristics. Figure 6A shows that the proportions of Th22 cells were associated with the patients’ tumor node metastasis (TNM) stages and intrahepatic metastasis in the tumor samples but not in the peritumoral liver tissue, which implies an unfavorable effect on patient survival. We investigated this possibility by immunohistochemical staining of IL-22 in paraffin-embedded tissues from 197 HCC patients (Supplemental Table 2). Kaplan-Meier analysis showed inverse associations between IL-22⁺ cell density in the tumor area and both overall survival (OS) (P < 0.001) and disease-free survival (DFS) (P < 0.0001) (Figure 6, B and C). In multivariate analysis, the number of IL-22⁺ cells in the tumor region was an independent prognostic factor for both OS and DFS (Supplemental Table 3).

Figure 6

Protumorigenic roles of IL-22 and IL-17 in HCC. (A) Correlations between proportions of Th22 cells in tumor tissue from 26 patients and TNM or intrahepatic metastasis in patients. (B and C) Paraffin-embedded hepatoma samples were stained with an anti–IL-22 Ab. Scale bar: 50 μm. Based on the median value of IL-22⁺ cell density, patients were divided into low-density (n = 99) and high-density (n = 98) groups; cumulative OS and DFS times were calculated by the Kaplan-Meier method and analyzed by a log-rank test (C). (D and H) IL-22 plus IL-17 promoted the survival, proliferation, and proangiogenic activities of hepatoma cells. HepG2 cells undergoing serum starvation were stimulated with IL-22 alone or combined with IL-17 or IFN-γ (50 ng/ml for all) for 10 hours. Expression of the indicated proteins and genes (given as the mean) was determined by immunoblotting and real-time PCR, respectively. Results are representative of 4 separate experiments (n = 4–6). (E) Injection of T cells polarized by tumor monocytes significantly enhanced HepG2 hepatoma growth in mice; the effects were attenuated by additional injection of neutralizing Abs against IL-22 and IL-17. Data represent the mean ± SEM of 5 independent experiments (n = 6 for each). *P < 0.05, compared with the Medium group; ^#P < 0.05, compared with the T cell group (without B7-H1 mAb). (F and G) IL-22 plus IL-17 promoted the growth and angiogenesis of hepatoma in mice. Data represent the mean ± SEM (n = 5 for each group). *P < 0.05. Paraffin-embedded hepatoma samples were stained with an anti-CD34 Ab (G). Scale bar: 100 μm.

Th22 cells in peritumoral liver tissue produced higher levels of IFN-γ, whereas those in tumor tissue were selectively associated with IL-17 (Figure 5B). Therefore, we investigated the role of IL-22 alone or in combination with IFN-γ or IL-17 in regulating the biological behavior of hepatoma. We exposed hepatoma cells to IL-22 for 10 hours under serum starvation, which resulted in a markedly altered phenotype exhibiting active phosphorylation of the STAT3 oncogene and pronounced upregulation of survival-associated Mcl-1 and cell cycle–related CDK4, C-Myc, and cyclin D1 (Figure 6D and Supplemental Figure 6). However, adding IFN-γ completely abolished such proproliferative and prosurvival roles of IL-22. Notably, IFN-γ also increased expression of proapoptotic BAX and BID in IL-22–treated hepatoma cells, whereas the addition of IL-17 had only a marginal effect on hepatoma cell survival and proliferation (Figure 6D and Supplemental Figure 6). Consistent with these in vitro findings, we observed that injection of T cells polarized by tumor monocytes in the presence of anti–B7-H1 mAb (25.1 ± 3.4% and 21.3 ± 2.9% of them were IFN-γ⁺ Th22 cells and IL-17⁺ Th22 cells, respectively; n = 6) only marginally affected the growth of human hepatoma. In contrast, the injection of T cells polarized by tumor monocytes without anti–B7-H1 mAb (10.5 ± 2.7% and 23.2 ± 3.2% of them were IFN-γ⁺ Th22 cells and IL-17⁺ Th22 cells, respectively) significantly enhanced the tumor growth (Figure 6E). Supporting our hypothesis that IL-22 and IL-17 signals contribute to disease progression, administration of anti–IL-22– and anti–IL-17–neutralizing Abs effectively suppressed the protumorigenic effect of T cells polarized in the absence of B7-H1 mAb (Figure 6E). Similar results were obtained in another in vivo system, showing that the protumorigenic effect of IL-22 was completely reversed by adding IFN-γ, but was accelerated by IL-17 supplement (Figure 6F). Data from immunohistochemical staining revealed that IL-17 additionally increased IL-22–mediated angiogenesis (Figure 6G).

We further investigated the potential role of IL-22 in angiogenesis using real-time PCR analysis. Compared with untreated hepatoma cells, those exposed to IL-22 exhibited distinct upregulation of a set of proangiogenic genes including CCL2, CCL5, CCL20, CXCL1, CXCL2, CXCL3, CXCL5, CXCL8, CXCL9, CXCL10, CXCL11, and MMP2, and this process was enhanced by the addition of IL-17 but not IFN-γ (Figure 6H and Supplemental Figure 7). In support of this, we also found a significant correlation between microvessel density (MVD) and the level of Th22 cells in human HCC tissue (r = 0.728, P < 0.001), whereas no such association was observed in peritumoral liver tissue (r = 0.293) (Supplemental Figure 8). Together, these results demonstrate that, in HCC, IL-22 plays a protumorigenic role that is potentiated by IL-17 in tumor tissue but counteracted by IFN-γ in peritumoral liver tissue.

Despite recent success in demonstrating the existence of Th22 cells, little is known about the mechanisms underlying the regulation of different Th22 subsets in humans (8). The current study delineated the phenotype, distribution, generation, and functional relevance of various Th22 subsets in blood, liver, and HCC tissue.

Studies have identified a subset of epithelial Th effector cells that produce IL-22 (and are thus designated Th22) (9, 21), and it has been suggested that these cells participate in adaptive immunity and reorganization of epithelia via IL-22–mediated induction of antimicrobial peptides and genes involved in keratinocyte differentiation (27). In our investigation, the substantial proportion of Th22 cells in peritumoral liver and tumor tissues from HCC patients also secreted IFN-γ and/or IL-17, and reports have described such plasticity in other T cell phenotypes as well: for example, Th17 cells expressing IFN-γ and Tregs expressing IL-17 (28–31). Interestingly, we found that IFN-γ–expressing Th22 subsets were enriched predominantly in peritumoral liver tissue, and IL-17–producing Th22 subsets were selectively accumulated in HCC tissue. Further investigations revealed that the proportion of IL-17–producing Th22 cells in tumors correlated with aggressive angiogenesis, and the IL-17 cytokine could effectively enhance the proangiogenic effect of IL-22 both in vitro and in vivo. These findings imply that Th22 cells in tumor tissue, particularly those producing IL-17, do not represent the host defense against the malignancy, but instead reflect effects that are rerouted in a tumor-promoting direction by fostering angiogenesis. This notion is supported by our observations that the occurrence of Th22 cells in HCC tumor tissue, but not in peritumoral liver tissue, was positively associated with intrahepatic metastasis and TNM stages and also predicted shorter survival of the patients.

It is fairly well established that effector T cells are extremely heterogeneous in terms of cytokine production. We found that, along with classical Th22 subsets, nonclassical Th22 subsets constituted a major part of the IL-22–producing Th cells in tumor and peritumoral tissues from human liver, and most of these cells were expanded from memory T cells and spontaneously acquired the ability to produce IL-17 and IFN-γ. This finding agrees with studies suggesting that even though polarized T cells retain memory of the “imprinted” cytokine, they undergo further differentiation in response to polarizing cues (18, 32, 33). Alternatively, it is possible that classical Th22 cells are “programmed” to differentiate into Th17 cells, as has been shown for cells precommitted to Th1 differentiation (34). The commitment and flexibility of effector T cell populations are probably controlled by the expression and balance of lineage-specifying transcription factors (35). In support of that assumption, our results showed that the nonclassical, but not the classical, Th22 cells in blood and peritumoral liver and tumor tissues from HCC patients expressed T-bet and/or RORγt at a level comparable to that exhibited by Th1 and/or Th17 cells. It is plausible that under certain conditions of local stimulation, Th22 cells in tissues can differentiate into polyfunctional T cells that trigger additional responses. Thus, whereas IL-22 and IL-17 may induce cancer growth and angiogenesis, it is possible that IFN-γ suppresses the proliferation and survival of tumor cells.

An unexpected finding is that Th22 expansion and polarization were efficiently elicited by activated monocytes and, to a lesser extent, also by monocyte-derived Mϕ or DCs. This difference in T cell–polarizing capacity correlates with the pattern of cytokines produced by the various stimulated APCs. Indeed, when activated by LPS and Pam3CysSK4, monocytes produced larger amounts of TNF-α, IL-1β, IL-23, IL-6, and TGF-β. One of those molecules, IL-1β, proved to play a dominant role in the induction of IFN-γ–related Th22 subsets, whereas TGF-β mainly stimulated the IL-17–expressing Th22 subsets. These findings agree with the widespread views that IL-1β efficiently activates Th1 responses, whereas TGF-β is involved in Th17 cell development (18, 36, 37). Our results also suggest that nonclassical Th22 subsets are expanded from memory CD4⁺ T cells, which concurred with our earlier study showing that activated monocytes from HCC tissue can intensively expand Th17/Th1 cells only from the memory effector T cells (18). The Th22 cells expanded from naive T cells were not well marked but exhibited a classical phenotype, thus revealing that extra induction processes are required for such cell development. Further research is needed to identify the APCs and maturation stimuli other than microbial products that can induce Th22 responses in vivo under other physiological and pathological conditions.

B7-H1 is a cell-surface glycoprotein that belongs to the B7 family of cosignaling molecules and has a profound regulatory effect on T cell responses (38–40). The present study reveals what we believe to be a novel role of B7-H1 in regulating Th22 subset polarization in human tumors, and this conclusion is supported by the results of 3 sets of experiments. First, although both peritumoral liver and paired tumor monocytes had a marked impact on cytokine production, only monocytes isolated from tumor tissue showed pronounced B7-H1 expression, and most Th22 cells isolated from corresponding tissue did not produce IFN-γ. Second, the monocytes isolated from peritumoral liver tissue were more potent in promoting the expansion of IFN-γ–producing Th22 cells, whereas paired tumor monocytes were selectively linked to the IL-17–expressing Th22 subsets; blocking B7-H1 by preincubating tumor monocytes with the mAb MIH1 directed Th22 differentiation toward IFN-γ. Third, expression of the B7-H1 receptor PD-1 was significantly increased on tumor-infiltrating Th cells; briefly, in HCC tissue, a striking inverse association was found between the levels of PD-1⁺ Th cells and IFN-γ–producing, but not Th17-expressing, Th22 cells. Therefore, if B7-H1⁺–activated monocytes can defeat the formation of IFN-γ–related Th22 subsets, it may reflect a novel immunoediting mechanism in human tumors.

Recent clinical trials with anti–B7-H1/PD-1 mAbs have demonstrated clinical activity in patients with advanced solid tumors (41, 42). Interestingly, B7-H1 expression in tumor specimens was found to be associated with the objective response in patients treated with anti–PD-1 mAbs (43). It should be noted that the upregulation of B7-H1 in tumor tissues can also be viewed as a reflection of endogenous inflammatory immune responses, but not simply as the dominance of immune suppression (44). In a previous study, we observed that the activated monocytes in HCC tissue express B7-H1 to suppress tumor-specific T cell function (38, 39). The present study showed that B7-H1–expressing monocytes could skew Th22 polarization away from IFN-γ and toward IL-17, which may promote tumor progression by fostering angiogenesis. Thus, in addition to reversing immune tolerance, blockade of the B7-H1/PD-1 axis may also regulate tumor angiogenesis. A better understanding of the mechanisms of B7-H1/PD-1 interaction in tumor environments would be helpful for developing rational treatment combinations with mAbs blocking PD-1 or B7-H1 and other potentially synergistic therapeutic agents, including those targeting angiogenesis.

View Supplemental data

We thank Patricia Ödman for editing the manuscript. The study was supported by project grants from the National Natural Science Foundation of China (81422036, 81171982, and 81230073); the Ministry of Health of China (2012ZX10002-011); the Guangdong Natural Science Funds for Distinguished Young Scholar (S2013050014639); the Program for New Century Excellent Talents in University of the Ministry of Education of China (NCET-11-0534); and the Foundation for the Author of National Excellent Doctoral Dissertation of PR China (201230).

Address correspondence to: Limin Zheng or Dong-Ming Kuang, School of Life Sciences, Sun Yat-sen University, 135 Xin Gang Xi Lu, Guangzhou 510275, China. Phone: 86.20.84112163; E-mail: zhenglm@mail.sysu.edu.cn (L. Zheng). Phone: 86.20.84112157; E-mail: kdming@mail.sysu.edu.cn (D.M. Kuang).

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2014;124(10):4657–4667. doi:10.1172/JCI74381.

1. Wolk K, Sabat R. Interleukin-22: a novel T- and NK-cell derived cytokine that regulates the biology of tissue cells. Cytokine Growth Factor Rev. 2006;17(5):367–380.
   View this article via: PubMed CrossRef Google Scholar
2. Sonnenberg GF, Fouser LA, Artis D. Border patrol: regulation of immunity, inflammation and tissue homeostasis at barrier surfaces by IL-22. Nat Immunol. 2011;12(5):383–390.
   View this article via: PubMed CrossRef Google Scholar
3. Rubino SJ, Geddes K, Girardin SE. Innate IL-17 and IL-22 responses to enteric bacterial pathogens. Trends Immunol. 2012;33(3):112–118.
   View this article via: PubMed CrossRef Google Scholar
4. Kamanaka M, et al. Memory/effector (CD45RB(lo)) CD4 T cells are controlled directly by IL-10 and cause IL-22-dependent intestinal pathology. J Exp Med. 2011;208(5):1027–1040.
   View this article via: PubMed CrossRef Google Scholar
5. Qin WZ, et al. Expressions of IL-22 in circulating CD4+/CD8+ T cells and their correlation with disease activity in SLE patients. Clin Exp Med. 2011;11(4):245–250.
   View this article via: PubMed CrossRef Google Scholar
6. Feng D, et al. Interleukin-22 promotes proliferation of liver stem/progenitor cells in mice and patients with chronic hepatitis B virus infection. Gastroenterology. 2012;143(1):188–198.
   View this article via: PubMed CrossRef Google Scholar
7. Mitra A, Raychaudhuri SK, Raychaudhuri SP. Functional role of IL-22 in psoriatic arthritis. Arthritis Res Ther. 2012;14(2):R65.
   View this article via: PubMed CrossRef Google Scholar
8. Ma HL, et al. IL-22 is required for Th17 cell-mediated pathology in a mouse model of psoriasis-like skin inflammation. J Clin Invest. 2008;118(2):597–607.
   View this article via: JCI PubMed CrossRef Google Scholar
9. Eyerich S, et al. Th22 cells represent a distinct human T cell subset involved in epidermal immunity and remodeling. J Clin Invest. 2009;119(12):3573–3585.
   View this article via: JCI PubMed CrossRef Google Scholar
10. Gurney AL. IL-22, a Th1 cytokine that targets the pancreas and select other peripheral tissues. Int Immunopharmacol. 2004;4(5):669–677.
    View this article via: PubMed CrossRef Google Scholar
11. Zheng Y, et al. Interleukin-22, a TH17 cytokine, mediates IL-23-induced dermal inflammation and acanthosis. Nature. 2007;445(7128):648–651.
    View this article via: PubMed CrossRef Google Scholar
12. Ma HL, et al. IL-22 is required for Th17 cell-mediated pathology in a mouse model of psoriasis-like skin inflammation. J Clin Invest. 2008;118(2):597–607.
    View this article via: JCI PubMed CrossRef Google Scholar
13. Liang SC, et al. Interleukin (IL)-22 and IL-17 are coexpressed by Th17 cells and cooperatively enhance expression of antimicrobial peptides. J Exp Med. 2006;203(10):2271–2279.
    View this article via: PubMed CrossRef Google Scholar
14. Duhen T, Geiger R, Jarrossay D, Lanzavecchia A, Sallusto F. Production of interleukin 22 but not interleukin 17 by a subset of human skin-homing memory T cells. Nat Immunol. 2009;10(8):857–863.
    View this article via: PubMed CrossRef Google Scholar
15. Trifari S, Kaplan CD, Tran EH, Crellin NK, Spits H. Identification of a human helper T cell population that has abundant production of interleukin 22 and is distinct from T(H)-17, T(H)1, and T(H)2 cells. Nat Immunol. 2009;10(8):864871.
    View this article via: PubMed Google Scholar
16. Reis e Sousa C. Activation of dendritic cells: translating innate into adaptive immunity. Curr Opin Immunol. 2004;16(1):21–25.
    View this article via: PubMed CrossRef Google Scholar
17. Zhou J, Ding T, Pan W, Zhu LY, Li L, Zheng L. Increased intratumoral regulatory T cells are related to intratumoral macrophages and poor prognosis in hepatocellular carcinoma patients. Int J Cancer. 2009;125(7):1640–1648.
    View this article via: PubMed CrossRef Google Scholar
18. Kuang DM, Peng C, Zhao Q, Wu Y, Chen MS, Zheng L. Activated monocytes in peritumoral stroma of hepatocellular carcinoma promote expansion of memory T helper 17 cells. Hepatology. 2010;51(1):154–164.
    View this article via: PubMed CrossRef Google Scholar
19. Kuang DM, Wu Y, Chen N, Cheng J, Zhuang SM, Zheng L. Tumor-derived hyaluronan induces formation of immunosuppressive macrophages through transient early activation of monocytes. Blood. 2007;110(2):587–595.
    View this article via: PubMed CrossRef Google Scholar
20. Kuang DM, Zhao Q, Xu J, Yun JP, Wu C, Zheng L. Tumor-educated tolerogenic dendritic cells induce CD3epsilon down-regulation and apoptosis of T cells through oxygen-dependent pathways. J Immunol. 2008;181(5):3089–3098.
    View this article via: PubMed CrossRef Google Scholar
21. Liu T, et al. Increased circulating Th22 and Th17 cells are associated with tumor progression and patient survival in human gastric cancer. J Clin Immunol. 2012;32(6):1332–1339.
    View this article via: PubMed CrossRef Google Scholar
22. Kirchberger S, et al. Innate lymphoid cells sustain colon cancer through production of interleukin-22 in a mouse model. J Exp Med. 2013;210(5):917–931.
    View this article via: PubMed CrossRef Google Scholar
23. Kim K, Kim G, Kim JY, Yun HJ, Lim SC, Choi HS. Interleukin-22 promotes epithelial cell transformation and breast tumorigenesis via MAP3K8 activation. Carcinogenesis. 2014;35(6):1352–1361.
    View this article via: PubMed CrossRef Google Scholar
24. Kryczek I, et al. IL-22(+)CD4(+) T cells promote colorectal cancer stemness via STAT3 transcription factor activation and induction of the methyltransferase DOT1L. Immunity. 2014;40(5):772–784.
    View this article via: PubMed CrossRef Google Scholar
25. Xu X, et al. Increased intratumoral interleukin 22 levels and frequencies of interleukin 22-producing CD4⁺ T cells correlate with pancreatic cancer progression. Pancreas. 2014;43(3):470–477.
    View this article via: PubMed CrossRef Google Scholar
26. Jiang R, et al. Interleukin-22 promotes human hepatocellular carcinoma by activation of STAT3. Hepatology. 2011;54(3):900–909.
    View this article via: PubMed CrossRef Google Scholar
27. Kong X, et al. Interleukin-22 induces hepatic stellate cell senescence and restricts liver fibrosis. Hepatology. 2012;56(3):1150–1159.
    View this article via: PubMed CrossRef Google Scholar
28. Wolk K, et al. IL-22 regulates the expression of genes responsible for antimicrobial defense, cellular differentiation, and mobility in keratinocytes: a potential role in psoriasis. Eur J Immunol. 2006;36(5):1309–1323.
    View this article via: PubMed CrossRef Google Scholar
29. Lee YK, et al. Late developmental plasticity in the T helper 17 lineage. Immunity. 2009;30(1):92–107.
    View this article via: PubMed CrossRef Google Scholar
30. Siegmund K, et al. Unique phenotype of human tonsillar and in vitro-induced FOXP3⁺ CD8⁺ T cells. J Immunol. 2009;182(4):2124–2130.
    View this article via: PubMed CrossRef Google Scholar
31. Kryczek I, et al. IL-17⁺ regulatory T cells in the microenvironments of chronic inflammation and cancer. J Immunol. 2011;186(7):4388–4395.
    View this article via: PubMed CrossRef Google Scholar
32. Annunziato F, et al. Phenotypic and functional features of human Th17 cells. J Exp Med. 2007;204(8):1849–1861.
    View this article via: PubMed CrossRef Google Scholar
33. Messi M, Giacchetto I, Nagata K, Lanzavecchia A, Natoli G, Sallusto F. Memory and flexibility of cytokine gene expression as separable properties of human TH1 and TH2 lymphocytes. Nat Immunol. 2003;4(1):78–86.
    View this article via: PubMed CrossRef Google Scholar
34. Rivino L, Messi M, Jarrossay D, Lanzavecchia A, Sallusto F, Geginat J. Chemokine receptor expression identifies Pre-T helper (Th)1, Pre-Th2, and nonpolarized cells among human CD4⁺ central memory T cells. J Exp Med. 2004;200(6):725–735.
    View this article via: PubMed CrossRef Google Scholar
35. Zhou L, et al. TGF-β-induced Foxp3 inhibits TH17 cell differentiation by antagonizing RORgammat function. Nature. 2008;453(7192):236–240.
    View this article via: PubMed CrossRef Google Scholar
36. McGeachy MJ, et al. TGF-β and IL-6 drive the production of IL-17 and IL-10 by T cells and restrain T(H)-17 cell-mediated pathology. Nat Immunol. 2007;8(12):1390–1397.
    View this article via: PubMed CrossRef Google Scholar
37. Volpe E, et al. A critical function for transforming growth factor-beta, interleukin 23 and proinflammatory cytokines in driving and modulating human T(H)-17 responses. Nat Immunol. 2008;9(6):650–657.
    View this article via: PubMed CrossRef Google Scholar
38. Zhao Q, et al. Interleukin-17-educated monocytes suppress cytotoxic T-cell function through B7-H1 in hepatocellular carcinoma patients. Eur J Immunol. 2011;41(8):2314–2322.
    View this article via: PubMed CrossRef Google Scholar
39. Kuang DM, et al. Activated monocytes in peritumoral stroma of hepatocellular carcinoma foster immune privilege and disease progression through PD-L1. J Exp Med. 2009;206(6):1327–1337.
    View this article via: PubMed CrossRef Google Scholar
40. Zou W, Chen L. Inhibitory B7-family molecules in the tumour microenvironment. Nat Rev Immunol. 2008;8(6):467–477.
    View this article via: PubMed CrossRef Google Scholar
41. Brahmer JR, et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J Med. 2012;366(26):2455–2465.
    View this article via: PubMed CrossRef Google Scholar
42. McDermott DF, Atkins MB. PD-1 as a potential target in cancer therapy. Cancer Med. 2013;2(5):662–673.
    View this article via: PubMed Google Scholar
43. Topalian SL, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med. 2012;366(26):2443–2454.
    View this article via: PubMed CrossRef Google Scholar
44. Taube JM, et al. Colocalization of inflammatory response with B7-h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape. Sci Transl Med. 2012;4(127):127ra37.
    View this article via: PubMed Google Scholar
45. Zhao Q, et al. Activated CD69⁺ T cells foster immune privilege by regulating IDO expression in tumor-associated macrophages. J Immunol. 2012;188(3):1117–1124.
    View this article via: PubMed CrossRef Google Scholar
46. Kuang DM, et al. Tumor-activated monocytes promote expansion of IL-17-producing CD8⁺ T cells in hepatocellular carcinoma patients. J Immunol. 2010;185(3):1544–1549.
    View this article via: PubMed CrossRef Google Scholar
47. Wu Y, Zhao Q, Peng C, Sun L, Li XF, Kuang DM. Neutrophils promote motility of cancer cells via a hyaluronan-mediated TLR4/PI3K activation loop. J Pathol. 2011;225(3):438–447.
    View this article via: PubMed CrossRef Google Scholar
48. Kuang DM, et al. Peritumoral neutrophils link inflammatory response to disease progression by fostering angiogenesis in hepatocellular carcinoma. J Hepatol. 2011;54(5):948–955.
    View this article via: PubMed CrossRef Google Scholar