Cancer stem cells have been detected and isolated from multiple solid and hematopoietic malignancies [1, 2]. They express different patterns of markers in accordance with their phenotypes and the microenvironment they survive. However, the most common biomarkers in this population include membrane- bound antigens, such as CD44, CD24, CD133, CD34, and CD38 and cytoplasmic enzymes like ALDH1 [11]. Nevertheless, finding a highly specific and comprehensive biomarker for diagnosing CSCs remains unanswered and needs further researches.
According to literature, the CSCs pool in a tumor mass can be originated by a complex mixture of hierarchical and stochastic models that are not completely two separate entities and they may even complement each other in some stages during cancer progression. Based on these models, all tumor cells have the same potential to be a new cancer initiating cell because of accumulated mutations or favorable microenvironment which provides tumor- permissive cytokines and growth factors [6, 12, 13]. This dynamic evolution of tumor mass associated with multilineage differentiation potential of CSCs may result in real- time intratumoral heterogeneity which leads to various clinical consequences of responses to conventional chemoradiotherapy and immunotherapy ranging from resistance to complete cure [13, 14].
Major stem-related signaling pathways including Wnt, Hedgehog, JAK/STAT, Myc, Notch, PTEN/PI3K, and EGF/EGFR, WNT/β-Catenin, and NFκB are aberrantly expressed and deregulated in CSCs which promote cancer progression and figure out heterogeneity in clinical behavior of a cancer [2, 3, 5, 6, 10]. Intriguingly, three important signaling pathways, namely Notch, Wnt/β-catenin, and Hh have wide range of functions in both CSCs and immune cells, which makes them hot topics for many researches [8, 10]. Since they not only play multitude roles in differentiation and maturation of distinct immune cells, but also regulate self-replication and immune avoidance potential in CSCs [10, 11, 15]. In accordance with this point, many other overlaps have been mentioned in the context of regulatory markers and signaling profile. For instance, although CSCs can be easily identified based on cell surface biomarkers such as CD44 and CD24, designing target therapy against these markers impairs their functions on other normal cells like immune cells [10, 16–20]. Since CD44 as a well-known CSC marker in breast, colon and prostate cancers, plays critical roles in T helper1 survival and functions [10]. Another molecule released frequently by CSCs, IL-6, is associated with higher expression rate of stemness genes including Notch and Oct-4 in colorectal cancers and also contributes to self-renewal and proliferation of CSCs in head and neck squamous cell carcinoma. While its receptor called IL-6R control the development, maintenance, and function of memory Tcells and inhibit FOXP3+ Treg activity [16–20].
The other pathway like c-Myc may have been involved in CSCs immunoresistance mechanisms. It enhances expression rate of inhibitory molecules of innate immunity, CD47 and adaptive immunity, PD- L1 [1, 10, 21]. CD47 is the ligand for SIRPα receptors on phagocytic cells like macrophages and dendritic cells (DCs). Hence, in order to avoid phagocytosis, CSCs derived from multiple malignancies has increased its CD47 signaling. Preclinical works demonstrated that the CD47- SIRPα signaling prevention resulted in recovered phagocytosis of macrophages [22–25].
In uterine leiomyosarcoma, loss of PTEN is associated with decreased neoantigens and reduced tumor immunogenicity, which result in immune checkpoint blockade therapy resistance [26]. In glioblastoma CSCs, inhibition of STAT3 pathway leads to recovery of T cell mediated responses against CSCs implying the immunosuppressive effect of this pathway [10]. Recent studies presented mechanistic association between autophagy induced by Nanog in CSCs which contributes to aberrant EGFR-AKT signaling activation and subsequent immunoresistance [27].
Cancer stem cells also harness transcription factors involving in differentiation and proliferation pathways, including, Nanog, Oct4, Klf4, Sox2 in order to keep their population and maintain their niche habitable [5]. Moreover, the other transcription factor at the crosstalk of stem like feature acquisition and immunoresistance is called Yin Yang 1 (YY1) in glioblastoma. It works with other transcription regulators and involves in conventional therapy failure. In addition, YY1 makes transcriptional axis with CDK9 which recruit Treg cells, curb interferon- mediated immune responses, and minimize the potency of immune checkpoint blockades. Given all of these events, cooperation between intrinsic transcription factor and extrinsically stimulated epigenetic alterations provide cancer cells with immune avoidance power [28].
Despite sharing common signaling pathways and transcriptional factors for their unique characteristics with respect to oncogenesis and self-replication, the extreme heterogeneity can be seen in the way of CSCs’ regulation [6]. Hence, this complexity is added to the other challenges of therapy. For instance, when CSCs encounter the hostile environment, they trans-differentiate to endothelial cells in order to supply nutrients to tumor mass, induce epigenetic reprogramming in immune cells in surrounding TME to provoke protumor phenotypes of them [9, 12].
Immunotherapy is the fourth line of treatment against cancer which has enormous successes particularly in advanced and resistant ones to conventional chemoradiotherapies [10]. However, T cell-based immunotherapy targeting CSCs in solid tumors is at its beginning and faces lots of challenges which should be tackled. So, understanding the precise mechanisms used by CSCs to escape from immunosurveilance and to edit their immunogenicity is vitally important. In following parts, we present strategies exploited by this population in more details.
Most of tumor associated antigens (TAA) expressed by CSCs and recognized by immune system have been categorized as following: i. Ag with high expression in cancer cell, but no or limited expression in normal tissues (like hTERT and survivin), ii. Ag restricted to normal placenta and testis, but expresses abnormally in cancers as cancer/testis Ag (like MAGE-A4, MAGE-A3, and NY-ESO1), iii. Neoantigens originated from somatic mutations in cancer cells which are tumor specific and completely novel for immune system, iv. Tissue differentiation antigens expressed in both normal and cancer cells (like MART-1 in melanoma, PSA in prostate cancer) [1, 10, 11].
As TAAs can provoke the immune responses, particularly T cell, different platform of cancer vaccines have been proposed for boosting one’s immune system to fight better against their cancers [29]. In spite of T cell response induction by these antigens, most studies have been shown no promising clinical outcomes which can stem from ineffective eliminating of the main origin of therapy resistance, CSCs [8, 30]. Although antigen expression by CSCs is one of the main prerequisite to develop immunization against them through vaccination, it cannot be a determinant factor in CSCs eradication, since none of these markers (ALDH1, CD133, CEP55, COA-1, EpCam, HEATR1, or SOX) are exclusively expressed by CSCs. On the other hand, for most of these molecules the immune system has been evoked and in turn developed tolerogenic T cells which leads to more decrease in their immunogenicity [1].
Antigen processing is the other crucial step in arising immune response against an antigen. Recently, many investigators have reported deregulated expression and dysfunction of molecules involving in antigen processing machinery (APM) in CSCs [1, 8, 12, 31]. This system includes multiple immunoproteasomes, transporters, and chaperones which act in concert in order to release the assembled MHC- Ag to the cell surface of antigen presenting cells (APCs) [1, 30]. This multistep process have been depicted impairments in each steps in different solid cancer stem cells, including colorectal, melanoma, and glioblastoma which can be exacerbated with chemotherapy interventions [1, 32]. Although some authors have proposed HLA1 deficiency as a discriminative factor for isolating CSCs from the whole tumor cells, it should be noted that this trait may be seen even in some differentiated cancer cells [1]. According to experimental data, limited or no expression of HLA1 antigen by CSCs confers them the immunoresistance due to impaired detection and eradication by T cells. This conclusion has been made based on those experiments in which HLA1 expression improvement in CSCs resulted in refreshing T cell ability in recognition and elimination of these cells [1, 31, 33].
The molecular mechanisms of HLA1 expression defects have been studied recently. Some researchers have presented the epigenetically deregulated causes [1, 8]. However, in some other studies genetic mutations in genes related to APM components have been considered as a main mechanism [1]. In order to assess which immune cell can be selected and induced by CSCs, researchers have been designed experimental studies in which peripheral blood lymphocytes cultured with autologous CSCs. They showed T-helper 2 lymphocyte selection and propagation which emphasizes the immunoprivileged potential of CSCs [34, 35].
However, the immunological makeup of CSCs still remains unclear, recent findings with respect to downregulation of MHC1 APM components and lack of universal and specific CSCs’ TAAs underscore the challenge of applying T cell therapy for fighting against CSCs in clinical practice [1, 21]. So, elucidating the immunological characteristics of CSCs and comprehension of their complex network with immune cells and other immunological factors could pave the way of designing better therapeutic strategies based on individuals’ immune system to counteract their cancers, particularly CSCs.
Notably, HLA1 antigen deficiency makes CSCs susceptible to natural killer (NK) cell recognition, if these cells express the sufficient level of NK cell activating ligands (NKG2D) or receptors [31]. But, this tumorigenic subpopulation has been shown lower level of these ligands or receptors in order to evade innate immunity which contributes to immunoresistance [1]. In patients bearing gastric cancers with advanced clinical stages, lymph nodes involvement, and shorter overall survival, NK cell frequency and cytotoxic functions have been dropped [36, 37]. In this regard, cytotoxic potential of NK cells has been declined in the microenvironment with high gradients of immunosuppressive agents like IL-10 and TGF- β and M2 macrophages [12]. Although CSCs have been demonstrated with high expression level of ligands necessary for NK cell induction, like MICA/B, Fas, DR5, NKp30, and NKp44, the simultaneous presence of CSCs and M2 act as a potent source of immunomodulators and anti-inflammatory factors helping cancer cells elude immune system [38].
Dendritic cells mediate critical roles in stimulating T cell-based responses against cancer cells by operating as professional APCs [12, 21]. Although, their antitumor functions can be confined by the cancer, since tumor cells prevent DCs recruitment, impair their maturity, and incite their differentiation toward tolerogenic ones [21]. These cells not only do not inhibit tumor growth, but also promote its progression and invasion through expansion of Treg cell population [39, 40]. In this context, the exact role of CSCs has not been yet completely elucidated, but growing evidence points the multidirectional interactions between CSCs and DCs [21]. For instance, renal cancers enriching CSCs with high expression of CD105 hindered in vitro maturation of monocyte- derived DCs compared to their non-CSCs counterparts [41]. This inhibitory effect of CSCs was explained by producing and releasing vesicular MHC1-AgG complex which operates as a ligand for an inhibitory receptor named ILT on DCs. This interplay results in DC differentiation arrest by IL-6–STAT3 pathway [21, 41]. Interestingly, the phenotype of CSCs can also impact on DCs’ functions, since wide variety of immunomodulatory effects has been reported based on different CSCs phenotypes. For example, it has been reported that CD44+ CSCs can provoke tumor repressive phenotype of DCs, while CD44−CD133+ CSCs may subvert their anti- tumor abilities [12].
According to reciprocal connections that CSCs make by individual components in TME, DCs also supply CSCs with signals supporting their survival and maintenance. In colon cancer samples, isolated DCs with downregulated IL-12 and upregulated CXCL1 secretion amplify stem-related transcription factors including MYC, NANOG, SOX2, and OCT4 [42]. Furthermore, there is evidence indicating that protumorogenic DCs can interfere in different downstream pathways critical for cancer clinical fate, in this regard they participate in premetastatic environment preparation by CXCL1 expression which subsequently attracts CXCR2 + MDSCs [43].
Apart from lowering the sensitivity to cytotoxic T cell responses, CSCs provoke immunosuppressive population of immune system by releasing chemokines like CCL-2 leading to recruiting and propagation of Tregs and employing Th2 responses [11]. In melanoma, CSCs undermine the potency of effector T cell responses by increasing the immune inhibitory factors such as PD-L1 and CD-86 and also by decreasing their Ag expressions like MAGE-A and MART-1 [8, 10]. Interestingly, PD-L1 can impact on stem cell enrichment in a cancer by either immune inhibition or through modulating some signaling pathways like PI3K/AKT and MAPK [12]. In other cancers, including prostate carcinomas, significant association have been reported among higher level of immune inhibitory molecules like TGF- β 1/2 and lower level of HLA molecules which result in immune escape and therapy resistance [44]. In fact, prostate CSCs with defective HLA and aberrant expression of Notch and Hedgehog signaling members are able to be hidden from immune detection and be extremely aggressive with poor clinical outcome [10].
Another mechanism stated for immunoprivilege nature of CSCs is that they secrete galectin-3 which produces necessary signals for naïve and activated T cell apoptosis [11, 45]. CSCs originating from different cancers share the same strategy to provide immunosuppressive milieu including releasing soluble growth factors and cytokines supportive for Tregs expansion, M2 phenotype switching, and MDSC recruiting [10, 46]. A portion of these cellular activities are done through TME remodeling which results in further releasing of ECM- bound factors, neoangiogenesis, and that all of them provide CSCs the root of escape and invasion [10, 47].
The other way used by CSCs to overwhelm immune responses is changing the expression pattern of cell surface markers by these cells. In line with this notion, in breast CSCs, augmented level of CD200 expression has been related to declined Th1 activities, TAN functions and higher levels of immunosuppressive cytokines including IL-4, IL-10, and IL-13 [8, 10, 11]. Additionally, in squamous cell carcinoma, CSCs express CD80 which has more inclination toward CTLA4 rather than CD28 on CTL and following defective antitumor response [48].
Cancer stem cells apply their inherent competence of phenotypic and functional plasticity to conquer any restricting factors [3, 8]. In accordance with this point, vasculogenic mimicry and neuronal cell simulation are acquired by CSCs to support their survival and amplify their invasion [28, 49]. In the same way, it has been demonstrated that CSCs in glioblastoma can show functions equivalent to immune cells which prompt immunoevasion and stem-like cells expansion [28, 50]. This way of immune escape has not been established by genetic preference of special group of cancer cells, but anti-tumor immune responses impose CSCs to reprogram epigenetically and reconstruct the tumor immune landscape [51, 52]. Mechanistically, infiltrated TAMs in glioblastoma models may probably transduce signals that trigger epigenetic adjustments in transcription pathways crucially significant for immunoediting in cancer cells and these transcriptional complexes are persistent which subsequently are heritable to offspring [9]. Interestingly, glioblastoma CSCs have been shown increased expression of Irf8 which is well-known as a myeloid- specific transcription factor. This obtained proficiency of behaving as myeloid cells in hostile tumor environment can elucidate the high frequency of tumor promoting and immune suppressing cells [9].
Additionally, DNA methylation pattern has been prominently altered in this myeloid cell mimicry process and transcriptional makeup depicted in these CSCs is analogous to Treg cells. Two stem related transcription factors engaged in this process are Oct4 and Sox2 which support the development of immunosuppressive environment through restraining antitumor effector T responses by canonical Treg gene signatures, including TGFB1 and CD274 [8, 28, 50]. Importantly, unveiling the potential of immune mimicry in CSCs has emphasized the confounding effect of this trans-differentiation in immune cell infiltration data analysis and patient outcome prediction, since it is necessary to exclude these data from the original recruited immune cell data. In order to rule out the cancer cells immune mimicry transcriptome, researchers have suggested the optimized immune infiltration data analyzing method called oIRS which provides more reliable prognostic information [50]. Although genetic alterations may be the first contributor to the generation of CSCs in different malignancies, the epigenetic modifications can also exert significant driving strength in cellular adjustment to sustain their dynamic plasticity and heterogeneity [28, 51–53]. Gaining insight into how these mechanisms can contribute to tumor supporting immune profile and acquisition the most compatible state by CSCs will pave the way for enhancing the target immunotherapy efficiency.
Reciprocal Interactions among CSCs and TME Constituents Contributing to Immune Evasion:
The environment surrounding CSCs is composed of different cellular population like endothelial cells, fibroblasts, inflammatory cells, and immune cells associated with extracellular matrix (ECM). This microenvironment exerts its regulatory effects on CSCs function and phenotype as an extrinsic factor [54, 55]. Among these cells, immune cells play crucially significant role in this complex interaction [6]. In order to target this complex interaction and impede tumor relapse due to inherent therapy resistance and immune evasion capacity of CSCs, clarification of the immunological characteristics of CSCs is the first and fundamental step.
Reciprocal interplays between CSCs and immune system compartments bring about development of tumor promotive immune cells and intensify the stem-like features and frequency of CSCs which all of these events are determinative agents in shaping cellular fate in cancer [12]. CSCs with immunomodulatory competencies and persisting in quiescence can avoid immune detection and clearance, consequently lead to immunotherapy failure and tumor recurrence [56].
Mounting evidence represents the role of evolutionary pressure in shaping the complexity and configuration of immune environment in a cancer [57]. Since nutritional, anabolic, and therapeutic pressures put all cancer constituents including neoplastic cells and non-neoplastic ones (such as immune cells, fibroblasts, and endothelial cells) into cellular adaptation conditions which may lead to selection of most robust, aggressive, with low nutritional demands resulted in intratumoral heterogeneity [28, 57]. In tandem with this natural selective behavior, CSCs can cooperate synergistically with TME to trigger aggressive propagation of tumor cells and therapy failure [58]. Accordingly, neoplastic cells, particularly CSCs, are endowed with the metabolic flexibility pertinent to their surrounding nourishing environment [28]. Heterogeneous CSCs states with diverse metabolic demands and priorities can reconfigure their metabolic programs by means of post-transcriptional epigenetic changes, reformed enzymatic and transporters expression, and adjusted HIF-1a expression [1, 28, 59]. It is noteworthy that modulating CSCs metabolic programs also makes them more prone to acquiring the pro-tumor immunosuppressive phenotypes [28]. For example, in glioblastoma, CSCs present in the vicinity of immune cell infiltration reveal transcriptionally immunoresistance gene signatures [60].
The causative role of inflammation and immune response in cancer emergence has been investigated for decades. Nowadays, what makes this field of research critical is the role of inflammation and immune system in CSCs development. Since inflammation creates a milieu with high NOS, ROS gradients, high propensity for mutations, cellular reprogramming as a result of mutual interactions among present constituents, and nourishing cells with growth factors which all provide condition for a CSC generation based on proposed CSCs models [12].
1) Cellular compartments:
As macrophages are the essential cellular components of normal tissue stem cell niche, it has been postulated that CSCs might harness the intrinsic homeostatic potentials of macrophages for own survival, self-propagation, and imunoevasion [33, 61, 62]. The dichotomic classification of macrophages to M1 and M2 may lead to misinterpretation since it cannot present the dynamic plasticity of their functions based on signals they gained from surroundings [21, 57]. The TAM pool in a tumor is supplied by prenatal tissue residential and blood-borne macrophages [7, 21, 63]. As macrophages have been well known for their decisive roles in different steps of carcinogenesis, TAMs are tightly associated with CSCs in determining the cancer patients’ prognosis [12, 31, 64]. Furthermore, the prognostic value of M2/M1 ratio in tumor microenvironment has been reported in several malignancies [12, 62]. It is noteworthy that M2 macrophages have potential to exert effects mutually on CSCs’ development and fate, as these macrophages secrete immunosuppressive molecules, MFG-E8 and IL6, which induce therapy and immune resistance in CSCs by STAT3 and SHh pathways activation [10]. On one hand TAMs with anti-inflammatory and tumor-promoting features has been correlated with downregulation of CD3-ζ chain of T cells which directly impedes their efficiency [12]. Reciprocally, TAMs have been participated in modulation of some challenging stem features including chemoradioresistance, selfproducing, and pluripotency of CSCs. They exert these effects through releasing MFG-E8 and IL-6 which induce STAT3 and SHh signaling pathways in CSCs [8, 12, 21, 62]. Macrophages also interact with CSCs through PTPRZ1 and EPHA4 receptors enhancing tumor formation capacity of CSCs by activating NF- κB, AKT, STAT3, and Src pathways [21].
The concurrent presences of CSCs and TAMs in gastric cancer models have been associated with metastatic spread highlighting the role of MCP-1, IL-10, IFN-γ, and VEGF in this cooperation [12]. According to findings in different malignancies, CSCs produce IL-6, IL-8, IL-10, VEGF-A, and neurotensin acting on macrophages and polarizing them to M2 like phenotype whose productions and functions favor the acquiring stemness and EMT plasticity in cancer cells [12, 62]. One of the main signaling pathways playing significant roles in such communications between CSCs and TAMs is STAT3. In addition, this pathway regulates survival and expansion of CSCs [65].Additionally in prostate cancer models, TAMs by releasing CCL5 chemokine and subsequent the β-catenin/STAT3 signaling activation stimulate invasion, epithelial mesenchymal plasticity, and clonigenecity all of which are suggestive of CSCs presence [12].
Interestingly, recruiting macrophages to the tumor by releasing immunomodulatory agents from CSCs is closely interconnected to epigenetic reprogramming of stem-like features [8, 21]. In gastric and squamous cell carcinoma models, transcriptomic analysis has been revealed that CSCs produce IL-33 which is paramount for recruiting FcεRIα + TAMs to the vicinity of CSCs in a milieu which is dependent to these macrophages for their TGF-β gradients. Since TAMs stimulated by IL-33 create the paracrine TGF-β signaling loop triggering EMT plasticity and aggressiveness in CSCs [12, 64]. Additionally, in colorectal cancer models, IL-33 produced by other cellular components can attract macrophages to the CSCs niche. These TAMs release PGE2 which prompts CRCSCs clonigenic potential and therapy resistance [3, 66].
Several factors released by malignant and nonmalignant compartments drive recruitment and maintenance of TAMs in tumor mass, among which CCL2 and CSF1 are of significant mediators [7, 21]. Interestingly, the immunoevasion potential of a malignancy is a matter established in an early evolutionary step in a cancer [15, 21, 57]. In this sense, studies in mouse mammary cancer models have reported monocytes flowing into tumors at an early step of carcinogenesis through CCL2/ CCR2 signaling pathway [21].
Supernatants derived from human or mouse tumors, such as glioblastoma and hepatocellular carcinomas, co-cultured with CSCs have demonstrated higher levels of macrophage factors supporting tumor progression, including CCL2, CCL5, CSF1, TGFβ, CCN4, GDF15, and IL-13 compared to non-CSCs group [67–69]. Furthermore, transcriptome analysis has shown upregulation of immunosuppressive markers in macrophages cultured in media enriching CSCs, indicative of great impact of CSCs in macrophage polarity in the continuum of M1-like to M2-like [7, 21]. Thus, CSCs and recruited macrophages can operate synergistically to reduce immunosensitivity of a tumor.
Mesenchymal constituents of a tumor coupled with their products are essential cues for CSCs survival [12]. Since there are evidence demonstrating fusion of various types of microenvironmental cells like mesenchymal stem cells, fibroblasts and macrophages with cancer cells which leads to forming cells with stem cell characteristics, including tumorogenesis and migration [8, 12, 70, 71]. It has been proposed that MSCs derived from bone marrow secrete exosomes holding ubiquitin protein ligase E3 component n-Recognin 2 (UBR2). Gastric cancer cells engulf these exosomes and under the stimulation of UBR2, incite their Wnt/β-catenin pathway and initiate stem- related programs [12]. Intriguingly, CSCs applying exosomal route of communication not only can evade immune system, but also prepare their future fostering premetastatic environment [3, 8].
2 − 1) Myeloid- derived suppressor cells (MDSCs), Tumor associated Neutrophils (TANs), and CSCs:
MDSCs are a common infiltrating population seen in multiple malignancies and divided into two subtypes: Polymorphonuclear MDSCs (PMN- MDSCs) and Monocytic MDSCs (M- MDSCs) [21, 62, 64]. According to the immunosuppressive milieu created by these two categories of MDSCs, they have prognostic value in measuring the patient outcome since they can compromise the effectiveness of anticancer therapies [21, 62]. In several malignancies, the frequency of MDSCs forming the tumor immune configuration is directly correlated to its CSCs content [7, 62]. In addition, the predictive role of MDSCs in determining the clinical response to immunotherapy has been described by researchers in melanoma [1]. MDSCs induce miR-101 activity and inhibit CTBP2 function in ovarian cancer cells which result in enhancing tumor sphere formation and ALDH enzyme expression in cancer cells [21, 62]. Furthermore, in pancreatic and breast cancer cells, monocytic MDSCs play predominant role in generating CSC phenotype in cancer cells. M-MDSCs promote their NOS2 and NO production which incite Notch and STAT3 signaling axes in cancer cells and subsequently augment the potential of EMT induction and drive ALDH+ CSCs enrichment [21, 31, 62, 65]. In fact, CSCs take advantage of immunosuppressive potentials of MDSCs to maintain the environment favoring tumor progression [7, 21]. Mechanistically, this achievement may be gained by secretion of CCL15, G-CSF, MIF, CCL2, and CXCL12 from CSCs, subsequent MDSC recruiting to the TME, and increasing the STAT3 and arginase expression levels in MDSCs [21, 31, 62].
Notably, MDSCs can incite TCR impairment and T lymphocyte apoptosis through NO and ROS generation and stimulate Treg differentiation via secretion IL-10, TGF-β, and PD-L1 [8, 62]. In addition, it has been pointed that MDSCs infiltrated into colorectal cancer reinforce stem cell density by releasing exosomes carrying S100A9 factor [31].
As tumor- associated neutrophils (TANs) and PMN- MDSCs are not exactly the same subsets, the functional variation based on tumor type has been reported between them [21, 64]. CSCs also make complex functional network with TANs in different cancers, like lung cancers and hepatocellular carcinoma. In animal models, TGFβ secreted by cancer cells is involved in TAN polarization [6, 21]. In addition, CXCL5 originated from CSCs attracted TANs to the TME in hepatocellular carcinoma [21]. It is noteworthy to point the two- directional action of TGFβ in CSCs and TANs dependency. For instance, experimental studies have reported that TANs can produce both TGFβ and BMP2 which contribute to cancer cell propensity to acquire stemness features like tumor initiation potential [72].
Cancer stem cells have been reported to communicate with bone marrow derived neutrophils through exosome secretion whose RNA content can interact with gene signatures involving in protumoral reprogramming of neutrophils, including producing IL-1β, TNF-α, IL-6, and IL-8, and JAK2/STAT3 and ERK1/2 signaling pathways activation [73]. Notably, TNFα as a protumoral cytokines can enhance frequency of stem cell subpopulation in a cancer through interaction with its receptor TNFR1 activating Noxo1 and Gna14 signaling pathways [74]. Furthermore, CSCs by releasing CXCL1 and CXCL2 attract these neutrophils to the tumor environment where they can provoke EMT in cancer cells and exacerbate metastatic potential of them [75].
3 − 1) T Lymphocytes, NK cells, and CSCs:
CSCs can contribute to the development of tumor promoting environment by influencing on the immune cell proportions and their characteristics towards immunosuppressive nature [7]. Indeed, exploiting the elements found in ECM by CSCs is most commonly done in order to create tumor permissive niche. Accordingly, it has been reported that tenascin-C in the environment surrounding CSCs not only impedes T cell response induction, but also reinforces self- propagating potential of these aggressive subpopulation [3, 76, 77].
Although CSCs originated from leukemia and glioblastoma can evade from NK cell attacks by decreasing NKG2D ligand and MHC-1 expression, respectively, the lower level of MHC1 in colorectal and lung CSCs make them susceptible to cytotoxic killing in experimental studies [78, 79]. These variations can stem from the existence of confounding elements which direct the type of interconnection between CSCs and NK cells, or increase the natural selection of NK cell- resistant CSCs based on tumor type and tumor evolving demands [21]. CSCs can educate NK cell to do what, in this context, glioblastoma CSCs produce TGFβ to confine expression of NK cell activating ligands in circulating immune cells [80]. In addition, CSCs can shift to quiescence and hide themselves from NK cells’ detection by using a WNT inhibitor factor, DKK1, which can also suppress the NK cell activation ligands expression [10].
Recent studies in mouse model of epidermal SCCs have been reported positive correlation between CSCs enrichment and adoptive T cell therapy resistance which can be explained by TGFβ- dependent CD80 upregulation in CSCs, and subsequently inhibiting the cytotoxic killing effect of CTLs [81]. On one hand, CSCs in an elaborate and multidirected network can influence on T cell plasticity and functions and can mutually be influenced by them [21]. One of these possible ways is upregulating the inhibitory checkpoints by CSCs to dysregulate CTL function, for instance CSCs in breast, colorectal, and HNSCC cancers express high levels of PDL1 and in glioblastoma cancers they express high levels of VTCN1 [8]. Other mechanisms are prohibition of T lymphocytes expansion in tumor environment and confining the co-stimulatory signals [21]. Dysfunction of T cell can be in part resulted from the absorbing of CSCs-derived vesicules by peripheral mononuclear cells and producing IL-6, IL-10 and IL-1β. All of these cytokines may trigger STAT3 axis which contributes to enhancing VTCN1 and PDL1 levels and creating immunosuppressive milieu [82]. There is evidence suggestive of direct interruption of effector T cell abilities due to CSC secretome without myeloid cell mediation [21]. In this context, tenascin C (TNC) interacts with α5β1 on T cells which leads to decrease in AKT and ERK phosphorylation, so impairing the T cell proliferation in media culturing CSCs and T cells simultaneously [76, 83]. However, this pathway was just one of the potential suppressive mechanisms mediated by CSCs- originated soluble factors which guarantee further research on clarifying their precise mechanisms. Significantly, growing body of evidence has been reported incompetent T lymphocytes in detecting and eliminating CSCs may foster stem- like features in cell cultures with unknown processes [21].
Among different types of lymphocyte responsible for immunosurveillance, Th 17 cells have been reported the most common lymphocytes in gastric cancers and are adversely correlated with patient survival [12]. The main interleukin produced by Th17 is IL-17 which can also be produced by other immune cells like Treg. It is noteworthy that in most cases Th17 cells have pro-inflammatory characteristics in contrast to Treg cell functions, they have potential to amplify stem-like population in tumors by producing IL-17 and IL-22 [12, 21]. In fact Th17 and Treg cells originate from the common progenitor cells that confer them the ability of transdifferentiation to each other [21]. In order to maintain the tissue homeostasis with respect to inflammatory response, the ratio of these two cell types should be always in a balance [21]. CSCs can manipulate this immune cells’ phenotypic plasticity by secretion cytokines like IDO1 and TGFβ. Although IDO1 suppresses the Treg cell transdifferentiation to Th17 cell, TGFβ may support this conversion in association with IL-6–STAT3 axis [8, 21].
Nevertheless, there is evidence suggestive of transforming the IL17− Treg into IL-17+ Treg with gaining pro-inflammatory functions. On the other hand, FOXP3− Th17 has been shown to turn into FOXP3+ Th17 with the functions like regulatory T lymphocytes [84, 85]. Lymphocytes with both IL-17 and FOXP3 expression in colorectal cancer provoke the CSCs development and expansion in hypoxic niche through activating the MAPK and AKT kinases signaling pathways [86]. Additionally, in gastric cancer the co-culture of IL-17+FOXP3+ T cells and bone marrow- derived cells resulted in the generation of cells with stem cell features. This finding is in accordance with the hypothesis of gastric CSCs generation [87]. Furthermore, the regulatory networks composed of bone marrow derived MSC, GCSCs, T lymphocytes, and progenitor tumor cells may determine the extent and ratio of Th17 and Treg population in a mass which can influence on the clinical behavior of it [88]. Since it has been postulated that BM derived MSCs release TGF-β under the influence of GCSCs which probably tip the balance in favor of Treg cell differentiation. In contrast, the communication between progenitor tumor cells and BM derived MSCs stimulate lymphocytes to differentiate to Th17 [12].
Th17 lymphocytes exert their protumorogenic effects by producing IL-17β which activates EMT core programs and AKT/GSK3-β/β-catenin signaling pathway which culminating in acquiring stemness characteristics [89]. Additionally, IL-17β can enhance invasive potential of dormant CSCs by STAT3 signaling pathway activation. It has been demonstrated that IL-17 has wide range of functions on multiple cells like tumoral mesenchymal stem cells which can modulate the proportion of immune cells in tumor microenvironment [12].
Treg is the main tumor suppressive contributor present in the TME. CSCs can recruit them directly by releasing CCL1, CCL2 and CCL5 chemokines acting on Tregs or indirectly and through hiring firstly myeloid cells, by which attracting Tregs to the tumor [21]. Two other significant promoters of Treg cells trafficking are IDO1 and TGFβ produced by CSCs and reciprocally these Tregs enhance stemness pathways and features regarding tumorogenesis and colony formation [8, 21]. One of the main pathways involved in stemness induction and advent of related characteristics is TGF-β/Smad signaling axis [12]. In gastric cancer, TGF-β released by Treg cells induces LGR5 upregulation through canonical Wnt signaling pathway concluding in patients’ poor prognosis [90]. Mechanistically, Wnt pathway contributes to PGE2 production which causes Th1 response switching to Th2 responses and limiting the antitumor reactions [8].
2) Soluble mediators, MicroRNAs, and Exosomes participating in immune avoidance capacity of CSCs:
Cancer stem cells like their normal counterparts produce and release a wide range of soluble factors with multiple immunomodulatory abilities, including IL-10, IL-13, GDF-15, Galectin-3, and TGF-β [8]. All of these factors can contribute to differentiation of Treg or MDSC from immune progenitors recruited to the tumor environment [1, 91]. Cancer stem cells together with these immunosuppressive cells create a vicious cycle in which their secreted products drive more immune avoidance followed by more multiplication of CSCs and more pro-inflammatory and tumor permissive cells [1, 62]. This signifies the ultimate effort of CSCs in order to preserve their population and supportive microenvironment [62, 64]. According to the immune contexture of a cancer, recruited macrophages and dendritic cells might differentiate to anti-tumor or protumor phenotypes. It is noteworthy that the tumor permissive phenotype of TAMs and DCs are conducive to increase in frequency and aggressiveness of CSCs in different cancers [1]. Cancer stem cells originated from hematopoietic cancers by using TIM3/ Galectin 9 signaling pathway disrupt T cell antitumor responses by developing immunosuppressive environment with extensive MDSCs and TAMs recruitment and differentiation [92]. Hence, exploring the underlying mechanisms of such complicated interactions among innate and acquired immune ingredients with CSCs will provide us with essential knowledge to design more precise experiment based on CSCs’ immunologic features.
Other signaling molecules including chemokine CXCL12 and its receptor CXCR4 expressed by wide variety of cells in a tumor mass support CSCs to maintain their self-producing ability and also mediate the MDSCs and protumoral DCs recruitment to TME, and increase Tregs differentiation [1].
Cancer stem cells derived from several epithelial tumors present IL-4 and CD200 which affect negatively on T cell expansion and cytotoxic functions [1, 8]. Interestingly, aberrantly produced CD200 by CSCs in different solid tumors can both mediate immunomodulatory effects and contribute to maintenance of stemness mode of cancer cells by involving multiple transcription factors inducing EMT process and CSCs signaling circuits [93, 94]. The immunoresistance ability of CSCs may also stem from upregulation of IDO (indoleamine 2,3-dioxigenase) in these cells compared to other differentiated cancer cells which act through influencing on MDSCs to exert immunosuppressive effects on T cell responses [1]. Additionally, CSCs overexpress immune checkpoint factors including B7-H3, B7-H4, and PD-L1 in order to curb the effector T cell differentiation and weaken the extent of T cell- based responses in one hand, and intensify their own stem related characteristics in the other hand [1, 12]. In contrast, downregulation of PD-L1 has been correlated with reduced tumorogenesis and therapy resistance in CSCs. [12].
Experimental studies depicted the positive correlation among reprogramming transcription factors, particularly Sox2 and Oct4, and immunopriviledge competency in CSCs. As Sox2 can induce CD39 expression in these cells, consequently contribute to blocking the immune response by sequestering the ATP and limiting its availability to T lymphocytes [95]. Another transcription factor, Oct4, together with Sox2, stimulate inhibitory immune checkpoints production, including PD-L1, TDO, CD70, and A2aR which culminate in increasing immunomodulatory agents like CXCL3, IL8, CCL20, and SSP1 and decreasing chemokines such as CCL5, CXCL9, and CXCL10. This set of alterations has been occurred epigenetically and it depends on BRD4 protein which is defined as an epigenetic reader and remodels chromatin following histone modifications [28].
MicroRNAs are noncoding mRNAs with 18–23 nucleotides length regulating posttranscriptional gene expression and based on target genes they can suppress or motivate their expression. In CSCs, different groups of miRNAs modulate their tumorogenesis, immunoevasion, dormancy, and phenotypic plasticity [3]. According to their roles, some of miRNAs might be downregulated, including MiR451, MiR199b-5p and some others can be upregulated in CSCs such as MiR31, MiR34a, and MiR205 [1, 15]. It is noteworthy that microRNAs can effect on both T cell effector functions and CSCs immunoprofile which makes them a crossing point in studying mechanisms used by CSCs to overwhelm T cell attacks [1]. Recently it has been shown that breast CSCs enhance their stemness potentials by means of MicroRNA 199a upregulation, since this miRNA inhibits differentiation and senescenece pathways induced by IFNs produced by immune system [97]. In other cancer models, hypoxic TME enriched with high level of miRNA-210 can modulate the immunosensitivity of CSCs to T cell- based cytotoxic attacks [1, 97].
Growing evidence has been described exosomes as an important way of communication among different cell types, particularly between CSCs and cells of immune system. Exosomes are small vesicles originated from multivesicular endosomes and may carry multiple cargos, such as miRNAs, mRNAs, TFs, growth factors, and DNA [1, 3, 8]. Exosome release their contents by assembling with the cell membrane. Their production and secretion from one cell can be affected by environmental alterations created by other cells and soluble factors which subsequently can exert changes in other cells’ functions receiving them [3]. Additionally, exosomes play crucially significant role in immune makeup of CSCs. This aggressive cellular fraction transduces their communicating signals by these carriers and consequently remodels their niche in order to provide themselves with a way of escape from active innate and acquired immune contributors [1]. For instance, exosomes enriched in FASL, TRAIL, and Galectin-9 can reprogram APCs like monocytes and DCs toward MDSCs or suppress CTLs or stimulate effector T cell apoptosis [98]. Further analyzing the cellular production of PD-L1, B7-H3 and IDO by CSCs has been unveiled that all of these molecules were stored in microvesicles, unlike in other cancer cells they were gathered in the cytoplasm [99]. Accordingly, CSCs exploit all the current mechanisms to be the immunopriviledge component of a cancer, escape from immunosurveilance, be quiescent, and regrow in a second place as a metastatic foci. So further research should be done to investigate the genomic and epigenomic regulations of immunological characteristics of CSCs, since discovering such complicated interplays will contribute to recognizing the manipulating strategies to recover CSCs’ sensitivity to immunodetection and elimination by immune system.