Oral squamous cell carcinoma (OSCC) is a common malignant neoplasm (80%–90%) of the oral cavity, derived from the head and neck region of the body. It is associated with the common risk factors of smoking and alcohol consumption (1, 2). Contributing the highest incidence and mortality rate in both males and females, there were 354,864 new cases and 177,384 deaths worldwide in 2018 and was the leading cause of mortality in Central Asia (3). Whilst there is an improvement in advancing therapies such as surgery and chemotherapy, the 5-year survival rate remains 50% in various countries over the past 3 decades (4). This insufficient improvement in prognosis could be explained by the lack of consideration of immunological parameters in prognostic classification and treatment of OSCC (5). Moreover, poor prognosis in OSCC may be a result of its aggressive local invasion and metastasis, leading to an uncontrollable recurrence (6).

Metastasis is achieved through the interaction of tumour cells and the surrounding tumour microenvironment (TME) (7). This TME plays a critical role in tumourigenesis, tumour progression, invasion and tumour tissue infiltration by tumour-associated macrophages (TAMs) (8). TAMs are abundant in both the tumour and tumour stroma, playing a significant role in cancer progression (9, 10). MCP-1 (Monocyte chemoattractant protein-1 or CCL2) plays a role in recruiting and attracting TAMs to tumour sites (11, 12). These TAMs may exhibit either of two functional phenotypes, M1 or M2, dependent on cytokine, chemokine, chemokine receptor and other regulator expression (13). M1 TAMs exhibit pro-inflammatory and anti-tumoural properties, mediated by IL-12, TNF-α, IFN-γ, and stimulate strong Th1 IFNγ-driven cell mediated responses resulting in tumouricidal function. M2 TAMs are generally anti-inflammatory and pro-tumoural, expressing IL-10, IL-13, MR (mannose receptor) and are capable of inducing humoral Th2-driven cytokine responses, secreting IL-4, IL-13 and high levels of chemokines and growth factors such as VEGF, TGF-β, FGF and uPA, promoting angiogenesis, immunosuppression, tumour invasion and metastasis (14, 15).

TAM polarisation to distinct M1 and M2 subsets however, remains unclear, as recent evidence suggests functional plasticity and the ability to repolarise from one phenotype to the other (16). Human ovarian cancer TAMs have been observed to repolarise from M2 to M1-like phenotype suppressing levels of CCL18, MMP9 and VEGF when exposed to IFN-γ (17). TME TAMs have been shown to favour tumourigenesis, tumour survival and angiogenesis (18). This role in the TME however, is controversial; in colorectal cancer, for example, TAMs exhibit pro-inflammatory anti-tumour effects, leading to a favourable prognosis (19, 20). This may be explained by these M1 TAMs inducing the secretion of galectin-3 in human colon cells which further induces TAM infiltration and release of the pro-inflammatory cytokines, IL-1β and TNF-α, causing a strong anti-tumour response (20). Nevertheless, whilst TAMs may exhibit either phenotype, studies recognise TAMs to be predominantly of the M2 phenotype and correlate with a poor prognosis (15, 21).

TAMs may thus serve as potential biomarkers for the prognosis and therapeutic targeting of several cancers, particularly OSCC. Interestingly, over 80% of studies reveal a high number of TAMs correlates with poor patient prognosis (21). Investigations have shown CD163 (M2 macrophage class B scavenger receptor) (22), as a biomarker for macrophage activation in lung, breast and hepatocellular carcinoma (23-25). Recently, overexpression of M2-like CD163⁺ TAMs in head and neck squamous cell carcinoma (HNSCC) patients, revealed a poor clinical prognosis in both overall survival (OS) and progression-free survival (PFS) (26). CD163 functionality involves eradicating and endocytosing the haemoglobin/haptoglobin complex, thereby protecting tissues from oxidative damage (27). Used as a biomarker for M2c deactivated macrophages, it presents both anti-inflammatory and pro-tumoural functions (28).

Monocyte/macrophages, specifically M2 macrophages, abundantly express CD68, a glycosylated type I transmembrane glycoprotein belonging to the LAMP (lysosomal-associated membrane proteins) family (29). Its primary function is poorly understood, but as a class D scavenger receptor, it plays a role in promoting phagocytosis, clearing cellular debris and mediates recruitment/activation of macrophages (30). In contrast, CD68 is considered to be a pan-macrophage marker expressed by both M1 and M2 subsets, derived from anti-CD14-purified peripheral blood monocytes (31). This may explain observations where overexpression of CD68⁺ TAMs was associated with poor overall survival and disease-free survival (DFS) in breast cancer patients (32), whereas conversely, high CD68⁺ TAM expression conferred a longer overall survival and disease-free survival in hepatocellular carcinoma patients (33). Therefore, its prognostic relevance to OSCC needs clarification.

PD-L1 (programmed death ligand-1 also known as CD274, B7-H1) is a cell surface type I glycoprotein expressed on antigen-presenting cells (APCs) and located in dendritic cells and macrophages. Belonging to the B7 family, it is a co-inhibitory ligand which binds PD-1 (programmed death receptor-1). PD-1 functions as a T-cell checkpoint protein, regulating T-cell suppression (34). PD-1 is a member of CTLA-4 (cytotoxic T lymphocyte-associated protein 4) family, primarily expressed by cytotoxic T cells (Tc), which predominate anti-tumour responses. PD-1 ligation suppresses T-cell function via an inhibitory signal involving SHP-2 which inhibits CD28-mediated PI3K and Akt activity (35). In various cancers, PD-1 can be expressed on tumour infiltrating lymphocytes (TILs), where CD4⁺ and CD8⁺ TILs exhibit an increased PD-1 expression on Treg and Tc, effectively resulting in Treg-mediated immunosuppression and Tc anergy/loss of CTL function (36). In order to maintain homeostasis, PD-1/PD-L1 induces immune tolerance and effectively suppresses excessive tissue inflammation and autoimmune disease. In tumours, however, binding of PD-L1 to its PD-1 receptor on activated T cells results in T-cell suppression and immune escape by inhibiting perforin/granzyme production, suppressing IL-2 and IFN-γ production and promoting apoptosis, effectively inducing tumour growth (37). Several studies investigated the relationship between TAM PD-L1 expression and cancer patient prognosis. High PD-L1 expression revealed a poor clinical prognosis in malignant pleural mesothelioma (MPM) and renal cell carcinoma patients (38, 39), whereas other studies reached controversial and inconsistent conclusions. Conversely, high PD-L1 expression was associated with longer overall and disease-free survival (40, 41). Therefore, its prognostic relevance needs further clarification.

This study aims to investigate the prognostic role of CD68⁺, CD163⁺ TAMs and PD-L1 expression in OSCC, through a meta-analysis of the current literature. Furthermore, the prognostic role of these biomarkers was investigated in different sub-locations in OSCC (tumour versus stroma). This study hypothesised a high expression of CD68⁺, CD163⁺ TAMs and PD-L1 would lead to worse survival in OSCC patients, whereas concluded that CD163⁺ TAMs located in both tumour and stroma were predictive of a poor prognosis in OSCC and that PD-L1 may prove to be indicative of a positive outcome in these patients.

Studies that had met the following inclusion criteria were included in the meta-analysis: 1) English language publication. 2) Studies that reported the prognostic significance and role of CD163, CD68 and PD-L1 in OSCC 3) Studies analysing the protein level expression of CD163, CD68 and PD-L1 in clinical analysis such as immunohistochemistry (IHC) sections in OSCC. 4) Evaluate the association of CD163, CD68 and PD-L1 and patient prognosis according to the following parameters: overall survival (OS). 5) Provided sufficient survival data which included only hazard ratio (HR) with 95% confidence interval (Cl), p-value P) alongside Kaplan Meier survival graphs. Studies that had less than 30 patients and did not meet the parameters were excluded from the meta-analysis.

This meta-analysis has reviewed the current literature on the prognostic potential of tissue biopsy tumour-associated macrophages; CD163⁺ TAMs and CD68⁺ TAMs as well as PD-L1 expression in OSCC. This meta-analysis demonstrated a significant association between high CD163⁺ TAMs with poor survival/prognosis in OSCC patients. Additional results revealed an insignificant association of CD68⁺ TAMs with OS, whereas PD-L1 expression approaches significance, indicating a potential positive prognosis associated with OSCC patients. Therefore, it reveals CD163⁺ TAMs to exhibit the best prognostic potential of macrophage subsets in both intra-tumour and stromal OSCC biopsies.

The present pooled analysis revealed a high density of CD163⁺ TAMs was associated with worse overall survival in OSCC (HR = 2.64; 95% Cl: [1.65, 4.23]; p < 0.0001). These findings are consistent in several other studies, reporting the correlation of CD163⁺ macrophages and worse survival in breast, gastric, colorectal and hepatocellular cancers (24, 55–57). Also, pooled analysis found CD163⁺ located within the tumour stroma to be associated with poor survival in OSCC patients (HR = 3.56; 95% Cl: [2.33, 5.44]; p < 0.00001). These findings were consistent with high CD163⁺ TAM densities in the tumour stroma and associated with poor survival in SCCHN (squamous cell carcinoma of the head and neck) (26). Similarly, high levels of tumour stroma CD163⁺ TAMs were associated with lymph node metastasis in OSCC (58). These high levels of OSCC CD163⁺ TAMs could be explained by the ability of TAMs to directly stimulate EGF (epidermal-growth factor) as well as anti-inflammatory cytokines (such as IL-6, IL-10), pro-inflammatory cytokines such as TNF-α, and chemokines such as CXCL12, CCL16, CCL18. Collectively, these factors induce tumour cell growth and survival factors which enhance tumour cell proliferation, migration and metastasis (59, 60).

Interestingly, the expression of CD163 is not only restricted to TAMs but may also be associated with cell fusion where the fusion of cancer cells and TAMs can increase metastatic potential with migratory leukocytes in cancer patients and plays a role in cancer progression. This can lead to a more aggressive and metastatic phenotype causing an unfavourable prognosis as demonstrated in OSCC (61, 62). Therefore, it is imperative to discriminate CD163⁺ malignant cells and macrophages when examining the influence of CD163⁺ on prognosis. Nevertheless, many studies reveal consistent findings with the present results demonstrating CD163 may serve as a significant prognostic biomarker in OSCC. Interestingly, the results presented about CD163⁺ TAMs may provide clinical implications. More specifically, they may serve as therapeutic targets for anticancer therapeutic regimens which may include the repolarisation of TAMs from M2-like TAMs to an M1-like phenotype, to restrain tumour progression (63). Therefore, a greater understanding of TAM function and OSCC progression is critical for future research in TAM-targeted therapies.

Compared to CD163⁺ TAMs, pooled results demonstrated high CD68⁺ TAMs were not statistically significant in revealing poor overall survival in OSCC patients (HR = 1.26; 95% Cl: [0.76, 2.07]; p = 0.37). However, other meta-analysis, has revealed high CD68⁺ TAM densities were associated with worse overall survival and disease-free survival in nasopharyngeal carcinoma (NPC) patients (64). Similar studies revealed CD68⁺ TAMs were associated with poor survival in gastric cancer and hepatocellular carcinoma, respectively (65, 66). These findings may be indicative of CD68⁺ TAMs possessing immunosuppressive and pro-tumour responses, favouring cancer progression. Interestingly, CD68⁺ TAMs have been shown to suppress cytotoxic activity of CD8⁺ T-cells and increase tumour growth (67). On the other hand, in the case of oesophageal squamous cell carcinoma, CD68⁺ TAMs were correlated with a favourable prognosis (68), suggestive that CD68⁺ TAMs may also function as M1 macrophages, revealing pro-inflammatory and anti-tumour effects. Whilst these molecular mechanisms are not fully understood, studies reveal TAMs may exert tumoricidal activity in vitro; more specifically, to polarise into M1 TAMs orchestrated by the production of IFN-γ which also activates cytotoxic CD8⁺ T and NK cell responses to initiate tumour cell killing (12, 19).

Similar to findings in this study, CD68⁺ TAMs did not significantly correlate with overall survival and recurrence-free survival (RFS) in multivariate analysis in basal-like breast cancer (BLBC) and triple-negative cancer of the breast (24, 69). In addition, no prognostic utility was found between CD68⁺ TAMs and OS in SCCHN patients (26). These findings reveal CD68⁺ TAMs may serve as a poor prognostic biomarker as demonstrated in this study focussed on OSCC, but more importantly, may indicate CD68 as a pan-macrophage marker expressed by both M1-like and M2-like TAMs, capable of exhibiting opposing effects on the tumour microenvironment (70). Therefore, this study’s findings and the mounting conflicting evidence between different cancers, indicates that CD68⁺ TAMs may be a poor prognostic biomarker in OSCC or at least requires further investigation across a variety of cancers/tumours as well as their TMEs.

Contrary to expectation, pooled results also revealed high PD-L1 expression had a non-significant positive impact on overall survival in OSCC patients (HR = 0.64; 95% Cl: [0.35, 1.18]; p = 0.15). A high level of heterogeneity (I² = 70%) among the included studies were revealed, demonstrating conflicting results with each other. In contrast to these data however, numerous studies reveal PD-L1 expression was associated with poor prognosis and overall survival (OS) in solid cancers, such as head and neck squamous cell carcinoma (HNSCC) (71), breast cancer (72), non-small cell lung cancer (NSCLC) (73) and bladder cancer (74). This association with poor prognosis may be suggested by PD-L1/PD-1 binding to suppress CD8⁺ cytotoxic T-lymphocyte activation, leading to the evasion of the host immune anti-tumour response, thereby decreasing the survival rate in many cancers (37, 75). Additionally, whilst PD-L1 expression may protect macrophages from cell death, OSCC tumour cells induce TAM PD-L1 expression via IL-10 and induce T-cell apoptosis, further reinforcing an unfavourable prognosis (76).

Contradictory to these studies, yet consistent with findings in this investigation, PD-L1 expression in primary tumour cells was associated with prolonged DFS (Disease-free survival) in HNcSCC (head and neck cutaneous squamous cell carcinoma) (77). Similarly, high PD-L1 expression correlated better OS and DFS in breast cancer patients (78). This observation in prolonging survival in patients with PD-L1 expression may be due to the induction of an anti-tumour immune response. More specifically, IFN-γ, released by tumour-infiltrating lymphocytes as an adaptive immune-resistance mechanism to inhibit local effector T-cell function, can upregulate PD-L1 expression in tumour cells (36). Interestingly, IFN-γ also induces protein kinase D isoform 2 (PKD2), an important negative regulator of PD-L1 expression in OSCC. Thus PDK2 inhibits PD-L1 expression and promotes anti-tumour effects (blocking PD-1/PD-L1 dependent tumour antigen-specific CD8⁺ T cell apoptosis) (79). In addition, a high expression of PD-L1 was not statistically associated with OS in oesophageal squamous cell carcinoma (80), and more recently, pooled analysis of high PD-L1 expression did not have a statistically significant association with OS, DFS, DSS (Disease-specific survival) in OSCC patients (81). The findings of this investigation (Figure 5) are consistent with these studies focussed on OSCC and oesophageal SCC. In addition, further subgroup analysis suggested that stromal expression of PD-L1 may be associated with improved survival, whereas intra-tumour PD-L1 expression may be associated with poor prognosis and overall survival. This may be indicative of PD-L1⁺ cell location is predictive of survival and may reflect stage of cancer, or, when contrasted with the poor survival observed for CD163⁺ TAMs, is suggestive that stromal PD-L1 and CD163 are expressed on different TAM subsets or that PD-L1 may not be expressed on TAMs at all. Thus, this current investigation goes some way to indicating PD-L1 as a prognostic marker of survival, or indeed stage of cancer progression in OSCC which may reach statistical significance with the inclusion of more clinical studies. Further investigation may also potentially validate PD-1/PD-L1 interaction as a future therapeutic target for OSCC.

In conclusion, this meta-analysis confirmed the prognostic role of CD163⁺ TAMs, where a high cell number was associated with poor overall survival in OSCC. This indicates CD163⁺ TAMs may be a useful novel prognostic biomarker for OSCC and may suggest TAMs as a potential therapeutic target. Both CD68⁺ TAMs and PD-L1 revealed an insignificant correlation with overall survival in OSCC patients and limits the prognostic value of both biomarkers in OSCC, however the fact that the OS approached significance for PD-L1 is potentially indicative of PD-L1 being revealed as a positive prognostic indicator in the future.

• TAMs, and their location, are indeed, indicative of OSCC survival; where both tumour and stromal located CD163⁺ TAMs are indicative of poor prognosis whereas stromal PD-L1 expression may be indicative of a better prognosis when compared to tumour expressed PD-L1.