Anticancer Activity of Pyrimethamine via Ubiquitin Mediated Degradation of AIMP2-DX2

While aminoacyl-tRNA synthetase-interacting multifunctional protein 2 (AIMP2) is a tumor suppressor, its exon 2-depleted splice variant (AIMP2-DX2 or shortly DX2) is highly expressed in human lung cancer, and the ratio of DX2 to AIMP2 increases according to the progression of lung cancer. In this study, pyrimethamine inhibited the level of DX2 (IC50 = 0.73 µM) in A549 cells expressing nanoluciferase-tagged DX2. In a panel of 5 lung cancer cell lines with various DX2 levels, pyrimethamine most potently suppressed the growth of H460 cells, which express high levels of DX2 (GI50 = 0.01 µM). An immunoblot assay in H460 cells showed that pyrimethamine decreased the DX2 level dose-dependently but did not affect the AIMP2 level. Further experiments confirmed that pyrimethamine resulted in ubiquitination-mediated DX2 degradation. In an in vivo mouse xenograft assay using H460 cells, intraperitoneal administration of pyrimethamine significantly reduced the tumor size and weight, comparable with the effects of taxol, without affecting body weight. Analysis of tumor tissue showed a considerably high concentration of pyrimethamine with a decreased levels of DX2. These results suggest that pyrimethamine, currently used as anti-parasite drug, could be repurposed to treat lung cancer patients expressing high level of DX2.


Introduction
Aminoacyl-tRNA synthetases (ARSs) are the enzymes that catalyze the ligation of specific amino acids to their cognate tRNAs in the first step of protein translation [1]. In higher eukaryotic systems, nine ARSs and three aminoacyl-tRNA synthetase-interacting multifunctional proteins (AIMPs) are assembled into a multi-tRNA synthetase complex (MSC) that serves as a hub for various regulatory pathways in addition to playing a role in protein synthesis. While three AIMPs, AIMP1/p43, AIMP2/p38 and AIMP3/p18, function as scaffold proteins for the assembly and integrity of MSCs [2][3][4], they also participate in diverse regulatory roles that are not directly linked to protein synthesis.
Among the AIMPs, AIMP2 works as a potent tumor suppressor and induces growth arrest by transforming growth factor-β (TGF β) signaling via ubiquitin-mediated degradation of FUSE-binding protein (FBP) and downregulation of c-myc [5]. In addition, AIMP2 suppresses tumor necrosis factor-α

Identification of Pyrimethamine as a Suppressor of DX2 Level
Previously, we established a firefly luciferase assay system with A549 cells expressing luciferase-tagged DX2 to find small molecules that could reduce the cellular level of DX2 [11]. For establishment of more sensitive high-throughput screening, we changed the luciferase tag to a nanoluciferase tag because the nanoluciferase is three times smaller and 2,500,000 times brighter than the previously used luciferase, and 763 compounds were screened in the drug repositioning library from KCB at KRICT by our screening system. We found 30 compounds that reduced the cellular level of DX2 at 5 µM by more than 50% compared with the level in control cells ( Figure 1). Pyrimethamine ( Figure 2) showed a good suppressive effect on the level of DX2 (IC 50 = 0.73 µM) but no inhibitory effect on the level of AIMP2 (IC 50 > 100 µM), indicating that it selectively decreased the expression level of DX2.
In this paper, we describe the effects of pyrimethamine in a nanoluciferase assay, a viability assay in 5 lung cancer cell lines with different expression levels of DX2, a mechanism of action study, and an in vivo xenograft assay. Figure 1. Screening to identify DX2 inhibitors. 763 chemicals (5 µM) from KCB at KRICT was used for first screening. Thirty chemicals showing greater than 50% inhibition were subjected to secondary screening via nanoluciferase and cell viability assays using A549 and WI-26 cells.

Identification of Pyrimethamine as a Suppressor of DX2 level
Previously, we established a firefly luciferase assay system with A549 cells expressing luciferasetagged DX2 to find small molecules that could reduce the cellular level of DX2 [11]. For establishment of more sensitive high-throughput screening, we changed the luciferase tag to a nanoluciferase tag because the nanoluciferase is three times smaller and 2,500,000 times brighter than the previously used luciferase, and 763 compounds were screened in the drug repositioning library from KCB at KRICT by our screening system. We found 30 compounds that reduced the cellular level of DX2 at 5 µM by more than 50% compared with the level in control cells ( Figure 1). Pyrimethamine ( Figure 2)

Antiproliferative Effect of Pyrimethamine on Cancer Cells but not Normal Cells
A549 and WI-26 cells [9,24] were subjected to a cell viability assay to determine the GI50 values for 50% inhibition of cell growth. Pyrimethamine displayed good inhibition of cell growth in A549 cells (GI50 = 0.8 µM) similar to its IC50 (0.73 µM), which show a moderate level of DX2, but no inhibitory effect in WI-26 cells (GI50 > 100 µM), which have a low level of DX2 expression (Table 1). Additionally, cell viability experiments were performed using 4 other lung cancer cell lines (H460, HCC-1359, HCC-366, and H2087) with various levels of DX2 [9,24]. Pyrimethamine showed potent anti-proliferative activity on H460 cells with high levels of DX2 (GI50 = 0.01 µM), implying that the anti-proliferative potency of pyrimethamine is well correlated with the DX2 level, while taxol showed similar potencies on all 6 tested cell lines as (GI50 = 0.02~0.08 µM) shown in Table 1.

DX2 Level-Dependent Antiproliferative Effect of Pyrimethamine
Based on the results above, the cell lines showing high level of DX2 appear to be highly susceptible for pyrimethamine-dependent cell death. To confirm whether anti-proliferative activity of pyrimethamine is mediated by DX2, we set up the cell proliferation assay using DX2-inducible system [24]. For negative control against the expressed plasmid, empty vector (EV) was used. DX2or EV-inducible stable A549 cells were treated with doxycycline (Dox) to induce the expression of DX2 and the cells were subsequently treated with pyrimethamine. As expected, the DX2-induced A549 cell lines showed the increased cell viability. After treatment with pyrimethamine, we observed that pyrimethamine significantly inhibited the growth of DX2-inducible A549 cells compared to EVinducible A549 cells (Figure 3). Pyrimethamine was reported to show the anti-tumor effect against cancer cells [15][16][17][18][19][20][21][22][23], but its mechanism was still veiled [19]. Our results showed that the increased level of DX2 leads the strong anti-tumor effect, suggesting that DX2 could be one of the targets for

Antiproliferative Effect of Pyrimethamine on Cancer Cells but not Normal Cells
A549 and WI-26 cells [9,24] were subjected to a cell viability assay to determine the GI 50 values for 50% inhibition of cell growth. Pyrimethamine displayed good inhibition of cell growth in A549 cells (GI 50 = 0.8 µM) similar to its IC 50 (0.73 µM), which show a moderate level of DX2, but no inhibitory effect in WI-26 cells (GI 50 > 100 µM), which have a low level of DX2 expression (Table 1). Additionally, cell viability experiments were performed using 4 other lung cancer cell lines (H460, HCC-1359, HCC-366, and H2087) with various levels of DX2 [9,24]. Pyrimethamine showed potent anti-proliferative activity on H460 cells with high levels of DX2 (GI 50 = 0.01 µM), implying that the anti-proliferative potency of pyrimethamine is well correlated with the DX2 level, while taxol showed similar potencies on all 6 tested cell lines as (GI 50 = 0.02~0.08 µM) shown in Table 1. or EV-inducible stable A549 cells were treated with doxycycline (Dox) to induce the expression of DX2 and the cells were subsequently treated with pyrimethamine. As expected, the DX2-induced A549 cell lines showed the increased cell viability. After treatment with pyrimethamine, we observed that pyrimethamine significantly inhibited the growth of DX2-inducible A549 cells compared to EV-inducible A549 cells ( Figure 3). Pyrimethamine was reported to show the anti-tumor effect against cancer cells [15][16][17][18][19][20][21][22][23], but its mechanism was still veiled [19]. Our results showed that the increased level of DX2 leads the strong anti-tumor effect, suggesting that DX2 could be one of the targets for pyrimethamine-mediated anti-tumoral effect. This results indicated that anti-proliferative activity of pyrimethamine would depend on the expression level of DX2.

Specific Suppression of DX2 Level by Pyrimethamine
To examine the concentration-dependent inhibitory effect of pyrimethamine on the protein and mRNA levels of endogenous DX2 and AIMP2, H460 cells were treated with 0, 0.6, 1.2, 2.5 and 5.0 µM pyrimethamine for 12 h. The amounts of DX2 and AIMP2 protein were detected by immunoblotting, and the effect of pyrimethamine on the mRNA transcript levels was monitored via our previously established RT-PCR [9,11]. Pyrimethamine reduced the protein level of DX2, but not AIMP2, ( Figure  4a, upper), which is consistent with the result of the nanoluciferase assay described above. However, pyrimethamine did not affect the mRNA transcript level of either DX2 or AIMP2 (Figure 4a, bottom), while our previous hit, BC-DXI01, inhibited DX2 level through decreasing the mRNA transcript level. These results suggest that pyrimethamine selectively decreased the protein level of DX2 with no effect on the mRNA transcript levels. Additionally, we performed an ubiquitination assay [9,11] to confirm pyrimethamine-mediated ubiquitination of DX2 because our previous study reported that DX2 protein was degraded by ubiquitination [24]. As shown in Figure 4b, DX2 was degraded by ubiquitination upon pyrimethamine treatment in a dose-dependent manner. Thus, pyrimethamine appears to induce ubiquitination of DX2 for its degradation.

Specific Suppression of DX2 Level by Pyrimethamine
To examine the concentration-dependent inhibitory effect of pyrimethamine on the protein and mRNA levels of endogenous DX2 and AIMP2, H460 cells were treated with 0, 0.6, 1.2, 2.5 and 5.0 µM pyrimethamine for 12 h. The amounts of DX2 and AIMP2 protein were detected by immunoblotting, and the effect of pyrimethamine on the mRNA transcript levels was monitored via our previously established RT-PCR [9,11]. Pyrimethamine reduced the protein level of DX2, but not AIMP2, (Figure 4a, upper), which is consistent with the result of the nanoluciferase assay described above. However, pyrimethamine did not affect the mRNA transcript level of either DX2 or AIMP2 (Figure 4a, bottom), while our previous hit, BC-DXI01, inhibited DX2 level through decreasing the mRNA transcript level. These results suggest that pyrimethamine selectively decreased the protein level of DX2 with no effect on the mRNA transcript levels. Additionally, we performed an ubiquitination assay [9,11] to confirm pyrimethamine-mediated ubiquitination of DX2 because our previous study reported that DX2 protein was degraded by ubiquitination [24]. As shown in Figure 4b, DX2 was degraded by ubiquitination upon pyrimethamine treatment in a dose-dependent manner. Thus, pyrimethamine appears to induce ubiquitination of DX2 for its degradation.

Inhibitory Effect of Pyrimethamine on Tumorigenesis in vivo
We next examined the in vivo antitumor activities of pyrimethamine in a BALB/cSLC-nu/nu mouse xenograft model using H460 cells, which exhibited the most potent growth inhibition among the tested lung cancer cells. When the embedded tumor mass reached a volume of approximately 100 mm 3 , pyrimethamine (20 mg/kg/day, at 1-5 and 8-12 day) and taxol (15 mg/kg/day, at 1, 3, 5, and 8, 10, 12 day) were intraperitoneally injected five and three times a week, respectively, for two weeks. We decided the dosage of pyrimethamine as 20 mg/kg/day, because BC-DXI-01 with a half efficacy of pyrimethamine on cell showed in vivo efficacy at 50 mg/kg/day [11]. Taxol was treated as the published method [25]. Tumor growth in the pyrimethamine-and taxol-treated groups was apparently inhibited compared to that in the control group (Figure 5a). Tumor volumes and weights were reduced by up to 40% following pyrimethamine treatment (Figure 5b and 5c), while pyrimethamine did not affect the body weights of treated mice (Figure 5d). The concentration of the injected pyrimethamine in tumor tissue was 142 ng/g, high enough to decrease the endogenous DX2 level (Figure 5e). The effects of pyrimethamine on endogenous DX2 or AIMP2 were also analyzed by immunoblotting, and the specific decrease of DX2, not AIMP2, protein level was frequently observed in the pyrimethamine-treated tumor tissues (Figure 5f and 5g).

Inhibitory Effect of Pyrimethamine on Tumorigenesis In Vivo
We next examined the in vivo antitumor activities of pyrimethamine in a BALB/cSLC-nu/nu mouse xenograft model using H460 cells, which exhibited the most potent growth inhibition among the tested lung cancer cells. When the embedded tumor mass reached a volume of approximately 100 mm 3 , pyrimethamine (20 mg/kg/day, at 1-5 and 8-12 day) and taxol (15 mg/kg/day, at 1, 3, 5, and 8, 10, 12 day) were intraperitoneally injected five and three times a week, respectively, for two weeks. We decided the dosage of pyrimethamine as 20 mg/kg/day, because BC-DXI-01 with a half efficacy of pyrimethamine on cell showed in vivo efficacy at 50 mg/kg/day [11]. Taxol was treated as the published method [25]. Tumor growth in the pyrimethamine-and taxol-treated groups was apparently inhibited compared to that in the control group (Figure 5a). Tumor volumes and weights were reduced by up to 40% following pyrimethamine treatment (Figure 5b,c), while pyrimethamine did not affect the body weights of treated mice (Figure 5d). The concentration of the injected pyrimethamine in tumor tissue was 142 ng/g, high enough to decrease the endogenous DX2 level (Figure 5e). The effects of pyrimethamine on endogenous DX2 or AIMP2 were also analyzed by immunoblotting, and the specific decrease of DX2, not AIMP2, protein level was frequently observed in the pyrimethamine-treated tumor tissues (Figure 5f,g).

Discussion
Through the screening of a drug repositioning library from KCB, pyrimethamine was identified as a hit chemical for the treatment of lung cancer, selectively inhibiting DX2 over AIMP2 (IC50 > 100 µM) with an IC50 value of 0.73 µM in A549 cells expressing nanoluciferase-tagged DX2. It also decreased the cell viability of DX2-inducible A549 cells but showed no effect on empty vectorinducible A549 cells, indicating that the DX2 level is important for pyrimethamine-mediated suppression of A549 cell viability. Additionally, cell viability assays on 5 lung cancer cell lines with various expression levels of DX2 were performed. Pyrimethamine suppressed cell growth most potently in H460 cells, which express a high level of DX2 (GI50 = 0.01 µM), followed by A549 (

Discussion
Through the screening of a drug repositioning library from KCB, pyrimethamine was identified as a hit chemical for the treatment of lung cancer, selectively inhibiting DX2 over AIMP2 (IC 50 > 100 µM) with an IC 50 value of 0.73 µM in A549 cells expressing nanoluciferase-tagged DX2. It also decreased the cell viability of DX2-inducible A549 cells but showed no effect on empty vector-inducible A549 cells, indicating that the DX2 level is important for pyrimethamine-mediated suppression of A549 cell viability. Additionally, cell viability assays on 5 lung cancer cell lines with various expression levels of DX2 were performed. Pyrimethamine suppressed cell growth most potently in H460 cells, which express a high level of DX2 (GI 50 = 0.01 µM), followed by A549 (GI 50  with pyrimethamine showed decreased level of DX2 in dose dependent manner. While pyrimethamine decreased the protein level of DX2, it did not decrease the DX2 mRNA transcript level, differently from our previous hit, BC-DXI-01 to decrease the DX2 mRNA [11]. Pyrimethamine also showed increased potency than BC-DXI-01 with better pharmacokinetic properties. Pyrimethamine was confirmed to induce ubiquitin-mediated DX2 degradation. Finally, we evaluated in vivo efficacy of pyrimethamine in a mouse xenograft assay using H460 cells. It significantly reduced the tumor size and weight, comparable with the effects of taxol, without affecting body weight. Analysis of tumor tissue showed a considerably high concentration of pyrimethamine with a decreased level of DX2. In this experiments, pyrimethamine showed potent antitumorigenic efficacy against H460 lung cancer cells through decreasing the DX2 protein level. Currently, pyrimethamine is used for the treatment of parasite diseases including toxoplasmosis, actinomycosis, and isosporiasis, and for the treatment and prevention of Pneumocystis jirovecii pneumonia through the competitive inhibition of DHFR, thereby interfering with the regeneration of tetrahydrofolic acid from dihydrofolate [26]. Tetrahydrofolic acid is essential for DNA and RNA synthesis in many species, including protozoa. Antitumor effects of pyrimethamine through targeting DHFR were reported [16,20]. STAT3 has been intensively studied as a therapeutic target for cancer of pyrimethamine including lung cancer [21,23]. There is a report that DHFR reduced STAT3 activity [27], while the exact mechanism of action on antitumor effects of pyrimethamine is not still clear. In this study, we showed that pyrimethamine induced the ubiquitination of DX2, which might provide one of the plausible mechanism on the antitumor efficacy of pyrimethamine against cancer cells with high level of DX2. Pyrimethamine can cause gastrointestinal symptoms such as nausea, vomiting, anorexia, and diarrhea when higher doses are used (www.drugs.com). Because pyrimethamine inhibited lung cancer cells potently, it may be repositioned to treat lung cancer patients with an enough safety margin. Also, it needs further safety study to be used in cancer because the duration of treatment on cancer would be longer compared to infectious diseases. We are continuously trying to validate the efficacy and safety of pyrimethamine as an anti-cancer agent, and elucidate its mode of action.

Library Screening Based on a Nanoluciferase Assay
A549 cells (4 × 10 4 cells/mL) expressing nanoluciferase-tagged DX2 via transient transfection were seeded and cultured in 96-well white flat bottom plates (#3903 Corning) for 12 h. Cells were incubated in serum-free medium containing 763 compounds (5 µM) in a drug repositioning library from Korea Chemical bank (KCB) for 4 h, and the luminescence signal was measured following the manufacturer's protocol (Promega, Madison, WI, USA). The instrument used for measuring the luminescence is GloMax discover (Promega). Thirty chemicals showing greater than 50% inhibition in the primary screen at 5 µM were subjected to secondary screening via a luciferase assay using nanoluciferase-tagged AIMP2 for negative screening, as described above. The IC 50 values, were calculated using the Prism (GraphPad, San Diego, CA, USA).

Assessment of Cell Viability Using DX2-and EV (Empty Vector)-Inducible A549
To assess the DX2-dependent effect of pyrimethamine on cell viability, DX2 was cloned into the EcoRI/NotI sites of pTetOne vector (Promega). pTetOne vector cloned with DX2 or vector only for expressing empty vector was transfected into A549 cells using transfection reagent, lipofectamine2000 (Invitrogen, Carlsbad, CA, USA) following the instruction. A549 cells were subjected to selection by treatment with puromycin (Sigma) every two days for two weeks. EV-and DX2-inducible A549 cells (1 × 10 4 cells/mL) were cultured in 96-well plates for 24 h, and doxycycline (1 µg/mL) was then added for induction. After induction for 6 days, the isogenic cell lines were incubated in serum-free medium with pyrimethamine (1 µM). Cell viability was assessed with a cell viability assay.

Immunoblotting and RT-PC
Cells were lysed with 50 mM Tris-HCl (pH 7.4) lysis buffer containing 100 mM NaCl, 0.5% Triton X-100, 0.1% SDS, 10% glycerol, 1 mM EDTA, and a protease inhibitor (Calbiochem, San Diego, CA, USA), for 45 min. Supernatants from cell lysates were subjected to SDS-PAGE for protein separation. Proteins of interest were detected by immunoblotting using a corresponding specific antibodies. Total RNA was extracted from H460 cells with an RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and reverse transcriptase polymerase chain reaction (RT-PCR) was conducted with 1 µg of the extracted RNA, dNTPs, random hexamers and Moloney murine leukemia virus (MMLV) transcriptase. To measure the mRNA expression of DX2, AIMP2 and Actin, 2 µL of cDNA was used for PCR with the corresponding specific primers. The specific primer sequences were as follows. Simultaneous detection of both DX2 and AIMP2: ATGCCGATGTACCAGGTAAAG and CTTAAGGAGCTTGAGGGCCGT; Actin: CCTTCCTGGGCATGGAGTCCT and GGAGCAATGATCTTGATCTT. Actin was used as a loading control.

Ubiquitination Assay
The detailed procedures were following our previous report [12]. 293T cells expressing HA-ubiquitin and Strep-DX2 were treated with various doses of pyrimethamine and MG-132 (50 µM) for 12 h. Cells were lysed with 50 mM Tris-HCl (pH 7.4) lysis buffer containing 100 mM NaCl, 0.5% Triton X-100, 0.1% SDS, 10% glycerol, 1 mM EDTA, and a protease inhibitor (Calbiochem). Total cell lysates were precipitated with a Strep-Tactin column, and the eluates were separated by SDS-PAGE. The level of ubiquitinated DX2 was determined by immunoblotting using a specific antibody against HA.

Animals
Seven-week-old female BALB/cSLC-nu/nu mice (Central Lab. Animal Inc., Seoul, Korea), were used in this study. Mice were bred in specific pathogen free (SPF) condition. Animal experiments were performed in compliance with the guidelines of the University Animal Care and Use Committee at Seoul National University (SNU-150901-2, SNU-181205-9).

In Vivo Mouse Xenograft
H460 cells (1 × 10 7 ) were subcutaneously injected into the backs of 7-week-old female BALB/cSLC-nu/nu mice (Central Lab. Animal Inc., Korea) (n = 5 mice/group). Pyrimethamine (20 mg/kg) and Taxol (15 mg/kg) were intraperitoneally administered five and three times a week, respectively, for 2 weeks. Pyrimethamine and Taxol were treated at 0, 24,48,72,96,168,192,216,240, and 264 h (10 times) and at 0, 48, 96, 168, 216, and 264 h (6 times), respectively. The used carrier for control was DTPP (DMSO: Tween 80: PEG400: PBS = 20:1:30:49). The embedded tumor sizes and body weight were measured four times during the experimental period. After 13 days, all mice were sacrificed, and the mice with embedded tumors, along with the excised tumors, were photographed. After the harvested tumors were weighed, the endogenous protein levels of DX2 and AIMP2 were determined by immunoblotting. The tumors were lysed in PBS containing 1% Triton X-100, 0.1% SDS and a protease inhibitor (Calbiochem), and supernatants from the extracts were subjected to SDS-PAGE and immunoblotting using specific antibodies against DX2 and AIMP2. To measure the concentration of pyrimethamine in the excised tumors, the harvested tumors were homogenized in three volumes of PBS buffer and mixed with nine volumes of cold acetonitrile containing the internal standard (disopyramide). After centrifugation at 15,000 rpm and 4 • C for 10 min, 100 µL of the supernatants was subjected to LC-MS/MS analysis [28]. Mass spectrometry (Agilent 6460) with HPLC (Agilent 1261) was used for mass analysis. Standard samples were prepared via the same procedure described above.

Statistics
Statistical tests were performed with Prism (GraphPad). A value of p < 0.05 was considered statistically significant. All error bars means standard deviation (S.D.). Statistical parameters are presented in the figure legends.

Conclusions
In this study, pyrimethamine displayed selective inhibitory activity against level of DX2 over AIMP2 in A549 cells expressing nanoluciferase-tagged DX2, and most potently suppressed the growth of H460 cells, which express high levels of DX2. Pyrimethamine also decreased the DX2 level dose-dependently but did not affect the AIMP2 level in an immunoblot assay using H460 cells. Further experiments confirmed that pyrimethamine resulted in ubiquitination-mediated DX2 degradation. In an in vivo mouse xenograft assay using H460 cells, pyrimethamine significantly reduced the tumor size and weight, comparable with the effects of taxol, without affecting body weight. In conclusion, pyrimethamine showed potent antitumorigenic efficacy against H460 lung cancer cells through decreasing the DX2 protein level.