sign in sign up
<<
Home , Sodium selenite

Chemico-Biological Interactions 291 (2018) 29–39

Contents lists available at ScienceDirect

Chemico-Biological Interactions

journal homepage: www.elsevier.com/locate/chembioint

From epidemiology to treatment: 's prevention of brain and breast- T cancer and cardioprotection may associate with its metabolite gentisic acid

∗ Meric A. Altinoza,b, , Ilhan Elmacic, Salih Cengizd, Ebru Emekli-Alturfane, Aysel Ozpinarf a Neuroacademy Group, Istanbul, Turkey b Department of Psychiatry, Maastricht University, Holland, Netherlands c Department of Neurosurgery, Memorial Hospital, Istanbul, Turkey d Insitute of Forensics, Istanbul University, Turkey e Department of Biochemistry, Marmara University, Istanbul, Turkey f Department of Medical Biochemistry, Acibadem University, Istanbul, Turkey

ARTICLE INFO ABSTRACT

Keywords: Background: Epidemiological studies indicate that aspirin consumption reduces the risk of tumors, which is Aspirin especially relevant for colonic adenoma and carcinoma. Similar observations were made for glial brain tumors Gentisic acid and breast cancers, yet the results are inconsistent. Gentisic acid (GA) is a minor catabolite of aspirin; yet Cancer chemoprevention humans carrying CYP2C9-variants incapable to catabolize aspirin to GA do not benefit from aspirin in prevention Cancer treatment against colonic adenoma. GA blocks binding of Fibroblastic Growth Factor to its and its sulphonate metabolite dobesilic acid blocks growth of C6 glioblastoma in vivo. GA is also an endogenously produced Anthracycline cardiotoxicity siderophore in mammalians for the transport of , a trace element which stimulates tumor growth and en- hances anthracycline cardiotoxicity. Materials and Methods: In this study, we assessed whether GA exerts direct antitumor activity on C6 glioma cells in vitro (cytotoxicity, colony growth, 3H-thymidine labeling analysis of DNA synthesis); and whether it can modify growth of Ehrlich breast ascites carcinoma (EAC) and solid tumors (EST) in vivo. GA and antitumoral trace element selenium block 12-lipoxygenase activity and aspirin's paradoxical inflammatory effects are seen in selenium-deficient humans; thus, we also investigated antitumor interactions between GA and selenite. Lastly, we evaluated whether GA could protect against doxorubicin cardiotoxicity due to its function to chelate iron. Results: Clinically achievable doses of GA blocked growth, colony formation and DNA synthesis of C6 glioma in vitro with high significance. GA enhanced the survival of EAC-bearing mice at a dosage of 0.4 mg/mice/day, in which 33% of the treated animals survived more than 3-weeks, when all untreated mice succumbed to their tumors. Selenium decreased EST volumes initially, yet increased tumor volumes at later stages in surviving mice. GA alone reduced solid tumor growth and did not modify selenite antineoplasticity initially, but blocked the late tumor-stimulating effects of selenite. Lastly, doxorubicin-induced cardiac myofibrillary and endothelial damage and hyalinization necrosis were attenuated with GA treatment. Conclusions: GA highly merits to be studied in further animal models as an anticancer and chemoprotective drug.

1. Introduction inconsistent results are not easily explicable; yet we propose that re- duced catabolism of aspirin to gentisic acid (GA, 2,5-dihydroxybenzoic Conflicting findings exist regarding effects of aspirin on the risk of acid) may be responsible. The first cancer-chemopreventive efficacy of brain and breast cancers. Some studies demonstrated that usage of as- aspirin (acetylsalicylic acid) is demonstrated in colon cancer; and in- pirin is associated with reduced risk of high grade glial tumors flammatory colitis enhances risk of colon cancer which can be blocked [28,70,80], yet other studies failed to demonstrate such a protective by aspirin [37]. Aminosalicylic acid (5-ASA) is also used in the treat- efficacy [5,23]. Similarly; while some studies showed that aspirin had ment of inflammatory colitis and is a scavenger of the oxidants released no effect on the risk of breast cancer [80]; others reported reduction of by neutrophils in inflammatory bowel disease [49]. Oxidation of 5-ASA breast cancer risk with regular aspirin intake [51,78,79]. These by HOCl and reactive intermediates were defined by mass spectral

∗ Corresponding author. Guven Sk. No: 5 D:6 Kagithane, Istanbul, Turkey. E-mail address: [email protected] (M.A. Altinoz). https://doi.org/10.1016/j.cbi.2018.05.016 Received 1 March 2018; Received in revised form 17 May 2018; Accepted 29 May 2018 Available online 07 June 2018 0009-2797/ © 2018 Published by Elsevier B.V. M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39 analysis which revealed the formation of iminoquinone and quinone week. Following 1 week of incubation, colonies containing more then intermediates including GA [49]. GA is a byproduct of aspirin catabo- 50 cells were counted under an inverted microscope. Colony inhibition lism and in carriers of the CYP2C9 variant alleles, salicylate conversion was calculated with the following formula: Clonogenic Cell Death to GA is hampered, causing in a reduced inhibition of % = 100 × (1-colony number of drug treated group/colony number of synthesis and lack of risk reduction against colon adenoma [8]. control group). GA may also involve in aspirin's chemopreventive efficacy against breast and brain cancers and polymorphisms of involving in 2.3. S-Phase Fractions assased by 3H-thymidine-labelling index aspirin catabolism to GA may cause different epidemiological results. Some preliminary clues support our proposal. Solanum nigrum Linn (SN) Half an hour before the end of each 24-h period of monitoring plant has been used in South Asian folk medicine for anticancer effects, plating efficacy, cells were incubated with 1 μCi/ml 3H-thymidine which contains high amounts of GA and its extract induces autophagic (Amersham, England, sp. act. 185 GBq/mmol, 5 Ci/mmol). Following death of human AU565 breast cancer cells [40]. In a similar manner, 30 min of incubation, cells were treated with trypsin and smear slides extracts of Leonurus sibiricus L. rich in GA exerts antiproliferative and were prepared. Then slides were fixed with Carnoy's fixative (a 3:1 ratio apoptotic effect on high grade glioma cell cultures [69]. More than 65 of /glacial acetic acid) and unbound radioactive materials were years have passed, since Morel et al. demonstrated that GA retention washed twice with 2% perchloric acid at 4 °C for 30 min. Cover slips occurs in patients with cancer [56]; but only a handful studies exist on were coated with gel emulsion film (Ilford K2, England). After 3 days of the anticancer potencies of this simple and cheap agent. In this study, exposure at 4 °C, autoradiograms were washed with D-19 developer and we assessed antitumoral efficacies of GA on C6 glioma and Ehrlich slides were evaluated after Giemsa staining. With 100 × 12.5 magni- mammary carcinoma and also determined its interaction with sele- fication 3 × 103 cells were evaluated in 100 areas, cells with at least nium's antitumoral potential and cardiotoxicity of anthracycline. five grain in their nuclei were considered to be labelled. For the study, Ehrlich ascites and solid tumor models were selected because Ehrlich carcinoma cells are still considered as breast carcinoma 2.4. Determination of the antitumor efficacy of GA in Ehrlich Ascites Tumor and respond to tamoxifen treatment, despite they are derived from an model old tumor model [59,74,76]. Moreover, they are very suitable to assess anticancer activity of antiinflammatory compounds because in- All animal experimentations were carried out according to the traperitoneal injection of Ehrlich tumor cells induces guidelines of the Declaration of Helsinki. Male Balb/C mice (10-weeks- synthesis, neutrophil and macrophage infiltration [32], iNOS and COX- old, 20–25 g) were purchased from the Experimental Animal Research 1 synthesis and induction of the synthesis of MCP-1 and IL-1β, which Center-Cerrahpasa Faculty of Medicine (Istanbul), and kept there at are pro-inflammatory and protumorigenic cytokines [33]. Similarly, C6 least 1 week before use. The animals were fed with a standard pellet glioma cells significantly produce PGE2 in vitro, which is sensitive to diet and drinking water ad libitum. Ehrlich Ascites Tumor (EAT) cell aspirin and indomethazine suppression [41]; and additionally, their in lines were obtained from the same center and were propagated as vivo growth can be suppressed with aspirin treatment [2]. transplantable ascites tumors in female Balb/C mice. EAT grows both in female and male mice, yet EAT is maintained in female mice to allow 2. Materials and Methods that female hormones would stimulate the growth of this tumor ori- gined from a mammary carcinoma. Male mice were used to assess the 2.1. Cell culture, assessment of growth inhibition and drug treatments effects of drug treatments to avoid the likely effects of fluctuating fe- male sex steroids (eg. due to menstrual cycle) which may differ between C6 glioma cells obtained from American Type Culture Collection different experimental groups. (ATCC) were maintained using RPMI-1640 (Biological Industries, For ascites tumor model, tumor cells (EAT) aspirated from the Haemek, Israel) with 15% heat inactivated fetal calf serum (FCS), peritoneal cavity of mice were washed with saline and 1 × 106 tumor 0.2 mM glutamine, 50 μg/ml neomycin and 100 μg/ml streptomycin. cells were given intraperitoneally to four group of animals (14 mice per The flasks were kept in an incubator with a humidified atmosphere of group). Treatments were began following intraperitoneal tumor in-

5% CO2 at 37 °C. We used early passages (4–7 t h) of C6 cells after oculation. In the first (control) group, mice (n = 14) were given saline obtaining the cell line. Cells prepared in 5 ml of RPMI-1640 were plated solution by gavage every day. GA was dissolved in pure water and given into a six-well plate in 1 × 105/ml concentration with 100% vitality. by intragastric gavage. In the experimental groups, GA treatments were GA was obtained from Sigma-Aldrich (85707 – Sigma, purity > 99%, applied at dosages of 0.1, 0.2, 0.4 and 0.8 mg/day to second, third and CAS Number 490-79-9) and dissolved in pure water. GA was given in fourth group of mice, respectively. There were 14 mice in each dosage equal volumes of 100 μl into plates, and after 24, 48, 72 and 96 h, cells groups. The mice weighed around 20 g and the average weights of mice were harvested with trypsin-EDTA C (Biological Industries, Haemek, did not differ between these groups; hence, these treatments roughly Israel). Vital cells were identified by trypan blue exclusion and counted corresponded to 5, 10, 20 and 40 mg/kg/day of GA. Animals were on a Thoma chamber. To reveal changes of the cell proliferation, only observed and weighed everyday for the development of ascites tumor; vital cells are compared between different groups. and deaths due to tumor burden were recorded. Weight changes were calculated as percentage changes by comparison to the mean weight of 2.2. Colony growth animals in the same groups.

Stock agar (Gibco, 152–00391 M) solution of 0.3% concentration 2.5. Determination of the antitumor efficacy of GA and sodium selenite in was prepared at 45 °C. Four volumes of RPMI-1640 was rapidly mixed ehrlich solid tumor model with one volume of stock agar solution and 1 ml of this solution was filled into each well of a six-well plate and used as an underlayer. For the solid tumor model, one million EAT cells were injected in to Subsequently, plates were incubated at 4 °C for 15 min. For each well the right hind limb of male Balb/C mice. Formation of solid tumors 3×103, 100% vital cells were suspended in 1 ml RPMI-1640 media were monitored until the 10 t h day of inoculation and mice were dis- and mixed with 2 ml of (4 vol. RPMI+1 vol. Stock agar solution mix). tributed to different experimental groups which had similar size of tu- One millilitre of this suspension was put onto each under layer of six- mors which did not differ significantly from each other. Animals in well plates. Drugs were added in 100 μl final volume to the corre- group-I were kept as control, which received the vehicle. Control sponding groups. Six-well plates were then kept in an electronic in- groups (n = 14) were given saline solution intraperitoneally every day. cubator under a humidified atmosphere of 4.5% CO2 and air mix for 1 Selenium group animals (n = 15) were given 40 μg/per mice (2 mg/kg)

30 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39 of sodium selenite intraperitoneally every day. GA group animals (n = 15) were given 0.4 mg/per mice (20 mg/kg) of GA in- traperitoneally every day. GA and selenium combination group (n = 15) received 40 μg/per mice (2 mg/kg) of sodium selenite plus 0.4 mg/per mice (20 mg/kg) of GA intraperitoneally every day. Tumor diameters were measured every second day using vernier calipers and the multiplication of the diameters of the largest and shortest diameters were calculated and compared between different groups.

2.6. Treatment of Balb-C mice with EA and doxorubicin (DOX) and sampling of heart tissues for microscopical analysis

Adult male Balb/C mice (n = 20) weighing approximately 20 g were used in the experiments following ethical approval by the Cerrahpasa School of Medicine. They were fed ad libitum. Control group (n = 4) received intraperitoneal injections of the physiological saline solution at the 7 t h, 8 t h, 14 t h, 15 t h, 21 t h and 22nd days of the study. DOX group received intraperitoneal injections of DOX (5 mg/kg) at the 7 t h, 8 t h, 14 t h, 15 t h, 21 t h and 22nd days of the study. DOX + GA group received intraperitoneal injections of doxorubicin (5 mg/kg)+GA (0.4 mg/mice – (weighing 20 g, corresponding to 20 mg/kg) at the 7 t h, 8 t h, 14 t h, 15 t h, 21 t h and 22nd days of the study. At the 23 t h day, all animals were sacrified and heart tissue Fig. 1. GA Induced Growth Inhibition in C6 Glioma Cell Line. Until the 48 t h h, samples were taken into 10% tamponated formaline solution. Two only 8 and 16 μg/ml of GA were capable to significantly block cell growth μ heart samples from each group were taken into glutaraldehyde for (p < 0.005 for both); but after the 72 t h h, even low doses of GA (2 and 4 g/ fi further processing for the transmission electron microscopical studies. ml) signi cantly blocked cell growth (p < 0.005 for both). Results of triplicate experiments as assessed by Student's T test. Heart specimens were cut into small pieces and each sample (about 1mm3) was fixed in 2.0% phosphate-buffered (pH 7.2) glutaraldehyde for 2 h at 48 °C, and post-fixed in 1% phosphate-buffered (pH 7.2) os- mium tetroxide for 1 h at 48 °C. Dehydration was in graded series of ethanol solutions. After embedding in Epon 812, samples were sec- tioned to a thickness of 700–800 Å on a Reichert Ultramicrotom (OMS, Austria). Thin sections were collected on grids and stained with uranyl acetate and lead citrate before examination with a JEM 100C electron microscope.

2.7. Statistical analysis

Survival data was analyzed both by log rank test of Kaplan-Meier survival curves and by two-tailed Student's T-Test comparison of mean survival days. T-Tests were also employed for the remaining statistical analyses and p values below 0.05 were considered as significant.

3. Results

3.1. Antiproliferative activity of GA in C6 glioma (24–96 h)

Effects of various doses of GA on C6 glioma cell proliferation is Fig. 2. GA Inhibition of S-Phase Fractions in C6 Glioma Cell Line in Control and shown in Fig. 1. In the first 24 h, 2 and 4 μg/ml of GA blocked growth GA-Treated Groups (3H-Thymidine Labeling Index). A dose-response pattern 8.7% and 23.5% of the cell population, respectively; and the differences was witnessed during GA-inhibition of S-phase fraction in C6 glioma at early between total cell numbers were insignificant (p = 0.44) and borderly intervals; but prolonged incubation with low doses of GA were also efficient in ff significant (p = 0.063) in comparison to control, respectively. In the suppressing DNA synthesis. At the 72 h time point, TLI-suppressing e ect of all fi first 24 h, 8 and 16 μg/ml of GA blocked the growth of 40% and 33.4% dosages of GA were highly signi cant (p < 0.00001), which was also relevant ff for the 96 h time point, in which TLI's of all GA treated cells reduced to much of the cell population, respectively; and the di erences in cell numbers μ fi narrower values of 5.7%, 5.8%, 5.5% and 4.9% for dosages of 2, 4, 8 and 16 g/ between control and treatment groups were signi cant (p < 0.05). In ml, respectively. Please refer to the main text for more detailed statistical re- μ the 48 t h h, 2 and 4 g/ml of GA acid blocked cell growth by 26.6% and sults. Results of triplicate experiments as assessed by Student's T test. 22%, respectively (p = 0.088 and p = 0.095, respectively). In the 48 t h h, 8 and 16 μg/ml of GA blocked the growth 42.7% and 46.9% of were higly significantly different (p < 0.01) from control, respectively. the cell population, and the differences in cell numbers between control and treatment groups were significant, with p < 0.001, and p < 0.005, respectively. In the 72 t h h, even low doses of GA, 2 and 3.2. GA inhibition of S-Phase fraction in C6 glioma 4 μg/ml considerably blocked cell growth by 36% and 40.5%, respec- tively; and the differences in cell numbers between control and treat- Effects of GA on the S-phase (DNA synthesis) of C6 glioma cells were ment groups were significant (p < 0.005 for both). Lastly, in the shown in Fig. 2. Control group cells had a healthy S-phase pattern with 96 t h h, high doses of GA 8 and 16 μg/ml blocked the growth 49.5% TLI values equal to 33.7%, 31.5%, 20.7% and 22.4% at 24, 48, 72 and and 74.2% of the cell population, reducing cell numbers to levels which 96 h, respectively. The TLI values decreased during later stages of

31 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39 logaritmic growth due to the consumption of metabolites from a de- treated with GA at the dosage of 0.8 mg/per mice in comparison to creasing nutritional pool in the medium by an increasing population of control (p = 0.066). In the group treated with GA at the dosage of 0.4 cells. After 24 h, 2 and 4 μg/ml of GA decreased TLI slightly to 30.2% mg/per mice, the average value of survival days was significantly and 31.9%, which were not significantly different from control. But 8 higher in comparison to control (p = 0.0024). As shown in Fig. 3C, and 16 μg/ml of GA brought TLI values to 15.3% and 9.6, which were percentage of mice who survived longer than 3 weeks were 0%, 0%, highly different from the control by p < 0.0001 and p < 0.00001, 9%, 31% and 15%; in the control, 0.1 GA, 0.2 GA, 0.4 GA and 0.8 GA respectively. Interestingly, 48 h after treatment even low doses of GA groups, respectively. As shown in Fig. 5 weight changes of mice in each were highly inhibitory for S-phase fraction, which brought TLI's to different group did not differ from eachother. 14.9% and 14.3% with p < 0.0001 at the concentrations of 2 and 4 μg/ ml, respectively. Peculiarly, at 48 h a dosage of 8 μg/ml was less ef- 3.5. GA effects on the growth of solid ehrlich tumors alone and with sodium fective in reducing the S-phase than the lower dosages with a TLI value selenite of 20.6%, albeit still very significantly different from the control (p < 0.0001). At the same time interval, 16 μg/ml of GA brought the Here, we will provide data of early (12–20 days – Fig. 6A) and late TLI value to 12%, which was very highly different from control period (32–40 days – Fig. 6B) of tumor growth. The reason for this is (p < 0.00001). After 72 h of incubation, TLI values of all GA treated that the differences between different treatment groups decline as the cells begin to approach similar levels, and they were 11.3%, 10.4%, mice with bigger tumors die and perturbate the evaluation of the sta- 9.2% and 8.8% for the GA doses of 2, 4, 8 and 16 μg/ml, respectively. tistical analyses. We observed that both GA and selenium reduced At the 72 h time point, TLI-suppressing effect of all dosages of GA were growth of Ehrlich tumors at the early period, whereas after 30 days, the highly significant (p < 0.00001), which was also relevant for the 96 h surviving animals in the selenium group had larger tumors. On the time point, in which TLI's of all GA treated cells reduced to much other hand, such enlargement of tumors was not observed in the narrower values of 5.7%, 5.8%, 5.5% and 4.9% for dosages of 2, 4, 8 GA + selenium group. At the 16 t h and 18 t h day, GA reduction of and 16 μg/ml, respectively. A dose-response effect was observed during tumor growth occurred with statistical significance with p values of GA-inhibition of S-phase fraction in C6 glioma at early intervals; but a 0,0089 and 0,02648, respectively. Selenium was more efficient in prolonged incubation with low doses of GA were also highly effective in blocking tumor growth at the 14 t h, 16 t h, 18 t h and 20 t h days with p suppressing DNA synthesis in C6 glioma in vitro. values of 0,0013; 0,00012; 0,00253 and 0,0059, respectively. GA + se- lenium group also exerted significant growth reduction in comparison 3.3. GA Induced Inhibition of soft agar colony growth in C6 glioma to control at the 14 t h, 16 t h, 18 t h and 20 t h days with p values of 0,043817; 0,00017; 0,00049; 0,0291, respectively. At the early period – fi ff GA Induced Inhibition of Soft Agar Colony Growth in C6 Glioma (12 20 days), there were no statistically signi cant di erences between was shown in Fig. 3. Percentage inhibition of colony growth were single selenium versus GA + selenium groups. Nonetheless, compar- 2.54%, 6.52%, 5.28% and 42.6% in C6 glioma cells exposed to 2, 4, 8 ison of GA and GA + selenium groups revealed that the combination and 16 μg/ml, respectively. The difference between the control group exerted higher growth inhibitions at the 16 t h and 18 t h days with p – and the group treated with 16 μg/ml of GA was highly significant values of 0,0197 and 0,0093, respectively. At the late period (32 40 (p = 0.00022). The inhibitions observed with lower doses did not reach days), slight growth stimulations and reductions with GA in comparison fi statistical significance. to the control group did not reach statistical signi cance. The tumors in the late surviving animals in the selenium groups were larger; none- theless, as very few animals remained alive at the late periods, statis- 3.4. GA effects on the survival of mice inoculated with ehrlich ascites tical analysis demonstrated significance only at the 32 t h day carcinoma (p = 0,0378). Adding GA to selenium treatment reduced late-phase growth promoting efficacy of selenium; but no statistical significances As shown in Fig. 4A, we observed different survivals of Ehrlich as- were found due to few numbers of surviving mice. cites tumor bearing mice in GA-treated groups, yet Kaplan-Meier sta- tistics did not show statistical significance. As shown in Fig. 4B, mean 3.6. Light and electron microscopical findings in the heart tissue survival days of Ehrlich ascites tumor bearing mice were 11, 12.86, 12.07, 17.71 and 15 in groups treated with GA at doses 0.1, 0.2, 0.4 and Fig. 7A, B and 7C depict pictures taken with ×200 magnification on 0.8 mg/per mice, respectively. Student's T test revealed that average H&E stained slides. As shown in Fig. 7A, interstitial edema and lym- value of survival days exerted a trend of significance in the group phomonocytic infiltration of cardiac tissue was evident in doxorubicin treated mice. Hyalinization was also witnessed in doxorubicin treated groups (Fig. 7B), while hyalinization and interstitial edema was lesser in GE + doxorubicin treated groups (Fig. 7C). Fig. 8A, B and 8C depict pictures of healthy cardiac muscle in control group mice. Note the healthy subsarcolemmal mitochondria (curved arrow), proper myofi- brillary arrangement with healty appearing Z and M lines (zl, ml) in Figure A (×30000 magnification). A healthy intercalated disk structure (i) with discernible desmosome (parenthesis sign) is seen in Fig. 8B (×23300 magnification). In Fig. 8C, note the healthy mitochondria (m) between the myofibrils and juxtaposing cardiomyocyte nuclei (n) (×13200 magnification). Fig. 8D depicts details of a cardiac en- dothelial cell in doxorubicin treated mice; note the hypervesiculation (curved arrow), enhanced glycogen granules (g) and the likely damage of the endothelial basal membrane with leakage of the endothelial cy- toplasmic components (×46000 magnification). Fig. 3. Percentage Inhibition of C6 Glioma Growth with Various Doses of GA. Fig. 9A and B also depict cardiomyocyte damage in doxorubicin- The difference between the control group and the group treated with 16 μg/ml treated groups. On the left upper side of Fig. 9A, note the chromation of GA was highly significant (p = 0.00022). Results of triplicate experiments as condensation of the cardiomyocyte nucleus (c), myofibrillary disin- assessed by Student's T test. tegration and likely enhanced cellular debris is witnessed (curved

32 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39

Fig. 4. GA Effects on the Growth of Ehrlich Ascites Tumor (0.1, 0.2, 0.4 and 0.8 represent mg treatment per mice). A) Kaplan-Meier Graphs of Survival. B) Mean Survival Days. C) Percentage of Long-Term Survivors. Mean survival days between the control group and 0.4 mg GA-receiving group were significantly different according to the Student's T test (p = 0.0024). arrow) with vanishing of M line (vm). Mitochondria (m) were distorted were witnessed. In Fig. 11C, intercalated disk (i) structures are normal with occasional vacuolization (curved thick line) (×16600 magnifica- without osmiophilic thickening as seen in doxorubicin-treatment tion). In Fig. 9B, note that the intercalated disk (i) structure is aberrant, groups; mitochondria (m) were also normal. In the lower side of in which the osmiophilic component is highly thickened which is the Fig. 10D, note the healthy endothelial cytoplasm with proper granula- site of actin attachment (“+” large plus sign) (×26000 magnification). tion (arrow) and glycogen (g) content. Fig. 10A, B, 10C and 10D depict pictures of heart tissues obtained the single GA-treated groups. Note proper enlighnment of myofibrils, 4. Discussion proper subsarcolemmal mitochondria (ssm), intramyofibrillary sparse glycogen granules (g) and proper vesiculation of cardiac endothelia In this study, we revealed that i) GA is capable to block growth, (arrow on the upper right corner) in Fig. 9A (×40000 magnification). colony formation and DNA synthesis of C6 glioma cells in easily Connective tissue (co) and fascia adherens of the intercalated disks (i) achievable concentrations during its clinical application, ii) in vitro are seen in Fig. 9B (×13200 magnification). In Fig. 10C, fascia ad- concentrations of GA capable to block tumor growth are also within the herens is seen more closely (thick “+”-plus sign), mitochondria (m) are blood level limits which can be achieved with regular high dose aspirin intact (×40000 magnification). Note the proper alignement of myofi- treatment indicating that GA might be a component of aspirin chemo- brils with normal Z and M lines (signed as “zl” and “ml” on the figure) prevention against cancer, iii) GA reduces ascites and solid growth of (×52000 magnification). Ehrlich mammary carcinoma, which is a highly inflammatory tumor, Fig. 11A, B, 11C and 11D are pictures taken from the hearts of iii) GA hindered tumor propagating effects of high dose selenium at doxorubicin + GA treated mice (×40000 magnifications for 11 A, prolonged exposure suggesting that it can be an important component 11 B, 11C and 11D, respectively). Proper alignement of myofibrils with of redox therapies against cancer, iv) GA efficiently prevented doxor- healthy Z and M lines (signed as “zl” and “ml” on Fig. 11A, B and 11C) ubicin cardiotoxicity, suggesting a potential that it might alleviate dose

Fig. 5. Weight Gain Ratios. No statistical difference was noted between different groups.

33 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39

Fig. 6. GA and Sodium Selenite Effects on the Solid Tumor Growth of Ehrlich Mammary Carcinoma. A) Early Period (12–20 Days) and B) Late Period (32–40 Days). – Please refer to the main text for statistical significance values. At the 16 t h and 18 t h day, GA reduction of tumor growth occurred with statistical sig- nificance with p values of 0,0089 and 0,02648, respectively. Selenium was more efficient in blocking tumor growth at the 14 t h, 16 t h, 18 t h and 20 t h days with p values of 0,0013; 0,00012; 0,00253 and 0,0059, respectively. Comparison of GA and GA + selenium groups revealed that the combination exerted higher growth inhibitions at the 16 t h and 18 t h days with p values of 0,0197 and 0,0093, respectively. The tumors in the late surviving animals in the selenium groups were larger; nonetheless, as very few animals remained alive at the late periods, statistical analysis demonstrated significance only at the 32 t h day (p = 0,0378). Adding GA to selenium treatment reduced late-phase growth promoting efficacy of selenium; but no statistical significances were found due to few numbers of surviving mice.

4.1. Relevance of the selected GA doses in cell culture studies. GA May very likely contribute to the anticancer activity of aspirin and even higher levels can Be achieved in clinical application

Aspirin is first metabolized to and eliminated from the body by a combination of processes including the formation of GA [50]. In human liver microsomes, salicylic acid generates 2,3- and 2,5-DHBA (GA) but acetylsalicylic acid generates only GA [9]. Free radical attack of OH upon salicylate also produces GA and 2,3-dihydroxybenzoate [36]. In patients receiving six to eight 650 mg aspirin tablets daily (3.9–5.2 g), plasma levels between 5 and 25 μM GAwere measured [15]. In human plasma, levels of 2 mM salicylate, which can be reached under aspirin therapy, lead to levels of about 20 μMGA[31]. GA has a molecular weight of 154,12 g/mol; hence 2 μg/ml of GA corresponds to 12,9 μM. In our study, even 2 μg/ml of GA at the 96 t h h reduced S- phase from 22,4 to 5,7%. Hence, it is very plausible that GA may contribute to the antineoplasticity of aspirin, when employed at high – yet clinically applicable – doses. Salicylate blocks hyaluronidase at very high dosages in vitro, whereas GA exerts the same effect at a few μg/ml [3,55]. Since hyaluronidase involves in damage to the cartilage, GA has been introduced as an antirheumatic agent in 1948 [55] and was used about two decades. In the past, very high levels of GA up to 18–20 g per day were safely given to patients and the obtained levels in blood varied Fig. 7. Light Microscopical Analyses of the Cardiac Tissue (H&E Staining). A), B) Doxorubicin Treatment.C)GE + Doxorubicin treated group.– Please refer to according to the applied doses. Daily blood levels of GA between 40 and μ the main text for details. 80 g/ml were measured in subjects given GA at a dose of 20 g per day in divided doses [54]. In our study, continuous exposure of C6 glioma cells to 16 μg/ml (103,2 μ M) GA reduced cell growth by 74.2% and limiting toxicity of anthracyclines, which are powerful and indis- decreased its clonogenic growth by %42,6. Hence, these dosages – even pensable chemotherapy agents for a variety of cancer treatment pro- maybe higher levels – can also be very easily achieved as GA is quite tocols. Among the important features of GA antineoplasticity is its time safe at high dosages. dependency. While there was a dose response pattern in earlier inter- vals to block growth and S-phase of C6 glioma in vitro, at late intervals (72 h–96 h) and with continous exposure even low doses of GA acted 4.2. GA inhibition of Fibroblastic Growth Factor (FGF) pathway. Potentials efficient to do so. In soft agar colony growth assay, there was a dose- for blocking breast cancer and glioblastoma growth response pattern, which was measured one week after cell implantation into agar. In Ehrlich ascites carcinoma, there was a dose-response The fibroblast growth factors exert versatile protumorigenic func- pattern until 0.4 mg/day GA application, but a higher dose of GA tions by enhancing tumor cell motility, angiogenesis and also mediate (0.8 mg/day) acted less efficient in enhancing survival, which is not resistance of cancer cells against chemotherapy and radiotherapy by easily explicable. Not every member of the prostaglandin family is increasing the apoptotic threshold [34]. FGF family members behave protumorigenic, in fact J-series of (prostamides) could like oncogenes in virally induced mouse breast cancer [25,42]. In induce endoplasmic reticulum stress and cell death in skin cancer cells parallel, lucitanib, a FGFR-inhibitor exerts antitumor activity in human [72]. It is possible that higher doses of GA blocked other eicosanoids breast cancer [19]. Robust vascularization is a major characteristic of than tumor-promoting prostaglandins, which may partly limit tumor high grade gliomas; FGF-1 and FGF-2 are abundant in most glio- growth and hence; there might exist an U-shaped curve of GA in acting blastomas, the former is associated to malignant astrocytes, the latter is anticancer in vivo. associated with the matrix surrounding neovascularization [34]. GA was discovered to belong to a novel chemical group of FGF inhibitors, which includes homogentisic acid (HGA; 2-(2,5-dihydroxyphenyl)acetic acid), the toxic agent in alkaptonuria [34]. These compounds dissociate

34 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39

Fig. 8. Electron Microscopical Analyses of the Cardiac Tissue. A), B), C) – Control Groups. D) Doxorubicin Treated Groups. – Please refer to the main text for details. the FGF-receptor·growth factor signaling complex [34]. The potential [20]. There exist conflicting data on the effects of VIP on glioma inhibition of FGF mitogenic activity by GAwas first assayed in vitro growth. Some authors proposed that VIP acted proliferative on glioma using FGF-1 and 3T3 fibroblasts, where it displayed a half-maximum [6,71]. Nonetheless, there exist substantial opposite evidence sug- inhibitory activity (IC50)of∼36 μM[34]. Substitution of carboxylate gesting that VIP might act antiproliferative in human [77]; and serum by sulfonate (2,5-dihydroxybenzene sulfonate (2,5-DHBS), or so called starved rat C6 glioma cells [21]. Moreover, more recent studies showed dobesilic acid) caused a prominent decrease in the IC50 (3 μM). Hence, prominent antiinvasive potential of VIP on human glial tumor cells both the investigators inoculated rats with C6 glioblastoma cells and five in normoxia [17]; and hypoxia [52]. We think that 3-dimensional days after implantation, when the tumors protruded throughout the spheroid cultures of human glioma may reveal the net effect of GA on skin, they treated rats with either vehicle only or 2,5-DHBS (100 mg/ VIP-associated pathways in glioma invasion. kg). 2,5-DHBS treatment blocked the growth of C6 rat glioma by around 50% [34]. In our future studies, we will assess the effects of high dose 4.4. GA as a constituent of mammalian iron chelating-siderophore. GA (upto 300 mg/kg) on the intracranially implanted C6 glioma, as GA Relevance for the anticancer effects and anthracycline cardiotoxicity may easily traverse the blood-brain barrier due to its quite small mo- lecular structure and . Iron maintains the growth of cancer cells and microbial pathogens; for instance, breast cancer cells exert 5–15 times more transferrin re- 4.3. GA inhibition of cyclin dependent kinases and agonism of Vasoactive ceptors than the normal breast tissue [64]. In humans, increased body Intestinal Peptide pathway iron enhances the risk of several cancers including breast cancer [47]. Reactive oxygen species (ROS) produced by aerobic metabolism cause Cyclin dependent kinases (CDKs) are master regulators of the cell the release of free iron from ferritin, and in the presence of superoxide 3+ cycle, and get activated by binding with temporally expressed cyclins radical and H2O2, stored ferric iron (Fe ) is reduced to ferrous iron during different cell cycle steps. CDK1 and CDK2 are activated via (Fe2+), which catalyzes the formation of the hydroxyl radical (.OH). binding to cyclins A and B and provide the progression through G2 and .OH in turn stimulates lipid peroxidation, oncogene activation and M phases [22]. CDK1 enhances anchorage-independent growth of tumor suppressor inhibition, enhancing the breast cancer risk. Iron le- glioblastoma and its escape from temozolomide-induced senescence vels in the mitochondria shall be tightly regulated; an insufficient de- [73]. GA and GA isomers 2,3-dihydroxybenzoic acid (2,3-DHBA), 2,4- livery of iron can hinder the its metabolic functions, whereas supra- dihydroxybenzoic acid and 2,6-dihydroxybenzoic acid could block physiological levels of mitochondrial iron could also trigger ROS [24]. CDK1 activity, which may also explain its anticancer potential The mammalian antibacterial protein siderocalin/lipocalin 24p3 func- in glioma and mammary carcinoma [22]. tions by sequestering iron to reduce iron levels necessary for bacterial In a guinea pig isolated trachea model, GA induces relaxation via growth and its labile component was defined as GA [18]. Salicylate acting as an agonist of Vasoactive Intestinal Peptide (VIR)-receptor analogues form stable chelates with biologically important cations

35 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39

Fig. 9. Electron Microscopical Analyses of the Cardiac Tissue. Both A) and B) Depict Heart Pieces from Doxorubicin Treated Groups. – Please refer to the main text for details.

Cu2+,Fe3+,Al3+ with the following orders of chelate stability: salicylic progressive tumor growth, as well as chemo- and radiotherapy re- acid > GA > salicyluric acid, and Fe+3 >Al+3 >Cu+2 [62]. As sistance of cancer cells by inducing repopulation of cancer stem cells iron is exploited in cancers for growth [30] and excess iron contributes [60]. Formation of GA from salicylate acid might happen in an in- to the anthracycline cardiotoxicity [29,53]; GA – as an endogenous iron flammatory micromilieu as a consequence of “co-oxidation” coupled to chelating siderophore – may uncouple antitumoral and cardiotoxic prostaglandin-H-Synthase (PGH-Synthase)-peroxidase activity, which is actions of chemotherapy. commenly observed with a variety of phenolic compounds [39]. GA in this oxidative micromilieu might undergo further oxidation to reactive quinonoid compounds, which could irreversible inhibit COX via cova- 4.5. Antioxidant activity of GA with emphases to the inhibition of lent binding. Salicylate does not, unlike its acetylated derivative as- and anticancer activity pirin, inhibit COX-1 and COX-2 activity in vitro but demonstrates a comparable analgesic and anti-inflammatory action as aspirin [38]. It is suggested that GA blockage of COX (, also However, the salicylate metabolite GA (10–100 μM) significantly in- known as prostaglandin endoperoxide synthase) activity may also as- hibited LPS-induced PGE2 production [38]. Albeit the aspirin fraction sociate with its antioxidant activity [10]. GA could inhibit COX via metabolized to GA is rather small, GA is predominantly formed under quenching the tyrosyl radical at the active site of COX, which catalyzes inflammatory conditions by neutrophils (which highly produce ROS) the oxygenation and cyclization of to produce PGG2 and febrile patients undergoing salicylate treatment excrete higher GA [10]. COX-2 is overexpressed in malignant cells and contributes to

36 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39

Fig. 10. Electron Microscopical Analyses of the Cardiac Tissue. A), B), C) and D) Depict Heart Pieces from GA-Treated Groups. – Please refer to the main text for details.

levels [38]. solid tumor growth in Ehrlich carcinoma bearing mice; but in longer- surviving animals selenite-treated mice had larger tumors. GA did not 4.6. GA interactions with selenium effect selenite's anticancer activity at early period, yet suppressed its increase of tumor volumes at later periods. Selenium is a trace In our study, we observed that sodium selenite strongly reduced which selectively accumulates within tumor tissues and is depleted in

Fig. 11. Electron Microscopical Analyses of the Cardiac Tissue. A), B), C) and D) Depict Heart Pieces from GA + Doxorubicin Treated Groups. – Please refer to the main text for details.

37 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39 benign tissues during tumor development which is relevant for systemic Trace Elem. Res. 13 (1) (1987 Aug) 55–67, http://dx.doi.org/10.1007/ cancers and brain malignancies [7][68]; [4]; [81]. Selenium is the BF02796621. [2] O. Arrieta, P. Guevara, S. Reyes, G. Palencia, E. Rivera, J. Sotelo, Paradoxical effect cofactor of glutathione peroxidase, which involves in the recycle of of aspirin on the growth of C6 rat glioma and on time of development of ENU- glutathione. Glutathione recycle involves glutathione reductase which induced tumors of the nervous system, J. Canc. Res. Clin. Oncol. 127 (11) (2001 provides the renewal of non-oxidized glutathione, while glutathione Nov) 681–686. [3] B.D. Astill, D.W. Fassett, R.L. Roudabush, The metabolism of phenolic antioxidants. peroxidase oxidizes glutathione to reduce intracellular oxidant mole- 4. The metabolites of gentisic acid in the dog, Biochem. J. 90 (1) (1964 Jan) cules. Hence, despite selenium is involved in an antioxidant recyle 194–201. process, it shall be kept in mind that it involves in the oxidizing side of [4] G.A. Babenko, I.P. Pogribnyĭ, Dynamics of the selenium content and selenium-de- this cycle. Therefore, in conditions when intracellular oxidant mole- pendent glutathione peroxidase activity in mice and rats with transplantable tu- mors, Vopr. Onkol. 32 (6) (1986) 93–98. cules increase to very high levels, selenite may act as prooxidant; and [5] F.J. Bannon, M.A. O'Rorke, L.J. Murray, C.M. Hughes, A.T. Gavin, S.J. Fleming, indeed, high levels of selenium per se may act prooxidant [26]. Sodium C.R. Cardwell, Non-steroidal anti-inflammatory drug use and brain tumour risk: a selenite is among the selenium compounds with the greatest anticancer case-control study within the clinical practice research datalink, Cancer Causes Control 24 (11) (2013) 2027-34. ffi +4 e cacy with Se in the oxidation state and could directly oxidize [6] A. Barbarin, P. Séité, J. Godet, S. Bensalma, J.M. Muller, C. Chadéneau, Atypical thiol-containing cellular substrates. nuclear localization of VIP receptors in glioma cell lines and patients, Biochem. Biophys. Res. Commun. 454 (4) (2014) 524–530. Reduced glutathione reduces cyclooxygenase activity; vice versa, ff fi [7] W.A. Baumgartner, V.A. Hill, E.T. Wright, Antioxidant e ects in the development of selenium de ciency hinders aspirin potential to block cyclooxygenases Ehrlich ascites carcinoma, Am. J. Clin. Nutr. 31 (3) (1978 Mar) 457–465. [11]. Hence, our working hypothesis is that selenium acts antitumoral [8] J. Bigler, J. Whitton, J.W. Lampe, L. Fosdick, R.M. Bostick, J.D. Potter, CYP2C9 and at low doses via antioxidant and antiinflammatory mechanisms, yet UGT1A6 genotypes modulate the protective effect of aspirin on colon adenoma risk, Canc. Res. 61 (9) (2001 May 1) 3566–3569. when its intratumoral levels increase to very high levels in animals [9] M. Bojić, C.A. Sedgeman, L.D. Nagy, F.P. Guengerich, Aromatic hydroxylation of which survived longer, it acted prooxidant and proinflammatory. At salicylic acid and aspirin by human cytochromes P450, Eur. J. Pharmaceut. Sci. 73 this stage, presence of GA would block putative prooxidant and (2015) 49–56. fl ff [10] R.S. Borges, S.L. Castle, The antioxidant properties of salicylate derivatives: a proin ammatory e ects of selenite and prevent its propagation of possible new mechanism of anti-inflammatory activity, Bioorg. Med. Chem. Lett 25 tumor growth. Here, it shall be remembered that high dose and pro- (21) (2015 Nov 1) 4808–4811, http://dx.doi.org/10.1016/j.bmcl.2015.07.001. longed aspirin consumption reduces levels of selenium [44] and vice [11] J.H. Capdevila, J.D. Morrow, Y.Y. Belosludtsev, D.R. Beauchamp, R.N. DuBois, versa, unexplained detrimental and proinflammatory effects of aspirin J.R. Falck, The catalytic outcomes of the constitutive and the mitogen inducible isoforms of synthase are markedly affected by glutathione and in Reye Syndrome and asthma is observed in patients with selenium glutathione peroxidase(s), Biochemistry 34 (10) (1995 Mar 14) 3325–3337. deficiency [1,61]. Hence, in the future it would be very tempting to [15] L.G. Cleland, P.J. Lowthian, D. Imhoff, F. Bochner, W.H. Betts, J. O'Callaghan, Plasma and synovial fluid gentisate in patients receiving salicylate therapy, J. study whether selenium and aspirin metabolites may synergy in – fl Rheumatol. 12 (1) (1985) 136 139 Feb. blocking in ammatory processes and cancer. [17] S. Cochaud, A.C. Meunier, A. Monvoisin, S. Bensalma, J.M. Muller, C. Chadéneau, Here, we shall underline that another aspirin metabolite, 2,3-dihy- Neuropeptides of the VIP family inhibit glioblastoma cell invasion, J. Neuro Oncol. – droxybenzoic acid (also called: hypogallic acid, 2-pyrocatechuic acid, o- 122 (1) (2015) 63 73, http://dx.doi.org/10.1007/s11060-014-1697-6. [18] C. Correnti, V. Richardson, A.K. Sia, A.D. Bandaranayake, M. Ruiz, et al., pyrocatechuic acid) may also act antitumoral as demonstrated in mel- Siderocalin/Lcn2/NGAL/24p3 does not drive apoptosis through gentisic acid anoma cells [43]. In pancreas cancer cells, 2,3-dihydroxybenzoic acid mediated iron withdrawal in hematopoietic cell lines, PLoS One 7 (8) (2012) triggers oxidative stress and DNA damage, which are suppressed by e43696. [19] C. Criscitiello, A. Esposito, S. De Placido, G. Curigliano, Targeting fibroblast growth bathocuproine, a Cu(I) chelator, and catalase indicating that this aspirin factor receptor pathway in breast cancer, Curr. Opin. Oncol. 27 (6) (2015 Nov) metabolite induced reaction of H2O2 with Cu(I) participate in triggering 452–456, http://dx.doi.org/10.1097/CCO.0000000000000224. oxidative DNA injury [58]. The authors interpreted these findings in [20] J.F. Cunha, F.D. Campestrini, J.B. Calixto, A. Scremin, N. Paulino, The mechanism of gentisic acid-induced relaxation of the Guinea pig isolated trachea: the role of terms of the carcinogenic potential of aspirin and 2,3-dihydroxybenzoic channels and vasoactive intestinal peptide receptors, Braz. J. Med. Biol. acid. But we think the opposite as they have shown these effects in Res. 34 (3) (2001 Mar) 381–388. malignant cells. The main action mechanisms of orthodox cancer [21] A.G. D'Amico, S. Scuderi, S. Saccone, A. Castorina, F. Drago, V. D'Agata, ff treatments including chemotherapy and radiotherapy are based on in- Antiproliferative e ects of PACAP and VIP in serum-starved glioma cells, J. Mol. Neurosci. 51 (2) (2013 Oct) 503–513. ducing oxidative injury and DNA damage in cancer cells; and hence, we [22] R. Dachineni, D.R. Kumar, E. Callegari, S.S. Kesharwani, R. Sankaranarayanan, think that 2,3-dihydroxybenzoic acid may even act as a chemo-radio- T. Seefeldt, H. Tummala, G.J. Bhat, Salicylic acid metabolites and derivatives in- ff sensitizer in cancer. hibit CDK activity: novel insights into aspirin's chemopreventive e ects against colorectal cancer, Int. J. Oncol. 51 (6) (2017 Dec) 1661–1673. [23] S.E. Daugherty, S.C. Moore, R.M. Pfeiffer, P.D. Inskip, Y. Park, A. Hollenbeck, 5. Conclusions P. Rajaraman, Nonsteroidal anti-inflammatory drugs and glioma in the NIH-AARP Diet and Health Study cohort, Canc. Prev. Res. 4 (12) (2011) 2027-34. ff [24] L.R. Devireddy, D.O. Hart, D.H. Goetz, M.R. Green, A mammalian siderophore It is without doubt that the e orts to discover novel molecules carry synthesized by an enzyme with a bacterial homolog involved in enterobactin pro- tremendous importance for the development of new cancer treatment duction, Cell 141 (6) (2010 Jun 11) 1006–1017. strategies. Nonetheless, repurposal of known drugs for cancer treatment [25] C. Dickson, B. Spencer-Dene, C. Dillon, V. Fantl, Tyrosine kinase signalling in breast fi fi cancer: broblast growth factors and their receptors, Breast Cancer Res. 2 (3) may also provide novel and bene cial results. Drugs with simple mo- (2000) 191–196. lecular structure may be useful; and among these, GA highly merits to [26] E.N. Drake, Cancer chemoprevention: selenium as a prooxidant, not an antioxidant, be studied in different cancer models as an anticancer adjuvant. Med. Hypotheses 67 (2) (2006) 318–322. [28] K.M. Egan, L.B. Nabors, Z.J. Thompson, C.M. Rozmeski, G.A. Anic, J.J. Olson, Moreover, GA also carries the potential to uncouple chemotherapy R.V. LaRocca, S.A. Chowdhary, P.A. Forsyth, R.C. Thompson, Analgesic use and the toxicity against malignant versus benign tissues. It would be very risk of primaryadult brain tumor, Eur. J. Epidemiol. 31 (9) (2016 Sep) 917–925. plausible to evaluate combinations of GA with anthracyclines to test [29] N. Elihu, S. Anandasbapathy, W.H. Frishman, Chelation therapy in cardiovascular this hypothesis. Lastly, interactions of GA and aspirin with selenium disease: ethylenediaminetetraacetic acid, deferoxamine, and dexrazoxane, J. Clin. Pharmacol. 38 (2) (1998 Feb) 101–105. compounds also deserve to be studied both for the discovery of novel [30] R.L. Elliott, M.C. Elliott, F. Wang, J.F. Head, Breast carcinoma and the role of iron chemotherapy strategies and also to find safer antiinflammatory path- metabolism. A cytochemical, tissue culture, and ultrastructural study, Ann. N. Y. – ways in non-malignant diseases (aspirin sensitive asthma, Reye syn- Acad. Sci. 698 (1993 Nov 30) 159 166. [31] M. Exner, M. Hermann, R. Hofbauer, S. Kapiotis, W. Speiser, I. Held, C. Seelos, drome etc.). B.M. Gmeiner, The salicylate metabolite gentisic acid, but not the parent drug, inhibits glucose autoxidation-mediated atherogenic modification of low li- References poprotein, FEBS Lett. 470 (1) (2000) 47–50. [32] D. Fecchio, P. Sirois, M. Russo, S. Jancar, Studies on inflammatory response induced by Ehrlich tumor in mice peritoneal cavity, Inflammation 14 (1) (1990 Feb) [1] J.M. Andres, R.W. Hurd, H.A. Van Rinsvelt, Application of trace metal analysis to 125–132. basic problems in neurobiology studies of patients with Reye's syndrome, Biol. [33] P.D. Fernandes, F.S. Guerra, N.M. Sales, T.B. Sardella, S. Jancar, J.S. Neves,

38 M.A. Altinoz et al. Chemico-Biological Interactions 291 (2018) 29–39

Characterization of the inflammatory response during Ehrlich ascitic tumor devel- cellular and isolated DNA induced by a metabolite of aspirin, Mutat. Res. 661 (1–2) opment, J. Pharmacol. Toxicol. Meth. 71 (2015 Jan-Feb) 83–89. (2009) 93–100. [34] I.S. Fernández, P. Cuevas, J. Angulo, P. López-Navajas, A. Canales-Mayordomo, [59] G. Ozcan Arican, A. Ozalpan, Evaluation of the effect of paclitaxel, epirubicin and R. González-Corrochano, R.M. Lozano, S. Valverde, J. Jiménez-Barbero, A. Romero, tamoxifen by cell kinetics parameters in estrogen-receptor-positive ehrlich ascites G. Giménez-Gallego, Gentisic acid, a compound associated with plant defense and a tumor (eat) cells growing in vitro, Acta Biol. Hung. 58 (1) (2007 Mar) 49–59. metabolite of aspirin, heads a new class of in vivo fibroblast growth factor in- [60] L.Y. Pang, E.A. Hurst, D.J. Argyle, Cyclooxygenase-2: a role in cancer stem cell hibitors, J. Biol. Chem. 285 (15) (2010 Apr 9) 11714–11729. survival and repopulation of cancer cells during therapy, Stem Cell. Int. 2016 [36] M. Grootveld, B. Halliwell, Aromatic hydroxylation as a potential measure of hy- (2016) 2048731. droxyl-radical formation in vivo. Identification of hydroxylated derivatives of sal- [61] D.J. Pearson, V.J. Suarez-Mendez, J.P. Day, P.F. Miller, Selenium status in relation icylate in human body fluids, Biochem. J. 237 (2) (1986 Jul 15) 499–504. to reduced glutathione peroxidase activity in aspirin-sensitive asthma, Clin. Exp. [37] Y. Guo, Z.Y. Su, C. Zhang, J.M. Gaspar, R. Wang, R.P. Hart, M.P. Verzi, A.N. Kong, Allergy 21 (2) (1991 Mar) 203–208. Mechanisms of colitis-accelerated colon carcinogenesis and its prevention with the [62] J. Pecci, W.O. Foye, The avidity of salicylic, gentisic, and salicyluric acids for heavy combination of aspirin and : transcriptomic analysis using RNA-seq, metal cations, J Am Pharm Assoc Am Pharm Assoc. 49 (1960 Jul) 411–414. Biochem. Pharmacol. 135 (2017) 22–34. [64] P. Reizenstein, Iron, free radicals and cancer, Med. Oncol. Tumor Pharmacother. 8 [38] B. Hinz, V. Kraus, A. Pahl, K. Brune, Salicylate metabolites inhibit cyclooxygenase- (4) (1991) 229–233. 2-dependent prostaglandin E(2) synthesis in murine macrophages, Biochem. [68] O.P. Savel'eva, N.P. Fadeev, Study of the tropism of sodium Selenium-75-selenite to Biophys. Res. Commun. 274 (1) (2000) 197–202. experimental cancer of the mammary gland, Vopr. Onkol. 25 (9) (1979) 50–53. [39] T.J. Holmes Jr., J.L. Vennerstrom, V. John, Inhibition of cyclooxygenase by elec- [69] P. Sitarek, E. Skała, M. Toma, M. Wielanek, J. Szemraj, T. Skorski, A.J. Białas, trochemical oxidation of gentisic acid, J. Biol. Chem. 260 (26) (1985) 14092–14095 T. Sakowicz, T. Kowalczyk, M. Radek, H. Wysokińska, T. Śliwiński, Transformed Nov 15. root extract of Leonurus sibiricus induces apoptosis through intrinsic and extrinsic [40] H.C. Huang, K.Y. Syu, J.K. Lin, Chemical composition of Solanum nigrum linn ex- pathways in various grades of human glioma cells, Pathol. Oncol. Res. 23 (3) (2017 tract and induction of autophagy by leaf water extract and its major flavonoids in Jul) 679–687. AU565 breast cancer cells, J. Agric. Food Chem. 58 (15) (2010 Aug 11) 8699–8708. [70] N.R. Sivak-Sears, J.A. Schwartzbaum, R. Miike, M. Moghadassi, M. Wrensch, Case- [41] S.L. Hwang, K.S. Lee, C.L. Lin, A.S. Lieu, C.Y. Cheng, J.K. Loh, Y.F. Hwang, Y.F. Su, control study of use of nonsteroidal antiinflammatory drugs and glioblastoma S.L. Howng, Effect of aspirin and indomethacin on prostaglandin E2 synthesis in C6 multiforme, Am. J. Epidemiol. 159 (12) (2004) 1131–1139. glioma cells, Kaohsiung J. Med. Sci. 20 (1) (2004 Jan) 1–5. [71] P. Sokolowska, J.Z. Nowak, Effects of PACAP and VIP on cAMP-generating system [42] D. Jackson, J. Bresnick, C. Dickson, A role for fibroblast growth factor signaling in and proliferation of C6 glioma cells, J. Mol. Neurosci. 36 (1–3) (2008 Nov) the lobuloalveolar development of the mammary gland, J. Mammary Gland Biol. 286–291. Neoplasia 2 (4) (1997 Oct) 385–392. [72] E. Soliman, K.L. Henderson, A.S. Danell, R. Van Dross, Arachidonoyl-ethanolamide [43] E.P. Kable, P.G. Parsons, Potency, selectivity and cell cycle dependence of catechols activates endoplasmic reticulum stress-apoptosis in tumorigenic keratinocytes: role in human tumour cells in vitro, Biochem. Pharmacol. 37 (9) (1988) 1711–1715. of cyclooxygenase-2 and novel J-series prostamides, Mol. Carcinog. 55 (2016) [44] V. Kesik, M.K. Lenk, A.E. Kurekci, C.H. Acikel, E.O. Akgul, A. Aydin, O. Erdem, 117–130. M. Gamsizkan, Do and selenium prevent the antioxidant, hepatic and renal [73] Z. Song, Y. Pan, G. Ling, S. Wang, M. Huang, X. Jiang, Y. Ke, Escape of U251 glioma system impairment caused by aspirin in rats? Biol. Trace Elem. Res. 123 (1–3) cells from temozolomide-induced senescence was modulated by CDK1/survivin (2008) 168–178, http://dx.doi.org/10.1007/s12011-008-8115-x Summer. signaling, Am J Transl Res. 9 (5) (2017 May 15) 2163–2180. [47] J.G. Liehr, J.S. Jones, Role of iron in estrogen-induced cancer, Curr. Med. Chem. 8 [74] G.M. Suddek, Allicin enhances chemotherapeutic response and ameliorates ta- (7) (2001 Jun) 839–849. moxifen-induced liver injury in experimental animals, Pharm. Biol. 52 (8) (2014 [49] Z.C. Liu, R.A. McClelland, J.P. Uetrecht, Oxidation of 5-aminosalicylic acid by Aug) 1009–1014, http://dx.doi.org/10.3109/13880209.2013.876053. hypochlorous acid to a reactive iminoquinone. Possible role in the treatment of [76] P. Tomšík, S. Mičuda, D. Muthná, E. Čermáková, R. Havelek, E. Rudolf, M. Hroch, inflammatory bowel diseases, Drug Metab. Dispos. 23 (2) (1995) 246–250. Z. Kadová, M. Řezáčová, J. Ćmielová, P. Živný, Boldine inhibits mouse mammary [50] A. Lorico, P. Masturzo, S. Villa, M. Salmona, N. Semeraro, G. de Gaetano, Gentisic carcinoma in vivo and human MCF-7 breast cancer cells in vitro, Planta Med 82 acid: an aspirin metabolite with multiple effects on human blood polymorpho- (16) (2016 Nov) 1416–1424. nuclear leukocytes, Biochem. Pharmacol. 35 (14) (1986 Jul 15) 2443–2445. [77] P. Vertongen, I. Camby, F. Darro, R. Kiss, P. Robberecht, VIP and pituitary ade- [51] L. Lu, L. Shi, J. Zeng, Z. Wen, Aspirin as a potential modality for the chemopre- nylate cyclase activating polypeptide (PACAP) have an antiproliferative effect on vention of breast cancer: a dose-response meta-analysis of cohort studies from the T98G human glioblastoma cell line through interaction with VIP2 receptor, 857,831 participants, Oncotarget 8 (25) (2017 Jun 20) 40389–40401. Neuropeptides 30 (5) (1996) 491–496. [52] G. Maugeri, A. Grazia D'Amico, R. Reitano, G. Magro, S. Cavallaro, S. Salomone, [78] Y.S. Yang, E. Kornelius, J.Y. Chiou, Y.R. Lai, S.C. Lo, C.H. Peng, C.N. Huang, Low- V. D'Agata, PACAP and VIP inhibit the invasiveness of glioblastoma cells exposed to dose aspirin reduces breast cancer risk in women with diabetes: a nationwide ret- hypoxia through the regulation of HIFs and EGFR expression, Front. Pharmacol. 7 rospective cohort study in taiwan, J Womens Health (Larchmt) 26 (12) (2017 Dec) (2016 May 31) 139, http://dx.doi.org/10.3389/fphar.2016.00139. 1278–1284. [53] J.V. McGowan, R. Chung, A. Maulik, I. Piotrowska, J.M. Walker, D.M. Yellon, [79] X. Zhang, S.A. Smith-Warner, L.C. Collins, B. Rosner, W.C. Willett, S.E. Hankinson, Anthracycline chemotherapy and cardiotoxicity, Cardiovasc Drug Ther. 31 (1) Use of aspirin, other nonsteroidal anti-inflammatory drugs, and acetaminophen and (2017) 63–75, http://dx.doi.org/10.1007/s10557-016-6711-0. postmenopausal breast cancer incidence, J. Clin. Oncol. 30 (28) (2012 Oct 1) [54] B.W. Meade, M.J. Smith, The estimation of sodium gentisate in plasma and urine, J. 3468–3477. Clin. Pathol. 4 (2) (1951 May) 226–230. [80] T. Zhang, X. Yang, P. Liu, J. Zhou, J. Luo, H. Wang, A. Li, Y. Zhou, Association [55] K. Meyer, C. Ragan, The antirheumatic effect of sodium gentisate, Science 108 between nonsteroidal anti-inflammatory drugs use and risk of central nervous (2802) (1948 Sep) 281. system tumors: a dose-response meta analysis, Oncotarget 8 (60) (2017 Oct 11) [56] A. Morel, A. Josserand, J. Chappet, Sodium gentisate retention in a cancerous or- 102486–102498. ganism; possible relation with collagen pathology, C. R. Seances Soc. Biol. Fil. 146 [81] Z. Zhang, Y. Chinen, Z. Zhu, M. Kimura, Y. Itokawa, Uptake and distribution of (17–18) (1952 Sep) 1341–1342. sodium selenite in rat brain tumor, Biol. Trace Elem. Res. 48 (1) (1995 Apr) 45–50. [58] S. Oikawa, H. Kobayashi, S. Tada-Oikawa, Y. Isono, S. Kawanishi, Damage to

39