HMGB1 is a classical RAGE ligand with recognized roles in tumor biology. Of note, HMGB1 is overexpressed in gastric tumors, and a recent systematic meta-analysis demonstrated that HMGB1 plays an important role in GC, and its expression significantly correlates with tumor pT stage[74].

HMGB1-mediated RAGE activation triggers either MAPK or PI3K/AKT signaling and thus enhances tumor cell proliferation in different cancer cell types[75,76]. Additionally, data derived from in vitro experiments support the role of HMGB1 in promoting GC cell proliferation and migration via activation of RAGE-mediated Akt-mechanistic target of rapamycin (mTOR) and ERK-CREB signaling pathways, as well by initiation of a positive feedback loop that increases RAGE expression via these pathways. Furthermore, HMGB1 increases the expression levels of cyclin D1, cyclin E1, and proliferating cell nuclear antigen (PCNA), and HMGB1 knockdown reduces the expression of these proteins in GC cells[77].

The role in cancer cell proliferation of the S100 Calcium Binding Protein A16 (S100A16) remains controversial in some cancer types[78,79]. However, this alarmin is reported to promote cancer cell proliferation, migration, and invasion. In vivo and in vitro analyses revealed that S100A16 promotes both tumor cell proliferation and migration, whereas S100A16 knockdown abolished both[80]. Autophagy is a lysosome-mediated, self-degradation process that has emerged as key in protecting normal cells via the elimination of damaged organelles and protein aggregates and supporting bioenergetic homeostasis[81]. Autophagy can suppress the initiation of tumor development, or, in contrast, activate tumor cell survival, growth, and malignancy by facilitating access to metabolic demands during tumor progression and supporting tumor cell survival under stress[82-84].

The role of autophagy in GC progression, metastasis, and overall prognosis has been extensively documented[85,86]. Of note, HMGB1 released by autophagic GC cells through a RAGE-dependent mechanism acts as a pro-survival signal for remaining tumor cells, supporting cancer cell survival and promoting resistance to chemotherapy[87] (Figure 2).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Overview of RAGE-dependent intracellular signaling. RAGE binding triggers myriad mechanisms that support tumor proliferation and survival, invasion, and metastasis. AGE: Advanced glycation end product; AP-1: Activator protein-1; CREB: cyclic-AMP response element binding; ERK: Extracellular signal-regulated kinase; HMGB1: High mobility group box 1; JAK: Janus kinase; JNK: c-Jun n-terminal kinase; NF-ҡB: Nuclear factor-kappa B; PAMP: Pathogen-associated molecular patterns; RAGE: Receptor for advanced glycation end products; SAPK: Stress-activated protein kinase; STAT: Signal transducer and activator of transcription; TIRAP: Toll-interleukin 1 receptor domain-containing adaptor protein.

Decades of research have demonstrated that the RAGE axis and its associated proinflammatory downstream pathways act as key mediators of carcinogenesis[88], especially in gastrointestinal tumor development[89].

Among different cancer cell types, RAGE overexpression has been largely recognized as a pathological feature associated with increased tumor development, high invasiveness of cancer cells, and overall poor clinical prognosis[90-92].

The first experimental evidence supporting a correlation between RAGE overexpression and increased invasive and metastatic activity of GC cells was reported in 2002 by Kuniyasu et al[93]. This study revealed that RAGE axis signaling in gastric tumors may promote the acquisition of an invasive phenotype by increasing MMP-2 and MMP-9 expression, decreasing E-cadherin expression, and promoting tumor growth in a MAPkinase-dependent manner.

More recently, the analysis of RAGE expression in primary GC tissue reported by Wang et al[94] demonstrates that the upregulation of RAGE is associated with histological grade, nodal status, metastasis status, and American Joint Committee on Cancer (AJCC) stage in GC. RAGE upregulation is also associated with shorter overall survival rates. These data suggest that RAGE expression could be considered an independent predictor of GC aggressiveness and prognosis.

RAGE axis activation can have great influence on the invasive capacity of GC tumor cells through the modulation of cellular adhesion, motility, and production of MMPs. These functions are critical for epithelial-to-mesenchymal transition (EMT)[9,10], a complex process by which epithelial cells transdifferentiate to mesenchymal phenotypes. EMT is normally present during embryogenesis, tissue morphogenesis, wound healing, and regulation of stem cell behavior[95]. During this process, epithelial cells lose basal adhesion and demonstrate downregulation of epithelial cell markers (e.g., E-cadherin) and upregulation of mesenchymal markers (e.g., vimentin and fibronectin). These changes are also accompanied by increased expression of MMP-2 and MMP-9, which favors the degradation of ECM proteins[96]. EMT is a key mediator of the invasive and dissemination capacities of almost all types of cancer cells[97], including GC[98,99].

RAGE ligands have also been associated with cancer cell invasion and progression[100,101]. For example, the cancer-promoting role of HMGB1 has been extensively reported as a prognostic factor in many cancer types[74]. Importantly, differential expression analysis has shown that HMGB1 is increased in gastric adenocarcinoma cells[100] and has been directly associated with higher tumor stage, lymph node metastasis, depth of invasion, macrophage infiltration in the tumor microenvironment, and antineoplastic drug resistance[102-105]. Moreover, increased HMGB1 has also been correlated with poor prognosis and overall survival in GC[106-109].

In vitro and in vivo assays have demonstrated that high concentrations of extracellular HMGB1 can directly impact the progression of GC cells through the induction of IL-8, which promotes tumor angiogenesis via increasing vascular endothelial growth factor (VEGF) and TNF-α expression[110]. Increased HMGBI also works to favor tumor invasiveness and promote EMT through RAGE-dependent activation of ERK 1/2 and NF-κB[111]. Recent experiments have also confirmed that HMGB1 activation of the RAGE axis can strongly regulate direct EMT mediators in GC cells via downregulation of E-cadherin, as well as by upregulation of transcription factors of Snail and Slug[111] and MMPs (e.g., MMP-9 and MMP-2), whereas the knockdown of HMGB1 showed opposite effects. These results implicate HMGB1 is a key mediator of RAGE-dependent EMT, acting to favor the expression of more invasive GC cell phenotypes[77]. To this end, compelling data have revealed that RAGE knockdown in GC cells suppresses cell growth, invasion, and metastasis through the downregulation of RAGE-related pro-tumoral mediators of EMT (e.g., Akt, PCNA, and MMP-2)[112]. Additionally, the knockdown of the HMGB1 gene inhibits cell proliferation and the invasive potential of GC cells via suppression of RAGE-dependent NF-κB activation[113].

The S100 calcium-binding protein family member S100A8/A9 is a proinflammatory heterodimer produced by both myeloid-derived suppressor cells (MDSCs) and tumor cells[114]. S100A8 and S100A9 form a heterodimeric complex to interact with RAGE, promoting pro-tumoral signaling pathways within cancer cells[115]. Of note, serial analyses of gene expression reported significant overexpression of the genes encoding S100A8 and S100A9 in GC cells when compared to normal gastric epithelial cells[116]. In addition, in vitro evidence has revealed that these calcium-binding proteins are critical in promoting the invasive and migratory phenotype in GC cells through the engagement of RAGE and its pro-tumoral signaling pathways (e.g., p38, MAPK, and NF-ҡB). These RAGE-activated pathways further mediate S100A8/A9-induced migration and invasion in GC[116,117]. For example, Yong and Moon[118] studied siRNA blockade of S100A8/A9 in a human GC cell line, and found a drastic decrease in the invasive ability of malignant cells. This finding suggests the critical role of S100A8 and S100A9 and their heterodimer in promoting progression and invasion in GC. Novel evidence has also shown a significant increase in MDSC-induced S100A8/A9 plasma levels in GC patients, which is thought to induce immunosuppressive effects through RAGE-dependent T-cell proliferation and interferon-γ production, which directly correlate with advanced cancer stage and reduced survival[119].

Interestingly, the emerging role of calprotectin as a modulator of apoptosis has been recently reported. Calprotectin has been shown to be involved in the regulation of apoptotic mediators such as Bax, Bcl-2 family proteins, and ERK signaling pathway proteins through RAGE engagement in GC cell lines[120].

The dissemination of malignant cells from their primary site and the subsequent development of secondary tumors in distant organs, a process called metastasis, is a main consequence of cancer therapy failure and a major cause of cancer-related mortality[1,14]. The complex process of metastasis is supported by different mechanisms that require primary tumor cells to leave their origin, circulate within the microvasculature, escape immune surveillance, and colonize distant niches to promote the formation of secondary tumors[121].

A pivotal step in the development of secondary metastasis is the establishment of a complex interplay of tumor-derived factors, a favorable microenvironment in the distant site (i.e. a “pre-metastatic niche”). The pre-metastatic niche favors tumor cell colonization and promotes micro- and macrometastasis. Processes that contribute to the cultivation of the niche include inflammation, immunosuppression, neoangiogenesis, lymphangiogenesis, and organotropism[122].

Strikingly, the niche-promoting molecules described in various malignancies, including GC, are closely regulated by RAGE axis activation[71]. Examples of niche-promoting factors include S100A8/A9 ligand[39,123], VEGF-A, TNF-α, TGF-β, MMPs[43,46,124], extracellular vesicles, exosomes, and various microRNAs[5,7]; these mediators sculpt the cellular crosstalk that allows the initiation of metastatic secondary tumor formation[125,126]. In GC, it has been documented that patients with advanced stage and distant dissemination have a poor prognosis with a 5-year survival of approximately 20% depending on the population analyzed, tumor histological features, and the antineoplastic therapy used[1-3].

Experimental results suggest that RAGE overexpression is closely associated with the disruption of cell-to-cell adhesion and acceleration of GC cell motility, both of which support the metastatic activity of GC cells[93]. To this end, RAGE knockdown reduces GC cell proliferation and invasion and decreases the expression of pro-tumoral mediators (e.g., Akt, PCNA, and MMP-2) consequently inducing aberrant cell apoptosis and cell cycle arrest[112]. For example, in vivo and in vitro evidence has shown that AGE binding to RAGE can accelerate gastric tumor progression and dissemination through the dose-dependent upregulation of the pro-tumoral mediators Specificity Protein 1 and MMP-2 via activation of ERK signaling[126,127].

In addition to its actions discussed above, HMGB1 also facilitates the progression and metastasis of various types of cancer[128], including GC[105,107]. In recent years, a growing body of evidence has begun to unravel the molecular mechanisms behind this contribution HMGB1 to CG metastasis[106]. For example, several studies have revealed increased expression of RAGE and HMGB1 in gastric tumor samples, which correlates with local and distant advanced disease and poor overall prognosis[102,107,109,110,113]. This association is accentuated under hyperglycemic conditions such as in diabetes mellitus, which enhance endogenous AGE formation and RAGE signaling[105]. Furthermore, other experimental research has shown that the overexpression of HMGB1 and the subsequent induction of IL-8 in the gastric tumor microenvironment (TME) may contribute to EMT and GC micrometastasis through the activation of downstream pathways of the RAGE axis, including ERK 1/2 and NF-ҡB[110]. In addition, overexpression of HMGB1 promotes the recruitment of neutrophils to the TME; this inflammation and associated production of ROS further promotes GC proliferation and metastasis[128,129].

Another recent investigation has demonstrated that the HMGB1/RAGE axis can regulate GC cell proliferation and migration via the enhanced expression of cyclins and MMPs, the induction of EMT, and activation of several cellular proliferation pathways including the Akt/mTOR and ERK/CREB signaling pathways[77].

Moreover, the in vivo or in vitro knockdown of HMGB1 leads to downregulation of NF-ҡB, PCNA, and MMP-9 expression in gastric tumor cells. This finding suggests that targeting the HMGB1 pathway may be a promising approach to inhibit the growth and metastasis of GC cells through the regulation of RAGE-dependent pathways[113]. Also, the overexpression of HMGB1 has been shown to enhance gastric tumor cell secretion of IL-8, supporting EMT during the early stages of tumor progression. Conversely, reduction in HMGB1/RAGE signaling by in vivo targeting of extracellular HMGB1 produces a significant reduction in tumor growth[111].

The overexpression of calgranulins in invasive GC cell lines has also been demonstrated[116]. These recognized RAGE ligands can promote not only more invasive phenotypes of the malignant cells, but can also enhance their migration capacity via RAGE-dependent mechanisms (e.g., p38 MAPK-dependent NF-ҡB activation and MMP2 upregulation)[117].

The expression level of RAGE has been associated with invasion, metastasis, and poor prognosis in GC[93,94]. RAGE levels also correlate with the severity of gastric mucosal lesions in patients infected with H. pylori, where the highest RAGE expression was observed in precancerous lesions (e.g., atrophy or IM), dysplastic lesions, or in situ adenocarcinoma[130].

Compelling data have identified diverse polymorphisms within exons, introns, and regulatory regions of the gene encoding RAGE[131]. Interestingly, some of these polymorphisms may affect either RAGE gene transcriptional activity or the binding affinity of RAGE for its ligands[55,132]. Thus, RAGE polymorphisms are now potentially considered as susceptibility or poor prognostic factors in different cancer types. In this regard, several research groups have investigated the association between RAGE gene polymorphisms and the risk of various cancers[133-137]. For example, studies have demonstrated the association of certain RAGE polymorphisms and GC. The pioneering work of Wang et al[136] demonstrated that the rs2070600 RAGE polymorphism [Gly82Ser] confers an increased risk of GC in the Chinese population[138]. Of note, this polymorphism is associated with enhanced RAGE signaling[132].

More recently, Liu et al[125] further confirmed that the rs2070600 variant AG genotype plays a predominant role in the development of GC. On the contrary, the rs184003 GT genotype represents a significantly reduced risk for GC. Additionally, the rs2070600 and rs184003 polymorphisms affect the serum levels of the splice variant of sRAGE, where the first is associated with a decreased level of while the latter with an increased serum level of sRAGE[139].

In addition to the protease-mediated cleavage soluble RAGE variants, sRAGE can arise from alternative splicing of the RAGE gene. These soluble variants lack the transmembrane and cytoplasmic domains, allowing them to function as decoy receptors, that do not lead to the activation of RAGE-associated signaling[46]. Reduced sRAGE has been associated with increased cancer risk and tumor progression in many other cancer types[140,141].

In summary, a diverse and compelling body of data shows the wide range of mechanistic contributions of the RAGE axis to gastric carcinogenesis (Table 1).