Despite the discovery of numerous diagnosis processes and modalities of cancer therapy, there is still scope for developing more effective and targeted therapeutic options of cancer. Revealing the origin of cancer, therefore, has become an extremely important issue in that it might shed light on successful and relapse-free treatment of this deadly disease. During the last 25 years, a plethora of evidences have suggested the candidature of a small subpopulation of cancer cells, named cancer stem cells (CSCs) with tumor-initiating capability, as the core origin of tumorigenesis [1, 2]. Like normal stem cells, CSCs take charge of a cancer hierarchy, harbouring stem cell-like properties involving self-renewal, to form identical daughter cells, and differentiation potential to create various types of progenies that finally generate non-stem cancer cells (NSCCs) to build up the total tumor mass [3]. Besides, CSCs have also been characterized as the drivers of metastasis [4, 5]. Enhanced invasiveness and metastatic capability of any tumor that ensure infiltration to the surrounding area and forming new tumor to the secondary organs, have been reported to be related to CSCs [4, 6]. In fact, quiescent CSCs that have travelled to distant places may cause metastases that arise many years after curative surgical treatment of a primary tumor. CSC hypothesis could, therefore, implies that targeting CSCs might be the most promising therapeutic strategy. However, despite frequent efforts to target CSCs, current anti-cancer therapies are still obstructed by highly drug-resistant CSCs. In fact, by application of therapeutic regimens such as chemo- or radiotherapy, NSCCs with a high proliferative capacity are targeted and killed causing a decrease in tumor size while rare CSCs survive these therapeutic regimens [7], explaining why local recurrence is usually always the result of effective treatment of solid tumors with radiation or chemotherapy [8]. Such high resistance thus provides CSCs not only a survival advantage over differentiated NSCCs [7], but also the chance to “recreate” the tumor after anti-cancer therapy [9]. At this point, it will not be out of context to mention that in a tumor mass, since majority of the tumor cells are NSCCs, they become more effective in immune evasion than CSCs. This information, together, signify the robust functionalities of this small subpopulation of tumor-initiating cells (TICs) as the major contributors in tumor immune evasion during tumor initiation, metastasis, and relapse.

It is acknowledged that during the early stages of tumor initiation, immune cells get activated and infiltrate to tumor microenvironment (TME) to mount an effector immune response, eliminating immunogenic tumor cells [10, 11]. Although above discussion portrayed CSCs as the TICs during primary initiation, metastasis, and relapse [2, 12], their survival during the initial ‘elimination phase’ of immune attack [13] still remains elusive. This tempted us to understand whether due to change in the immune system by CSCs ensures their own survival during elimination phase to finally build the whole tumor mass either during primary initiation, or during metastasis or recurrence.

This review discusses the recent studies investigating immuno-modulation of CSCs during tumor initiation, and the underlying molecular mechanisms as well as the impact of immune cells on CSCs. A deeper understanding of this bidirectional crosstalk that reshapes the immunological landscape and governs therapeutic responses, will expedite the improvement of cancer immunotherapy, in which CSCs might apply the same strategy of overpowering immune attack by suppressing the infiltrating active immune cells, thus generating immunosuppressive milieu.

A growing body of evidence suggests that malignant tumors are initiated and maintained by CSCs which have similar biological properties to normal adult stem cells [14]. The CSCs have specific surface signature-markers depending on their tissue of origin. The CSC hypothesis states that tumors, like adult tissues, arise from cells with the ability to self-renew and/or give rise to differentiated tissue cells [14]. According to the canonical hardwired stem cell/CSC hierarchy model, stem cells/CSCs are rare, relatively dormant, and defined primarily by intrinsic properties. When they divide asymmetrically, they produce one stem cell and one transient amplifying cell [15]. As the tumor evolves, these progenitor cells differentiate to NSCCs that proliferate quickly to form the tumor mass. Therefore, according to the CSC theory, at the early stages of tumor initiation very few CSCs are engaged to sculpt the tumor mass, which will eventually be built with the division of NSCCs generated by the asymmetric division of CSCs [15, 16].

To metastasize and initiate new tumor at the secondary site, cancer cells go through a “metastatic cascade”, consisting of extracellular matrix (ECM) disassembly by enzymes such as matrix metalloproteinases (MMPs), followed by remodelling of the ECM to drive growth in the metastatic niche [17, 18]. These obstacles make metastasis an inefficient process in which most tumor cells die in circulation or upon arrival at the distant tissue [11]. Only a subset of disseminated cells with CSC-like properties [19], overcome such hurdles and initiate tumor growth at distant tissues. These “metastatic stem cells” [20, 21] like primary tumor CSCs, have the ability to restart tumor growth at distant sites. According to our findings, non-migratory tumorigenic intrinsic CSCs generate C-X-C chemokine receptor type-4 (CXCR4)-positive migrating CSCs in the tumor’s periphery, possibly to help the migrating fraction withstand the rigours of the metastatic voyage, because depletion of this population resulted in the complete tumor’s metastatic activity being halted [22]. These reports support our hypothesis that CSCs adapt to new environments, initiate the formation of heterogeneous tumors, and explain the clinical course of metastasis and distant relapse [23, 24].

Tumor recurrence is a significant impediment to cancer treatment, even though its underlying causes are mostly unknown. The most commonly used antitumor therapies, such as chemotherapy and radiotherapy, have lethal effects on cancer cells that are highly proliferating. CSCs are slow-cycling by nature, in contrast to their transitory amplifying progeny, which actively proliferates prior to differentiation [25, 26]. CSCs have the ability to enter the quiescent cell cycle phase, which represents a reversible latent state in which cells are alive but not multiplying [25, 27]. As a result, non- or slow-proliferative latent CSCs escape chemotherapeutic cytotoxicity and persist after treatment. As the cytotoxic effects of the chemotherapeutics wear off, they re-appear and initiate tumors in the same or at distant new locations [13, 28, 29], eventually leading to tumor relapse. Moreover, this group has also reported the contribution of CSCs to acquired chemoresistance and higher residual risk of recurrence in cancer patients having acquired chemoresistance [13] since as a result of their inherent properties and adaptability to their respective TME, “drug-spared” CSCs play pivotal roles in fuelling tumor growth during relapse.

This discussion signifies the possibility of involvement of CSCs in initiating tumor during initiation phase, after metastasis, and during tumor relapse (Figure 1).

A typical tumor shape is made up of the tumor core, as well as the surrounding stromal and lymphoid components [30, 31]. Immune infiltration can vary from patient to patient. Macrophages, dendritic cells (DCs), natural killer (NK) cells, mast cells, B lymphocytes, and T lymphocytes including T-helper-1 (Th1), T-helper-2 (Th2), T-regulatory (Treg) cells, and cytotoxic T lymphocytes (CTLs), are all common immune cell types found in TME [30, 32]. Within this immunological milieu, there are components that are advantageous to the patient as well as components that are harmful [32, 33] (Figure 2). Tumors create an array of neo-antigens, which allow the immune system to distinguish tumor cells from normal cells [34, 35]. Immune surveillance is the process by which malignant cells are identified, damaged and are then targeted for eradication by the immune system.

Growing evidences suggest that immune cells can both prevent the formation of precancerous cells and facilitate the regression of established tumor burden [34]. Substantial evidences claim that certain innate and adaptive immune cell types, effector chemical molecules, and immunogenic pathways can sometimes operate as extrinsic tumor-suppressor mechanisms in concert. Antitumor effector immune cells such as macrophages, DCs, and other myeloid-derived polynuclear cells not only phagocytose and present neoplastic altered cells (and/or foreign pathogens) to Th cells, but also mount antitumor immunity by leaking tumoricidal cytokines [36, 37]. Continuous approaches and technologies have precipitated several important discoveries related to the adaptive handle of tumor immune response. These reports demonstrating T cell mobility within different lymphoid and non-lymphoid compartments, effector functions after antigen presentation, priming other immune cells for eliminating foreign pathogen and/or neoplastic cells via secreting plethora of cytokines, advanced our understanding of effective immunosurveillance [38, 39].

This paradigm has recently been expanded to include the concept of immunoediting, in which altered self-cells give rise to tumor formation despite having selection pressure from immune system [40, 41]. However, some tumors escape the immune clearance. This incident highlights the seminal role of tumor in reshaping the immune cells into immune-suppressive pro-tumorigenic cells to ensure their own survival [42].

Early detection of neoantigen-expressing cells engaging several immune cell types, is a complex process [35]. Stemming examples demonstrate the innate immune system’s instrumental and decisive functions in eliciting antitumor immunity within the tumor onset window period [34, 36, 38, 39]. APC subtypes coordinate and produce an immune response against neoantigen-expressing cells even in the absence of PAMPs which are generally necessary to licence APCs to present foreign or self-antigens as immunogens [33]. When mutated gene products are broken down and displayed as neoepitopes on the patient’s MHC molecules, they can behave as tumor neoantigens. This immunological complex allows surveying monocytes to present neoantigens to CD4⁺ T cells as immunogens. CD4⁺ Th cells use CD40L to licence cross-presenting DCs to induce a CTL response against neoantigen-expressing cells [138]. Th cells direct the development of antigen-primed B cells in secondary lymphoid tissues which secrete natural IgM required for the detection of neoantigen-expressing cells and the beginning of adaptive immunity [138]. Thus, adaptive immune responses take some time to get primed, expand, and secrete antibodies against an antigen. But the number of cancerous cells increase during this time course since they have endless proliferating capacity. Without making any delay, our body triggers innate immune responses to engulf the changed self-cells or eliminate the premalignant cells via perforin-granzyme mediated cytotoxicity [34, 36].

So far, this review has discussed that CSCs are often responsible for tumor initiation and progression. However, the question is always raised that despite having strong immune surveillance mechanisms, how could CSCs get undetected by the immune system? Generally, CSCs play comprehensive roles in modifying the body’s immune response in favor of its own survival by creating immunosuppressive conditions via multiple mechanisms (Figure 3). By producing GM-CSF, periostin, and cyclooxygenase-2, CSCs recruit TAMs and polarize them into the pro-tumor M2 phenotype [139–141]. Studies reported that CSCs produce various substances that aid in infiltration as well as formation of MDSCs to generate an immune-suppressive milieu [87, 88, 142]. Tumor-initiating CSCs also express various chemokines, like CCL1 and recruit immunosuppressive Treg cells via NFκB-CCL1 signaling axis [100]. Such findings shed light on the functional relationship between Tregs and CSCs, thereby pointing towards a promising future cancer therapeutic method. It has been already mentioned that cancer cells attract immunosuppressive cells via generating numerous factors such as IDO [143], TGF [63, 144], IL10, and PDL1, thereby avoiding immune responses [63, 66, 96]. Remarkably, current studies show that CSCs, express the same set of markers as IL10 [145], IL4 [61], IDO [101], and PDL1 [118, 119]. Therefore, the above-mentioned immunosuppressive environment might also be generated by the CSCs during initiation.

Recent report showed that EMT-β-catenin-STT3 (oligosaccharyltransferase complex)-PDL1 signaling-axis enriches PDL1 in CSCs [119]. This axis is also used by the general cancer cell population, although it has a far greater impact on CSCs than NSCCs to promote immune evasion [119]. CSCs have been shown to downregulate the expression of transporter associated with antigen processing (TAP), a vital player in antigen processing and presentation machinery in order to mask the recognition by adaptive immune system [66, 146]. CSCs also avoid the cytolytic effects of NK cells by downregulating NK group-2, member D ligands (NKG2DL) expression [147]. Such immune evasion is essential for the replication of TICs which further gives rise to differentiated tumor cells, resulting in formation of a heterogeneous tumor bulk [5]. Similar incidents take place when a tumor relapses after chemotherapy. Besides the immune-regulatory effects of CSCs, as mentioned above, the stressed immune system following chemotherapeutic shock, allows drug-resistant CSCs to resurface from lymph nodules, and initiate tumor [9]. As the arena of research involving CSC-immune cross talk is a recently emerged field, the information regarding how CSCs re-educate the antitumor immune system to generate pro-tumor one, is not yet complete. More intense research is needed to know in depth the full mechanisms underlying CSC-mediated immune-suppression during tumor development and progression.

Strong evidences suggest that CSCs give unique inputs to the TME in order to positively govern their survival and behaviour, dictating their ability to remain stem-like, invasive, and resistant to drug-induced apoptotic signals. Because CSCs have different treatment resistance modalities by nature, eliminating CSC pools is a valuable strategy for achieving a relapse-free cancer cure [63, 148] (Figure 4). Exploiting a variety of factors that make immune effectors more immunogenic towards CSCs and other resting cells in the tumor niche could aid in the development of novel therapeutic approaches. Several treatments have been developed with the goal of eliminating CSCs, including (i) specifically targeting CSCs, (ii) modifying their stemness, and (iii) influencing the microenvironment. In this case, monoclonal antibody-based therapies are best suited to target the markers commonly used to isolate and detect CSCs [149].

CSCs can also be selectively suppressed by compounds that inhibit drug-efflux pumps and detoxifying enzymes [1]. By targeting CSC-specific pathways including Hedgehog, Wnt, Notch, and focal adhesion kinase by CSC-specific cytotoxic drugs, positive results have been obtained [150]. All-trans retinoic acid [151, 152], histone deacetylase inhibitor [153], bone morphogenetic proteins [154, 155], and cyclopamine [156] are a few of the compounds that induce cell differentiation while suppressing CSCs’ specialized stemness. Furthermore, various small molecule inhibitors like mithramycin A, curcumin, andrographolide [157–160] or conventional drugs have been repurposed to sensitize CSCs prior to chemotherapy [161]. In this regard, repurposing aspirin for successfully sensitizing breast CSCs towards chemotherapy has been reported recently from this group [13, 29, 162, 163]. Targeting the interaction between tumor-stroma and nearby cancerous cells by inhibiting paracrine signaling molecules such as VEGF, hypoxia-inducible factor, CD44v, and CXCR4, also holds alluring prospects for the successful eradication of CSCs [164]. Furthermore, the chemotherapeutic drug gemcitabine and the ALDH1A1 inhibitor disulfiram have been reported to suppress breast tumor development by increasing T-cell immunity and eliminating ALDH⁺ TICs [87]. All these findings reveal novel therapeutic approaches for targeting cancer from its “root”.

Cancer immunotherapy is an emerging treatment strategy that re-educates host’s immune system to fight against cancer [165]. Although several immunotherapeutic strategies are employed in cancer treatment, low patient response rates remain a significant obstacle. In fact, plethora of basic science discoveries and subsequent clinical translation finally could undeniably demonstrate the supremacy of immune system modulation in tumor regression. Further research involving regulation of T cells and other immune cells, e.g., APCs and NK cells, might allow the scientists to improve the efficacy of this approach [166–168]. The effects observed in clinical trials of checkpoint blockade agents [169, 170], adoptive T cell transfer therapies [170], and cancer vaccines [171, 172] in “difficult to treat” tumors were far greater than the most effective chemotherapeutic agents. Even though immune-related side effects are common, these novel immuno-targeting therapies are more well-tolerated than traditional chemotherapeutic agents but definitely is cost-effective [173].

The field of NK cell-based cancer therapy has advanced significantly and is now a critical area of immunotherapy innovation. The development of NK cell-directed therapeutics should prioritize two objectives: increasing the source of therapeutic NK cells for adoptive transfer and increasing their in vivo cytotoxicity and durability while avoiding CSC-mediated suppression [174–176]. Immunotherapy that combines the benefits of cellular and antibody-based therapies employs chimeric antigen receptor-expressing T (CAR-T) cells [177, 178]. CAR-T therapy involves the administration of autologous CD4⁺ and CD8⁺ lymphocytes that have been modified to express CAR on their cellular surfaces. This allows the treated cells to recognize and respond to a specific antigen without requiring an earlier antigen-presenting procedure. There is also evidence that CAR-Ts that specifically target markers such as CD44v6, CD133, EpCAM, and CD54 can effectively eradicate CSCs [179–182]. But CSCs can limit the activity of CAR-T cells regardless of whether they are targeted or sensitized.

This review highlights the role of tumor-initiating CSCs in modulating the immune-landscape of the cancer patients, raise the important issue that any effort to boost host via present-day available cancer immunotherapy might finally activate the pro-tumor immunity and not the antitumor immune response of the patient through CSC-induced re-education of the same, thereby leading to failure of the high-cost therapy. Disrupting the deadly liaison between CSCs and immune cells should therefore be the cutting-edge research target to finally achieve fully successful immune-therapy of cancer in near future.

This review establishes the instrumental role of CSCs during tumor initiation, metastasis to a secondary organ, and tumor relapse. CSC-regulated re-education of immune cells for ensuring the survival of these TICs during tumor initiating phases bolsters the paramount significance that CSCs uphold in tumor formation, propagation, and immune-modulation. This review article has discussed how the anti-cancer immune-subsets infiltrate the tumor site and eradicate cancer cells via multitude of mechanisms. However, tumor-initiating CSCs manipulate the TME in its favor by promoting infiltration of pro-tumor immune-subsets such as Treg and TAMs, as well as, by re-educating the phenotype from pro-inflammatory to an immune-suppressive one, thereby, establishing a deadly liaison between CSCs and pro-tumor immune cells. As a result, the tolerogenic immune milieu is formed, further supporting CSCs, thereby ensuring tumor initiation and growth. Hence, it becomes imperative to develop therapies targeting CSCs and their interaction with immune cells rather than already established conventional therapeutic regimens for getting successful and relapse-free survival of cancer patients. Consequently, it highlights few treatment avenues which either specifically sensitize CSCs or attack the CSC-immune cross-talks. However, future studies are essential to be conducted for effective targeting of CSCs to successfully eliminate tumor. To that end, appropriate model systems employing authenticated cell surface markers and/or reporter systems, properly validated for CSC activities, are required to investigate the interaction of these cells with the immune system. Without acquiring the modalities for targeting CSCs and understanding CSC-mediated complexities of the TME, it will be difficult to develop effective immune-modulatory treatment procedures of cancer including the flourishing field of cancer immunotherapy. In fact, the prospective success stories of cancer immunotherapy, emphasize the requirement of coherence between basic CSC research and clinical practice. This review demonstrates how a bench-to-bedside approach, based on CSC research, will pave the way for a successful fight against one of human’s most feared diseases.