Biological Evaluation of Novel Pyridine derived Compounds - Exploring Cytotoxicity and Gene Targeting Activities

Biological Evaluation of Novel Pyridine derived

Compounds - Exploring Cytotoxicity and Gene

Targeting Activities

2

Biological Evaluation of Novel Pyridine derived

Compounds - Exploring Cytotoxicity and Gene

Targeting Activities

Author: Mike Filius

Email address: mike_filius@hotmail.com Student nr: 2048683

University:

Avans University of Applied Sciences

School of Life science and Environmental Technologies Biology and medical laboratory research

Lovensdijkstraat 61-63 4818 AJ Breda Mentor: Arjen Bakker

Email address: ahf.bakker@avans.nl Supervisor: Dr. João Carvalho

Email address: "J.F. dos Santos Carvalho" <j.dossantoscarvalho@erasmusmc.nl> Internship Institute

Department of Genetics

Erasmus University Medical Center 3000 CA Rotterdam The Netherlands

Table of Content

Table of Content  ...  3  

Abstract  ...  4  

Chapter 1 Introduction  ...  5  

1.1  DNA  Damage  ...  5  

1.2  Non-­‐Homologous  End  Joining  ...  6  

1.3  Homologous  Recombination  ...  8  

1.4  Gene  Targeting  ...  11  

1.5  hERG  channel  ...  13  

1.6  hERG  channel  and  Cancer  ...  13  

Chapter 2 Materials and Methods  ...  15  

2.1  Transformation  of  Bacteria  ...  15  

2.2 Large Scale Plasmid Preparation (Maxi-prep)  ...  15  

2.3 Agarose gel electrophoresis  ...  15  

2.4 Restriction Endonuclease Digestion of DNA  ...  16  

2.5 Cell Culture  ...  16  

2.6 Cytotoxic assay  ...  16  

2.7  Gene  Targeting  Assay  (Figure  3)  ...  17  

2.8  Chemistry  ...  18  

Chapter 3 Toxicity of Pyridine Derivatives in Mouse Embryonic Stem Cells  ...  21  

Results  and  Discussion  ...  21  

Chapter 4  Potential Anti-Cancer Activity of Positively Charged Pyridines  ...  26  

Results  and  Discussion  ...  26  

Chapter 5 Positively Charged Pyridines: Correlation between Toxicity and hERG Channel Affinity  ...  35  

Results  and  Discussion  ...  35  

Chapter 6 Neutral Pyridines Cytotoxicity and Gene Targeting effect  ...  39  

Results  and  Discussion  ...  39  

4

Abstract

Over the last decades, medicinal chemists have been searching for compounds that could specifically target cancer cells and not healthy cells. The human ether-a-go-go-related gene K+ channel (hERG) has recently been proposed as an anticancer target [1]. Structural activity relationships of this ion channel allowed the identification of positively charged pyridines that could effectively interact with the hERG channel [2]. In the present study, a library of positively charged pyridine derivatives was assed for cytotoxicity in leukemia cell line (K562) and an immortalized fibroblast cell line (C5RO). Structural modifications of the pyridine moiety, side chain and peripheral aromatic moieties were evaluated. Most positively charged pyridines were cytotoxic and only some showed great selectivity toward cancer cells, K562 (leukemia cells) being particularly sensitive to compounds 8, 10b, 10p, 12a, 12i and 12k (Selectivity Index >6). The structural requirements to induce selective toxicity are discussed to shed light on the development of new anticancer drugs.

The ability to achieve site-specific manipulation of the mammalian genome has widespread implications for basic and applied research. Gene targeting is a process in which a DNA molecule introduced into a cell replaces the chromosomal segment by homologous recombination, and thus presents a precise way to manipulate the genome. Recent results have demonstrated that zinc fingers or other mega-nucleases can stimulate gene targeting up to 2-fold. However, inducing double stranded breaks in the genome can lead to chromosome loss, chromosomal rearrangements, apoptosis, or carcinogenesis. In the present study we demonstrate that chemical modulation is a valuable alternative of enhancing gene targeting. Our results demonstrated that neutral pyridines are significantly less toxic then positively charged analogous. We describe the structural requirements to enhance gene targeting in mouse embryonic stem cells, allowing the identification of a non toxic-neutral pyridine 13c that is capable of stimulating gene targeting up to 2,4 fold proving that chemical modulation is a valuable alternative of stimulating gene targeting.

Chapter 1

Introduction

Double Stranded Breaks (DSBs) are the most critical cellular damage. Failure to repair DSBs or their miss-repair can result in chromosome loss, chromosomal rearrangements, apoptosis, or carcinogenesis [3]. DSBs can be generated by exogenous sources such as ionizing radiation (IR), by intermediary metabolic products of Streptomycetes and topoisomerase inhibitors [4, 5]. Cells exposed to IR, present a wide range of DNA damage, including DSBs. Intermediary metabolic products of Streptomycetes, including bleomicyn, mitomycin and related compounds, have been successfully used in anti-tumour therapy. They directly induce DSBs by attacking specific carbons in deoxyribose, leaving non-standard end-groups[3]. The third class of DNA-based anti-tumour therapeutics is represented by topoisomerase inhibitors [4]. Topoisomerases are enzymes that open and close strands of DNA: type I topoisomerases open/close one strand, inducing in this way temporary single stranded breaks, whereas type II topoisomerases open/close both strands at a time, therefore producing intermediate DSB [6]. Camptothecin and etoposide are examples of inhibitors of topoisomerase I and II, respectively leading to DNA damage.

If DNA damage is recognized by cellular machinery several responses may occur to prevent replication in the presence of genetic erors. At the cellular level, checkpoints can be activated to arrest the cell cycle; transcription can be up regulated to compensate for the damage, or the cell can go into apoptosis. In this way, the damage can be repaired at the DNA level enabling the cell to replicate as planned. Complex pathways involving numerous molecules have evolved to perform such repair. Because of the importance of maintaining genomic stability in the prevention of carcinogenesis, genes coding for DNA repair proteins have been proposed as candidate cancer-susceptibility genes [7, 8]. There are two major pathways for the repair of DSBs in mammalian cells, homologous recombination (HR) and non-homologous end joining (NHEJ).

6 1.2  Non-­‐Homologous  End  Joining  

After detection of a DSB, non-homologous end joining (NHEJ) repair pathway promotes the ligation of the broken ends directly to each other without the need of a homologous template. Therefore, NHEJ is not error free and often causes insertion or deleterious mutations in the site of the DNA break [9]. In mammalian cells, NHEJ has a dominant role in repairing DSBs in the G1 phase of the cell cycle, whereas homologous recombination is used predominantly in S phase, when the sister chromatid can be used as template for repair. In the NHEJ reaction the binding of the Ku70/80 heterodimer to a DNA end (Figure 1) attracts the catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs) and activates its protein kinase activity [15]. DNA-PKcs can phosphorylate several cellular targeting proteins including the Ku polypeptides and itself [10], regulating the NHEJ reaction. DNA ends covered with the Ku/DNA-PKCS complex can now be joined by the ligase IV/XRCC4 complex. However, in many circumstances the DNA ends are not compatible. For instances, Ionizing radiationcreates a large number of ends that contain damaged bases and/or DNA backbone sugars that need processing before ligation. In this manner, nucleases, DNA polymerases, polynucleotide kinases and other enzymes that help such ends to be ligated by the ligase IV/XRCC4 complex can participate in NHEJ pathway generating mutations such as deletions and insertions.

NHEJ is not only important for repair of DSBs that result from exogenous and endogenous DNA damaging agent, but also for the repair of DSBs generated during V(D)J recombination. This takes place during B- and T-cell differentiation to generate antigen specific receptors [11].

Figure 1. Double-stranded break repair by non-homologous end joining. NHEJ brings

the ends of the broken DNA molecule together by the formation of a synaptic complex, consisting of two DNA ends, two Ku70/80 and two DNA-PKCS molecules. Non-compatible

DNA ends are processed to make ligation feasible followed by repair of the break by the DNA ligase IV/XRCC4 complex. Adapted from literature [10].

8 1.3  Homologous  Recombination  

Homologous Recombination (HR) is an essential biological process involving the exchange of genetic information between two homologous DNA molecules. It promotes genome stability through the accurate repair of deleterious DNA lesions such as DSBs. HR can repair DSBs by using the undamaged sister chromatid as template. Therefore, HR generally results in a more accurate repair of a DSB then NHEJ. In many eukaryotic cells, including mammals, HR is mediated through the so-called RAD52 group of proteins [12], which includes RAD50, RAD51, RAD54 and meiotic recombination enzyme 11(MRE11). A brief summary of the HR mechanism is shown in Figure 2.

DNA-end recognition by RAD52 [13] is the first step (pre-synapsis) in the DSB repair by HR. Subsequently the broken DNA ends are processed into 3’ single stranded DNA overhangs, which is performed by nucleotylic enzymes. RAD51 protein will bind to the resulting single stranded overhang, forming a RAD51-ssDNA filament (one filament consists of six RAD51 molecules and 18 nucleotides per helical turn [14][). The RAD51-ssDNA filament formation is a crucial step in HR, because its aim is to search for the homologous DNA template and promotes strand exchange allowing the free error repair [10]. More precisely, when the filament finds the homologous DNA duplex, RAD51 will form a physical connection between the invading DNA substrate and the homologous template, leading to the generation of heteroduplex DNA intermediate (D-Loop) [14]. The homology search and DNA strand invasion are collectively called synapsis. Finally, during the post-synapsis RAD51 dissociates from dsDNA exposing the 3’-OH end which is used as a primer for DNA synthesis..

Once DNA synthesis is initiated at least three different routes can be adopted (Figure 2B-D). As illustrated in the double-strand break repair model (DSBR) the second end of DSB can be engaged to stabilize the D-loop structure (second-end capture), leading to the generation of a double-Holliday Junction (dHJ, Figure 2B). A

dHJ is then resolved to produce crossover or non-crossover products (Figure 2B) or

dissolved to exclusively generate non-crossover products. Alternatively, the invading strand can be displaced from the D-loop and anneals either with its complementary strand as in gap repair or with the complementary strand associating with the other end of the DSB. This represents the synthesis-dependent strand-annealing model of

HR (SDSA)[15](Figure 2C). SDSA mechanism is preferred over DSBR during

mitosis. During meiosis, crossovers are formed by resolution of dHJs via the DSBR mechanism, while non-crossovers are primarily produced via SDSA mechanism. Finally, it is also possible that the D-loop structure assembles into a replication fork and copy the entire chromosome arm in a process called break-induced replication (BIR)(Figure 2D). All the above pathways require RAD51, with the exception of some

forms of BIR. However, DSBs may also be repaired by pathways independent of RAD51 (Figure 2E and F). One of these pathways is the single-strand annealing

pathway (SSA). In SSA, ssDNA sequences generated during DSB processing contain regions of homology at both sides of DSB and can be annealed and ligated (Figure 2E).. Another RAD51-independent pathway that operates at DSBs is NHEJ, which

10

Figure 2 Models for the repair of DNA double-stranded breaks DNA DSBs are resected to

generate 3’_{-protruding ends followed by formation of the RAD51 filament that invades the homologous} template forming D-loop intermediate structures. (A) After priming DNA synthesis, three pathways can be invoked. In the DSBR pathway, the second end is captured and a dHJ intermediate is formed. (B) Resolution of dHJs can occur in either plane to generate crossover or non-crossover products. Alternatively, dHJs can be dissolved by the action of Sgs1–Top1–Rmi1 complex to generate only non-crossovers. In the SDSA pathway (C), the extended nascent strand is displaced, followed by pairing with the other 3’-single-stranded tail, and DNA synthesis completes repair. Nucleolytic trimming might be also required. Alternatively, when the second end is absent BIR (D) can occur. In this pathway the D-loop intermediate turns into a replication fork capable of both lagging and leading strand synthesis. Two other RAD51-independent recombinational repair pathways are also depicted. In SSA (E), extensive resection can reveal complementary sequences at two repeats, allowing annealing. The 3’-tails are removed nucleolytically and the nicks are ligated. SSA leads to the deletion of one of the repeats and the intervening DNA. Finally, the ends of DSB can be directly ligated resulting in NHEJ. Adapted from literature [14].

1.4  Gene  Targeting  

Gene targeting (GT) relies on HR to mediate precise genomic alterations. It represents the most powerful tool for genetic engineering; however random integration of DNA occurs more frequently than HR, hampering the application of GT. Gene Targeting in mammalian cells is not very efficient, since it is laborious, time consuming and not always successful. As described earlier, inducing a DSB in the DNA will promote cellular repair mechanisms such as NHEJ or HR. In this way GT can be boosted by inducing a precise DSB due to stimulation of HR pathway. In fact, inducing a DSB in the target locus is the best-known approach for the stimulation of GT by HR [16]. This is typically done using endonucleases that recognize and cleave specific sites with considerable specificity and efficiency. Three classes of rare-cutting endonucleases are currently used in gene targeting: the naturally occurring homing endonucleases and two classes of engineered chimeric proteins, the zinc-finger nucleases (ZFNs), and the transcription activator-like (TAL) effector nucleases. The long target site recognized by these enzymes (15–30 bp) provides considerable sequence specificity for DNA cleavage, even in larger genomes. The haploid human genome is 3 × 109 bp in length, so a nuclease that recognizes a 16 bp sequence is predicted in principle to cleave approximately one site, on average (416 = 4.3 × 109). In practice, however, rare-cutting endonucleases are not truly sequence specific. Significant off-target activity has been documented [16], and one of the continuing challenges has been to minimize harmful effects associated with rare-cutting nucleases.

In this study an integration assay for GT is used by targeting the RAD54 locus. This GT assay has been developed and validated in our laboratory. Specifically, an embryonic stem cell line was designed to contain an I-Scel binding site replacing exon 4 (5.1 ES cells)[17]. Only upon targeted integration and consequently replacement by HR of a region of the RAD54 gene with a targeting construct engineered to express a RAD54-GFP protein under the control of the RAD54 promoter (Figure 3) generates green cells (RAD54-GFP), which can be quantified by

12 Determining the frequency of RAD54-GFP cells amongst the neomycin resistant cells gives a quantitative estimation of the GT efficiency.

Figure 3 Rad54 gene targeting assay. Integration of the RAD54-GFP knock-in construct

generates neomycin resistant cells, but only the homologous integration will yield GFP positive cells. The reporter consists a neomycin resistance gene under the control of a constitutive promoter (PGK promoter) and a gfp gene, which will only be expressed when homologous recombination occurs. The #5.1 cell line contains an I-Sce I site instead of exon 4. To determine the percentage of gene targeting efficiency, results were also plotted in a fluorescence (GFP) histogram. In non-transfected ES cells only one peak is visible, while in transfected cells with the RAD54-GFP knock-in construct two peaks are visible, representing non-GFP- and GFP-expressing cells.

I" III" V" IX" X" XI" XVI" XVIII"

I"Sce &

="

1.5  hERG  channel  

Ion channels have been implicated in signaling pathways leading to cell proliferation or apoptosis [18]. Their identification and functional characterization in tumor cells suggest a potential significance in anticancer therapy [19]. Potassium K+ _channels represent the largest family of ion channels involved in cell death and proliferation [20, 21]. The voltage-sensitive human ether-a-go-go-related gene (hERG) potassium channels are widely expressed and their functions differ according to their localization. The hERG channel has a dominant presence in normal human cardiac muscle cells where it is involved in the repolarization phase of the cardiac action potential [22]. Mutation of this channel causes long QT syndrome leading to irregular heartbeat and sudden death. Gain of function mutations in this channel lead to short QT syndrome and sudden infant death [23]. Some studies have demonstrated that HERG is expressed in a variety of cancer cell lines of different histogenesis, but HERG is absent in the healthy cells from which the respective cancer cells are derived [24].

1.6  hERG  channel  and  Cancer  

The functional expression and role of K+ channels in tumor cells has bee deciphered in a broad array of tumors. On the whole, K+ channels can accomplish diverse functions: the type of channel involved and the function which is affected depend of the cell type, the subtype of channel, the stimulus tested, as well the stage of tumor progression [25, 26]. Considering that K+ _{channels are involved in cell cycle} progression, abundant expression of K+ channels is expected to cause loss of proliferative control if endogenous pathways fail to block excessively expressed K+ channels [27]. Interestingly, the promoter region of the hERG gene harbors multiple binding sites for oncoproteins [28]. It was hypothesized that mutations in oncoproteins constitutively activate hERG gene expression, shifting resting membrane potentials of cancerous cells toward more depolarized values and repolarizing them at the end of G1 phase, thereby facilitating cell cycle progression and thus leading to cell proliferation [19].

14 Leukemic cell lines express hERG K+ channels whereas non-cancerous lymphocytes do not exhibit hERG protein. Selective hERG channel blockade reduced proliferation in cancerous cell lines [29]. Cell cycle analysis of FLG29.1 leukemia cells revealed accumulation of cells in G1 phase following treatment with hERG channel blockers [30]. Blocking the hERG channel could effectively reduce cell proliferation in cancerous cell lines and therefore represents a potential anti-cancer target. It was already demonstrated that the positively charged pyridines, used in the present study, could effectively block the hERG channel [2].

Chapter 2

Materials and Methods

Chemically competent DH5α cells were thawed on ice (from -80°C) and 50 µl of cells were added to 10 µg of plasmid DNA. The mixture was incubated on ice for 20 minutes, heat- shocked for 90 seconds at 42°C and then returned to ice for 2 minutes. 450 µl Luria Broth medium was added and the mixture was incubating at 37°C for 1 hour, then pelleting for 30 seconds at 6,500 rpm. 400 µl of the supernatant was removed and the cells were gently resuspended in the remaining LB medium, plated onto L-agar plates with appropriate supplements and incubated at 37°C over night.

2.2 Large Scale Plasmid Preparation (Maxi-prep)

Large-scale plasmid preparations were performed using the QIAprep®Maxiprep Kit (Qiagen) following the manufacturer’s instructions.

2.3 Agarose gel electrophoresis

Analysis of DNA and separation of DNA fragments was performed using agarose gel electrophoresis. Gel slabs were prepared by melting 1% (w/v) agarose in TAE buffer (40 mM Tris, 20 mM NaAc, 10 mM EDTA pH 8.2) containing Etidium Bromide (EtBr) with a final concentration of 0.5 µg/ml. DNA samples (200-400 ng) were loaded into wells by mixing with 20 µl of 5x DNA loading buffer OrangeG. Samples were electrophoresed in TAE buffer at 100V for 1-2 hours. DNA was visualized and analyzed on a Typhoon laser scanner (GE healthcare).

16 2.4 Restriction Endonuclease Digestion of DNA

Typically, 400 µg of plasmid DNA was linearized overnight at 37oC using an appropriate restriction enzyme, at a concentration that allows theoretically complete digestion in about 4 to 5 hours. The buffer conditions for the reaction were used as recommended by the supplier of the enzyme.

2.5 Cell Culture

The mouse embryonic stem cells (mES) were cultured on 0.1% (w/v) gelatin-coated tissue culture dishes with complete ES medium containing 50% (v/v) DMEM, 40% (v/v) BRL-medium (DMEM+10% FCS+ 1% PS, conditioned by confluent Buffalo rat liver cells (BRL) for a week), 10% (v/v) fetal calf serum (FCS), 50 µM β-mercaptoethanol, 1% (v/v) streptomycin-penicillin, 1% (v/v) nonessential amino acids, and 0.001% (v/v) leukemia inhibitory factor (LIF) (106 U/ml). Human leukemia cells (K562) were cultured in culture flasks containing complete DMEM medium (DMEM + 10% FCS + 1% PS). Human fibroblasts (C5RO) were cultured in complete F10 medium (F10 medium + 15% FCS + 1% PS). All cell lines were grown at 37 oC in a humidified 5% CO2 incubator.

2.6 Cytotoxic assay

mES cells (5.1#, Figure 3) and C5RO cells were plated in flat-bottom 96-well plates at cell density of 8 x 103 cells/well and 10 x 103 cells/well, respectively, using 100 µL of appropriate cell culture medium per well. K562 cells were plated in v-shaped 96-well plates at cell density of 7,5 x 103 _{cells/well. Cells were incubated overnight and} then exposed to different compound concentrations dissolved in culture medium containing DMSO in 100 µL. The maximal DMSO concentration in the total medium volume was always lower than 0.6% (including controls). Compound-induced cytotoxicity as quantified after 48 h of exposure by a cell viability colorimetric redox assay using resazurin[31]. With this method the conversion of resazurin (blue-absorbance at 600nm) to its reduced form, resorufin (pink-(blue-absorbance at 560 nm), by viable cells is measured and correlated with cell viability. After 48 h incubation, the medium was removed and replaced by fresh medium, depleted in FCS and SP, but

containing 10% of resazurin solution (0.1 mg/mL). Cells were further incubated for about 2h at 37 oC, and absorbance measured at 560 and 600 nm with a plate spectrophotometer. Cell viability (CV) was calculated for each concentration tested, using the following formula:

D, compound treated cells; C, control cells; R, resazurin control.

The inhibitory concentration in which 50% of the cell die (IC50) was determined for each compound by plotting the CV observed for each concentration versus the logarithmized concentration and fitting a nonlinear regression curve (using GraphPad Prism software). All experiments were performed in triplicates and were repeated at least 3 times.

2.7  Gene  Targeting  Assay  (Figure  3)  

Approximately 6 x 106 mES cells were transfected with 8 µg of Rad54-GFP knock-in construct (JF3, Figure 3) containing a neomycin selectable marker. After transfection, cells were exposed to the compounds for 48 hours at non-cytotoxic concentrations (cell viability ≥70%). Selection with complete ES medium containing additionally 0.2 mg/ml G418 was started after the 48-hour compound treatment. After 8-10 days of G418 selection, plates were trypsinized, resuspended to single-cell suspensions in a 10% FCS-PBS solution and fixed with appropriate volume of 1% paraformaldehyde solution. After permeabilization with the same volume of a 0.1% Triton X-100 solution, cells were analysed in a Becton Dickinson FacsCalibur on a green fluorescence (FL1) versus forward scatter (FSC-H) plot, as describe earlier [17].

18 2.8  Chemistry  

The chemistry and synthesis of the pyridines used in the present study were described earlier [2]. In this section chemical structures of the pyridine derivatives were presented.

Scheme 1: Chemical structure of asymmetric neutral compounds 1a, 1b and asymmetric positively charged compounds 2a, 2b.

Scheme 2: Chemical Structure of Symmetric long side chained and rigid compounds.

N O HCl 1a N O HCl 1b N O 2a N O 2b N O O R R 3a R=CH3 3b R=H N O (n) O (n) 4a n=1 4b n=2 4c n=3 N O O 5 N O O R R 6a R=CH3 6b R=H N O (n) O (n) 7a n=1 7b n=2 7c n=3 N O O 8

Scheme 3: Chemical structures of Pyridines bearing side chains with different degrees of rigidity and different peripheral aromatic substituents.

Scheme 4: Chemical structure of long chained pyridine 4a, 4b and 4c and their oxidized analogous 13a, 13b and 13c.

N R R N _R R _{R N R} N R R OMe Cl Cl Cl O O O O OMe O Cl O Cl O Cl a... b... c... d... e... f... g... h... i... j... k... l... m... n... o... p... 9 10 11 ₁₂ N O (n) O (n) 4a n=1 4b n=2 4c n=3 N O (n) O (n) O 13a n=1 13b n=2 13c n=3

20 Scheme 5: Chemical structures of long chained asymmetric pyridines

Scheme 6: Chemical structure of symmetric pyridines containing a nitrogen at a different position in the pyridine moiety (compounds 20-23). Compounds lacking the central pyridine moiety were displayed at the bottom (compound 24 and 25).

N O N O N O O N 14 15 16 17 N O O _N O O 18 19 N O O N O O N O O N O O 20 21 22 23 24 25

Chapter 3

Toxicity of Pyridine Derivatives in

Mouse Embryonic Stem Cells

Pyridine compounds are defined by the presence of a six-membered heterocyclic ring consisting of five carbon atoms and one nitrogen atom. The arrangement of atoms is similar to benzene except that one of the carbon−hydrogen groups has been replaced by a nitrogen atom. Some important chemical compounds contain the pyridine ring structure, such as vitamins [32] and the anti-tuberculosis drug isoniazid [33]. Positively charged pyridines have a wide range of biological effects, [34, 35] and some of the pyridines derivatives used in the present study, have been shown to effectively block the HERG channel [2].

Preliminary screening allowed the identification of pyridines that could stimulate gene targeting (Figure 4). Comparatively to the control, the positively charged pyridine 12a enhanced gene targeting by about 1,9 fold at 1 µM, a concentration that inhibited around 30% of the cell viability on mES cells. On the other hand, at 50 µM the neutral pyridine 13c stimulated gene targeting by about 2,4 fold, being less toxic than positively charged pyridine 12a. Based on this initial screening positively charged pyridine compounds appear to be more toxic than neutral ones.

22

Figure 4 Toxicity and Gene Targeting of Pyridines 12a and 13c. A) Chemical structure of

compounds, B) Toxicity and C) Gene Targeting Efficiency. The targeting efficiencies of the GT experiments treated with the compounds were compared to the controls (no treatment), this gives a relative GT efficiency Data was represented as mean±SEM for at least three separated experiments after 48 h of exposure.

This observation can be easily realized by determining the cytotoxicity of neutral pyridines 1a and 1b and positively charged analogues 2a and 2b. Neutral compounds 1a and 1b did not affect cell viability of mES cells at 20 µM (Figure 5). Interestingly the position of the nitrogen in the pyridine moiety did not interfere with the biological outcome of these neutral pyridines. However, when positively charged, a decrease in cell viability was determined for compounds 2a and 2b at 20 µM. Again, apparently the position of the positive charge in the pyridine moiety did not seem to interfere much with the biological outcome.

12a [1] _{13c [50]} 0 10 20 30 40 50 60 70 80 90 100 Ce ll V ia b ili ty (%) Compound [µM] B 12a [1] _{13c [50]} 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Re la tiv e G T E ff ic ie n c y Compound [µM] C N O O O N 12a 13c A

Figure 5: Toxicity of Neutral -and Positively Charged Pyridines. A) Chemical structures

of compounds 1a, 1b and 2a, 2b. B) Toxicity, neutral compounds 5a and 5b did not show considerable toxicity towards mES cells at 20 µM (cell viability >70%). The position of the nitrogen in the pyridine moiety did not change toxicity. However, when positively charged a decrease in toxicity was observed (compounds 2a and 2b). Data was represented as mean±SEM for at least three separated experiments after 48 h of exposure.

Having established that some pyridine can effectively reduce cell viability in mES cells, particularly positively charged pyridines, a library of 77 pyridine derivatives was selected for cytotoxic evaluation and structure activity relationship (SAR) studies. Pyridine derivatives were divided into two groups accordingly to their charge. Hetero-aromatic compounds are generally lipophilic and thus not easily soluble in water, therefore all the compounds were dissolved in DMSO. Solubility testing of the compounds was performed in order to identify the maximal concentration of the compound to be tested in the biological assays using about 0.5% DMSO. Compounds

1a 2a 1b 2b 0 10 20 30 40 50 60 70 80 90 100 110 Ce ll V ia b ili ty ( %) Compound [20µM]

B

A

24 The cytotoxicity of both neutral and positively charged pyridines was evaluated in mES cells. Most neutral pyridines did not present considerable cytotoxicity at 20 µM (cell viability ≥70%)(Figure 6). However, introducing rigidity in the side chain of long-chained neutral pyridines reveals toxicity up to 30% (compounds 5 and 9j). Additionally, the position of the triple bonds in the side chain does not affect toxicity, where triple bonds were introduced close to the pyridines moiety of compound 9j, compound 5 contains triple bonds close to the peripheral aromatic group. Changing the position of the ‘chloro’ electron-withdrawing group could lead to increasing toxicity (compare 9l, 9n and 9o), where the p-chloro derivative 9l was less toxic then its m-and o-chloride analogues 9n and 9o.

Figure 6: Toxicity of Neutral Pyridines. Cell viability assay was performed on the mES

cells at 20 µM, as described in the Material and Methods. Data was represented as mean±SEM for at least three separated experiments after 48 h of exposure.

Consistent with previous observation, when pyridine compounds were methylated and consequently positively charged an increase in toxicity was observed for most of the compounds. In fact, screening for cytotoxicity of positively charged pyridines at 20 µM did not allow good SAR, because most of the compounds kill already too much at such concentration. Therefore, positively charged compounds were screened for cytotoxicity in mES cells at 10 µM (Figure 7).

1a 1b 3a 4b 4c 5 9c 9d 9f 9h 9j 9l 9n 9o 9p11a11b 11i 11j11k 11l11m11o11p13a13b13c 16 18 20 22 24 25 0 10 20 30 40 50 60 70 80 90 100 110 120 130 Compound [20µM] Ce ll V ia b ili ty ( %)

Figure 7: Toxicity of Positively Charged Pyridines. Cell viability assay was performed on

the mES cells, as described in the Material and Methods. Data was represented as mean±SEM for at least three separated experiments after 48 h of exposure

A great variation of toxicity was observed for positively charged pyridine derivatives, interestingly, not all positively charged pyridines reduced cell viability at 10 µM. Introducing rigidity close to the pyridine moiety in short chained pyridines reveals a increase in toxicity, as demonstrated by compounds 10a – 10e. However, when the side chain length was increased in positively charged rigid pyridines, a decrease in toxicity was determined by compounds 10f – 10o. Flexible compound 12 and derivatives revealed an overall increase in toxicity when compared to rigid analogous 10 and derivatives. In order to explore the cytotoxic potential of positively charged pyridine derivatives as anticancer agents some compounds were selected for cytotoxic screening in leukemia cell line (K562) and an immortalized fibroblast cell line (C5RO) as discussed in the next chapter. In contrast, neutral pyridines were significantly less toxic compared to positively charged pyridines and were screened for their ability to stimulate gene targeting at 50 µM.

2a 2b 6a 6b 7a 7b 7c 810a10b10c10d10e 10f10g10h 10i 10j10k 10l10m10n10o10p12a12b12i 12j12k 12l12m12o12p 15 17 19 21 23

0 10 20 30 40 50 60 70 80 90 100 110 120 Compound [10µM] Ce ll V ia b ili ty ( %)

26

Chapter 4

Potential Anti-Cancer Activity of

Positively Charged Pyridines

Results  and  Discussion  

Chronic myeloid leukemia is a cancer of the white blood cells. It is a form of leukemia characterized by the increased and unregulated growth of myeloid cells in the bone marrow and the accumulation of these cells in the blood. Over the last decades, medicinal chemists have been searching for compounds that could specifically target cancer cells and not healthy cells. Imatinib is a great example that designed targeted anticancer therapy works. Like all tyrosine-kinase inhibitors, Imatinib works by inhibiting the BCR-ABL tyrosine kinase activity for cellular transformation. The essential role of BCR-ABL tyrosine kinase activity for cellular transformation provided the rational for targeting this function therapeutically Because the BCR-ABL tyrosine kinase enzyme exists only in cancer cells and not in healthy cells, Imatinib works as a form of targeted therapy —only cancer cells are killed through the drug's action.

Figure 8 Toxicity of Imatinib in K562 and C5RO cells: Dose-dependent effects of Imatinib

on cell viability of leukemia K562 and non-cancer C5RO cell lines. Data shown are mean ± standard error of at least 3 independent experiments.

0 5 10 15 20 25 30 35 40 45 50 55 0 20 40 60 80 100 Concentration (µM) Cell V iability (%) K562 C5RO

Based on the selectivity of Imatinib towards cancer cells, the in vitro anticancer activity of Imatinib was evaluated at different concentrations in human leukemia cells (K562) and human fibroblasts (C5RO). The IC50 values (±SEM) for Imatinib in K562 cells and C5RO cells were 0.69±0.43 µM and 21.11±1.8 µM, respectively (Figure 8). The degree of selectivity a compound shows for leukemia cells can be expressed by its selectivity index (SI) value:

The higher the SI, the more promise a compound holds, due to its selectivity for the K562 cells. However, an SI below 1 indicates that while a compound may possess strong anticancer activity, it may be a general toxin, due to cytotoxicity in normal cells as well. Based on this, the SI data shown in Table 1 indicates that Imatinib exhibits a high degree of cytotoxic selectivity due to IC50 values. Comparison of individual concentrations reveals a decrease in selectivity of Imatinib.

Table 1 Cell viability results of Imatinib in C5RO and K562 cells Affect of Imatinib on

cell viability of leukemia K562 and non-cancer C5RO cell lines at different concentrations. Selectivity Index represents cell viability of C5RO/ cell viability of K562. Data shown are mean ± standard error of at least 3 independent experiments.

K562 cells Selectivity index

1 86,06 ± 1,1 40,47 ± 3,8 2,2 10 86,79 ± 3,1 35,29 ± 1,9 2,5 30 20,74 ± 6,36 21,27 ± 1,3 0,9 IC50 21,11 ± 1,8 0,69 ± 0,4 30,6 Concentration (uM) C5RO cells Cell viability (%) Concentration (uM)

28 Having established that the selectivity index can have great variations between the IC50 and an individual concentration of a compound. Initially, a structurally diverse library of 24 positively charged pyridines were tested for their ability to reduce cell viability at a standard concentration (10 µM) in C5RO and K652 cells. It was already been shown that the positively charged pyridines, used in the present study, could effectively block the HERG channel [2] and therefore have potential anticancer activity. Not all positively charged pyridines could block the HERG channel and therefore it is unlikely that the toxicity, of the compounds used in the present study, is due to their ability to block the HERG channel. For example, compound 2a did not interact with the HERG channel, the derivative 2b bound effectively to the HERG channel (the correlation between toxicity and hERG affinity were discussed later). However, there was no difference in toxicity observed of these two compounds (Table 2) meaning that the position of the nitrogen in the pyridine moiety does not affect the toxicity of these positively charged pyridines.

Table 2 Toxicity of Positively charged pyridines. Affect of positively charged pyridines on

cell viability of leukemia K562 and non-cancer C5RO cell lines at 10 µM. Selectivity Index represents cell viability of C5RO/ cell viability of K562. Data shown are mean ± standard error of at least 3 independent experiments.

Compound C5RO K562 2a 88,02 ± 4,7 41,23 ± 7,3 2,13 2b 86,12 ± 1,9 40,98 ± 3,3 2,10 6a 94,81 ± 2,2 33,38 ± 0,1 2,84 7b 48,29 ± 1,6 35,38 ± 3,2 1,37 7c 8,44 ± 5,6 4,28 ± 0,4 1,97 8 68,81 ± 10,9 4,95 ± 4,9 13,90 Imatinib 10 uM 86,79 ± 3,1 35,29 ± 1,9 2,5 Cell Viability (%)

Structure Selectivity index (SI)

N O O N O O N O O N O N O N O O

Side-chain length (Table 2)

If flexibility is a limiting parameter for toxicity, then increasing side-chain length of the positively charged pyridine 2a results in more flexible molecules. The results of these extended pyridines (6a, 7b and 7c) revealed increased toxicity in both cell lines. The flexibility of the molecules results into greater toxicity to both cell lines, allowing the molecule to get into optimal position on its target of the cell

30

Table 3)

Increasing side-chain rigidity may, therefore, decrease toxicity. Introducing triple bonds close to the aromatic group, such as pyridine 8 resulted in great selectivity towards K562 cells. Interestingly, when compound 8 compared to compound 7c, a significant increase in selectivity towards K562 was observed. Compound 8 is also more potent in selectivity then Imatinib at a concentration of 10 µM. The toxicity is related to the position of the rigidity and the length of the side-chain (compare 10a and 10i with 12a and 12i), introducing triple bonds close to the pyridine moiety in shorter side-chained compounds (two carbon side-chains, compare 10a and 12a), a great increase in toxicity was determined towards C5RO cells, however toxicity remains the same when no triple bond was introduced. Interestingly, introducing triple bonds close to the pyridine moiety in large side-chained compounds did not result in toxicity of the compounds (four carbon side-chain, compare 10i and 12i). Additionally, flexibility of the side chain resulted in selectivity of the compound towards K562 cells. Adding two more carbons to the side-chain allowed recovery of the selectivity index towards K562 cells (compare 12i and 12j).

Table 3 Affect of Flexibility and Peripheral Aromaticity on Toxicity. Affect of Flexibility

and Peripheral Aromaticity on cell viability of leukemia K562 and non-cancer C5RO cell lines at 10 µM. Selectivity Index represents cell viability of C5RO/ cell viability of K562. Data shown are mean ± standard error of at least 3 independent experiments.

Peripheral Aromaticity (

Compound Rest group

C5RO K562 10a 4,49 ± 1,6 15,02 ± 7,5 0,28 10i 96,07 ± 3,7 86,63 ± 3,3 1,11 10p 93,01 ± 4,8 13,42 ± 8,8 6,93 12a 77,85 ± 13,2 19,18 ± 19,1 4,06 12i 64,70 ± 5,8 18,25 ± 5,2 3,55 12j 58,65 ± 8,2 27,32 ± 9,7 2,15 12p 1,24 ± 0,7 0,81 ± 0,8 1,53 Imatinib 10 uM 86,79 ± 3,1 35,29 ± 1,9 2,5

Selectivity index (SI) Cell Viability (%) O O O N

R

R

10

R

R

12

32

Table 3)

Compounds 10p and 12p, aliphatic analogous of compounds 10i and 12i, were screened for toxicity to evaluate the importance of peripheral aromaticity. Interestingly, compounds bearing aliphatic side chains and a triple bond close to the pyridine moiety were found to be ~ 6-fold more selective towards K562 cells (compare 10i and 10p). Interestingly, flexible derivative 12p displayed one of the highest toxicities towards both cell lines in this study. However, introducing triple bonds in aliphatic derivative 10p increased SI but did not affect SI in aromatic side-chained derivatives 10i.

Different Peripheral Aromatic Groups of Rigid Pyridines (Table 4)

Introducing triple bonds and varying in side-chain length can lead to different toxicity and selectivity of the compounds. Therefore, the peripheral aromatic groups were chemically modified to study their toxicity and selectivity.

The short-chained pyridine 10a has a high toxicity towards both cell lines but is more toxic towards C5RO cells then K562 cells (SI = 0,28). Interestingly, removing the p-methyl substituent leads to a significant decrease in toxicity towards C5RO, without affecting toxicity against K562 (compare 10a and 10b). Previously, it was demonstrated for the aliphatic long side-chained pyridine 16p (Table 3) that rididification of the side chain leads to increased selectivity for K562 cells. Toxicity of aliphatic derivative 10c demonstrates that, in contrast to the long-chained pyridine 10p, for short-chained rigid compounds aromaticity is better for selectivity, while aliphaticity is better for general toxicity. However substituents can clearly influence this. For instance, we would expect based only on aromaticity that compound 10f would be as toxic as compound 10c, which is not the case, meaning that although recovery of SI was observed when compared to compound 10a, introducing a chloro substituent also decrease the van der Waal interactions, or maybe hydrogen bonding with other regions of the proteins destabilize the required interaction for potent cytotoxicity. Using the same argument, if aromaticity is really important for Selectivity, then compound 10e, due to the electron donating properties of the methoxy substituent, would be even better than compound 10b, which is not the case, maybe due to the polarity introduced or the possibility of hydrogen bonding (methoxy is a hydrogen bond acceptor). Would be interesting to see if this effect is related to the position of the substituents, for which chloride derivatives are available

Table 4 Toxicity of Rigid and Short-Chained Pyridines Bearing Different Substituents.

Affect of different substituents on cell viability of leukemia K562 and non-cancer C5RO cell lines at 10 µM. Selectivity Index represents cell viability of C5RO/ cell viability of K562. Data shown are mean ± standard error of at least 3 independent experiments

Long-Chained Pyridines (Table 5)

It was already demonstrated that increasing side-chain length of non-rigid positively charged compounds could increase toxicity in C5RO cells (compare 12a, 12i and 12j). However, removing the p-methyl substituent of compound 12i leads to the highest selectivity towards K562 cells shown in this study (compound 12k). Compound 12k also have more potential of anticancer activity then Imatinib, and is the best compound found in this study due to selectivity towards K562 cells. Introducing the p-methoxy substituent leads to decreased activity, while in the short chained pyridines some recover of the SI is obtained, in more extended pyridines there is a decrease in SI. However, comparing 10e and 12l, extending the side chain and removing rigidity result in less toxicity towards normal cells and consequently better SI for compound 18l. But this is not really a fair comparison, because we are changing more than one parameter in the compound. Changing the position of the ‘chloro’ electron-withdrawing group could lead to increasing toxicity (compare 12m

Compound Rest group

C5RO K562 10a 4,49 ± 1,6 15,02 ± 7,5 0,28 10b 87,31 ± 2,7 13,67 ± 1,3 6,39 10c 4,25 ± 2,2 2,16 ± 1,7 1,97 10e 24,16 ± 14,9 27,80 ± 0,4 0,87 10f 99,98 ± 5,9 85,49 ± 12,9 1,17 Imatinib 10 uM 86,79 ± 3,1 35,29 ± 1,9 2,5

Selectivity index (SI) Cell Viability (%) OMe Cl N

R

R

10

34 and 12o), where the p-chloro derivative 12m was less toxic then its o-chloride analogue 12o.

Table 5 Toxicity of Long-Chained Pyridines Bearing Different Substituents. Affect of

different substituents on cell viability of leukemia K562 and non-cancer C5RO cell lines at 10 µM. Selectivity Index represents cell viability of C5RO/ cell viability of K562. Data shown are mean ± standard error of at least 3 independent experiments

Compounds that were potent in selectivity towards K562 cells at 10 µM were selected for generating concentration effect curves in different cell lines and IC50 values were determined. Concentration effect experiments are still in progress and therefore did not allow the presentation of pyridines 8, 10b and 12k, which is unfortunate because positively charged pyridine 12k demonstrate the highest selectivity towards K562 cells in this study. As shown in Figure 9, the concentration-effect curves of pyridines 10p, 12a and 12i were obtained and allowed the identification of two potential anti-cancer agents. Additionally, toxicity of compound 10p reveals no toxic affect on C5RO cells (IC50 >50 µM). However, when compound 10p was evaluated for toxicity in K562 and mES cells, a significant increase in toxicity was observed, IC50 values of 9,78 ± 1,4 µM and 3,95 ± 0,7 µM. This allowed the determination of a Selectivity Index based on IC50 values for compound 10p of 5,1.

Compound Rest group

C5RO K562 12a 77,85 ± 13,15 19,18 ± 19,1 4,06 12i 64,70 ± 5,9 18,25 ± 5,2 3,55 12j 58,65 ± 8,2 27,32 ± 9,7 2,15 12k 33,36 ± 3,4 0,25 ± 0,2 133,42 12l 91,65 ± 7,9 33,51 ± 14,7 2,74 12m 87,30 ± 9,2 61,76 ± 12,5 1,41 12o 32,01 ± 11,4 14,02 ± 9,7 2,28 12p 1,24 ± 0,7 0,81 ± 0,8 1,53 Imatinib 10 uM 86,79 ± 3,1 35,29 ± 1,9 2,5 Cell Viability (%)

Selectivity index (SI)

O O O OMe O Cl O O Cl N

R

R

12

However, selectivity of compound 10p is not greater towards K562 cells then selectivity of Imatinib, it needs to be noted that toxicity of compound 10p does not show toxic effect towards C5RO cells. In contrast, our toxicity results of Imatinib (Figure 8) show that Imatinib is more toxic towards C5RO cells then K562 at 30 µM.

Figure 9 Dose-Depended Toxicity Curves of compounds 10p, 12a and 12i. Concentration

effect curves were performed in three different cell lines; human immortalized fibroblasts (C5RO, blue line), human myeloid leukemia cells (K562, red line) and mouse embryonic stem cells (#5.1 ES, orange line). Data shown are mean ± standard error of at least 3 independent experiments.

Increasing concentration of compound 12a more than 10 µM did not allow any increased toxicity towards leukemia cells. It needs to be noted that the IC50 value for compound 12a is not determined yet in C5RO cells. In contrast to leukemia cells, increasing the concentration of compound 12a in both C5RO and mES cells leads to increased toxicity. Therefore, compound 12a has not the same potential of being an

0 5 10 15 20 25 30 35 40 45 50 0 20 40 60 80 100 120 Concentration (µM) Cell V iability (%) K562 C5RO #5.1 ES Compound 12a 0 5 10 15 20 25 30 35 40 45 50 0 20 40 60 80 100 120 Concentration [µM] Cell V iability (%) K562 C5RO #5.1 ES Compound 12i 0 5 10 15 20 25 30 35 40 45 50 0 20 40 60 80 100 120 Concentration (µM) Cell V iability (%) K562 C5RO # 5.1 cells Compound 10p N N N O O

36 of compound 12i allowed the identification of IC50 in all the cell lines used. Allowing the identification of a compound that can also be highly toxic towards C5RO cells.

Chapter 5

Positively Charged Pyridines:

Correlation between Toxicity and

hERG Channel Affinity

Results  and  Discussion  

Ion channel encoding genes form a large superfamily, which comprises numerous different members.[36] With 78 members, potassium channels make up about half of this superfamily. Potassium K+ channels represent the largest family of ion channels involved in cell death and proliferation.[20, 21] The voltage-sensitive human ether-a-go-go-related gene (hERG) potassium channels are widely expressed and their functions differ according to their localization. A recent study revealed that hERG is expressed in a number of cancer cell lines of different histogenesis but absent from the normal cells from which the respective cancer cells are derived. [24] Based upon the notion that ion channels and transporters control many cancer hallmarks in different types of human cancer, blocking the hERG channel activity has been suggested as a new target to be explored for selective anticancer therapy[19]. It was already demonstrated that the positively charged pyridines, used in the present study, could effectively block the HERG channel [2]. To investigate the correlation between toxicity of the positively charged pyridines and hERG channel affinity, results of the positively charged pyridines of cytotoxicity and their ability bind to the HERG channel were compared.

Cytotoxicity of the positively charged pyridines was evaluated at 10 µM in myeloid leukemia cells (K562), which have been shown to express the hERG channel.[30] As publish elsewhere and regarding the hERG channel activity, [2] compounds were initially screened at 10 µM for the ability to displace radioligand [3_{H]astemizole from HERG K}+ _{channels in HEK}

293 cells. Those capable of displacing more then 50% of the [3_{H]astemizole were then selected for further testing in order to}

generate concentration-effect curves allowing the determination of the IC50values. In this way it was showed that the hERG channel affinity is highly dependent on the positive charge of the pyridine moiety, while here a same sort of observation is also verified in the context of cytotoxicity. This observation raises the question about the correlation between toxicity and hERG channel affinity. If toxicity of the positively charged pyridines is due to hERG channel affinity, this could be demonstrated by comparing potent hERG channel blockers and their toxicity. It was previously demonstrated; using a radioligand-binding assay that flexible positively charged pyridines can effectively fit in the hERG channel. To investigate if permanently positively charged pyridine derivatives can be a strategy to overcome hERG liability, three compounds were selected for patch clump studies on intact cells (Table 6). All the permanently positively charged pyridines (compounds 7c, 8 and 12i) showed significant loss of hERG activity when compared to radioligand binding assay. When results of patch clamp assay were compared to toxicity results of the compounds towards K562, a correlation of hERG affinity and toxicity was suggested. Additionally, both toxicity and hERG affinity were similar of most potent compounds 7c and 8. A decrease in toxicity was observed for the less potent hERG blocker 12i.

Table 6 Comparison of hERG Affinity of Selected Compounds As determined by the Radioligand Binding Assay on Cell Membranes and the Patch Clamp assay on intact cells And Toxicity of the Compounds

IC50%(nM)%of% PCA Compound Structure 7c 194 8 214 12i 962 24 ± 12 nM 40 ± 4 nM 86%±%31%nM IC50 ±SEM (nM) in K562 cells of RBA [29] 4,95 ± 4,9 18,25%±%5,2 4,28 ± 0,4 Cell viability (%) N O O N O O N O O

38 In contrast to this observation, some positive charged pyridines were found to have less hERG affinity in radioligand-binding assay but could significantly be more potent in toxicity towards K562 cells. The positively charged pyridine 6a did not interact with the hERG channel (Table 7), while the positively charged pyridine 6b bounds effectively to the channel. However in the cytotoxicity assay both compounds showed the same potency against the leukemia cells. Results of both toxicity and radioligand binding of compounds 10c and 10f, suggests that there is no direct correlation between hERG affinity and toxicity in K562 cells. Additionally, compound 10c showed to be about 4-fold less active in hERG affinity then derivative 10f, however, 10c was about 40-fold more potent in toxicity then 10f. Comparing two of the most potent hERG blockers 18m and 18p (IC50 ~ 15 nM) with their toxic profile in K562 cells indicates once more that there is no correlation since both compounds present significantly different cell viabilities at 10 µM.

Table 7: Toxicity and HERG Channel Affinities of Positively Charged Pyridines

Compound Structure 2a 21 ± 7 % 2b 346 ± 125 nM 10c 4340&±&1264&nM 10f 1159&±&368&nM 12m 14&±&4&nM 12p 15&±&5&nM 85,49&±&12,9 2,16&±&1,7 0,81&±&0,8 Displacement of radioligand at 10 µM ±SEM in K562 cells 10 µM (%) or IC50 ±SEM (nM) [29]

40,98 ± 3,3 61,76&±&12,5 41,23 ± 7,3 Cell viability (%) at N O N O N O O Cl Cl N N N Cl Cl

Comparing data from patch clamp assay and toxicity suggests a direct correlation between hERG affinity and toxicity. Additionally, when reduced hERG affinity is observed a similar decrease in toxicity was also determined (compare 8 and 12i, Table 6). In contrast, compounds 2a and 2b reveal no direct correlation between hERG affinity and toxicity. Since compounds 2a and 2b are structurally almost identical, we would expect that these compounds could cross the membrane in the same magnitude and therefore have similar hERG affinity, and if there is a correlation between hERG affinity and toxicity, also similar toxicity. However, this is not the case, toxicity of compounds 2a and 2b reveal similar toxicity but hERG affinity is different, suggesting a different target for toxicity.

40

Chapter 6

Neutral Pyridines Cytotoxicity and

Gene Targeting effect

Results  and  Discussion  

Gene targeting represents the most powerful tool for genetic engineering; however random integration of DNA occurs more frequently than HR, hampering the application of gene targeting. As mentioned earlier, inducing a DSB in the target locus is the best approach for stimulating gene targeting by HR.[16] However, DSBs are highly toxic as a single unrepaired DSB can lead to cell death or genomic instability. The potential for off-target cleavage presents an important safety concern for the use of customized endonucleases (Zinc fingers and TALENs). Significant off-target activity has been documented [16], and one of the continuing challenges has been to minimize harmful effects associated with rare-cutting nucleases. Preliminary screening allowed the identification of chemical compounds that could stimulate gene targeting without inducing DSBs. In the present study, we are particularly interested in finding enhancers of this error free DSB repair mechanism and prove that chemical modulation is a viable alternative to boost genetic engineering by gene targeting. To date, only one compound is known to stimulate the human homologous recombination protein RAD51,[37] and therefore has the potential of enhancing gene targeting. It was demonstrated that RS-1 could stimulate strand assimilation activity of hRAD51 at concentrations of 10 µM. However, in contrast to those findings, our results demonstrate that RS-1 can stimulate gene targeting up to 2-fold at 20 µM (Figure 10), a concentration that does not reduce cell viability more than 30%. Remarkably, RS-1 was able to stimulate gene targeting at toxic concentrations of 35 and 50 µM. In dose-response curves, colonies were counted to determine the correlation between toxicity and inhibiting colony formation. As expected, a decrease in colonies was observed when RS-1 concentration was increased in gene targeting experiments (Figure 10).

Figure 10 Toxicity and Gene Targeting efficiency of RS-1. A) Dose-dependent effects of

RS-1 on cell viability of engineered #5.1 mouse Embryonic Stem Cells. B) Dose-dependent effects of RS-1 on gene targeting efficiency of engineered #5.1 mouse Embryonic Stem Cells. The targeting efficiencies of the GT experiments treated with the compounds were compared to the controls (no treatment), this gives a relative GT efficiency. Data shown are mean ± standard error of at least 3 independent experiments.

Having established that a toxic compound can stimulate gene targeting, neutral pyridines were evaluated for cytotoxicity and the ability of enhancing gene targeting in mES stem cells at 10, 20 and 50 µM. It was found that SAR regarding GT of neutral pyridines was best found at a concentration of 50 µM, therefore results on cytotoxicity and GT presented regard initially 50 µM and later for the best

0 20 40 60 80 100 20 40 60 80 100 120 Concentration (µM) Ce ll V ia b ili ty (%)

A

B

42

Stimulating Gene Targeting by Inducing a Double Stranded Break (Table 8)

Murine Embryonic Stem (ES) cells were engineered to contain a binding site for the endonuclease I-SceI in one of the alleles of RAD54 locus. This engineered cell line allows the comparison between GT stimulation induced by the chemical compounds and by a precise DSB, which is the best tool know so far for stimulating GT[16]. Our results demonstrate that inducing a DSB leads to gene targeting efficiencies of ~1.8 fold. However, compound RS-1 showed to stimulate gene targeting 2,3 fold, demonstrating the validity of chemical modulation of gene targeting as an alternative approach. Nonetheless, RS-1 showed to be toxic at 50 µM, which is a limiting parameter for RS1 in enhancing gene targeting.

Table 8: Toxicity and Gene Targeting Efficiency of Neutral Pyridines Affect of neutral

pyridines on cell viability and gene targeting of engineered #5.1 mouse Embryonic Stem Cells at 50 µM. The targeting efficiencies of the GT experiments treated with the compounds were compared to the controls (no treatment), this resulted in relative GT efficiency. Data shown are mean ± standard error of at least 3 independent experiments.

Position of Nitrogen in the Pyridine Moiety (Table 8)

Compounds 1a and 1b, which differ only in the position of the nitrogen in the pyridine moiety, present similar stimulation of gene targeting at 50 µM, indicating that the position of the nitrogen is not very relevant for this biological effect. However, regarding toxicity, changing the position of the nitrogen results in a different biological outcome. While pyridine 1b did not affect cell viability, pyridine 1a reduced cell viability by about 30%.

Compound 1a 1b 4b 4c RS(1 DSB 12,83/±5,7 2,39/±/0,46

Cell Viability (%) ±SEM at 50 µM in mES cells Gene Targeting Efficiency ±SEM Structure 1,91/±/0,28 ( 1,78/±/0,07 70,55/±/2,9 99,34//±/9,9/ 73,75/±/2,9 81,91/±/4,3 1,47/±/0,21 1,59/±/0,11 1,09/±/0,19 N O HCl N O HCl N O O N O O N Br H _O S N O O H Br

Side chain length (Table 8)

Increasing side-chain length could increase toxic activity for positively charged pyridines, as demonstrated earlier. Therefore, side-chain length was also evaluated for neutral pyridines (compare 4b and 4c). In contrast to the positively charged analogues, increasing side-chain length leads to a decrease in toxicity. On the other hand, while neutral pyridine 4b did not stimulate gene targeting, neutral pyridine 4c showed to stimulate gene targeting by 1,9 fold. These results demonstrate that increasing side chain length can effectively enhance gene targeting, suggesting that flexibility is an important parameter to promote GT stimulation.

Side Chain Rigidity (Table 9)

Increasing side-chain length leads to a more flexible molecule. This flexibility translates into gene targeting stimulation, which might be related to the fact that a more flexible molecule can better assume an optimal position on its binding site. To evaluate the influence of rigidity in gene targeting, triple bonds were introduced at different positions of the side chain. Introducing a triple bond close to the peripheral aromatic groups, such as in compound 5, slightly decreases both cell viability and gene targeting, when compared to 4c. In contrast, when the triple bond was introduced close to the pyridine moiety, interestingly an increase in gene targeting stimulation while a decrease in cell viability were noticed (compare 9j and 11j). Chemical modulating of gene targeting with pyridine compounds proved to be related to the position of the triple bonds in the side-chain, and introducing triple bonds close to the pyridine moiety leads to an increase of gene targeting.

Peripheral Aromaticity (Table 9)

Compounds 9p and 11p, aliphatic analogous of compounds 9j and 11j, were screened for gene targeting to evaluate the importance of peripheral aromaticity. Neutral pyridines bearing aliphatic side chains were found to be less active in gene targeting than the corresponding aromatic derivatives, revealing the importance of peripheral aromaticity for the biological effect. Interestingly, introducing triple bonds close to the pyridines moiety increases gene targeting efficiency in both cases revealing the

44

Table 9 Affect of Flexibility and Peripheral Aromaticity on Toxicity and Gene Targeting

Affect of flexibility and peripheral aromaticity on cell viability and gene targeting of engineered #5.1 mouse Embryonic Stem Cells at 50 µM. The targeting efficiencies of the GT experiments treated with the compounds were compared to the controls (no treatment), this resulted in relative GT efficiency. Data shown are mean ± standard error of at least 3 independent experiments.

Different Peripheral Aromatic Groups

Our results demonstrated that neutral pyridines containing a peripheral aromatic group are more potent in stimulating gene targeting than their aliphatic analogous. It was also demonstrated that chemical modifications on the peripheral aromatic group could affect the toxicity of positively charged pyridines (Chapter 4). Therefore, neutral pyridines containing different peripheral aromatic groups were studied regarding their ability to induce cellular toxicity and gene targeting stimulation on mES cells. Compound 4c 5 9j 9p 11j 11p 72,73,±,4,5 1,39,±,0,27 81,91,±,4,3 1,91,±,0,28 75,79,±,3,0 1,54,±,0,07 59,28,±,2,6 1,79,±,0,15

Cell Viability (%) ±SEM Gene Targeting Structure at 50 µM in mES cells Efficiency ±SEM

89,24,±,3,9 1,23,±,0,23 0,99,±,0,04 113,09,±,3,9 N O O N O O N O O N O O N N