A novel naphthalimide that selectively targets breast cancer via the arylhydrocarbon receptor pathway

We report that the naphthalimide analogue 2-(2-aminophenyl)-1H-benzo[de]isoquinoline-1,3(2H)-dione (NAP-6) is a highly potent and selective breast cancer targeting molecule. These effects are mediated via the aryl hydrocarbon receptor (AHR) pathway and the subsequent induction of CYP1 metabolising monooxygenases in breast cancer cell line models. Indeed the triple negative breast cancer cell line MDA-MB-468 with a GI50 value of 100 nM is greater than 500-fold more sensitive to NAP-6 compared with other tumour derived cell models. Within 1 h exposure of these cells to NAP-6, CYP1A1 expression increases 25-fold, rising to 250-fold by 24 h. A smaller concurrent increase in CYP1A2 and CYP1B1 is also observed. Within 24 h these cells present with DNA damage as evident by enhanced H2AXγ expression, cell cycle checkpoint activation via increased CHK2 expression, S-phase cell cycle arrest and cell death. Specific small molecule inhibitors of the AHR and CYP1 family ameliorate these events. A positive luciferase reporter assay for NAP-6 induced XRE binding further confirms the role of the AHR in this phenomenon. Non-sensitive cell lines fail to show these biological effects. For the first time we identify 2-(2-aminophenyl)-1H-benzo[de]isoquinoline-1,3(2H)-dione as a new AHR ligand that selectively targets breast cancer.


NAP-6 selectively targets breast cancer cells.
We have previously shown that ANI-7 (Fig. 1a), a HAH ligand of the AHR, is a potent (GI 50 = 0.16 µM) and breast cancer selective (up to 300-fold when compared with other tumour types) growth inhibitor 22,23 . In the present study, we set out to examine the breast cancer selectivity of a new class of compound, discovered during the phenotypic screening of multiple chemical libraries for dynamin and clathrin inhibitors. The chemical libraries included a number of naphthalimide based compounds including 2-(2-aminophenyl)-1H-benzo[de]isoquinoline-1,3(2H)-dione (NAP-6) (Fig. 1b) and Pitstop-1 (Fig. 1c). While Pitstop-1 was found to be a potent and non-toxic inhibitor of clathrin 26 , NAP-6 was not an inhibitor of clathrin or dynamin but did produce a unique growth inhibition profile in breast cancer cell lines worthy of further investigation.
Analysis of the results in Table 1 show the initial growth inhibition screen of NAP-6 and Pitstop-1 in a panel of cancer cell lines derived from a variety of tumour types. At a concentration of 25 µM NAP-6 induced more than 100% growth inhibition in MCF-7 breast cancer cells and a considerably lower effect in the other tumour types. Pitstop-1 was relatively non-toxic at 25 µM; however a preference towards the growth inhibition of MCF-7 cells was noted (28% inhibition). The magnitude of the breast cancer selectivity is more evident when NAP-6 was screened in a broader panel of breast derived cell lines including a non-tumour derived normal MCF10A cell line (Fig. 2). Indeed the GI 50 values calculated from these dose response curves (Tables 2, 3) show that NAP-6 is active in the MCF-7, BT474, T47D, ZR-75-1, SKBR3 and MDA-MB-468 cells with GI 50 values of 0.1-0.7 µM.      Concurrently, NAP-6 (1 µM) induced a substantial increase in H2AXɣ (more than twofold) in MDA-MB-468 cells under the same conditions, indicative of DNA double strand damage (Fig. 4c). The propensity of NAP-6 to induce S-phase cell cycle arrest and selectively target breast cancer cells is typical of compounds that target the AHR pathway 23,27 .
Inhibition of the AHR pathway ameliorates the effects of NAP-6. To confirm the role of NAP-6 in activating the AHR pathway we examined the effect of the AHR antagonist, CH223191 (5 µM) 28 , on the ability of NAP-6 to induce growth inhibition. The results in Fig. 5a show a substantial increase in survival of MDA-MB-468 cells in response to NAP-6 (0.1 μM) from 10 to 50% in the presence of CH223191 (5 µM). Since the AHR pathway is also known to induce the expression of CYP1 cytochrome P450 enzymes the effect of the specific CYP1 inhibitor α-naphthoflavone was also examined. Indeed, α-naphthoflavone (10 µM) ameliorated the growth inhibitory effects of NAP-6 (0.5 µM) from total growth inhibition to 58% survival (Fig. 5b). Furthermore, siRNA knockdown of AHR expression by 60% (Fig. 5c) enhanced the survival of NAP-6 (0.5 µM) treated cells from 14 to 54% (Fig. 5d). For assay optimisation the highest concentration of inhibitor that induced the least effect on cell growth was used; similarly the concentration of NAP-6 was optimised by titrating NAP-6 with each inhibitor and observing the maximal impact on growth amelioration. Collectively, these observations confirm the role of the AHR and CYPs in mediating the effects of NAP-6 in sensitive breast cancer cell line models. Supplementary Fig. S2 shows that NAP-6 with or without CH223191 or α-naphthoflavone had no effect on cell growth in the normal MCF10A cells.
NAP-6 activates XRE activity and expression of CYP1 family members. An XRE reporter assay was exploited to determine the ability of NAP-6 to induce binding of the AHR with the XRE promotor. The NAP-6 sensitive cell line MDA-MB-468 was transfected with an XRE reporter plasmid, together with control  (Fig. 6c). Concurrent, albeit smaller, increases in CYP1A2 and CYP1B1, up to 18-fold and sixfold respectively were observed under the same conditions (Fig. 6d,e). Interestingly, no changes in the expression of the phase 2 metabolising enzyme SULT1A1 was observed following treatment (Fig. 6f). A similar profile of enhanced CYP expression (fold-change after 4 h of treatment) was observed in the other NAP-6 sensitive MCF-7, BT474, SKBR3, T47D, ZR-75-1 breast cancer cell lines, while the non-sensitive MCF10A and MDA-MB-231 failed to induce this response (Fig. 7). Collectively the ability of the sensitive cell lines to respond to NAP-6 primarily correlated with the induction of CYP1A1 and CYP1A2 expression.
Potential off-targets of NAP-6. In order to further characterise the activity of NAP-6 we also evaluated its ability to inhibit the ER receptor, aromatase activity and enzymatic activity of a broad panel of protein kinases. An eight point dose response curve of NAP-6 failed to show ER binding even at the highest concentration of 2 µM. In contrast, tamoxifen and 4-hydroxytamoxifen substantially bound to the receptor with IC 50 values of 0.012 ± 0.009 µM and 0.0017 ± 0.007 µM, respectively (Table 4). NAP-6 also failed to inhibit aromatase activity at concentrations up to 100 µM while anastrozole showed substantial inhibition with an IC 50 value of 0.12 ± 0.07 µM (Table 4). In a broad panel of kinase enzymes including tyrosine kinase receptors, lipid kinases, or those specific to the PI3K/mTOR or MAP Kinase pathway, NAP-6 at a concentration of 10 µM did not substantially alter kinase activity (Table 5).

Discussion and conclusion
We report that the naphthalimide analogue 2- is a potent (nM) and selective (more than 500-fold) inhibitor of cell growth in breast cancer cell line models. Importantly, the cell lines sensitive to treatment are derived from breast cancer molecular subtypes of luminal A (MCF-7, T47D, ZR-75-1), luminal B (BT474), basal (MDA-MB-468, BT20) and HER2 (SKBR3) with varying receptor status for ER, PR, and HER2. Also included is the MCF-7/VP16 cell line which presents with a drug resistant phenotype (MCF-7/VP16) and overexpresses the p-glycoprotein drug transporter ABCC1.
Notably, all ER positive lines including MCF-7, BT474, T47D, ZR-75-1 and MCF7/VP16 were sensitive to growth inhibition of NAP-6. Similar sensitivity was also shown for two ER negative cell lines, i.e. SKBR3 and MDA-MB-468 cells, with the latter the most sensitive. This profile of sensitivity to AHR ligands has previously been described by us and others 23,[29][30][31] . The non-sensitive MDA-MB-231 cell line is triple negative for receptor status and represents a basal subtype of breast cancer with amplifying mutations in KRas and BRaf activity 32 . Such mutations are rarely found in breast cancer 33 . The resistance of MDA-MB-231 cells to AHR activation has been observed with other ligands including aminoflavone and ANI-7 19,23,34 . The only non-breast cancer cell line that showed appreciable sensitivity to NAP-6 was the A431 skin-vulva cell line, an ER positive cell line www.nature.com/scientificreports/ overexpressing the EGFR growth receptor 35 . These observations in cell line sensitivity mirror those observed for the HAH AHR ligand ANI-7 and aminoflavone previously reported by us, albeit of a differing chemical class 23 .
The naphthalimide core structure of NAP-6 is observed in other chemotherapy treatments including amonafide, and mitonafide 36 . Amonafide, a topoisomerase II inhibitor was assessed in a Phase III clinical trial in combination with cytarabine for treatment of secondary acute myeloid leukemia. This disease is characterised by overexpression of drug efflux mechanisms including the gp170 protein, of which amonafide is not a substrate. Unfortunately this combination was no better than standard treatment 36 . Sankara et al. 37 , designed various naphthalimide benzothiazole/cinnamide analogues, all of which showed DNA intercalation, topoisomerase inhibition and low µM cytotoxicity using the MTT assay; however, none showed breast cancer selectivity in MCF-7 cells. Li et al. 17 designed various naphthalimide-pyrazolyl derivatives that again showed DNA intercalation properties; the most potent of these showed a slight selectively towards MCF-7 cells (MTT assay 0.7 µM) compared with HeLa (3 µM) and A549 (5 µM) cells. Further to these observations a comprehensive review of more than 750 1,8-naphthalimide analogues was conducted in 2018, while many were effective at inhibiting the growth of cancer cell lines, none presented with the breast selectivity observed in the present study nor was the AHR pathway noted in their mode of action 38 .
Using standard cell biology methods we show that NAP-6 binds to the AHR, induces translocation to the nucleus, activates the XRE (Fig. 6a), induces CYP1 activity (Fig. 6c-e), culminating in cell cycle arrest (Fig. 3), www.nature.com/scientificreports/ checkpoint activation (Fig. 4), DNA damage (Fig. 4) and cell death (Fig. 2). Of note, is the substantial induction of CYP1 expression within 1 h following treatment, with CYP1A1 dominating the effect (Fig. 6c). The amelioration of this effect by AHR antagonists, AHR siRNA and CYP inhibition (Fig. 5) further supports the role of the AHR pathway. The enhanced expression of CYP1A1, CYP1A2, and CYP1B1 in the other NAP-6 sensitive cells lines (Fig. 7) confirms that this phenomenon is not specific for the MDA-MB-468 cells but rather the mode-ofaction across all sensitive populations.
In previous studies we have examined the inherent expression of the AHR family members in the current panel of cell lines 23 . Comparisons show that the ability of NAP-6 to activate the AHR is not dependent upon the inherent expression of the pathway members (AHR, ARNT and CYP1), underscoring their inducible nature rather than constitutive activity 23 . However, comparison with the inherent expression of SULT1A1 did predict for NAP-6 sensitivity and its expression was not altered following treatment (Fig. 6). The role of the phase 2 metabolising enzyme SULT1A1 has been noted in the breast cancer selectivity of ANI-7 and aminoflavone 23,27 . Moreover, the transfection of SULT1A1 into MDA-MB-231 aminoflavone-resistant cells restored sensitivity 27 , highlighting the multiple steps of drug metabolism. Structurally, aryl amine groups are metabolised by CYPs to form N-hydroxyl metabolites that are then substrates for bioactivation by sulphur transferase (SULT1A1). N-sulfoxy groups are further converted to active nitrenium ions, which form DNA adducts and induce cell death 27 . The presence of the DNA damage marker H2AXɣ and cell cycle checkpoint activation suggest that NAP-6 is metabolised to a DNA interacting compound; although the identity of this metabolite is unknown.
For the first time we have identified the naphthalimide analogue 2-(2-aminophenyl)-1H-benzo[de]isoquinoline-1,3(2H)-dione (NAP-6) as a potent and selective inhibitor of breast cancer cell growth. This is the first 1,8-naphthalimide analogue to show this effect. This study builds upon our knowledge of selectively targeting breast cancers and significantly adds to the development of molecules exploiting the AHR pathway.  Table 2).  Growth inhibition. Growth inhibition was determined by plating cells in duplicate in medium (100 µL) at a density of 2,500-4,000 cells per well in 96-well plates. On day 0 (24 h after plating), when the cells are in logarithmic growth, medium (100 µL) with or without the test agent was added to each well. After 72 h drug exposure, growth inhibitory effects were evaluated using the MTT (3-(4,5-dimethyltiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay and absorbance read at 540 nm 39 . The percentage growth inhibition was calculated at a fixed concentration of 25 µM (Table 1). The GI 50 value was calculated from an eight-point dose-response curve as shown in Fig. 2 using MS Excel software. Each data point is the mean ± the standard error of the mean (SEM) calculated from 4-5 replicates which were performed on separate occasions and separate cell line passages. The GI 50 value represents the drug concentration at which cell growth was inhibited by 50% based on the difference between the optical density values on day 0 and those at the end of drug exposure (Tables 2, 3).
cell cycle analysis. Cells in logarithmic growth were transferred to 6 well plates at a density of 2 × 10 5 -2.5 × 10 5 cells/well. On day 0 (24 h after plating), the cells were treated with or without NAP-6. The cells were harvested 24 h after drug treatment and washed twice in phosphate buffered saline (PBS), fixed in 70% ethanol and stored overnight at − 20 °C. The cell pellet was incubated in 600 µL of PBS containing propidium iodide (40 µg mL −1 ) and RNase (200 µg mL −1 ) for at least 30 min at room temperature. The samples (1.5 × 10 4 events) were www.nature.com/scientificreports/ analysed for fluorescence (FL2 detector, filter 575/30 nm band pass) using a FACScan (Becton Dickinson) 23 . Cell cycle distribution was assessed using Cell Quest software. Experiments were each performed on three separate occasions. Values are the percentage distribution for each phase of the cell cycle (Fig. 3).
Western blotting. Cells (3 × 10 5 ) were plated in 6 well plates in DMEM media containing test agent. At the indicated times the cells were harvested and protein content determined (Lowry Modified/Biorad Protein Assay). Equal aliquots (20 µg) of total protein from whole cell lysates were fractionated on a 10% denaturing sodium dodecyl sulfate (SDS) polyacrylamide gel and transferred to polyvinylidine difluoride membranes. Nonspecific interactions were blocked with 5% nonfat milk/0.05% Tween 20. Proteins were identified using rabbit monoclonal antibodies against H2AXγ, and pCHK2 (Cell signaling) and mouse monoclonal antibody CHK2. Membrane-bound antibodies were detected using goat anti-rabbit and anti-mouse secondary antibodies (Abcam) and Clarity Western ECL (Bio-Rad) 23 . Full length gels are shown in Supplementary Fig. S1. eR binding. Competition binding assays were performed by using an enzyme fragment complementation (EFC) method described in the HitHunter (Freemont, CA) EFC Estrogen Chemiluminescence Assay kit according to the manufacturer's instructions. Briefly, competing ligands at final concentrations ranging from 25 pM to 2 µM were incubated with 5 nM recombinant ERα (Invitrogen) and 17β-estradiol-conjugated enzyme donor for 1.5 h. The enzyme acceptor was then added followed by the chemiluminescence substrate and incubated for 1 h. Relative luminescence was determined by using a GloMax Explorer plate reader (Promega). Sigmoidal standard curves were created by Excel 23 .
Aromatase assay. Aromatase reactions were carried out as previously described 40  www.nature.com/scientificreports/ Company, Ann Arbor) according to the manufacturer's instructions. Absorbance of each sample was proportional to the amount of bound estradiol tracer which was inversely proportional to the amount of estradiol 23 .

Kinase inhibition.
A dry sample of NAP-6 was sent to Reaction Biology Corp (PA, USA) and The International Centre for Kinase Profiling (The University of Dundee, UK) for kinase inhibition assays. Both organisations use the 33 P ATP radioactive filter binding assay 41 . A stock solution of NAP-6 was prepared in DMSO and kinase inhibition assays were conducted in duplicate in the presence of a single concentration of NAP-6 (10 µM). Data represents percentage kinase enzyme activity, the lower the value the greater the enzyme inhibition 23 .