Nowadays, there are CSC markers that are widely used to identify several tumor types. Such markers have been reported in CSC-enrichment culture models from cell lines or primary cultures derived from patient samples and serial xenotransplantation of putative CSCs in mouse models, which must be able to recapitulate the original heterogeneous populations and be directly validated in human tumor samples. It is important to note that the use of a single marker to define a CSC population is not recommended. For this purpose, a phenotypic profile that combines various markers should be established, as well as carrying out self-renewal assays (Figure 1)[2,25].

CD133, also known as prominin-1, is a transmembrane cell surface glycoprotein traditionally used as a hematopoietic stem cell marker that is effective for detection of non-stem cells from various tumor and tissue samples. The Dirks laboratory used the CSC marker CD133 for brain CSC identification. The purified CD133⁺ population from primary human brain tumors samples showed higher proliferation and self-renewal capacity in neurosphere formation assays than CD133^- cells[10]. Moreover, the inoculation of only a few CD133⁺ cells was sufficient to produce a tumor, which was then successfully transplanted[11]. In 2013, the Pelicci laboratory reported that CD133 was found in an interconvertible state in glioblastoma patient-derived neurospheres and that the use of short hairpin RNA (shRNA) against CD133 diminished their self-renewal and tumorigenicity potential[18]. Interestingly, some studies have proposed that CD133 could maintain CSC properties through the Wnt/β-catenin signaling pathway[38].

CD133 has also been tested in colorectal cancer cell lines and tumor tissue samples[39,40] through the use of various techniques, including flow cytometry and serial xenotransplantation in mice[41]. Additionally, CD133⁺ CSCs have been reported in many other solid cancer models, including endometrial cancer[42], lung cancer[43], small cell lung cancer[44], laryngeal cancer[45,46], liver cancer[47], colorectal cancer[48], and gastric cancer[49].

CD133 has been found in samples that represent higher stage tumors and are predictors of poor prognosis. For this reason, CD133 is considered a promising therapeutic target. This year, a phase I trial for testing the efficacy of CD133-directed CAR-T cells showed that CD133⁺ cells were successfully eliminated after CART-133 infusion[50].

CD44 is a multifunctional glycoprotein involved in cell adhesion, signaling, proliferation, migration, hematopoiesis, and lymphocyte activation[51]. It functions as a receptor for hyaluronan and other extracellular matrix components[52]. CD44 is widely used as a CSC marker, especially for tumors of epithelial origin, and it is used alone or in combination with CD24 for the identification of breast CSCs[5]. CD24 is a small surface protein that is found in many tumor types. However, reports from cancer cell lines show that there is a substantial variation in CD24 expression even among the same tumor types[53].

Though CD24^- cells are commonly associated with CSC phenotypes, there are some cases in which CD24⁺ has been found to be a marker for cell populations with CSC features. For example, in nasopharyngeal carcinoma (NPC) cell lines[54] and in HPV-16 SiHa cervical cancer cells, isolated CD44⁺CD24⁺ cells were radioresistant and more tumorigenic than those negative for the same markers[55]. The same CD44⁺CD24⁺ phenotype was used to identify gastric CSCs[56].

A known classic publication demonstrated that only a small population isolated from breast tumors, defined as CD44⁺CD24^-/low, has the capacity to sustain tumor growth in NOD/SCID mice and generate heterogeneous cell populations as the original breast tumor[5]. Later, in human prostate cancer samples, CSCs characterized through immunofluorescence with the CD44⁺/β₂β₁^hi/CD133⁺ phenotype were identified and characterized[57]. The next year, CD44⁺ prostate cancer cell populations were obtained[58]. Also, CD44 and CD133 expression was evaluated in gastric adenocarcinoma tumors by immunohistochemistry, and it was found that both markers could be correlated with clinical and pathological parameters[51].

Although CD44 is widely reported as a CSC marker, it is very important to note that it is a ubiquitously expressed molecule derived from a gene with 18 exons. When all variable exons are spliced out, the standard form (CD44s) is expressed, and when alternative splicing occurs, variant forms (CD44v) are expressed[59]. In spite of this, there are only a few reports in which CD44 isoforms are considered when evaluating CSCs. In 2005, Mackenzie and his group demonstrated the existence of two CSC populations, both expressing CD44^high (and CD44⁺), derived from head and neck cutaneous squamous cell carcinoma. One was associated with EMT properties and the other one possessed an epithelial phenotype[60]. They demonstrated that the CD44^high cells that undergo EMT preferably expressed the CD44s isoform; while the epithelial CD44^high cells expressed the CD44v isoform. Using RNAseq, another group later confirmed these results. The CD44v6 isoform was identified as the predominant isoform in a prostate cancer epithelial cell line[61].

A very important contribution from the Mackenzie laboratory is that they demonstrated that the use of enzymes (for example, trypsin or collagenase) for cell extraction from tissues caused destruction of cell surface CD44v isoforms, leaving only the CD44s isoform[62]. Moreover, CD44-specific antibodies are not able to distinguish between all isoforms. Specifically, in breast cancer, CD44v was found to be associated with better prognosis while CD44s was related to poor prognosis[63]. As a consequence, CD44 is the most frequently found CSC marker[64,65]. Other examples are found in colorectal cancer, in which CD44 was found together with CD133[66,67], head and neck squamous cell carcinoma[68,69], ovarian CSCs[70], and gastric cancer using the specific isoform CD44v8-10[71].

CD49f or integrin α6, is a transmembrane glycoprotein that functions as the receptor for the extracellular matrix protein laminin[72,73]. CD49f is widely distributed in stem cells and in the brain[73]; because of its role in tumor cell proliferation, survival, self-renewal and tumor growth, it has been proposed that it could be used as a CSC marker[73].

In sphere colony forming cell culture using prostate cancer cells, CD49f was shown to be a better marker than CD133 and CD44[74]. In gastric cancer, CD49^high cells displayed CSC characteristics, including resistance to doxorubicin, 5-fluorouracil and doxifluridine[75].This has also been reported in breast[76] and colon cancer[77].

Besides the examples mentioned above, there are other surface markers that have been proposed as CSC markers, such as CXCR4 and LGR5, among others.

Another strategy for CSC identification and purification is the use of functional or intracellular markers (Figure 1), which are considered to be more stable than surface markers. The principal functional CSC marker is aldehyde dehydrogenase or ALDH, part of an enzyme superfamily encoded by 19 genes that metabolize endogenous and exogenous aldehydes. It is present in practically all organisms, and its levels and isozymes vary depending on tissue and organ[78].

For ALDH identification, specific ALDH antibodies are available; nonetheless, we suggest that the most appropriate way for ALDH identification is the measurement of its activity using the commercial ALDH fluorescent substrate ALDEFLUOR^® kit assay by Stem Cells Technologies, Inc. (Vancouver, BC, Canada). Cells that display high ALDH activity, (named ALDH^high or ALDH⁺ or ALDH^br), can be identified and isolated using flow cytometry[79]. Several works have shown that high ALDH activity is often associated with CSCs derived from solid tumor types[80]. These cells are generally characterized by a higher proliferation potential, colony-forming capacity, self-renewal, in vivo tumorigenic capacity, metastasis, and drug resistance. For instance, ALDH^high CSCs have been identified in colon cancer[81,82], lung cancer[83], cervical cancer[14,84,85], breast cancer[86], pancreatic cancer[87,88], and melanoma[89,90], to mention some examples.

As for surface markers, ALDH is often reported in combination with other cell markers to increase the accuracy of CSC validation. In some cases, high ALDH activity is found together with high expression of markers like CD133. Some cases have been identified in ovarian cancer[91,92], invasive ductal breast carcinoma tumors[93], and lung cancer[94]. The combination ALDH⁺/CD44⁺ has been evaluated in various tumors such as breast cancer[95] and lung cancer[96].

Despite the broad variety of CSC publications in the last years, the discovery of effective therapies has remained elusive. However, some advances have been made in the field that could be getting us closer to direct CSC elimination. A brief outline of some of these strategies is showed in Figure 2.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Drugs that may target cancer stem cells. Promising therapeutics to treat cancer patients. The flowchart highlights the new and more promising cancer therapies that can be directed toward cancer stem cells to eliminate them. CSC: Cancer stem cell.

Targeting deregulated pathways in CSCs aims at developing effective strategies against CSCs. In adult pancreas, the Hedgehog (Hh) signaling pathway is dormant, but it is upregulated in pancreatic ductal adenocarcinoma, specifically in CD44⁺/CD24⁺/ESA⁺CSCs. In a phase I study, 68 patients were treated with GDC-0449 or Vismodegib, a Hh pathway antagonist[112], alone or in combination with gemcitabine. GDC-0449 inhibited Hh signaling, but there was no correlation with survival or other parameters[113]. Other drugs that show promising results in inhibiting this pathway are PF-044449913[114] and thiostrepon, which attenuates CD44⁺/CD24^- triple-negative breast CSCs[115].

In addition, γ–secretase inhibitors target the Notch pathway and possess a stronger anti-neoplastic activity when combined with chemotherapeutic agents[116]. Nevertheless, adverse effects have been reported, as patients developed cutaneous rash in phase I clinical trials[117,118].

Several drugs that aim to inhibit the Wnt/β-catenin signaling pathway are being developed. One such drug is Celecoxib, a non-steroidal anti-inflammatory drug that inhibits β-catenin signaling by cyclo-oxygenase (commonly known as COX)-dependent and COX-independent mechanisms[116]. This drug downregulates CD133 expression in colon cancer cells by inhibiting Wnt signaling[119] and intestinal cancer growth[120]. The Wnt inhibitor LGK-974 inhibits porcupine, an O-acyltransferase required for Wnt secretion. In liver cancer cells, LGK-974 blocks secretion of the Wnt3A protein, and as a consequence, cells become more sensitive to radiation[121]. A recent study showed that LGK-974 downregulates ALDH1A3 and reduces chemoresistance in glioblastoma cells[122].

Curcumin is an antioxidant derived from turmeric whose anti-cancer effect is well documented. Referring specifically to CSCs, curcumin has shown the potential to regulate the CSC self-renewal pathways, as well as specific microRNAs[123]. In CD133⁺ lung CSCs, curcumin suppresses the activation of Wnt/β-catenin and Shh pathways, as well as other CSC traits[124]. It has been demonstrated that in bladder cancer, curcumin suppress the Shh pathway[125] and in laryngeal carcinoma treatment, curcumin enhances the effectiveness of cisplatin, reducing CD133⁺ cells in vitro[46]. Additionally, a combination of curcumin and FOLFOX chemotherapy inhibits colorectal CSCs in ex vivo models[126].

An interesting strategy is to target CSCs using nanoparticles to reduce side effects on surrounding normal cells. In 2015, construction of glucose-coated gold nanoparticles (Glu-GNPs) that used glucose to facilitate GNP entry into leukemic stem cells overexpressing CD44 (TH1-P) was reported. Leukemic cells were cultured for one hour in the absence of glucose for better Glu-GNP uptake, and then X-ray irradiation tests were performed. Results showed that Glu-GNPs enhanced cell death compared to either irradiation or GNPs alone[127]. Formulated mangostin-encapsulated poly(lactic-co-glycolic acid) nanoparticles (Mang-NPs) successfully downregulated the known stemness genes c-Myc, Nanog and Oct4, two CSC markers, CD24 and CD133, and the Shh pathway[128]. Salinomycin and paclitaxel nanoparticles are also being used to eliminate breast cancer cells including CD44 breast CSCs[129].

Interestingly, CSCs have a strict dependence on mitochondrial biogenesis. Five classes of FDA-approved antibiotics that inhibit mitochondrial biogenesis were used on eight different cancer cell lines, and the results suggested that the observed therapeutic effects were infection-independent[130]. Clinical trials using doxycycline showed positive results in cancer patients[131]. Another drug that has been shown to specifically eliminate CSCs is metformin, and its effects are enhanced when it is used in combination with doxorubicin[132]. Moreover, it has been observed that metformin reduces metastasis by targeting both EMT and CSCs[133]. In the ovarian cancer cell line SKOV3, low doses of metformin diminished CD44⁺CD117⁺ CSCs in xenograft tissue and enhanced the effect of cisplatin[134]. In esophageal cancer, metformin reduced the number of ALDH+ cells, tumor growth in vivo[135], and in pancreatic cancer, it increased radiation sensitivity[136].

Using antibodies is another strategy to block CSC signaling pathways and reduce tumor activity in different models. For instance, the anti-DLL4 (Enoticumab) antibody that targets the dominant Notch ligand DLL4 has shown anti-tumor activity, especially in VEGF-resistant tumors in human phase I studies[137]. Furthermore, another anti-DLL4 antibody (Demcizumab) is effective in decreasing tumor size but produces hypertension[138]. In colon cancer patients, increased progastrin levels in the blood have been observed, which is a tumor-promoting peptide that participates in colon CSC self-renewal and is also a direct target gene of β–catenin/Tcf4. Based on this information, specific anti-progastrin antibodies have been developed and tested in colon cancer cell lines and in mice. The antibodies, alone or in combination with chemotherapy, decreased self-renewal, migration and invasion. Moreover, they mitigated Wnt-driven intestinal neoplasia and induced tumor cell differentiation in vivo[139]. H90 is a mouse IgG1 mAb against human CD44 that directly targets CSCs to induce differentiation and proliferation in AML xenograft mouse models[140]. Additionally, anti-CD44s-specific antibodies are effective in eliminating pancreatic stem cells[141]. For more extensive information about antibodies against CSCs, we recommend reference[142].

ALDH is an important CSC marker that is overexpressed in several cancers. Specific ALDH inhibitors are effective in modulating cell growth, apoptosis and differentiation. Additionally, increased chemo- and radio-sensitivity is usually observed. All-trans retinoic acid (commonly known as ATRA) is a first generation systemic retinoid that promotes cell differentiation[143,144] and has been used in clinical trials[145]. ATRA has also been tested in breast cancer cells[106,146,147] and in gastric cancer, where it inhibited tumor growth[148], and in head and neck cancer, where it suppressed Wnt/β-catenin signaling[149]. In a phase I/II trial, advanced breast cancer patients did not show a significant improvement when treated with ATRA and tamoxifen compared with tamoxifen alone[150].

Disulfiram is a drug used for treating alcoholism, and it shows anti-cancer activity in vitro and in vivo, further potentiating the chemotherapeutic response. Its effectiveness has been demonstrated on paclitaxel-resistant triple-negative breast cancer cells[151], in non-small cell lung cancer cells[152], and glioblastoma.