Co-Prodrugs of 7-Ethyl-10-hydroxycamptothecin and Vorinostat with in Vitro Hydrolysis and Anticancer Effects

7-Ethyl-10-hydroxycamptothecin (SN38) and vorinostat (SAHA) are quite promising combination therapy agents applied to the clinical treatment of cancer. In this study, we designed and synthesized a series of novel SN38-SAHA co-prodrugs, which were conjugated by four different amino acids including glycine, alanine, aminobutyric acid, and 6-aminocaproic acid. The hydrolytic reconversion rate to SN38 and SAHA critically depended on the carbon chain length, which were evaluated in PBS (pH 6.0/7.4) and plasma (human/mouse). With decreasing amino acid chain length, the hydrolytic reconversion rate increased gradually. The in vitro cytotoxicity test was evaluated by the sulforhodamine B (SRB) assay on the human lung adenocarcinoma cell line A549 and human colorectal cancer cell line HCT116. With the evaluation of stability and in vitro cytotoxicity, an appropriate linker was found, and the active drug can be released efficiently from compound 3a, which exhibited strong antiproliferative activity in A549 and HCT-116 cell lines correspondingly. These results indicated that the well-designed co-prodrug 3a and this kind of strategy can be a promising approach for anticancer therapy.


■ INTRODUCTION
7-Ethyl-10-hydroxycamptothecin (SN-38), as a highly active topoisomerase (Topo) I inhibitor, is the active metabolic product of its prodrug irinotecan (CPT-11) commercially available as Camptosar. 1,2 By combining with the Topo I-DNA covalent conjugate, the camptothecin analogue forms a ternary complex that blocks the repolymerization of the DNA strands during their replication leading to the cell apoptosis eventually. 3−5 Although SN-38 is about 100-to 1000-fold more potent than irinotecan, its clinical application is limited by its poor solubility and nonspecific toxicity. 1 To improve these deficiencies, many efforts have been made to develop prodrugs of SN-38 such as polymer-, albumin-, and immune conjugates, 6,7 and a number of these prodrugs have reached different stages of preclinical or clinical trials. 8−11 It has been suggested that cells arrested in G2/M by camptothecin were most sensitive to subsequent addition of histone deacetylase inhibitors (HDACIs). 12,13 The mechanism of this phenomenon is possibly caused by the HDACI-induced decreases in cyclin B levels of the antiapoptotic proteins. 14,15 In this regard, reduced expression of these pivotal antiapoptotic factors by HDACIs, which does not prevent tumor cell progression through the S phase, may increase the cytotoxicity of Topo I inhibitors.
HDACIs can weaken the interaction between DNA and histones by inducing accumulation of acetylated histones. The weakness leads to the loosening of the chromosome structure and disturbs transcriptional machinery, which could cause cell apoptosis ultimately. 16−19 Several HDACIs exhibit little or no anticancer effects in preclinical studies, and some are in clinical studies either as single agents 16,20,21 or in combination with conventional chemotherapy. Suberoylanilide hydroxamic acid (SAHA, vorinostat), which is an oral HDACI approved by the FDA as a treatment for cutaneous T-cell lymphoma (CTCL), has also shown strong anticancer effects in some sensitive cell lines. 22,23 Despite promising clinical effects reported, SAHA is not efficient for the treatment of solid cancers. 20,24 Therefore, SAHA has been used for drug combinations with other traditional chemotherapeutic drugs like gemcitabine and camptothecin against solid cancers to exert more cytotoxicity. 25 −27 In addition, SAHA has a tendency to hydrolyze to the corresponding carboxylic acid derivative quickly in plasma, which is completely inactive. 28,29 For this kind of situation, it is necessary to improve the stability of SAHA through structural modification or other means while in clinical use. To overcome these obstacles, SAHA has been designed in various forms, such as clickable pH-responsive prodrug, 30 selective enzymatic cleavable prodrug, 28 carbamate prodrug, 31 thiol-sensitive prodrug, 32 redox-responsive prodrug, 33 and so on.
Chemotherapy is one of the most commonly used treatments for both hematological diseases and solid tumors. However, it is usually found that chemotherapy cannot achieve expected therapeutic efficacy but can express high systemic toxicity. 34 Besides, drug resistance in tumors is a key reason for the low anticancer activity of many drugs. 35 Therefore, combination therapy has been adopted to improve the therapeutic efficacy of drugs while simultaneously reducing drug-related side effects by lowering the respective dosage.
Employing multiple drugs possessing different mechanisms of action could induce cell death more effectively than single medication. 36 As mentioned previously, SN38 and SAHA have been shown synergistic antitumor effects in many works. 27,37 Nevertheless, the clinical application was limited by the different administrations of SN38 and SAHA. SN38 is metabolized from irinotecan after intravenous injection, and SAHA is an oral preparation. 22 The different administration of SN-38 and SAHA may cause complex interaction. A promising strategy is conjugating SN38 and SAHA to generate a coprodrug that can hydrolyze to the two active compounds at the same time under certain conditions ( Figure 1). The mechanism of release has been fully studied and characterized by other groups (Figure 2). 38 The co-prodrug contains a specialized nontoxic linker to hide the activity of the parent molecule temporarily.
In this manner, we can enhance the stability of SAHA and obtain synergistic therapy after cellular uptake of the SN38-SAHA co-prodrug. The hydroxamic acid group is essential for SAHA, which can inhibit the activity of HDAC by binding to the active site Zn 2+ . 22,39,40 The design of the co-prodrug is illustrated in Figure 1, which blocks the active parts of both SAHA and SN-38. Four SN38-SAHA co-prodrugs were synthesized with different aliphatic chain lengths (glycine (Gly) (3a), alanine (Ala) (3b), aminobutyric acid (Abu) (3c), and 6-aminocaproic acid (Eaca) (3d)). Our hypothesis of the release mechanism is that the carbonic ester between the amino acid and SAHA could be simply hydrolyzed or enzymatically hydrolyzed at first.
Subsequently, the SN38 release mechanism ( Figure 2) proceeds via nucleophilic terminal amine attack at the SN38 Ering carbonyl lactone producing a circular intermediate. Finally, the SN38 was produced by the intermediate through the attack of the hydroxide ion in alkaline conditions. 41 With increasing amino acid chain length, the formation of the circular intermediates becomes more difficult and the SN-38 release of 3b−d mainly relies on the enzymatic hydrolysis. Therefore, the reconversion rate of SN38 and SAHA may increase with decreasing chain length of the amino acid.
Generally, the novel concept co-prodrug was presented there, and four SN38-SAHA co-prodrugs were synthesized. Briefly, we conjugated the hydroxamic acid of SAHA and the 20-OH of SN-38 through a series of amino acids to form inactive co-prodrugs. Compared with combination therapy, such co-prodrugs can improve the stability of SN38 and SAHA, increasing the anticancer effects while the co-prodrug releases two active drugs inside the cancer cells simultaneously. Co-prodrug reconversion was studied at pH 6.0/7.4 phosphate buffer solution (PBS) and human/mouse plasma. The metabolic transformation of the co-prodrugs to the drugs was catalyzed by hydrolysis and enzymolysis. 42 Cytotoxicity was evaluated in HCT116 and A549 cell lines using the SRB assay. These results suggested that 3a, which is conjugated by glycine, was the best candidate.

■ RESULTS AND DISCUSSION
Synthesis of the SN38-SAHA Co-Prodrugs. The 3a−d co-prodrugs were synthesized through a four-step procedure with CH 2 Cl 2 as the solvent. First, the 10-OH of SN-38 was protected by the butoxycarbonyl group (Boc), and then the 10-Boc-SN38 was coupled to four different Boc-protected amino acids resulting in compounds 1a−d (Scheme 1). Cleavage of the Boc group from compounds 1a−d was performed using 30% TFA in CH 2 Cl 2 , resulting in 2a−d. Finally, we activated the SAHA to a reactive intermediate by reacting with CDI and then using the terminal amine of 2a−d to attack the ring of the SAHA reactive intermediate produced compounds 3a−d. 39 The progress of the reaction was monitored by thin layer chromatography (TLC). All reactions were processed with good yields and purified to high purity. Each compound and the final product were confirmed by 1 H NMR and mass spectrometry.
In Vitro Degradation and Release of SN38-SAHA Co-Prodrugs in PBS. The standard curves of various compounds were prepared for the convenience of calculation. The hydrolysis studies of 3a−d were first performed in PBS at pH 6.0/7.4 at 37°C for a period of 24 h ( Figure 3). The disappearance of 3a−d and the formation of SN38 and SAHA were quantified using HPLC. The percentages of remaining 3a−d and converted SN38 and SAHA were plotted as a function of time. The cumulative release and formation curves in Figure 3A−C showed that 3a has the fastest rate to reconversion of SN38 and SAHA. The longer the chain length of the co-prodrugs, the faster the hydrolysis of the co-prodrugs and the formation of SN38 were. The half-lives (t 1/2 ) of 3b−d were found to be beyond 24 h, while the t 1/2 of 3a was merely 2 h at pH 7.4. 24 h later, the formation percentages of SN38 were 26.3 and 3.4% corresponding to 3a and 3b, respectively, while 3c and 3d almost have no SN38 generated; besides, the formation of SAHA decreased with increasing carbon chain length. In pH 6.0 PBS, the t 1/2 of 3a−d were found to be far beyond 24 h, and the reconversions of SN38 and SAHA were

ACS Omega
Article little either as illustrated in Figure 3D. As our hypothesis before, increasing the chain length of amino acid generally increases the co-prodrug reconversion time because of the difficulty to form the cyclic intermediate. Moreover, the coprodrugs were more stable in weak acid conditions.
In Vitro Degradation and Release of SN38-SAHA Co-Prodrugs in Plasma. The hydrolysis studies of 3a−d were then performed in human plasma and mouse plasma at 37°C ( Figure 4). The disappearance of 3a−d and the formation of SN38/SAHA were quantified using HPLC. The percentages of the remaining 3a−d and converted SN-38/SAHA were plotted as a function of time. The co-prodrugs had a short t 1/2 in both human and mouse plasma. The t 1/2 of 3a−d were found to be between 0.5 and 2 h, and the highest reconversions of SN38 were 37.1, 19.2, 22.9, and 0.0%, respectively, in mouse plasma. In contrast, the t 1/2 of 3a−d were extended in human plasma compared to mouse plasma. The t 1/2 of 3a−d in human plasma were 1, 3, 5, and 5 h, and the highest reconversions of SN38 were 39.8, 20.5, 16.3, and 3.2%, respectively. The HPLC peak of SAHA was according with some substance in the plasma so that the conversion rate of SAHA had not listed out. Compared to PBS stability experiments, the rates of degradation were much faster in plasma, suggesting that the enzymatic hydrolysis is an important pathway here. Besides, the formation of SN38/SAHA seemed easier and faster in plasma, which means that the co-prodrug may release the active drug more completely in vivo.

ACS Omega
Article Cytotoxicity of 3a−d Co-Prodrugs. After assuring that SN38-SAHA co-prodrugs could hydrolyze efficiently, the in vitro biological efficacy of 3a−d was evaluated in A549 and HCT116 by the SRB assay. The IC 50 values of 3a−c were more potent than SAHA but less potent than SN38, and the IC 50 value of 3d fell far below 3c on A549 and HCT116. As expected, the IC 50 of 3a was almost equivalent to that of the SN38, suggesting that the toxicity mainly depended on the stability (i.e., co-prodrug hydrolysis and SN38/SAHA formation). Furthermore, the cytotoxicity of 3a−d increased nearly proportionally as the amino acid carbon chain length decreased, and the order of the co-prodrugs tested was 3a > 3b > 3c > 3d (Table 1). Co-prodrugs have slow degradation processes to release active compounds. Therefore, the most active synthesized compound 3a was less cytotoxic than SN38. We could possess a longer action time than the original drugs in this way. Besides, we intended to increase the targeting ability by the hydroxamic acid group of SAHA.

■ CONCLUSIONS
In summary, we synthesized the SN38-SAHA co-prodrugs with four different chain lengths of amino acid, and the strategy was a promising approach for anticancer therapy. After SAHA was released readily, the rest of the prodrugs containing terminal amines occurred, and essentially complete hydrolysis of SN38

ACS Omega
Article was through the formation of the cyclic intermediate or the esterase hydrolysis directly. The hydrolysis rule of 3a−d was found to be pH-dependent and approximately proportional to the amino acid carbon chain length. Among these co-prodrugs, compound 3a showed similar antiproliferative activity to SN38. The results of in vitro stability confirmed the complete conversion from compound 3a to SN-38 in different plasma. Therefore, this kind of strategy has potential in the cancer therapy for different drug combinations. Considering the effect of the co-prodrug on the cytotoxicity test, animal studies on mice bearing lung or colorectal cancer will be conducted in future.

■ EXPERIMENTAL SECTION
Materials and Methods. 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded on a Bruker DRX-400 MHz spectrometer (400 and 101 MHz, respectively) using CDCl 3 or DMSO-d 6  Synthesis of the SN38-SAHA Co-Prodrugs. Synthesis of 10-Boc-SN38. SN-38 (3 g, 7.65 mmol) was dispersed in CH 2 Cl 2 (300 mL), and then (Boc) 2 O (2 g, 9.18 mmol) and pyridine (1.85 mL, 22.9 mmol) were added sequentially. After stirring for 12 h at 25°C, the reaction mixture was washed with a saturated aqueous solution of NaHCO 3 , water, and brine. The mixture was then dried over Na 2 SO 4 and filtered, and the filtrate was concentrated under reduced pressure to give 10-Boc-SN38 (3.7 g, 7.5 mmol, 98%) as a pale-yellow solid.
General Method for Synthesis of 1a−d. 10-Boc-SN38 (0.5 g, 1.01 mmol) was dissolved in CH 2 Cl 2 (10 mL), the reaction mixture was cooled to 0°C, and Boc-glycine (or Boc-groupprotected alanine/4-aminobutyric acid/6-aminocaproic acid) (2.02 mmol), DIC (0.31 mL, 2.02 mmol), and DMAP (0.15 g, 1.21 mmol) were added subsequently. After stirring for 15 h at 25°C, the reaction mixture was diluted with CH 2 Cl 2 and washed with 0.1 M hydrochloric acid, water, and brine. The mixture was then dried over Na 2 SO 4 and filtered, and the filtrate was concentrated under reduced pressure. The residue was purified using silica gel column chromatography with CH 2 Cl 2 and MeOH to give 1a−d (0.69 mmol, 68%) as yellow solids.
In Vitro Degradation and Release of SN38-SAHA Co-Prodrugs in PBS. SN38-SAHA co-prodrug solutions with a concentration of 100 μM/L were diluted to 0.05 μM/L with 10 mM PBS buffer (pH 7.4/6.0). After incubation at 37°C for 0.5, 1, 2, 3, 4, 6, 12, 24, and 48 h, the areas of SN38-SAHA coprodrugs and other species were detected by HPLC as well. Then each compound was calculated by the concentration− area standard curve, which was obtained by the external standard method.
In Vitro Degradation and Release of SN38-SAHA Co-Prodrugs in Plasma. SN38-SAHA co-prodrug solutions with

ACS Omega
Article a concentration of 100 μM/L were diluted to 0.1 μM/L by plasma (human/mouse). After incubation at 37°C for 0.5, 1, 2, 3, 4, and 6 h, 50 μL of reaction mixture was quenched with 50 μL of cold ACN and centrifuged at 30000 rpm × 5 min. The supernatant (50 μL) was detected by HPLC. Then each compound was calculated by the concentration−area standard curve, which was obtained by the external standard method.
Cytotoxicity of 3a−d Co-Prodrugs. The in vitro biological efficacy of 3a−d was evaluated in A549 and HCT116 cell lines. Briefly, the cells were seeded into 96-well microtest plates (1.0 × 10 4 cells per well) in 100 μL of culture medium. Cells were treated in triplicate with gradient concentrations of test drugs and incubated at 37°C for 72 h. The cell viabilities on the cell lines were determined by the SRB assay subsequently. The drug concentration required for 50% growth inhibition (IC 50 ) of tumor cells was determined from the dose−response curves.