Decreased APE-1 by Nitroxoline Enhances Therapeutic Effect in a Temozolomide-resistant Glioblastoma: Correlation with Diffusion Weighted Imaging

Glioblastoma multiforme (GBM) is one of the most aggressive human tumors with poor survival rates. The current standard treatment includes chemotherapy with temozolomide (TMZ), but acquisition of resistance is a persistent clinical problem limiting the successful treatment of GBM. The purpose of our study was to investigate therapeutic effects of nitroxoline (NTX) against TMZ-resistant GBM in vitro and in vivo in TMZ-resistant GBM-bearing mouse model, which was correlated with diffusion-weighted imaging (DWI). For in vitro study, we used TMZ-resistant GBM cell lines and evaluated therapeutic effects of NTX by clonogenic and migration assays. Quantitative RT-PCR was used to investigate the expression level of TMZ-resistant genes after NTX treatment. For in vivo study, we performed 9.4 T MR imaging to obtain T2WI for tumor volume measurement and DWI for assessment of apparent diffusion coefficient (ADC) changes by NTX in TMZ-resistant GBM mice (n = 8). Moreover, we performed regression analysis for the relationship between ADC and histological findings, which reflects the changes in cellularity and apurinic/apyrimidinic endonuclease-1 (APE-1) expression. We observed the recovery of TMZ-induced morphological changes, a reduced number of colonies and a decreased rate of migration capacity in TMZ-resistant cells after NTX treatment. The expression of APE-1 was significantly decreased in TMZ-resistant cells after NTX treatment compared with those without treatment. In an in vivo study, NTX reduced tumor growth in TMZ-resistant GBM mice (P = 0.0122). Moreover, ADC was increased in the NTX-treated TMZ-resistant GBM mice compared to the control group (P = 0.0079), which was prior to a tumor volume decrease. The cellularity and APE-1 expression by histology were negatively correlated with the ADC value, which in turn resulted in longer survival in NTX group. The decreased expression of APE-1 by NTX leads to therapeutic effects and is inversely correlated with ADC in TMZ-resistant GBM. Therefore, NTX is suggested as potential therapeutic candidate against TMZ-resistant GBM.

www.nature.com/scientificreports www.nature.com/scientificreports/ to TMZ 5,6 . One study showed that the DNA base excision repair enzyme, APNG confers resistance to TMZ and the downregulation of APNG using siRNA increased the TMZ sensitivity in several established as well as primary GBM cell lines. The authors also showed that the presence of APNG is associated with poor survival in patients and that the loss of APNG improves the survival in xenograft models. Another study reported that the use of a small molecule BER inhibitor potentiates the TMZ tumor cell killing and strongly sensitizes glioma cells to TMZ. Bertucci et al. used the mesoporous silica nanoparticles, decorated with polyarginine-peptide nucleic acid to target miRNA, loaded with TMZ and showed a strong induction of apoptosis in TMZ-resistant glioma cells 7 .
Nitroxoline (NTX), is an FDA-approved antibiotic against urinary tract infections that has been repurposed for cancer treatment and could have anti-cancer effects in patients with various kinds of cancer, including leukemia, lymphoma, bladder cancer and pancreatic cancer 8,9 . Mirkovic et al. reported the NTX as a reversible inhibitor of cathepsin B, a protein responsible for tumor invasion and showed that NTX inhibited the extracellular and intracellular degradation of the extracellular matrix 10 . Furthermore, another study demonstrated that the NTX induces apoptosis in vitro and inhibits the glioma growth in genetically engineered Pten/Kras mice in vivo 11 . Since NTX is already being used in humans for UTIs with tolerable side effects, it has the capability to be used against GBM.
Diffusion-weighted imaging (DWI) is a promising technique to evaluate the response to treatment in tumors in both clinical and research settings. The apparent diffusion coefficient (ADC) measured from DWI is sensitive to the tumor microenvironment 12 and is a potential noninvasive biomarker for the prediction and monitoring of the response to therapy 13 . ADC maps can provide valuable information as they reflect the mobility of water molecules within the tissues and can sensitively detect changed cellularity in tumors 14 . Sugahara et al. reported that the ADC is negatively correlated with the tumor cellularity, with a high cell density and extracellular tortuosity resulting in an increased restriction of the diffusion of water molecules 15 .
Considering all these findings, our purpose was to investigate the therapeutic effects of NTX against TMZ-resistant GBM in vitro and in vivo in a TMZ-resistant GBM bearing mouse model, which was correlated with DWI.

NTX treatment recovers cell morphology change and inhibits colony formation and migration in TMZ-resistant GBM cells. TMZ-resistance induced morphological transformations in both of the resist-
ant cell lines as the cells changed from round or oval shapes to elongated and spindle shapes, as demonstrated by α-tubulin staining under a fluorescence microscope (Fig. 1A). However, the cytoskeleton was disrupted and recovered to the original parental morphology after 10 µg/ml of NTX treatment for 24 hours in both of the TMZresistant cell lines (Fig. 1B). To evaluate the therapeutic effects of NTX in vitro, we performed a colony formation and migration assay. NTX (0.5 µg/ml) significantly reduced the formation of colonies in both of the TMZresistant cell lines after 10 days of treatment. Approximately 60% and 80% of the colonies were inhibited in TMZresistant LN229 (LN229R) (P = 0.0180) and TMZ-resistant U87 (U87R) (P = 0.0172), respectively (Fig. 1C). In a migration assay, we found that 10 µg/ml of NTX inhibited approximately 84% and 80% of the migration over 48 hours in LN229R (P = 0.0180) and over 24 hours in U87R (P = 0.0180), respectively, when compared to their respective nontreated groups (Fig. 1D).  (Fig. 2). In addition, an increased expression of cleaved caspase-3 and PARP after NTX treatment was noted in both TMZ-resistant cell lines ( Supplementary Fig. S2).

Discussion
Our study demonstrates that NTX can act as an anticancer compound against TMZ-resistant GBM through the in vitro and in vivo studies. First, NTX could reduce the expression of APE-1, an essential protein for DNA base excision repair and redox regulation, in TMZ-resistant GBM cells. Second, we observed an increase in the ADC value before a significant tumor volume change soon after NTX treatment in the TMZ-resistant GBM mouse model, which indicates a prior decrease in GBM cellularity before the shrinkage of tumor burden, as well as a long-term survival gain of NTX treatment in the TMZ-resistant GBM mouse model. Third, both APE-1 expression and tumor cellularity were inversely correlated with the ADC value after NTX treatment in the TMZ-resistant GBM mouse model, which suggests that DWI can be used as an imaging surrogate marker of early response to chemotherapy.
We selected several genes related to TMZ-resistance through previous studies 4, [16][17][18][19][20] and investigated regulated the gene levels after NTX treatment in two TMZ-resistant cell lines. Interestingly, the increased APE-1 level of both TMZ-resistant cells compared to the parental cells was significantly decreased in TMZ-resistant cells after NTX treatment. This pattern was consistent with the morphological changes, which showed a recovery of cell morphology to parental cells after NTX treatment in both TMZ-resistant cell lines (Figs 1B and 2). However, other genes including MSH-2, MSH-6, MGMT, and Cx43 did not show similar changes (Supplementary Fig. S1).
APE-1 activity is elevated in human gliomas and confers resistance to ionizing radiation and alkylating agents 21,22 . Demple et al. reported the interfered cell proliferation by the strong downregulation of APE-1 in cancer cell lines, which leads to the activation of apoptosis 23 . Related to these reports, our data showed an increased expression level of cleaved caspase-3 (CC3) and cleaved PARP, which are associated with the concentration of NTX in both TMZ-resistant cell lines (Supplementary Fig. S2). Previous reports have revealed that the cytoskeleton is composed of αand β-tubulin, which form the microtubules that are required for cell differentiation and the maintenance of structural integrity 24 . In addition, the migrative and invasive properties of GBM cells were found to be potentially blocked by a microtubule inhibitor 25 . It is also obvious that the migrative and invasive features of GBM cells contribute to the development of resistance to apoptosis 26,27 . In agreement with previous reports, by the use of immunofluorescence staining, we found that NTX disrupted the microtubule structures of  www.nature.com/scientificreports www.nature.com/scientificreports/ both TMZ-resistant GBM cell lines, leading to impairment of the colony formation capacity and the migrative and invasive properties of cancer cells. Based on these findings, the downregulated APE-1 by NTX was able to increase the therapeutic effect against TMZ resistance, which can be supported by the reduced colony formation and migration capacities after NTX treatment as well (Fig. 1).
To investigate the therapeutic effect of TMZ resistant GBM by NTX in vivo, we developed a TMZ-resistant GBM-bearing mouse model 4,28,29 and assessed cellularity changes using the ADC parameter 30 , respectively. In our study, the TMZ-resistant GBM-bearing mouse model was confirmed by investigating the tumor size and ADC value after TMZ treatment for 1 week. All control, NTX, and TMZ groups showed an increased tumor size which indicates the development of resistance even after TMZ. In our present study, to assess the early therapeutic effect of NTX against TMZ-resistant GBM, we measured changes in tumor volume and ADC values, where we observed a more prominent ADC increase of TMZ-resistant GBM in the NTX group than control group, which was prior to the tumor volume decrease. We also confirmed the long-term survival gain of NTX in the TMZ-resistant GBM model as well. These results suggest that NTX has a potent anti-tumor activity even within the early period, which resulted in the extended survival of the TMZ-resistant GBM model. One might assume that the 1-week TMZ exposure may be insufficient to develop a TMZ-resistant GBM model. To further scrutinize this possibility, we investigated the TMZ group, which was given TMZ for one more week after the confirmation of a TMZ-resistant GBM-bearing mouse model (total time TMZ received: 2 weeks), and we also observed subsequent increased tumor volume and decreased ADC values after persistent TMZ treatment (Supplementary Fig. S3). This finding clearly reflects that GBM had further progressed as resistance developed to TMZ in the TMZ group 31,32 .
The ADC changes after NTX treatment were also correlated with the histological findings including significantly decreased cellularity and cell proliferation in the NTX group of the TMZ-resistant GBM model compared with the control group (Fig. 4), which are, in particular, supported by the decreased APE-1 expression level and the increased TUNEL-positive cells, well correlated with the in vitro results in our study (Figs 2 and 4 and Supplementary Fig. S2). Moreover, numerical data of cellularity change after NTX treatment by H&E which were consistent with a previous report 33 and the APE-1 expression change by immunohistochemistry were negatively correlated with the ADC value in both the control and NTX groups (Fig. 4F,G). This demonstrated that the ADC value could be used as a noninvasive biomarker to evaluate the early treatment effect of NTX in TMZ-resistant GBM, which could result in long-term survival gain.
Our study has a limitation. In a TMZ-resistant GBM mouse model, we used a limited number of mice in each group for the evaluation of an early response to NTX as well as long-term survival to minimize the animal number according to the guideline of the institutional animal care and use committee in our institute.
In conclusion, we have demonstrated that the NTX could induce apoptosis through the decreased expression of APE-1 in TMZ-resistant GBM in vitro and inhibit the growth of a TMZ-resistant GBM tumor model, which is confirmed by a decreased tumor volume and an increased ADC after short-term NTX therapy in addition to extended survival. Particularly, since MGMT is overexpressed in ~70% of GBM, our study supports that NTX is a promising candidate for future clinical trials against TMZ-resistant GBM as a second-line treatment regimen.

Materials and Methods
Cell lines and Drugs. The human GBM cell lines LN229 and U87MG were purchased from the American Type Culture Collection (ATCC) and used for our study. LN229 and U87MG were cultured in DMEM and RPMI supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37 °C in the presence of 5% CO 2 . TMZ-resistant cell lines were established from the parental cells (LN229 and U87) by continuously exposing them to low doses of TMZ and designated LN229R and U87R for our study.
NTX and TMZ were obtained from Sigma Aldrich and dissolved in dimethyl sulfoxide (DMSO) to prepare stock concentrations, which were further diluted with culture medium to working concentrations as required.
TMZ-resistant GBM mouse model. This study was approved by the institutional animal care and use committee of the Seoul National University Hospital. All research was performed in accordance with the relevant guidelines/regulations in our institute (16-0130-C1A0). To prepare the TMZ-resistant GBM mouse model, 6-week-old male BALB/c nude mice (n = 8) were anesthetized by the intraperitoneal injection of a mixture of Zoletil (zolazepam) and Rompun (xylazine) and placed in a stereotaxic device. LN229 GBM cells (3 × 10 6 cells) were injected into the caudate/putamen region of the brain by using a Hamilton syringe fitted with a 28-gauge needle. Two weeks after tumor implantation, the required tumor size was confirmed by pretreatment MRI. For TMZ-resistant models, GBM-bearing mice were developed by administering successive high doses of TMZ (100 mg/kg/day) intraperitoneally until tumor growth showed no inhibition by TMZ for 7 days, as described in previous studies 4,28,29 . The models were further confirmed by post-1 MRI and T2WI, which is an MR protocol that was used for the assessment of tumor size.

Short-term in vivo treatment response study in a TMZ-resistant GBM mouse model. For the
in vivo animal MRI, the animals were anesthetized with 1.5-2% isoflurane/oxygen (v/v), and then scanned by using a 9.4T MR scanner (Agilent Technologies, Santa Clara, CA, USA). Throughout each imaging session, the animals were wrapped in warm water blankets, and their oxygen saturation and heart rates were monitored. For the short-term treatment response study, we divided the animals into 2 groups after the confirmation of TMZ-resistance by post-1 MRI. Four TMZ-resistant GBM mice were intraperitoneally injected with 1% saline for 7 days (control group). The remaining four TMZ-resistant GBM mice were intraperitoneally treated with 100 mg/kg/day of NTX for 7 days (NTX group). The post-2 MRI was conducted after the saline or NTX treatment to evaluate the therapeutic effects on tumor growth. DWI was also acquired sequentially. www.nature.com/scientificreports www.nature.com/scientificreports/ MRI protocol and analysis. In the T2WI in the coronal plane, fast spin-echo multi-slices methods were employed (TR n 3000 ms, effective TE = 31.18 ms, ETL = 4, average = 2, data matrix size = 256 × 256, and FOV (field of view) = 25.0 × 25.0 mm 2 ). The echo-planar DWI in the coronal plane was also obtained (TR/ TE = 4000/60.04 ms, with shots = 2, repetitions = 1, average = 2, matrix = 128 × 128, FOV = 24.0 × 24.0 mm 2 , b-value = 0, 100, 200, 400, 700 and 1000 s/mm 2 , 1 mm slice thickness). In addition, apparent diffusion coefficient (ADC) maps were generated, and image analysis was performed by using our in-house software developed with a commercial analysis package (MATLAB version R2007b, MathWorks Inc., Natick, MA, USA).
One investigator (N. K.) who was blinded to the experimental data drew the regions of interest (ROIs) that contained the entire tumor on every continuous section of the coregistered T2WI and ADC maps. Tumor boundaries were defined with reference to the high-signal intensity areas thought to represent tumor tissue on the T2WI. Finally, we calculated the tumor volume and mean ADC for each tumor, as well as the ratios of the tumor volumes and the ADC values (post-2 value/post-1 value × 100).
Long-term survival analysis in a TMZ-resistant GBM mouse model. For the long-term survival analysis of 70 days, we used TMZ-resistant GBM mice (n = 18) generated by a previously described method. The mice were divided into two groups [control, (n = 9, saline-treated) and NTX, (n = 9)] according to tumor sizes, which were equally distributed in each group. The saline solution or NTX were injected in the same manner as described above in the control and NTX groups, respectively, during the follow-up period. All mice were observed until euthanasia or the survival endpoint of 70 days.
Immunohistology. For immunohistological analysis, all animals were sacrificed at the end of the MRI studies. The brain tissues were extracted, preserved in 10% formalin and embedded in paraffin using standard protocols. We prepared 4-µm thick tissue sections and then stained them with APE-1 (Novus Biologicals; for DNA-repair) and Ki-67 (UltraMab; for cancer cell proliferation) as primary antibodies. Then, the sections were rinsed with washing buffer and incubated with secondary antibodies (Santa Cruz Biotechnology), then evaluated by DAB. The APE-1-and Ki-67-positive cells were analyzed by using ImageJ software. Additionally, the TUNEL assay was used to measure apoptotic cells.
Statistical analysis. All statistical analyses were performed using MedCalc Software (MedCalc version 13.1.0.0). The Kolmogorov-Smirnov test was used to determine the normal distribution of the noncategorical variables. The nonparametric data are presented as the median and interquartile range (IQR, range from the 25 th to the 75 th percentile), and the parametric data are shown as the mean ± standard deviation. According to the results of the Kolmogorov-Smirnov test, a paired or unpaired Student's t-test, Mann-Whitney U-test or Wilcoxon test was performed, as appropriate, to compare the values between two groups. For correlation studies, we used Spearman's correlation. Survival data in the GBM mouse model were analyzed by using the Kaplan-Meier with a log rank method and the Cox proportional hazards model in the GBM mouse model.