Development of VPC-70619, a Small-Molecule N-Myc Inhibitor as a Potential Therapy for Neuroendocrine Prostate Cancer

The Myc family of transcription factors are involved in the development and progression of numerous cancers, including prostate cancer (PCa). Under the pressure of androgen receptor (AR)-directed therapies resistance can occur, leading to the lethal form of PCa known as neuroendocrine prostate cancer (NEPC), characterized among other features by N-Myc overexpression. There are no clinically approved treatments for NEPC, translating into poor patient prognosis and survival. Therefore, there is a pressing need to develop novel therapeutic avenues to treat NEPC patients. In this study, we investigate the N-Myc-Max DNA binding domain (DBD) as a potential target for small molecule inhibitors and utilize computer-aided drug design (CADD) approaches to discover prospective hits. Through further exploration and optimization, a compound, VPC-70619, was identified with notable anti-N-Myc potency and strong antiproliferative activity against numerous N-Myc expressing cell lines, including those representing NEPC.


Introduction
The Myc protein family are transcription factors regulating communication networks within cells by binding to DNA and modulating the expression of numerous genes involved in cell growth and division. Dysregulation of Myc, especially N-Myc, is strongly associated with the development of the most resistant malignancies, including neuroendocrine prostate cancer (NEPC) [1][2][3][4] for which efficient therapeutic options are limited [5,6]. Hence, development of novel molecules targeting N-Myc onco-driver has a great potential as therapeutic intervention for NEPC [7].
Myc has always been considered a difficult target and often classified as undruggable, due to its intrinsically disordered nature and the lack of distinct binding pockets on its surface [8][9][10][11]. In addition, directly blocking Myc transcription might trigger unwanted side effects because of its central role in controlling gene expression in cells [12]. Deployment of cutting-edge computer-aided drug design (CADD) approaches and refined experimental methodologies have been critical in recent years to overcome the challenges of targeting Myc with small molecules [13][14][15]. Novel strategies were recently developed to inhibit Myc-driven effects by disrupting the Myc-Max heterodimer, or by interfering with the formation of the Myc-Max-DNA complex.
For example, compound EN4 was recently designed as a covalent ligand targeting the Cys171 residue within the intrinsically disordered region of Myc [16]. By binding to this unique residue, EN4 reduces stability of Myc and its obligate partner, Max, thus reducing the DNA binding ability of the Myc-Max complex and its capability to carry out proliferation and tumorigenesis activity in cancer cells. Another compound MYCi361 proliferation and tumorigenesis activity in cancer cells. Another compound MYCi361 and its improved analogue, MYCi975, were developed to directly bind to and inhibit Myc activity [17]. These small molecules regulate Myc Thr58 phosphorylation and increase its proteasomal degradation, and therefore decrease Myc protein stability. Finally, L755507 is a recent preclinical chemical developed to target heterodimerization of Myc and Max to inhibit Myc transcriptional activation [18]. In that study, Singh et al. combined CADD approaches and experimental validation to design a new small molecule that disrupts the c-Myc-Max heterodimer, thus restricting growth of diverse Myc-expressing cells. Given the high structural and functional similarity between the c-Myc and N-Myc oncoproteins [19,20], we used our previously-developed chemical series targeting Myc-Max/DBD complex as templates for the further development of novel N-Myc specific inhibitors [21,22]. We subsequently designed the next generation of compounds through synergetic use of structure-based screening, modeling, docking, medicinal chemistry efforts and experimental evaluation. Through this work we have identified compound VPC-70619, which blocks the N-Myc-Max heterocomplex from binding to DNA E-boxes and demonstrated strong inhibition activity against N-Myc-dependent cell lines as well as high bioavailability in both oral and intraperitoneal administration.

Large-Scale Structure-Based Similarity Search and Scaffold Tuning
Previously, we identified a potentially druggable pocket where the Myc bHLHLZ domain interacts with its homologous domain from the Max protein, forming a stable helical configuration which binds specifically to DNA E-boxes at enhancers and promoters of Myc target genes [21]. By screening molecules targeting this site, we identified the initial hit VPC-70063 as a potential lead to inhibit Myc-downstream effects. Using the ROCS program, we performed ligand-based similarity searches against the drug-like purchasable chemical space of the ZINC15 library with VPC-70063 and related hits as query molecules and identified compounds with distinct chemical frameworks, shown in Table  1. proliferation and tumorigenesis activity in cancer cells. Another compound MYCi361 and its improved analogue, MYCi975, were developed to directly bind to and inhibit Myc activity [17]. These small molecules regulate Myc Thr58 phosphorylation and increase its proteasomal degradation, and therefore decrease Myc protein stability. Finally, L755507 is a recent preclinical chemical developed to target heterodimerization of Myc and Max to inhibit Myc transcriptional activation [18]. In that study, Singh et al. combined CADD approaches and experimental validation to design a new small molecule that disrupts the c-Myc-Max heterodimer, thus restricting growth of diverse Myc-expressing cells. Given the high structural and functional similarity between the c-Myc and N-Myc oncoproteins [19,20], we used our previously-developed chemical series targeting Myc-Max/DBD complex as templates for the further development of novel N-Myc specific inhibitors [21,22]. We subsequently designed the next generation of compounds through synergetic use of structure-based screening, modeling, docking, medicinal chemistry efforts and experimental evaluation. Through this work we have identified compound VPC-70619, which blocks the N-Myc-Max heterocomplex from binding to DNA E-boxes and demonstrated strong inhibition activity against N-Myc-dependent cell lines as well as high bioavailability in both oral and intraperitoneal administration.

Large-Scale Structure-Based Similarity Search and Scaffold Tuning
Previously, we identified a potentially druggable pocket where the Myc bHLHLZ domain interacts with its homologous domain from the Max protein, forming a stable helical configuration which binds specifically to DNA E-boxes at enhancers and promoters of Myc target genes [21]. By screening molecules targeting this site, we identified the initial hit VPC-70063 as a potential lead to inhibit Myc-downstream effects. Using the ROCS program, we performed ligand-based similarity searches against the drug-like purchasable chemical space of the ZINC15 library with VPC-70063 and related hits as query molecules and identified compounds with distinct chemical frameworks, shown in Table  1. proliferation and tumorigenesis activity in cancer cells. Another compound MYCi361 and its improved analogue, MYCi975, were developed to directly bind to and inhibit Myc activity [17]. These small molecules regulate Myc Thr58 phosphorylation and increase its proteasomal degradation, and therefore decrease Myc protein stability. Finally, L755507 is a recent preclinical chemical developed to target heterodimerization of Myc and Max to inhibit Myc transcriptional activation [18]. In that study, Singh et al. combined CADD approaches and experimental validation to design a new small molecule that disrupts the c-Myc-Max heterodimer, thus restricting growth of diverse Myc-expressing cells. Given the high structural and functional similarity between the c-Myc and N-Myc oncoproteins [19,20], we used our previously-developed chemical series targeting Myc-Max/DBD complex as templates for the further development of novel N-Myc specific inhibitors [21,22]. We subsequently designed the next generation of compounds through synergetic use of structure-based screening, modeling, docking, medicinal chemistry efforts and experimental evaluation. Through this work we have identified compound VPC-70619, which blocks the N-Myc-Max heterocomplex from binding to DNA E-boxes and demonstrated strong inhibition activity against N-Myc-dependent cell lines as well as high bioavailability in both oral and intraperitoneal administration.

Large-Scale Structure-Based Similarity Search and Scaffold Tuning
Previously, we identified a potentially druggable pocket where the Myc bHLHLZ domain interacts with its homologous domain from the Max protein, forming a stable helical configuration which binds specifically to DNA E-boxes at enhancers and promoters of Myc target genes [21]. By screening molecules targeting this site, we identified the initial hit VPC-70063 as a potential lead to inhibit Myc-downstream effects. Using the ROCS program, we performed ligand-based similarity searches against the drug-like purchasable chemical space of the ZINC15 library with VPC-70063 and related hits as query molecules and identified compounds with distinct chemical frameworks, shown in Table  1.  This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counter-screen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a. This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counter-screen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a.  This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counter-screen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a. This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counter-screen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a. This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counter-screen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a. This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counter-screen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a. This similarity search yielded the hit compound VPC-70127, which is characterized by the presence of a 2-nitro-4-(trifluoromethyl)phenyl and a pyrazine connected via a hydrazide linker. Although VPC-70127 showed strong Myc inhibitory activity in our primary transcriptional assay, initial assessment demonstrated that VPC-70127 only weakly disrupts DNA binding to Myc-Max complex. Similarly, although VPC-70511 showed promising inhibitory activity in the transcription assay in LNCaP cells, further testing in a counterscreen using Myc-negative HO15.19, revealed that VPC-70511 had a highly cytotoxic profile. Thus, we pursued another hit, VPC-70551, which contains a similar hydrazide linker connecting two substituted phenyl rings together. The 2-nitro-4-(trifluoromethyl)phenyl was replaced by 4-cyano-2-(trifluoromethyl)phenyl and the pyrazine was replaced by a (trifluoromethyl)phenyl, as shown in Figure 1a. To confirm the new scaffold's specificity toward N-Myc, we performed a round of substructure searches using N'-phenylbenzohydrazide as a query against the ZINC15 database and selected 34 analogues for experimental validation (Table 2). We then observed that the absence of the 4-cyano group or 2-trifluoromethyl significantly reduced inhibition activity (VPC-70586 and -70611). Compounds with the cyano-and trifluoromethyl substitutions are significantly more active than their counterparts without (VPC-70593, -70596, -70599, and -70604). To determine the interactions responsible for the reported activity, we docked the active compounds to a homology model of N-Myc-Max. For our binding pose predictions, the same homology model of N-Myc-Max DBD as described in our previous publication was used [22].  To confirm the new scaffold's specificity toward N-Myc, we performed a round of substructure searches using N'-phenylbenzohydrazide as a query against the ZINC15 database and selected 34 analogues for experimental validation (Table 2). We then observed that the absence of the 4-cyano group or 2-trifluoromethyl significantly reduced inhibition activity (VPC-70586 and -70611). Compounds with the cyano-and trifluoromethyl substitutions are significantly more active than their counterparts without (VPC-70593, -70596, -70599, and -70604). To determine the interactions responsible for the reported activity, we docked the active compounds to a homology model of N-Myc-Max. For our binding pose predictions, the same homology model of N-Myc-Max DBD as described in our previous publication was used [22].
The docking pose of VPC-70551 inside N-Myc-Max/DBD complex revealed that the compound forms an H-bond with the backbone of Max's Asp215 through its hydrazide linker, and a pair of weaker H-bond with sidechain Lys419 and Arg214 of N-Myc through its cyano group. The 4-cyano-2-(trifluoromethyl)phenyl group is involved in numerous hydrophobic contacts with N-Myc residues Leu397, Phe401, and Lys419, and with Max residues Arg212, Arg215 Ile218, Lys219, Phe222, and Arg239. These contacts anchor VPC-70551 into its position in the DBD pocket ( Figure 1c). A comparison between chemical structures of the analogues suggested that the N'-[4-Cyano-2-(trifluoromethyl)phenyl] group is the substituent responsible for the observed higher potency of VPC-70551 and its analogues in inhibiting Myc-Max transcriptional activity. Therefore, we utilized the N'-[4-cyano-2-(trifluoromethyl)phenyl]benzohydrazide scaffold of VPC-70551 to further explore improved hits. that the absence of the 4-cyano group or 2-trifluoromethyl significantly reduced inhibition activity (VPC-70586 and -70611). Compounds with the cyano-and trifluoromethyl substitutions are significantly more active than their counterparts without (VPC-70593, -70596, -70599, and -70604). To determine the interactions responsible for the reported activity, we docked the active compounds to a homology model of N-Myc-Max. For our binding pose predictions, the same homology model of N-Myc-Max DBD as described in our previous publication was used [22].

Extensive N-Myc Specific SAR Exploration
To determine the optimal substitutions on the free phenyl ring, we used the active scaffold for a second substructure search against the ZINC15 database to identify a set of 181 analogues for further validation (VPC-70617 to -70799, presented in Supplementary Table S1). Compounds were tested for their inhibition of a N-Myc-Max luciferase transcription activity assay at 5 and 10 µM Figure 2a [21]. Herein, we inspected all the proposed substitutions on the scaffold to discern the features required to construct a possible SAR model for inhibitors targeting N-Myc-Max/DBD complex (Figure 2b).  Table S1). Compounds were tested for their inhibition of a N-Myc-Max luciferase transcription activity assay at 5 and 10 µ M Figure 2a [21]. Herein, we inspected all the proposed substitutions on the scaffold to discern the features required to construct a possible SAR model for inhibitors targeting N-Myc-Max/DBD complex (Figure 2b).  Table 3. para-position. R3 substitution shifts ligands towards exposed surfaces of the N-Myc-Max pocket and can negatively affect the inhibitory activity. This negative effect is reported in bicyclic compounds and the majority of compounds with large or extended chemical groups that do not interact directly with the N-Myc-Max DBD (VPC-70620 . Bulky and large chemical groups were deemed unfavorable at the meta-substitutions (VPC-70649, 70748, 70795) as they would displace the scaffold completely from its proposed binding position and shift the compound to the solvent-exposed areas.

Sulfonyl and Oxy-Linkers Did Not Improve Scaffold Potency
We then tested another series of substitutions with sulfonyl linkers at the R3-position. Combinations of different groups with the linker, such as methyl, ethyl, amine, or methylaniline with the sulfonyl linker, did not help their inhibitory activity (VPC-70623, 70641, 70655, 70665, 70692, 70694, 70697, 70717, 70764, 70789, 70790, 70793, and 70796). Although the sulfonyl linkers are interacting with the Max protein, they cannot compensate for the unfavorable position of the ligands' groups in the solvent-exposed area of the pocket. This lack of activity is observed in most compounds containing sulfur groups at para and meta The inhibitory potency of the analogues appeared to be sensitive to substitutions on the phenyl ring, especially at the R3 para-position. R3 substitution shifts ligands towards exposed surfaces of the N-Myc-Max pocket and can negatively affect the inhibitory activity. This negative effect is reported in bicyclic compounds and the majority of compounds with large or extended chemical groups that do not interact directly with the N- . Bulky and large chemical groups were deemed unfavorable at the metasubstitutions (VPC-70649, 70748, 70795) as they would displace the scaffold completely from its proposed binding position and shift the compound to the solvent-exposed areas.

Sulfonyl and Oxy-Linkers Did Not Improve Scaffold Potency
We then tested another series of substitutions with sulfonyl linkers at the R3-position. Combinations of different groups with the linker, such as methyl, ethyl, amine, or methylaniline with the sulfonyl linker, did not help their inhibitory activity (VPC-70623, 70641,  70655, 70665, 70692, 70694, 70697, 70717, 70764, 70789, 70790, 70793, and 70796). Although the sulfonyl linkers are interacting with the Max protein, they cannot compensate for the unfavorable position of the ligands' groups in the solvent-exposed area of the pocket. This lack of activity is observed in most compounds containing sulfur groups at para and meta positions (VPC-70630, 70676, 70682, 70688, 70729). Interestingly, three compounds (VPC-70658, 70733, and 70782) with sulfur or large linkers were active, and all three had fluorine groups.
Compounds with simultaneous halogen substitutions with at least one -Cl at the metapositions of the scaffold are shown to be significantly more active. Combining -Cl with either -Cl, -Br, and -CF 3 substitutions demonstrated very strong transcription inhibition (VPC-70650, 70652, and 70722). We observed that -Cl led to a strong effect that could compensate for other substitutions that were deemed inactive in other compounds. This is the case for VPC-70619 which substituted the meta -OH group of VPC-70673, for a -Cl, restoring the inhibitory activity at 5 µM.
Based on the above SAR considerations we have concluded that due to the presence of the strong electronegative -Cl group, mainly at the R3 position and to a lesser extent -F and -Br, there is a strong case to have at least one halogenic group or a functional group with significant electronegativity as a substituent on the free phenyl ring. We stipulated that halogenic groups might have a crucial role as a hydrophobic contact to anchor the ligand to Max. Consequently, we carried out further viability assays to determine whether the reported activity in -Cl compounds emerged from toxicity effects.

VPC-70619 Shows a Good Balance of Potency and Viability in Multiple N-Myc Expressing Cell Lines
We shortlisted 54 active compounds requiring further testing in cell viability assays to determine their inhibition effect on N-Myc-dependant cellular growth. The shortlisted compounds included all derivatives of the N'-[4-Cyano-2-(trifluoromethyl)phenyl]benzohydrazide scaffold which were active at both 10 µM and 5 µM in our LNCaP-NMYC transcription inhibition assays. The compounds are mainly characterized by substitutions at the paraand meta-positions, with at least one of the substituted groups being a halogenic or a strong electronegative group.
The 54 compounds were tested at 10 µM for their inhibition of the IMR32 human neuroblastoma cell line, in which the MYCN gene is amplified and actively expressed ( Table 3). HO15.19 cell line (Myc-negative) was used in parallel as a compound specificity control. Compounds showing an inhibition of HO15.19 cell line greater than 20% at 10 µM were excluded for toxicity.      We reported 13 compounds with minimal toxicity effects in the Myc-negative HO15.19 cell lines. From the 13 compounds, only one compound (VPC-70619) was effective in the IMR32 cell lines with an inhibition higher than 50%. At 10 µM, VPC-70619 had an inhibition of 99.4% in IMR32 cells while having a minimal effect of 14% in HO15. 19

VPC-70619 Interaction with N-Myc-Max Complex Interferes with Its Binding to the DNA
Based on a promising activity profile of VPC-70619, we studied its binding pose in greater details. We observed that VPC-70619 has a similar docking profile to VPC-70551 thanks to their common shared scaffold. The compound maintained most of the same interactions, including engagement to the side chains of Lys419 of N-Myc. Additionally, side chains Lys219, Asp216, Arg212, and Arg215 act as clamps to anchor the phenyl ring into a stable conformation in the pocket through hydrophobic interactions (Figure 4). Although the cyano group itself does not interact with Myc-Max, the group is important to maintain for microsomal stability, and clashes directly with the DNA E-box. (c) Microsomal stability for compounds 70063, 70551, and 70619 were also determined. 70619 was shown to be highly stable with a T 1/2 of 2310 min while we reported T 1/2 of 141 and 69 min for 70551 and 70063, respectively.

VPC-70619 Interaction with N-Myc-Max Complex Interferes with Its Binding to the DNA
Based on a promising activity profile of VPC-70619, we studied its binding pose in greater details. We observed that VPC-70619 has a similar docking profile to VPC-70551 thanks to their common shared scaffold. The compound maintained most of the same interactions, including engagement to the side chains of Lys419 of N-Myc. Additionally, side chains Lys219, Asp216, Arg212, and Arg215 act as clamps to anchor the phenyl ring into a stable conformation in the pocket through hydrophobic interactions (Figure 4). Although the cyano group itself does not interact with Myc-Max, the group is important to maintain for microsomal stability, and clashes directly with the DNA E-box.
As seen in the docking poses, the 3-chloro-4-cyano benzohydrazide portion of VPC-70619 overlaps significantly with the DNA backbone (Figure 4a  As seen in the docking poses, the 3-chloro-4-cyano benzohydrazide portion of VPC-70619 overlaps significantly with the DNA backbone (Figure 4a). These groups are in direct contact with DNA E-box recognition sequences. More specifically, the chloro and cyano groups are colliding significantly with the phosphate group of DC110 (third position in the E-box recognition sequence), disrupting its H-bond with Max side chains Arg215 and Arg212. The carbonyl oxygen of the hydrazide linker sterically clashes with the phosphate group DA109 (2nd position), disrupting it from interacting with Arg239 of Myc and Lys219 of Max. The proposed binding mode of VPC-70619 suggests that the compound can compete with DNA binding to disrupt N-Myc-Max-DNA interaction.
We first performed a microscale thermophoresis (MST) assay to validate the direct binding of VPC-70619 to the recombinant N-Myc-Max complex and to calculate the binding affinity. We evaluated the effect of increasing concentrations of 70619 on the movement of a fluorescent labelled N-Myc-Max complex through a temperature gradient induced by an infrared laser. VPC-70619 demonstrated a dose-dependent shift of thermophoresis; however, due to the limited solubility of VPC-70619 in the assay buffer, we could only estimate a dissociation constant above 130 μM (Figure 4b). We first performed a microscale thermophoresis (MST) assay to validate the direct binding of VPC-70619 to the recombinant N-Myc-Max complex and to calculate the binding affinity. We evaluated the effect of increasing concentrations of 70619 on the movement of a fluorescent labelled N-Myc-Max complex through a temperature gradient induced by an infrared laser. VPC-70619 demonstrated a dose-dependent shift of thermophoresis; however, due to the limited solubility of VPC-70619 in the assay buffer, we could only estimate a dissociation constant above 130 µM (Figure 4b).
To confirm that 70619 could disrupt the binding of N-Myc-Max DBD to DNA, we performed a bio-layer interferometry (BLI) assay where we immobilized biotinylated E-box sequence on streptavidin sensors, and we evaluated recombinant N-Myc-Max binding to the DNA in the presence of increasing concentrations of VPC-70619. We found that the compound was effective at blocking N-Myc-Max binding to DNA in a dose-dependent manner (Figure 4c). We finally tested VPC-70619 in a proximity-ligation assay (PLA) to determine whether the compound was interfering with the N-Myc-Max heterodimer, and demonstrated that the chemical does not disrupt the N-Myc-Max heterodimer formation in LNCaP cells. We observed a similar number of interactions when comparing the ratio of signal per nuclei in the presence or absence of the compound. Comparatively, in the presence of the Myc-Max dimer inhibitor 10074-G5, the number of interactions is reduced, indicating a decrease in N-Myc-Max heterodimer formation (Figure 4d).

Pharmacokinetic Study of VPC-70619 Reveals Good Intraperitoneal and Peroral Tolerance
In order to characterize VPC-70619 as a drug candidate against N-myc, its pharmacokinetic (PK) characteristics were evaluated in Balb/c mice following intraperitoneal (IP) and peroral (PO) administrations. As presented in Figure 5, we determined the PK properties of VPC-70619 and compared it to the parental molecule, VPC-70551. For the latter, the peak plasma concentrations (T max ) occurred at 480 min for PO and 60 min for IP, with peak concentrations (C max ) of 2600 ng/mL and 6220 ng/mL, respectively, following administration in 75:25 propylene glycol (PG)/poly-(ethylene glycol) 400 (PEG) formulation at 10 mg/kg dosing. Plasma concentrations of 70551 decreased slowly, with a half-life (T 1/2 ) of 332 min for IP administration. The clearance of 70551 for PO delivery was extremely slow, and we could not determine the T 1/2 in the experimental conditions used for this assay. No obvious adverse effects of 70551 were observed in the tested animals. Regarding 70619, it presented much higher bioavailability, with a C max of 29,500 ng/mL and 24,000 ng/mL for IP and PO, respectively, following administration in a 20:80 Kolliphor EL formulation. Plasma concentrations decreased slowly, with a reported terminal half-life of 330 min for IP, while PO had a T 1/2 of 427 min. No signs of toxicity were observed in any control or treated animals. Therefore, VPC-70619 was deemed to have appropriate efficacy in vivo, and a better pharmacokinetic profile due to higher bioavailability for both oral and intraperitoneal administration.

Discussion
While N-Myc is not significantly expressed in adult tissues, it was reported to be amplified and overexpressed in NEPC tumors [23]. The presence of N-Myc alteration and deregulation drives tumor proliferation and progression, leading to a low survival rate in

Discussion
While N-Myc is not significantly expressed in adult tissues, it was reported to be amplified and overexpressed in NEPC tumors [23]. The presence of N-Myc alteration and deregulation drives tumor proliferation and progression, leading to a low survival rate in patients affected by NEPC [23]. Significant efforts were carried out to indirectly inhibit Myc-driven tumors by either interfering with the expression of Myc at the DNA (gene transcription), RNA (mRNA translation) and protein level (stability), or by targeting Myc genes with synthetic lethality [24]. Successful inhibition of the N-Myc-Max complex formation has shown to be an effective therapeutic strategy [25]; however, there are currently no clinically-approved compounds that were designed to interact directly with the N-Myc-Max DBD to treat PCa. Targeting N-Myc directly could be advantageous to avoid the likelihood of resistance emerging due to redundancy and parallelism in the oncogenic pathways [26,27]. Therefore, successful inhibition of "undruggable" and yet major oncoproteins like Myc would constitute a critical step forward for the plausibility of modulating difficult transcription factor proteins, opening the door for the possibility to develop therapeutics for a wider range of cancers. Consequently, integrating novel CADD approaches would play an important role in targeting novel sites without classical properties, such as those with unusual and convex pocket shapes or transient protein-protein interactions [28].
In our previous studies, we described a novel anti-Myc compound series targeting the proposed Myc-Max heterodimer DBD pocket [21], and reported the development of a dual N-Myc and AURKA inhibitor using the currently identified scaffold [22]. Due to toxicity effects and microsomal instability in our initial compound series, we sought to optimize our scaffold's N-Myc specificity to reduce cytotoxicity, and explored ways to increase microsomal stability of chemicals. For our CADD experiments, we utilized the same homology model of the N-Myc-Max DBD complex as previously described, and established an active scaffold based on similarity and substructure searches. The large-scale structure-based similarity searches allowed us to identify the N'-phenylbenzohydrazide scaffold of VPC-70551 as a potential starting point for optimized N-Myc inhibitors. We validated a refined scaffold through a second structural similarity search. This fine-tuned search confirmed that the addition of a 4-cyano-2-(trifluoromethyl) substitution to the phenylbenzohydrazide enhanced the inhibitory activity.
To improve the potency in the series, we explored the enhanced N-Myc scaffold by searching for new compounds with substitutions on the empty phenyl ring. Extensive SAR analysis revealed that successful inhibition relies on the presence of an electronegative group such as halogens at the para-or meta position of the phenyl ring. The halogen substitutions, combined with another small group, maintained and enhanced activity or specificity. Docking simulations show that the halogens at the para or meta positions interact with the N-Myc-Max DBD site through a network of hydrophobic contacts, while halogen substitutions at the ortho position do not impart any activity as they are not able to interact with the pocket or are displaced by the compound completely when bound. Similarly, large, extended, and bulky substitution groups are detrimental to the scaffold's anti-N-Myc activity, as they often displace the compound from the pocket in our docking simulations. We tested multiple series of ligands, including sulfonyl, methyl, ethyl, and oxy linkers, and the majority of them did not return significant improvement over the parental compound, VPC-70551. We stipulate that they could not enhance the proposed interaction network that was already in place, or could not compensate for the unfavorable solvation of the ligands' exposed groups. Our binding and SAR model provides valuable insight into the proposed N-Myc-Max DBD pocket, and suggests that active compounds must bind tightly to both N-Myc and Max and minimize solvent-exposed groups.
Using cell-based screening, we could exclude molecules presenting off-target effects or toxicity. We identified a potent compound, VPC-70619, characterized by the same 4cyano-2-(trifluoromethyl)phenyl-benzohydrazide scaffold which anchors the compound into the hydrophobic core of the N-Myc-Max DBD pocket, as well as a 2-chlorobenzonitrile group which clashes with the DNA E-box binding to N-Myc-Max. Interestingly, 70619 has 15-fold increased microsomal stability, with a half-life of 2310 min compared to 140 min for VPC-70551. The lead compound, 70619, inhibits cell proliferation levels in the N-Mycdriven NCI-H660, with a low micromolar IC 50 of 7.0 µM. Importantly, 70619 was effective at inhibiting growth in N-Myc-specific cells (IMR32 and NCI-H660) while presenting minimal cytotoxicity in the Myc-negative HO15.19 cell lines. The compound was shown to bind to recombinant N-Myc-Max complex and to block DNA binding without disrupting the N-Myc-Max heterodimer itself. However, VPC-70619 presented somewhat weak affinity to N-Myc-Max; this might be due to the nature of the proposed pocket, and thus further optimization of the molecule is needed to ensure the success of our structural determination of the N-Myc-Max interaction with the identified inhibitors. Nonetheless, we propose that 70619 s mode of action in inhibiting N-myc cell proliferation is via the blocking of the N-Myc-Max heterocomplex from binding to target genes.
Based on the proposed SAR model, including a halogenic group in our series is highly beneficial to the reported anti-N-Myc activity. As previously highlighted, their inclusion as hydrophobic moieties could therefore be important as molecular interaction with weak to medium strength [29], specifically with Max's sidechain in the proposed pocket. Therefore, active compounds in the series could be relying primarily on hydrophobic contacts to interact with N-Myc-Max, and the addition of H-bonds could be only slightly beneficial to the overall interaction network. Critically, the presence of a halogenic group seemed to be significantly involved in disrupting DNA binding to N-Myc-Max DBD. For our lead compound, VPC-70619, both halogenic and nitrile groups interfere with the DNA at multiple positions. We observed that inclusion of nitriles is important for overall microsomal stability, in accordance with what has been proposed elsewhere [30]. Consequently, it will be important to fine-tune the balance between the potency and cytotoxicity of the reported anti-N-Myc compound series for better optimization.
Thus, we propose compound VPC-70619 as a single agent that could be effective with tolerable toxic effects within a specific therapeutic window, or as a chemotherapy agent which can be part of a combined oncogenic treatment regimen [31,32]. Pharmacokinetic studies revealed that VPC-70619 has high bioavailability by the intraperitoneal and, importantly, oral administration, which both have distinct advantages as alternate routes over intravenous administration, most notably by avoiding the first-pass effect of hepatic metabolism [33]. The pharmacokinetic profile of VPC-70619 is a significant improvement over our previous hit compounds, and represents an important step forward. Significant efforts should now be devoted to improving its viability profile. The reported scaffold of VPC-70619 has successfully guided the CADD pipeline and our hit selection. It provided new insight into the proposed N-Myc-Max DBD pocket and could represent possible starts for therapeutic development of a direct anti-N-Myc drug candidate with high potency and selectivity. This new class of N-Myc inhibitor demonstrates that interfering directly with N-Myc-Max's ability to interact with DNA E-boxes could be a viable mechanism for the design of a novel type of anti-cancer drugs to treat lethal NEPC.

Chemical Similarity Searches
To identify potentially improved derivatives, high-similarity 3D structural searches were performed on the ZINC15 database [34] using the ROCS program from OpenEye [35]. Omega2 from OpenEye (Santa Fe, NM, USA) was used to generate the different conformers of each searchable compound [36], and 2D substructure-based searches were performed against the ZINC15 database. Hit compounds from similarity and substructure searches were ranked per their Tanimoto score. Specialized small molecules were received from ENAMINE-REAL from our similarity searches, and custom-made molecules received from Life Chemicals were prepared for docking simulations.

Protein and Ligand Preparation
The Protein Preparation Wizard tool within Maestro 9.3 from Schrödinger LLC (New York, NY, USA) [37] was used to prepare the N-Myc-Max DNA-binding pocket homology model and to add adequate hydrogen atoms to the proteins. The protein structures were then submitted to energy minimization using the OPLS3e forcefield until RMSD reached convergence at 0.3 Å, relative to the starting geometry [38].

Docking
Docking analysis of all compounds obtained from similarity searches or medicinal chemistry synthesis was performed with the Glide docking program from Schrödinger LLC [39]. A docking grid was generated by setting a 20 Å box centered in the suggested binding region of the N-Myc-Max DBD structure. The compounds were docked into the N-Myc-Max DBD pocket using Glide SP (Standard Precision) mode. The docking poses were then scored and ranked accordingly.

Myc Transcription Assay
LNCaP-NMYC cells were transfected using TransIT-2020 transfection reagents per the manufacturer's protocol. Cells were plated at a density of 10,000 cells per well of a 96-well plate. Following 24 h treatment with 5 µM or 10 µM of the compounds, the Myc reporter activity was measured using the Cignal Myc Reporter Assay Kit from Qiagen (#336841) (Hilden, Germany) per the manufacturer's instructions.

Cell Viability
Viability of N-Myc positive (IMR32and NCI-H660) and negative (HO15.19) cells were determined using the CellTiter-Glo Luminescent Cell Viability Assay (Promega cat. G7570) (Madison, WI, USA). Cells were seeded into a 96-well white plate with a clear bottom: 10,000 cells per well for the IMR32 cells cultured in EMEM supplemented with 10% FBS, 2000 HO15.19 cells per well in Eagle's minimum essential medium (EMEM) supplemented with 10% FBS, and 2000 cells per well for the NCI-H660 neuroendocrine cells cultured in RPMI supplemented 5% FBS, 1% Insulin-Transferrin-Selenium (Thermo Fisher, 41400-045), 10nM b-estradiol (Sigma, E8875), 10nM hydrocortisone (Sigma H0888) and 1% matrigel. Following 24 h incubation, the cells were treated with serial dilutions of the derivatives along with the respective parental compound starting at 25 µM for 72 h. Cell Titer Glo was added to each well at a ratio of 1:1 and incubated on a shaker at RT for 2 min. Luminescence and IC 50 was measured with the TECAN InfiniteM200 plate reader (Männedorf, Switzerland).

Protein Purification
The bHLHLZ domain of N-Myc (amino acids 309-394) and Max (amino acids 12-93) protein sequences were cloned into pETDuet with an N-terminal 6×His. pETDuet-His-NMyc-Max was transformed and expressed in E. coli BL21 and induced with 0.5 mM 4.2.9. Thermophoresis (MST) An MST assay was conducted to determine the binding affinity of 70619 to the purified N-Myc-Max protein complex. N-Myc-Max protein samples were labeled with the red fluorescent dye NT647 using the Monolith NT Protein labeling kit RED-NHS amine-reactive (NanoTemper Technologies, München, Germany); 70619 was serial diluted with 100% DMSO and mixed with the labeled fluorescent N-Myc-Max (final concentration of protein: 5 nM) in the assay reaction buffer (20 mM Tris, pH = 8, 100 mM NaCl, 0.2 mM TCEP, 0.1 mM PMSF, 5% glycerol) starting at 600 µM, with a final concentration of DMSO of 5%. The protein/inhibitor solution was mixed and incubated at room temperature in the dark for 5 min before being filled into the capillaries. MST assays were performed with 20% LED/excitation power and medium MST power using premium capillaries for Monolith NT.115. Analysis of the data for Kd estimation was calculated with the MO Affinity Analysis software from Nanotemper (München, Germany).

Conclusions
In this study, we determined that N'-[4-Cyano-2-(trifluoromethyl)phenyl]benzohydrazide derivatives developed by CADD modelling synergized with experimental validation are potent and selective N-Myc inhibitors. We further reported that the previously-identified binding site on the N-Myc-Max/DBD complex represents a viable drug target for inhibition with small molecules. The lead compound VPC-70619, identified here from multiple rounds of structural similarity searches and molecular docking, demonstrated strong antiproliferative activity against a neuroendocrine cell line. While we identified a potential therapeutic window that could situate VPC-70619 as a potential candidate in anti-Myc treatments, supplemental optimization studies will be required to determine its full potential for future clinical applications. We identified novel derivatives through extensive SAR studies of structural analogues, providing valuable insight into N-Myc-Max DBD site topology. Finally, we have proposed that interfering directly with the ability of N-Myc-Max to interact with DNA E-boxes could be a viable mechanism for the design of potent small molecules to treat NEPC patients.