Synthesis of 2-Phenylazonaphtho[1,8-ef][1,4]diazepines and 9-(3-Arylhydrazono)pyrrolo[1,2-a]perimidines as Antitumor Agents

Two series of naphtho[1,8-ef][1,4]diazepines and pyrrolo[1,2-a]perimidines were prepared starting from 1,8-diaminonaphthalene and hydrazonoyl chlorides. The structures of the products were determined on the basis of their spectral data and elemental analyses. The mechanism of formation of such products was also discussed. The prepared compounds were screened for their antitumor activity against three cell lines, namely, MCF-7, TK-10 and UACC-62, and some derivatives showed promising activity.


Chemistry
Treatment of 1,8-diaminonaphthalene (1) with the acetylhydrazonoyl chloride 2a in dioxane in the presence of triethylamine at reflux temperature furnished a single product to which the structure 2-(4-methoxyphenylhydrazono)-3-methyl-1H-naphtho [1,8-ef] [1,4] diazepine (3a) was assigned based on the elemental analyses and spectral data of the isolated product. The IR spectrum of the reaction product showed two NH absorption bands at 3,394 and 3,286 cm −1 . Moreover, the mass spectrum of the isolated product revealed a molecular ion peak at m/z 330. 1 H-NMR displayed two singlet peaks resonating at 2.25 and 3.58 due to the CH 3 and OCH 3 protons, respectively, in addition to two broad D 2 O-exchangeable signals resonating at 9.98 and 10.10 due to two NH protons, in addition to an aromatic multiplet in the 6.65-7.92 ppm region.
Analogous reactions of the acetyl hydrazonoyl chlorides 2b-j with 1,8-diaminonaphthalene (1) furnished the corresponding 2-(arylhydrazono)-3-methyl-1H-naphtho [1,8-ef] [1,4] diazepine derivatives 3b-j in high yields, as outlined in Scheme 1. The structures of the isolated products were assigned on the basis of their elemental analyses and spectral data (see Experimental section). Compounds 3 have three tautomeric forms 3A-C (Scheme 1). The electronic absorption spectra of 3 in dioxane revealed, in each case, one characteristic absorption band in the region 381-345 nm (Table 1) and the electronic absorption spectra of compound 3d, taken in different solvents as a typical example of the series prepared, exhibited little, if any, solvent dependence (Table 1). These data indicate that compound 3 is found mainly in the hydrazone-form 3A [33,34].
Similar reaction of 1,8-diaminonaphthalene (1) with the ester hydrazonoyl chlorides 4a,b, under the typical experimental reaction conditions mentioned above, led to the formation of the corresponding 2-(arylhydrazono)-1,4-dihydro-naphtho [1,8-ef] [1,4]diazepine-3-ones 5a,b. The structure elucidation of the reaction products 5a,b was based on their elemental analyses and spectral data (IR, MS and 1 H-NMR). The IR spectra of the reaction products exhibited three NH absorptions in the 3,325-3,120 cm −1 region in addition to an amide -C=O absorption band around 1,660 cm −1 , as well as three singlet D 2 Oexchangeable-NH proton signals resonating in the 9.95-10.91 region in their 1 H-NMR spectra. Moreover, their mass spectra revealed in each case a peak due to the molecular ion (M + ). The electronic absorption spectra data of 5 in dioxane revealed, in each case, one characteristic absorption band in the 378 and 385 nm region corresponding to the hydrazone chromophore 5A [33,34] (Table 1) From the literature [35,36], we found that reaction of 4-hydroxycoumarin with 1,2-phenylenediamine furnished 1,5-benzodiazepin-2-ones. This finding attracted our intention to study the reaction of 4-hydroxycoumarin with 1,8-diaminonaphthalene. Thus, heating a 1:1 molar mixture of 1,8-diaminonaphthalene (1) with 4-hydroxycoumarin (6) in refluxing ethanol furnished a single product for which two possible structures; the naphtho [1,8-bc] [1,5]diazocin-2-one structure 7 or the 1,2-dihydroperimidine 8 derivatives can be postulated (Scheme 2). Based on its spectral data (MS, IR, 1 H-NMR and 13 C-NMR) the 2-(2-hydroxybenzoylmethylene)-1,2-dihydroperimidine structure 8 was assigned to the reaction product. For example, its 13 C-NMR revealed a characteristic signal at δ 186.22 due to a ketonic carbonyl-carbon rather than an amidic carbonyl-carbon where the latter one resonates in the region δ 160-166 [25] and the signal for methene carbon at 77.15 [37]. Also, the carbon signals of the perimidine moiety were clearly observed in the 13 C-NMR spectrum of compound 8 [37]. Further evidence for the assigned structure 8 is provided by its mass spectrum, which revealed ion peaks at m/z 168 and 121, corresponding to the perimidine and HO-C 6 H 4 -CO fragments, respectively. This assignment is in good agreement with literature data indicating the stability of perimidines to be greater than that of diazocines [37][38][39]. To account for the formation of the product 8, it is suggested, as depicted in Scheme 2, that the reaction of 1 with 6 starts with two nucleophilic addition reactions to give the intermediates 9 and 10 which underwent in situ cyclization via elimination of water molecules to give 8 as the end product. Compound 8 has two possible tautomeric forms, 8A and 8B, however the presence of a singlet signal resonating at 5.62 integrated for 1 H due to (-C=CH-) and the absence of any aliphatic protons in the 2-4 region of the 1 H-NMR confirmed the tautomer 8A as the main product (Scheme 2).
The reactivity of the multifunctional -enaminoketone 8A towards the hydrazonoyl chlorides 4 was also investigated. Both the -enaminoketone 8A and hydrazonoyl chlorides 4 have reaction different sites that may attack each other, thus, when compound 8A was treated with the ester hydrazonoyl chlorides 4a-h in refluxing dioxane in the presence of triethylamine, it afforded, in each case, only one isolable product whose elemental analyses and mass spectral data were consistent with compounds 11, 12 or 13 (Scheme 3), where all products have identical molecular ion peaks due to elimination of HCl and ethanol molecules.  The IR spectra of the products in the solid state showed carbonyl group bands in the 1693-1687 cm −1 (cyclic amide C=O group) and 1647-1639 cm −1 (benzoyl C=O group) regions, which are consistent with the structure of compound 11 rather than its isomeric structures 12 or 13 (Scheme 3). The mechanism outlined in Scheme 4 accounts for the formation of compound 11. Due to the conjugation effect of the electron-donating amino group of compound 4 and the electron-withdrawing substituent (2-hydroxybenzoyl group), the double bond is highly polarized and the electron density on the α-carbon is greater than that of the nitrogen atom so the reaction is assumed to proceed through carbon-carbon cross-coupling via initial attack of the activated -enamine's carbon (nucleophile) of 8A on the electrophilic carbon of nitrilimine 4A, followed by elimination of ethanol to give the desired product 11 as depicted in Scheme 4. As products 11 can have five possible tautomeric structures 11A-E, their electronic absorption spectra were studied in dioxane and showed in each case three absorption bands in the 509-472, 361-336 and 296-231 nm regions ( Figure 1). This absorption pattern is similar to that reported for analogous azo chromophore [40]. Such data indicate that the actual tautomeric structure of such compounds in solution phase is the arylazohydroxy form 11B, whereas their IR spectral data indicate that they exist in the solid state as the diketo-hydrazone tautomeric form 11A (Figure 1).

Effect on the Growth of Human Cancer Cell Lines
The effects of the 16 newly synthesized compounds 3c-f, 3h-j, 8, 11a-h on the growth of three human cancer cell lines, TK-10, MCF-7 and UACC-62, given in concentrations that were able to cause 50% of cell growth inhibition, are summarized in Table 2. Regarding the activities of the naphtha [1,8ef]diazepine derivatives 3, no capacity to inhibit the growth of the human cancer cell lines was observed with 3e and 3h, while compound 3i inhibited the GI 50 on all the cell lines (TK-10 = 96.73 µM/mL, MCF-7 = 81.29 µM/mL and UACC-62 = 41.97 µM/mL). Compounds 3c and 3j showed an inhibitory effect against the growth of the TK-10 and MCF-7 cells, but did not possess any activity on the UACC-62 cell line. Compounds 3d and 3f showed an inhibitory effect against one cell line only (MCF-7 and TK-10 cells, respectively). On the other hand, regarding the activities of the pyrrolo[1,2α]perimidine derivatives 11a-h, compound 11c showed growth inhibitory effects that were stronger to TK-10 and MCF-7 cell lines and its inhibitory parameters (GI 50 , TGI and LC 50 ) demonstrated total growth inhibition (TGI = 25.28, 21.87 µM/mL, respectively) compared to those presented by the rest of the group which affected only on one or two of the experimented cell lines. Compound 8 showed an inhibitory effect against the growth of the TK-10 and MCF-7 cells, while the other derivatives 11a,b,e-h showed no activities. Table 2. Concentration (µM/mL) required to inhibit cell growth by 50% (GI 50 ), to produce total growth inhibition (TGI) and to cause 50% net cell killing (LC 50  The range of doses assayed was 0.01-100 µM/mL; Results are mean ± SE (standard error of the mean) (n = 3); The GI 50 , TGI and LC 50 for the other products are >100.
In conclusion, concerning the effect on the growth of human cancer cell lines, 11c was found to have the best inhibitory activity towards all the three cell lines, followed by 3i that caused 50% cell growth inhibition on all the cancer cell lines. This activity could be attributed to heterocyclic fused rings system such as naphtha [1,8-ef]diazepines and pyrrolo[1,2-a]perimidine derivatives containing multiple centers with hydrogen accepting properties (COMe, C=O and nitrogen atoms), are essential for DNA intercalating activity and antitumor activity.

Oxygen Radical Absorbance Capacity (ORAC) Assay
All the samples show weak antioxidant (ORAC) activity as scavengers. Compounds with antioxidant potential can directly interact with ROS, thereby reducing oxidative damage [41]. To investigate further the mechanism of action by which compounds 3, 8 and 11 exert a protective effect under oxidative stress conditions, the potential radical-scavenging activity using the oxygen radical absorbance capacity (ORAC) chemical assay was evaluated.
The values of the area under the curve (AUC, Table 3) showed also that compounds 3, 8 and 11 are much weaker scavengers than the reference compound Trolox, as the decline in fluorescence over time was significantly more rapid for all of the tested compounds than for Trolox (assayed at 8 μM concentration), indicating that they are not as effective as Trolox in scavenging peroxyl radicals generated by the thermal decomposition of 2,2'-azobis(2-amidinopropane) dihydrochloride (AAPH). Thus, because peroxy radicals are not effectively neutralized by the studied compounds, fluorescein is degraded, resulting in non-fluorescent products.

General Information
All melting points were determined on an electrothermal Gallenkamp apparatus. Solvents were generally distilled and dried by standard literature procedures prior to use. The IR spectra were measured on a Pye-Unicam SP 300 instrument in potassium bromide discs. The 1 H-NMR spectra were recorded on a Varian Mercury VXR-300 MHz spectrometer (300 MHz for 1 H-NMR and 75 MHz for 13 C-NMR) and reported as the chemical shifts  (ppm) downfield from tetramethylsilane (TMS) used as an internal standard. The mass spectra were recorded on a GCMS-Q1000-EX Shimadzu or GCMS 5988-A HP spectrometer; the ionizing voltage was 70 eV. Elemental analyses were carried out by the Microanalytical Center of Cairo University, Giza, Egypt. TLC was run on silica gel G coated plates and iodine vapour was used as visualizing agent.

ORAC Assay
Scavenging activity was measured by using an ORAC assay [42]. Dilutions of the samples 3c-f,h-j, 8, and 11a-h and Trolox (as antioxidant reference compound) were incubated for 10 min at 37 °C in fluorescein (70 nM final concentration). After this incubation period, 2,2''-azobis(2-amidinopropane) dihydrochloride (12 mM final concentration) was added to the mixture. Fluorescence was read every 56 s for 98 min using a FLUOstar Optima (BMG Labtech, Headquarters, Germany) fluorometer. Area under curve values were calculated for each sample and compared with the AUC corresponding to Trolox. Results are expressed as µM of Trolox equivalents/mg of sample.

Human Cell Lines for Cytotoxicity Assays
The human renal denocarcinoma (TK-10), the human breast adenocarcinoma (MCF-7) and the human melanoma (UACC-62) Cell lines were used in these experiments. The human tumour cytotoxicities were determined following protocols established by the NCI [43]. TK-10, MCF-7 and UACC-62 cell lines were cultured in RPMI 1640 medium (Bio Whittaker, Basel, Switzerland) containing 20% foetal calf serum (FCS), 2 mM L-glutamine, 100 µg/mL penicillin and 100 µg/mL streptomycin. All cell lines were maintained at 37 °C in a 5% CO 2 atmosphere with 95% humidity. Maintenance cultures were passaged weekly, and the culture medium was changed twice a week.

Testing Procedure and Data Processing
The sulforhodamine B (SRB) assay is used for cell density determination, based on the measurement of cellular protein content. The method described here has been optimized for the toxicity screening of compounds to adherent cells in a 96-well format. Viable cells were counted using a Coulter counter and diluted with medium to give final densities of 15 × 10 4 , 5 × 10 4 and 100 × 10 4 cells/mL for TK-10, MCF-7and UACC-62, respectively. After 24 h, the cells were treated with the serial concentrations of extracts. One hundred microlitres per well of each concentration was added to the plates to obtain final concentration of 10 −4 , 10 −5 , 10 −6 , 10 −7 and 10 −8 M. The final volume in each well was 200 µg/mL. The plates were incubated for 48 h.

Sulphorhodamine B Method
After an incubation period for 48 h, cell monolayers are fixed with 50 µL of cold 50% (wt/vol) trichloroacetic acid (TAC) and stained for 60 min at 4 °C, after which the excess dye (SRB solution 0.4% w/v in 1% acetic acid) is removed after 30 min incubated, by washing repeatedly with 1% (vol/vol) acetic acid. The protein-bound dye is dissolved in 10 mM Tris base solution for OD determination at 492 nm using a microplate reader. At the end, GI 50 values (concentrations required to inhibit cell growth by 50%), TGI (concentration resulting in total growth inhibition) and LC 50 (concentration causing 50% of net cell killing) were calculated according to the previously described protocols [43]. Two or three experiments were carried out for each extract or compound. The data are given as the mean of three different assays ± SE.

Conclusions
Highly efficient and simple methods were described for the preparation of two series of naphtho [1,8-ef] [1,4]diazepines and pyrrolo[1,2-a]perimidines from 1,8-diaminonaphthalene and hydrazonoyl chlorides. The mechanism of formation of such products was discussed. The newly synthesized compounds were examined for in vitro activity against several human cancer cell lines, including TK-10, MCF-7 and UACC-62. Compound 11c was found to have the best inhibitory activity towards all the three cell lines, followed by 3i that caused inhibition of 50% of cell growth on all the cancer cell lines.