NF-κB activation is an early event of changes in gene regulation for acquiring drug resistance in human adenocarcinoma PC-9 cells

Gefitinib and erlotinib are epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKIs). Although EGFR-TKIs are effective as anti-cancer drugs, cancer cells sometimes gain tolerance to the drugs. Previous studies suggested that the fibroblast growth factor receptor (FGFR)-signaling pathway could serve as compensation for the EGFR-signaling pathway inhibited by EGFR-TKIs. Our study further suggested that FGF2, a FGFR ligand, leaked out from naïve cells killed by gefitinib could initiate the FGFR-signaling pathway in surviving cells; i.e., altruistic survival may occur in naïve cells immediately after EGFR-TKI treatment. Altruistic survival may be temporal, and cells need to change their gene regulation toward gaining resistance to EGFR-TKIs. Changes in such gene regulation after EGFR-TKI treatment are poorly understood. In this study, we examined early events of such gene regulation changes in human adenocarcinoma PC-9 cells that are capable of changing their nature from susceptibility to resistance to EFGR-TKIs. Our study indicated that activation of nuclear factor-kappa B (NF-κB) occurred in the cells immediately after EGFR-TKI treatment and also by gene silencing against oncogenic EGFR; and, MG132 treatment for inhibiting NF-κB activation affected cell viability. Taken together, our findings (including the previous study) suggest that altruistic survival and NF-κB activation might be vital for initiating the acquisition of EGFR-TKI resistance.


Introduction
Epidermal growth factor receptor (EGFR) is a transmembrane protein containing a tyrosine kinase domain in its intracellular region. Binding of epidermal growth factor (EGF) to EGFR triggers the dimerization of EGFR followed by autophosphorylation of its tyrosine residues, thereby activating its downstream signaling-pathway involving ERK/MAPK and PI3K-AKT-mTOR [1,2]. The EGFR-signaling system plays important roles in various vital functions such as cell proliferation and cell survival.

Transfection of reporter plasmids and reporter assay
The day before transfection, cells were trypsinized, diluted with medium without antibiotics, and seeded onto 96-well culture plates (0.5x10 4 cells/well). The pGL4 vectors carrying various response elements followed by the Photinus luciferase reporter gene (40ng/well) (Promega, Fitchburg, WI, USA) were introduced together with the phRL-TK plasmid (10ng/ well) (Promega) into cells by using a Lipofectamine2000 transfection reagent (Thermo Fisher Scientific) according to the manufacturer's instructions. After 24h-incubation, culture media were replaced with fresh medium containing 10μM of gefitinib or erlotinib, and further incubation was carried out for 6h. After the treatment, cells were lysed, and the expression levels of the Photinus and Renilla luciferase reporter genes were examined by a Dual-Luciferase Reporter Assay system (Promega) according to the manufacturer's instructions. The luminescent signals were measured using a Synergy H1 Multi-Mode Reader (BioTek, Winooski, VT, USA). The pGL4 vectors used in this study were as follows (abbreviated name used in this study):

Gene silencing by RNAi
To knockdown EGFR, a small interfering RNA (siRNA) against the gene as a whole was purchased from Thermo Fisher Scientific (catalog-number: s564). In addition, si746/50_3D10 was used for oncogenic allele-specific silencing against a deletion-type of EGFR in PC-9 cells according to the previous study [17]. The sequences of si746/50_3D10 are as follows: The siRNAs (10-20nM final concentration) were transfected into cells by a Lipofecta-mine2000 transfection reagent (Thermo Fisher Scientific) according to the manufacturer's instructions. The cells were incubated for 24h and subjected to subsequent experiments.

Total RNA preparation and reverse transcription (RT)-quantitative polymerase chain reaction (qPCR)
Total RNAs were extracted from cells using a TRI Reagent (MRC, Cincinnati, OH, USA) according to the manufacturer's instructions, and subjected to complementary DNA (cDNA) synthesis using oligo (dT) 15 primers (Promega) and a Superscript III reverse transcriptase (Thermo Fisher Scientific) according to the manufacturer's instructions. The resultant cDNAs were subjected to qPCR using a StepOne Plus Real-Time PCR system (Thermo Fisher Scientific) with a FastStart Universal SYBR Green Master (Roche Diagnostics, Basel, Switzerland) and Perfect Real Time primers (TAKARA BIO, Kusatsu, shiga, Japan) according to the manufacturers' instructions. The Perfect Real Time primers used were as follows (TAKARA BIO primer-set IDs): EGFR (HA159668), GAPDH (HA067812), GFOD1 (HA205357).

Western blot analysis
Cells were washed with D-PBS (Wako) and lysed in RIPA buffer (Thermo Fisher Scientific) containing 1x Protease/Phosphatase Inhibitor Cocktail (Cell signaling technology). The lysate was incubated on ice for 5min, passed 10 times through a 26G needle using a 1ml-syringe and centrifuged at 14,000xg for 15min at 4˚C. The resultant supernatant (cell lysate) was collected.

Cell viability assay
Cell viability was measured by a CellTiter 96 1 Aqueous Non-Radioactive Cell Proliferation Assay (Promega) according to the manufacturer's instructions.

ELISA analysis
Conditioned media from PC-9 cells were collected and centrifuged at 2,000xg for 15min at room temperature. The supernatant was transferred into an Amicon 1 Ultra centrifugal filter 10k (Merck Millipore), and subjected to concentration by centrifugation at 14,000xg for 15min. The level of FGF2 in the concentrated medium was measured by a Human FGF basic Quantikine ELISA kit (R&D SYSTEMS, Minneapolis, MN, USA) according to the manufacturer's instructions.

Prediction of transcription factor binding sites
Transcription factor binding sites in a putative promoter region of the GFOD1 gene was predicted by using the TFBIND software according to the instructions [18].

Reporter assay for activated transcription factors in PC-9 cells after EGFR-TKI treatment
When naïve cells acquire tolerance under the presence of harmful drugs, gene expression changes should occur in the cells. Identification of such gene expression changes immediately after drug treatment is of particular importance, but not well performed. To explore such a change in gene regulation, human adenocarcinoma PC-9 cells may be a good cellular model: this is because the cells are capable of changing their nature from susceptibility to resistance to EFGR-TKIs [8,9]. To search for early events of gene regulation immediately after EGFR-TKIs treatment, we focused on transcription factors and examined the effects of EGFR-TKIs on transcription factors by means of a reporter-gene assay. We used the pGL4 reporter plasmids carrying various transcription factor response elements linked to the luciferase reporter gene, and transfected them into naïve PC-9 cells. The expression of the reporter gene was examined after the cells were treated with gefitinib, an EGFR-TKI, for 6h. As a result, the expression of the reporter gene carrying the response element of nuclear factor-kappa B (NF-κB) was significantly upregulated under the presence of gefitinib (Fig 1a). A slight increase in the expression of the reporter gene carrying the cAMP response element (CRE) was also detected after gefitinib treatment. When the reporter genes carrying the antioxidant response element (ARE), serum response element (SRE) and sis-inducible element (SIE) were examined in the presence of gefitinib, a significant decrease in the reporter gene expression was seen. Note that not only gefitinib but also erlotinib had an effect on the NF-κB response element and the serum response element (SRE) (Fig 1b and 1c); and the dose-dependent effect of erlotinib on NF-κB activation was also detected (Fig 1b). Consequently, gefitinib and erlotinib exhibited similar results in the reporter gene assay.

Effects of extracellular FGF2 on cell survival and on activation of transcription factors
The previous study suggested that FGF2, a ligand of FGFR, could leak out from neighboring cells killed by gefitinib and could trigger the initiation of the FGFR-signaling pathway compensating the EGFR-signaling pathway inhibited by gefitinib [8]. Consistently, the gefitinibdose-dependent increase of extracellular FGF2 that leaked from gefitinib-treated cells (Fig 2a), and also the positive contribution of FGF2/FGFR to cell survival (Fig 2b and 2c) were reproducibly observed in this study. In addition, little or no increase in endogenous FGF2 was detected in gefitinib-treated cells (S1 Fig), suggesting that the increase in extracellular FGF2 after gefitinib treatment was not caused by increasing endogenous FGF2. To further prove the (a) Luciferase reporter assay. The pGL4 vectors, which encode response elements for transcription factors (indicated) or a minimal promoter (minP; a negative control) linked with the Photinus luciferase gene, were transfected together with phRL-TK carrying the Renilla luciferase gene (as a reference) into PC-9 cells. After 24h-incubation, the cells were exposed to 10μM gefitinib or DMSO (a vehicle) for 6h and subjected to a dual-luciferase assay. The activity of the Photinus luciferase (P luc.) was normalized to that of the Renilla luciferase (R luc.). Data are shown as mean ± standard deviation (SD) (n = 4). Significant difference between gefitinib and DMSO was examined by means of two-tailed Student's t-test ( Ã <0.05, ÃÃ p<0.01, ÃÃÃ p<0.001). (b, c) Effects of gefitinib or erlotinib on transcription factors. The GL4-plasmid carrying the NF-κB response element (b) or the serum response element (SRE) (c) was co-transfected with phRL-TK into PC-9 cells, and after 24h-incubation, the cells were treated with 0, 0.1, 1.0 and 10μM of gefitinib or erlotinib for 6h. Dual-luciferase assay followed by normalization were carried out as in a. Normalized data were further normalized to the data obtained from cells treated with no EGFR-TKIs (0 μM) as 1. Data are shown as mean ± standard deviation (SD) (n = 3). vital effect of leaked-out (dead-cell-derived) FGF2, PC-9 cells were treated with gefitinib together with anti-FGF2 antibodies for neutralizing extracellular FGF2. The results indicated that cell viability was significantly reduced with increasing in the dose of anti-FGF2 antibodies (Fig 2d); thus, the neutralization of leaked-out FGF2 resulted in worsening of cell viability. Therefore, the findings strongly suggested that extracellular FGF2 leaked out from dead cells was essential to survive cells which were still alive under the presence of gefitinib.
Based on the positive contribution of FGF2 to cell survival, we examined whether extracellular FGF2 was capable of activating NF-κB in the absence of gefitinib. PC-9 cells that were transfected with the reporter gene carrying the NF-κB response element (NF-κB-RE) were treated with various concentrations of FGF2, and the expression of the reporter gene was examined. As a result, no significant difference in the reporter gene expression among the various doses of FGF2 was detected (Fig 2e). When the reporter gene carrying SRE was examined, an increased tendency of the expression with increasing in FGF2 was observed (Fig 2f), suggesting SRE reactive to FGF2. To further confirm, we examined the effect of gefitinib on NF-κB activation under the presence of PD173074, a FGFR-tyrosine kinase inhibitor. The results indicated that the activation of NF-κB was hardly affected by FGFR inhibition (Fig 2g). Therefore, the findings suggested that neither FGF2 nor FGFR-signaling pathway was implicated in NF-κB activation under the presence of gefitinib.

Effects of EGFR knockdown on activation of NF-κB
The contribution of gefitinib to the activation of NF-κB was obvious from the current findings described above. However, it was still unclear whether gefitinib triggered the activation of NF-κB through loss of EGFR function or else by unknown function(s) of gefitinib itself. To address the question, we conducted gene silencing against EGFR by RNA interference (RNAi). Because PC-9 cells carry oncogenic E746_A750del EGFR and normal EGFR alleles, we performed RNAi specific for the oncogenic EGFR allele as well as for all EGFR alleles (including the oncogenic EGFR allele). After RNAi knockdown, the reporter gene carrying NF-κB-RE was introduced. If functional knockdown of EGFR were related to the activation of NF-κB, gene silencing of EGFR should result similar to that obtained from gefitinib treatment. If gefitinib itself had unknown functions involved in NF-κB activation, RNAi knockdown could lead to little activation of NF-κB.
As shown in Fig 3c, the results indicated that the reporter gene was upregulated under knockdown of the oncogenic EGFR (si3D10). Interestingly, total EGFR knockdown exhibited no upregulation of the reporter gene (siEGFR in Fig 3c), and gefitinib treatment under such total EGFR knockdown had little effect on the expression of the reporter gene (Fig 3d); thus, suggesting little possibility of gefitinib's own unknown function. Taken together, the findings suggested that the loss of oncogenic EGFR function might be involved in the activation of NF-κB.

Cell viability under inhibition of NF-κB activation
Inhibition of NF-κB might have some influence on cell survival if activated NF-κB had some relationship with the survival of PC-9 cells under the presence of EGFR-TKIs. We investigated between α-FGF2 and a control IgG was examined by Student's t-test (two-tailed; Ã <0.05, ÃÃ p<0.01, ÃÃÃ p<0.001). (e, f) Effects of extracellular FGF2 on transcription factors. The pGL4-NF-κB-RE and pGL4-SRE plasmids were introduced together with phRL-Tk into PC-9 cells. After 24h-incubation, the cells were treated with various levels of FGF2, and further incubation was carried out. After 6h, dual-luciferase assay and data analyses were carried out as in Fig 1b  (n = 3). (g) Influence of PD173074 on NF-κB activation. The pGL4-NF-κB-RE and phRL-Tk plasmids were transfected into PC9 cells as in e, and the cells were treated with PD173074 for 1h prior to gefitinib treatment. After 6h gefitinib treatment, dual-luciferase assay was carried out as in Fig 1b (n = 3). https://doi.org/10.1371/journal.pone.0201796.g002

Fig 3. Effect of gene silencing against EGFR on NF-κB activation.
(a) RNAi knockdown. PC-9 cells were subjected to gene silencing with siRNAs against total EGFR (siEGFR) or oncogenic mutant EGFR alleles [si746/50_3D10 (si3D10) targeting the 746/750 deletion in exon 19] in PC-9 cells. Non-silencing siRNAs (siControl: siC) were also examined as a negative control. 24h after introduction of siRNAs, EGFR was examined by western blotting. Alphatubulin was examined as an internal control. (b) EGFR mRNA levels. Indicated siRNAs were transfected into PC-9 cells as in a. The level of EGFR was examined by RT-qPCR and analyzed by the delta-delta Ct method using the data of Gapdh as a reference. The data were further normalized to the data obtained by siControl (siC) as 1. Data are shown as mean ± SD (n = 3). (c) Activation of NF-κB. The pGL4-NF-κB-RE and phRL-Tk plasmids together with indicated siRNAs (20nM) were introduced into PC-9 cells. After 24h incubation, Dual-luciferase assay was carried out as in Fig  1b (n = 6). (d) Effect of gefitinib on NF-κB activation under EGFR knockdown. Reporter plasmids and siRNAs were introduced into PC9 cells as in c. The cells were treated with 0, 0.1, 1.0 and 10μM of gefitinib (Gef). Dual-luciferase assay was carried out as in Fig 1b (n = 6). Significant difference in cell viability between gefitinib and DMSO treatment was examined by one-way analysis of variance (ANOVA) (Dunnett's test; Ã p<0.05). n.s.: not significant.
if the cell viability of PC-9 cells was affected by inhibiting NF-κB activation under the presence of gefitinib. NF-κB generally forms a protein-complex together with one of the Rel proteins, and the complex was inactivated and sequestered in the cytoplasm by binding of Inhibitory κB (IκB). When cells receive stimuli, e.g., a proinflammatory cytokine such as tumor necrosis factor-alpha (TNFα), the stimulation triggers the activation of IκB kinases (IKKs) followed by phosphorylation of IκB. The phosphorylated IκB becomes a target of ubiquitination and/or proteasomal degradation, thereby freeing the NF-κB complex as an active form from IκB; and then, the NF-κB complex can move into the nucleus and work as a transcription factor (refer to review articles [19][20][21][22]).
The phosphorylation of IκB immediately after gefitinib treatment was detected (Fig 4a). Thus, the result suggested that a common process of activation of NF-κB occurred in PC-9 cells after gefitinib treatment. To block the activation process, we used MG132, a proteasome inhibitor capable of inhibiting the degradation of IκB. The inhibitory effect of MG132 on IκB degradation was consistently detected (Fig 4b); and, MG132 treatment caused reduction of the expression of the reporter gene carrying NF-κB-RE even in the presence of gefitinib (Fig 4c), indicating that NF-κB activation due to gefitinib was suppressed by MG132. As additional data, neither the presence nor absence of PD173074 had influence in the results, suggesting little contribution of FGFR to NF-κB activation.
When PC-9 cells were treated with gefitinib and MG132, in other words, when the cells were treated with gefitinib without NF-κB-activation, a significant decrease in cell viability was detected at a lower concentration of gefitinib (0.1μM) under the presence of 0.3μM of MG132 (Fig 4d). Therefore, the findings suggested that NF-κB activation immediately after EGFR-TKIs treatment might be implicated in cell survival.

Discussion
Drug resistance is a major problem in chemotherapy, and eventually reduces therapeutic options to patients. To solve the problem, elucidation of the mechanism by which cells can acquire drug resistance is vital; and intracellular changes immediately after drug treatment are of particular importance. PC-9 cells examined in this study are capable of developing resistance to EGFR-TKIs by exposure of the drugs [8,9]. Thus, PC-9 cells may be a useful cellular model for examining the mechanism of acquired drug resistance and provide us with the opportunity to investigate intracellular changes in the early stages of EGFR-TKIs treatment.
Previous studies suggested that the FGFR-signaling pathway was capable of working as a surrogate pathway for the EGFR-signaling pathway inhibited by EGFR-TKIs [8,15,16,23], and our study further suggested that FGF2, a key ligand that triggers the initiation of the surrogate FGFR-signaling pathway, would be supplied from neighboring cells killed by EGFR-TKIs [8]. Thus, altruistic survival based on neighbors' self-sacrifices may have an important role in cell survival early in the treatment of EGFR-TKIs (Fig 2).
Even if cells could escape from the first exposure of EGFR-TKIs by temporal altruistic survival, the cells would be killed by the drugs unless they could develop resistance to the drugs. Therefore, surviving cells need to change their properties to be resistant to EGFR-TKIs as soon as possible; and of such changes, changes in gene regulation would be vital.
Our current study indicated NF-κB activation immediately after EGFR-TKIs treatment as an initial event of such changes in PC-9 cells (Fig 1). The activation of NF-κB was independent of the FGFR-signaling pathway, which works as a surrogate pathway for the EGFR-signaling pathway inhibited by EGFR-TKIs (Fig 2). The arrest of EGFR-tyrosine kinase activity appeared to be a key step in the activation of NF-κB.
PC-9 cells possess oncogenic E746_A750del EGFR and normal EGFR, and the oncogenic EGFR promotes tyrosine kinase activity. RNAi knockdown specific for the oncogenic EGFR resulted in the activation of NF-κB (Fig 3). Since the oncogenic E746_A750del EGFR is the major target of EGFR-TKIs, inhibition of oncogenic EGFR function might trigger the onset of NF-κB activation. The interesting point to note here is that total EGFR knockdown by RNAi exhibited little activation of NF-κB even under the presence of gefitinib (Fig 3c and 3d). The finding led us to the possibility that normal EGFR might have some role in the activation of NF-κB after inhibition of oncogenic EGFR. To clarify the possibility, further studies need to be carried out in the future.
It is of importance to find evidence on an association between NF-κB activation and cell survival (viability) under the presence of EGFR-TKIs. MG132 is a proteasome inhibitor that blocks the degradation of IκB binding to NF-κB and consequently inhibits NF-κB activation. When cells were treated with MG132 together with gefitinib, the cells exhibited a decrease in cell viability (Fig 4). This suggests that NF-κB activation might be implicated in cell survival under the presence of EGFR-TKIs. In addition, the findings were consistent with the previous study [13] that identified many genes working in NF-κB-activating cascades as genes associated with resistance to EGFR-TKI. Taken together, activation of NF-κB might participate in early steps to develop resistance to EGFR-TKIs, and might trigger genetic changes for acquiring drug resistance.
NF-κB is a transcription factor and is involved in various gene regulations. Likewise, NF-κB that is activated by EGFR-TKIs presumably participates in gene regulation. As an instance, the expression of the Glucose-fructose oxidoreductase domain containing 1 (GFOD1) gene, which has many predicted NF-κB binding sites in its putative promoter region (S1 Table), was markedly upregulated 24h after gefitinib treatment (S2 Fig) and also increased in gefitinibresistant PC-9 cells relative to naïve cells [8]. Thus, activated NF-κB may contribute to alteration of gene expression during early stages of EGFR-TKIs treatment and over the course of development of EGFR-TKI resistance.
Based on the findings of the previous and present studies concerning the early stages of EGFR-TKIs treatment, a hypothetical scenario may be considered as follows (Fig 5): after EGFR-TKIs treatment, many cells are killed and the cells leak out survival factors such as FGF2; and simultaneously, neighboring (alive) cells may manage to survive from the first exposure of EGFR-TKIs by leaked-out FGF2 and also by activated NF-κB. FGF2 can initiate the FGFR-signaling pathway as a surrogate pathway for the EGFR-signaling pathway inhibited by EGFR-TKIs. NF-κB and FGF2 may independently contribute to cells for survival in the early stages of EGFR-TKIs treatment. Cells which survived successfully, thereafter, change gene regulation to develop and maintain tolerance to EGFR-TKIs; and NF-κB might play an important role as an initial transcription factor there. Finally, the cells would acquire drug resistance, in which FGF2-FGFR autocrine system may work as a major vital pathway [15,16,23]. In conclusion, altruistic survival and NF-κB activation, both of which are induced by EGFR-TKIs treatment, may be essential for cells to survive early in EGFR-TKIs treatment and then to obtain resistance to EGFR-TKIs.
Our current study shed light on the activation of NF-κB immediately after EGFR-TKIs treatment in PC-9 cells capable of acquiring EGFR-TKI resistance. For further investigation, the putative factors mediating the activation of NF-κB under EGFR inhibition still remain unknown. To further clarify the mechanism of acquired EGFR-TKIs resistance, more extensive studies including identification of such factors need to be carried out in the future.
Supporting information S1 Fig. EGFR and FGF2 under the presence of gefitinib. PC9 cells were treated with 0, 0.1, 10μM of gefitinib. 24h after gefitinib treatment, EGFR and FGF2 were examined by western blotting. GAPDH was examined as an internal control. (PDF) Fig 5. Schematic drawing of hypothetical early events in EGFR-TKIs-exposed PC-9 cells. After EGFR-TKIs treatment, many PC-9 cells are killed and the dead cells leak out FGF2 that works as a survival factor; and simultaneously, neighboring (alive) cells manage to survive from the first exposure of EGFR-TKIs by leaked-out FGF2 and also by NF-kB activation. FGF2 can trigger the initiation of the FGFR-signaling pathway that is a surrogate pathway for suppressed EGFR pathway. The FGFRsignaling pathway and NF-kB independently contribute to cells for survival in the early stages of EGFR-TKIs treatment. Cells that survived successfully, thereafter, need to change gene regulation to maintain resistance to EGFR-TKIs; and NF-kB may play an important role as an initial transcription factor there. Finally, the cells acquire complete drug resistance, in which FGF2-FGFR autocrine system may work as a major vital pathway.
https://doi.org/10.1371/journal.pone.0201796.g005 S2 Fig. GFOD1 expression after gefitinib treatment. PC-9 cells were treated with gefitinib (indicated) for 24h. The expression level of GFOD1 was examined by RT-qPCR. The data were analyzed by the delta-delta Ct method using the data of GAPDH as an internal reference and with the delta Ct obtained from the data of DMSO as a control as 1. Data are shown as mean +/-SD (n = 3). (PDF)  Table. Predicted NF-κB binding sites in a putative GFOD1 promoter region. NF-κB binding sites in a putative promoter region (1,500bp-long) of GFOD1 were predicted by the TFBIND software. (PDF)