PRIMA-1Met suppresses colorectal cancer independent of p53 by targeting MEK

PRIMA-1Met is the methylated PRIMA-1 (p53 reactivation and induction of massive apoptosis) and could restore tumor suppressor function of mutant p53 and induce p53 dependent apoptosis in cancer cells harboring mutant p53. However, p53 independent activity of PRIMA-1Met remains elusive. Here we reported that PRIMA-1Met attenuated colorectal cancer cell growth irrespective of p53 status. Kinase profiling revealed that mitogen-activated or extracellular signal-related protein kinase (MEK) might be a potential target of PRIMA-1Met. Pull-down binding and ATP competitive assay showed that PRIMA-1Met directly bound MEK in vitro and in cells. Furthermore, the direct binding sites of PRIMA-1Met were explored by using a computational docking model. Treatment of colorectal cancer cells with PRIMA-1Met inhibited p53-independent phosphorylation of MEK, which in turn impaired anchorage-independent cell growth in vitro. Moreover, PRIMA-1Met suppressed colorectal cancer growth in xenograft mouse model by inhibiting MEK1 activity. Taken together, our findings demonstrate a novel p53-independent activity of PRIMA-1Met to inhibit MEK and suppress colorectal cancer growth.


INTRODUCTION
Colorectal cancer (CRC) is the third most common cancer with nearly 1.4 million new cases in 2012. Based on the lesions within colon or rectum, CRC varies in terms of biological characteristics and carcinogenic mechanisms [1]. Several intracellular signaling pathways such as Ras/Raf/MEK/ERK, PI3k/Akt, Wnt/β-catenin and p53 are frequently dysregulated in CRC [2][3][4]. Notably, p53 missense or nonsense mutations constantly occur in approximately 40-50% of sporadic CRC, causing therapeutic resistance and poor prognosis [5].
p53 is a key tumor suppressor regarded as the "guardian of the genome". Once activated in response to a variety of stress, p53 elicits potent inhibition of tumorigenesis, leading to cell cycle arrest, apoptosis, senescence and autophagy [6][7][8]. However, mutant p53 not only abrogates tumor suppressor functions of wild-type p53 but also acquires an oncogenic "gain-of-function" defined as the ability for malignant proliferation, invasion, metastasis and anti-apoptotic effect [9]. Therefore, a promising strategy to treat cancer with mutant p53 via restoring wild-type p53 function has been pursued. Several small molecules such as PRIMA-1, APR-246 (PRIMA-
PRIMA-1 is one of chemical compounds originally identified by screening a library of low-molecularweight compounds [10]. PRIMA-1 Met , as methylated analog of PRIMA-1, could enhance killing efficacy. Biochemical analysis demonstrated that either PRIMA-1 or its methylated version is converted into active product named methylene quinuclidinone (MQ), which in turn covalently binds to several cysteine (Cys) residues in the DNA-binding domains of p53 and restores functional conformation of wild-type p53 [14]. Recent reports revealed other targets of PRIMA-1 Met such as oxidosqualene cyclase and selenoprotein thioredoxin reductase 1 (TrxR1) which are sufficient to suppress cancer cell growth irrespective of p53 mutation status [15,16]. Therefore, identifying potential targets of PRIMA-1 Met is important to the development of new cancer therapy approaches.
The mitogen-activated protein kinase (MAPK) pathway Ras/Raf/MEK/ERK is often aberrantly activated in CRC and plays a critical role in multiple cellular processes including proliferation, transformation, apoptosis and senescence [17]. Generally, constitutive activation of MAPK cascades is stimulated by mutant Ras because activating K-Ras mutations occur in approximately 37-45% CRC [18]. So far, no K-Ras mutant inhibitors are available for clinical practice [19,20]. MEK is a core downstream effector of Ras/Raf kinases and appears as a promising alternative for therapeutic strategy targeting Ras-activated MAPK pathway [21].
In the present study, we reported that PRIMA-1 Met attenuated colorectal cancer cell growth irrespective of mutant p53 status. In p53 null or p53 wild-type HCT116 colorectal cancer cells, we demonstrated that PRIMA-1 Met effectively inhibited cell proliferation and anchorage independent growth. Furthermore, we showed that PRIMA-1 Met directly bound MEK kinase and suppressed its activity both in vitro and in cells. Moreover, in vivo animal experiments confirmed that PRIMA-1 Met inhibited MEK activity to suppress the growth of colorectal cancer xenografts.

PRIMA-1 Met inhibits the proliferation and growth of CRC cells independent of p53 status
The chemical structure of PRIMA-1 Met was shown in Figure 1A. To evaluate p53-independent efficacy of PRIMA-1 Met , we selected a series of CRC cell lines representative of different TP53 heterogeneity, including TP53 wt (HCT116 wt and LOVO), TP53 mut (SW480, DLD-1 and HT29) and TP53 neg (HCT116 neg ). p53 protein expression level and genotypes of these cell lines were presented (Supplementary Figure S1 and Supplementary  Table S1). MTS assay showed that the treatment of different cells with 25 uM PRIMA-1 Met for 12 h or 24 h led to similar suppression of cell growth, indicating that CRC cells with different TP53 status were generally sensitive to PRIMA-1 Met ( Figure 1B). Meanwhile, we treated the cells with different concentrations of PRIMA-1 Met for 24 h. MTS assay showed that PRIMA-1 Met reduced the viability of all six CRC cell lines in a dose-dependent manner ( Figure 1C). Notably, the inhibitory effect of PRIMA-1 Met at concentrations less than 50 umol/L showed no significant difference among the cells ( Figure 1C). However, high dosage of PRIMA-1 Met (75 uM) triggered apoptosis, especially in cells harboring mutant TP53 (Supplementary Figure S2A-S2D).
Next we examined the effect of PRIMA-1 Met on CRC cell growth, and the results showed that PRIMA-1 Met inhibited the growth of three cell lines: TP53 wt HCT116 wt , TP53 mut SW480, and TP53 neg HCT116 neg ( Figure 1D), confirming that the inhibitory effect of PRIMA-1 Met on CRC cell growth is independent of TP53 status.
To further investigate whether the effect of PRIMA-1 Met on cell proliferation is dependent on TP53, we compared the effects of PRIMA-1 Met in SW480 or DLD-1 cells transfected with sh-mock or sh-TP53 plasmid ( Figure  1E and Supplementary Figure S3A). The viability of shmock or sh-TP53 cells exhibited no significant difference ( Figure 1F and Supplementary Figure S3B). Taken together, these results suggest that the inhibitory effects of PRIMA-1 Met on CRC cell proliferation and growth are independent of p53 status.

PRIMA-1 Met inhibits CRC cell colony formation and EGF-induced cell transformation independent of p53 status
Next, we determined the effect of PRIMA-1 Met on anchorage-independent growth of cancer cells. Colony formation assay showed that PRIMA-1 Met markedly inhibited anchorage-independent growth of HCT116 wt , SW480 and HCT116 neg cells in a concentration dependent manner (Figure 2A-2C). In particular, PRIMA-1 Met at 50 uM significantly decreased more than 70% of colony formation in all 3 cell lines compared to control (P<0.05). Even at the lower concentration of 25 uM, PRIMA-1 Met caused approximate 30% decrease of colony formation in soft agar. Our data showed no significant differences in dose-dependent inhibition by PRIMA-1 Met among TP53 wt , TP53 mut and TP53 neg cells.
We further accessed whether PRIMA-1 Met could inhibit EGF-promoted neoplastic cell transformation in immortalized normal cells. We found that PRIMA-1 Met significantly inhibited EGF induced anchorage-independent growth of JB6 P+, an immortalized normal skin cell line with impaired p53 function ( Figure 2D). These data strongly suggest that p53 status is not indispensable for PRIMA-1 Met mediated inhibition of anchorage-independent growth and EGF-induced cell transformation.

Kinase profiling identifies PRIMA-1 Met as a potential MEK inhibitor
To understand the mode of action of anti-tumor activity of PRIMA-1 Met , we performed in vitro kinase profiling assay using Millipore's Kinase Profiler ( Table 1). The results showed that MEK1 or MEK2 kinase activity was significantly suppressed by PRIMA-1 Met at the concentration of 25 uM. At the same concentration, PRIMA-1 Met had little effect on EGFR, MAPK1/2 and RSK1/2, which were key kinases in Ras/Raf/MEK/ERK pathway. These data suggest that MEK is a strong candidate target of PRIMA-1 Met compared to other screened kinases.

PRIMA-1 Met inhibits MEK kinase activity in vitro and in cells
To confirm that PRIMA-1 Met could inhibit MEK activity, we conducted in vitro kinase assay by using inactive ERK1 as a substrate of MEK and MEK inhibitor AZD6244 as a positive control. We found that the phosphorylation of ERK was potently decreased by PRIMA-1 Met in a dose dependent manner ( Figure 3A).
Next, we detected MEK kinase activity in colorectal cancer cells and found that its activity was significantly elevated in six colorectal cancer cells compared with normal colon cells ( Figure 3B). When colon cancer cell lines HCT116 wt and HCT116 neg were treated with PRIMA-1 Met , The phosphorylation of downstream molecules of MEK such as ERK1/2 and RSK2 was dramatically attenuated with PRIMA-1 Met concentration more than 25 uM ( Figure 3C, 3D). However, there were no significant differences in MEK inhibition between HCT116 wt and HCT116 neg cells, indicating that p53 status does not interfere with PRIMA-1 Met induced MEK inactivity.
To further exclude p53-dependent regulatory mechanisms, we depleted the expression of mutant p53 in SW480 cells by shRNA and found that depletion of p53 showed minimal effect on the attenuation of MEK activity after PRIMA-1 Met treatment ( Figure 3E). Collectively, these data confirm that PRIMA-1 Met inhibits MEK activity independent of p53 status.

PRIMA-1 Met binds MEK directly in vitro and ex vivo
To better understand PRIMA-1 Met -induced inhibition of MEK activity, we wondered whether PRIMA-1 Met could interact with MEK. We docked PRIMA-1 Met to MEK monomer using automated docking tool AutoDock 4 [22]. The results showed that PRIMA-1 Met bound to the ATP-binding pocket of MEK1 monomer ( Figure 4A). The oxygen atom of PRIMA-1 Met could form hydrogen bond Kinase screening assay was performed according to Millipore's Kinase Profiler protocols. Briefly, the reaction concentration of PRIMA-1 Met was set at 100 µM and the concentration of ATP at 10 µM. Scores for each kinase were represented as the percent (%) of kinase activity after PRIMA-1 Met treatment.
with the sulfur group of Cys207 and Met143, and the carboxylate group of Asp208 ( Figure 4B). Next we performed ATP competitive binding assay by using PRIMA-1 Met -conjugated Sepharose 4B beads in vitro. Compared with Sepharose 4B beads alone, PRIMA-1 Met -conjugated Sepharose 4B beads could pull-down MEK and their affinity was impaired by the presence of ATP ( Figure 4C). Furthermore, immunoprecipitation assay in HCT116 neg cells confirmed the binding of PRIMA-1 Met to endogenous MEK ( Figure 4D). Collectively, these data provide strong evidence that PRIMA-1 Met could directly bind MEK in ATP-binding pocket in vitro and in cells. Inactive ERK1 protein was used as a substrate and mixed with active MEK kinase and different doses of PRIMA-1 Met . The phosphorylation level of ERK1 (Thr202/Tyr204) was detected by Western blot analysis. Total MEK was used as loading control and MEK inhibitor AZD6244 was used as a positive control. B. Western blot analysis of MEK protein expression level in six colorectal cancer cell lines and normal colon epithelial cell line FHC. The activation of MEK signaling was detected in HCT116 neg C., HCT116 wt D. and SW480 E. cells. After starvation in serum-free medium for 24 h, cells were treated with different doses of PRIMA-1 Met for 12 h and then treated with 5 ng/ml EGF for 30 min. AZD6244 was used as a positive control. Data were representative results from three independent experiments. www.impactjournals.com/oncotarget

MEK is indispensable for the anti-tumor activity of PRIMA-1 Met in vitro
To confirm that PRIMA-1 Met attenuated cancer cell proliferation and anchorage-independent growth via the inhibition of MEK kinase activity, we used shRNA to knockdown MEK1 expression in HCT116 neg or HCT116 wt cells ( Figure 5A). The depletion of MEK abrogated PRIMA-1 Met -induced suppression of anchorageindependent growth of HCT116 neg and HCT116 wt cells, and the magnitude of growth reduction was correlated with the efficiency of MEK1 knockdown ( Figure 5B).
PRIMA-1 Met could induce reactive oxygen species (ROS) to mediate p53-independent cell apoptosis [16]. Thus it is possible that PRIMA-1 Met inhibits CRC cell growth by promoting ROS-induced cell apoptosis. To test this possibility, we measured endogenous ROS in HCT116 neg and HCT116 wt cells treated with 25 uM PRIMA-1 Met . PRIMA-1 Met induced massive production of ROS and this was significantly inhibited by ROS scavenger NAC ( Figure 5C). However, NAC had little effect on PRIMA-1 Met induced attenuation of anchorage-independent growth of HCT116 neg and HCT116 wt cells ( Figure 5D). These results indicate that MEK but not ROS predominantly contributes to p53-independent anti-tumor activity of PRIMA-1 Met .

PRIMA-1 Met suppresses CRC growth in xenograft mouse model by inhibiting MEK activity
To confirm that the anti-tumor activity of PRIMA-1 Met is mediated by MEK in vivo, we established subcutaneous xenograft nude mice model implanted with HCT116 neg or HCT116 wt cells. Once the tumor xenografts grew to approximately 50 mm 3 , the mice were randomly assigned to three groups: vehicle (treated with PBS); lower dose of PRIMA-1 Met (20 mg/kg); and higher dose of PRIMA-1 Met (100 mg/kg). PRIMA-1 Met markedly reduced tumor size in both HCT116 neg and HCT116 wt derived xenografts in a dose dependent manner ( Figure 6A). Compared to vehicle group, PRIMA-1 Met (20 mg/kg or 100 mg/kg) significantly inhibited average tumor volume (P < 0.05). In addition, body weight was not significantly affected by PRIMA-1 Met (Figure 6B), indicating that treatment dose of PRIMA-1 Met was well tolerated by the mice.
Immunohistochemical analysis of tumor samples showed that the phosphorylation of ERK1/2 was substantially reduced in treatment groups compared to vehicle group. Additionally, PRIMA-1 Met inhibited HCT116 neg and HCT116 wt cell proliferation because   fewer cells were positively stained with proliferation marker Ki-67 compared to vehicle group ( Figure 6C). Western blot analysis of tumor samples showed that the phosphorylation of ERK1/2 was substantially suppressed in treatment groups compared to vehicle group in a dose dependent manner ( Figure 6D). Taken together, our data indicate that PRIMA-1 Met inhibits CRC growth in vivo independent of p53 status by suppressing MEK activity.

DISCUSSION
Tumor suppressor p53 is encoded by TP53 gene which is frequently mutated in half of all human cancers, including breast cancer, multiple myeloma, lung cancer and colorectal cancer [23]. The most mutations in TP53 such as Y220C or R175H are clustered within the DNAbinding domain and thus impair transcription activity of p53, resulting in the inactivation of p53 targets such as p21, Bax, PUMA and Noxa [24]. As a consequence, loss of p53 tumor suppressor activity prevents cancer cells from p53 mediated cell senescence, cell cycle arrest and apoptosis. Moreover, mutant p53 acquires gain-of-function to promote malignant phenotypes of cancer [25]. Thus it is urgent to develop novel anti-cancer strategies targeting mutant p53.
PRIMA-1 and its methylated form PRIMA-1 Met were originally isolated from the low-molecular-weight compounds library for the ability to restore sequencespecific DNA binding and transcription-dependent apoptotic function to mutant p53 in vitro and in vivo [10]. Although PRIMA-1 Met has demonstrated preferential inhibitory effect on cancer cells harboring mutant p53, recent studies suggest p53-independent activity of PRIMA-1 Met . Saha et al. found that PRIMA-1 Met induced apoptosis via p53-independent Noxa induction [26]. Grinter et al. found that PRIMA-1 Met exerted anti-cancer effect by the inhibition of oxidosqualene cyclase (OSC), a key enzyme involved in cholesterol synthetic pathway [15]. Because p53 family members p63 and p73 have similar sequence homology to p53 in DNA binding domains, PRIMA-1 Met was found to restore mutant p63 and p73 with apoptotic function irrespective of p53 status [27]. Meanwhile, PRIMA-1 Met could induce ROS generation and cause p53-independent massive cell death by converting TrxR1 to NADPH oxidase or impairing GSH metabolism [16,28]. These observations prompted us to evaluate p53 independent ant-tumor activities of PRIMA-1 Met .
In the present study, we first used PRIMA-1 Met to treat six colorectal cancer cell lines with different p53 status. We found that PRIMA-1 Met inhibited the proliferation and growthof all cell lines in a dose and time dependent manner, and high dose of PRIMA-1 Met triggered cell apoptosis regardless of p53 status. In addition, we found that treatment with 25 or 50 uM PRIMA-1 Met caused G2/M arrest in HCT116 neg cells (Supplementary Figure S4), consistent with previous study [29]. Moreover, SW480 and DLD-1 cell lines with depleted mutant p53 were as sensitive as their parental cells to PRIMA-1 Met treatment, demonstrating that PRIMA-1 Met could suppress CRC cell growth in a p53-independent manner.
Kinase signaling pathways, such as Ras/Raf/ MEK and PI3-K/Akt/mTOR cascades, are generally hyperactivated and contribute to cancer cell proliferation, migration, invasion and chemoresistance [30]. Therefore, we performed kinase profiling assay by screening total 40 kinases and identified MEK kinase as a potential target that mediates p53-independent anti-tumor activity of PRIMA-1 Met . MEK kinase and its downstream effectors are consistently activated owing to aberrant mutant activation of Ras and B-Raf in colorectal cancer [31]. Once activated, MEK1 and MEK2 form heterodimer kinase and consecutively mediate downstream ERK1/ ERK2 phosphorylation. Several selective and allosteric MEK inhibitors that bind non-ATP competitive site, such as Trametinib, Cobimetinib, MEK162, AZD6244 and PD-0325901, have shown promise in preclinical and clinical studies [32][33][34]. In contrast, we speculated that PRIMA-1 Met directly binds MEK in the ATP-binding pocket based on the results of automated docking analysis, pull-down binding and ATP competitive assay. Thus the mode of action of PRIMA-1 Met in inhibiting MEK activity seems to be different from that of MEK inhibitors binding non-ATP competitive site. Further studies are needed to compare the efficacy of PRIMA-1 Met and other MEK inhibitors.
By detecting the phosphorylation of MEK substrate ERK, we showed that PRIMA-1 Met potently inhibited MEK kinase activity in CRC cells in a dose dependent manner. More importantly, knockdown of mutant p53 in SW480 did not affect the inhibition of MEK activity by PRIMA-1 Met , confirming that PRIMA-1 Met could inhibit MEK kinase activity in cells with different p53 status. Moreover, in vivo study showed that PRIMA-1 Met inhibited MEK activity, leading to significant tumor regression in HCT116 neg and HCT116 wt xenograft model compared with vehicle group. These in vitro and in vivo data indicate that PRIMA-1 Met is a novel MEK1 inhibitor.
In this study we depleted endogenous MEK1 in CRC cells to investigate whether MEK1 is indispensable for anti-tumor activity of PRIMA-1 Met since many small molecules that targeting the main target have off targets [35,36]. As expected, the depletion of MEK1 abrogated PRIMA-1 Met induced suppression of CRC cell growth. To avoid the confounding potential of ROS, which could be induced by PRIMA-1 Met and mediate p53-independent cancer cell apoptosis [16], we further assessed ROS level and found that ROS production was induced after PRIMA-1 Met treatment. However, ROS scavenger NAC only slightly compensated for PRIMA-1 Met induced CRC cell growth suppression. Collectively, these results demonstrate that direct binding to MEK and inhibiting its activity is the main, if not exclusive, mechanism by which PRIMA-1 Met suppresses CRC independent of p53 status.
In conclusion, we provide the first lines of evidence that PRIMA-1 Met inhibits colorectal cancer cell proliferation and tumor growth predominately by directly targeting MEK in p53 independent manner. Consequently, targeting both MEK and mutant p53 by PRIMA-1 Met may allow us to initiate new strategy for the treatment of malignant colorectal cancer that harbors both hyperactivated Ras/Raf/MEK signaling and p53 mutation.

MTS assay
The cells were seeded (1×10 3 /well) in 96-well plates and cultured overnight. Cells were treated with different concentrations of PRIMA-1 Met and cell proliferation inhibition was measured by MTS assay (Promega, WI, USA) according to the manufacturer's instructions.

Anchorage-independent cell growth assay
A total of 8,000 cells were suspended in 1 ml solidified Basal Medium Eagle containing 10% FBS and added to 0.33% top agarose with various concentrations of PRIMA-1 Met . A base layer of 0.6% agarose was placed at the bottom of top agar. Three weeks after seeding, colonies were counted in low-power field of light microscope using the Image-Pro Plus Software (vs.4) program (Media Cybernetics).

Kinase profiling
Kinase profile assay was performed according to Millipore's Kinase Profiler protocols. The concentration of PRIMA-1 Met and ATP for each kinase analysis was set at 100 uM and 10 uM, respectively.

Molecular modeling
Briefly, the PDB files of compound PRIMA-1 Met was generated from the PRODRG server [37]. The X-ray crystallographic structures of MEK-1/2 (PDB code: 1S9I [38]) were downloaded from Protein Data Bank and prepared by using the Protein Preparation Wizard in Maestro v9.2. Hydrogen atoms were added to the complexes consistently with a pH of 7 and water molecules were deleted. Ligand docking the compound to MEK-1/2 monomer was accomplished using the automated docking tool AutoDock 4. Ten independent runs were carried out to yield ten AutoDock docking solutions, which were ranked by the calculated energy favorable scores. Top hits based on hydrogen-bond interactions were considered as topranked conformations of docked complex.

In vitro kinase assay
Active MEK1 (100 ng) and its inactive substrate ERK1 (1 ug) were purchased from Millipore (Temecula, CA, USA), mixed in the presence of various doses of PRIMA-1 Met , and incubated in 100 uM ATP and ×1 kinase buffer (10 mM MgCl 2 , 1 mM EDTA, 50 mM Tris-HCl pH 7.5, 0.01% Brij35, 1 mM DTT; Cell Signaling Technology) at 30°C for 30 min. Then the samples were resolved in ×6 SDS loading buffer. Phosphorylated ERK1 and total MEK protein level were analyzed by Western blot analysis.

Measurement of ROS
HCT116 neg and HCT116 wt cells were cultured in sixwell plate and grown to subconfluency. Cells were treated with PRIMA-1 Met in the presence or absence of NAC for indicated time. Then cells were stained with 20 uM DCFHDA for 30 min and fluorescence was analyzed using flow cytometer in the FITC channel. The fluorescence intensity of 15,000 cells was measured by WinMDI 2.8 software and normalized to vehicle group.

Xenograft mouse model
Athymic nude mice (6-9 weeks old) were provided by SLAC Laboratory Animal Co. Ltd. (Shanghai, China). Briefly, cells in logarithmic phase were harvested and suspended in cold PBS at 5×10 6 /mL. 1×10 6 cells (in 200 μl of PBS) were injected subcutaneously into right flank of each mouse. Once the tumor volume reached around 50 mm 3 , mice were randomly divided into three groups (n=5 each): (i) vehicle group; (ii) 20 mg/kg PRIMA-1 Met -treated group; (iii) 100 mg/kg PRIMA-1 Met -treated group. PRIMA-1 Met or vehicle (PBS) was intraperitoneally injected daily for 5 days, stopped for 2 days and reinjected for additional 5 days. Tumor volume was measured twice a week based on the following formula: volume = (Length×Width 2 )/2. Mice body weight was measured and recorded once a week. Mice were euthanized on day 28 and tumor tissues were dissected for further analysis.

Immunohistochemistry analysis
The tumor samples were fixed in 10% neutral formalin solution and embedded in paraffin. Then sections were cut into fragments of 6 um, deparaffinized and hydrated. Tissues on slides were permeabilized by 0.5% Triton X-100 and incubated overnight at 4ºC with primary antibodies against Ki67 and phosphorylated ERK1/2. The slides were washed by PBS for 5 times and incubated with 1:200 dilution of secondary biotinylated antibody for 2 h at room temperature. All slides were developed using horseradish peroxidase-streptavidin (Santa Cruz Biotechnology, CA, USA). Images were taken under microscope and signal densities were analyzed using the Image-Pro Plus Software (Media Cybernetics).

Statistical analysis
All data were presented as mean values ± SD or SE as indicated. The Student's t test (two groups only) or a one-way ANOVA (more than two groups) was used for the comparison. p < 0.05 was accepted as statistical significance.