Antitumour immunity and tumour microenvironment

The principle of immunotherapy is based on the fundamental concept, ‘cancer immunoediting’, stating that the immune system can be protective to the host meanwhile prompting cancer progression.17 During immunoediting, the immune system shapes the immunogenicity of tumours in three phases—elimination, equilibrium and escape.17 In the first phase, both innate and adaptive immune systems detect and destroy early tumours before they become clinically visible.17 To eliminate tumour cells, dendritic cells are required to cross-present tumour neoantigens to naïve T cells, thereby subsequently allowing their activation.18 These activated effector T cells then release interferon-γ to mediate antitumour effects.18 Moving on to the equilibrium phase, the balance of antitumour immunity starts sloping to immune-mediated tumour dormancy, in parallel with tumour growth.16 Due to the pressure on antitumour immunity, tumour cells with less immunogenicity are more likely to survive the immune attack. These survived tumour cells can then undergo genetic and epigenetic modifying, leading to increased resistance of immune recognition in the tumour microenvironment and eventually immunosuppression,19 whereas the balance between tumour-suppressing and tumour-promoting effects of antitumour immunity is mainly mediated by adaptive but not innate immune system.19 Finally, given the accumulation of less immunogenic tumour cells, the antitumour immunity becomes exhausted and hence ineffective to eliminate tumour cells. Tumour cells evade immune recognition via multiple mechanisms, including the loss of surface markers such as tumour neoantigens and co-stimulatory molecules which can be readily recognised by immune cells. These mechanisms collectively shift the balance of antitumour immunity from being protective of the host towards favouring tumour progression.20

A solid tumour consists of different cells with tumour cells being the predominant cell type, while these tumour cells interact immune cells and their derivatives to form the tumour microenvironment.21 There are a variety of immune cell types in the tumour microenvironment. Some of them such as natural killer (NK) cells and active CD8⁺ T cells stimulate antitumour immune response to directly eliminate or induce cytotoxicity on tumour cells.22 In contrast, other immune cells suppress the antitumour immunity. For example, tumour-associated macrophages can differentiate into the anti-inflammatory M2-like phenotype, thereby inhibiting the antitumour function of cytotoxic CD8⁺ T cells and promoting intratumoral infiltration of immunosuppressive regulatory T cells (T_reg).23 Meanwhile, myeloid-derived suppressor cells (MDSCs) can coordinate with other immune cells to maintain an immunosuppressive microenvironment, which in turn favours the survival of tumour cells.24 In general, antitumour immunity is often over-ridden by tumour cells as a mechanism of immunoediting, while recent evidence has illustrated the crucial crosstalk between microbes and immune cells in cancer development.

Gut microbiota and immune alteration in CRC

Among all different cancer types, the correlation between the gut microbiota and CRC is undoubtedly being the most extensively studied. For instance, our team previously identified that the transplantation of gut microbiota from patients with CRC could induce colorectal tumourigenesis in conventional and germ-free mice by promoting pro-inflammatory cytokines expression and increasing helper T cells (Th)-1 and Th17 infiltration into the colon.25 Th17 in the small intestinal lamina propria could also been induced by gut-segmented filamentous bacteria accompanied with the release of serum amyloid A, a protein acting on dendritic cells to promote Th17 differentiation.26

Several bacterial species that are capable of promoting colorectal tumourigenesis have been identified. F. nucleatum recruits tumour-infiltrating myeloid cells, which can promote tumour progression.27 Moreover, F. nucleatum inhibits T cells and NK cells, which contribute to CRC progression.28 Peptostreptococcus anaerobius is a pathogenic bacterium that is highly enriched in patients with CRC. Consistent to the clinical findings, P. anaerobius could induce a higher incidence of intestinal dysplasia in CRC mouse models.29 Through single-cell sequencing, P. anaerobius was found to be able to suppress the host antitumour immunity, with the expansion of MDSCs, tumour-associated macrophages and neutrophils associated with inflammation and tumourigenesis.30 Parvimonas micra is another pathogenic bacterium that is enriched in patients with CRC, and it has been identified by several studies as an independent risk factor of CRC with poor survival. Mechanistically, P. micra could induce the differentiation of CD4⁺ T cells to Th17 phenotype and increased Th17 cells-secreted cytokines, thereby promoting inflammation.31 Collectively, these pathogenic bacteria promote CRC development through regulating immune microenvironment.

Gut microbiota and immune alteration in HCC

The gut microbiota not only affects the local intestines, but can also impact distal organs including liver though the gut–liver axis.32 The human liver has a double blood supply: the hepatic artery from the coeliac axis, and the portal vein bringing venous blood from the intestines and spleen. Of note, the portal venous blood flow from the mesenterium accounts for 75% of the blood supply to the liver.33 In healthy or non-disease conditions, the gut–liver axis is maintained by the intestinal barrier, which strictly limits the access of gut bacteria and their derivatives into the portal vein circulation. However, the intestinal barrier can be interrupted by (1) dysbiosis or abnormal composition of the gut microbiota, (2) an impaired mucosal barrier and (3) loss of integrity of the immune system, leading to increased intestinal permeability.32 As a consequence, gut bacteria and their derivatives now have easier access to the portal vein circulation and further reaching the liver. On the other hand, some endotoxins including lipopolysaccharide (LPS) and peptidoglycans can accumulate in the liver, eventually contributing to hepatic inflammation and pathogenesis of liver diseases.32

The gut microbiota is now increasingly acknowledged for its crucial role in hepatocarcinogenesis. We recently reported that the gut microbiota and metabolites contribute to the development and progression of non-alcoholic fatty liver disease (NAFLD)-associated HCC, of which transplanting faeces from NAFLD-HCC mice could induce liver dysplasia with increased inflammation in germ-free mice.34 Using different HCC mouse models, Ma et al showed that commensal gut bacteria especially Clostridium spp secrete bile acids to increase hepatic CXCR6⁺ NKT cells, thereby strengthening the host antitumour immunity and suppressing tumour progression in the liver.33 Moreover, another study by Behary et al showed that the gut microbiota from patients with NAFLD-HCC induce a T cell immunosuppressive phenotype with decreased CD8⁺ T cells and increased T_reg.35 The deficiency of Akkermansia muciniphila may induce MDSC infiltration, which promotes liver inflammation, fibrosis and HCC. Taken together, the interplay between gut microbiota and host immune cells plays an important role in NAFLD-HCC progression.

Hepatitis viral infection is another common aetiology of HCC. The gut microbiota in patients with chronic hepatitis B was found to be greatly altered with the significant depletion of beneficial lactic acid bacteria and other probiotics such as Faecalibacterium prausnitzii, accompanied with increased serum level of pro-inflammatory cytokine tumour necrosis factor (TNF)-α.36 Meanwhile, excess alcohol consumption can lead to the onset of alcoholic fatty disease (ALD), which may subsequently progress to liver malignancy. The development of ALD-associated HCC is closely related to the gut microbiota, of which gut microbes can metabolise ethanol to produce metabolite acetaldehyde, a compound able to destroy the integrity of intestinal barrier.37 Moreover, similar to viral hepatitis, a marked reduction of gut probiotics including Lactobacillus and Lactococcus was observed in patients with ALD.38 Hence, the dysbiotic microbiota and increased microbial translocation collectively promote the progression from liver premalignancy to HCC.

Gut microbiota and immune modulation in melanoma

Similar to the gastrointestinal tract, human skin surface is covered by a large number of microbes which also provide a unique milieu for host–microbes interaction. For example, the skin microbiota can communicate with innate immune system to produce antimicrobial molecules as a primary defence against foreign pathogens.39 More specifically, the increased interleukin-17 produced by Th17 protects the host against microbial invasion that passes through the skin.39 Of note, microbiota alteration in the skin can affect the immune system, potentially contributing to melanoma development. On the other hand, various recent studies have demonstrated the association of gut microbiota with the response to melanoma immunotherapy,40 suggesting the complex relationship between skin and gut microbes in melanoma. Nevertheless, the mechanism of how gut microbiota modulates immune response during melanoma development needs further investigation.

Immune checkpoint inhibitors

ICIs play an important role in regulating the host immune system. Under normal conditions, these checkpoints help to maintain a balance between co-stimulatory and inhibitory signals of T cell-mediated immunity, thereby ensuring peripheral tolerance and preventing collateral tissue damage caused by overstimulated immune responses against invaded pathogens.41 However, characteristics of these immunosuppressive checkpoints can be manipulated by tumour cells and used as a means to evade surveillance of antitumour immunity. To restore and improve antitumour immune response, it is plausible to develop therapeutic approaches that aim at suppressing the immune checkpoints. To date, there are multiple monoclonal antibodies approved by the US FDA as ICIs that target the two major negative regulators of immune system, programmed death 1 (PD-1) and its ligand (PD-L1), as well as cytotoxic T lymphocyte-associated antigen 4 (CTLA-4).

PD-1 is a major checkpoint receptor whose expression is upregulated in immune cell lines undergoing programmed cell death.42 The ligand PD-L1 is highly expressed in different human cancer types and involved in immune evasion, of which mouse tumour cells with forced PD-L1 expression could not be eliminated by immune response. Monoclonal antibodies against PD-L1 could promote antitumour immune responses, resulting in tumour rejection, hence lays the foundation of anti-PD-1 blockade.43 PD-1 immunosuppressive pathway is initiated by PD-L1/PD-L2 ligation, and its downstream signalling leads to decreased production of pro-inflammatory cytokines and T cell receptor (TCR) inhibition through dephosphorylating intermediates in TCR signalling pathway.43 Taken together, PD-1 pathway causes T cells to become anergic and express an ‘exhausted’ phenotype.44

CTLA-4 is expressed on effector T cells in a regulated fashion while is constitutively expressed on T_reg.45 CTLA-4 is homologous to CD28, the protein expressed on T cells that provides co-stimulatory signal for T cell activation and proliferation upon interaction with CD80 and CD86 situated on antigen-presenting cells. CTLA-4 has a higher affinity for CD80/CD86 than CD28; hence, it is able to outcompete CD28 for binding, resulting in the disruption of co-stimulatory signal.45 Meanwhile, CTLA-4 exerts immunosuppressive function via other mechanisms. The critical role of CTLA-4 in maintaining immune tolerance has been demonstrated, while the efficacy of anti-CTLA-4 is largely dependent on gut microbiota.46

Manipulation of microbiota in tumour microenvironment

Increasing evidence has illustrated the presence of microbes within the tumour microenvironment.47 48 Tumour growth involves the development of new vasculatures due to the increasing demand for oxygen in rapidly proliferating tumour cells; hence, the tumour microenvironment is widely considered to be necrosis and hypoxia. Such hypoxic condition in the tumour niche could attract the entry of microbes into tumours and favour their replication.49 These approaches mainly aim to facilitate the detection of microbial structural components such as LPS, peptidoglycan, flagellin and other pathogen-associated molecular patterns by innate pattern recognition receptors, subsequently stimulating antitumour immune responses.50 Moreover, the antitumour effects of tumour-colonising anaerobic bacteria have been investigated. The presence of bacteria in tumours could cause inflammation which further induces potent immune responses and activates effector CD8⁺ T cells to contribute enhanced tumour immunosurveillance and clearance.51 Nevertheless, whether approaches that target intratumoral bacteria to promote inflammation are safe and effective in human patients requires further investigations. In contrast, several pathogenic bacteria were found to be enriched in the tumour microenvironment, particularly F. nucleatum which is closely linked to colorectal adenocarcinoma.52 F. nucleatum can inhibit the cytotoxicity of NK cells and antitumour effector T cells, thereby restraining their ability to kill tumour cells.53 In this case, elimination of bacteria, if achievable, should be considered instead to enhance antitumour immune response.

Gut microbiota and immunotherapy response in gastrointestinal cancer

Recent studies have suggested that the gut microbiota is closely related to the efficacy of ICIs in gastrointestinal cancer. For instance, an observational study including 74 patients with different gastrointestinal cancers found that the microbiota composition is significantly different between responders and non-responders of immunotherapy.54 In particular, short-chain fatty acid-producing bacteria including Eubacterium, Lactobacillus and Streptococcus were positively associated with ICI response.54 Moreover, using CRC mouse models, Mager et al identified that Bifidobacterium pseudolongum could modulate ICI response through generating the metabolite inosine.13

Gut microbiota and immunotherapy response in HCC

The anti-PD-L1 antibody atezolizumab and the vascular endothelial growth factor-neutralising antibody bevacizumab combined therapy has been approved as the first-line therapy for HCC by the FDA, European Medicines Agency and other regulatory agencies.55 However, the efficacy of immunotherapy for patients with HCC is largely varied among patients with a response rate of 15%–30%.56 57 This suggests that immunotherapy efficacy is dependent on the aetiology of HCC, with diverse hepatic environments distinctly regulating HCC formation and immune response.58 Particularly, patients with NAFLD-induced HCC were found to be resistant to ICIs. ICI treatment explicitly increased tumour growth and tumour number in mouse models of NAFLD-HCC induced by choline-deficient high-fat diet. ICIs induce the activation of exhausted CD8⁺ PD1⁺ CXCR6⁺ TNF⁺ T cells, which promote non-alcoholic steatohepatitis-HCC instead of regressing tumour growth.55

Gut microbiota and immunotherapy response in melanoma

In the landmark study by Sivan et al, mice from two different providers Jackson Laboratory (JAX) or Taconic Farms (TAC), which had distinct microbiota compositions, displayed different rates of subcutaneous melanoma growth of which TAC mice had fewer intratumoral infiltrating CD8⁺ T cells with faster tumour growth than JAX mice.59 Through microbial profiling, Bifidobacterium was identified to be responsible for a more effective antitumour immune system. When combining Bifidobacterium administration with PD-L1 blockade, the efficacy of anti-PD-L1 was improved with enhanced dendritic cell maturation as well as their ability to process antigen and produce cytokines.59 At the same time, several clinical studies involving patients with melanoma showed the difference in microbiota composition between responders and non-responders receiving anti-PD-1 treatment.11 60 61 In responders, the gut microbiota is more diverse with the enrichment of Faecalibacterium, in contrast to lower diversity and enriched Bacteroidales observed in non-responders.11 Similarly, B. longum, Collinsella aerofaciens and Enterococcus faecium were enriched in responders of patients with metastatic melanoma.60 Faecal microbiota transplantation induced the alteration of immune cells and promotes the efficacy of anti-PD-1 therapy in patients with immunotherapy-refractory melanoma.61 Further mechanistic study showed that the bacterial genus Enterococcus improves ICI immunotherapy by producing SagA.62 Together, these findings suggest that altering the gut microbiota by increasing the abundance of several specific microbes could be a potential approach to enhance the efficacy of PD-1 blockade.