PAN-811 Blocks Chemotherapy Drug-Induced In Vitro Neurotoxicity, While Not Affecting Suppression of Cancer Cell Growth

Chemotherapy often results in cognitive impairment, and no neuroprotective drug is now available. This study aimed to understand underlying neurotoxicological mechanisms of anticancer drugs and to evaluate neuroprotective effects of PAN-811. Primary neurons in different concentrations of antioxidants (AOs) were insulted for 3 days with methotrexate (MTX), 5-fluorouracil (5-FU), or cisplatin (CDDP) in the absence or presence of PAN-811·Cl·H2O. The effect of PAN-811 on the anticancer activity of tested drugs was also examined using mouse and human cancer cells (BNLT3 and H460) to assess any negative interference. Cell membrane integrity, survival, and death and intramitochondrial reactive oxygen species (ROS) were measured. All tested anticancer drugs elicited neurotoxicity only under low levels of AO and elicited a ROS increase. These results suggested that ROS mediates neurotoxicity of tested anticancer drugs. PAN-811 dose-dependently suppressed increased ROS and blocked the neurotoxicity when neurons were insulted with a tested anticancer drug. PAN-811 did not interfere with anticancer activity of anticancer drugs against BNLT3 cells. PAN-811 did not inhibit MTX-induced death of H460 cells but, interestingly, demonstrated a synergistic effect with 5-FU or CDDP in reducing cancer cell viability. Thus, PAN-811 can be a potent drug candidate for chemotherapy-induced cognitive impairment.


Introduction
One of the most common complications of chemotherapeutic drugs is toxicity to the central nervous system (CNS), namely, chemotherapy-induced cognitive impairment or chemobrain. This toxicity can present in many ways, including encephalopathy syndromes and confusional states, seizure activity, headache, cerebrovascular complications and stroke, visual and hearing loss, cerebellar dysfunction, and spinal cord damage with myelopathy [1]. Mild to moderate effects of chemotherapy on cognitive performance occur in 15-50% of the survivors after treatment [2,3]. The cognitive problems can last for many years after the completion of chemotherapy in a subset of cancer survivors. Up to 70% of patients with cancer report that these cognitive difficulties persist well beyond the duration of treatment [4][5][6]. Chemobrain can seriously affect quality of life and life itself in cancer patients.
Among possible candidate mechanisms, oxidative stress (OS) may play a key role in cognitive disorders caused by broad types of anticancer drugs, such as antimetabolites, mitotic inhibitors, topoisomerase inhibitors, and paclitaxel [7]. These chemotherapeutic agents are not known to rely on oxidative mechanisms for their anticancer effects. Among the antimetabolite drugs, methotrexate (MTX) and 5-fluorouracil (5-FU), widely used chemotherapeutic agents, are most likely to cause CNS toxicity [1]. Although there are yet no reports of 5-FU increasing CNS OS, it has been observed to induce apoptosis in rat cardiocytes through intracellular OS [8], to increase OS in the plasma of liver cancer patients [9], and to decrease glutathione (GSH) in bone marrow cells [10]. MTX can also cross the bloodbrain barrier as well [11] and result in an increase of OS in cerebral spinal fluid and executive dysfunction in MTXtreated patients of pediatric acute lymphoblastic leukemia [12,13]. It is well known that ROS, such as H 2 O 2 , can result in neuronal cell death [14,15]. Cisplatin (CDDP) is an alkylating agent. Its cytotoxic effect is thought to be mediated primarily by the generation of nuclear DNA adducts, which, 2 Oxidative Medicine and Cellular Longevity if not repaired, cause cell death as a consequence of DNA replication and transcription blockage. However, oxidative damage has been observed in vivo following exposure to CDDP in several tissues including nervous tissue, suggesting a role for OS in the pathogenesis of CDDP-induced doselimiting toxicities [16][17][18]. Cotreatment with antioxidants (AOs) suppresses the toxic effects of CDDP on several organs [19,20].
Currently, there are no proven treatments for chemotherapy-induced cognitive impairment. Some efforts have been focused on correcting cognitive deficits rather blocking the neurotoxic pathway of chemotherapeutic drugs [21]. Since ROS mediates neurotoxicity in a number of neurodegenerative disorders, one strategy in disease control has been focused on development of antioxidants as preventive and therapeutic molecules. These include vitamin C, vitamin E, glutathione, coenzyme Q (CoQ), carotenoids, melatonin, and green tea extract [22,23]. In contrast to the minimal positive effects of these efforts, antioxidative therapy could be a promising strategy for the treatment of neurotoxicity. Several preclinical studies have shown that AO treatment prevents chemotherapy-induced OS and cognitive deficits when administered prior to and during chemotherapy [24,25].
Our previous research has demonstrated that PAN-811 (known as 3-aminopyridine-2-carboxaldehyde thiosemicarbazone or Triapine), a bioavailable small molecule (MW 195) currently in phase II clinical trials for the treatment of patients with cancer, can efficiently block neurodegeneration. Major underlying mechanisms for the neuroprotection of PAN-811 are blockage of both excitatory pathway and OS [22]. Hence we hypothesized that PAN-811 could protect neurons from anticancer drugs, such as MTX, 5-FU, and CDDP. Since PAN-811 is an anticancer drug targeting ribonucleotide reductase, which is distinctive from intracellular targets of MTX, 5-FU, or CDDP, coadministration of PAN-811 with any of these may also have a synergistic effect on suppression of cancer cell growth.

Neuronal Cell Culture.
Mixed cortical and striatal neurons from embryonic day 17 male Sprague-Dawley rats (tissue obtained from NIH) were seeded into poly-D-lysine coated 96-well plates at density of 50,000 cells/well and initially cultured at 37 ∘ C, 5% CO 2 , in neurobasal medium (NB) with B27 supplement (Invitrogen) containing full strength of AOs to obtain highly enriched (95%) neurons [26]. Since AOs, including vitamin E, vitamin E acetate, superoxide dismutase (SOD), catalase (CAT), and GSH, are additives to culture medium, reduction of AO concentration in culture medium provides an approach to determine the level of OS involvement in a neurotoxic process. In our study, the culture medium was replaced at a 50% ratio with NB plus B27 minus AOs twice at days 7 and 9 to set AO concentrations as 50% and 25%, respectively. At 16 days in vitro (d.i.v.), a fraction of the culture medium was harvested for lactate dehydrogenase (LDH) assay, and then AO concentration was reduced to 12.5% or 17.5%, and cultured for a further 5 hours prior to ending the experiment.

Cancer Cell Culture.
The mouse liver cancer cell line BNLT3 (gift of Dr. Jack Wands, Brown University) and the human lung cancer cell line H460 (ATCC) were seeded into 96-well plates at a density of 4,000 cells/well and cultured at 37 ∘ C, 5% CO 2 , in DMEM (11965, Gibco) supplemented with 10% fetal bovine serum, 20 mM HEPES, 1 mM sodium pyruvate, and 24 ng/mL gentamycin (all reagents came from Gibco).

Quantitative Assays and Morphological Assessment.
Cell membrane integrity and mitochondrial function of either neurons or cancer cells were measured with LDH and 3-(4,5dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4sulfophenyl)-2H-tetrazolium [MTS] analyses, respectively. The latter has been used to quantify cell survival. For the LDH assay, a mixture of a 35 L aliquot of culture supernatant and 17.5 L of Mixed Substrate, Enzyme and Dye Solutions (Sigma) was incubated at room temperature (RT) for 30 minutes. For the MTS assay, 10 L of MTS reagent (Promega) was added to a culture well containing neurons in 50 L of medium. The preparations were incubated at 37 ∘ C for 2 hours. The preparations for both assays were then spectrophotometrically measured at 490 nm using a 96-well plate reader (Mode 550, Bio-Rad). Neuronal cell death was morphologically determined based on the integrity of the cell soma and continuity of neuronal processes. The change in number of cancer cells was judged directly by cell density. Cells were photographed under an inverted phase contrast microscope (IX 70, Olympus) using 10x or 20x objective.

ROS Examination.
Neurons were incubated in 15 M dihydrorhodamine 123 (DHR123, Molecular Probes) for 30 min at 37 ∘ C to determine intramitochondrial ROS levels. Fluorescence was photographed by using a fluorescent microscope and quantified by excitation at 485 nm and emission at 520 nm using a 96-well plate reader (Model 550, Bio-Rad).

PAN-811 Shows No Antagonistic Effect on MTX-, 5-FU-, or CDDP-Induced Cytotoxicity in BNLT3 Cells.
To understand whether PAN-811 could interfere with anticancer efficacy of tested anticancer drugs, the mouse liver cancer cell line BNLT3 was cotreated with each anticancer drug at the concentrations used for elicitation of neurotoxicity and 10 M PAN-811, the highest concentration used for neuronal protection in these experiments. A 3-day insult with 100 M MTX severely reduced the cancer cell number (Figure 4(a)). In the culture treated with 10 M PAN-811 alone or cotreated with 100 M MTX and 10 M PAN-811, cell density was also much lower than that in no-insult control. Quantitatively, MTX at 100 M reduced MTS reading by 85% (Figure 4(d)), while PAN-811 at 10 M reduced MTS reading to the same level as MTX. A cotreatment with both did not cause any further reduction in MTS reading when comparing with MTX alone.
Similarly, 5-FU at 25 M significantly reduced the cell density of the cancer cells, and a cotreatment with both 25 M 5-FU and 10 M PAN-811 significantly decreased the cell number as well (Figure 4(b)). Quantitatively, 5-FU at 25 M reduced MTS reading by 84%, which was less efficient than 10 M PAN-811 group (Figure 4(e)). A cotreatment with both caused a further reduction in MTS reading when comparing with 5-FU alone. No synergistic effect between 5-FU and PAN-811 could be detected.
An insult with 3.5 M CDDP also caused a decrease in the cell density (Figure 4 alone, despite showing much lower reading than CDDP alone ( < 0.01).
In general, PAN-811 did not show any inhibition in the effect of MTX, 5-FU, or CDDP on BNLT3 cells, neither did it demonstrate any synergistic effect with each tested anticancer drug on BNLT3 cell growth.

Discussion
The fate of neurons under an OS condition is dependent on the balance between production of ROS and strength of AO defense systems in vivo. Enzymatic AOs, such as SOD and CAT, and nonenzymatic AOs, exemplified with vitamin E and GSH, are both involved in the defenses [23]. Loss of the balance, under condition such as chemotherapy, can elicit cytotoxicity and organ toxicity in experimental animals [24,36] and in humans [9,12,13]. Administration of anticancer drugs is accompanied by not only an increase in ROS level, but also a decrease in antioxidative enzymes [37,38]. In comparison with the in vivo studies, it is rare to find an in vitro study that examines the direct effects of anticancer drug on neurons in an enriched neuronal culture system. The presence of AOs in the culture medium may sufficiently block the effect of an anticancer drug and therefore the system is not suitable for examining ROS-mediated neurotoxicity and may not reflect the real conditions under chemotherapy in animals and humans. To mimic in vivo conditions under chemotherapy, we reduced AO concentrations in a double-diluted manner to a final AO concentration of 12.5%. It was observed that neurotoxicity of MTX, 5-FU, or CDDP occurred only when neurons were bathed in low AO-containing medium.
Cell membranes seemed to be more fragile to these anticancer drugs under these conditions. When the AO content was reduced to 25%, MTX, 5-FU, and CDDP all resulted in robust LDH release, but neither notable morphological changes nor MTS reading differences in anticancer drug-insulted groups were detected. Only when the AO content was further reduced to 12.5% was there observation of morphological cell death in 5-FU-or CDDP-insulted groups and corresponding 33% and 66% reductions in MTS readings, respectively. These data, together with the phenomenon where coadministration of MTX and 5-FU resulted in a significant intramitochondrial ROS increase, indicate a key role of OS in mediation of the in vitro neurotoxicity. Our study demonstrated that under low AO conditions MTX insult only resulted in membrane leakage but did not show significant detrimental effects on cell viability of neurons in our neuron-enriched culture. This is identical to the previous findings that excitatory neurotoxicity marks MTX-mediated cell death, which only occurs in the presence of glial cells, and is protected by N-methyl-D-aspartate receptor antagonists MK-801 and memantine [39].
PAN-811 can suppress neurotoxicity of all tested anticancer drugs in the present study. Under 12.5% AO condition, PAN-811 dose-dependently blocked MTX-, 5-FU-, or CDDPinduced membrane leakage. In addition, PAN-811 at 10 M fully inhibited 5-FU-induced MTS reduction and elevated MTS reading by 48% for CDDP-insulted neurons under 12.5% AO condition. Furthermore, neurons that were treated with PAN-811 looked to have a healthy appearance even when they were insulted with MTX, 5-FU, or CDDP. Our previous studies have demonstrated that, besides inhibiting excitatory neurotoxicity, PAN-811 can protect neurons from cell death under different OS-involved conditions, such as hypoxia [22], and hydrogen peroxide insult [14,15]. In a cellfree and metal-free system, PAN-811 demonstrated an activity in direct scavenging of stable radical diphenylpicrylhydrazyl (DPPH) [22]. Taken together, the neuroprotection provided by PAN-811 in the anticancer drug-insulted condition is most likely due to its activity in inhibition of intracellular ROS accumulation.
In this study, the blockage of oxidative damage by PAN-811 was shown by not only its neuroprotective effect but also its inhibitory role in anticancer drug-induced membrane leakage. Our results showed that each of MTX, 5-FU, and CDDP can induce membrane leakage of cancer cell H460. Theoretically, ROS can be produced in plasma membrane and other cell compartments [40]. Free radicals can pass freely through cellular and nuclear membranes and oxidize biomacromolecules, including lipids. Lipid peroxidation caused by ROS leads to membrane leakage [41]. Efficient inhibition of membrane leakage of H460 cells by PAN-811 indicates its role in suppression of ROS signal not only in neurons but also in other cell types.
Our result demonstrated that PAN-811 did not suppress anticancer efficacy of anticancer drugs MTX, 5-FU, and CDDP despite suppressing anticancer drug-induced membrane leakage. This indicates that the anticancer activity of the tested anticancer drugs does not rely on intramitochondrial ROS accumulation they induced, which provides a basis for using PAN-811 as a neuroprotectant in chemotherapy. In addition, PAN-811 manifested a synergistic effect with 5-FU or CDDP on suppression of cancer cell growth. Both MTX and 5-FU are antimetabolites or antifolate drugs. MTX inhibits DNA synthesis by competitively binding to dihydrofolate reductase, an enzyme that converts dihydrofolate into tetrahydrofolate [42]. 5-FU acts predominantly as a thymidylate synthase inhibitor and suppresses synthesis of the pyrimidine thymidine, which is a nucleoside required for DNA replication [43]. CDDP is an alkylating agent. It binds to and causes cross-linking of DNA, which ultimately triggers apoptosis [44]. PAN-811 is an anticancer drug by itself with a different intracellular target from MTX, 5-FU, and CDDP. PAN-811 divalently chelates the ferrous ions of ribonucleotide reductase and blocks its bioactivity in conversion of ribonucleotides to deoxynucleotides and therefore inhibits DNA synthesis [45]. The synergistic effect by cotreatment with PAN-811 and 5-FU or CDDP may be due to affecting more than one intracellular target. The synergistic effect by cotreatment with PAN-811 may provide an opportunity in reduction of dose usage of 5-FU or CDDP. In this way, the neurotoxicity of 5-FU or CDDP could be further reduced while retaining equal strength of anticancer efficacy. PAN-811 can be a potential neuroprotective drug for chemotherapyinduced cognitive impairment due to its in vitro inhibition of MTX-, 5-FU-, or CDDP-induced neurotoxicity. A pharmacodynamic study for the effect of PAN-811 on cognitive functions will be carried out in a chemobrain animal model in near future.

Conflict of Interests
Zhi-Gang Jiang, Steven A. Fuller, and Hossein A. Ghanbari are employees of and own stock in Panacea Pharmaceuticals, Inc.