Colorectal cancer (CRC), which includes cancers of the colon and rectum, ranks as the third most prevalent cancer worldwide, constituting 10% of all new cancer cases, and is the second leading cause of cancer-related mortality, resulting in >930,000 deaths annually (1,2). Currently, surgical intervention, chemotherapy and targeted therapy constitute the cornerstone of CRC treatment. However, at the time of diagnosis, approximately one-quarter of patients with CRC are diagnosed with advanced disease, and an additional 20% develop distant metastases, frequently rendering surgical methods insufficient for improving prognosis and ineffective for identifying cancer that has propagated to adjacent organs (3). Thus, improving the prognosis of patients with locally advanced rectal cancer and metastatic CRC (mCRC) remains a pivotal and formidable challenge. Survival rates have demonstrated that 70-75% of patients with mCRC survive beyond 1 year, 30-35% survive >3 years and <20% survive >5 years (4,5). Furthermore, after neoadjuvant chemoradiation therapy and total mesorectal excision surgery, 54% of patients with rectal cancer relapse. Adjuvant chemotherapy or radiotherapy is required for ~66% of patients with stage II-III colon cancer and 50% of patients with stage II-III rectal cancer (6). The challenge of treating advanced CRC is compounded by resistance to chemotherapy, radiotherapy, targeted therapy and immunotherapy. Encouragingly, clinical evidence has demonstrated the effectiveness of neoantigens (cancer-specific abnormal peptides) in generating immune responses (7). Therefore, identifying novel components of the host immune system as relevant biomarkers and therapeutic targets is crucial. A previous study has emphasized the role of the tumour microenvironment (TME) in treatment resistance (8). Components of the TME, such as tumour-associated macrophages (TAMs), regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs), have been shown to dampen the immune response and influence CRC progression (9-11). As such, the potential roles of these cells in cancer metastasis and therapy have garnered significant interest.

MDSCs constitute diverse immature cell populations originating from the bone marrow, all of which exhibit immunosuppressive activity (12). As outlined by the dual signalling mode (13), three key processes occur within MDSCs, namely, migration, expansion and activation, each of which partially overlap and are facilitated by factors boosting myelopoiesis and hindering differentiation of mature myeloid cells and factors promoting the activation of MDSCs. Before these cells can become functional within the TME, these steps are imperative (14). Additionally, MDSCs display several biochemical characteristics crucial for suppressing immune responses, including enhanced signal transducer and activator of transcription 3 (STAT3) expression, the induction of endoplasmic reticulum (ER) stress (which can independently induce apoptosis and regulate ER chaperones, protecting cells by ensuring proper protein folding and preventing the accumulation of misfolded proteins) and the expression of arginase 1 (ARG-1) and S100 calcium binding protein A8/A9 (S100A8/A9) (15,16).

In healthy individuals, immature myeloid cells (IMCs) differentiate into granulocytes, macrophages or dendritic cells (DCs), eventually migrating to specific organs and tissues to perform regular immune functions (17). However, under pathological conditions or chronic inflammation, IMCs deviate from their typical differentiation pathway (18), evolving into granulocyte-monocyte progenitor cells under the influence of granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte CSF (G-CSF) and interleukin (IL)-6 (19). These cells then gradually migrate from the bone marrow into the peripheral blood and spleen where they are activated by proinflammatory cytokines such as interferon-γ (IFN-γ), IL-1β and IL-4, consequently forming relatively immature cells termed MDSCs, which exhibit the morphology of neutrophils and monocytes and possess immunosuppressive functions (20-22).

MDSCs are typically divided into two subsets: Granulocytic or polymorphonuclear-MDSCs (PMN-MDSCs) and monocytic-MDSCs (M-MDSCs) (12,23,24), with considerably more PMN-MDSCs than M-MDSCs noted in most tumours (18,25). MDSCs have an important role in the malignant progression of tumours (26,27), and the presence of MDSCs has been shown to obstruct the infiltration of CD8⁺ T cells into tumours, promote Tregs and modulate natural killer (NK) cell activity, among other functions (28,29). As such, the suppressive actions of MDSCs contribute to tumour growth. Furthermore, the concentration of MDSCs in the peripheral blood is positively associated with the cancer stage; specifically, elevated levels indicate a more advanced cancer stage (30). MDSCs also promote tumour angiogenesis and create premetastatic niches by suppressing immune cells (26,27). This dual activity encourages the spread of cancer cells while concurrently diminishing the efficacy of antitumour drugs and fostering drug resistance, consequently influencing the inception, progression and prognosis of tumours (31-33). Research has demonstrated that distinct MDSC populations, which are involved in CRC (34) metastasis and progression through multiple mechanisms, are detected in tumour tissues and peripheral blood of patients with CRC (35).

The present review describes the mechanisms underlying MDSC aggregation and migration and illustrates the role of MDSCs in the CRC TME. The present review also surveys current therapeutic strategies, providing insights that may benefit the diagnosis, prognosis and treatment of CRC.

The phenomenon of extramedullary haematopoiesis and neutrophilia frequently accompanies malignancy (such as CRC), and these manifestations include cells with suppressive activity that were initially termed IMCs (17). With advances in research clarifying their origin and function, these cells were ultimately designated MDSCs (36) (Fig. 1).

Initially, to distinguish immature MDSCs from normal cells, researchers detected MDSCs in mice using surface biomarkers, with the goal of gaining insight into human MDSCs. It was found that MDSCs in mice are characterized by the markers, CD11b and Gr-1, and that the Gr-1 antigen complex comprises the components Ly-6G and Ly-6C. Specifically, mouse PMN-MDSCs are CD11b⁺Gr-1⁺Ly6G⁺Ly6C^low and are phenotypically and morphologically similar to neutrophils. M-MDSCs are CD11b⁺Gr-1⁺Ly6G⁻Ly6C^high and are phenotypically and morphologically similar to monocytes (37-39). However, a subsequent study has shown that the two characteristic markers of mouse MDSCs, CD11b and Gr-1, are not optimal markers for distinguishing different species of MDSCs in humans (40).

Researchers have identified key surface markers, such as human leukocyte antigen DR (HLA-DR), CD11b and CD14, which are intrinsic to monocytes, as pivotal in distinguishing human MDSCs (41). Another signature of M-MDSCs is the positive expression of C-C motif chemokine receptor 2 (CCR2) and C-X3-C motif chemokine receptor 1 (CX3CR1) (42,43). A subset of cells bearing the phenotype of HLA-DR^low/−CD11b⁺CD14⁻CD15⁺ has been categorized as granulocytic-MDSCs (12,23). Given their morphological and phenotypic resemblance to neutrophils, they are also aptly referred to as PMN-MDSCs. Conversely, cells characterized as HLA-DR^low/−CD11b⁺CD14⁺CD15⁻ are termed M-MDSCs, given their resemblance to monocytes (12).

The aforementioned subtypes also exhibit distinct T-cell suppressive activities and mechanisms. Additionally, human M-MDSCs express more CD33 than PMN-MDSCs (44). The distinction among human monocytes, neutrophils and MDSCs is achieved through the expression of surface molecules such as HLA-DR and lectin-type oxidized LDL receptor 1 or CD14, which serve as differentiating markers (45-48). Previous investigations have identified early-stage MDSCs (12,49) and fibrocytic MDSCs (50-52), with the former capable of colony formation and distinguished by a lineage⁻ (Lin⁻) (including CD3, CD14, CD15, CD19 and CD56) HLA-DR CD33⁺ phenotype. The latter, a subpopulation with fibrocytic characteristics and immunosuppressive functions, differs from umbilical cord blood precursors and can be identified by a CD11b^lowCD11c^lowCD33 IL-4Ra⁺ phenotype. Furthermore, CD49d is a potential marker that complements CD11b for MDSC categorization (53). A study by Alshetaiwi et al (54) revealed that high concentrations of MDSCs in the spleen and primary tumour sites correspond to significant expression of both CD84 and junction adhesion molecule-like, suggesting that CD84, which is typically a lymphoid marker, may represent a novel and specific marker for MDSCs in CRC.

Chemotaxis and the accumulation of MDSCs in the colorectum are critical for colorectal tumorigenesis. Furthermore, gene sequence and epigenetic variations are significant risk factors for MDSC accumulation (33). In CRC, mutations in the oncogene, Kirsten rat sarcoma viral oncogene homologue gene (KRAS), are observed in up to 40% of cases and are associated with increased tumour aggressiveness and metastasis (55-57). Previous studies have highlighted the essential role of the SLC25A22 gene in enhancing KRAS mutation-induced CRC immunosuppression by facilitating asparagine binding and SRC phosphorylation activation (58,59). This process promotes ERK/ETS proto-oncogene 2 signalling and drives C-X-C motif chemokine ligand 1 (CXCL1) transcription. The secreted CXCL1 then acts as a chemoattractant for MDSCs through the chemokine C-X-C motif receptor 2 (CXCR2), creating an immunosuppressive microenvironment (60). Furthermore, MDSCs suppress interferon regulatory factor 2 expression, resulting in increased CXCL3 binding to CXCR2 on MDSCs and promoting MDSC migration into the TME, leading to MDSC overaccumulation (58,61). YTH N6-methyladenosine RNA binding protein 1 (YTHDF1) regulates CXCL1 expression via promoting p65 protein expression to activate NF-κB signal transduction, impacting MDSC migration through the CXCL1/CXCR2 axis; furthermore, a reduction in MDSC number is observed upon YTHDF1 gene knockdown (62).

N6-methyladenosine (m6A) RNA methylation, a posttranscriptional RNA modification, involves m6A methylase complexes (writers), demethylases (erasers) and binding proteins (readers) that regulate their respective axes, promoting normal biological processes. However, aberrant m6A RNA methylation promotes MDSC recruitment at the gene expression level, thereby increasing CRC risk. For instance, the methyltransferase 3 methylase N6-adenosine-methyltransferase complex catalytic subunit promotes m6A methylation of the transcription factor, basic helix-loop-helix family member e41, which stimulates CXCL1 expression and mediates MDSC migration and aggregation in the TME of CRC, forming an immunosuppressive environment and weakening the immune system, promoting CRC development (63,64). Furthermore, the demethylase, Alk B homologue 5 (Alkbh5), recruits MDSCs and diminishes NK and CD8⁺ T cell populations by decreasing the mRNA stability and demethylation of axis inhibition protein 2 mRNA, leading to its dissociation from the m6A reader, insulin like growth factor 2 mRNA binding protein 1 (65), thereby resulting in Wnt/β-catenin pathway overactivation. Human Alkbh5 also contributes to gene splicing by placing m6A modifications near splice sites, influencing MDSC migration and aggregation in the TME (66-68). In addition, Alkbh5 affects changes in metabolite content. Alkbh5 modulates monocarboxylate transporter 4 expression (a key enzyme that rapidly transports lactic acid to plasma membranes) and lactate concentration in the TME (69). Additionally, lactate can increase the frequency of MDSCs and upregulate expression of FoxP3, an important transcription factor in Treg development and function (68-70).

The pathogenesis, progression and metastasis of CRC are intricately linked to genetic factors within the host as well as the specific tissue milieu in which the cancer resides. This local milieu, consisting of immune cells, malignant cells, fibroblasts, various non-immune cells and connective tissues, constitutes the TME (71). In contrast to being a static entity, the TME is dynamic and complex. Within this setting, common myeloid progenitor cells (CMPs) located in the bone marrow can differentiate into MDSCs during tumour progression. In response to inflammatory signals or tumour-released factors, including growth factors, chemokines and inflammatory cytokines, CMPs proliferate, promoting the recruitment and expansion of MDSCs to the tumour locus (72). This process not only facilitates the influx of IMCs but also inhibits the innate antitumour response of the host.

Since their initial discovery in 1986 (73), over 50 human chemokines, small secreted proteins that direct immune cell movement, have been identified (74). These chemokines are divided into four subfamilies, including CXC, CC, C and CX3C, based on variations in the sequence of the first two conserved cysteines at the N-terminus (75). Chemokines expressed by tumour cells play a pivotal role in recruiting MDSCs into the TME by binding to specific receptors on the MDSC surface. In CRC, MDSC recruitment is predominantly associated with two chemokines, one from the CC subfamily (CCL2, CCL5 and CSF1) and the other from the CXC subfamily (CXCL1/2/3/5/6/8/12).

CCL2, also known as monocyte chemotactic protein-1, is a vital component of the CC chemokine family. This protein significantly impacts the growth, progression and metastasis of various tumours, including CRC (76-78). CCL2-mediated recruitment of the CCR2 receptor, which has a high affinity for MDSCs, supports CRC cell proliferation (79). Evidence suggests that MDSC accumulation induced by CCL2 enhances their immunosuppressive capabilities during colorectal carcinogenesis, with a corresponding increase in CCL2 levels as the cancer progresses (76,80). Moreover, reactive nitrogen species produced by MDSCs can modify chemokines such as CCL2 to nitro-CCL2, which preferentially recruits myeloid cells, in contrast to unmodified CCL2, which attracts CD8⁺ T cells (81).

Conversely, MDSCs positive for CXCR2 can engage with CXCL1/2/3, facilitating the recruitment of MDSC clusters in the colon and rectum (61,82). The migration and recruitment orchestrated by members of the CC family, including CCL2, CCL5 and CSF1, predominantly involve M-MDSCs and monocytes. By contrast, CXC family members, such as CXCL1, CXCL5, CXCL6, CXCL8 and CXCL12, are more closely involved with the recruitment of PMN-MDSCs and granulocytes in CRC. MDSCs express various chemokine receptors, including CCR2, CCR5 and CXCR2 (21,47). Additionally, CXCL1 and CXCL2 are crucial in the context of colitis-associated tumours and chronic colonic inflammation (83).

The regulation of MDSCs extends beyond that of chemokines to include other inflammatory mediators, such as prostaglandins and histamine. For instance, histamine enhances the proliferation of M-MDSCs and increases the expression of enzymes such as inducible nitric oxide synthase (iNOS) and ARG-1 in the CRC (84). Conversely, histamine suppresses the expression of these enzymes in PMN-MDSCs while promoting the production of the interleukins, IL-13 and IL-4 (85). Hence, histamine differentially modulates M-MDSC and PMN-MDSC activities (86).

Additionally, cyclooxygenase-2 (COX-2)-synthesized prostaglandin E2 (PGE2), an inflammatory mediator, is upregulated in CRC (87). PGE2 not only induces myeloid cells to secrete procarcinogenic factors, such as IL-6, CXCL1 and G-CSF, that create a tumour-friendly microenvironment, but also promotes tumour proliferation and MDSC activation via STAT3 phosphorylation (87-91). The PGE2 receptor EP1-4, a set of related G protein-coupled receptors, influences MDSC differentiation in CRC (92). In particular, interactions with EP4 promote the differentiation of immunosuppressive M2 macrophages and MDSCs while diminishing the proliferation of immunostimulatory M1 macrophages (93).

Leukotriene B4, a 5-lipoxygenase (5LO) derivative of arachidonic acid, has a chemotactic impact on MDSCs, affecting their aggregation. Reduced 5LO expression is associated with decreased MDSC levels in circulation and diminished ARG-1 and iNOS activity (94,95).

Hypoxia is a common condition within the CRC TME and plays a crucial role in promoting MDSC proliferation. Hypoxia-inducible factor 1α (HIF-1α) stimulates the production of ectonucleoside triphosphate diphosphohydrolase 2 (ENTPD2), and ENTPD2 functions as an exonuclease on MDSCs, significantly contributing to their expansion within the CRC TME (17). Furthermore, hypoxia elevates the levels of vascular endothelial growth factor (VEGF) and its associated molecules. In addition to their known role in angiogenesis, these factors also promote MDSC migration from the bone marrow to the tumour bed (the original tumour site prior to any treatment) thereby enhancing MDSC proliferation in the CRC TME (96,97). Additionally, extracellular vesicles such as exosomes, which are secreted by tumour cells, directly stimulate the formation of MDSCs while modulating their activity, a phenomenon observed in various tumour types, including CRC (98,99).

The presence of gut microorganisms is important for MDSC amplification and normal metabolism in the gastrointestinal tract. A recent study has shown that the accumulation of anaerobic Pseudomonas aeruginosa (Peptostreptococcus anaerobius) in tumour lesions can mediate MDSC recruitment into the CRC microenvironment and promote IL-23 secretion by MDSCs, which leads to epithelial-mesenchymal transition (EMT) and CRC cell resistance to chemotherapy (100). Fusobacterium nucleatum plays a role in the early development of inflammatory bowel disease and colorectal adenomas (101). Patients with CRC with higher levels of F. nucleatum in the TME have elevated numbers of both M-MDSCs and PMN-MDSCs and suppressed NK cell numbers, conditions that promote CRC development (102-104).

The immunosuppressive functions of activated MDSCs are mediated by intercellular contacts and paracrine or exosomal mechanisms. These functions inhibit immune cell activity, including the degradation of L-arginine, the production of reactive oxygen and nitrogen species (RONS), engagement of the programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) axis, the production of immunosuppressive cytokines such as IL-10 and transforming growth factor-β (TGF-β), the suppression of T cells and the induction of other immunosuppressive cells (105-107) (Fig. 2). Among the subtypes of MDSCs, M-MDSCs are characterized by elevated secretion of TGF-β and the inhibitory cytokines, IL-10, ARG-1 and iNOS, which collectively suppress non-specific immune responses. By contrast, PMN-MDSCs predominantly produce high levels of reactive oxygen species (ROS) and PGE2, and influence antigen-specific immune responses through direct cell-to-cell interactions (47).

The inhibitory effects of MDSCs on T-cell responses and multiple cellular functions are critical for antitumour immune responses. MDSCs are positively correlated with the development and spread of cancer, and their presence typically decreases the effectiveness of immunotherapy (151). Given that the accumulation of MDSCs in the TME is considered to be a major obstacle to tumour immunotherapy (152), targeting MDSCs has become a new strategy for tumour immunotherapy. Current clinical studies have focused on four directions (153,154): Depletion of MDSCs, inhibition of MDSC recruitment, suppression of MDSC immunosuppressive function, induction of MDSC differentiation (Fig. 3).

The establishment of an immunosuppressive TME is one of the key mechanisms by which these CRC cells evade the immune system. In recent years, MDSCs have been shown to play a crucial role as major contributors to this process. Studies targeting MDSCs are mainly focused on the multiple major directions of MDSC expansion, namely migration, activation, recruitment, action and metastasis (131). Based on the studies published thus far, targeting MDSCs has been shown to reduce the tumour load and represent a promising therapeutic strategy for CRC (153). In the present review, the clinicaltrials. gov website was searched and the therapeutic approaches involving MDSCs and different treatment regimens for CRC are selectively summarized in Table I.

Although studies targeting MDSCs have reported notable results, the existence of potential limitations constrains the continued advancement of strategies targeting MDSCs. First, MDSCs have a short half-life (200-202), and the strategy of MDSC depletion in the peripheral blood negatively feeds back to the bone marrow (203,204), which increases the risk of MDSC recurrence and accumulation and promotes CRC development. More effective therapies may aim to block MDSC differentiation in the bone marrow, inhibit MDSC migration to affected tissues or manipulate the tissue microenvironment. Second, current targeted therapies for MDSCs are focused on reducing the number of MDSCs that have been generated or on inhibiting their function, with fewer strategies targeting MDSC-induced migration (33). Third, due to the large phenotypic heterogeneity of MDSCs, complex amplification and functional networks, different subgroups of MDSCs use different mechanisms to suppress the immune system, making current therapeutic strategies targeting MDSCs only partially effective. Therefore, further investigations into the molecular phenotypes of MDSCs and their mechanisms in tumour tissues are urgently needed, and the development of therapeutic strategies targeting MDSC subgroups is essential for improving the effectiveness of tumour therapy (12). In addition, based on the summary provided in the present review, cancer cells and MDSCs in primary tumours can directly or indirectly promote metastatic premetastatic niche formation, but the mechanistic similarities and differences in the interactions between the TME and the premetastatic niche remain topics worth exploring (143). Moreover, the TME is very complex and is composed of multiple immune cells and cytokines that form a complex network. It is therefore difficult to achieve the expected effect by exclusively targeting MDSCs, and a more desirable therapeutic effect may be achieved by co-targeting other cells or targets to enhance the immune system or utilizing special contact modalities such as exosomes (33).

Iron-mediated death is a unique form of cell death driven by the accumulation of lipid peroxides. A study on iron death induction and gastric cancer reported that inhibition of the Wnt/β-catenin signalling pathway reduced the expression of the transcription factor, TCF4, inhibited glutathione peroxidase 4 gene transcription and promoted iron death, resulting in a decrease in both tumour weight and volume (205), suggesting the feasibility of the induction of iron-mediated death for the treatment of gastrointestinal tumours. The triple combination of MDSC blockade with iron-mediated death induction and ICB therapy greatly restored tissue immune surveillance and promoted a normal immune response. This treatment inhibited tumour growth and represented a promising strategy for targeted therapy of CRC (206).

Exosomes in the TME act as carriers of molecular delivery exerting indirect effects on MDSC expansion, formation of an immunosuppressive environment (48), cancer cell growth and invasion (207,208) and therapy resistance (209), and direct promotion of CRC cell stemness through S100A9 (210). Under IL-6 induction, PMN-MDSCs synthesize increased exosomal miR-93-5p, which promotes the differentiation of M-MDSC into M2 macrophages and increases CRC risk (211). Compared with healthy individuals, patients with CRC have significantly increased exosomal miR-19a levels in the serum and have a poorer prognosis (212,213). This suggests that exosomes in the TME have the potential to be another potential target in treating CRC.

Therefore, from the perspective of patients with CRC, the strategy of targeting MDSCs requires more comprehensive and refined research and clinical trials, but it remains a promising class of therapeutic strategy due to its special immunosuppressive effects in the TME.

In the present review, the nomenclature history of MDSCs along with the two main subtypes, PMN-MDSCs and M-MDSCs, were systematically and comprehensively reviewed and the recruitment, role and metastatic mechanisms of MDSCs in the CRC microenvironment were revealed. Mutations in genes such as KRAS, adenomatous polyposis coli, BRAF and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α (214,215), epigenetic modifications of m6A, chemokines in the CC and CXC families, inflammatory mediators, alterations in intestinal microbes and hypoxic environments may cause local aggregation of MDSCs and increase the risk of CRC.

The present review proposes that MDSCs possess two main features: Immunosuppression and metastasis. MDSCs achieve immunosuppressive functions by generating RONS, degrading L-arginine, secreting immunosuppressive cytokines via intercellular contact, paracrine or exocrine mechanisms (216), inhibiting T cell and NK cell toxicity and inducing other immunosuppressive cells, such as Tregs, to achieve immunosuppressive functions and provide favourable conditions for cancer cell survival and growth (217). MDSCs are recruited and expanded to support immunosuppression. MDSCs inhibit normal T cell proliferation in patients with CRC, resulting in poor prognosis. The recruitment of expanded MDSCs enhances cancer stem cell gene expression, directly promotes metastatic cancer cell survival, antagonizes cancer cell senescence and acts as a crossroads between tumour angiogenesis and immunosuppression to provide precursor support for CRC metastasis; therefore, the presence of MDSCs is important for CRC cells to colonize, survive and grow in target organs. Increased MDSC and tumour miRNA101 expression also predicts poor survival.

In addition, the present review comprehensively summarized four types of potential therapeutic strategies for CRC: Depletion of produced MDSCs, inhibition of MDSC recruitment, suppression of MDSC immunosuppression and induction of MDSC differentiation. The primary TME can promote premetastatic niche formation, but the mechanism underlying the interaction between the primary TME and premetastatic niche formation during the process of metastasis initiation requires further investigated.