Chinese Chemical Letters  2018, Vol. 29 Issue (12): 1741-1756   PDF    
Naphthalimides and analogues as antitumor agents: A review on molecular design, bioactivity and mechanism of action
Zhuo Chen, Yufang Xu, Xuhong Qian*     
State Key Laboratory of Bioreactor Engineering, Shanghai Key Laboratory of Chemical Biology, Shanghai Key Laboratory of New Drug Design, School of Pharmacy, East China University of Science and Technology, Shanghai 200237, China
Abstract: In this review, we retrospect our progress in biological active naphthalimide and analogues as antitumor agents in the past 20 years. On one hand, various derivations in naphthalimide pharmacophores were developed to enhance their DNA binding affinity and antitumor property thereby. Heterocyclic fused naphthalimides, bis-naphthalimides, non-fused substituted naphthalimides and the carboxamide derivatives were synthesized. For example, thio-heterocyclic fused-naphthalimides were designed and evaluated in comparison with their oxo-heterocyclic fused analogues. Extended or created heterocyclebased skeleton were also developed as antitumor agents. On the other hand, we broaden the design strategy of naphthalimide antitumor agents besides DNA intercalation and topo Ⅱ poison. We have introduced more drug design methods, such as prodrugs, multitarget drugs, computer-aided drug design, photodynamic therapy. For example, we have got naphthalimide derivatives which inhibited topo Ⅱ and induced LMP by introducing long alkyl chain and polyamines. Several representative compounds were clarified of their antitumor mechanism of action. In all, our research improves the structure diversity of naphthalimide antitumor agents and distinct variances of antitumor targets and mechanism of action.
Keywords: Naphthalimides     Antitumor     Structure activity relationship     Drug design strategy     Mechanism of action    
1. Introduction

Naphthalimide and its analogues are conventional fluorescent dyes for fibers. In the past decades, our group was devoted to the application of naphthalimides derivatives as sensors and medicine. On one hand, the π-conjugated polycyclic ring system of naphthalimides is structural base for chromophores and fluorophores. And we have already summarized our exploration on the development of naphthalimides as fluorescent sensors in the featured article of 2010 [1]. An important application of these sensors is cancer diagnose and prediction [2, 3]. For example, we have designed N-acetyltransferase 2 (NAT2) probe to predict the activity of NAT2 [4]. On the other hand, the structural features of naphthalimide skeleton, i.e. rigidity, planarity and hydrophobicity, allow them to insert into the cavities of biomacromolecules, for example, intercalation with DNA was responsible for their cytotoxicity and application in cancer treatment [5].

Cancer is the second leading cause of death in the world. It accounted for 8.2 million deaths all over the world in 2012, and the amount is likely to rise to 13.1 million in 2030. Cancer is treated by surgery, radiation, chemotherapy, photodynamic therapy, biological therapy and so on. Chemotherapy is an effective treatment using chemical to destroy cancer cells. It could be used along or in combination with the other therapies. DNA plays an important role in the uncontrolled growth of cancer cells and are consequently continues to be one of the effective molecular targets in the design of antineoplastic agents. Intercalation is one of the major categories of drug-DNA interactions. Small molecules which bind to the DNA double helical structures by intercalation normally consist of a flat, generally π-deficient aromatic or heteroaromatic system and a basic side chain [6]. They would insert between the base pairs of the double helix, distort the DNA conformation (like decrease in the DNA helical twist and lengthening of the DNA) and interfere with DNA-protein interaction.

In 1973, Brana et al. reported the first series of naphthalimide derivatives on their high antitumor activity upon a variety of murine and human tumor cells. Among them, Mitonafide () and Amonafide () have entered into clinical trials (Fig. 1). They could bind to double-stranded DNA by intercalation and induce a topoisomerase Ⅱ (topo Ⅱ)-mediated DNA cleavage at nucleotide No. 1830 on pBR322 DNA [7]. The clinical development of terminated in Phase Ⅱ trials due to its inefficacy in the solid tumors in slow administration schedule and central neurotoxicity in short administration schedule [7]. As for its analogue , it was extensively metabolized to N-acetyl-amonafide via N-acetylation by NAT2. This metabolite caused a high-variable, unpredictable toxicity because of the inter-individual differences in N-acetylation and greatly obstructed its clinical development.

Download:
Fig. 1. The structures of Amonafide, Mitonafide, Azonafide and Elinafide.

From then on, several structural modification strategies were developed to avoid the N-acetylation and improve the antitumor potency of naphthalimide derivatives. Conventionally, these compounds were designed by the following methods: 1) fused ring instead of free primary-amino group; 2) bisnaphthalimides; 3) blocking the primary-amino group with N-substitution on the aromatic ring; 4) 3, 6-disubstituted naphthalimides [7, 8]. Several of the derivatives have also entered the clinical trial, such as UNBS5162 [9, 10], Azonafide () [11, 12] and Elinafide () [6, 13]. Normally speaking, these compounds preserved the structural features of DNA intercalators, including a flat, generally π-deficient aromatic or heteroaromatic system and the basic side chains.

Meanwhile, more and more novel naphthalimide derivatives with favorable pharmacologic profiles have been synthesized and evaluated. From the point of view of antitumor drug targets, various mechanisms of action were proved in the past decade. For example, 4-amino-1, 8-naphthalimide was found to be potent PARP-1 inhibitor in 1992 [14]. Kiss et al. reported UNBS3157 as a pan-antagonist of CXCL chemokine expression [15]. Besides, more structural modifications were proved to efficient for naphthalimide antitumor agents recently. For example, 4-substituted analogues, which was believed to be weak intercalator or antitumor agent before [7], were synthesized to avoid N-acetylation and approved to be potent antitumor agents as their 3-substituted analogues [16]. More functional groups were introduced to N-imide position instead of basic side chains, and the derivatives were no longer limited by the structure feature of DNA intercalators.

In the past decades, our group did a lot work to improve the structure diversity of naphthalimide antitumor agents, which were designed by various drug design strategies, such as prodrugs, multitarget drugs, computer-aided drug design, photodynamic therapy. The antitumor mechanism was studied and more antitumor drug targets were proved except for DNA binding. In this review, we classified our work by their structural designing strategies.

2. Heterocyclic fused naphthalimides as ring-fused strategy

Our research on naphthalimide derived antitumor agents started from the exploration on naphthalimide-based DNA photocleavers since 1996. These compounds could intercalate into the DNA duplex in the dark and efficiently cleave DNA upon photochemical activation. As shown in Fig. 2, various naphthalimide derivatives were designed and synthesized, including the hydroperoxide and N-aroyloxy derived naphathlimides (1-7) [17-24], thiodiazole fused naphthalimides (8) [25], angular thiozole-fused naphthalimides (9-11) [26, 27], '4-1' pentacyclic naphthalimides (12) [28], phenazine-fused naphthalimides (13) [29], acridine-fused naphthalimides (14) [30]. The structureactivity relationship of naphthalimides was studied on their DNA binding activity. It was found extension of the conjugated systems not only caused a bathochromic shift of the UV-absorption band close to the irradiation wavelength, but also improved the DNA intercalating capacity, which in all resulted in the enhanced DNA cleaving capabilities [19]. Besides, basic side chains, like N, N-dimethylaminoethyl were introduced to improve their DNA intercalation activities as DNA photocleavers, and found to be able to improve the anticancer activities at the same time. For example, compound 11a was about 30-fold more cytotoxic than that of in A549 cells (0.032 μmol/L vs. 1.10 μmol/L). Because of the limitation of the length of the review, there will be no more tautology here on the detailed evaluation of the DNA photocleaving activity of compounds in Fig. 2. Based on the structure activity relationship (SAR) studies of naphthalimide-based DNA photocleavers, more heterocyclic fused naphthalimide derivatives were designed and studied of their antitumor activities. The representative compounds were also studied of their antitumor mechanism of action.

Download:
Fig. 2. (A) Hydroperoxide and N-aroyloxy derived naphthalimides, (B) thiodiazole fused naphthalimides, (C) angular thiozole-fused naphthalimides, (D) '4-1' pentacyclic naphthalimides, (E) phenazine-fused naphthalimides and (F) acridine-fused naphthalimide derivatives.

2.1. S- strategy: thio- vs. oxo- heterocyclic fused naphthalimides as DNA photocleavers and antitumor agents

In our development of DNA photocleavers, the thiono-naphthalimides exhibited greater DNA intercalating and photocleaving ability than their oxo-counterpart, such as 15b vs. 15a (Fig. 3A) [31]. Then the thio-heterocyclic fused naphthalimides (16a-f, 17a-e, 18a-d) were designed and obtained [32-34]. They were tested of their DNA photocleaving activities and antitumor activities in comparison with their oxo-heterocyclic analogues (19a-f, 20a-d, 21a-d). The chemical structures of the synthesized compounds were listed in Fig. 3B. Herein, we will not describe the detail photocleaving activities in this review. As for the antitumor activities, the thio-heterocyclic fused derivatives could normally induce enhanced cytotoxicity than their oxo-containing analogues. Derivatives with N, N-dimethylethane-1, 2-diamine basic side chains are generally more potent than their analogues derivative, which was in consistence with previous reported naphthalimi-dederived DNA intercalators. For example, 17a and 18a exhibited significant antiproliferative activities with IC50 values at 0.14 μmol/L and 0.007 μmol/L for A549, respectively. For 19a-f and their surfur-containing analogues 16a-c, electron-donating groups improved the antitumor efficacy, like 19c vs. 19e. Besides, less cytotoxicity was observed against the normal cell line LO2 in contrast to tumor cell lines A549 and P388.

Download:
Fig. 3. (A) Thiono- or thio- naphthyl heterocycles and (B) thio vs. oxo-heterocyclic fused naphthalimides.

According to circular dichroism (CD) spectra and fluorescent spectra, the sulfur-containing analogues exhibited better DNA affinity which was consistent with their antiproliferative potency. Furthermore, the representative compounds 17a, 18a, 20a and 21a could induce significant inhibition activity to both topoisomerase Ⅰ (topo Ⅰ) and topo Ⅱ in cell-free system. They also exhibited GC sequence preference, and certain sequence selectivity against designed double-strand oligodeoxynucleotides. In all, dual topo Ⅰ/topo Ⅱ inhibition and DNA sequence preference might contribute to enhancing tumor selectivity and overcoming drug resistance [35].

The antitumor efficiency of 18a was evaluated both in vitro and in vivo with in comparison with . In a panel of various human tumor cell lines, 18a was more cytotoxic than its parent compound . It was also effective against multidrug-resistant cells. Importantly, the i.p. administration of 18a inhibited tumor growth in mice implanted with S-180 sarcoma and H22 hepatoma. The molecular and cellular machinery studies showed that the 18a functions as a topo Ⅱ poison via binding to the ATPase domain of human topo Ⅱ α. The superior cytotoxicity of 18a to was ascribed to its potent effects on trapping topo Ⅱ-DNA cleavage complexes. Moreover, using a topo Ⅱ catalytic inhibitor Aclarubicin, ataxiatelangiectasia-mutated (ATM)/ATM- and Rad3-related (ATR) kinase inhibitor caffeine and topo Ⅱ-deficient HL-60/MX2 cells, it was further shown that 18a-triggered DNA double-strand breaks, tumor cell cycle arrest, and apoptosis were in a topo Ⅱ-dependent manner. Taken together, 18a stood out by its improved anticancer activity, appreciable anti-multidrug resistance activities, and welldefined topo Ⅱ poisoning mechanisms, as comparable with the parent compound . All these collectively promise the potential value of 18a as an anticancer drug candidate, which deserves further development [36].

2.2. Linear fused strategy: angular or linear heterocyclic fused naphthalimides

It is reported that was averagely 40-fold cytotoxic against a panel of human tumor cell lines than its unfused parent compound [37]. Interestingly, also showed much stronger cytotoxicity than its angular phenanthracene analogue M, indicating that the linear chromophore is preferred [38]. Unfortunately, only one kind of such linear agent, N, bearing a tetrahydroanthracene nucleus instead of anthracene, was further reported, which exhibited averagely lower cytotoxicity than [38]. Our group synthesized compounds 23a-f with an aminothiazole ring linearly fused to the naphthalimide skeletons in comparison with their angular thiazonaphthalimide counterparts 22a and 22b. (Fig. 4)

Download:
Fig. 4. Novel linear thionaphthalimides as antitumor agents.

The linear compound 23a, was found to be the most cytotoxic one, which was about 600-fold more active against A549, 200-fold against P388 than its angular isomer 22a. It is in good agreement with the comparison of and its angular phenantherene analogue M, confirming the prevalent antitumor advantage of the linear chromophore. Also 23a was about 10-fold more active against A549 (0.001 μmol/L) than the lead compound (0.010 μmol/L), indicating the more important role of the aminothiazole moiety than phenyl moiety in cytotoxicity evaluation. Furthermore, both the angular and linear sulfur containing analogues (22a, 22b, 23a, 23b) were found to be more cytotoxic against A549 than against P388. These results showed that the introduction of an aminothiazole moiety to the naphthalimide skeleton result in higher cytotoxic and selectivity. Also, cytotoxicities of these four analogues against two tumor cells were highly dependent on the length of the side chains. 22a or 23a, with two methylene units in the side chain, was more cytotoxic than its corresponding homologue 22b or 23b, with one more methylene unit. The similar result has been found for or homologues.

23c-f also showed highly efficient antitumor activities at the range of 10-8-10-9 mol/L. They are much stronger cytotoxic agents than most of the other reported heterocyclic fused mononaphthalimides. This proved the importance of the linear thiazonaphthalene chromophores. Compound 23f, with two aminoalkyl side chains, was more cytotoxic than the other three analogues 23c-e, indicating that the introduction of a second basic side chain to the chromophore would lead to a higher cytotoxicity, similarly reported in other polycyclic system. However, none of these modified analogues was more cytotoxic against A549 and P388 than the parent compound 23a, indicating a possibility that the formation of hydrogen bond between the amino group and the sugar phosphate chain assisted the stabilization of DNA-23a complex to exert its most efficient antitumor activity [39].

Compound 23a was proved to be able to inhibit the growth of HeLa cells by induction of cell cycle arrest and apoptosis in a doseand time-dependent manner via the p53-dependent pathway. The release of cytochrome c from mitochondria was detected using confocal microscopy in HeLa cells treated with 23a. Therefore, antitumor activity of 23a is associated with the activation of p53 and the release of cytochrome c [40]. Treatment with naphthalimide-based DNA intercalators, normally lead to a decrease in Bcl-2 intracellular amounts. However, the mechanism underlying changes in Bcl-2 expression remains poorly understood. We reported later that p53 contributed to the Bcl-2 transcriptional down-regulation induced by 23a in MCF-7 cells. We found a significant increase of p53 binding to P2 promoter TATA box in MCF-7 cells by chromatin immunoprecipitation. Our research proved that 23a-induced capase-independent apoptosis in MCF-7 cells is associated with the activation of p53 and the downregulation of Bcl-2 [41].

3. Bis- and tris- naphthalimide

The dimeric heterocyclic-fused larger ring system attracted attention due to their high DNA affinity and antitumor activity [42]. As reported, bis-naphthalimide derivatives (like ) have been studied for their cytotoxicity [43]. We designed novel mono- and bis-isoquinolino[4, 5-bc]acridine derivatives 24a-f which combined the structural feature of both naphthlimide and acridine chromophores [44]. As seen in Fig. 5, naphathalimide active group was remained and the acridine unit was effectively fused with electron-deficiency group. Besides, several polyamine chains ranging from 7.3-12.3 Å were used as linkers to bridge two heterocyclic-fused acridine chromophores. The target compounds were evaluated of their antitumor activities against A549 and P388. It is found that the dimers exhibited more active cytotoxicity than the mononuclear analogue, except for 24b. Besides, the linker greatly altered the antitumor activities of the target compounds. Compound 24f with polyamino alkyl chain linker exhibited decent antitumor activity against both A549 and P388 (0.025 μmol/L and 0.281 μmol/L, respectively). DNA binding study and molecular modeling of the 24e/24f DNA complexes indicated that 24f with the long enough linker could exhibit the higher DNA affinity than 24e, which contributed to its higher antitumor activity.

Download:
Fig. 5. Mono- and bis-acridine-fused naphthalimide derivatives.

Considering the high antitumor potency of some mono- and bis-naphthalimides, we also attempted to extend these ongoing efforts on new class of "trisnaphthalimides", which contain three naphthalimide centers in one molecule. Therefore, trisnaphthalimides with different substituents on the naphthalimide ring system were synthesized and evaluated of their antiproliferative activities against MCF-7 breast adenocarcinoma and HT-29 colon carcinoma cells in comparison to related mono- and bisnaphthalimide analogues.

It is confirmed that derivatives containing a nitro substituent were the most active compounds, e.g., 25b vs. 25a, 26b vs. 26a. (Fig. 6) The bis-naphthalimide analogue 26b exhibited the great antitumor potency of this series against both MCF-7 and HT-29 (0.071 μmol/L and 0.015 μmol/L, respectively). However, trisnaphthalimides were not as potent antiproliferative agents as we expected probably due to the solubility problems in the used cell culture media. Therefore, for a further development of trisnaphthalimides as anticancer drugs, we need to improve their solubility in biological media by introduction of solubilizing substituents on the naphthalimide core or in the aminoalkyl spacers [45].

Download:
Fig. 6. Mono-, bis- and tris-naphthalimide derivatives.

4. 3- or 4- Substituted naphthalimide as non-fused strategy

Kiss et al. synthesized a series of N-substituted derivatives to reduce the high-variable, unpredictable toxicity of in clinical trials due to its acetylation by N-acetyltransferase 2 (NAT2) [7, 46]. The amino group at position 3 was reacted with acyl chloride, aldehyde, isocyanates and isothiocyanates. Among the products, UNBS3157, bearing a urea group was found to have a 3-4 fold higher maximum tolerated dose compared to and not to provoke hematotoxicity in mice at doses that display significant antitumor effects [15]. Our group designed and synthesized more 3-, and 4- substituted napthalimides based on the above discovery.

4.1. 3-Substituted naphthalimides

As introduction of alkylamino groups into the pharmacophore, such as DACA, can improve the antitumor activity [47], various aliphatic amine groups were intergrated into the 3-position of naphthalimide to avoid side effects owing to lacking of primary amine at the 3-position and retain the biological activities. Nucleophilic substitution reaction at the 3-position of naphthalimide is difficult to occur by traditional chemical method, due to the electron-charge accumulation induced by strong electronwithdrawing ability of carbonyl. Therefore, recently developed transition metal catalyzed C-N bond-forming processes were applied in the synthesis of 3-alkylamino substituted naphthalimide derivatives.

Compounds 28a-h were synthesized by three steps including bromination, amination and Cu/L-proline catalyzed coupling reaction from naphthalic anhydride (Fig. 7). They showed potential anticancer activities against HeLa and P388D1 cell lines in vitro. Compounds 28a, 28b and 28h exhibited even better activity than with the IC50 values mainly at submicromolar range, which was consistent with their Scatchard binding constants (Kb). The DNAbinding properties of 28a and 28h were evaluated via fluorescent spectra, UV-vis absorption spectra, CD spectra, and viscosity titration with calf thymus DNA (CT-DNA) in comparison with . The DNA-binding studies revealed that they bind DNA via similar mode of [48].

Download:
Fig. 7. 3-Alkylamino substituted naphthalimide analogues.

Apart from the fusing heterocycles to naphthalene nucleus as we summarized before, linkage of functional rings with a single covalent bond also drew more and more attention. Brana et al. designed and reported five "non-fused" arylnaphthalimides with good antitumor activity [49]. Our group also contributed to the synthesis of more appended arylnaphthalimides and suggested the application of "non-fused" strategy as an efficient measure to design antitumor naphthalimides.

3-Non-amino aromatic substituted naphthalimides 29a-g were synthesized starting from naphthalic anhydride by three steps including bromination, amination and Suzuki reaction by Pdcatalyst to improve antitumor activity and avoid side effects (Fig. 8). The Suzuki-Miyaura cross-coupling reaction (also named Suzuki reaction) has evolved into a most effective strategy for carbon-carbon (C—C) bond formation. It is less expensive and less toxic compared with Stille reaction employed by Brana's synthetic method. All the new naphthalimides except 29b exhibited similar or stronger antiproliferative activity than against HeLa and P388D1 cell lines in vitro. Therefore, it can be deduced that the "nonfused" strategy can be applied to develop the anticancer naphthalimides. These could avoid the side effect of as they lack amino group at the 3-position. Compound 29a and its heteroacryl substituted analogue 29g (thiophene is a bioisostere for benzene) were studied of their DNA-binding studies. The representative compounds could bind to DNA via intercalation in different fashion [50].

Download:
Fig. 8. Non-fused aromatic substituted naphthalimides.

Furthermore, an easy approach of linkage of 1, 2, 3-triazole to the 3-site or 4-site of naphthalimide by using "click reaction" was established. 1, 2, 3-Triazole was regarded as bioisosteres of amide for their similar space structure and electronic effect to amide. Its features in metabolic stability, hydrogen bonding and reactivity with proteins enabled its application in medicinal chemistry. In recent years, more and more attention has been paid to their antitumor activities. By modification with the triazole ring, the reported compounds exhibited promising DNA-binding affinity and anticancer activity in selected human cancer cell lines.

On one hand, 1, 2, 3-triazole instead of amino was linked to 3-site of naphthalimide as counterparts of to improve the biological activities and reduce drug resistance (Fig. 9). The target compounds 30a-e and 31a-e were obtained by azide-alkyne click reaction with high yields and simple purification steps from the starting material 3-nitro-1, 8-naphthalic anhydride. The antitumor activities of the targets compounds were evaluated against MCF-7, HeLa, and 7721, which proved that this reconstruction improved the cytotoxicity of the compounds (30a, 30b, 30e, 31a, 31b and 31e) compared to . The potency was closely related with the side chains, and the presence of a basic terminal group in the side chain was essential for cytotoxic activity. In particular, compound 30a were 5.6-fold, 2.4-fold, and 8.5-fold more potent than that of against the three tested cell lines, respectively. The representative compounds 30a and 31a were evaluated of their DNA intercalating properties. Compound 30a had a good ability of DNA affinity. The binding constant of 30a with DNA was 5.10 × 105 L/mol, by UV-vis spectra due to its weak fluorescence and the Scatchard binding constant of 31a was determined to be 3.51 × 105 L/mol by fluorescence spectroscopy technique. The results suggested that the 30-DNA complex was more stable with the aid of large π-conjugated systems formed by the phenyl linked to the 1, 2, 3-triazole of 30a. The result from viscosity measurement was also parallel to their DNA binding constants [51].

Download:
Fig. 9. 3- or 4-Position 1, 2, 3-triazole substituted naphthlimides.

On the other hand, increasing evidence proved that 4-position naphthalimide derivatives not only retained the antitumor activities of their 3-position counterparts, but also avoided the side effects. Therefore, 32a-e were also designed and synthesized via Cu (Ⅱ) catalyzed click reaction from 4-bromo-1, 8-naphthalimide anhydride. In accordance with the above research, the presence of a basic terminal group in the side chain was essential for cytotoxic activity. 32a, 32b and 32e exhibited comparable cytotoxicity with their 3-site counterparts and were more potent than , especially 32a with an IC50 amounting to 10-7 mol/L. According to UV-vis titration, the binding constant of 32a with DNA was 2.95 × 105 L/mol, showing good DNA affinity as other naphthalimides. However, the excellent cytotoxicity might result from the efficient action on the enzyme relating with DNA replication and transcription. Based on the CD titration, the binding mode of 32b with CT-DNA was characterized as two steps depending on the ratio (R) of [32b]/[DNA]. When R[32b]/[DNA] < 0.20, 32b intercalated into DNA and the intercalation orientations were heterogeneous. While 0.20 < R[32b]/[DNA] < 0.80, the stacking of 32b on DNA was available while intercalation was still the dominant force. When R[32b]/[DNA] > 0.80, surface stacking became the only binding mode. According to viscosity measurement, 32a had a higher capacity of DNA binding than 32d, which was parallel to their cytotoxicity [52].

After all, by application of important reactions, including Suzuki reaction, click reaction, Cu/L-proline catalyzed coupling reaction, some previous inaccessible non-fused substituted naphthalimides were designed and synthesized by our group.

4.2. 4-Substituted naphthalimides 4.2.1. Conjugated naphthalimide apoptosis inducers

Apoptosis was an evolutionarily conserved and highly regulated process, which was used to eliminate defective and unnecessary cells. Disorder of apoptosis was strongly associated with cancer, and induction of apoptosis was one of the most potent means to fight against cancer. Therefore, design and discovery of highly efficient apoptotic inducers, which could specifically target the abnormal cell death pathway found in cancer cells, have been paid great attention.

Compound and were two of the representative reported apoptotic inducers [53, 54]. As they both have the characteristics of a conformational flexible aryl moiety, naphthalimide derivatives with alkyl/aryl moiety linked by various chains were supposed to have improved or different biological activities.

Compounds 33a-c, 34a-e, 35a and 35b were designed and synthesized [55] (Fig. 10). The basic hydrophilic amine chain was introduced to the N-imide position to maintain its cytotoxicity and the amino substituents were introduced to the 4-position of naphthalimide ring to avoid the side effect from N-acetylation and involve arrest of cell cycle. Their antitumor activities were evaluated against HeLa, A549, P388, HL-60, MCF-7, HCT-8 and A375 cancer cell lines in vitro. Most of these compounds showed comparable cytotoxicities over against tested cancer cell lines except for HCT-8 with the IC50 values of 10-6 to 10-5 mol/L. In most cases, the cytotoxicities increased in sequence of 35, 34 and 33, which indicated that the magnitude and conformation of alkyl/aryl substituents had intense influence on the cytotoxic activities of these compounds. For example, 34a was more cytotoxic than 35b but less cytotoxic than 33b. Moreover, the linkers in position 4 influenced their cytotoxicity. Compounds 33c and 35a bearing flexible alkyl linker showed a striking contrast compared to 34e bearing semi-rigid piperazin linker. Additionally, the cytotoxic activities of 34a-d were also influenced by the substituents on aryl moiety. When the substituents were phenyl and nitrile groups, relatively favorable cytotoxic activities were obtained.

Download:
Fig. 10. Naphthalimide-alkyl/aryl moiety conjugates.

Besides, flow cytometric analysis indicated that these derivatives could effectively induce G2/M arrest and progress to apoptosis in HL-60 cell line after double staining with annexin Ⅴ and propidium iodide. Therefore, the target compounds might inhibit the growth of HL-60 cell line by induction of apoptosis.

4.2.2. Thio-substituted naphthalimide derivatives as DNA photocleaving inducers

According to our previous research, thio-heterocyclic fused naphthalimides induced promising antiproliferative efficiency. However, little attention has been paid to simple thio-substituted naphthalimide derivatives. Based on the structure of 18a and structurally closely related thio-heterocyclic compounds described by Brana et al., we developed a series of non-ring fused derivatives containing sulfur in position 4 of the naphthalimide ring system 36a-e. Compounds 36f and 36g, lacking either N, N-dimethylaminoethyl side chain as an optimized pharmacophore concerning the interaction with the DNA backbone and with the topoisomerase enzyme or sulfur-containing substitution, were prepared as references (Fig. 11). Besides, due to the significant DNA photocleaving activities of thio or thiono naphthalimide derivatives in comparison with corresponding oxygen derivatives, the target compounds were further evaluated as phototoxic antitumor agents in vitro, which might widen their applicability concerning the use in photodynamic therapy.

Download:
Fig. 11. Naphthalimides in photodynamic use.

Initial DNA interaction studies (including fluorescence-quenching, UV-vis and CD spectroscopy, thermal denaturation, topoisomerase western blot analysis and the photochemical DNA-cleaving abilities experiments) indicated intercalation into the DNA and interaction with the topoisomerase system as possible modes of action between target and 36a-e as well as 36g. The reference compounds 36f lacking N, N-dimethylaminoethyl side chain showed no significant activity. Based on fluorescence spectroscopic and microscopic experiments, the existence of N, N-dimethylaminoethyl side chain was important for celluar uptake. The target compounds 36c and 36e were susceptible to fluorescence induction in cellular environment and showed enhanced plasmid DNA photocleaving properties. They were significant phototoxic agents in cultured tumor cells with 3-5 times lower IC50 values after short-time UV irradiation, which suggested the further development of sulfur substituted naphthalimides for potential use in photodynamic tumor therapy (PTD). It is noticed that 36c and 36e also showed antiproliferative efficacy under nonirradiation conditions. Therefore, they could be useful lead compounds that are active under "normal" therapeutic conditions and could be additionally activated in a controlled manner by the use of advanced optical fiber technology due to the low tissue penetration of low wavelength irradiation [56].

Considering the positive influence of sulfur substituents with non-fused aromatic rings, the thiophenyl substituted compound 36c were also examined of its application in photodynamic therapy. It had an efficient cellular uptake into cultured tumor cells according to fluorescence microscopic experiments. In further extended proliferation inhibition studies under irradiation conditions, the compound were withdrawn after an initial 24 h exposure, exposed to a short irradiation at 365 nm and then further incubated under cell culture conditions. In control experiments the irradiation period was omitted. According to our results, the IC50 value for 36c dropped from 3.0 ± 0.4 μmol/L (without irradiation) to 0.8 ± 0.2 μmol/L (with irradiation) for MCF-7 cells and from 3.5 ± 0.9 μmol/L (without irradiation) to 1.4 ± 0.9 μmol/L (with irradiation) for HT-29 cells [45].

Gold complexes have recently gained attention because of their strong antiproliferative effects [57]. Furthermore, a strong inhibition of the enzyme thioredoxin reductase (TrxRa), which is involved in tumor cell proliferation, has been noted for many derivatives and an antimitochondrial mode of action for these complexes has been proposed [58]. The studies on the mode of action for Auranofin and other related gold complexes (e.g., the chloro analogue Et3PAuCl) indicated that the active species was the gold ion itself and the ligands were more relevant for the biodistribution and kinetic properties of the agents [59] (Fig. 12). Besides, the presence of the phosphine ligand is important for the biological potency of the complexes. Consequently, we considered creating an agent with the presence of the central gold atom, retention of the triethylphosphine moiety and the addition of an antiproliferative compound as ligand. N-(N', N'-dimethylaminoethyl)-4-mercapto-1, 8-napthalimide were chosen as the ligand which contains a heterocyclic naphthalimide core for DNA intercalation and a side chain containing a protonable nitrogen, which enables an initial contact to the DNA phosphate backbone.

Download:
Fig. 12. Naphthalimide-gold(Ⅰ) conjugates as antiproliferative agent and angiogenesis inhibitors.

The target compound 37 was investigated for its primary biological properties. Cell culture experiments revealed the strong antiproliferative effects and induction of apoptosis via mitochondrial pathways. Biodistribution studies by fluorescence microscopy and atomic absorption spectroscopy showed the uptake into cell organelles, an accumulation in the nuclei of tumor cells, and a homogeneous distribution in zebrafish embryos. As the nonnaphthalimide Et3PAuCl showed a much lower nuclear uptake, it may be speculated that the naphthalimide ligand of 37 might be a useful vector to facilitate transport of metals into the nucleus. In vivo monitoring of vascularization in developing zebrafish embryos revealed a significant antiangiogenic potency of the complex. However, Et3PAuCl was almost inactive in this assay. Consequently, the antiangiogenic properties can be attributed to the naphthalimide ligand of 37, which is not present in Et3PAuCl. Finally, mechanistic experiments indicated that the inhibition of thioredoxin reductase (based on the covalent binding of a gold triethylphosphine fragment) might be involved in the pharmacodynamic behavior of this novel gold species [60].

4.2.3. Multitarget antitumor agents

Cancer is a complex disease with redundant and robust biological networks. Agents modulating more than one disease related targets are believed to have superior efficacy compared to single target drugs. Multitarget drugs were defined according to Richard Morphy as "compounds that are designed to modulate multiple targets of relevance to a disease, with the overall goal of enhancing efficacy and/or improving safety" [61]. Generally, two approaches were used to design multitarget drugs: one was knowledge-based combination of pharmacophores and the other was screening of compound libraries or known drugs.

Lysosomal membrane permeabilization (LMP)-inducing molecules were novel cancer therapeutic agents as a result of their potency to induce mitochondrial outer membrane permeabilization (MOMP)-mediated intrinsic apoptosis and caspase- and mitochondrion-independent programmed cell death in apoptosis-deficient cancer cells [62-64]. Several lysosomotropic detergents, such as Siramesine and O-methylserine dodecylamide hydrochloride (MSDH), shared the structure feature of highly lipophilic weak bases and could induce LMP [65, 66]. By introducing the structural features of lysosomotropic detergents, our group synthesized naphthalimide derivatives which could inhibit topo Ⅱ and induce LMP, and ultimately cause apoptosis and cell death.

Naphthalimide derivatives 38a-d and 39a-d were functionalized at the N-imide and 4-position with polyamines and long alkyl chains. Compounds 40a-d were prepared as reference compounds (Fig. 13). All compounds except 40a and 40c exhibited better cytotoxic activity than in the tested cancer cell lines, which strongly demonstrated for the rationale of the design. Moreover, the introduction of polyamines in 38a-d and 39a-d reduced cell proliferation more significantly in the majority of these five cell lines than their monoamine precursors 40b and 40d.

Download:
Fig. 13. Multitarget naphthalimides.

The eight compounds were fairly weak DNA binders in contrast to their precursor according to CD and fluorescent spectra, but were modest topo Ⅱ inhibitor by kDNA decatenation assay, among which 39c and 39d exhibited similar topo Ⅱ poisoning potency with . Meanwhile, 38b-d could induce intense LMP in HeLa cells at the concentrations around IC50 values according to LMP induction assay based on acridine orange (AO) method. In contrast, the analogues exhibited no obvious LMP inducing activity. These results were in consistent with the potent antiproliferative activity of 38b-d comparing with their analogs, which primarily verified the advantages of multitarget drugs. Further study clarified druginduced cell death process. All the newly synthesized compounds could induce almost 100% apoptosis of HeLa cells via intrinsic mitochondrial pathway at the concentrations around IC50 values according to cytochrome c staining and annexin Ⅴ-FITC apoptosis detection. Accordingly, it provided a new paradigm for the design of novel multitarget anticancer drugs [67].

Further research was also carried on antitumor mechanism of 38b and 38c. For 38b, its effects and molecular mechanisms on Burkitt's lymphoma proliferation, cell cycle progression, apoptosis activity and oxidative stress levels of lymphoma Raji cells in vitro were observed. Our results showed that 38b inhibited the proliferation of Raji cells and induced G1 cell cycle arrest in a dose-dependent manner. Moreover, 38b treatment triggered programmed cell death, production of reactive oxygen species (ROS) and alteration of the mitochondrial membrane potential. Altogether 38b mediated its growth inhibitory effects on Raji cells via the activation of a ROS-mediated mitochondrial pathway and cell cycle checkpoint signaling pathway [68]. For 38c, it could also induce ROS generation, lysosome rupture as well as cathepsin B release. Subsequent mitochondrial damages including mitochondrial membrane permeabilization and the release of cytochrome c were also found in Hela cells when treating with 38c [69].

Angiogenesis is essential in tumor progression by providing both oxygen and nutrition for tumors beyond the size of 1-2 mm3, it has been proved to be a promising strategy for the treatment of cancers with few side effects by modulating cytokines and specific tyrosine kinases like VEGFR, PDGFR and FGFR [70]. Due to the advantages of combination of antiangiogenic agents with conventional cytotoxic agents in cancer treatment, designing single agents with both antiangiogenic and cytotoxic activity was a promising strategy.

Based on our previous research, derivatives of 39c and 39d with different lengths of alkyl chains were designed and synthesized (Fig. 14). Most target compounds exhibited effective and selective antiproliferative activities against three cancer cell lines (HL60, MDA-MB-231, A549) by inhibiting topo Ⅱ. In detail, 41a, 41d, 41e, 39c and 39d could inhibit topo Ⅱ activity as potent as in kDNA decatenation assay. In accordance with their topo Ⅱ inhibition activity, 41d, 39c and 39d inhibited the growth of cancer cell lines comparably with . Besides, 41a and 41e exhibited less potent antiproliferative activity because of their poor membrane penetration or aqueous solubility.

Download:
Fig. 14. Naphthalimides inhibited topo Ⅱ and TKs.

On the other hand, we assumed that naphthalimides might target tyrosine kinases and thereby down-regulate AKT/mTOR pathway according to recent reports. We investigated the tyrosine kinases inhibition activity of the derivatives using ELISA method. 39c and 39d were proved to moderately inhibit various angiogenesis-related RTKs, including FGFR1, VEGFR2 and PDGFRa. The representative compound 39c was then proved to possess antiangiogenic activity, which was evidenced by the inhibition of migration and tube formation activities of HMEC-1 cells. To our knowledge, it is the first time naphthalimides were identified as tyrosine kinases inhibitors (TKIs) besides their conventional cytotoxicity [71].

Due to the antitumor potency of azomacrocycles, cyclam was introduced to the N-imide and 4-position of naphthalimide instead of flexible basic side chain. Besides, lipophilic long alkyl chains were introduced to naphthalimide-cyclam conjugates in comparison with methyl and butyl substitution. Compounds 42c, 42d, 43c and 43d exhibited comparable or even more potent antiproliferative activity than , which stand for the rationale of naphthalimide-cyclam conjugates (Fig. 15). Besides, the introduction of long alkyl chain could improve the cytotoxic activity. For example, 43c and 43d were 4-6 fold potent than 43a and 43b on HeLa and HCT116 cell lines. Furthermore, 42c, 42d, 43c and 43d showed strong inhibition against both topo Ⅰ and Ⅱ. The representative compound 43c exhibited moderate DNA intercalation activity. Molecular modeling studies identified the possible interaction of 43c with molecular target by forming topoisomerase/DNA/drug ternary complex. Finally, derivatives with long alkyl chains could efficiently induce apoptosis [72].

Download:
Fig. 15. Naphthalimide-cyclam conjugates.

Cyclam can complex various transition metal ions with high kinetic and thermodynamic stability because of its strong affinity. It is reported that cyclam could chelate with Zn(Ⅱ) and Cu(Ⅱ) to inhibit cancer cell growth. The metal complexes (44a-t) (Fig. 16) of naphthalimide-cyclam conjugates (42a-d) were synthesized. According to our research, the Zn(Ⅱ) and Cr(Ⅲ) complexes exhibited comparable antiproliferative activities with via multiple tyrosine kinase inhibition. The representative compound 44p displayed broad-spectrum antiproliferative activity against 15 cancer cell lines with average IC50 value 10.18 ± 3.25 μmol/L, and effective antiangiogenic activity on human microvascular endothelial cells (HMEC-1) [73].

Download:
Fig. 16. Metal complex of naphthalimide-cyclam conjugates.

5. N-Imide substitution

Various drug targets were found to be related with the antitumor activity of naphthalimide antitumor agents in the process of developing heterocyclic fused and 3-/4-substituted derivatives. Therefore, more N-imide substitutions besides N, N-dimethylaminoethyl were introduced, such as chiral aminos, amino acids, benzoic acid.

5.1. Naphthalimide-chiral amino conjugates

Several heterocyclic-fused naphthalimides (45a-d, Fig. 17) with chiral amino side chains were investigated. Their side chains' chiral configuration determines DNA binding activities in the order: S-enantiomers > R-enantiomers. And their DNA photodamage activites were in good agreement with their DNA binding constants, the S-enantiomers could photocleave circular supercoiled pBR322 DNA more efficiently than their R-enantiomers. S-Enantiomer 45b could photodamage DNA at 0.2 μmol/L and cleave supercoiled plasmid DNA from form Ⅰ to form Ⅱ completely at 50 μmol/L. Almost all of these intercalators showed effective cytoxicities against A549 and P388. S-Enantiomers showed different antitumor cytotoxicity by comparison with R-enantiomers [74].

Download:
Fig. 17. S- vs. R-Enantiomers substituted naphthalimides.

5.2. Naphthalimide-amino acid conjugates

Various drug-amino acid conjugates were reported as promising antitumor agents [75-79]. Amino acids were essential starting materials for cell growth and building blocks for protein synthesis, and thereby improve the cell uptake of antitumor agents. Conjugation with amino acids was a possible strategy for increasing the transport of lipophilic compounds of biological interest across cell membrane [80]. As fused multi-aromatic ring system encounter problems on poor solubility, naphthalimides conjugated with flexible leucine moiety as side chain were supposed to have some improved or different biological activity.

The target compounds were obtained by conjugating the leucine, arylamine and aliphatic amine functional groups to the naphthalimide ring. The amino substituents were introduced to the 4-position of naphthalimide ring to avoid acetylation and thereby to reduce the side effects. The introduction of arylamine to leucine moiety might result in favorable solubility and pharmacologic profiles. Furthermore, the introduction of thio group, known to facilitate the interaction of naphthyl heterocycles with DNA, might lead to a concomitant increase in the cytotoxicity against tumor cell lines. The antitumor activities of naphthalimides 46a-c and 47a-c were evaluated against HeLa, A549, P388, HL-60, MCF-7, HCT-8 and A375 cancer cell lines in vitro (Fig. 18). Most of the derivatives had moderate antitumor activities with the IC50 values of 10-6-10-5 mol/L. More importantly, 47a-c exhibited exclusive antitumor activities against MCF-7 cell line. DNA binding showed that these derivatives behaved as DNA intercalating agents. Serum albumin is the most abundant protein in plasma and functioned in the binding and transportation of various ligands such as fatty acids, hormones, and drugs. The distribution, free concentration, and metabolism of these ligands strongly depended on their binding properties with serum albumin. The verified formation of compound-BSA complex might provide some suggestions on the understanding of their exclusive antitumor selectivity against MCF-7 cell line [81].

Download:
Fig. 18. Naphthalimide-amino acid conjugates.

5.3. Naphthalimide conjugated derivatives as apoptosis-inducing agents

The 4-[4-(3, 3-diphenylallyl)piperazin-1-yl] benzoic acid moiety of the apoptotic inducer (Fig. 10) was introduced to naphthalimide scaffold for enhanced tumor specificity and efficacy (Fig. 19). Naphthalimide scaffold was utilized as key prototype structural unit, and substituted benzoic acid and aliphatic amine functional groups were conjugated to naphthalimide. Amino substituent introduced to position-4 of naphthalimide was difficult to be acetylated and might involve arrest of cell cycle. Different types of linkers, ethanolamine and sulfanilamide, connected naphthalimide with substituted benzoic acid to investigate their effects on biological activity.

Download:
Fig. 19. Naphthalimide-benzoic acid conjugates.

The antitumor activities of target compounds were evaluated against a variety of cancer cell lines in vitro. Preliminary results showed that most of the derivatives had cytotoxic activity comparable with that of , with IC50 values of 10-6-10-5 mol/L. Interestingly, 48e had the selective antitumor activity against MCF-7 among the cancer cell lines tested. More importantly, flow cytometric analysis indicated that compared with , the target compounds could effectively induce G2/M arrest and progress to apoptosis in HL-60 cells after double staining with annexin Ⅴ-FITC and propidium iodide. The present work provided a novel class of naphthalimide-based derivatives with potential apoptosis-inducing and improved antitumor activity for further optimization [82].

Compound ⅩⅠ (Indomethacin) is a member of the nonsteroidal anti-inflammatory drugs, which are widely applied in treatment of arthritis and cardiovascular diseases, cancer prevention, etc. [83-85]. It has been proven that indomethacin can induce G1 arrest and apoptosis of human colorectal cancer cells. Accordingly, naphthalimide-indomethacin hybrids with different linkers might have improved or different biological activity (Fig. 20). The naphthalimide scaffold was utilized as the key prototype structural unit, and dimethylamine and indomethacin functional groups were conjugated to the naphthalene ring. The dimethylamino substituent introduced at 4-position of naphthalimide is difficult to acetylate and might be involved in DNA synthesis arrest. Various (hydrophobic or hydrophilic) linkers were designed between naphthalimide and indomethacin in order to investigate their effects on biological activity, which might lead to concomitant increase in cytotoxicity against tumor cell lines. Introduction of an amide/ester moiety might result in different pharmacologic profiles and optimal therapeutic window for hybrids.

Download:
Fig. 20. Naphthalimide-indomethacin conjugates.

The antitumor activity of target compounds was evaluated against a variety of cancer cell lines in vitro. Preliminary results showed that the hybrids had moderate cytotoxic activity and could effectively induce apoptosis in HeLa cells. More importantly, the amide derivatives 50c and 50d had better cytotoxic and proapoptotic activity than their ester analogues 50a and 50b, whereas the ester derivatives exhibited hypoxic preferred cytotoxicity and might be used as promising candidate prodrug against hypoxic tumor cells. The results indicated that linkers played important roles in biological activity. This work provides a novel class of hybrid lead compounds with improved bioactivity for further optimization. Detailed biological studies on the molecular mechanism of action of the hybrids are in progress [86].

5.4. N-Oxide potential prodrugs

had inappropriate central nervous system (CNS) toxicity and had myelosuppression, vomiting, and erythra side effects, which hampered further studies. The cationic tertiary amine side chains on these agents played an important role on electrostatic binding affinity with DNA, which also ensured good uptake into cells, then interfered with the topoisomerase function to inhibit the tumors. In order to lower the toxicity and improve extravascular drug transport properties, the oxidized tertiary amine was introduced into the naphthalimide backbone to form a prodrug lead instead of original amino side chain. When the compounds were introduced into hypoxic cells, the tertiary amine N-oxides could be bioreduced to the corresponding tertiary amine and showed the high bioactivity of the amine. Of course, this process could be inhibited by oxygen. So, we designed, synthesized, and evaluated a series of novel potential anticancer agents as prodrug leads against hypoxic solid tumor [87].

The N-oxides and their corresponding amines were initially screened for their cytotoxicities in vitro against usual A549 and P388 cell lines, respectively (Fig. 21). It was found that for oxic cells lines, the bioactivities of 51b, 51c, 51d, and 51e were decreased about 5-, 36-, 44-, 675-fold against A549, respectively, compared to the corresponding amines , 18a, 17a, 20a. And 51c, 51d, 51e decreased around 11-, 8-, 165-fold against P388, respectively. Fewer cytotoxicity of 51b against P388 was observed.

Download:
Fig. 21. Novel design compounds for potential hypoxia selective antitumor agents and markers.

The antiproliferative activities for 51a-e and their corresponding amines were then measured in vitro against A375 and V79 solid tumors. Compared with the corresponding naphthalimides, the NOxides showed less cytotoxicity in oxic A375 cell cultures. The results at least implied that the N—O group in hypoxic cell cultures might be bioreduced to corresponding amine and led to the localization of bioreductive metabolites (active cytotoxins) in hypoxic cells and should further interfere with topoisomerase function. The hypoxic cytotoxicity ratios (HCR) for 51a-e were 1.9, 1.2, 1.6, 2.4, and 1.1-fold. However, the HCR was very low, which might be caused by their insufficient bioreduction revealed by the fluorescence changes. On the other hand, the prodrugs did not show hypoxia selectivity against V79 cells except for 51a and 51e with HCR of 1.7 and 1.47, respectively, which indicated that this series of N-Oxides showed hypoxia-selective antitumor activity for A375 solid tumor cells, but not for V79.

6. Naphthalimide analogues

Our group also designed some naphthalimide analogues and discovered their antitumor activities, including naphthalene carboxamides, 1-oxo-1H-phenalene-2, 3-dicarbonitrile derivatives, naphthostyrils. Their antitumor activities were evaluated and a preliminary antitumor mechanism of action was studied.

6.1. Naphthalene carboxamides

Naphthalimides incorporated with fused phenyl or heterocyclic rings constituted an important class of compounds as the presence of the planar chromophore were capable of intercalating into the base pairs of DNA. Therefore, they were not only potent antitumor agents against a variety of murine and human tumor cells, but also DNA cleavers by generating various reactive intermediates. The larger aromatic ring system was proved to account for the higher affinity for DNA and consequently for higher antitumor and photocleaving activities. In our continued efforts to develop simple but efficient antitumor and photocleaving agents, we proposed a novel molecular design of modifying the reported naphthalimide DNA intercalators to corresponding ring-opened models, or naphthothiazole carboxamides via mercuric oxide mediated decarbonylation. The planar tri-cyclic ring systems may also intercalate into DNA to show medical or biological potential. Besides, the aminoalkyl side chain serving as DNA groove binder and/or external electrostatic binder was inferred to be more flexible to increase the affinity with DNA.

Phenyl thiazole and thiadiazole conjugated naphthalimides have been reported to bind to DNA and act as antitumor agents and DNA photonucleases. Firstly, phenyl naphthothiazole carboxamides 52a-d with aminoalkyl side chains at 1-position were synthesized (Fig. 22). Their DNA-intercalating properties were evaluated by fluorescence quenching technique. Compound 52b exhibited higher Scatchard binding constant value (1.22 × 105 L/mol). It was also the most cytotoxic compound against A549 with IC50 of 12 μmol/L. The DNA cleavage activities were also studied [88].

Download:
Fig. 22. Phenyl naphthothiazole carboxamides and thiadiazo naphthalene carboxamides.

Later, thiadiazo naphthalene carboxamides 53a and 53b with aminoalkyl side chains at 1-position and the 8-position isomers of 52a, 52b, 53a and 53b were synthesized. According to fluorescence quenching method, the Scatchard binding constants for 52a, 52b, 52e, 52f were determined to be 6.39 × 104, 1.22 × 105, 6.73 × 104, 1.29 × 105 L/mol. The slight difference between 52a and 52b and their isomers 52e and 52f suggested that position of the carboxalimide side chain in the chromophores have no effect on their corresponding DNA binding activities. Besides, the constant of 52b and 52f was as twice higher as that of 52a and 52e, which suggests the aid of hydrogen bond probably formed between the oxygen atom on the phenyl ring of 52b and 52f and hydrogen atoms in base pairs of DNA molecules. Compounds 52e and 52f showed stronger antitumor abilities than their isomers 52a and 52b. Compound 52f exhibited the highest cytotoxicty with the IC50 of 2.53 μmol/L and 0.11 μmol/L against A549 and P388 cell lines. Besides, the phenylthiazole conjugates showed stronger cytotoxicity than the thiadiazole conjugates 53a-d. Though they were less potent antitumor agents, the thiazo and thiadiazo naphthalene carboxamides could photocleave DNA more efficiently than their corresponding naphthalimides [89].

Considering the efficient DNA photocleaving activities and antitumor potency in vitro of thio-heterocyclic fused naphthalimides, their naphthalene carboxamide analogues 54a-d were synthesized (Fig. 23). The aminoalkyl side chain derived 1-position isomers were not obtained due to the slight yield or too similar molecular polarities. The Scatchard binding constants were 4.46 × 105, 6.86 × 105, 1.95 × 104, 4.02 × 104 L/mol, which were much higher than their naphthalimide analogues (2.8 × 105) and (8.27 × 103). Compounds 54a or 54b, with the thiophene ring, could intercalated into DNA more strongly than 54c and 54d having the thioxanthene ring, which was consistant with their naphthalimide analogues. Meanwhile, the length of aminoalkyl side chains could alter the intercalating abilities. The antitumor activities of 54a-d were evaluated in vitro. All these exhibited efficient antitumor activitieswith IC50rangingfrom0.176 μmol/L to29 μmol/L. Also, all of them were more cytotoxic against P388 than against A549, reflecting selectivity for a special murine or leukemia cell type. Cytotoxic potencies of these compounds against two tumor cells were highly dependent on the length of side chains. The compound with two methylene units in the side chain between two nitrogen atoms was more cytotoxic than corresponding homologue with one more methylene unit, as indicated by the cytotoxicity orders of 54a > 54b, 54c > 54d. Compound 54c was found to be the strongest inhibitor for P388 with IC50 of 0.176 μmol/L, while 54a was the most cytotoxic one against A549 with IC50 of 1.16 μmol/L. Stronger DNA cleaving activities than corresponding naphthalimides were also displayed [90].

Download:
Fig. 23. Thio-heterocyclic fused naphthalene carboxamide.

A series of 11H-benzo[a]carbazole-5-carboxamides (55a-l, Fig. 24) were designed by our co-operators to improve the aqueous solubility and maintain the antitumor activity of 18a. In comparison with 55b-d, 55a was more potent in both A549 and HCT116 cell line. Then, the basic side chain at 5-position was replaced with various amines for compounds 55e-l. Most of the derivatives exhibited moderate to good antitumor potency against A549 and HCT116 cell lines. N, N-dimethyl substituted analogues (55a and 55e) are generally more potent. Compound 55e bearing N, N-dimethyl substitution is the most potent derivatives of this series. It is less potent than 18a while has similar activity to . Further research proved that 55e could inhibit the circulation of the T4 ligase like 55a and 18a. Compounds 55a and 55e also induced G2/M arrest in HL-60 cells in consistent with 18a. Compound 55e exhibited comparable potency with against HCT-116 xenografts in nude mice [91].

Download:
Fig. 24. Benzocarbazole carboxamide.

6.2. 1-Oxo-1H-phenalene-2, 3-dicarbonitrile derivatives

Discovery novel structure leads is a major challenge and of great importance in developing novel, anticancer agents with improved efficacy and specificity for medicinal chemistry researchers. As electron-deficient polycyclic chromophoric systems play an important role in cytotoxic antitumor agents, such as the anthraquinone ring system in mitoxantron and doxorubicin, naphthalimide in and acridine in DACA, to discover novel electron-deficient heterocycles which can be easily derived or functionalized gets great attention.

1-Oxo-1H-phenalene-2, 3-dicarbonitrile, initially misidentified as 8-oxo-8H-acenaphtho[1, 2-b]pyrrol-9-carbonitrile, was discovered by our group in 2005 [92]. In 2014, two international teams (Wang and Qian et al., and Lebteton et al.) corrected the identifications through the analyses of X-ray crystallographic structures and 2D NMR spectra. During the past 10 years, the parent compound (abbreviated as phenalene dicarbonitrile) was highlighted for its convenient synthesis and easy derivation, and was functionalized for ion sensing, biomolecule imaging, and tumor diagonosis as we have overviewed in the lately published featured article. Meanwhile, phenalene dicarbonitriles and naphtho[1, 8-ef]isoindole-7, 8, 10(9H)-triones (abbreviated as phenalene imide) were also discovered as potential antitumor agents and made great progress from the very beginning.

6.2.1. Phenalene dicarbonitrile and its analogues

Firstly, phenalene dicarbonitrile derivatives with various basic amino chains (Fig. 25) were synthesized via SNArH reaction under mild conditions. They were evaluated of their antitumor activities against human cervical carcinoma (HeLa) cell line. 6-Substitued phenalene dicarbonitrile derivatives 56a-i exhibited higher antitumor activity than 6, 9-site di-substituted compounds 56j, 56m. Besides, 56k and 56i with alkyl chain amino substituents displayed weaker cytotoxicity. Furthermore, compounds with sulfur-heterocycles showed higher bioactivity than those with oxo-heterocycles, for example, 56e vs. 56d and 56h vs. 56g. Among all the tested compounds, 56e shows the highest activity (IC50, 0.17 mmol/L). Considering their potent activity and easy derivation, the compounds are potential leading compounds for the finding of valuable antitumor agents [93].

Download:
Fig. 25. Phenalene dicarbonitrile derivatives.

We proved phenalene dicarbonitrile 56e as Bcl-2 inhibitor [94]. Zhang found later that 56e inactivated not only Bcl-2, but also Mcl-1. Therefore, it acted as pan-Bcl-2 inhibitor. More efficient pan-Bcl-2 inhibitors derived from phenalene dicarbonitrile skeleton were designed and synthesized [95]. This work was summarized in our lately published review [96].

6.2.2. N-Alkylated phenalene imides and their derivatives

To improve the solubility of phenalene dicarbonitrile and its derivatives, the cyano group were hydrolyzed and alkylated to give the N-alkylated phenalene imide derivatives. Besides, various amino substituents were introduced to the 6-position of the scaffold (Fig. 26). The imides and their amino derivatives were evaluated of their antitumor activities against A549 and P388. According to the results, both N-alkylated phenalene imides and their 6-substituted amino derivatives exhibited desirable antitumor activities. Most of the 6-amino derivatives of 57a displayed weaker activities than that of their precursor. Besides, in comparison with 57l and 57m, the analogues derived from 57f bearing Br in imide group chain was much greater, namely 0.60 μmol/L and 0.032 μmol/L for 57q, 0.14 μmol/L and 0.019 μmol/L for 57r against A549 and P388, respectively. This is probably due to the alkylating activity of the 2-bromoethyl imide moiety, because leaving groups such as chloride and bromide are required for cytotoxicity in an alkylation mechanism. The 6-substituents bearing aminoalkyl amines (including 57h, 57l, 57m, 57q and 57r) possessed higher antitumor activities, which was in consistence with conventional DNA intercalators, such as acridine, anthracycline derivatives. Unexpectedly, compounds without Ndialkyl basic side chains also showed moderate activity, which suggested the existence of different antitumor mechanism [97].

Download:
Fig. 26. N-Alkylated phenalene imide and their derivatives.

Several phenalene dicarbonitrile and N-alkylated phenalene imide derivatives (Fig. 27) were synthesized and investigated of their DNA intercalation geometries and antitumor activities. The target compounds resembled the structure features of DNA intercalators, that is, they have both electron-deficient planer polycycles and basic side chains. It was proved that all the compounds behaved as DNA intercalators by SYBR Green-DNA melt curve, fluorescence titration, absorption and CD studies. According to SYBR Green-DNA melt curve, compounds 58a-c, 59b, 59c, 60b, and 60c could compete with all of the SYBR Green molecules for the intercalation sites and therefore were much stronger DNA intercalators. Compounds 59a and 60a were weaker than their analogues. The binding constants Kb were determined through fluorescence titration. It was also demonstrated that series a containing methylpiperazine substitution exhibited weaker DNA affinity than series b and c with alkylamine substitution. Besides, higher affinity was observed for phenalene dicarbonitrile derivatives as the stronger electron-withdrawing capacity of dicarbonitrile group would lead to higher electron-deficiency of planar ring system. The stacking of π-bond interactions between the electron-deficient chromophore and the electron-rich purinepyrimidine base plays an important role in DNA binding affinity. However, CH2CH2Br introduced compounds displayed weaker DNA affinity than CH3 derivatives. The more steric hindrance of CH2CH2Br moiety might be responsible for it. As demonstrated by absorption titration, 58a-DNA revealed a classical interaction, which 58b and 58c bind to DNA in a two-step binding process with the increase of DNA concentration. At low DNA/compound ratio, the interaction and electrostatic attraction both contributed to the complexation and stacking of the compounds along DNA surface. At high DNA/compound ratio, a complete intercalation mode was formed. Both absorption titration and induced circular dichroism revealed that compounds 58b and 58c behaved different intercalation modes with 58a. Compound 58a with the methylpiperazine substitution induced negative ICD signal, suggesting that it intercalated into DNA with its long axis parallel to the base-pair long axis. And the alkylamine-substituted compounds 58b and 58c exhibited positive ICD signal, which indicated a vertical orientation in the intercalation pocket. According to CD spectrum which illustrated the DNA conformational changes, 58a, whose binding affinity was among the lowest, led to the most significant base pair stacking at 275 nmol/L.

Download:
Fig. 27. Phenalene dicarbonitrile and N-alkylated phenalene imide derivatives as DNA intercalators.

The target compounds were tested of their cytotoxicity against HeLa and MCF-7 cancer cell lines. All of them, especially compounds containing the CH2CH2Br group, exhibited outstanding antitumor activity with IC50 ranging from 10-7 mol/L to 10-6 mol/L. It is found in this series of compounds (60a-c), no obvious correlation was found between their DNA binding affinities and antitumor activities. Compound 58a could alter DNA conformation most significantly but was with lowest binding affinity, displayed much lower IC50 value than other compounds. This might be due to the interference with some critical DNA binding protein. Finally, 58a induced MCF-7 cell apoptosis, while compounds 58b and 58c could only necrotize tumor cells. The different DNA binding geometry probably resulted in their different mechanism of killing tumor cells [98].

6.2.3. 6-Aliphatic amino substituted phenalene imides

The 6-aliphatic amino substituted phenalene imides were also synthesized (Fig. 28). They were evaluated of their DNA-binding affinity, cytotoxicity and cell cycle arrest. Among the targeted compounds 61b-j, 61g was the most active which exhibited moderate cytotoxicity against the tested cancer cell lines HeLa, HL-60, HCT-8, A375 and MCF-7 with the IC50 values from 14 μmol/L to 24 μmol/L. Compound 61d exhibited selective growth inhibition activities against HeLa and HL60. The sulfur-containing analogues 61f and 61i were not as potent as their phenalene dicarbonitrile and N-alkylated phenalene imide derivatives.

Download:
Fig. 28. 6-Aliphatic amino substituted phenalene imides.

According to the UV-vis absorption, 61d binds to DNA via intercalation and external contacts, like surface binding, which was similar with 61g. The calculated K (intrinsic binding constant) values have no direct relationship with the IC50 values of the new derivatives, which means there is no direct relationship between DNA-binding and cytotoxicity. Therefore, DNA is not the unique target for anticancer activity. Based on further CD and viscosity measurement, compounds 61d and 61g could bind to DNA in the same way, though the advanced DNA conformation was different. Besides, 61d could induce S-phase arrest which was followed by apoptosis, while 61g induced apoptosis by flow cytometry at various concentrations. In all, the study of 6-aliphatic amino substituted phenalene imide derivatives extended the structureactivity relationship of the phenalene aromatic-heterocycle compounds [99].

6.2.4. Phenalene imides as FGFR1 inhibitors

FGFRs are precisely regulated in normal cells but are constitutively activated in transformed cells involved in various human cancers [100]. Aberrant activation of FGFRs has been observed in the process of tumorgenesis and tumor development. Thus, inhibition of FGFRS represents an attractive tumor therapeutic strategy [101, 102].

Phenalene derivatives were confirmed of their antiproliferative activity against cancer cell lines as DNA intercalators. There are also exceptions. For example, 63a had great in vitro antiproliferative effects, but the exact molecular mechanism was not clearly understood. To identify potential target candidates of 63a, we first preformed the in silico targets screening using PharmMapper server, a reverse pharmacophore mapping approach, against an inhouse pharmaphore database (PharmTargetDB) (Fig. 29). According to the top prediction results, kinases may be involved in the suppression of proliferation of the tumor cells. The efficient drug screening platform with enzyme-linked immunosorbent assay (ELISA) determined 63a as a potent and selective FGFR1 inhibitor. To illuminate the SAR of phenalene derivatives as FGFR1 inhibitors, 3-thiol derivatives of 63a and its precursor 62a were designed based on not only sulfur-containing heterocycles derived antitumor agents reported before, but also the principle sulfurcontaining constituents, like DAS, DADS, DATS, which were responsible for tumor prevention and suppression of garlic. It is proved that N-alkylated phenalene imides (63-66a) are potent inhibitors of FGFR1 with IC50 values ranging from 19 nmol/L to 77 nmol/L. Their allylmercaptan substituted derivatives (63-66c) were also potent FGFR1 inhibitors with submicromolar IC50 values. Molecular docking simulation demonstrated that the hydrophobic aromatic ring that binds to the neutral region of ATP-binding site and the 9-substituted imide side chains that bind to the negative electricity region were essential for FGFR1 inhibition.

Download:
Fig. 29. Phenalene derivatives as FGFR1 inhibitors.

Furthermore, these compounds exhibited favorable growth inhition property against FGFR-expressing cancer cell lines with IC50 values ranging from micromolar to submicromolar. Western blotting analysis of compounds 63a-c and 64a were in agreement with the ELISA results, and 63a, 64a and 63c could reduce the phosphorylation of FGFR1 and Erk. It is the first time phenalene derivatives were reported as FGFR1 inhibitors [103].

6.3. Naphthostyril

Naphthostyril derivatives were originally developed as dyes and later as electronic typing materials since 1920s. Their prominent bioactivities have been discovered recently as thymidylate synthase inhibitors and 5-HT7R antagonists. Our group designed naphthostyril DNA intercalators based on the similarity between naphthostyril and naphthalimide. The design concept is that the lactam group on the tricyclic rings can form hydrogen bond with its targets and its planarity would facilitate its embedding to DNA base pairs.

Amino side chains were introduced as N, N-dimethyl ethyldiamino group and its analogues commonly applied in antitumor agents especially in DNA intercalators (Fig. 30). Compound 67c were designed as new antitumor agents due to the better antitumor activity and selectivity of polyamine conjugated intercalators. Compound 67d was designed as hybrid compound to crosslink with DNA as alkylating agents. Considering the potency of bis-intercalators like , 67e were synthesized. The DNA interaction property was studies for 67a and 67c by UV-vis absorption and the fluorescent spectra, which demonstrated their DNA intercalation activities. Compound 67c exhibited distinctive interaction with DNA. The antitumor activities of 67a-e were evaluated against cell lines A549 and P388, respectively. It was found that the structures of the amine side chains greatly affected the antitumor activities. Compound 67c with the polyamine side chain was the strongest growth inhibitor against A549 and P388 with IC50 of 0.489 μmol/L and 1.69 μmol/L, respectively. Thus, 67c also reflected the selectivity for A549, a special human lung cell type in comparison with its analogues [104].

Download:
Fig. 30. Naphthostyril derivatives.

Subsequently, naphthostyril derivatives without basic side chains were designed as novel DNA intercalators. Dicyano moiety was introduced to the site of oxygen atom to enlarge the conjugated plane and to reduce the electron density so as to facilitate binding of the electron-rich DNA (Fig. 31). For 68a-c, unsaturated hydrocarbons were linked to nitrogen atom to increase electrostatic interaction between compounds and acidic DNA molecules instead of amine side chains. Compounds 68d and 68e were synthesized to behave as potential DNA alkylating agents and result in inhibition of tumor growth by cross-linking of basic groups. Furthermore, 68f-g were designed as new anticancer agents by changing lactam to acylamide, as the intercalator conjugated with acylamide had been reported to have better antitumor activity.

Download:
Fig. 31. Dicycano substituted naphthostyril derivatives.

The affinities of the target compounds for CT-DNA were proved by spectroscopic technique and viscosity measurement to be efficient DNA intercalators. Compound 68f could intercalate into DNA entirely due to the good conjugation of carbonyl group with benzo[c, d]indol moiety. All the target compounds could cleave DNA both in darkness and by light irradiation. The antitumor activities of the targeted compounds were evaluated against 7721 and MCF-7 cells. All the compounds exhibited efficient antitumor activities with preference against cell line 7721 rather than MCF-7. The IC50 of most of the compounds was as low as 10-7 mol/L against 7721 cell line. Compound 68f was the most potent antitumor agents, with IC50 against MCF-7 at 0.003 μmol/L and 7721 at 0.115 μmol/L.

In all, dicyano group were introduced to modify the carbonyl group of benzo[c, d]indol-2(1H)-one as DNA poisons. The forceful electron-withdrawing ability of dicyano group played an important role in their antitumor effects. These compounds contained no basic side chains and also showed attractive biological activities. These might expand the diversity of intercalators' structures and substituting groups [105].

7. Conclusion

DNA intercalating agents are normally characterized by the presence of a tri- or tetracylic annelated planar and aromatic ring capable of inserting the nucleic acid bases and one or two flexible amino side chains for promoting DNA affinity through electrostatic or hydrophobic interactions. Due to their π-conjugated polycyclic ring system, naphthalimides were endowed to be not only chromophores and fluorophores as we reviewed before, but also antitumor drug candidates based on DNA intercalation. Our group enhanced DNA intercalating activities of naphthalimides via fusing heterocycles, dimers, non-fused substitution and carboxamide derivatives. For example, the thio-heterocyclic fused naphthalimides were normally more cytotoxic than their oxo-analogues according to our research. The linear heterocyclic fused naphthalimides were more potent antitumor agents than angular analogues. The 3- or 4-dialkylamino or alkyl substituted naphthalimides as non-fused derivatives, also showed promising antitumor activities. The antitumor potency was normally in accordance with DNA intercalation capacity. We have also synthesized other polycyclic systems and evaluated their antitumor activities. Among them, 1-oxo-1H-phenalene discovered by our group and its derivatives displayed great antitumor potency.

Our research was not limited to promote antitumor potency by enhancing DNA intercalation. We tried to regulate hallmarks of cancer, such as sustaining proliferation, resisting cell death, inducing angiogenesis. More naphthalimides were designed and proved to be with other antitumor mechanism. We have got naphthalimide-derived prodrugs, multitarget drugs, CADD drugs, photodynamic drugs. For these compounds, the antitumor drug targets were much more complex.

In all, apart from its application as fluorescent probes, our group have applied various drug design strategies to acquire more potential naphthalimide-derived antitumor agents with different antitumor targets and mechanism of action. Improved cancer specificity and lower side effects were acquired. Our attempts were benefit in the discovery of heterocyclic compounds as antitumor agents [106-109]. On one hand, improving drug-like properties is now one of our main targets of structural modification. Enlarged aromatic rings decreased the solubility of napthalimide derivatives, which would affect the absorption of the antitumor agents in vivo. We recently have synthesized a series of nonplanar and rigid naphthalimides by fusing a seven-membered heterocycle onto a naphthalimide skeleton [110]. The new series of compound were proved to be with greatly improved solubility and good photo stability and brightness in water. The antitumor efficient of the new compounds is evaluated now both in vitro and in vivo. On the other hand, naphthalimide scaffold is both a fluorophore and a pharmacophore. Therefore, our work would help to develop theranostic agents. We have made some attempts in this aspect [2, 3, 111-113]. For example, Yi Xiao et al. report the two-photon imaging using naphthalimide probe [114], and the organelles-targetable phenalene dicarbonitriles [115]. We believed that the application of naphthalimide derivatives has bright prospects.

Acknowledgments

We are grateful for the financial supports from the National Natural Science Foundation of China (No. 20536010), National Key Project for Basic Research (No. 2003CB114400), the Program of Shanghai Subject Chief Scientist and the Science and Technology Foundation of Shanghai.

References
[1]
X. Qian, Y. Xiao, Y. Xu, et al., Chem. Commun. 46 (2010) 6418-6436. DOI:10.1039/c0cc00686f
[2]
H. Duan, Y. Ding, C. Huang, et al., Chin. Chem. Lett. (2018). DOI:10.1016/j.cclet.2018.03.016
[3]
[4]
L. Cui, Y. Zhong, W.P. Zhu, et al., Chem. Commun. 46 (2010) 7121-7123. DOI:10.1039/c0cc01000f
[5]
A. Kamal, N.R. Bolla, P.S. Srikanth, A.K. Srivastava, Expert Opin. Ther. Pat. 23 (2013) 299-317. DOI:10.1517/13543776.2013.746313
[6]
M.F. Brana, M. Cacho, A. Gradillas, et al., Curr. Pharm. Des. 7 (2001) 1745-1780.
[7]
M.F. Brana, A. Ramos, Curr. Med. Chem. Anticancer Agents 1 (2001) 237-255. DOI:10.2174/1568011013354624
[8]
M.F. Brana, M. Cacho, A. Ramos, et al., Org. Biomol. Chem. 1 (2003) 648-654. DOI:10.1039/b209042b
[9]
T. Mahieu, T. Mijatovic, E. van Quaquebeke, et al., Mol. Cancer Ther. 6 (2007) A119. DOI:10.4161/cbt.6.1.3824
[10]
D. Mahadevan, D.W. Northfelt, P. Chalasani, et al., Int. J. Clin. Oncol. 18 (2013) 934-941. DOI:10.1007/s10147-012-0475-8
[11]
R.T. Dorr, J.D. Liddil, S.M. Sami, et al., Anticancer Drugs 12 (2001) 213-220. DOI:10.1097/00001813-200103000-00007
[12]
S.M. Sami, R.T. Dorr, D.S. Alberts, et al., J. Med. Chem. 43 (2000) 3067-3073. DOI:10.1021/jm9905817
[13]
M.F. Braña, M. Cacho, M.A. García, et al., J. Med. Chem. 47 (2004) 1391-1399. DOI:10.1021/jm0308850
[14]
M. Banasik, H. Komura, M. Shimoyama, K. Ueda, J. Biol. Chem. 267 (1992) 1569-1575.
[15]
E. Van Quaquebeke, T. Mahieu, P. Dumont, et al., J. Med. Chem. 50 (2007) 4122-4134. DOI:10.1021/jm070315q
[16]
J.T. Norton, M.A. Witsch, L. Luong, et al., Anti-Cancer Drug 19 (2008) 23-36. DOI:10.1097/CAD.0b013e3282f00e17
[17]
Z.F. Tao, X.H. Qian, D.Z. Wei, Dye. Pigment. 31 (1996) 245-251. DOI:10.1016/0143-7208(95)00105-0
[18]
Z.F. Tao, X.H. Qian, Dye. Pigment. 43 (1999) 139-145. DOI:10.1016/S0143-7208(99)00037-6
[19]
W. Yao, X.H. Qian, Dye. Pigment. 48 (2001) 43-47. DOI:10.1016/S0143-7208(00)00078-4
[20]
W. Yao, X.H. Qian, J. Fluorine Chem. 106 (2000) 69-72. DOI:10.1016/S0022-1139(00)00304-3
[21]
W. Yao, X.H. Qian, Q.Y. Hu, Tetrahedron Lett. 41 (2000) 7711-7715. DOI:10.1016/S0040-4039(00)01304-6
[22]
Y.F. Xu, X.Y. Huang, X.H. Qian, W. Yao, Biorg. Med. Chem. 12 (2004) 2335-2341. DOI:10.1016/j.bmc.2004.02.013
[23]
Y.F. Xu, X.H. Qian, W. Yao, et al., Biorg. Med. Chem. 11 (2003) 5427-5433. DOI:10.1016/j.bmc.2003.09.026
[24]
X.H. Qian, P. Mao, W. Yao, X.F. Guo, Tetrahedron Lett. 43 (2002) 2995-2998. DOI:10.1016/S0040-4039(02)00442-2
[25]
Y.G. Li, Y.F. Xu, X.H. Qian, B.Y. Qu, Bioorg. Med. Chem. Lett. 13 (2003) 3513-3515. DOI:10.1016/S0960-894X(03)00737-6
[26]
Z.G. Li, Q. Yang, X.H. Qian, Bioorg. Med. Chem. Lett. 15 (2005) 3143-3146. DOI:10.1016/j.bmcl.2005.04.012
[27]
Z.G. Li, Q. Yang, X.H. Qian, Biorg. Med. Chem. 13 (2005) 4864-4870. DOI:10.1016/j.bmc.2005.05.006
[28]
F. Li, J.N. Cui, L.Y. Guo, et al., Bioorg. Med. Chem. Lett. 15 (2007) 5114-5121. DOI:10.1016/j.bmc.2007.05.032
[29]
P. Yang, Q. Yang, X.H. Qian, J.N. Cui, Bioorg. Med. Chem. Lett. 13 (2005) 5909-5914. DOI:10.1016/j.bmc.2005.07.029
[30]
P. Yang, Q. Yang, X.H. Qian, et al., Biol. 84 (2006) 221-226.
[31]
X.H. Qian, T.B. Huang, D.Z. Wei, et al., J. Chem. Soc. Perkin Trans. I 2 (2000) 715-718.
[32]
Y.G. Li, Y.F. Xu, X.H. Qian, B. Qu, Tetrahedron Lett. 45 (2004) 1247-1251. DOI:10.1016/j.tetlet.2003.11.145
[33]
X.H. Qian, Y.G. Li, Y.F. Xu, et al., Bioorg. Med. Chem. Lett. 14 (2004) 2665-2668. DOI:10.1016/j.bmcl.2004.02.059
[34]
Y.F. Xu, B.Y. Qu, X.H. Qian, Y.G. Li, Bioorg. Med. Chem. Lett. 15 (2005) 1139-1142. DOI:10.1016/j.bmcl.2004.12.011
[35]
S.Y. Tan, H. Yin, Z. Chen, et al., Eur. J. Med. Chem. 62 (2013) 130-138. DOI:10.1016/j.ejmech.2012.12.039
[36]
H. Zhu, M. Huang, F. Yang, et al., Mol. Cancer Ther. 6 (2007) 484-495. DOI:10.1158/1535-7163.MCT-06-0584
[37]
S.M. Sami, R.T. Dorr, D.S. Alberts, W.A. Remers, J. Med. Chem. 36 (1993) 765-770. DOI:10.1021/jm00058a014
[38]
S.M. Sami, R.T. Dorr, A.M. Solyom, et al., J. Med. Chem. 38 (1995) 983-993. DOI:10.1021/jm00006a018
[39]
X.H. Qian, Z.G. Li, Q. Yang, Bioorg. Med. Chem. Lett. 15 (2007) 6846-6851. DOI:10.1016/j.bmc.2007.07.008
[40]
X. Liang, A.B. Wu, Y.F. Xu, et al., Invest. New Drug. 29 (2011) 646-658. DOI:10.1007/s10637-010-9403-9
[41]
X. Liang, K. Xu, Y.F. Xu, et al., Toxicol. Appl. Pharm. 256 (2011) 52-61. DOI:10.1016/j.taap.2011.07.010
[42]
W.M. Cholody, L. Hernandez, L. Hassner, et al., J. Med. Chem. 38 (1995) 3043-3052. DOI:10.1021/jm00016a007
[43]
P.F. Bousquet, D. Conlon, K.M. Fitzgerald, et al., Cancer Res. 55 (1995) 1176-1180.
[44]
P. Yang, Q. Yang, X.H. Qian, Tetrahedron 61 (2005) 11895-11901. DOI:10.1016/j.tet.2005.09.065
[45]
I. Ott, Y.F. Xu, X.H. Qian, J. Photochem. Photobiol. B 105 (2011) 75-80. DOI:10.1016/j.jphotobiol.2011.06.012
[46]
E. van Quaquebeke, T. Mahieu, P. Dumont, et al., J. Med. Chem. 50 (2007) 4122-4134. DOI:10.1021/jm070315q
[47]
[48]
L.J. Xie, Y.F. Xu, F. Wang, et al., Biorg. Med. Chem. 17 (2009) 804-810. DOI:10.1016/j.bmc.2008.11.053
[49]
M.F. Brana, M. Cacho, M.A. Garcia, et al., J. Med. Chem. 47 (2004) 1391-1399. DOI:10.1021/jm0308850
[50]
L.J. Xie, J.N. Cui, X.H. Qian, et al., Bioorg. Med. Chem. 19 (2011) 961-967. DOI:10.1016/j.bmc.2010.11.055
[51]
X.L. Li, Y.J. Lin, Y.K. Yuan, et al., Tetrahedron 67 (2011) 2299-2304. DOI:10.1016/j.tet.2011.01.063
[52]
X.L. Li, Y.J. Lin, Q.Q. Wang, et al., Eur. J. Med. Chem. 46 (2011) 1274-1279. DOI:10.1016/j.ejmech.2011.01.050
[53]
C.M. Park, T. Oie, A.M. Petros, et al., J. Am. Chem. Soc. 128 (2006) 16206-16212. DOI:10.1021/ja0650347
[54]
Z. Zhang, L. Jin, X. Qian, et al., ChemBioChem 8 (2007) 113-121. DOI:10.1002/(ISSN)1439-7633
[55]
A.B. Wu, Y.F. Xu, X.H. Qian, et al., Eur. J. Med. Chem. 44 (2009) 4674-4680. DOI:10.1016/j.ejmech.2009.07.011
[56]
I. Ott, Y.F. Xu, J.W. Liu, et al., Biorg. Med. Chem. 16 (2008) 7107-7116. DOI:10.1016/j.bmc.2008.06.052
[57]
P.J. Barnard, S.J. Berners-Price, Coord. Chem. Rev. 251 (2007) 1889-1902. DOI:10.1016/j.ccr.2007.04.006
[58]
S. Gromer, L.D. Arscott, C.H. Williams Jr, et al., J. Biol. Chem. 273 (1998) 20096-20101. DOI:10.1074/jbc.273.32.20096
[59]
C.K. Mirabell, R.K. Johnson, D.T. Hill, et al., J. Med. Chem. 29 (1986) 218-223. DOI:10.1021/jm00152a009
[60]
I. Ott, X.H. Qian, Y.F. Xu, et al., J. Med. Chem. 52 (2009) 763-770. DOI:10.1021/jm8012135
[61]
R. Morphy, Z. Rankovic, J. Med. Chem. 48 (2005) 6523-6543. DOI:10.1021/jm058225d
[62]
L. Groth-Pedersen, M.S. Ostenfeld, M. Hoyer-Hansen, et al., Cancer Res. 67 (2007) 2217-2225. DOI:10.1158/0008-5472.CAN-06-3520
[63]
G. Kroemer, M. Jaattela, Nat. Rev. Cancer 5 (2005) 886-897. DOI:10.1038/nrc1738
[64]
H. Erdal, M. Berndtsson, J. Castro, et al., Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 192-197. DOI:10.1073/pnas.0408592102
[65]
M.S. Ostenfeld, M. Hoyer-Hansen, L. Bastholm, et al., Autophagy 4 (2008) 487-499. DOI:10.4161/auto.5774
[66]
W. Li, X. Yuan, G. Nordgren, et al., FEBS Lett. 470 (2000) 35-39. DOI:10.1016/S0014-5793(00)01286-2
[67]
Z. Chen, X. Liang, H.Y. Zhang, et al., J. Med. Chem. 53 (2010) 2589-2600. DOI:10.1021/jm100025u
[68]
B. Lin, Z. Chen, Y.F. Xu, et al., Leukemia Res. 35 (2011) 646-656. DOI:10.1016/j.leukres.2011.01.029
[69]
K. Shen, L.Y. Sun, H.Y. Zhang, et al., Cancer Lett. 333 (2013) 229-238. DOI:10.1016/j.canlet.2013.01.038
[70]
J. Marx, Science 301 (2003) 452-454. DOI:10.1126/science.301.5632.452
[71]
X. Wang, Z. Chen, L. Tong, et al., Eur. J. Med. Chem. 65 (2013) 477-486. DOI:10.1016/j.ejmech.2013.05.002
[72]
S. Tan, D. Sun, J. Lyu, et al., Bioorg. Med. Chem. 23 (2015) 5672-5680. DOI:10.1016/j.bmc.2015.07.011
[73]
S.Y. Tan, K. Han, Q. Li, et al., Eur. J. Med. Chem. 85 (2014) 207-214. DOI:10.1016/j.ejmech.2014.07.068
[74]
Q. Yang, P. Yang, X.H. Qian, L.P. Tong, Bioorg. Med. Chem. Lett. 18 (2008) 6210-6213. DOI:10.1016/j.bmcl.2008.09.104
[75]
N. Dias, J.F. Goossens, B. Baldeyrou, et al., Bioconjugate Chem. 16 (2005) 949-958. DOI:10.1021/bc050065x
[76]
L.W. Hsin, H.P. Wang, P.H. Kao, et al., Bioorg. Med. Chem. 16 (2008) 1006-1014. DOI:10.1016/j.bmc.2007.10.012
[77]
I. Hutchinson, S.A. Jennings, B.R. Vishnuvajjala, et al., J. Med. Chem. 45 (2002) 744-747. DOI:10.1021/jm011025r
[78]
S. Horvat, K. Mlinaric-Majerski, L. Glavas-Obrovac, et al., J. Med. Chem. 49 (2006) 3136-3142. DOI:10.1021/jm051026+
[79]
L. Song, R. Bevins, B.D. Anderson, J. Med. Chem. 49 (2006) 4344-4355. DOI:10.1021/jm060016l
[80]
N. Margiotta, P. Papadia, F. Lazzaro, et al., J. Med. Chem. 48 (2005) 7821-7828. DOI:10.1021/jm0504003
[81]
A.B. Wu, Y.F. Xu, X.H. Qian, Bioorg. Med. Chem. 17 (2009) 592-599. DOI:10.1016/j.bmc.2008.11.080
[82]
A.B. Wu, P. Mei, Y.F. Xu, X.H. Qian, Chem. Biol. Drug Des. 78 (2011) 941-947. DOI:10.1111/j.1747-0285.2011.01232.x
[83]
C. Rosenbaum, P. Baumhof, R. Mazitschek, et al., Angew. Chem. Int. Ed. 43 (2004) 224-228. DOI:10.1002/(ISSN)1521-3773
[84]
C.Y. Man, I.T.F. Cheung, P.A. Cameron, T.H. Rainer, Ann. Emerg. Med. 49 (2007) 670-677. DOI:10.1016/j.annemergmed.2006.11.014
[85]
J.Y. Liou, D. Ghelani, S. Yeh, K.K. Wu, Cancer Res. 67 (2007) 3185-3191. DOI:10.1158/0008-5472.CAN-06-3431
[86]
A.B. Wu, Y.F. Xu, X.H. Qian, Monatsh. Chem. 141 (2010) 893-899. DOI:10.1007/s00706-010-0337-x
[87]
H. Yin, Y.F. Xu, X.H. Qian, et al., Bioorg. Med. Chem. Lett. 17 (2007) 2166-2170. DOI:10.1016/j.bmcl.2007.02.015
[88]
Z.G. Li, Q. Yang, X.H. Qian, Biorg. Med. Chem. 13 (2005) 3149-3155. DOI:10.1016/j.bmc.2005.02.045
[89]
Z.G. Li, Q. Yang, X.H. Qian, Tetrahedron 61 (2005) 6634-6641. DOI:10.1016/j.tet.2005.03.135
[90]
Z.G. Li, Q. Yang, X.H. Qian, Tetrahedron 61 (2005) 8711-8717. DOI:10.1016/j.tet.2005.06.097
[91]
Y.Q. Wang, X.H. Li, Q. He, et al., Eur. J. Med. Chem. 46 (2011) 5878-5884. DOI:10.1016/j.ejmech.2011.09.050
[92]
Y. Xiao, F.Y. Liu, X.H. Qian, J.N. Cui, Chem. Commun. (2005) 239-241.
[93]
F.Y. Liu, Y. Xiao, X.H. Qian, et al., Tetrahedron 61 (2005) 11264-11269. DOI:10.1016/j.tet.2005.08.106
[94]
Z.C. Zhang, L.J. Jin, X.H. Qian, et al., ChemBioChem 8 (2007) 113-121. DOI:10.1002/(ISSN)1439-7633
[95]
Z.C. Zhang, G.Y. Wu, F.B. Xie, et al., J. Med. Chem. 54 (2011) 1101-1105. DOI:10.1021/jm101181u
[96]
Y. Xiao, F.Y. Liu, Z. Chen, et al., Chem. Commun. 51 (2015) 6480-6488. DOI:10.1039/C4CC09846C
[97]
F.Y. Liu, X.H. Qian, J.N. Cui, et al., Bioorg. Med. Chem. Lett. 14 (2006) 4639-4644. DOI:10.1016/j.bmc.2006.02.016
[98]
Z.C. Zhang, Y.Y. Yang, D.N. Zhang, et al., Bioorg. Med. Chem. Lett. 14 (2006) 6962-6970. DOI:10.1016/j.bmc.2006.06.029
[99]
L.J. Xie, Y. Xiao, F. Wang, et al., Bioorg. Med. Chem. Lett. 17 (2009) 7615-7621. DOI:10.1016/j.bmc.2009.02.031
[100]
A. Gschwind, O.M. Fischer, A. Ullrich, Nat. Rev. Cancer 4 (2004) 361-370. DOI:10.1038/nrc1360
[101]
W. Zhou, W. Hur, U. McDermott, et al., Chem. Biol. 17 (2010) 285-295. DOI:10.1016/j.chembiol.2010.02.007
[102]
V. Knights, S.J. Cook, Pharmacol. Ther. 125 (2010) 105-117. DOI:10.1016/j.pharmthera.2009.10.001
[103]
Z. Chen, X. Wang, W.P. Zhu, et al., J. Med. Chem. 54 (2011) 3732-3745. DOI:10.1021/jm200258t
[104]
H. Yin, Y.F. Xu, X.H. Qian, Bioorg. Med. Chem. Lett. 15 (2007) 1356-1362. DOI:10.1016/j.bmc.2006.11.016
[105]
X.L. Li, Q.Q. Wang, Y. Qing, et al., Bioorg. Med. Chem. Lett. 18 (2010) 3279-3284. DOI:10.1016/j.bmc.2010.03.017
[106]
B. Sun, J. Liu, Y. Gao, et al., Eur. J. Med. Chem. 136 (2017) 603-618. DOI:10.1016/j.ejmech.2017.05.050
[107]
Z.H. Li, T.Q. Zhao, X.Q. Liu, et al., Eur. J. Med. Chem. 143 (2018) 1396-1405. DOI:10.1016/j.ejmech.2017.10.037
[108]
P.H. Li, H. Jiang, W.J. Zhang, et al., Eur. J. Med. Chem. 145 (2018) 498-510. DOI:10.1016/j.ejmech.2018.01.010
[109]
R.X. Rong, S.S. Wang, X. Liu, et al., Bioorg. Med. Chem. Lett. 28 (2018) 742-747. DOI:10.1016/j.bmcl.2018.01.008
[110]
J.W. Zhang, C. Wang, L. Zhang, et al., RSC Adv. 6 (2016) 71624-71627. DOI:10.1039/C6RA13226J
[111]
[112]
[113]
[114]
[115]
J. Yu, H. Bian, H. Man, et al., Dye. Pigment. 149 (2018) 851-857. DOI:10.1016/j.dyepig.2017.12.001