The immune microenvironment and progression of immunotherapy and combination therapeutic strategies for hepatocellular carcinoma

Hepatocellular carcinoma (HCC) accounts for 75%-85% of all primary liver cancers and is the leading cause of cancer-related deaths. China accounts for almost half of the global incidence and deaths of HCC. The poor response of chemotherapeutics and targeted drugs may be due to the drug resistance, heterogeneity of HCC, severe chronic liver damage and cirrhosis. Restoration of the liver microenvironment changes caused by chronic injury is crucial. Immunotherapy recently seems to show promise for the treatment of HCC induced by inflammatory injury. However, the unique liver immune system and resident immune tolerance state also pose a challenge for HCC immunotherapy. Different combinations of strategies have been developed for enhancement of HCC treatment. Here, we will discuss the immune microenvironment and progression of immunotherapy and combination therapeutic strategies for HCC.


INTRODUCTION
Hepatocellular carcinoma (HCC) accounts for 75%-85% of all primary liver cancers. Due to the rapid progression of HCC, the lack of effective treatment programs, and poor prognosis makes it the fourth leading cause of cancer-related deaths [1] . Due to regional differences in medical diagnosis and treatment, more than half of the new cases and deaths of HCC each year occur in the Asia-pacific region. Patients with early HCC in Europe and United States can be diagnosed and effectively treated in time [2] . More than 70% of HCC patients do not benefit from medical therapy. The vast majority of HCC patients present with an advanced stage at diagnosis, and the most effective surgical programs are often challenging to implement. In the past ten years, dozens of promising chemotherapeutics have failed the phase III trial, with only sorafenib demonstrating a low objective response rate and a slight increase in survival [3] . Research on targeted drugs for cell proliferation, metastasis and angiogenesis are encouraging, such as regorafenib and lenvatinib, although the overall survival rate remains dissatisfactory [4] . The ineffectiveness of chemotherapeutics and targeted drugs may be due to drug resistance and heterogeneity of HCC. HCC is usually accompanied by severe chronic liver damage and cirrhosis. Hence, anti-HCC drugs require a good balance of therapeutic response and drug toxicity and this often limits the application of highly active compounds with high toxicity [5,6] . Therefore, restoration of the liver microenvironment caused by chronic injury should be incorporated in the holistic management of HCC. In recent years, immunotherapy has been used in the treatment of various solid tumors. This was observed through the checkpoint inhibition of programmed cell death 1/programmed cell death ligand 1 (PD-1/PD-L1) and cell toxic T lymphocyteassociated protein 4 (CTLA-4) while improving the tumor immune microenvironment which seems to be particularly relevant for the treatment of HCC. However, the unique liver immune system and resident immune tolerance state make it different from other organs. Besides, continuous matrix remodeling of the malignant hepatocyte transformation caused by chronic inflammation and scars has created an immunosuppressive microenvironment that promotes the development of HCC, posing a challenge for HCC immunotherapy [7] .

THE IMMUNE MICROENVIRONMENT OF THE LIVER AND HCC
The liver has a unique immune regulation and balance mechanism. On one hand, the portal vein system is directly exposed to gastrointestinal pathogens and requires an effective immune response. On the other hand, it needs to deal with a large number of harmless blood antigens and maintain the immune tolerance of the liver [8] . In most cases, the liver is in a physiological immune tolerance state [9] . Most non-parenchymal cells, such as live sinusoidal endothelial cells (LSEC), Kupffer cells (KCs) and hepatic dendritic cells (HDC), gather in the liver sinusoids. It constitutes the physiological basis of the liver's immunosuppressive microenvironment [10,11] . LSEC has the dual functions of immune surveillance and immune tolerance. It acts as an antigen-presenting cell (APC) to present pathogens or tumor antigens [12] . At the same time, it inhibits the excessive responses of DC and T lymphocytes to bacterial antigens from the portal system [13][14][15] . KCs maintain immune tolerance by engulfing pathogenic microorganisms derived from the intestine, secreting inhibitory factors (such as IL-10 and prostaglandins) and activating the proliferation of regulatory T cells (Tregs) [16][17][18][19] . Besides, HDC is also a component of the liver immune tolerance by reducing the expression of MHC II and co-stimulatory molecules [20] . In summary, this immune-tolerant physiological environment creates a huge obstacle to the host's anti-tumor immunity.
The pathogenesis of HCC is characterized by destruction of the sinusoidal structure by a viral infection and inflammatory injury, impairment of immune surveillance and immune tolerance functions leading to liver cirrhosis and liver cancer [6,10] . The high-risk factors of HCC (hepatitis virus, alcohol, aflatoxin, etc.) drive hepatocyte DNA damage, endoplasmic reticulum stress and necrosis, which in turn leads to the formation of regenerative nodules, proliferative nodules and ultimately HCC [21] . HCC has abundant immune cell infiltration, which is the immune response of the host trying to clear the tumor. Unfortunately, this immune response is often dysregulated [22] . Tumor-infiltrating lymphocytes (TILs) account for a high proportion of HCC [23,24] , but these ineffective TILs often prove to be insufficient to control tumor growth [25] . The increased FoxP3+ Treg may impair the effector function of CD8+ T cells, which exacerbates the immunosuppressive microenvironment in HCC and is associated with a poor prognosis [26] . In addition, adaptive immune cells (such as CD8 + T cells, Th17 cells and B cells) can also stimulate the development of HCC [27,28] . There are a large number of bone marrow-derived suppressor cells (MDSCs) and Tregs in the microenvironment of HCC, which evade immune surveillance through a variety of mechanisms, such as the expressing highlevels of SOCS3 and IL-10 to limit immune cell activation [29] and secretion fibrosis factor TGF-β. This is used to build an environment of immunosuppression and drug resistance [30] , and directly down-regulates the expression of T cells or NK cell activation ligands (MHC class I and NKG2D, etc.) [31,32] . Therefore, the immunosuppressive environment of HCC is an arduous challenge to the host's immune system, which makes immunotherapy a promising approach for HCC treatment in the future.

THE STRATEGIES OF HCC IMMUNOTHERAPY
According to the immunological basis of HCC, we divide current HCC immunotherapy into four categories, including immune checkpoint inhibitors, oncolytic virus therapy, HCC vaccine and chimeric antigen receptor T (CAR-T) cell therapy [ Figure 1]. Due to the destruction of the HCC sinusoidal structure, it is difficult for LSEC and HDC to complete the antigen presentation process. Therefore, the specific DC vaccine is obtained by impinging tumor-associated antigen (TAA) or tumor lysate into DC in vitro. It activates cytotoxic T lymphocytes (CTLs) through major histocompatibility complex (MHC) class II-TCR antigen presentation and CD40/CD80/CD86-CD28 interaction. CTLs recognize and destroy tumor cells containing HCC-related antigens on MHC class I molecules. In addition to blocking the antigen presentation process, cancer cells will evade CTLs by upregulating immune checkpoint ligands, such as PD-L1 binding to the PD-1 receptor on the surface of CTLs to exhaust it, and CTLA-4 blocking the interaction between CD40/CD80/CD86 and CD28. Therefore, antibodies against PD-L1/PD-1 and CTLA-4 are used for immune checkpoint inhibitor therapy of HCC. The other alternative is more direct, by cloning in vitro chimeric antigen receptor T cells that can target specific antigen genes [such as Glypican 3 (GPC3) or alpha-fetoprotein (AFP)] related antigens to directly kill tumor cells. Finally, genetically engineered

IMMUNE CHECKPOINT INHIBITORS
Tumor cells express a variety of immunosuppressive ligands on their surface, which bind to the indicated inhibitory receptors of activated T cells involved in the anti-tumor response. This process in turn reduces the intensity of the anti-tumor immune response, thereby evading immune surveillance [33] . Drugs that block these immunosuppressive targets to eliminate tumor immune escape are called immune checkpoint inhibitors (ICI). PD-1 is a member of the CD28 superfamily and is expressed on the surface of T cells and B cells. Its activation will lead to the phosphorylation of ITSM (Immunoreceptor tyrosine-based switch motif) in the cytoplasm of the cell, inhibiting energy metabolism in T cells, thereby hindering cell cloning proliferation and secretion of cytokines. In order to avoid the killing of T cells, tumor cells highly express PD-L1 and release the PD-L1 into the peripheral blood, which causes the exhaustion of T cells and the loss of tumor antigen presentation ability of myeloid immune cells [34] . Therefore, targeted inhibition of the interaction of PD-1 and PD-L1 is of great significance for the treatment of HCC.
Nivolumab, as the first PD-1 targeted drug to be used in clinical practice, was initially used in the treatment of melanoma, and its objective response rate and one-year survival rate were 40.0% and 72.9%, respectively [35] . Subsequently, Nivolumab was tried to treat advanced HCC. Among 144 HCC patients, 20% showed a good response to nivolumab, and 3 of them achieved complete remission (CR), highlighting the potential of Nivolumab to treat advanced HCC [36] . Another anti-PD-1 targeted drug Pembrolizumab has also shown effectiveness in the treatment of advanced HCC, with an objective response rate and a oneyear survival rate of 17% and 54% [37] . In fact, a single ICI is not satisfactory for the treatment of advanced HCC. The current ICI therapy is mostly performed in a variety of combinations (for example, anti-PD-L1 antibody plus anti-CTLA-4 antibody), which is more effective than a single agent. In the absence of targetable lymphocytes in the tumor microenvironment, inhibition of PD-1/PD-L1 cannot stimulate cancer immunity, and inhibition of the CTLA-4 can cause CD8 + T cells to proliferate in the lymph nodes and infiltrate the tumor tissue, thereby enhancing the efficacy of anti-tumor. In fact, combination therapy of molecularly targeted drugs and immune checkpoint inhibitors has received considerable attention. For example, immunosuppressive cytokines that cause the immunosuppressive liver environment of patients with liver cancer, such as interleukin (IL)-10, transforming growth factor (TGF)-β and vascular endothelial growth factor (VEGF) molecular targeted drugs [38,39] . Table 1 shows the ongoing use of ICI in combination with various interventions (such as kinase inhibitors, cytokine or receptor inhibitors, and embolotherapy).

ONCOLYTIC VIRUS THERAPY
The oncolytic virus can specifically host in cancer cells, replicate and destroy the cell structure and hence was not initially classified as immunotherapy. Subsequent studies confirmed that oncolytic viruses could induce anti-cancer immune responses and immunogenic cancer cell death, making them a form of immunotherapy [47] . Compared with traditional therapies, oncolytic virus therapy is safer, has the selective specificity of host cancer cells, and continuously self-replicates to lyse cancer cells [48] . In the tumor microenvironment, pathogen-associated molecular patterns (PAMP) of oncolytic viruses can be recognized by pattern recognition receptors (PRR) of immune cells, such as through TLR or MDA5 activation of macrophages or dendritic cells [49,50] . As a secondary effect, oncolytic viruses enhance the recognition and presentation of tumor antigens, and activate the infiltration of cytotoxic T cells into tumors [51] . Therefore, oncolytic virus therapy is a very interesting method to overcome HCC immunosuppression. Currently, oncolytic virus therapies used for HCC include dsDNA or ssRNA viruses, such as measles vaccine virus (MeV), herpes simplex virus (HSV), adenovirus (Adenovirus) and vaccinia virus (VV), etc., which are used to engineer infection vectors [52] . For example, inserting the overexpression sequence of granulocyte- macrophage colony-stimulating factor (GM-CSF) into the oncolytic virus sequence, GM-CSF recruits myeloid cells in the periphery to enhance the immune response in the tumor microenvironment [53] . So far, preclinical studies for HCC oncolytic virus therapy have been very encouraging. We have compiled preclinical studies on HCC oncolytic therapy for the past ten years, as shown in Table 2.
Although many preclinical research attempts have been made in oncolytic therapy in recent years, there are still very few programs that have entered the clinical stage. At present, the only HCC oncolytic virus entering clinical research is JX-549, with VV as an engineered vector. VV has the stability and efficiency of intravenous administration, is widely used in the safety of live vaccines, has the advantages of immuneinducing activity and better editability, and has become a carrier of various engineered tumor-melting viruses [73][74][75] . The thymidine kinase gene (TK) gene of JX-594 (also known as PexaVec; Jennerex Inc.) was deleted to make it more specific for cancer cell infection. In addition, hGM-CSF and β-galactosidase were inserted to enhance its immunostimulatory activity and replication capabilities [73,76,77] . JX-594 showed complete tumor response and systemic efficacy in a phase I clinical study [78] . In the phase II trial, lowdose JX-594 has significant anti-cancer effect and immune activation ability [79] , but this requires earlier interventional therapy [80] . Currently, a large-scale 600-person multicenter Phase 3 trial is still in progress (NCT02562755). More clinical studies of HCC oncovirus are shown in Table 3.

HCC VACCINE
Tumor vaccine is a treatment program to increase the specificity of tumor antigens, mainly antigen peptide vaccines and DCs vaccines, which are used to stimulate specific immune responses. The clinical trials of therapeutic vaccines for HCC are summarized in Table 4. At present, there are relatively few registered clinical trials for DCs vaccines in HCC, partly because of the unsatisfactory results of previous clinical trials of such vaccines [86] . On the other hand, the tumor heterogeneity of HCC also limits the development of a single antigen peptide or DCs vaccine. With the development of large-scale DNA sequencing technology, patient-specific multi-target peptide or DCs vaccine is still a promising strategy for the treatment of HCC. DC, as professional antigen-presenting cells (APC), recognize, process and present TAA. Allogeneic DC vaccines can provide T cells with antigens and co-stimulatory molecules needed for immune response.
In short, DCs are mobilized from peripheral blood and their expansion is stimulated with GM-CSF to produce DCs for reinfusion. Prior to this, DC needs to be exposed to TAA to trigger the specificity of the vaccine [87] . DCs can be transduced with DNA or RNA encoding known TAA, or directly co-cultured with patient tumor lysate [88] . Phase I clinical studies have shown that the allogeneic DCs vaccine can produce a specific immune response in 73% of HCC patients [89] .

CHIMERIC ANTIGEN RECEPTOR T CELL THERAPY
In addition to immune checkpoint inhibitors, oncolytic viruses and vaccines, adoptive therapy using genetically modified T cells have also become one of the potential immunotherapy options for HCC. T cells can be engineered to express a chimeric antigen receptor (CAR), which is composed of a T cell receptor CD3ζ chain and co-stimulatory receptors (e.g., CD28 and TNFRSF9) to form an antigen recognition domain [90] . The antigen recognition domain endows CAR-T cells with specificity for tumor-associated antigens, which shows promise in the treatment of HCC. Besides, CAR-T cells have a strong adaptive immunity and can recognize antigens that are not present in MHC molecules. CAR-T cell therapy has been used in the preclinical treatment of a variety of solid tumors, but there are few clinical studies on HCC, and more are still in the preclinical research stage. Like the HCC vaccine, the technical difficulty lies in the choice of tumor-specific antigens [91] . CD133 is expressed by cancer stem cells derived from various epithelial cells and is an attractive cancer treatment target. CAR-T cells targeting CD133 have shown the feasibility of treating advanced HCC, with controllable toxicity and effective activity [92] . Glypican-3 (GPC3) is a member of the heparan sulfate glycoprotein family and belongs to a transmembrane glycoprotein. It plays an important role in cell proliferation, differentiation and metastasis. CAR-T cells targeting glypican-3 can inhibit the growth of HCC [93,94] . Besides, there are HCC recognition antigens such as NKG2D [95] and CD147 [96] for CAR-T cell transformation. In addition, the CAR of CAR-T cells can be inserted into the expression of a variety of cytokine genes to overcome the immunosuppressive effects of the HCC microenvironment [97,98] . The clinical trials of CAR-T cell therapy for liver cancer are summarized in Table 5.

THE CURRENT COMBINATION OF THERAPEUTIC STRATEGIES FOR HCC
Currently, there are many immunotherapy and other target therapy drugs approved by the . Single agent therapy has historically shown poor results in HCC, leading to trials of combination therapy for a more efficacious outcome. For example, the FDA has approved Opdivo (nivolumab) + Yervoy (ipilimumab) based on the CheckMate 040 trial, atezolizumab + bevacizumab for patients with advanced HCC based on the IMbrave150 (NCT03434379) study. The CheckMate 040 is a multicentered, open-labelled, multicohort, phase 1/2 study. The result showed that nivolumab + ipilimumab had manageable safety, promising objective response rate, and durable responses. The arm A regimen (4 doses nivolumab 1 mg/kg + ipilimumab 3 mg/kg every 3 weeks then nivolumab 240 mg every 2 weeks) received accelerated approval in the US based on this study [99] . The IMbrave150a study is a global, open-labelled, phase 3 trial for patients with unresectable HCC who had not previously received systemic treatment. The study included 336 patients in the atezolizumab + bevacizumab group and 165 patients in the sorafenib group. The result showed that atezolizumab + bevacizumab resulted in better overall (overall survival at 12 months was 67.2% vs. 54.6%) and progression-free survival (6.8 months vs. 4.3 months) outcomes than sorafenib [100] . There are many different combinations of immune checkpoint inhibitors with other different therapeutic strategies under investigation. Some of the combination clinical trials are concluded in the Table 1.

CONCLUSION AND PROSPECT
Immunotherapy is a revolution in HCC treatment. Significant responses have been observed in various tumor types with immunotherapy, especially immune checkpoint inhibitors and CAR-T cells. However, it is clear that not all HCC patients are sensitive to current immunotherapy, and even in those who do respond, the effect is difficult to last. Lots of data indicate that most HCCs are immunosuppressive  tumors. Therefore, ongoing research using a multifaceted approach to enhance the activity of the immune environment remain underway to enhance current immunotherapy strategies.

Authors' contributions
Drafted the outline of this review: Feng ZY, Xia HP Drafted the manuscript: Feng ZY, Xu FG, Liu Y, Xu HJ, Wu FB, Chen XB, Xia HP Finalized the manuscript: Chen XB, Xia HP