Current perspectives on the immunosuppressive tumor microenvironment in hepatocellular carcinoma: challenges and opportunities

TAMs inhibit anti-tumor immunity and promote tumor progression by expressing cytokines and chemokines. Preclinical studies have identified key pathways for TAMs recruitment, polarization and metabolism during tumor progression, and new therapies for these pathways can indirectly stimulate cytotoxic T-cell activation and recruitment [16‐18]. Clinical trials of therapeutic agents currently promoting phagocytosis or inhibiting the survival, proliferation, transport or polarization of TAMs have shown improvement of cancer outcomes (Fig. 1).

TAMs are significant components of the microenvironment in HCC and associated with a poor prognosis of HCC patients. TAMs expression and density have been assessed by immunohistochemical staining from 253 HCC patients and results showed that CD68⁺ TAMs were not associated with clinicopathologic features and prognosis in HCC. However, low presence of CD86⁺ TAMs and high presence of CD206⁺ TAMs were significantly associated with invasive tumor phenotypes and with poorer overall survival (OS) as well as reduced time to recurrence [19]. TAMs develop from monocytes to functional macrophages and acquire various immunosuppressive functions at each stage of its differentiation to maintain the tumor microenvironment [20]. Repolarization of TAMs towards an antitumor phenotype was one approach to contributing to tumor regression [21]. Yang Y et al. first demonstrated that tumor cell-derived Wnt ligand stimulates M2 to transduce the polarization of TAMs via classical Wnt/β-catenin signaling, which results in immunosuppression in HCC [22]. Therefore, blocking Wnt secretion in tumor cells and/or activation of Wnt/β-catenin signaling in TAMs may be a potential strategy for future HCC treatment.

Previous literature has demonstrated that TAMs can produce a variety of chemokines, such as CCL17, CCL18 and CCL22, which attract Tregs cells to cancer sites, thereby impeding cytotoxic T cell activation [23, 24]. This may lead to a positive feedback loop between TAMs and Tregs, providing a new dimension to the immunosuppressive effects of cancer. TAMs in HCC can also attract Tregs to regulate the tumor microenvironment. This phenomenon was found by Zhou J that the intratumoral distribution of FoxP3⁺ Tregs were associated with high density TAMs. The anti-interleukin-10 (IL-10) antibody partially blocked this increase, suggesting that TAMs may trigger an increase in the population of FoxP3⁺ Treg cells in the tumor, thereby promoting the progression of HCC [25]. Mechanistically, Wu Q et al. recently confirmed that TREM-1⁺ TAMs can respond to hypoxia and tumor metabolites via the ERK/NF-κβ pathway leading to accumulation of CCR6⁺ Foxp3⁺ Tregs [26]. This study highlights hypoxia-induced tumor immunosuppression by TREM-1⁺ TAMs that attracts CCR6 ⁺ Foxp3 ⁺ Tregs, and TREM-1⁺ TAMs confers HCC resistance to Programmed cell death 1 ligand 1 (PD-L1) treatment.

TAMs are able to produce angiogenic factors, such as vascular endothelial growth factor (VEGF), platelet-derived growth factor and transforming growth factor β. TAMS also induce angiogenesis by expressing matrix metalloproteinase (MMPs) [27‐29]. Angiogenesis inhibition therapy has recently become a promising therapeutic strategy for HCC. Lipid-based nanoparticles (NP) targeting CXCR4 were developed to deliver VEGF siRNA specifically to HCC as an anti-angiogenic substance. The researchers demonstrated that AMD-modified NPs (AMD-NPs) can effectively transfer VEGF siRNAs into HCC and down-regulate VEGF expression in vitro and in vivo. Although SDF1α/CXCR4 axis is upregulation in hypoxia induction after anti-angiogenesis therapy, AMD-NPs combined with conventional sorafenib therapy or VEGF siRNA inhibition of CXCR4 can prevent the infiltration of TAMs [30]. However, has there been any association between TAMs and VEGF in HCC? One study has investigated the role of macrophages in HCC progression under sorafenib treatment and explored whether combination of drugs that deplete macrophages improved the antitumor effect of sorafenib. Results showed that although sorafenib significantly inhibited tumor growth and lung metastasis, it induced a significant increase in peripheral recruitment and intratumoral infiltration of F4/80- and CD11b-positive cells, which causes an increase of VEGF. Depletion of macrophages by clodrolip or zoledronic acid (ZA) in combination with sorafenib significantly inhibited tumor progression, tumor angiogenesis, and lung metastasis [31]. These results indicate macrophages may have an important role in tumor progression under sorafenib treatment.

Preclinical studies have established that autophagy may play a role in the manipulation of TAMs function and tumor-related immunity. Chang CP et al. recently found that toll-like receptor 2 (TLR2)-associated ligands derived from HCC can stimulate the differentiation of M2 macrophages by selectively autophagy to control the homeostasis of NF-κβ RELA/p65 protein. TLR2 signaling induces NF-κβ RELA cytoplasmic ubiquitination and leads to degradation by SQSTM1/p62-mediated autophagy. Inhibition of autophagy will save NF-κβ activity and shape the phenotype of HCC polarized M2 macrophages [32]. The findings reveal a new role for autophagy in TAMs, which controls cellular function. Previous studies reported that Abies georgei’s natural product, named 747, is associated with xenophenol structure and exhibits sensitivity and selectivity as a CCR2 antagonist. In animals, 747 increases the number of CD8⁺ T cells in tumors by blocking tumor-infiltrating macrophage-mediated immunosuppression and inhibits in situ and subcutaneous tumor growth in a CD8⁺ T cell-dependent manner. In addition, 747 enhances the therapeutic effect of low-dose sorafenib without significant toxicity [33]. As shown in Table 1, treatment of TAMs targets in HCC has been documented and could be a new approach for treating HCC [34‐38].

In cancer, the differentiation of myeloid cells often changes, producing a group of immature myeloid cells, which have strong immunosuppressive activity and impaired function as antigen-presenting cells (APCs) [39]. These cells are now known as MDSCs, a heterogeneous population of immature myeloid cell. MDSCs are also plastic and respond to microenvironment signals [40‐42]. MDSCs can differentiate into macrophages, granulocytes and dendritic cells (DCs) in vitro. Therefore, MDSCs have significant diversity and plasticity, and are capable of changing their functional status in response to a variety of cytokines and growth factors in the tumor microenvironment. It has been reported that the immunosuppressive activity of MDSCs in the tumor microenvironment mainly includes (a) inducing differentiation and expansion of Tregs; (b) inhibiting the polarization of DCs, NKs and macrophages to the M2 phenotype; (c) depriving T cells of essential amino acids; (d) inducing oxidative stress (Fig. 2) [40‐43]. We mainly reviewed the MDSCs in the tumor growth, angiogenesis, and metastasis of HCC in this context (Table 1) [44‐52]. These evidence suggest that MDSCs contribute to the immunosuppressive network through multiple mechanisms and are potential immunotherapy targets for HCC.

MDSCs from HCC patients are unable to stimulate an allogeneic T-cell response, suppress T-cell proliferation, and have high arginase activity. Liu YT et al. used a hydrodynamic jet and transposon system to create a model that introduces a protein kinase and an open reading frames (ORFs) chromosomal encoding agent for tumor antigens. The transposon-based Akt/N-Ras-induced HCC mouse model allows researchers to monitor tumor growth non-invasively, quantify and characterize endogenous or over-transferred CD8⁺ T cell responses [53]. These characteristics make it a convenient preclinical model for the evaluation of immunological checkpoint inhibitors and cellular immunotherapy HCC. MDSCs was found to cause inhibition of CD8⁺ T-cell responses. In addition, recent studies have reported that inhibition of tumor CCRK or liver IL-6 increases interferon γ⁺ tumor necrosis factor-α⁺CD8⁺ T-cell infiltration and impaired tumorigenicity and is restored by recovery of MDSCs. It is worth noting that the tumor CCRK depletion up-regulated the expression of PD-L1 and increased the expression of intratumoral CD8⁺ T cells, thereby enhancing the effect of PD-L1 blocking HCC [54].

Additionally, these MDSCs cocultured with autologous T cells induce Treg expansion, which mitigates effector T-cell function. Kalathil S et al. recorded an increase in the number of Tregs, MDSC, PD-1⁺-exhausted T cells, and an increase in immunosuppressive cytokine levels in HCC patients, revealing the potential mechanistic network for immune disorders in HCC patients compared with the normal control group [55]. Other mechanisms have been described, in which MDSCs affect T-cell function, survival, and trafficking. Similar to TAMs, MDSCs express galectin-9 that binds to TIM-3 on T cells, inducing T-cell apoptosis [56]. MDSCs can also impair natural killer (NK) cell function. In HCC, MDSCs inhibit NK cell cytotoxicity and cytokine release, which is mediated by the NKp30 receptor [57]. Regarding the interaction between MDSCs and DC cells, it has been reported that MDSC inhibits TLR-ligand-induced IL-12 by IL-10 production and inhibits T-cell stimulating activity of DCs in HCC [58].

Because of the tumor-promoting and immunosuppressive effects of bone marrow cells, it is of great interest to target them to enhance the efficacy of conventional cancer therapies. The recently approved chemotherapeutic agent, trabectedin, not only targets tumor cells, but induces rapid apoptosis of bone marrow cells [59]. Clinical trials have shown that trabectedin also has a strong cytotoxic effect on liver cancer cells [60]. Therefore, trabectedin may be a potential treatment for HCC. Another potential target is estrogen, which inhibits myeloid cell function in HCC [61]. Estrogen inhibits IL-6 exposure to macrophages exposed to necrotic hepatocytes and reduces the risk of liver cancer in DEN-treated female mice [62]. Estrogen inhibits tumor necrotic cell-STAT6 activation by inhibiting Janus kinase, resulting in a reduced HCC model in mice [63]. Thus, estrogen therapy may help disrupt the development and function of HCC bone marrow cells. Therefore, targeting bone marrow cells represents a point of further research as a possible adjuvant therapy to attenuate HCC progression.

It is clear that bone marrow-derived cells, including TAMs, TANs, and MDSCs, promote tumor progression [64‐66]. In recent years, many studies have shown that TANs not only promote tumor growth, but also anti-tumor effects on tumors, and can regulate their different phenotypes through tumor signal transduction. TANs can be divided into two major subtypes, N1 (anti-tumor) and N2 (pro-tumor) and the plasticity of these subtypes depends on the presence of TGF-β [67, 68]. Neutrophils can polarize TGF-β to the N2 phenotype while TGF-β together with increased inhibition of IFN-β induces N1phenotype [69].

Depending on the microenvironment of the tumor, TANs can promote or inhibit tumor progression by releasing cytokines. Zhou SL and his colleagues found that CCL2 and CCL17 are the most highly expressed cytokines in peripheral blood neutrophils (PBNs) activated by TANs and HCC cells. The number of CCL2⁺ or CCL17⁺ TANs is related to tumor size, microvascular infiltration, tumor embedding, tumor differentiation and staging. Compared to PBN-conditioned medium, TAN-conditioned media and recombinant CCL2 and CCL17 increased the migration activity of HCC cells or mouse macrophages and Tregs [70]. Taken together, this evidence suggests that TANs recruit macrophages and Treg cells into HCCs to promote their growth, progression, and resistance to sorafenib. Surprisingly, the same team reported that the secretion secreted BMP2 and TGF-β2 and triggered the expression of miR-301b-3p in HCC cells, followed by inhibition of the limbal gene expression membrane protein (LSAMP) and CYLD Lysine deubiquitinase and increased stem cell characteristics in HCC cells. These TAN-induced hepatoma stem cell-like cells are active in NF-κB signaling, a higher level of secretion of C-X-C-thematic chemokine 5 (CXCL5) and recruitment of more TANs infiltration, suggesting a positive feedback loop [71]. Additionally, we have literature some important molecules which might target TANs in HCC (Table 1) [72‐76].

Recent studies have shown that communication between cancer cells and fibroblasts is very important, and these fibroblasts are called CAFs [77, 78]. CAFs promote the proliferation, invasion and metastasis of tumors by secreting various growth factors and cytokines [79, 80]. In HCC, Mano Y et al. established that primary CAFs and non-cancerous liver fibroblasts from 15 patients undergoing HCC resection. They found that endogenous and exogenous BMP4 activate hepatic fibroblasts to gain the ability to secrete cytokines and enhance the invasiveness of cancer cells. BMP4 is one of the regulators of CAFs function in the HCC microenvironment [81].

CAFs train NK cells to obtain an inactive phenotype and produce an unresponsive state in the tumor. Li T et al. found that HCC-derived fibroblasts induced NK cell dysfunction significantly better than foreskin-derived fibroblasts, which were characterized by low expression of cytotoxic molecules and cell surface markers, and impaired production of cytokines. They also noted that PGE2 and IDO from activated fibroblasts inhibit the activation of NK cells, thereby creating favorable conditions for tumor progression [82]. In addition, CAF can recruit regulatory DCs through IL-6-mediated STAT3 activation and educate them to obtain a tolerogenic phenotype and up-regulate Treg production by secreting TGF-β in the tumor microenvironment [83]. In summary, CAFs play an important role in the development and progression of cancer cells and targeting CAFs may be effective in treating fibrosis and preventing HCC progression (Tables 1, [84‐90].

Anticancer therapies targeting CAFs or cytokine inhibitors secreted by CAFs have recently been actively studied. Inhibitors of TGF-beta signaling have been shown to block HCC growth and progression by modulating EMT in different experimental models [91]. Several studies have explored the active targeting of CAFs to provide therapeutic compounds. A recent report presented anti-fibrotic drugs such as PRI-724, conophylline, follistatin, salvianolic acid, Gliotoxin. Curcumin, sulfasalazine, and tanshinone I, which inhibit activated HSC and/or induce apoptosis [92‐96]. It is believed that these drugs can not only control fibrosis, but also inhibit HCC by controlling the function of CAFs. It has been reported that fibroblast growth factor receptor 1 is overexpressed in the fiber layer HCC [97]. CAFs stimulate tumor cells through FGF and produces fibrosis [98, 99]. Treatments that target CAFs may be effective in fibrolamellar HCC. CAFs are one of the most important components in the tumor microenvironment and promote cancer cell growth and invasion through various mechanisms. CAFs are important for the initiation and progression of cancer cells and for the treatment of CAFs to effectively treat fibrosis and prevent HCC progression.

Tregs cells are the subset of CD4⁺ T cells and identified with the CD25 marker. CD4⁺ CD25⁺ compartment of cells is approximately 1–2% of peripheral CD4⁺cells. Tregs cells do not only limit autoimmune responses, they also dampen responses against microbial and viral antigens, allergens, tumors and allografts. Tumor-associated Tregs directly promote tumor evasion by a number of contact-dependent and contact-independent mechanisms [100, 101].

The number of Tregs has been reported to increase in patients with HCC [102]. Yang et al. found that the proportion and absolute number of CD4⁺CD25⁺ T cells in the surrounding area of the tumor were significantly increased [103]. The other group showed a higher frequency of Tregs in peripheral blood from HCC patients compared to HCV patients and healthy subjects. Mechanistically, Huh7 culture supernatants appear to promote CD4⁺CD25⁺ T-cell proliferation and inhibit CD4⁺CD25⁻ T-cell proliferation [104]. In addition, the frequency of circulating Tregs was linked to the disease progression and had a potential of serving as a significant biomarker in HCC patients [105, 106]. Sorafenib is a multikinase inhibitor that could reduce the frequency of hepatic infiltrating Tregs by suppressing TGF-β signaling [107]. As a result,these evidence indicates suppressing Tregs might be one of the significant targets for the induction of an immune response for HCC.

CTLA-4 is member of immunoglobulin superfamily homologous to CD28 with higher affinity for B7. CTLA-4 inhibits early T cell immune response, mainly in lymphoid tissues [108]. In naive T cells, CTLA-4 localized in the intracellular space and expressed on the cell surface upon stimulatory signals. Whereas in Tregs, CTLA-4 is constitutively expressed and is involved in Treg suppressive functions [109]. Several studies investigated PD-1 and PD-L1 expression in the context of HCC and found that PD-L1 expression was localized mainly in neoplastic or intratumoral inflammatory cells [110, 111]. Moreover, PD-1/PD-L1 interaction was demonstrated to contribute to immune suppression in HCC [112]. High level of PD-L1 expression was shown in tumors with a high number of tumor infiltrating lymphocytes (TILs) and shorter overall survival [113]. There are several clinical trials investigating the efficacy of monoclonal antibodies (mAb) against CTLA-4 and PD-1/PD-L1 as single agent treatment. In this review, we have extensively studied the efficacy of the current proposed immunotherapy for Tregs in patients with HCC. (Fig. 3) (Tables 1,2) [114‐127].