Modeling of cancer stem cells and the tumor microenvironment Via NT2/D1 cells to probe pathology and treatment for cancer and beyond

Abstract

Introduction

Unique from the other tumor cells, tumorigenic cancer stem cells (CSCs) manifest as a subpopulation of cells within the tumor that exhibit genetic and phenotypic features and signaling processes, which escape traditional anti-oncogenic treatments, thereby triggering metastases and relapses of cancers. Critical to cancer biology is the crosstalk between CSCs and tumor microenvironment (TME), implicating a CSC-based cancer immunotherapy. Cognizant of CSCs’ significant role in cancer pathology and treatment, finding a biological model that recapitulates CSCs and TME may allow a better understanding of tumor onset and progression for testing CSC-based therapies. In this review paper, we examined the CSC and TME characteristics of the human embryonal carcinoma NTERA-2 clonal cell line called NTERA-2 cl.D1 or NT2/D1 cells and discussed their potential utility for research and development of treatments for cancer and central nervous system (CNS) disorders.

Methods

To probe our hypotheses that NT2/D1 cells display CSC and TME properties key to tumor development, which can serve as a screening platform to test cancer and CNS therapeutics, we conducted a literature review over a 10-year period (2014–2024), focusing on PUBMED and Science Direct published articles on cellular models of cancer, with emphasis on milestone research discoveries on NT2/D1 cells relevant to CSCs and TME. We categorized the studies under pre-clinical and clinical investigations in supporting the existence of CSC and TME features in NT2/D1 cells and providing a laboratory-to-clinic translational basis for cancer and CNS therapeutics.

Conclusions

NT2/D1 cells stand as a feasible biological model that recapitulates the crosstalk of CSCs and TME, which may critically contribute to our understanding of cancer and CNS biology and therapeutics. Designing therapeutics against CSCs' distinct self-renewal and differentiation capacities within the TME opens new avenues for treating cancers and CNS disorders.

Graphical Abstract

Introduction

Cancer is one of the most common causes of premature death, affecting more than 20 million people worldwide [1, 2]. The number of cancer patients is predicted to double over the next 50 years due to population aging [3]. Amidst the search for effective cancer therapy, cancer stem cells (CSCs) have been proposed as a new model of tumor initiation and a potential target for novel treatment. CSCs refer to a subpopulation within the tumor mass that exhibits unique genetic and phenotypic features similar to traditional stem cells. CSCs largely resist conventional chemotherapy, resulting in relapses and metastases [4–6].

Therapeutics targeting CSCs'distinct self-renewal and differentiation capacities expand possibilities for cancer treatment [7, 8]. In the past decades, there has been accumulating evidence demonstrating CSCs'mechanisms [4]. Preclinical studies showing positive results from targeting CSCs’ signaling pathways have pushed forward the clinical translation of CSC-based therapies [9, 10]. Crosstalk between CSCs and the tumor microenvironment (TME), including extracellular matrix and immune cells, is critical to cancer biology [11–13]. Recognizing CSCs’ essential role in cancer development and their pivotal potential for therapeutic development, the establishment of a biological model that recapitulates CSCs and TME will facilitate a better understanding of tumor initiation and progression, while also serving as a screening platform for CSC-based therapies.

In this review article, we explore the potential of the human embryonal carcinoma NTERA-2 clonal cell line, known as NTERA-2 cl.D1 or NT2/D1 [14, 15], which satisfies both CSC and TME properties. Equally important, NT2/D1 cells have been demonstrated to differentiate into non-cancerous cells under proper post-mitotic treatments [16, 17], thereby providing novel insights and treatment strategies to induce cell cycle exit in CSCs. Indeed, compelling evidence indicates that NT2/D1 cells primed to commit towards the neural lineage can be used biological models as well as cell transplantation source for central nervous system (CNS) disorders [18–23]. This review paper addresses two key questions: (1) Do NT2/D1 cells display CSC and TME properties suitable for modeling tumor development? (2) Can NT2/D1 cells be utilized as a screening platform for cancer therapeutics? Moreover, we also review the application of NT2/D1 cells as biological models and treatment screening platforms for diseases beyond cancer, i.e., CNS disorders to fully elucidate the applications of the NT2/D1 cells (Fig. 1). Our paper’s outline consists of reviewing the biological models of CSCs, introducing NT2/D1 cells as potential platform in investigating the crosstalk between CSCs and TME, and discussing NT2/D1 cells as tool for developing cancer therapeutics. Throughout the paper, our main scope focuses on cancer biology and therapeutics, with the objective of advancing NT2/D1 cells as a potent biological model of CSCs and TME.

Fig. 1.

NT2/D1 cells as a biological model and treatment screening platforms for cancer and beyond. NT2/D1 cells possess the unique properties of CSCs and have been utilized as a treatment screening platform for CSC-based cancer therapy, particularly representing chemotherapy-resistant tumors. NT2/D1 cells are also able to differentiate into neurons and astrocytes under specific induction condition, serving as convenient biological models and drug screening platforms for multiple neuropsychiatric disorders

Biological models of CSCs

The pathogenesis of cancer consists of eight hallmarks: sustaining proliferative signaling, evading growth suppressors, resisting cell death, enabling replicative immortality, inducing angiogenesis, activating invasion and metastasis, dysregulating cellular metabolism, and avoiding immune destruction [24]. Recently, two additional enabling characteristics, genome instability and tumor-promoting inflammation, have been included in this model [25]. CSCs, first discovered in 1877, are a subpopulation of tumor mass that originate from either differentiated cells or tissue-resident stem cells [26, 27]. CSCs drive tumor initiation as well as relapses, and they are frequently associated with aggressive, heterogeneous, and therapy-resistant tumor variants [28].

Widely recognized as a critical factor in tumorigenesis and progression, TME comprises diverse populations of cancer cells, including CSCs, and stromal cells [25]. Over the past decade, accumulating evidence has postulated mechanisms by which CSCs modify the TME, such as recruitment of macrophages, myeloid-derived suppressor cells (MDSCs), and regulatory T cells (Treg) via chemokine (C–C motif) ligand 2 (CCL2)-CCR2, colony stimulating factor 1 (CSF1)-CSF1R and IDO1 signaling [25, 29]. Other cytokines involved in CSC-TME crosstalk include interleukin (IL)−1β, IL-6, transforming growth factor β (TGFβ), and signal transducer and activator of transcription 3 (STAT3) [29]. A reciprocal interaction exists between CSCs and infiltrating immune cells in the TME, suggesting that these immune cells may serve as specific markers for CSCs [29]. In tandem, CSC-immune cell communication appears as an attractive signaling pathway [29] for cancer biology.

The first CSC model was developed in human myeloid leukemias [30]. In the 1990 s, cancer cell subpopulations were found to demonstrate a functional hierarchy similar to regular stem cells [31, 32]. In acute myeloid leukemia (AML), only a small subpopulation of leukemic cells, characterized as [CD34 +, CD38-] cells and later termed leukemic stem cells (LSCs), can engraft in non-obese diabetic severe combined immune deficiency (NOD/SCID) mice. These LSCs, with surface markers resemble normal multipotent progenitor cells, reproduce the phenotypic heterogeneity of the original leukemia donor once engrafted. However, CSC populations reveal heterogeneity as not all AML subtypes show the same CSC markers. For instance, acute promyelocytic leukemia (APL or AML-M3) predominantly consists of [CD34 +, CD38 +] cells rather than typical [CD34 +, CD38-] LSCs, suggesting that APL may originate from different CSCs [33]. Self-renewal genes in normal stem cells may function differently in CSCs. For instance, knockout of the tumor suppressor phosphatase and tensin homolog (PTEN) gene reduced self-renewal capacity in hematopoietic stem cells (HSCs) but enhanced spontaneous development of acute leukemias [34, 35]. In contrast, the BMI1 oncogene maintains self-renewal capacity in both HSCs and leukemia models [36, 37]. Collectively, these findings suggest that the two hallmarks of cancer, namely sustaining proliferative signaling and aberrant tendency, are independent properties. To this end, probing the cell survival and proliferative genes and transcription factors, such as BM1 in CD34 +, CD38- cells may reveal novel markers and signaling pathways associated with tumor biology [33, 36].

CSCs have also been detected in diverse types of solid tumors, including breast cancer, brain cancer, prostate cancer, lung cancer, melanoma, and multiple myeloma [30]. In human breast cancer, only a small subpopulation, labeled as [CD44 +, CD24 −/low] cells, can sustain tumor growth when transplanted in NOD-SCID mice, replicating the phenotypic diversity of the original tumor [38]. CD133 + cells have been identified as tumorigenic cells in glioblastoma multiforme and medulloblastoma [39–42], while CD44 cells are proposed to be CSCs in prostate cancer [43, 44], and CD138- cells exhibit high engraftment capability in multiple myeloma [45]. Based on these studies, identifying CSCs with CD138- B phenotype may reveal stem cells with the capacity to replicate and eventually become malignant CD138 + tumor cells [45].

Biological models accurately representing CSCs are indispensable to understanding intratumor heterogeneity, metastasis, and treatment-resistant tumors [30]. CSCs establish a new perspective in cancer biology, that the heterogeneous subpopulations of tumor cells may arise from CSCs at various differentiation stages, not only from genetic variation and environmental influences [46, 47]. Targeting pathways mediating CSC function, proliferation, and differentiation may be pivotal for effective tumor eradication. The conventional model of cancer proposes that metastases arise from monoclonal expansions of distinctive tumor subclones, resulting in metastatic tumors with significant differences in genotype and phenotype compared to the primary tumors. On the contrary, the CSC model posits that since primary and metastatic tumors have consistent intratumor heterogeneity and differentiation antigen expression, they should experience a similar differentiation paradigm possibly from CSCs [46, 48, 49]. CSCs have significant implications for cancer treatment design by introducing novel targets that are a small subset of cancer cells with stem cell-like properties [50, 51]. Treatments that fail to effectively eradicate CSCs may be unable to prevent relapse or metastasis. In addition, the CSC model may also explain why some novel cancer therapies, despite promising preclinical results, show limited clinical efficacy, possibly due to unsuccessful CSC targeting [30]. CSC-based therapy may suppress the tumors’ proliferation capacity, resulting in gradual tumor shrinkage, and can be used in combination with conventional treatments that target tumor bulk [30]. Accordingly, CSC-targeting using stem cell markers and signaling pathways may pose as a potent cancer therapeutic.

With CSCs’ pivotal role in cancer biology and their potential as a target for novel treatment development, establishing a biological model that recapitulates CSCs and TME will enable a better understanding of tumor initiation and progression, while also serving as a screening platform for CSC-based therapies. However, despite the promising potential of CSC-based treatments, biological models of CSCs are not yet fully established, warranting further exploration to find suitable platforms.

NTERA-2 clonal D1 (NT2/D1) cells

NT2/D1 cells are a human teratocarcinoma cell line derived from a testicular germ cell tumor [52, 53]. NT2/D1 cells possess the unique property among germ cell tumors that they differentiate exclusively into postmitotic, neuron-like cells after treatment with retinoic acid. Given the ease of differentiation, NT2/D1-derived neurons (NT2/D1-N) have been widely utilized for modeling neurological disorders. Interestingly, since the only differentiated phenotype of NT2/D1 cells is neuronal, undifferentiated NT2/D1 cells have been proposed to represent neuronal progenitor cells [54], providing a valuable tool for the investigation of CNS development.

As NT2/D1 cells are derived from germ cell tumors and exhibit embryonic markers such as nestin and SOX2, they can also be utilized as a biological model of CSCs [55, 56]. NT2/D1 cells are also capable of interacting with TME, an ability essential for CSCs. Examples of crosstalk between NT2/D1 cells and TME include extracellular matrix modification, recruitment of macrophages and fibroblasts, along with enhanced angiogenesis [57–59].

NT2/D1 cells, satisfying properties of both CSCs and TME, are a promising biological model to represent CSCs. Investigating the signaling pathways underlying NT2/D1 cells, particularly self-renewal and differentiation, will provide insights into CSC-based therapy, namely, inducing CSCs to exit the cell cycle through differentiation. Equally important, experiments focusing on the crosstalk between NT2/D1 cells and TME members will elucidate how CSCs interact with the surrounding environments and how to attenuate the modification. This timely field of research may reveal novel therapeutic targets that could reduce the rate of tumor growth and invasion.

NT2/D1 cells as a biological model of CSCs

In the past decade, laboratory investigations have demonstrated multiple signaling mechanisms in the NT2/D1 cells, which may also represent those in CSCs (Table 1). Transcription factors SOX2 and OCT3/4 play critical roles in tumor cell proliferation, migration, and invasion [82, 83]. Both in vitro and in vivo studies have shown elevated expression levels of SOX2 and SOX2 overlapping transcript (SOX2OT) long noncoding RNAs (lncRNAs) in NT2/D1 cells [76, 83]. Another In vitro study further demonstrated that SOX2 influenced NT2/D1 cells’ migration speed, mode, and path, but not cell adhesion [82]. Anterior gradient-2 (AGR2) was also found upregulated in NT2/D1 cells, due to hypomethylation of the AGR2 gene, promoting cell proliferation, invasion, and glycolysis [65]. Knockdown of AGR2 reduced tumor proliferation, invasion, and glycolysis in vitro while lessening tumor growth in vivo. Mechanically, AGR2 binds to annexin A2 (AnXA2), subsequently increasing epidermal growth factor receptor (EGFR) expression. Additionally, the transcription factor homeobox A10 (HOXA10), essential for testis development, exhibited less expression in NT2/D1 cells [84]. Under physiological conditions, HOXA10 inhibits cell proliferation and delays cell cycle progression at the G2/M phase through TP53, c-Kit, STAT3, AKT, and ERK signaling pathways. Reduced HOXA10 expression in NT2/D1 cells may contribute to the self-renewal properties of CSCs. Genome-wide association studies have identified SPRY4 as another potential oncogene in NT2/D1 cells, functioning through activation of the PI3 K/AKT pathway [78]. The proliferation of NT2/D1 cells also relies on hormonal signaling, particularly estrogen. Estradiol (E2) activates estrogen receptor 1/2 (ESR1/ESR2) in NT2/D1 cells, promoting cell proliferation and viability [67]. Human epidermal growth factor receptor 2 (HER2) is endogenously produced by NT2/D1 cells and can suppress the Wnt signaling pathway, one of the crucial differentiation mechanisms of NT2/D1 cells [68]. Moreover, demographic characteristics, such as biological gender, age, microbiome, and metabolic factors, also influence the treatment resistance of CSCs. For instance, the DLK1-MEG3 locus, containing paternally imprinted gene DLK1 and maternally imprinted gene IGF2, has been found to drive tumorigenicity of NT2/D1 cells [80].

Table 1.

NT2/D1 cells as a biological model of CSCs

Intervention/animal model	Characterization	References
In vitro
17β-estradiol, ERβ-selective agonist (DPN, PPT)	- ↑Migration, invasion and anchorage‑independent growth	[60]
Activin A, BMP4	- Activin A → ↑pluripotency, TGF-β, Notch, TP53, and Hippo signalling - BMP4 → ↑TGF-β, pluripotency, Hippo and Wnt signalling	[61]
piR-36249, DHX36	- piR-36249 and DHX36 inhibit the malignant phenotype of testicular cancer cells by ↑OAS2 protein	[62]
phorbol 12-myristate 13-acetate (PMA)	- ↑miR-630 → ↓NANOG	[63]
Retinoic acid	- Opposite Insulin-like growth factor 1 receptor (IGF-1R) and miR-6165 pattern	[64]
No intervention	- ↑Anterior gradient-2 (AGR2) → ↑proliferation, ↑invasion, ↑glycolysis	[65]
	- SOX9/TRPC3 promotes proliferation and controls cell morphology	[66]
	- E2 activate ESR1/ESR2 → ↑number and viability	[67]
	- Endogenous HER2-miR1 found in NT2 → ↓Wnt signaling	[68]
	- ↓miR-196a-5p	[69]
	- ↓TIG1 and SPINK2 - TIG1 inhibited cell invasion, migration, and epithelial–mesenchymal transition (EMT) through uPA/uPAR signaling pathway - SPINK2 augments TIG1 functions	[70]
	- Short-term activation of WNT signaling induced a differentiation - WNT signaling is distinct among types of germ cell tumors	[71]
	- ↑HOTTIP → ↑cell proliferation - HOTTIP binds to miR-128-3p to regulate HOXA13 expression	[72]
	- ↑Citric acid cycle/mitochondrial oxidative phosphorylation - ↑Sphingolipid biosynthesis	[73]
	- Loss of functions of HOXA10 → ↑proliferation of testicular cancer cells	[74]
	- DNA methylation promotes PIWI-LIKE 2 expression	[75]
	- ↑SOX2OT expression	[76]
	- Oxidative stress increased LRWD1 expression through a Nrf2-dependent mechanism	[77]
	- SPRY4 and SPRY4-IT1 may be oncogenes through activation of the PI3 K/Akt pathway	[78]
	- OCT4 A and OCT4B2 have similar expression - ↑OCT4B2 transcription after heat shock	[79]
	- DLK1–MEG3 locus drives the tumorigenicity - 5-azaD suppresses DLK1	[80]
	- Activation of PPARβ/δ inhibits RAR/RXR dimerization and tumor proliferation	[81]
	- SOX2 overexpression promote speed, mode and path of cell migration, but not the adhesion ability - ↑Expressions of tumor suppressor protein TP53 and the HDM2 oncogene	[82]
	- ↑SOX2 expression - Down-regulation of SOX2 promotes apoptosis - Inhibition of OCT3/4 induces differentiation	[83]
	- NF-Y-induced inhibition of cell growth is p53-independent	[56]
In vivo
Mouse	- Keratin, vimentin, neurofilament proteins and desmin	[15]
	- ↑SOX2 expression	[83]

NT2/D1 cells interact with TME through diverse mechanisms. NT2/D1 cells exhibited increased extracellular matrix metalloproteinase inducer (EMMPRIN) mRNA expression, leading to elevated MMP-2 production in the surrounding fibroblasts and contributing to matrix-mediated tumor progression [57]. NT2/D1 cells also interact with macrophages and cancer-associated fibroblasts (CAFs) to remodel the extracellular matrix and evade immune destruction. CAFs promote tumor cell proliferation, expression of cisplatin resistance factors, and modification of macrophage polarization through effector molecules such as IGFBP1, LGALS3BP, LYVE1, and PTX3 [59]. The CXCR4/CXCR7/CXCL12 axis represents another way of communication between NT2/D1 cells and TME, influencing tumor cell proliferation, migration, metastasis, and resistance to conventional treatment [58]. In addition, increased stromal CD44 expression induced by NT2/D1 cells promotes angiogenesis to support tumor growth and metastasis [85].

NT2/D1 cells serve as a promising biological model of CSCs, satisfying both stem cell properties and TME modification capabilities. However, studies on signaling mechanisms in NT2/D1 cells are limited, with a major focus on neural differentiation rather than cancer pathogenesis. Moreover, nearly all experiments are conducted in vitro, lacking the influence of TME in actual in vivo conditions. Novel CSC-based interventions, such as AGR2 and HOXA10 modification, remain in the preclinical stage. Further investigations to identify potential therapeutic targets based on the CSC model are warranted.

NT2/D1 cells as a screening platform for cancer therapeutics

Over the past decades, numerous chemotherapies and immunotherapies have been developed against increasing cancer prevalence. While these conventional therapeutic strategies can eradicate the majority of tumor mass, they are not effective enough to deplete CSCs, potentially due to the remodeling capacity of CSCs to evade immune destruction [29]. For instance, acquired resistance to receptor tyrosine kinase inhibitors occurs at the epigenetic level and can be reversed with histone deacetylase inhibition [86]. The crosstalk between CSCs and TME also contributes to chemotherapy resistance through paracrine signaling. Moreover, CSCs stimulate macrophages to secrete lactadherin and IL-6, which in turn activate STAT3 signaling [87]. Given these unique mechanisms of chemotherapy resistance in CSCs, using biological models of CSCs to identify therapeutic targets and evaluate treatment efficacy will enable us to sequester cancer more effectively.

Germ cell tumors, the original source of NT2/D1 cells, are primarily treated with chemotherapy, particularly cisplatin. However, some patients develop resistance to cisplatin through multiple mechanisms. NT2/D1 cells are utilized as a treatment screening platform to overcome such resistance (Table 2). Cyclin-dependent kinase 5 (CDK5), nuclear factor erythroid 2-related factor 2 (NRF2), and regulatory factor X1 (RFX1) contribute to cisplatin resistance in NT2/D1 cells. Combination therapies with dinaciclib or palbociclib have been shown to reduce cisplatin resistance along with suppressing tumor cell viability and proliferation in NT2/D1 cells [88, 91]. NRF2 upregulates multidrug resistance (MDR) genes in NT2/D1 cells, while RFX1 downregulates them. RFX1 expression is low in undifferentiated NT2/D1 cells but significantly elevates after differentiation with retinoic acid. Bexarotene, the retinoid X receptor (RXR) agonist that also inhibits the NRF2-ARE signaling pathway, augments RFX1 transcription, resulting in attenuated CSC properties and MDR protein expressions in NT2/D1 cells [89]. CSCs’ chemotherapy resistance is partially attributed to epigenetic modifications. Histone deacetylase inhibitors, belinostat and panobinostat, synergist with cisplatin to trigger cell cycle arrest and apoptosis in NT2/D1 cells [92]. Likewise, the demethylating agent guadecitabine upregulates TP53 target genes and downregulates pluripotency genes, leading to tumor regression and cisplatin sensitization [101]. Notably, cisplatin can induce resistance to itself and paclitaxel in NT2/D1 cells by triggering differentiation, suggesting another mechanism of chemotherapy resistance in CSCs [100].

Table 2.

NT2/D1 cells as a treatment screening platform for CSC-based therapy

Culture environment/animal model	Intervention	Outcomes	References
In vitro
No modification	sh-AGR2, sh-AnXA2, and sh-EGFR	- AGR2 knockdown ↓teratoma growth - AGR2 functions through AnXA2/EGFR signaling	[65]
	Dinaciclib, cisplatin	- Dinaciclib enhances the cisplatin effect - ↑CDK5 expression in the cisplantin-resistant model	[88]
	Regulatory factor X1 (RFX1), Bexarotene (RXR agonist)	- ↓Drug resistance proteins - ↑Cisplatin sensitivity	[89]
	bis-bibenzyls isolated from liverworts	- Perrottetin E and perrottetin F is cytotoxic against cancer cells	[90]
	Cisplatin, palbociclib	- ↓Cancer cell recovery - Additive effect when combined	[91]
	Belinostat and panobinostat	- ↑cell cycle arrest, Ki67 index, apoptosis - Effective in cisplatin-resistant cells	[92]
	miR-196a-5p	- miR-196a-5p inhibits NR6 A1/E-cadherin signaling axis → ↓cell proliferation, migration, invasion, and tumor neurogenesis	[69]
	SOX2OT siRNA	- ↓SOX2OT	[93]
	pTIG1-myc-his vector, pSPINK2-flag vector	- ↓TIG1 and SPINK2 - TIG1 inhibited cell invasion, migration, and epithelial–mesenchymal transition (EMT) through uPA/uPAR signaling pathway - SPINK2 augments TIG1 functions	[70]
	Benzothiazole-based carbamates and amides	- ↓Cell proliferation, migration and invasiveness	[94]
	imsnc761 and DDX6	- ↑Cell apoptosis - ↓Proliferation via the p53 pathway - ↓Mitochondrial function	[95]
	lipoplexes, nioplexes, and polyplexes	- Lipid-based vectors have a faster intracellular transportation	[96]
	Rana catesbeiana ribonuclease-6 (RC-6)	- ↑Cytotoxicity, caspase 9/3, and cellular senescence - ↑Neuron-like morphology	[97]
	Quercetin	- Block Wnt/β-catenin pathway - ↓SOX2, Oct4 and Nanog - ↓Proliferation, adhesion and migration	[98]
	MigR1-eGFP-hPPARβ/δ vector	- Activation of PPARβ/δ inhibits RAR/RXR dimerization and tumor proliferation	[81]
	Synthetic lipid A analog (ALA)	- Activate TLR4/IFNγ/NOS II pathway for tumor cell death induction	[99]
	Retinoic acid, cisplatin, paclitaxel	- ↓NANOG and POU5 F1 (Oct3/4) - Tetinoic acid ↑Resistance to both cisplatin and paclitaxel - Cisplatin induces resistance to itself and paclitaxel by triggering a differentiation response	[100]
	Guadecitabine	- Cisplatin-resistant tumor is sensitive to guadecitabine - The hypersensitivity depends on high levels of DNA methyltransferase 3B - ↑p53 target genes and ↓pluripotency genes	[101]
	Clathrodin	- ↑Apoptosis	[102]
In vivo
Mouse	sh-AGR2	- AGR2 knockdown ↓teratoma growth	[65]
	MigR1-eGFP-hPPARβ/δ vector	- Activation of PPARβ/δ inhibits RAR/RXR dimerization, tumor proliferation, and migration - ↑Tumor necrosis	[81]
	Synthetic lipid A analog (ALA)	- Activate TLR4/IFNγ/NOS II pathway for tumor cell death induction	[99]
	Guadecitabine	- Cisplatin-resistant tumor is sensitive to guadecitabine - ↑MAGE-A3 and MAGE-A1 expression	[101]
Zebrafish	Dinaciclib, cisplatin	- Dinaciclib enhances the cisplatin effect	[88]

Novel therapeutic targets have been proposed based on preclinical experiments on NT2/D1 cells. Knockdown of AGR2 and AnXA2 has been shown to inhibit the EGFR signaling pathway in NT2/D1 cells, interrupting TME modification by CSCs [65]. In addition, TIG1 and SPINK2, downregulated in NT2/D1 cells compared to normal testicular cells, have been demonstrated to inhibit cell invasion, migration, and epithelial-mesenchymal transition (EMT) [70]. Another potential treatment is miR-196a-5p, usually downregulated in NT2/D1 cells, but can inhibit the NR6 A1/E-cadherin pathway to attenuate tumor cell proliferation and migration [69]. Intermediate-sized non-coding RNA 761 (imsnc761) interacts with DEAD-box helicase 6 (DDX6) to trigger apoptosis and suppress cell proliferation in NT2/D1 cells through the TP53 signaling pathway [95]. Other substances with cytotoxic activity against NT2/D1 cells include bis-bibenzyl compounds from liverworts, benzothiazole-based carbamates, RC-6 ribonuclease, quercetin, and synthetic lipid A analogs (ALA) [90, 94, 97–99].

NT2/D1 cells offer a promising treatment screening platform representing chemotherapy-resistant cancer cells, particularly CSCs. However, studies focusing on cancer treatments in NT2/D1 cells remain limited. For instance, some key mediators in the crosstalk between CSCs and TME, such as STAT3, have not been studied in NT2/D1 cells [29]. Comprehensive analysis, such as whole-genome sequencing, may facilitate identifying additional mediators of CSCs’ signaling mechanisms and potential targets for the eradication of CSCs. Furthermore, most experiments with NT2/D1 cells are in vitro, lacking the component of TME interaction which is critical in cancer pathogenesis. In vivo or in vitro studies with cocultured cells representing TME may demonstrate more practical outcomes of CSC-based therapy.

NT2/D1 cells as biological models and treatment screening platforms beyond cancer

With the ease of neuronal differentiation in NT2/D1 cells, NT2/D1 cells and NT2/D1-Ns have been widely utilized as a neurodevelopmental model (Table 3). Retinoic acid is the well-established method for inducing neuronal differentiation in NT2/D1 cells. Diverse mechanisms have been shown to underly neuronal differentiation, involving both genetic and epigenetic levels. The Gαq/phospholipase C-β (PLCβ) signaling pathway triggers neuronal differentiation in NT2/D1 cells, as transfecting PLCβ1 into undifferentiated NT2/D1 cells can induce differentiation without external agents [105]. Activation of PPARα accelerates neuronal differentiation in NT2/D1 cells compared to only retinoic acid treatment, functioning through the ERK and p38 MAPK pathways [126]. In addition, BRCA1-associated ATM activator 1 (BRAT1) interacts with the integrator complex subunit 11 (INTS11) and INTS9 subunits, forming a trimeric complex essential for activating RE1 Silencing Transcription Factor (REST)-responsive genes required for differentiation [104, 109]. Epigenetic modifications play a crucial role in NT2/D1 cell differentiation, particularly histone H3 tail trimming and SMAD3/EZH2-based regulation [103, 124]. Beyond neurodevelopmental studies, NT2/D1 cells are also utilized as biological models for traumatic brain injury (TBI), neurological malignancies, and astrocytic senescence [64, 106, 137, 138].

Table 3.

NT2/D1 cells a biological models of diseases beyond cancer

Model	Differentiation agents	Characterization	References
In vitro
Neurodevelopment	Retinoic acid retinal, retinol, 5-bro- mouracil 2'deoxyribose (BUdR), 5-iodouracil 2'deoxyri- hose (IUdR), hexamethylene bisacetamide (HMBA), di- methylacetamide (DMA), and dimethylsulfoxide (DMSO), CAMP,butyrate, TPA, CDDP, or cytosine arabin- oside	- Neuronal morphology - ↑HCMV susceptibility - SSEA-3, A2B5, ALPp/Sp2/3, ALPp/Sp2/5, neurofilaments, GFP, desmin	[16]
	Retinoic acid	- ↓EGFR expression - SSEA-3	[14]
	Retinoic acid	- Neuronal morphology - ↑Cytokeratin - ↓Neurofilaments, glial filaments, vimentin	[17]
	Retinoic acid	- ↑Chromatin-associated histone H3 tail trimming due to ↑cathepsin B and cathepsin D	[103]
	Retinoic acid, shBRAT1	- BRAT1/INTS11/INTS9 involves in neuronal differentiation	[104]
	Retinoic acid, AraC, PLCβ1 shRNAs transfection	- PLCβ1 promote differentiation	[105]
	Retinoic acid	- SOX2 and SOX9 overexpression ↑proliferation	[106]
	Retinoic acid	- NeuroD, GluR, Tau	[107]
	Retinoic acid, CRISPR/Cas9	- GA-repeat/GPM6B reduce neuronal differentiation	[108]
	Retinoic acid	- BRAT1/INTS11/INTS9 complex involves in neuron differentiation	[109]
	AraC	- ↓pre-mir-106b and pre-mir-19b	[110]
	Flavonoids	- ↑HIF-1α → hypoxia response element (HRE) → ↑erythropoietin	[111]
	Retinoic acid	- ↓linc-ROR spliced variants	[112]
	Retinoic acid, hypoxia, nitric oxide	- Mitochondrial biogenesis is unchanged in differentiated cells	[113]
	Retinoic acid	- ↑DQ582359, ↑DQ596268 piRNAs → ↑MAP2 and TUBB3	[114]
	Retinoic acid, vitreous humor	- ↑SSEA3, TRA-1–81, OCT4 - ↑Growth rate and S phase	[115]
	Retinoic acid, siRNA (Nup93, Nup188, Nup205, CTCF)	- Nup93 and CTCF modulate HOXA gene locus during differentiation	[116]
	Retinoic acid	- TrkC-miR2 is reversely correlated to NGFR/TrkC in differentiation - ↓TrkC-miR2 down-regulation → ↓neural differentiation	[117]
	Retinoic acid	- ↓OCT4 A and ←→OCT4B5 after differentiation	[118]
	Retinoic acid	- ↑PIWIL4 - PIWIL4 regulates PTN and NLGN3 in glioma suppression - PIWIL4-silenced NT2 culture media ↓proliferation of glioma cells	[119]
	Neuron induction medium, TIAM2S shRNA	- TIAM2S is produced in NT2/N - TIAM2S-KD reduces serotonin level	[120]
	Retinoic acid, glucocorticoids	- PSORS1 C3 is regulated by glucocorticoids and could modulate OCT4 expression by acting as an enhancer	[121]
	Retinoic acid	- H2 AK119ub1 E3 ligase CUL4B is required for retinoic acid-inducible HOX gene - CUL4B suppresses RORγ activity	[122]
	Retinoic acid, BMP-2, UV irradiation	- ↑Nucleotide excision repair during differentiation - ↑Xeroderma pigmentosum (XP) complementation	[123]
	Retinoic acid	- Smad3 and EZH2 control epigenetic regulation of differentiation	[124]
	Retinoic acid	- Extended PHD domain (ePHD6) and zinc finger in MLL3/4 specifically recognizes an H4H18-containing histone H4 fragment - ePHD6 of MLL3/4 and histone H4 interact in nucleosomal methylation and MLL4-mediated neuronal differentiation	[74]
	Retinoic acid	- PPM1D maintains the undifferentiation state and suppresses retinoic acid-induced ERK regulation in cell differentiation	[125]
	Retinoic acid, Wy14643	- PPARα activation → ↑neuronal differentiation - ↑NeuroD, MAP2, and CDK5 - ↓Nestin	[126]
	Retinoic acid, 5-aminoimidazole-4-carboxamide-1-β-D-ribotide (AICAR)	- ↑MAP2, NRG1, NRP1, NRP2, NEUROG1 and EN1 by retinoic acid - ↑SHH, WNT2 and WNT7B by AICAR	[127]
	Retinoic acid, paraquat	- ROS decreases stemness-related genes (NANOG, OCT4 and TDGF1) - ROS promotes neural differentiation through MAPK-ERK1/2 pathway	[128]
	Retinoic acid, measles virus	- Neuron-to-neuron spreading of measles virus is dependent on hyperfusogenic F, hemagglutinin, and putative neuronal receptor for measles virus	[129]
	Retinoic acid, 1% O2 (hypoxia)	- Hypoxia promotes regression to stem cells with HIF-2α pathway	[130]
	Retinoic acid	- hsa-miR-6165 affects the cell cycle progression and increases apoptosis	[131]
	Ara-C	- ↑Differentiation efficiency (glutamatergic/cholinergic neurons)	[132]
	Retinoic acid	- Differentiation efficiency depends on the cell density at differentiation-initiation and the type of cell culture plastic ware	[133]
	Retinoic acid, paraoxon, mipafox	- Paraoxon disrupts chromatin assembly and nucleosome integrity - Mipafox teratogenicity is unclear	[134]
	Retinoic acid, methylmercury chloride, sodium arsenite, sodium valproate, and methylazoxymethanol	- ↓Differentiation - ↓Migration	[135]
	Retinoic acid	- SOX2 overexpression ↓MAP2-positive neurons while no difference in the number of GFAP-positive astrocytes	[136]
TBI	Retinoic acid	- SOX2 regulates proliferation of mature astrocytes - SOX9 overexpression promote migration and glutamate uptake	[106]
Neurological malignancy	Retinoic acid	- Opposite Insulin-like growth factor 1 receptor (IGF-1R) and miR-6165 pattern	[64]
	Retinoic acid	- PML gene expression changes depending on the stage of differentiation	[137]
Senescence	Retinoic acid, anti-miR-21	- miR-21 downregulation → Arrest of cell growth and premature cellular senescence	[138]

Neurons differentiated from NT2/D1 cells are also frequently used for neuropsychiatric drug screening, especially for bipolar disorder, schizophrenia, and Parkinson’s disease (Table 4). Furthermore, both undifferentiated NT2/D1 cells and NT2/D1-Ns are routinely used to evaluate the efficacy and safety of biological substances and drug delivery vehicles in preclinical studies [107, 139–141, 145–149].

Table 4.

NT2/D1 cells as a treatment screening platform for diseases beyond cancer

Model	Differentiation agent	Intervention	Outcomes	References
In vitro
Neuroprotection	Retinoic acid	Baicalin	- ↑Neurite outgrowth - ↑Glycolysis pathway	[107]
	No differentiation	Tectorigenin	- ↑Hypoxia-inducible factor (HIF)−1α → ↑EPO mRNA	[139]
	Retinoic acid	Nlackberry-digested polyphenols (BDP) and BDP major aglycones (hBDP)	- Neuroprotective effects through mTOR signaling and ASNS/ATF5 genes	[140]
	Retinoic acid	Bravo de Esmolfe extraction	- ↓Free radical production	[141]
Insecticide toxicity	Retinoic acid, rotenone, fipronil	Rho-kinase inhibitor Y-27632, n-acetyl cysteine	- ↑cell migration and neuronal differentiation	[142]
Glioma	Retinoic acid	PIWIL4 knockdown	- PIWIL4 regulates PTN and NLGN3 in glioma suppression - PIWIL4-silenced NT2 culture media ↓proliferation of glioma cells	[119]
Radiation-induced neuron injury	Retinoic acid	Radiation	- ↑Apoptosis, senescence and cell cycle arrest	[143]
Chemotherapy-induced peripheral neuropathy	Retinoic acid, cisplatin	Sodium azide, amifostine	- ↑Cell viability - ↓Oxidative stress and apoptosis	[144]
Drug-delivery vehicle	No differentiation	Niosome formulations without and with poloxamer 188	- ↑Cellular uptake and viability	[145]
	No differentiation	Niosomes, nioplexes	- Effective drug vehicles with low cytotoxicity	[146]
	No differentiation	Niosomes/nioplexes	- Stable transfection efficiency	[147]
	No differentiation	superparamagnetic iron oxide (SPION)	- Low cytotoxicity	[148]
	No differentiation	PEI-functionalized graphene oxide (PEI-GO)	- Low cytotoxicity with effective siRNA delivery	[149]
Psychiatric disorders	Retinoic acid	Amisulpride, aripiprazole, clozapine, and risperidone	- ↑Circadian rhythm and vascular endothelial growth factor signaling - ↓Adherens junction and cell cycle pathways	[150]
	Retinoic acid	Lithium chloride, valproate, lamotrigine and quetiapine	- Trimetazidine as a potential drug - ↑ATP production	[151]
	Retinoic acid	Amisulpride, aripiprazole, clozapine, risperidone	- Mechanisms involve Wnt signaling and action potential regulation - Protein–protein interaction (PPI) networks have clusters of PDCD10, ANK2, and AKT3	[152]
	Retinoic acid	Amisulpride, aripiprazole, clozapine, lamotrigine, lithium, quetiapine, risperidone, valproate	- ↓Ribosomal gene expression and protein synthesis - Propose protein synthesis inhibitors as novel agents for neuropsychiatric disorders	[153]
	Retinoic acid	Lamotrigine, lithium, quetiapine, valproate	- GAS6-AS1 and MIR100HG lncRNAs are key drivers in synaptic vesicle cycle, endoplasmic reticulum-related functions and neurodevelopment	[154]
	Retinoic acid	Aripiprazole, amisulpride, risperidone, quetiapine, clozapine	- Quetiapine ↑SREBF1 and SREBF2	[116]
	Retinoic acid	Amisulpride, aripiprazole, clozapine, lamotrigine, lithium, quetiapine, risperidone, valproate	- ↓Hippo pathway - ↓MAPK and NFκB pro-inflammatory signaling	[155]
	Retinoic acid	Lamotrigine, lithium, quetiapine, valproate	- ↓CHDH mitochondrial protein from quetiapine and lamotrigine	[156]
	Retinoic acid	Amisulpride, aripiprazole, clozapine, lamotrigine, lithium, quetiapine, risperidone, valproate	- ↓SOX2 and target genes (CCND1, BMP4, and DKK1) - ↓Neurogenesis	[157]
	Retinoic acid	Lithium, valproate, quetiapine, and lamotrigine	- Valproate: ↑OXPHOS genes, mitochondrial uncoupling and maximal respiratory capacity - Quetiapine: ↓OXPHOS genes and respiration - Lamotrigine: ↓OXPHOS genes, no effect on respiration - Lithium: ↓citric acid cycle	[158]
	Retinoic acid	Lithium, valproate, lamotrigine and quetiapine	- ↑Cholesterol biosynthesis pathway when combined medications - ↑Intracellular cholesterol transport genes - ↓Cholesterol efflux gene	[159]
	Retinoic acid	Lithium, valproate, quetiapine and lamotrigine	- Gene expression signature (GES) and connectivity map (Cmap) can be used as tools to repurpose drugs for BD	[160]
Parkinson’s disease	1-methyl-4-phenylpyridinium (MPP +), retinoic acid	Hexagonal boron nitride nanoparticles (hBNs)	- ↑Cell viability - ↑Antioxidant capacity - ↓Apoptosis	[161]
	Retinoic acid	Polysaccharide alginate capsulation	- Effective dopamine production capability	[162]
Alzheimer's disease, Parkinson's disease, and stroke	No differentiation	Tripeptide H-Gly-Pro-Glu-OH (GPE)	- Low cytotoxicity, oxidative, genotoxicity, and embryotoxicity	[163]

Future challenges

The extensive application of NT2/D1 cells beyond the field of cancer may yield valuable cross-disciplinary insights. For instance, some antipsychotic medications affect the Wnt signaling pathway, SOX expression, and intercellular protein synthesis, which may be applicable in cancer therapy [152–154, 157]. On the contrary, substances that promote the viability of NT2/D1 cells, such as hexagonal boron nitride nanoparticles, may require careful assessment of the safety profiles in cancer patients [161]. Additional factors, including lactylation index [164] and long non-coding RNAs [165], which affect CSCs and TME will need to be considered when contemplating NT2/D1 cells as a biological model. In the end, establishing cancer models and oncologic treatment should incorporate laboratory-to-clinical translational paradigms with emphasis on safety and efficacy in addressing CSCs and TME [166, 167].

Conclusion

Cancer is one of the most dominant contributors to premature death globally. CSCs pose a major challenge in cancer therapy due to their resistance to conventional chemotherapy, driving metastases and relapses [4–6]. Targeting CSCs opens new therapeutic avenues, particularly focusing on their self-renewal capacities and interactions with TME. NT2/D1 cells can be utilized as a biological model that accurately recapitulates CSCs and the TME, enhancing our understanding of tumor initiation and progression as well as serving as a valuable platform for screening CSC-based therapies. In addition, the ability of NT2 s/D1 cells to differentiate into non-cancerous cells provides novel insights into potential therapeutic strategies targeting CSCs. Furthermore, NT2/D1 cells have been employed as biological models and treatment screening platforms for CNS disorders. However, significant gaps remain in our understanding of CSCs’ signaling mechanisms. A major limitation of this report is it lacks original experimental data and relies solely on a literature review. Clearly, primary research experiments are warranted to probe these CSC signaling pathways, as well as the application of the insights gained from NT2/D1 cells in cancer and beyond, will expand the boundaries of CSC-based therapy for clinical indications.

Abbreviations

AGR2

Anterior gradient-2

ALA

Lipid A analogs

AML

Acute myeloid leukemia

AnXA2

Annexin A2

APL

Acute promyelocytic leukemia

BRAT1

BRCA1-associated ATM activator 1

CAFs

Cancer-associated fibroblasts

CCL2

Chemokine (C–C motif) ligand 2

CCR2

Chemokine (C–C motif) ligand 2 receptor

CDK5

Cyclin-dependent kinase 5

CNS

Central nervous system

CSCs

Cancer stem cells

CSF1

Colony stimulating factor 1

CSF1R

Colony stimulating factor 1 receptor

DDX6

DEAD-box helicase 6

E2

Estradiol

EGFR

Epidermal growth factor receptor

EMMPRIN

Extracellular matrix metalloproteinase inducer

EMT

Epithelial-mesenchymal transition

ESR1

Estrogen receptor 1

ESR2

Estrogen receptor 2

HER2

Human epidermal growth factor receptor 2

HOXA10

Homeobox A10

HSCs

Hematopoietic stem cells

IL

Interleukin

Imsnc

Intermediate-sized non-coding RNA

lncRNAs

Long noncoding RNAs

LSCs

Leukemic stem cells

MDR

Multidrug resistance

MDSCs

Myeloid-derived suppressor cells

NOD/SCID

Non-obese diabetic severe combined immune deficiency

NRF2

Nuclear factor erythroid 2-related factor 2

NT2/D1

NTERA-2 clonal D1 cell line

PLCβ

Phospholipase C-β

PTEN

Phosphatase and tensin homolog

REST

RE1 silencing transcription factor

RFX1

Regulatory factor X1

RXR

Retinoid X receptor

SOX2OT

SOX2 overlapping transcript

STAT3

Signal transducer and activator of transcription 3

TBI

Traumatic brain injury

TGFβ

Transforming growth factor β

TME

Tumor microenvironment

Treg

Regulatory T cells

Author contributions

MCB, TR, NP, HW and JYL drafted and finalized the manuscript. HW and JYL provided guidance and supervision. All authors read and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.