Anti-glioblastoma Effects of Structural Variants of Benzoylphenoxyacetamide (BPA): II. Synthesis Strategies for Phenolic Variants of BPAs With Potential for Blood Brain Barrier Penetration.

Glioblastomas are the most aggressive brain tumors for which therapeutic options are limited. Current therapies against glioblastoma include surgical resection, followed by radiotherapy plus concomitant and maintenance with temozolomide (TMZ), however, these standard therapies are often ineffective, and average survival time for glioblastoma patients is between 12 and 18 months. We have previously reported a strong antiglioblastoma activity of several metabolic compounds, which were synthetized based the chemical structure of a common lipid-lowering drug, fenofibrate, and share a general molecular skeleton of benzoylphenoxyacetamide (BPA). Extensive computational analyses of phenol and naphthol moieties added to the BPA skeleton were performed in this study with the objective of selecting new BPA variants for subsequent compound preparation and anti-glioblastoma testing. Initially, 81 structural variations were considered and their physical properties such as solubility (logS), blood-brain partitioning (logBB), and probability of entering the CNS calculated by the Central Nervous System – Multiparameter Optimization (MPO-CNS) algorithm were evaluated. From this initial list, 18 compounds were further evaluated for anti-glioblastoma activity in vitro . Nine compounds demonstrated desirable glioblastoma cell toxicity in cell culture, and two of them, HR51, and HR59 demonstrated significantly improved capability of crossing the model blood-brain-barrier (BBB) composed of endothelial cells, astrocytes and pericytes.


INTRODUCTION
Glioblastomas are the most aggressive brain tumors for which therapeutic options are very limited 1,2 . Current standard of care therapies include maximal surgical resection, followed by radiotherapy plus concomitant and maintenance with temozolomide (TMZ), however, these standard therapies are often ineffective, contributing to the dismal glioblastoma patient survival time of 12-18 months 3 . Multiple genetic and epigenetic abnormalities have been found in glioblastomas, among which p53, EGFR, PTEN, and IDH mutations are the most common [4][5][6] . In spite of these validated therapeutic targets, molecular, gene-therapy, and immunotherapy approaches are still ineffective 7,8 .
Therefore, new and more effective therapies for glioblastoma patients are desperately needed.
There are several reasons why it is difficult to treat glioblastoma. First, glioblastomas are characterized by many dysregulated pathways that cannot be blocked simultaneously with a single therapy 9 ; Second, glioblastomas are highly infiltrating and create heterogenous tumors that are very difficult to be removed by surgery without compromising the function of the surrounding brain areas 10 ; Third, early diagnosis of glioblastoma is rare, therefore, large highly infiltrating and vascularized tumors are often already present at the time of diagnosis 11 . Fourth, the optimization of clinical protocols for glioblastoma treatment requires the use of a reliable preclinical model/s. Unfortunately, commonly used rodent syngeneic and xenograft models have one major problem -the experimental tumors are typically ~10 3 -10 4 smaller than human tumors, and therefore, drug delivery, drug retention, and effective tissue penetration by the drug, cannot be tested in a reliable manner in small animal models 12 ; Fifth, the blood brain barrier (BBB) prevents the majority of anticancer drugs from reaching the tumor site, and current methods that enhance the BBB penetration are not effective for glioblastoma patients 13 .
One of the drugs that readily crosses the BBB is temozolomide (TMZ). Upon oral administration, TMZ maximum plasma concentration can be reached in about one hour, and the elimination half-life is approximately 2.1 hours. Importantly, penetration efficiency of TMZ into the CNS is experimentally estimated to be about 20% of the plasma levels 14 .
Applying this estimate to calculate the logBB (Brain-Blood Distribution) for TMZ, this equation produces a value of -0.7, which indicates sufficient capability of the compound to cross the BBB 15 . In spite of these positive features, TMZ-treated glioblastoma patients develop TMZ-resistance and recurrent tumors are practically incurable 16 . There are also several studies of the use of TMZ in combination with other drugs, which show beneficial therapeutic effects 17 18 . One interesting example is a combination of TMZ with lipid lowering drugs, including statins 19 . In addition, another class of lipid-lowering compounds, fibrates, have also attracted attention as a possible anticancer drugs [20][21][22][23] .
We have previously reported that 50µM fenofibrate (FF) has a strong anti-glioblastoma activity in cell culture, and in glioblastoma mouse models following intratumoral injection 24 (Scheme 1). However, FF does not cross the BBB, and is quickly processed by the blood and tissue esterases to form fenofibric acid (FFA), which is no longer effective in triggering tumor cell death 24,25 .
We have previously made several chemical modifications to the FF molecular skeleton, to address the FF low stability in human blood, low water solubility, and inability of penetrating the BBB. Indeed, one of the initial compounds, PP1, demonstrated improved water solubility and stability in human blood. In addition, PP1 was capable of triggering extensive glioblastoma cell death in vitro at concentrations over 4-fold lower than FF 26 (Scheme 1). To further improve anti-glioblastoma efficacy, we created other FF derivatives, which share the benzoyl-phenoxy-acetamide (BPA) molecular skeleton, and decided to test the addition of phenol and naphthol residues to the BPA structure due to the potential anti-cancer effects of these moieties [27][28][29][30] . As a result, 18 new compounds were generated and were analyzed during this study.

RESULTS and DISCUSSON
Overall Chemical Design: In our previous studies we have explored the importance of a basic BPA skeleton 31 , and concluded that BPA could serve as a "pharmacophore", necessary to retain anti-glioblastoma activity 32,33 . The amide part of the BPA skeleton can be specifically modelled to obtain a more desirable anti-tumor activity. This includes, among other properties, chemical and physical parameters (described below) that contribute to the increased BBB penetration, and possibly drug retention within the tumor tissue. In this regard, we have selected phenol and naphthol residues due to mounting evidence supporting the role of different derivatives of these compounds in health benefits 34 , including anti-cancer activities 28,29 . In this paper, three variants of BPA are discussed: a substituted phenol (Phenolic-BPA)], and two naphtholic BPAs (1-Naphtholic-BPA and 2-Naphtholic-BPA) (Scheme 2) that serve as prototype molecules for further modifications.

Scheme 2.
Phenol region of BPA skeleton selected for modification (circle) in search of the optimal antiglioblastoma drug.
The starting point for the preparation of all phenolic BPAs is fenofibric acid (FFA), and the corresponding aminophenol or aminonaphthol residues (Scheme 3) are added through amide (peptide) coupling reactions 35,36 . As previously reported 31 , due to the steric hindrance of the carboxylic group of FFA, which includes two methyl groups in the alpha position of carboxylic acid, combined with the lower amine nucleophilicity of anilines in DCC-or EDC-coupling, these reactions do not produce acceptable isolated yields. This occurs even with more reactive aminophenols and EDC or DCC, which are stronger nucleophiles compared to nonactivated anilines, which is expected to produce corresponding BPA compounds in acceptable yields 37 . However, we were able to detect only traces of the desirable products with these methods, and instead, decided to convert FFA into the more reactive fenofibrate chloride (FFC), followed by coupling with aminophenols or aminonaphthols (Scheme 3). We have explored several variations of this procedure and finally selected one that is very simple and can be applied to multigram and even multikilogram production scales. In particular, the FFC was prepared fresh and immediately used, for the next step of aminophenol addition (Scheme 3). The most common method of preparation of an acid chloride is by using thionyl chloride. This  As a result of this initial screening, we have identified two lead drug candidates, HR51 and HR59, with phenolic moieties that contain BPA structural skeleton similar to our prototype anti-glioblastoma compound PP1 26 . This is in addition to our recently reported BPA-based compounds (HR28, HR32, HR37, and HR46), which also demonstrated high potential as anti-glioblastoma drugs 31 . Anti-glioblastoma effects of HR51 and HR59 were subsequently confirmed using four different human glioblastoma cell lines, LN229, U-87 MG, U-118 MG, T98G, and the cytotoxicity data were compared to normal human astrocytes (NHA). Results in Fig. 6A demonstrate that all tested glioblastoma cells were partially responsive to 10µM HR51 and HR59, but were almost completely eliminated following 72 hour exposure to 25µM HR51 or HR59. In contrast, these two compounds were significantly less cytotoxic to normal human astrocytes (NHA), indicating that these two new compounds may have low CNS toxicity. In addition, results in Fig. 6B show that IC50 concentrations for HR51 and HR59 are below 10 µM, which is an acceptable therapeutic concentration for clinically relevant anticancer drugs.
We have also tested if the mechanism of action of HR51 and HR59 is similar to our prototype drugs, PP1 and fenofibrate, which have been previously shown to inhibit mitochondrial respiration at the level of Complex 1 of the electron transport chain (ETC) 24,26 . Indeed, results in Fig. 6C confirmed that both HR51 and HR59 inhibit mitochondrial respiration in the magnitude similar to PP1.
In addition to a strong in vitro anti-glioblastoma activity (Fig. 6), HR51 and HR59 have physical properties that may contribute to the improved brain tumor penetration.
Specifically, HR51 and HR59, have a minimal projection area (MPA) of 46.23 Å 2 and 43.73 Å 2 46 , respectively; water solubility (LogS) of -6.61 and -6.11 47 ; and brain to plasma concentration ratio (LogBB) of -0.15 and -0.49 15 , which are all considered as highly promising for compounds suspected of being capable of penetrating the brain tumor tissue.
Importantly, HR51 and HR59 can also cross the triple-coculture model of the blood brain barrier (BBB), which consists of astrocytes, pericytes and epithelial cells cultured on 24-well (3µm pores) transwell membranes (Fig. 7A Subsequently, aliquots of media from the corresponding inserts and from the wells were collected for HPLC measurements and to calculate BBB permeability (P=VA⋅CA/(t⋅S⋅CL) 52 . Results in Fig. 7C show that HR51, HR59 and caffeine cross the in vitro BBB at levels 4.4-fold, 3.5-fold and 22.0-fold higher compared to our internal negative control, fenofibrate, which although has a similar molecular weight and structure to the tested HR compounds (Fig. 1), its ability of crossing natural BBB is very low 25 .     All starting materials were reagent grade and purchased from Sigma-Aldrich, ArkPharm, and TCI America.   dichloromethane (50 ml) and water (50 ml). This final mixture was sonicated at room temperature until all solid was dissolved. From this bilayer solution, the water layer was discarded, and the dichloromethane layer was washed with 5% Na2CO3 (3x50 ml), water (50 ml), 5% HCl (3x50 ml), water (50 ml) and dried over anhydrous Na2CO3. After solvent evaporation, the final product was purified by crystallization from dichloromethane (~3ml) and hexane (20 ml and air-dried for 2 hrs. Next, the inserts were coated with 2µg/cm 2 poly-L-lysine (ScienCell) for 1 hr at 37°C, then washed twice with sterile H20 and air-dried for 2 hrs.
1.57x10 5 primary human astrocytes (ScienCell) and 3.125x10 4 primary human pericytes (ScienCell) were resuspended in 25µl of astrocyte medium and pericyte medium (ScienCell), respectively, then combined in a 1:1 ratio for 50µl total volume. Dried, coated inserts were turned upside down such that the basolateral surface was exposed at the top, and 50µl of the cell mixture was added to the membrane, covered with the plate lid, and incubated for 2 hrs at 37°C to allow cell adherence. Any medium remaining on top of the membrane was carefully removed before returning inserts to their upright position with the apical surface facing upward as they were placed in a 24-well plate containing 500µl per well of astrocyte/pericyte medium (1:1

Evaluation of metabolic parameters
Metabolic responses of human glioblastoma cells were evaluated with Extracellular Flux Analyzer XFe96 (Agilent Technologies). During the day prior to each assay the cells were plated at 2x10 4 cells/well in Agilent Seahorse 96-well XF cell culture microplates with growth supporting media and incubated overnight. At the time of measurement, growth media were replaced with serum-free XF assay medium (Seahorse XF Base Medium supplemented with 1 mM sodium pyruvate, 2 mM glutamine, and 5.5 mM glucose) and cartridges equipped with oxygen-sensitive and pH-sensitive fluorescent probes were placed above the cells. The oxygen consumption rate (OCR; indicative of mitochondrial respiration) was evaluated after injecting HR compounds or PP1 (all used at 25 μM), or DMSO (0.1%; vehicle control). These initial injections were followed by sequential injections of metabolic toxins to execute mitochondrial stress assay: oligomycin (inhibitor of ATP synthase; 0.5 μM); carbonylcyanide-p-