CSC cell-mediated drug resistance of cancers is a major cause leading to the failure in cancer therapies. Resistance to molecular targeted drugs of NSCLC is a common characteristic of CSC cells. Among EGFR non-T790M mutation NSCLC cases, approximately 20% of NSCLC patients with CSC features come into being a more resistant phenotype during12 months after an initial response to the EGFR-TKIs . Since potential relapse of NSCLC may occur due to the enrichment of CSCs following TKI initial therapy, the generation of more effective therapeutic interventions based on CSC cell-mediated resistance of NSCLC to EGFR-TKIs is an urgent requirement .
There is increasing evidence to show that ALDH1A1bright/CD44high CSC-mediated EGFR-TKI resistance may be a major obstacle for EGFR-TKIs maintenance therapy of NSCLC. Loss of responsiveness to EGFR-TKIs in NSCLC with non-T790M EGFR mutation can be explained in terms of EGFR-TKI-resistant ALDH1A1bright/CD44high CSC that evolutionally possesses drug resistance and is often referred to as a tumor-initiating cell and associated with EGFR-TKIs non-responder [5–10]. Recent studies showed that ALDH1A1bright CSCs promote EGFR-TKI resistance in NSCLC [13, 17]. With respect to CD44, it has been reported that modulation of CD44 is detrimental to CSCs self-renewal and differentiation and NSCLC cells expressing CD44 are enriched for stem-like properties, suggesting that ALDH1A1bright/CD44high CSC linked to tumor progression and EGFR-TKI resistance is associated with a significantly poor prognosis factor in NSCLC . Several clinical studies revealed significantly increased proportions of ALDH1A1bright NSCLC cells displaying resistance to EGFR-TKIs and chemotherapy drugs [26, 31, 32].
Differentiating CSCs may provide such an approach to modulating or converting the phenotypes of CSCs for sustained treatment response of NSCLC to EGFR-TKIs, although the mechanisms underlying CSCs contribution to resistance of NSCLC to TKIs remain unclear .
The ALDH-retinoic acid pathway plays an important role in differentiation of CSCs. It has been shown that treatment of lung adenocarcinoma A549 cell with ATRA led to the downregulation of ALDH1A1. RA can reduce the ALDH activity and CD44 expression, thus affecting cell proliferation, cancer invasiveness and sensitivity to various chemotherapy drugs [24, 35, 36]. ALDH1A1 has been shown to convert/oxidize retinaldehyde into retinoic acid (RA) in several tissues, and to be one of the target proteins of ATRA . Treatment with ATRA increased the C/EBP homologous protein (GADD153) and GADD153-CCAAT-enhancing binding protein-β (C/EBP-β) interaction resulting in a decreased cellular availability of C/EBP-β for binding to the Raldh1 CCAAT box and high ATRA levels inhibit Raldh1 gene expression by sequestering C/EBP-β through its interaction to GADD153 [37, 38]. CD44 expression was highly responsive to ATRA as it was down regulated following treatment. ATRA treatment also resulted in decreased migration and invasion of cancer cells and promoted tumor regression by inducing differentiation .
Retinoids prevent the development of several tumors and enhance the efficacy of cytotoxic drugs such as cisplatin and docetaxel [39, 40]. Retinoids bind to specific nuclear receptors, which function as transcriptional regulators controlling the expression of numerous genes. The retinoid X receptors (RXRs) and retinoic acid receptors (RARs) are selectively expressed in ALDHbright CSCs, indicating RA signaling mainly occurs via ALDHbright CSCs of lung cancer, which provides a mechanism to selectively target CSCs . RA signaling is modulated by two classes of nuclear retinoid receptors, RARs and RXRs. Both RXRs and RARs interact with multiple co-activator and co-repressor proteins to promote increased cell stemness or cell differentiation. Retinoic acid showed feedback inhibition of the ALDH1 gene through RARα and C/EBP-β . Specifically, RA signaling regulates ALDH via the binding of ATRA to RXR and RAR that transcriptionally control ALDH gene expression .
Loss of retinoid receptors expression happens frequently in the development of carcino- genesis and induction of resistance to apoptosis. The known effect of ATRA on differentiation of cells is mediated through RARβ. RARβ belongs to the nuclear receptor (NR) superfamily of transcription factors. Upregulation of RARβ within the drugresistant cancer cells, which exhibits loss of RARβ expression, has been shown to increase the susceptibility of cells to apoptosis induced by chemotherapeutic agents. Activation of RARs or RXRs contributes to induction of RARβ, growth inhibition and apoptosis by retinoids. It evidenced that the therapeutic anti-CSC and proapoptotic effects of ATRA are dependent on receptor class-selective retinoids and the expression of RARβ plays a role in mediating retinoid response in NSCLC cells. [38, 42]. The loss of RARβ might contribute to enhanced cancer stemness and the apoptosis resistance of CSCs to gefitinib in NSCLC cells. ATRA can induce the apoptosis of NSCLC CSCs through activation of RARβ and its ability to down-regulate the CSCs markers in lung cancer cells [42, 43]. The expression of RARβ as well as RXRβ was reported to be downregulated in NSCLC, which enabled the cancer cell to evade apoptosis [45, 46].The RARβ is also known as tumor suppressor and the major target gene of retinoid action, and an enhanced level of RARβ protein exhibited its growth inhibitory action of lung cancer cells [47, 48]. RARβ can mediate retinoid action in lung cancer cells by promoting apoptosis. However, a fundamental question that remains unanswered is how ATRA and RARβ trigger apoptosis in lung cancer cells. Studies showed the overexpression of RARβ was accompanied by an increase in c-Myc and Bax but not TP53 protein expression and associated with an increase in the Bax/Bcl2 ratio, and that ATRA enhanced G1 growth arrest, up-regulated p21and p27 and downregulated cyclin D1. These data suggest that the expression of RARβ is positively associated with ATRA-induced apoptosis and growth inhibition in lung cancer cells [38, 42–45]. It has been shown that RA inhibits EGFR expression at the transcriptional level by targeting the EGFR promoter leading to inhibition of lung cancer cell growth and arrests EGFR-TKI resistant NSCLC cells in the G0/G1 phase of the cell cycle by altering the expression of GATA-binding factor 6 (GATA6) and inhibits the activation of two important pathways involved in lung cancer progression namely EGFR and Wnt signaling to overcome TKI resistance . Combinatorial treatment of retinoids with EGFR-TKIs drugs in EGFR-TKIs resistance lung cancer cells promotes the activation of GATA6 and then inhibits the activation of EGFR/Wnt signaling pathways and favors the association of RXR, RARβ, and cellular retinoic acid binding protein-2 (CRABP2). This complex inhibits the proliferation and promotes the differentiation of lung tumor cells via inhibiting activating protein-2 (AP-2), which result in re-sensitization of EGFR-TKIs resistant lung cancer cells .
In this study, a short-term gefitinib treatment was used to enrich A549GSCs and H1650GSCs. FCM assay showed that A549GSCs and H1650GSCs have a significant increase in proportions of ALDH1A1bright/CD44high cells (Fig. 2), and we further confirmed that these ALDH1A1bright/CD44high GSCs exhibit increased IC50 values for gefitinib compared to that of their respective parental cells (Table 1 and Fig. 1A and B ), and are involved in CSCs but not in EGFR T790M-mediated gefitinib resistance (Supplementary Fig. 1), suggesting that ALDH1A1bright/CD44high CSCs in NSCLC/ADC contribute to resistance to gefitinib. Interestingly, treatment with ATRA significantly reduced ALDH1A1 and CD44 expression of A549GSCs and H1650GSCs, and their IC50 values for gefitinib, thus returning to sensitization to gefitinib (Table 1 and Fig. 1A and B). These results showed that in contrast to the known tendency of EGFR-TKIs, such as gefitinib, to target the non-stem-like ALDH1A1-negative cell population, ATRA can modulate the ALDH1A1bright/CD44high cell population in NSCLC/ADC. Therefore, the synergistic antitumor effect of ATRA in combination with gefitinib might be a promising therapeutic strategy to prevent or re-sensitize CSC-mediated gefitinib resistant NSCLC/ADC [49, 50].