Xanthatin synergizes with cisplatin to suppress homologous recombination through JAK2/STAT4/BARD1 axis in human NSCLC cells

Abstract Xanthatin (Xa) is a bicyclic sesquiterpene lactone identified from the plant Xanthium L. with impressive antitumor activity, but the role of Xa in non‐small cell lung cancer (NSCLC) is not known. Here we found that Xa inhibits proliferation, migration, invasion and induces apoptosis in NSCLC cells. RNA sequencing and Gene set enrichment analysis revealed that Xa significantly activates p53 pathway and suppresses E2F targets, G2M checkpoint and MYC targets in A549 cells. Among these changed genes, the down‐regulated gene BARD1 triggered by Xa was identified as a candidate involved in Xa’s antitumor effect because of its vital role in homologous recombination (HR). Further studies demonstrated that Xa inhibits HR through the BARD1/BRCA1/RAD51 axis, which enhances cell sensitivity to cisplatin. Mechanistic studies showed that Xa inhibits BARD1 through the JAK2/STAT4 pathway. Our study revealed that Xa is a promising drug to treat NSCLC, especially in combination with conventional chemotherapy.

Cisplatin (Cis) is one of the first line chemotherapeutic agent for treating advanced NSCLC. However, the clinical response rate of patients to Cis-based chemotherapy regimens is only 15% to 30%. 7,8 Although the drug resistant mechanism is complicated, the abnormal enhancement of DNA repair capacity in cancer cells is accounted, which undermines the therapeutic efficacy of DNA damage-inducing drugs such as Cis, etoposide and camptothecin. 9 The Cis sensitization in NSCLC cells is tightly correlated with the activation status of DNA repair pathways. 10 In terms of mechanism, cells activate DNA damage response (DDR) to safeguard genomic integrity in response to DNA damage. The DDR consists of six major DNA repair pathways. 11 Two of these pathways, non-homologous end joining (NHEJ) and homologous recombination (HR), dominate the repair of double-strand breaks (DSBs). 12 NHEJ directly ligates the broken DNA ends, which is considered to be error-prone. 13 By contrast, HR is an error free pathway that uses the information stored in a sister chromatin or homologue to repair the damage. 11 More importantly, HR is the most effective pathway activated in tumor cells to perform DSBs repair induced by Cis. 14,15 Hence, it is an effective strategy to overcome cancer drug resistance by targeting HR pathway. 16,17 Xa is a bicyclic sesquiterpene lactone and widely exists in the plant Xanthium L. 18 Xa has a broad spectrum of biological activities.
Recent studies indicated that Xa has significant antitumor effects on a number of malignancies including lung cancer, 19 hepatocellular carcinoma, 20 colon cancer 21 and glioma. 22 Xa triggers Chk1-mediated DNA damage response in lung cancer cells. 19 Furthermore, Xa induces endoplasmic reticulum (ER) stress, leading to activation of apoptosis in hepatoma cells. 20 Moreover, Xa activates ROS/XIAP signaling pathway to mediate G2/M cell cycle block and autophagy in colon cancer cells. 21 Nevertheless, the mechanisms of Xa's antitumor effects have not yet been defined.
In this study, we investigated whether Xa could increase the sensitivity to DNA damage and how Xa could regulate the DNA repair process in NSCLC. We found that Xa induces DNA damage, thus consequently leads to cell cycle arrest and apoptosis through the p53 signaling pathway. Moreover, our results clearly demonstrated that Xa has significant synergistic effects with Cis and powerful anticancer efficacy through suppressing HR via JAK2/ STAT4/BARD1 axis. These findings support potential applications of Xa as an anticancer therapeutic modality that targets HR pathway in NSCLC.

| Cell lines
The NSCLC cell lines A549 cells and H1299 cells were obtained from the Cell Bank of Chinese Academy of Sciences (Shanghai, China).
Cells were cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS) and then placed in a humidified atmosphere containing 5% CO 2 at 37°C.

| Chemicals and reagents
Xa was purchased from Yuannuotiancheng (Chengdu, China) and was dissolved in dimethyl sulfoxide (DMSO) as a stock solution at 40 mmol/L. JAK2 inhibitor TG101348 (GlpBio) was dissolved in DMSO as a stock solution at 10 mmol/L. All solutions stored at −80°C, and diluted in RPMI-1640 before using.

| Cell viability analysis
Equal number of cells were plated in 96-well plates and incubated at 37°C for 24 hours. Subsequently, the cells were treated with Xa (2.5, 5, 10, 20 and 40 μmol/L) or vehicle control for 12, 24 and 48 hours.
Finally, 10 μL CCK-8 reagent (Dojindo Laboratory, Japan) was added to each well and incubated at 37°C for another 2 hours. OD value (450 nm) was detected by a microplate reader for calculating cell viability.

| Cell wound scratch assay
Cells were plated in 6-well plates at a suitable density and were incubated at 37°C, 5% CO 2 . When cell confluence was 90%, cells were treated with serum-free medium overnight. Subsequently, 10 μL pipette tip was used to scratch the bottom of the six-well plate. Cells were cultured in no-serum medium with or without Xa. The width of the scarification worked as an indicator of migration distance and was photographed under a microscope at 0 and 48 hours.

| Transwell assay
The transwell insert with 8 μm pore (Corning) were placed on 24well plates. Subsequently, matrigel matrix was used to coat upper chambers (30 μL/chamber) and dried in air at 37°C for 2 hours. Cells (control or pre-treated with Xa) were added to the upper chamber in serum-free medium, while medium with 20% fetal bovine serum was added in the lower chamber. The 24-well plates with transwell insert were placed at 37°C with 5% CO 2 for 12-24 hours. Cells in the upper chambers were carefully removed by using cotton swabs.
Cells that migrated through the membrane were stained with 0.5% crystal violet and photographed under a microscope. The number of cells were calculated and used as indicator of invasion ability.

| Hoechst 33258 staining
Cells were plated in 96-well plates at a suitable density and incubated at 37°C, 5% CO 2 for 24 hours. Subsequently, cells were treated with DMSO (control) or Xa for another 24 hours. 4% paraformaldehyde was used to fix the cells, then 5 μg/mL Hoechst 33258 was added to each well for 15 minutes. Finally, the nuclei morphologies in different cells were stained with bright blue and observed under a fluorescence microscope.

| RNA sequencing (RNA-seq) and data analysis
Total RNA was extracted from cells for RNA-seq using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.), from three replicates of Xa treatment A549 cells (exposed to 10 μmol/L) and three replicates of controls. The quality and integrity of the RNA samples were examined as previously described. 23 RNA-seq was performed at Illumina Hiseq 2500 platform that included quality control, library preparation, fragmentation and PCR enrichment of target RNA according to standardized procedures. 150 bp pairedend raw reads were initially processed to obtain clean reads by removing adaptor sequences, low quality sequences, empty reads.
After quality control, the clean reads were mapped to human genome (hg38) using TopHat. 24 Genes expression level were quantitated by FPKM (Fragments per kilobase of exon per million reads mapped). Genes were considered to be expressed if they had an FPKM value greater than 1 in at least three or more samples.
Differentially expressed genes (DEGs) between control and Xa treatment cells were identified by using Cuffdiff. 25 An absolute fold change >2 and a FDR significance score <0.05 were used as thresholds to identify DEGs. Gene set enrichment analysis (GSEA) was performed using GSEA software (http://softw are.broad insti tute.org/gsea/) with default parameters.

| Validation of DEGs with quantitative realtime PCR (qRT-PCR)
In order to validate the reliability of the RNA-seq analyses, 12 candidate DEGs (NTRK3, HMOX1, ALDH1A3, DUSP6, BARD1, ZMAT3, FAS,   IGFBP3, CCNB3, CDKN1A, BBC3, SESN1) were selected for qRT-PCR tests both in A549 and H1299 cells. RNA expression levels were normalized to GAPDH. Each sample was analyzed in triplicate and the mean expression level was calculated. Primer sequences for these genes were shown in Table S1.

| Western blot (WB) analysis
Radioimmuno-precipitation assay (RIPA) lysis buffer was used to extract WB analysis total protein from in vitro cultured cells. Proteins

| Statistical analysis
All results were expressed as mean ± SD. and analyzed using Graphpad Prism7 software. All experiments were repeated at least three times. Student's t test was used to determine the significance between two groups. One-way analysis of variance with Tukey test was used for multiple comparisons. *P < 0.05, **P < 0.01 and ***P < 0.001 were considered statistically significant.

| Xa suppresses cell proliferation, migration and invasion in NSCLC cells
We firstly investigated the anti-proliferative effect of Xa on NSCLC cell lines via CCK-8 assays and found Xa dramatically suppressed cell growth in a dose-and time-dependent manner ( Figure 1A). The IC50 values were calculated for 12, 24 and 48 hours time points (Table S2).
Based on IC50 of 24 hours, we determined the Xa concentration used for the following experiment: the low concentration (Xa+) means 1 × 24 h IC50 (10 μmol/L for A549 and 4 μmol/L for H1299), the high concentration (Xa++) means 2× 24 hours IC50 (20 μmol/L for A549 and 8 μmol/L for H1299). We subsequently performed colony formation assays and confirmed that Xa treatment suppresses the growth of A549 and H1299 cells ( Figure 1B). Moreover, we found that Xa treatment remarkably changed cell morphology ( Figure 1C).
Likewise, we used Hoechst 33258 to identify the transitions of a cell's nuclear morphology and found that Xa treatment led to an obvious increase of nuclear pyknosis ( Figure 1D).
To further investigate whether Xa inhibits NSCLC cell proliferation by inducing cell cycle arrest or apoptosis, flow cytometry was performed to determine the cell cycle distribution and showed that A549 cells were mainly blocked in G0/G1 phase while H1299 cells were mainly blocked in G2/M phase when treated with Xa, which may be attributed to the different sensitivity to Xa ( Figure 1E). As shown in Figure 1F, a significant increase of apoptosis ratio was observed in both A549 and H1299 cells after treatment with Xa for 24 hours.
Additionally, we explored the effect of Xa on cell migration and invasion in NSCLC cells by using the wound scratch and transwell assays. Highly significant decreases in both migration (Figure 2A,B) and invasiveness ( Figure 2C,D) were noted in Xa treated cells compared with control. Altogether, our in vitro findings suggested that Xa may act as an important antitumor agent for NSCLC.

| Xa significantly activates p53 pathway and suppresses E2F targets in A549 cells
In order to explore molecular signature of NSCLC cells in responses to Xa, we constructed six cDNA libraries for RNA-seq, with three replicates of Xa treatment and three replicates of controls in A549 cells. In total, 259.8 million 150 bp paired-end clean reads were generated by RNA-seq after removing low-quality sequences (Table S3).
On average, ~40% of the clean reads were mapped to the human genome (Table S3). After filtering with quantitated FPKM, 16 770 genes were considered to be expressed and selected for further analysis (Table S4). Among these genes, more than 88 % had FPKM values in the range of 1-100 for each sample ( Figure S1A).
In total, we identified 78 up-regulated and 55 down-regulated DEGs in Xa treatment compared with control ( Figure S1B; Table S5).
GSEA indicated that the most up-regulated genes are enriched in p53 pathway, which means that Xa can activate p53 pathway in A549 cells ( Figure 3A). This is consistent with previous studies. Moreover, GSEA showed that Xa significantly suppresses E2F targets ( Figure 3B), G2M checkpoint ( Figure 3C) and MYC targets ( Figure 3D). It's worth noting that the down-regulated DEG BARD1 appears twice, which may be an uppermost contributor to this signature.

| Xa inhibits BARD1 both at the mRNA and protein level
To

| Xa inhibits HR pathway and synergizes with Cis
We next examined the effect of Xa on the HR pathway. γH2AX was used as an indicator of DNA damage. WB results showed that both Xa and Cis cause DNA damage in a dose-dependent manner ( Figure 5A). Besides BARD1, we also found Xa can suppress the expression of BRCA1 and RAD51 at the protein level ( Figure 5A), which indicates that Xa could inhibit HR pathway in NSCLC cells. However, cells treated with Cis did not possess significant effect on the expression level of these two proteins. (Figure 5A).

F I G U R E 2 Xa suppresses cell migration and invasion in NSCLC cells. (A and B)
Cell wound scratch assay showed the migration ability of A549 and H1299 cells were significantly decreased after treatment with Xa. (C and D) Transwell assay demonstrated Xa inhibited cell invasion of A549 and H1299. Experiments were repeated three times. *P < 0.05; **P < 0.01; ***P < 0.001 Considering the tight relationship between HR and Cis sensitivity, we hypothesized that Xa may increase NSCLC cells sensitivity to Cis by synergistic effect. We used Calcusyn to analyze the synergy between Xa and Cis. The results showed Xa and Cis exhibited a strong synergistic effect when they were applied to NSCLC cells at a concentration of 1:3 ( Figure 5B,C). These results indicated that this antitumor effect might be related to the inhibition of BARD1 and HR pathway. WB results confirmed that BARD1 was successfully overexpressed ( Figure 6A), and BARD1 OE partially reversed the down-regulation of BRCA1 and RAD51 caused by Xa ( Figure 6A). Moreover, CCK-8 assays demonstrated that BARD1 OE weakened the effect of Xa + Cis, but had no significant effect on single Cis treatment ( Figure 6B). Then, we performed an additional experiment through changing the fresh medium after 24 hours of drug treatment and testing after another 24 hours. The results preliminarily showed the suppressed state of DNA damage repair, which was almost accordance with the previous experiment ( Figure 6B). It is worth noting F I G U R E 3 The GSEA plots indicate Xa significantly activates p53 pathway (A) and suppresses E2F targets (B), G2M checkpoint (C) and MYC targets (D) in A549 cells (FDR < 0.05). The positive normalized enrichment score (NES) indicates higher gene expression in the Xa treatment cells compared with control, and the negative NES was opposite. Contributing DEGs are shown that the extremely low concentration of Xa we used did not significantly cause cell proliferation inhibition and DNA damage, but already had the ability to increase Cis-induced apoptosis ( Figure 6B).

| Xa inhibits HR pathway by down-regulating BARD1 via JAK2-STAT4 pathway
These data indicated a unique and pivotal role of BARD1 during Xa taking its effect.
In order to verify the crucial function of HR, we knocked down RAD51 (si-RAD51) on the basis of BARD1 OE to simulate a HRdeficient status. WB results showed that RAD51 was successfully knocked down ( Figure S2A). CCK-8 assays demonstrated that silencing RAD51 reversed the effects of BARD1 OE ( Figure S2B). These findings confirmed that BARD1 exerts its effects through the HR pathway.
We next focused on how Xa inhibits the expression of BARD1.
We used PROMO (alggen.lsi.upc.es) to identify STAT4 as a potential transcription factor for BARD1. Further experiment showed that after Xa treatment, expression of JAK2 and STAT4 did not change, but the phosphorylation of JAK2 and STAT4 was significantly reduced ( Figure 6C), which behaves similar to JAK2 inhibitor TG101348 ( Figure 6C). These data demonstrated that Xa may repress the expression of BARD1 via JAK2-STAT4 pathway.

| D ISCUSS I ON
Xa is a sesquiterpene lactone compound and it can work as an alkylating agent because of its alpha-methylene-gamma-butyrolactone moiety (αMγL) structure. This structural basis makes Xa possess a great potential for inducing DNA damage. 26 Previous investigations demonstrated that Xa exerts its anticancer effect by inducing DNA damage, reticulum stress, oxidative stress and so on. [19][20][21]27 However, the effects of Xa on the transcriptome of cancer cells have not been reported. In this study, we aimed to elucidate the antitumor effects of Xa and its underlying mechanisms in NSCLC cells. Firstly, we found that Xa inhibits cell proliferation, migration, invasion and progression of the cell cycle in NSCLC cells. Subsequently, we used RNA-seq to assess the changes of gene expression profiles induced by Xa treatment in NSCLC cells.
We identified 133 genes to be differentially expressed (78 up-regulated and 55 down-regulated in Xa treatment). And these DEGs were mainly involved in DNA damage response, cell apoptosis and p53 signaling pathway. Moreover, GSEA showing that Xa significantly activates p53 pathway and suppresses E2F targets, G2M checkpoint and MYC targets in A549 cells. It has been reported that Xa induces apoptosis via p53 signaling pathway, which was consistent with our results. 28 Therefore, we focused on exploring the underlying molecular mechanism of Xa's antitumor effects.
For this purpose, 12 DEGs which may be the potential targets of Xa were selected for qRT-PCR and WB to verify the reliability of RNA-seq. Among these genes, the down-regulated gene BARD1 seems to be particularly interesting in light of our results and its vital role in HR pathway. The relationship between Xa and HR pathway has not yet been reported. Thus, we chose it for further study.
Recent studies have revealed the important links between BARD1 and HR pathway. BARD1 and BRCA1 are known to function in HR by forming a stable heterodimer through their RING-finger F I G U R E 5 Xa inhibites HR pathway and synergizes with Cis. (A) WB analysis of γH2AX, BARD1, BRCA1 and RAD51, showing that both Xa (left panel) and Cis (right panel) caused DNA damage in a dose-dependent manner. Xa inhibited HR pathway while Cis did not. (B) CCK-8 assay, preliminary displaying the synergy between Xa and Cis. (C) Calcusyn was used to calculate the CI value of Xa and Cis, demonstrating that Xa and Cis exhibited a strong synergistic effect in NSCLC cells. Experiments were repeated four times. *P < 0.05; **P < 0.01; ***P < 0.001 domains then co-localize at nuclear foci. 11,29 The role of BRCA1-BARD1 during HR is to produce a single-stranded template by facilitating the nucleolytic resection of DNA ends, and then recruit other important tumor suppressor complex, including BRCA2-PALB2 and RAD51. [30][31][32] BRCA1 requires BARD1 to function, and BRCA1 is unstable and rapidly degrades in the absence of BARD1. 31 Meanwhile, BRCA1-BARD1 can interact with RAD51 and enhances its activity. 11 RAD51 executes an essential task in HR, interacts with other HR-related proteins and captures the backup DNA copy, which matching the sequence of the broken strand with a homologous sequence in the intact DNA double helix. 11 Considering the central role of RAD51, it is universal and reasonable that RAD51 can affect most proteins which are active in HR. [33][34][35][36] Strikingly, RAD51 is overexpressed in most cancer cells including NSCLC, and its overexpression generally leads to the genomic instability and resistance to DSB-inducing therapies. 37 In the present study, we highlighted that Xa could suppress BARD1 expression, and subsequently weaken the expression of BRCA1 and RAD51, indicating the involvement of HR pathway in Xa's antitumor effects.
It is known that DNA damage-inducing agents Cis had no significant effect on the HR pathway. Meanwhile, the inhibition of HR pathway in NSCLC cells is associated with the increased cell sensitivity to Cis. In fact, there are historical precedents for the successful treatment of cancer by inhibiting DNA repair pathway. For instance, poly ADP-ribose polymerase (PARP) inhibitors are effective drugs against breast cancer patients who have lost HR capability due to the mutations of BRCA1/2 but simultaneously with enhanced activity of PARP. 38,39 Therefore, we explored the synergistic effect between Cis and Xa and found that Xa synergized with Cis when the ratio of Xa/Cis kept 1:3. And this synergistic effect could be partly reversed after BARD1 OE ( Figure S2A,B), which further confirmed our speculation. Taken together, these results suggest that Xa can inhibit HR via BARD1/BRCA1/RAD51 axis and enhance NSCLC cell sensitivity to Cis. This finding may reduce cytotoxic effects and improve clinical outcomes of Cis in the future.
The Janus kinase (JAK)-signal transducer and activator of transcription (STAT) signaling pathway plays a vital role in many biological processes, including regulating angiogenesis, modulating inflammatory, immune responses, hematopoiesis and even oncogenesis. Recent studies have shown that STATs can function as transcription factors after phosphorylating, dimerizing and nucleus translocating. 40,41 Importantly, STAT4 is predicted as a transcription factor for BARD1 in NSCLC cells by using the PROMO, and there is evidence that the expression of STAT4 was persistently activated in NSCLC cells through its co-activator JAK2. 42,43 Interestingly, it has been reported that Xa is a covalent and selective inhibitor of JAKs. 44 Thus, to further gain insights into the mechanism of Xa-mediated inhibition of BARD1, we characterized the expressions of key proteins in JAK2-STAT4 pathway. Our investigation demonstrated that disruption of JAK2-STAT4 pathway by Xa contributed to its inhibitory F I G U R E 7 A schematic model of the molecular mechanism underlying the role of Xa in NSCLC cells. Left panel shows drug resistant to Cis of NSCLC due to the abnormal enhancement of DNA repair capacity in cancer cells, leading to cell survival. Right panel shows Xa suppresses HR and synergizes with Cis. Mechanistically, Xa inhibites the conversion of JAK2 to p-JAK2, making STAT4 unable to form its active form p-STAT4. Then, the expression of BARD1 is decreased due to the lack of the transcription factor p-STAT4, leading to reduction of the BARD1 protein. As a result, assembling of BARD1-BRCA1 complexes and the recruitment of RAD51 at double-strand breaks (DSBs) sites were decreased, resulted in degradation of BRCA1 and RAD51. Eventually, the down-regulation of BARD1 triggered by Xa decreases the ability of cells against DSBs and increases the NSCLC cells sensitivity to Cis, leading to cell apoptosis effects on the expression of BARD1 and HR pathway in NSCLC cells ( Figure 7).
In addition, it is worth noting that Xa can inhibit DNA damage repair as a DNA damage inducer, which might account for the different cell cycle arrests under the treatment of Xa in NSCLC cells. Actually, similar results had also been previously reported. 20 Mechanistically, overproduction of DSBs can cause a strong cell cycle arrest at the G0/G1 phase, 45 while HR-deficient cells are easily blocked at the G2/M phase because HR occurs between sister chromatids. 12 In the present studies, we speculated that high concentration of Xa potently induced cell cycle arrest at G0/G1 checkpoint in A549 cells via producing DSBs, which covered up the effect of HR deficient. And low concentration of Xa induced cell cycle arrest at G2/M checkpoint in H1299 cells via inhibiting HR pathway, because the IC50 value of H1299 is much lower than A549. These findings suggested that DNA damage induced by Xa should not be ignored. In fact, RNA-seq and validation results indicated that DNA damage caused by Xa leads to a series of gene expression changes through p53 pathway. Numerous p53 target genes were up-regulated by Xa treatment, including pro-apoptotic genes ZMAT3, FAS, 46 BBC3, 47 cell cycle regulatory genes CCNB3, 48 CDKN1A 49 and anti-stress gene SESN1. 50 However, similar cell proliferation inhibition were also observed in the p53-deficient H1299 cells, the effect of p53 on these changes remains to be determined. It had been reported that p53-deficient tumor cells rely much more on ATR/Chk1 to arrest cell progression. 19 Therefore, More experiments are needed to further elucidate the influence of Xa on gene expression.
In summary, here we first proposed the potential of Xa as a sensitizer for Cis because of its significant ability to inhibit HR.
Subsequently, our findings indicated that Xa inhibits HR via JAK2/ STAT4/BARD1 axis and enhances cell sensitivity to Cis (Figure 7). These data help further reveal the antitumor effect and mechanism of the sesquiterpene lactone compound Xa on NSCLC cells, which provides insight for Xa as a potential anticancer agent for NSCLC, especially in combination with routine chemotherapy. However, further investigations especially in vivo evidence are needed.

ACK N OWLED G EM ENTS
This study was supported by the grants from the National Natural

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supports the findings of this study are available in the supplementary material of this article.