Despite the advancement in drug development, chemoresistance and metastasis are the primary reasons for high cancer mortality. Recent research has identified that cancer stem cells (CSCs), a rare sub-population within cancer cells possessing self-renewal ability, drug resistance and high metastatic ability, are the reason for the relapse of the disease. Even though CSCs were identified in 1994 in AML (1), their relevance in solid cancer was widely explored after the identification of CD133⁺ population in colorectal cancer in 2007 (2, 3). Research in this field for more than a decade has reinforced the importance of this self-renewing population in cancer progression, metastasis, chemo-resistance and recurrence. The classical chemotherapeutic agents induce apoptosis by damaging DNA and/or by inhibiting mitotic division. As this therapeutic option is valid only for highly dividing cancer cells, slow dividing CSCs escape the drugs, leading to relapse. In addition, CSCs get a survival advantage by the high expression of the ATP-binding cassette (ABC) transporters, aldehyde dehydrogenases (ALDHs), and antiapoptotic molecules (4), acquiring drug resistance. Metastasis is a dynamic multistep process including the escape from the primary tumor, intravasation to the systemic circulation, extravasation at metastatic site, organ-seeding, and final metastatic colonization with the tumor outgrowth. Importantly, only a minority of the dispersed tumor cells expressing the stemness markers characteristic of CSCs, known as metastasis initiating cells (MICs), survives all these steps to initiate metastasis at a distant site. Given the importance of CSCs in bad prognosis, strategies for targeting them is gaining importance in cancer drug discovery (5–13).

The clinical relevance of CSCs is largely attributed to their resistance to chemotherapy and radiotherapy, which is mainly due to the quiescent nature of CSCs. The treatment-induced and microenvironment-induced de-differentiation of cancer cells to CSCs also poses a challenge in tumor eradication. CSC properties, including self-renewal ability, therapy resistance and metastasis, primarily depend on the aberrant expression of CSC-specific molecules and their interaction with CSC-niche. Thus, the current approaches for eradicating this population include differentiation of CSCs, targeting drug efflux molecules and other surface markers of CSCs, and inhibition of the signaling pathways sustaining CSCs (5). Various studies have been conducted to disrupt the CSC niche using inhibitors of CXCR4 or FAK ( Table 1 ). Inhibitors of drug resistant molecules, MDR proteins including ABCG2 have been targeted to abolish drug resistant CSCs ( Table 1 ). As reviewed recently, a majority of the attempts in clinical trials are for targeting the important signaling pathways like Notch, WNT, Hedgehog and Hippo along with immunotherapeutic approaches, both in solid cancers and hematological malignancies (12, 29)

A comprehensive analysis of the reported clinical trials in comparison to their preclinical analysis showed that the majority of the drugs that were tried in trials were already in use for other diseases or for cancer itself, but not specifically for CSCs ( Table 1 ). The CSC-targeting efficacy of Hedgehog inhibitor Vismodegib is established in preclinical studies of breast cancer (30). The Hedgehog inhibitors Vismodegib and Sonidegib are FDA approved drugs for basal cell carcinoma, and their clinical trials have shown some acceptable response, though CSC specific activities were not evaluated in those trials (31, 32). When these drugs were combined with conventional chemotherapeutic agents, there was no advantage observed for Hedgehog inhibitors (12, 33). The clinical trials for inhibitors of WNT pathway and Notch pathway have shown some positive aspects in early phase trials, but their therapeutic potential and efficacy in targeting CSCs are yet to be proved (12, 33). Thus, most of the inhibitors for the well-established self-renewal pathways, which influence many cellular functions are not promising therapeutic agents in clinics, as summarized in Table 1 . The clinical trials for targeting some other pathways, cell surface molecules, MDR proteins, niche interactions, etc. are ongoing. Since CXCR4-mediated interaction of CSCs to “CSC niche” is critical for their maintenance, targeting this molecule using its antagonist Plerixafor has shown a very promising result in acute myeloid leukemia with 46% of complete remission (34). The phase II/III trial of targeting the CSC marker, EpCAM using a monoclonal antibody Catumaxomab in ascites secondary to epithelial cancer has demonstrated a clear clinical benefit in patients, leading to its approval in Europe (35). Currently, it is being used for immunotherapy using CAR-T cells recognizing EpCAM for advanced solid tumors, the results of which are awaited. To conclude, several approaches are adopted to target CSCs, which have shown some promise but their therapeutic efficacy in larger clinical trials have to be proved. Irrespective of the strategy used, the efficiency of the drug is evaluated in clinical as a well as preclinical analysis using certain assays. Currently, CSCs are evaluated by FACS-based detection of the surface markers, sphere formation assay, soft agar colony formation assays, in vitro and in vivo serial passaging as well as in vivo limiting-dilution tumorigenicity assays using immunocompromised mice. Hence, ultimately, the screening assays define the efficiency of the drug. If the screening assay is inefficient in representing all the heterogeneous CSCs, it will lead to the failure of the drug in clinical trials.

In a classical drug discovery program, the majority of the drugs that enter the screening program fails in the “Valley of Death”, and a very few candidates emerge as drugs for clinical use. Since the in vitro assays used to ascertain CSC properties are not compatible with high-throughput screening platforms, it is very difficult to have a screening strategy for CSC-targeting drugs. Hence, the evolution of clinical trials for CSC targeting is by exploiting “drug repositioning”, which can bypass the problem of bioavailability and safety problems (13, 36). Yet, many of the drugs that entered clinical trials did not improve therapeutic efficacy (37). Hence, there is an increasing demand for new drugs for CSC-targeted therapy. In this context, we propose that CSC-targeting drug screening platforms can be developed using reporters of CSCs, which have been recently published (38–45). These fluorescent reporters can be easily adapted for large scale screening. Considering the heterogeneity of CSCs, including a quiescent population and a proliferative population with the metastatic ability (46), we propose the use of a combination of reporters. A dual reporter for a pluripotency marker and a marker like ALDH1A1 or ABCG2 will be useful in screening drugs that target CSCs. These dual reporters will ensure the representation of all the heterogeneous CSCs in the screening assay.

CSCs can be defined as a subpopulation that exhibits self-renewal property, while there are accumulating evidences to show that different pools of CSCs co-exist in a single tumor that can have varied properties including quiescence, drug resistance, epithelial to mesenchymal transition (EMT), invasion and metastasis (47). But, it is not necessary that all CSCs possess all the above characteristics. Recent reports underscore the induction of stemness in cancer cells according to the microenvironment in which the cancer cell resides. According to this induction theory, CSC is a state with high plasticity where all the CSC-associated properties are regulated by the “CSC niche”, which in turn can be attributed to the cytokine levels, hypoxia, acidic pH, cancer associated fibroblasts (CAFs) and the signaling pathways initiated thereby (33). Hypoxia, the low oxygen concentration in the tumor microenvironment, can influence the tumor cells, infiltrating blood cells and CAFs to alter the secreted cytokines and growth factors in a HIF-1α-dependent manner, which results in the acquisition of CSC properties (48). HIF-1α can impart drug resistance through transcriptional up-regulation of MDR1, ABCG2 and MRP2, while it induces stemness by regulating CD133 and Notch pathway (48). Hypoxia regulates epithelial to mesenchymal transition (EMT) which in turn, up-regulates EMT/β-catenin/STT3 axis as well as PD-L1, thereby helping in the immune evasion of CSCs (49). Hypoxia and other secreted factors collectively up-regulate certain oncogenic stemness molecules like Notch1, which is reported as a marker for HNSCC stem cells (50). The role of hypoxia in hematological malignancies is also studied in detail, and the important findings are reviewed elsewhere (51). In multiple myeloma cells, hypoxia activates EMT program and increases the expression of stem cell transcription factors, ABCG2 and ALDH1 to impart cancer stem cell phenotype and metastasis (52, 53). Hypoxia is known to impart drug resistance through the up-regulation of P-gp and FAK (54, 55). The developmental pathway up-regulated in CSCs are WNT, Hedgehog and Hippo, and the biological activities of CSCs are regulated by several pluripotent transcription factors, such as OCT4, SOX2, NANOG, KLF4, and MYC (5, 12). We have summarized how the factors and signaling pathways existing in “CSC niche” regulate properties associated with CSCs in Figure 1 . The role of these pathways in the regulation of CSCs is extensively reviewed recently (5, 12).

The “CSC niche” provides an immunosuppressive environment for the survival and growth of CSCs, which will enable tumor growth, chemoresistance and metastasis. In the primary site, the tumor microenvironment is adapted to induce stemness, acquire chemoresistance and induce metastatic ability for CSCs to convert them to MICs. The “CSC niche” in the metastatic site is also important in metastatic spread of the disease. Accumulating evidences suggest the existence of pre-metastatic niches that are permissive to the colonization of specific CSCs (56). These niches allow the CSCs to remain dormant during chemo/radiotherapy, and leads to recurrence after the withdrawal of the therapeutic agent. Bone marrow microenvironment is an important CSC-niche for both hematological malignancies as well as solid tumors. The leukemia stem cells are shown to hijack the hematopoietic stem cell niche to acquire and maintain CSC properties, including self-renewal, drug resistance and metastasis (57, 58). Bone marrow is also serving as survival place for breast cancer cells, where they metastasize to remain dormant till a favorable condition is regained (59–62).

Currently, based on the available literature, it can be summarized that the cancer cells acquire self-renewal property and become CSCs, depending on the signaling network existing in the “CSC niche”. They can further acquire drug resistance and metastatic ability in response to the tumor microenvironment, generating CSCs, exhibiting drug resistance and/or metastasis. This drug resistant CSC or MIC will be responsible for the bad prognosis, and thus should be targeted. Such reporters identifying metastatic CSCs or drug resistant CSCs should be used for screening drugs in preclinical models. The following sections will briefly examine the pathways and the consistent intermediate molecules and markers that regulate drug resistance and metastasis.

If we extrapolate the classical stem cell properties to cancer, a CSC should be a cell that is G₀-arrested, and divides asymmetrically to give rise to a non-dividing cell and a dividing differentiating cell. The G₀-arrested, low cycling CSC population in cancer will not be responsive to classical chemotherapeutic agents that target dividing cells. These cells that overcome therapy can re-establish the tumor since they possess the tumor initiating property, showing the role of CSCs in drug resistance and relapse. In another scenario, cancer cells up-regulate drug resistant molecules in response to therapy to overcome the effect of the drug. Though a sub-population of it can acquire self-renewal ability and become CSCs, all these drug resistant population may not have the self-renewal ability to re-create the entire tumor to cause recurrence. Therefore, it is important to segregate general drug resistant population from the drug resistant CSCs. Here, we postulate that molecules and pathways that regulate both self-renewal and drug resistance are important in sustaining the drug resistant CSCs. Even though many CSC markers are shown to regulate chemoresistance, there are two molecules, ABCG2 and ALDH that directly regulate drug resistance. The other self-renewal molecules regulate drug resistance by up-regulating either multidrug resistant proteins or ALDH ( Figure 2 ).

Over-expression of drug efflux molecules like ABC transporter molecules is one of the strategies adopted by cancer cells to overcome chemotherapy. Although all the different ABC transporters can do drug efflux, ABCG2 is usually associated with embryonic stem cells, adult stem cells and CSCs, but not exclusive to them (63). These drug efflux molecules transport substrates across the membrane to reduce their intracellular concentration, where the substrates can vary from xenobiotics to a variety of chemotherapeutics (63). ALDH is an enzyme catalyzing the conversion of cellular aldehydes to their corresponding carboxylic acids, which helps in detoxifying the chemotherapeutic agents by oxidizing the aldehyde groups of the drug. Though all ALDH forms can do this enzymatic detoxification, the self-renewal related isoforms like ALDH1A1 can regulate drug resistance by the up-regulation of drug efflux molecules through retinoic acid (RA) signaling. Thus it is evident that ABCG2 and ALDH1A1 are useful in identifying drug resistant CSCs.

Similar to the hierarchical clustering of stem cells, progenitors and differentiated cells observed during development, distinct states CSCs with different stages of self-renewal and differentiation are predicted for cancer also. There are functionally heterogenic CSCs exhibiting quiescence or dormancy as well as CSCs with a proliferative capacity (66, 160–162). This heterogeneous is considered to be resulting from the CSC plasticity, which is the result of a reprograming initiated by stemness signals in “CSC niche”. The presence of dormant CSCs was noted in several cancers and they are responsible for the relapse of the disease as these cells evade therapy and cause metastasis (163). These dormant CSCs responsible for metastasis in breast cancer can be marked with OCT4GFP reporter (38, 61). At the same time, the CSCs marked with ALDH1A1 are more proliferative in nature (164). ABCG2 is also implicated in the proliferation of stem cells (63). These proliferative CSCs also show multidrug resistance since they express high ALDH and drug efflux molecules. Additionally, CSCs with the hybrid EMT phenotype, marking MICs, express high ALDH (65). In a therapeutic approach, the dormant CSCs and the proliferative, drug resistant, and metastatic CSCs are critical because all these heterogeneous populations can lead to metastasis and recurrence. Considering all these factors, we advocate the use of two reporters, marking OCT4 and ALDH1A1, to identify all the CSCs populations.

For a high-throughput screening, we need to have reporters compatible with automated systems. Fluorescent and luciferase reporters are useful in quantifying the expression level in high-throughput platforms. But when analysis of the expansion or reduction of CSCs as a subpopulation is required, fluorescent reporters may be preferred over luciferase reporters. When we need to take this to preclinical animal models, luciferase reporters and fluorescent reporters of near infrared region, compatible with small animal imaging system, are extremely useful in tracing dormant drug resistant CSCs and metastatic CSCs.

As we discussed in the introduction, the success of a drug in the preclinical screening depends on the screening assay we use. A screening based on the dual reporter will ensure that we consider all the heterogenic CSC populations relevant for recurrence. Using this simple and high-throughput adaptable strategy, we can screen a large number of molecules for CSC-targeting efficiency. The most efficient drugs we obtain after screening has to go through the classical in vivo self-renewal assays, because some of the non-CSCs escaping the drug may become CSCs with the influence of a “CSC niche” persisting in vivo. The effect of this plasticity has to be taken in to account in drug screening. The introduction of dual reporters for drug screening can tremendously help to increase the number of molecules that enter the drug trial, so that the chance of getting a successful CSC-targeting drug is enhanced.

AM collected the data and drafted the manuscript. RR, GM, and PP contributed to the tabulation and graphical representation of the data. TT conceived the idea and critically revised the manuscript. All authors contributed to the article and approved the submitted version.

AM received financial support from University Grant Commission, India (19/06/2016 (1) EU-V/318501) and RR received funding from Council of Scientific & Industrial Research, India (09/716(0188)/2019-EMR-I). This work was supported by Department of Science and Technology, India (EMR/2015/001170).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.