SETD4 Confers Cancer Stem Cell Chemoresistance in Nonsmall Cell Lung Cancer Patients via the Epigenetic Regulation of Cellular Quiescence

Increasing evidence indicates that quiescent cancer stem cells (CSCs) are a root cause of chemoresistance. SET domain-containing protein 4 (SETD4) epigenetically regulates cell quiescence in breast cancer stem cells (BCSCs), and SETD4-positive BCSCs are chemoradioresistant. However, the role of SETD4 in chemoresistance, tumor progression, and prognosis in nonsmall cell lung cancer (NSCLC) patients is unclear. Here, SETD4-positive cells were identified as quiescent lung cancer stem cells (qLCSCs) since they expressed high levels of ALDH1 and CD133 and low levels of Ki67. SETD4 expression was significantly higher in advanced-stage NSCLC tissues than in early-stage NSCLC tissues and significantly higher in samples from the chemoresistant group than in those from the chemosensitive group. Patients with high SETD4 expression had shorter progression-free survival (PFS) times than those with low SETD4 expression. SETD4 facilitated heterochromatin formation via H4K20me3, thereby leading to cell quiescence. RNA-seq analysis showed upregulation of genes involved in cell proliferation, glucose metabolism, and PI3K-AKT signaling in activated qLCSCs (A-qLCSCs) compared with qLCSCs. In addition, SETD4 overexpression facilitated PTEN-mediated inhibition of the PI3K-mTOR pathway. In summary, SETD4 confers chemoresistance, tumor progression, and a poor prognosis by regulating CSCs in NSCLC patients.


Introduction
Lung cancer is a leading cause of cancer-related death worldwide, and nonsmall cell lung cancer (NSCLC) accounts for approximately 85% of all lung cancers [1]. The prognosis for NSCLC is still poor. For early-stage NSCLC patients, tumor recurrence or metastasis is a crucial problem despite the use of chemotherapy and/or radiotherapy after surgery. For patients with locally advanced or metastatic NSCLC, many treatment options have been identified. Beyond conventional treatment strategies, such as chemotherapy and radiotherapy, targeted therapy and immunotherapy have become important options in the treatment of advanced NSCLC [2]. However, these therapies are unable to eradicate all cancer cells. The 5-year survival rate of NSCLC patients with distant metastases is only approximately 6% [3]. The reexpansion of drug-resistant cancer cell clones facilitates tumor relapse and progression [4]. Thus, treatment strategies targeting these drug-resistant cancer cells are crucial to improving the survival of lung cancer patients.
Previous studies have identified that a small proportion of cancer stem cells (CSCs) in NSCLC can survive chemotherapy and radiotherapy, and these cells are responsible for therapeutic resistance, relapse, and tumor metastasis [5,6]. Accumulating evidence indicates that the root cause of drug resistance is a small subpopulation of quiescent cancer stem cells (qCSCs) [7][8][9][10][11]. qCSCs are maintained in a nondividing state, which enables them to resist harsh environmental conditions, including chemotherapy, radiotherapy, and hypoxia [12][13][14]. Importantly, this state is reversible, and qCSCs can be activated and reenter the cell cycle under favorable conditions, resulting in tumor relapse and metastasis [15,16]. Epigenetic mechanisms such as DNA methylation, histone modifications, and some key signaling pathways including the Wnt/β-catenin signaling pathway, hedgehog signaling pathway, and Notch signaling pathway promote CSC activation, tumorigenesis, and metastasis [17]. Alteration of DNA methylation and histone modifications could result in an expansion of CSCs [18]. Thus, qCSCs can survive during chemotherapy, and the activation of these cells is responsible for tumor regrowth after chemotherapy [19]. Similarly, the activation of qCSCs is necessary for tumor relapse or metastasis, which occurs long after initial treatment [20]. Current therapies fail to eliminate resistant qCSCs, making them attractive therapeutic targets in cancer therapy [21].
However, it remains unclear whether SETD4 could be an indicator of quiescent CSCs in different tumor types, and it also needs to be further verified whether SETD4-positive quiescent CSCs are correlated with tumor progression and prognosis. In this study, we assessed SETD4 expression and analyzed the relationship between SETD4 expression and chemoresistance and prognosis in NSCLC patients. . The human lung cancer cell lines  NCI-H1299 and NCI-H520 were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA) and cultured in RPMI-1640 with 10% FBS at 37°C in a humidified 5% CO2 incubator.

Clinical Sample
Acquisition. This study included 166 tumor specimens from patients with histologically confirmed stages I-IV NSCLC from January 2016 to July 2020 in our hospital. Clinical sample acquisition was approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University and conducted according to the Declaration of Helsinki. Most of the specimens from early-stage NSCLC patients were obtained by surgery, while specimens from advanced-stage NSCLC patients were obtained by bronchoscopic biopsy or percutaneous core-needle lung biopsy. Tumor specimens were fixed with 4% formalin, paraffin embedded, and sectioned at 3 μm thickness for IF analysis. In addition, fresh tumor tissues were stored in tissue storage solution at 4°C and processed immediately (within 2 hours) for subsequent primary culture and chemoradiotherapy in vitro.

Primary Culture of Lung Cancer Cells and
Chemoradiotherapy In Vitro. Surgical tumor specimens were chopped and digested in tumor dissociation solution (Miltenyi Biotec, 130-095-929) at 37°C for 1 hour. Then, the cells were filtered through 40 μm cell strainers and centrifuged at 300 × g for 5 minutes. The cells were resuspended in ultralow attachment 6-well plates and cultured in DMEM/F12 (Corning; 10-092-cv) for one week. Then, these cells were treated with 200 nmol/L TAX and 10 μmol/L CDDP and irradiated (30 Gy) approximately 10 days after chemotherapy initiation. The medium was replaced twice every week, and sustained drug treatment lasted for over 2 weeks. The surviving cells were obtained via a dead cell removal kit (Miltenyi Biotec; 130-090-101) according to the manufacturer's instructions and used for subsequent experiments. fields with a minimum of 2000 cells were analyzed for each sample by ImageJ software. Statistical analysis and graphic presentation were carried out using GraphPad Prism version 7.0 (GraphPad Software, San Diego, CA, USA) and SPSS 23.0 software (IBM Corporation, Chicago, IL, USA). The data are calculated as the mean ± SD or mean ± SEM. Statistical comparisons were performed using two-tailed Student's t tests. Progression-free survival (PFS) was defined as the time from the beginning of chemotherapy to the date that progressive disease (PD) was observed. Survival analysis was performed using Kaplan-Meier curves and log-rank tests. A Pearson r correlation analysis was performed to assess the correlation between the proportion of SETD4positive cells and PFS time. p < 0:05 was considered to indicate a statistically significant difference.

SETD4-Positive Lung Cancer Cells
Were Identified as qLCSCs in Tumor Specimens from NSCLC Patients. We first characterized SETD4-positive cancer cells in NSCLC patient specimens. Ye et al. previously confirmed that SETD4 was enriched in qBCSCs from breast cancer cell lines [33]. Consistently, IF analysis in NSCLC patient specimens revealed that approximately 97% of SETD4-positive lung cancer cells lacked a Ki67 signal (Figure 1(a)), suggesting that SETD4positive cells were maintained in a quiescent state. Additionally, most of these SETD4-positive cells showed positive ALDH1 staining (125 ALDH1 + cells/135 SETD4 + cells) (Figure 1(b)), positive CD133 staining (100 CD133 + cells/ 111 SETD4 + cells) (Figure 1(c)), and positive CD44 staining (80 CD44 + cells/91 SETD4 + cells) ( Figure S1), which are well-established features of CSCs; thus, SETD4-positive cells were identified as CSCs. We next tried to isolate qLCSCs from clinical tumor specimens according to methods established in our previous studies [33]. We obtained 34 surgical tumor specimens derived from NSCLC patients (23 cases of lung adenocarcinoma (LUAD) and 11 cases of lung squamous cell carcinoma (LUSC)) and dissociated the cancer cells. Primary lung cancer cells were treated with chemotherapy and radiotherapy. Trypan blue staining showed a small proportion of surviving cells in LUAD and LUSC cultures after chemoradiotherapy, and we selected these cells by live cell sorting for further analysis (Figure 1(d)). Quiescent CSCs can be activated, and these cells have been proven to play a major role in tumor repopulation and relapse [34,35]. Hence, we collected activated qLCSCs (A-qLCSCs) after inducing activation in tumorsphere formation medium after drug withdrawal for 48 hours (Figure 1(d)). Interestingly, A-qLCSCs exerted strong tumorsphere-forming properties ( Figure 1(d)), suggesting their extraordinary potential for facilitating tumor relapse. We also performed tumorsphere formation assay in H1299 cell lines and A-qLCSCs derived from H1299 cell lines. Tumorspheres more than 30 μm in size are calculated, and A-qLCSCs exhibited remarkable tumorsphere-forming capacity ( Figure S2). Consistent with our previous study, we found that a large proportion of the surviving cells were SETD4-positive and Ki67-negative    Dead cells were stained with trypan blue. Red arrows indicate surviving cells, which we referred to as qLCSCs. Then, qLCSCs were activated in tumorsphere formation medium after drug withdrawal for 48 hours. The middle panel shows a representative bright field image of A-qLCSCs. The right panel shows the tumorspheres formed by A-qLCSCs after cultivation for 7 days. Scale bar, 100 μm. Representative IF images of SETD4/Ki67 (e) and SETD4/ALDH1 (f) in qLCSCs and A-qLCSCs derived from clinical LUAD and LUSC specimens. DAPI, nuclear counterstaining. Scale bar, 50 μm. (g) Protein levels of SETD4, Ki67 and PCNA in qLCSCs and A-qLCSCs derived from the H1299 cell line, as detected by WB analysis. n = 3. GAPDH was used as an internal reference. The data are presented as the mean ± SD. * * p < 0:01; * p < 0:05.  Figure S3), supporting the notion that they were quiescent LCSCs. However, we observed low expression levels of SETD4 and high expression levels of Ki67 (Figure 1(e)), ALDH1 (Figure 1(f)), and CD133 ( Figure S3) in A-qLCSCs. Then, we isolated qLCSCs and A-qLCSCs from the H1299 cell line and the H520 cell line following the protocols described above. WB analysis showed that SETD4 was highly expressed in qLCSCs and downregulated in A-qLCSCs (Figure 1(g)). Concomitantly, proliferative markers, including Ki67 and PCNA, were significantly upregulated in A-qLCSCs compared with qLCSCs ( Figure 1(g)). Altogether, we identified and isolated SETD4-positive qLCSCs from clinical NSCLC specimens.

SETD4-Positive qLCSCs Are Highly Associated with
Chemoresistance in NSCLC Patients. Since SETD4 was highly expressed in chemoradioresistant qLCSCs, we speculated that SETD4 might predict the response to chemotherapy in NSCLC patients. To verify this hypothesis, we enrolled NSCLC patients who had received platinum-based combination chemotherapy. Treatment responses were evaluated according to RECIST version 1.1 after every 2 cycles of chemotherapy. Patients who experienced PD after 2 or 4 cycles of chemotherapy were considered chemotherapy resistant, whereas patients who did not experience PD after 4 cycles of chemotherapy were considered chemotherapy sensitive. Then, we selected 21 chemoresistant patients and 21 chemosensitive patients and retrospectively analyzed the proportion of SETD4-positive qLCSCs in tumor tissue obtained before chemotherapy. The clinical characteristics of these patients are shown in Table 2. The proportion of SETD4-positive qLCSCs was significantly higher in the chemoresistant group than in the chemosensitive group (Figures 3(a) and 3(b)). Moreover, patients with more SETD4-positive qLCSCs had a shorter PFS than those with fewer SETD4-positive qLCSCs (2.3 months vs. 5.8 months, HR = 1:95 (1.02-3.73), p = 0:028) (Figure 3(c)). Pearson r correlation analysis revealed that PFS was negatively correlated with the proportion of SETD4-positive cells (r = −0:326, p = 0:043) (Figure 3(d)). Figure 3(e) shows representative chest CT images for 4 patients. We observed significant enlargement of tumor lesions after 2 cycles of chemotherapy in case 1 and case 2, both of which had high expression of SETD4. Conversely, we observed tumor shrinkage in case 3 and case 4, both of which had low expression of SETD4. These results demonstrated that SETD4 is associated with chemotherapy resistance in NSCLC patients.

SETD4 Confers Resistance to Chemoradiotherapy by
Inducing Cellular Quiescence. To further verify that chemoradiotherapy resistance is regulated by SETD4, we overexpressed SETD4 in LCSCs. LCSCs were separated from H1299 cell lines by CD133 sorting, as CD133 is a commonly used cell surface marker to isolate CSCs from lung cancer [36][37][38]. We observed a reduction in proliferation markers, including Ki67 and PCNA, in SETD4-overexpressing LCSCs    Figure S4D). Additionally, a decrease in the caspase-3 signal was observed in SETD4-overexpressing LCSCs versus GFP control cells 24 hours after TAX (200 nmol/L) plus CDDP (10 μmol/L) treatment ( Figure S5), indicating reduced cell apoptosis after drug treatment in SETD4-overexpressing LCSCs. Thus, SETD4 confers resistance to chemoradiotherapy by inducing cell quiescence.

Quiescent LCSCs Suppressed Cell Proliferation and
Metabolism. To characterize the molecular signatures of SETD4-positive qLCSCs and investigate the underlying mechanism of SETD4-mediated chemoradioresistance, we performed RNA-seq analysis of qLCSCs and A-qLCSCs. Differentially expressed genes between qLCSCs and A-qLCSCs are presented in a heatmap ( Figure 5(a)). A total of 1054 genes were upregulated and 774 genes were downregulated in qLCSCs ( Figure 5(b)). GO enrichment analysis showed that the downregulated genes in qLCSCs were enriched in the glucose catabolic process, ATP generation from ADP, the glycolytic process, cell proliferation, and DNA replication (Figures 5(c) and 5(d)). GSEA showed downregulation of genes involved in cell proliferation ( Figure 5(e)), carbon metabolism ( Figure 5(f)), and the PI3K-AKT signaling pathway in qLCSCs ( Figure 5(g)).
3.6. SETD4 Controls Cell Quiescence via H4K20me3 Catalysis in NSCLC, and PI3K-mTOR Is a Possible Downstream Target. We previously found that SETD4 facilitated heterochromatin formation via H4K20me3 catalysis in Artemia diapause embryos and BCSCs, thereby leading to cell

12
Stem Cells International quiescence [32,33]. H4K20me3 was reported to play a key role in constitutive heterochromatin formation and cell quiescence [39][40][41], and heterochromatin protein 1-α (HP1-α) is a conserved biomarker of heterochromatin formation [42,43]. Accordingly, we hypothesized that SETD4 controls cell quiescence in NSCLC in the same manner. Consistently, H4K20me3 and HP1-α were highly expressed in qLCSCs compared with A-qLCSCs (Figures 6(a)-6(c)), indicating the enrichment of constitutive heterochromatin in qLCSCs. Moreover, there was no significant change in the level of H3K9ac ( Figure S6A and B), which is an indicator of chromatin. Then, we overexpressed SETD4 in A-qLCSCs, which exhibited low expression of SETD4, and we found that H4K20me3 and HP1-α and were markedly upregulated ( Figures  6(d)-6(f)). These results demonstrated that SETD4 facilitates heterochromatin formation via H4K20me3 catalysis in NSCLC, which strengthens previous findings and reveals an evolutionarily conserved mechanism of SETD4-mediated cell quiescence in different types of CSCs.
The above results suggest that SETD4 catalyzes H4K20me3 in specific gene regions and leads to gene silencing. Based on our RNA-seq analysis, the PI3K-AKT pathway was downregulated in qLCSCs; thus, we speculated that SETD4 epigenetically regulates the PI3K-AKT signaling pathway. PTEN, a negative regulatory factor of the PI3K-AKT pathway, has been reported to regulate cell quiescence in stem cells [25]. Mammalian target of rapamycin (mTOR), downstream of PI3K-AKT, has been identified as a key factor involved in stem cell quiescence [44]. Then, we overexpressed SETD4 in A-qLCSCs and detected the protein levels of PTEN, mTOR, PI3K, and phosphorylated PI3K (p-PI3K). As shown in Figure 6(g), overexpression of SETD4 led to a significant increase in the expression of PTEN as well as a significant decrease in p-PI3K and mTOR, indicating that the PI3K-mTOR pathway is a possible downstream target of SETD4.

Discussion
In the present study, we investigated the correlation between the abundance of SETD4-positive qLCSCs and clinical stage, chemotherapy sensitivity, and prognosis in NSCLC patients. We used SETD4 to define a subgroup of chemoradioresistant qLCSCs, thus revealing ideas for treatment strategies targeting SETD4-positive qLCSCs.
Quiescent CSCs play a critical role in tumor recurrence and chemoresistance. Approximately 30% to 55% of patients with NSCLC develop recurrence or distant metastasis after surgery [45]. Even stage I NSCLC reoccurs in approximately  15 Stem Cells International 25% of patients after curative resection [46], and reoccurrence is currently considered to be attributable to quiescent CSCs [15]. In the present study, we confirmed the existence of a group of SETD4-positive cells, defined as qLCSCs, in tumor tissues from early-stage NSCLC patients and found that the proportion of SETD4-positive qLCSCs in advanced-stage patients was significantly higher than that in early-stage patients. Hence, SETD4 could be used as a biomarker to predict the malignancy of NSCLC. There are currently no effective biomarkers for tumor recurrence risk after surgery or postoperative adjuvant chemotherapy in patients with early-stage NSCLC. SETD4 is a biomarker of qLCSCs, yet it remains unclear whether SETD4 can predict postoperative recurrence in patients with early-stage lung cancer. Further follow-up is needed to determine the time to tumor recurrence after surgery among early-stage patients. If the expression level of SETD4 is associated with postoperative tumor recurrence, then it may serve as a potential biomarker for predicting lung cancer recurrence and metastasis.
Platinum-based chemotherapy is one of the main treatment options for advanced NSCLC patients. However, the response rates to conventional platinum-based chemotherapy range from 19% to 36% in NSCLC patients [47][48][49]; these rates are unsatisfactory, and resistance results in a poor prognosis. It is difficult to identify patient subpopulations who are more likely to develop chemotherapy resistance in the clinic. Previous studies have demonstrated that chemoresistance is attributed to the increased proliferation of cancer cells [50][51][52]. In contrast to previous studies, this study proposes that quiescent CSCs are the root population  Figure 6: SETD4 controlled cell quiescence by inhibiting the PI3K-mTOR pathway via H4K20me3 catalysis. Representative IF images of H4K20me3 (a) and HP1-α (b) in qLCSCs and A-qLCSCs derived from H1299 and H520 cell lines. DAPI, nuclear counterstaining. Scale bar, 50 μm. (c) Protein levels of H4K20me3 and HP1-α in qLCSCs and A-qLCSCs. H4 was used as an internal reference. n = 3. * p < 0:05. Representative IF images of H4K20me3 (d) and HP1-α (e) in A-qLCSCs SETD4 and A-qLCSCs GFP . DAPI, nuclear counterstaining. Scale bar, 50 μm. (f) Protein levels of H4K20me3 and HP1-α in A-qLCSCs SETD4 and A-qLCSCs GFP . H4 was used as an internal reference. * p < 0:05. (g) Protein levels of PTEN, mTOR, PI3K, and p-PI3K in A-qLCSCs SETD4 and A-qLCSCs GFP . GAPDH was used as an internal reference. GAPDH was used as an internal reference. * p < 0:05. Data are presented as the mean ± SD.

16
Stem Cells International responsible for chemoresistance and that SETD4 is a unique marker of quiescent CSCs. In this study, SETD4-positive qLCSCs were examined in NSCLC patients who received platinum-based chemotherapy, and high SETD4 expression was found to be correlated with poor chemotherapy efficacy and prognosis. Thus, SETD4 may be emerging as a potential biomarker for predicting the efficacy of chemotherapy and may be a druggable target in the treatment of NSCLC. These findings provide an effective biomarker for identifying qCSCs in tumor tissues and have broad implications for clinical therapies. For patients with high levels of SETD4positive qLCSCs, conventional chemotherapy regimens have limited therapeutic effects, and new individualized treatment strategies need to be explored further. Studies with sample types ranging from Artemia embryos to human breast cancers have demonstrated the evolutionarily conserved mechanism of SETD4-regulated CSC quiescence across different species. In the present study, we found that SETD4 overexpression in CSCs derived from an NSCLC cell line led to cell quiescence and resistance to chemoradiotherapy. More importantly, we verified SETD4-mediated chemoresistance in NSCLC patients who received platinum-based chemotherapy. This is an extension of a previous study and validated that SETD4-positive qCSCs exert therapeutic resistance across different tumor types. Hence, SETD4 could be a potential therapeutic target to overcome drug resistance and might be applicable to a variety of tumors.
Quiescent CSCs can be activated under favorable conditions and divide asymmetrically into quiescent CSCs and proliferating CSCs. Proliferating CSCs divide continually and result in tumor relapse [33]. Hence, different treatment strategies should be adopted for different patients. For earlystage NSCLC patients who undergo surgery, the activation of qCSCs is a key factor that leads to tumor recurrence and metastasis after surgery. For these patients, some studies proposed that keeping these cells in a quiescent state for a long period could help prevent local progression or distant metastasis [53,54]. Quiescence can be maintained by the suppression of proliferative signaling or the activation of quiescence-related pathways. For example, inhibitors of CDK4/6 can block the transition from the G0/G1 phase to the S phase of the cell cycle, thus inhibiting tumor growth, recurrence, and metastasis [55,56]. On the other hand, for advanced-stage NSCLC patients who are resistant to chemotherapy, since chemotherapy treatment selects a subpopulation of chemoresistant qCSCs, strategies are urgently needed to completely eliminate these cells. Some studies have proposed that activating these cells and forcing them cells to reenter the cell cycle may sensitize them cells to chemotherapy drugs [53]. Our RNA-seq analysis revealed downregulation of genes involved in cell proliferation, glucose metabolism, and PI3K-AKT signaling pathway markers in qLCSCs than in A-qLCSCs. We thus inferred that drugs that alter the hypometabolic and low proliferative states of qLCSCs may activate these cells and that a combination of sequential chemotherapy can clear these activated cells. These hypotheses remain to be verified in future studies. Further study on the characteristics of SETD4 + CSCs and the epigenetic regulatory mechanism of SETD4 could pro-vide novel insight into methods for overcoming resistance to chemotherapy and radiotherapy.

Data Availability
The data that support the findings of this study are available upon request from the corresponding author.

Conflicts of Interest
The authors declare that there are no conflicts of interest.