Synthesis and Biological Evaluation of a Novel Pentagastrin-Toxin Conjugate Designed for a Targeted Prodrug Mono-therapy of Cancer

A novel carbamate prodrug 2 containing a pentagastrin moiety was synthesized. 2 was designed as a detoxified analogue of the highly cytotoxic natural antibiotic duocarmycin SA (1) for the use in a targeted prodrug monotherapy of cancers expressing cholecystokinin (CCK-B)/gastrin receptors. The synthesis of prodrug 2 was performed using a palladium-catalyzed carbonylation of bromide 6, followed by a radical cyclisation to give the pharmacophoric unit 10, coupling of 10 to the DNA-binding subunit 15 and transformation of the resulting seco-drug 3b into the carbamate 2 via addition of a pentagastrin moiety.


Introduction
One of the major problems in the chemotherapy of cancers is the usually low differentiation between normal and malignant cells by the known antiproliferating agents, resulting in severe side effects. Several approaches have been developed to overcome this problem like the antibody-directed enzyme prodrug therapy (ADEPT) [1,2] and the prodrug monotherapy (PMT) [3]. Whereas in ADEPT artificial antibody-enzyme conjugates are needed for targeting tumor cells, in PMT specific endogenous enzymes or receptors overexpressed in cancerous tissue are addressed to allow a selective killing of tumor cells. In both approaches, a relatively untoxic prodrug is used, which is then selectively converted into the corresponding cytotoxic drug in the cancer tissue; however, for PMT the prodrug is linked to a ligand which allows a targeting of cancer cells. Among these ligands small peptides play an important role having a low immunogenicity as well as high specifities and affinities to certain receptors which are overexpressed on certain tumour cells [4]. Some of these peptides that are already successfully applied in cancer therapy, belong to the gastrin family. For example, radiolabeled gastrin derivatives have shown a high therapeutic and diagnostic potential in targeting cholecystokinin (CCK-B)/gastrin receptor expressing tumors [5]. In addition, a gastrin derivative was linked to a triazene alkylating agent [6]. However, the observed receptor-mediated cytotoxicity of this conjugate was quite low. Better results were obtained with heptagastrin linked to an ellipticine derivative [7]. A high receptor-mediated cytotoxicity could be achieved with the anthracyclines daunorubicin, doxorubicin and 2-pyrrolinodoxorubicin as well as other cytotoxic agents like melphalan, cisplatin or methotrexate coupled to peptides of the LHRH [4,8], bombesin [4,8a,9], somatostatin [4,8a,10] and neuropeptide Y [4,11] type. Recently, we have developed the pentagastrintoxin conjugate 3a containing a seco-duocarmycin SA derivative (Scheme 1) [12]. Here, we report the synthesis of a novel pentagastrin conjugate 2 for the use in a targeted tumor therapy, which has the advantage over normal gastrin-toxin conjugates that a prodrug is used instead of a toxic drug (Scheme 1). In this concept, the pentagastrin moiety should serve not only as a targeting ligand for CCK-B/gastrin receptors, but also as a detoxifying unit. Thus, the corresponding drug 4b, which again is an analogue of the naturally occuring antibiotic (+)-duocarmycin SA (1) with an IC 50 value of 10 pM (L1210) [13], should be formed via 3b inside the tumor cells by cleavage of the carbamate by lysosomal enzymes after endocytosis. The antiproliferative effect of 1 and its analogues such as the CBI-drugs 4 derive most probably from a selective alkylation of N-3 of adenine in DNA by nucleophilic attack at the spirocyclopropyl-cyclohexadienone moiety as the pharmacophoric group [14]. Since we have previously shown that the formation of a drug as 4b from a seco-drug as 3b is a very fast process and that the blocking of the phenolic hydroxyl group of 3b allows a very strong reduction of its cytotoxicity, we used the seco-drug 3b as substrate for the conjugation with the pentagastrin moiety, performing the connection to the phenolic hydroxyl group via a carbamate moiety [1b,1c,15].
In our approach we did not employ the whole heptadecapeptide gastrin but the shorter β-alanine modified pentagastrin, because its β-Ala-Trp-Met-Asp-Phe-NH 2 sequence representing the C-terminal amide of the natural peptides restores the biological activity of gastrin in a comparable order of magnitude [4e,16]. As a consequence, the seco-duocarmycin moiety 3b had to be attached via the Nterminal amino functionality of pentagastrin using a carbamate. Such a carbamate substructure exists also in KW-2189 (5) (Figure 1) [17], an agent already investigated in clinical trials, and in several other anticancer agents [18].  For the formation of the carbamate moiety, we envisaged an addition of an isocyanate to the secodrug 3b. The TMSE ester moiety was introduced to allow a better comparison with the already prepared pentagastrin-conjugate 3a. Moreover, the handle could be used for the introduction of a fluorescence dye to allow an investigation of the mode of action of such a compound employing a confocal laser scanning microscope.

Synthesis
As starting material for the preparation of 2 we employed the known aminonaphthalin 6 [12]. 6 was converted into TMSE ester 7 in 56 % yield by a palladium-catalyzed carbonylation reaction using a CO atmosphere (1 bar) and Mo(CO) 6 as additional CO source [19] in a mixture of 2-(trimethylsilyl)ethanol and DMF (Scheme 2). The moderate yield of 56% of this carbonylation reaction might be due to the relatively high electron density of 6. Nevertheless, 6 had to be used in the carbonylation reaction as the Curtius rearrangement of the corresponding acid to the protected naphtholamine could not be achieved after the introduction of the TMSE ester moiety. Iodination of 7 employing NIS [20,21] with TsOH·H 2 O as catalyst followed by N-alkylation of the formed 8 with 1,3-dichloropropene and subsequent radical cyclization [22] using the untoxic tris-(trimethylsilyl)-silan (TTMSS) [23] as hydride source and AIBN as radical starter provided seco-CBI derivative 10 in 65 % yield over three steps.
Scheme 2. Synthesis of seco-CBI compound 10. a) Mo(CO) 6 , 1 bar CO, 5 mol% Pd(PPh 3 ) 2 Br 2 , 20 mol% dppf, nBu 3 N, TMSEOH, DMF, 120 °C, 7 h, 56%; b) NIS, TsOH·H 2 O, THF/MeOH, 50 °C, 1 h, 73%; c) NaH, 1,3-dichloropropene, DMF, 20 °C, 13.5 h, 97%; d) HSi(SiMe 3 ) 3 , AIBN, benzene, reflux, 2 h, 92%. In order to connect 10 with the peptide unit, we used the isocyanate 13 containing an ester moiety. This was first reacted with the phenolic hydroxyl group and then bound to the peptide via an amide linkage. The required isocyanate 13 was prepared as follows: first, β-alanine (11) was converted into the corresponding benzyl ester hydrochloride 12 employing TMSCl and benzylic alcohol [24]. Then, the isocyanate moiety was introduced using solid and thus easy to handle triphosgene in refluxing toluene to give 13 in 87 % yield over two steps (Scheme 3).  Hence, the synthesis of carbamate prodrug 2 was completed in seven further steps (Scheme 4). After deprotection of the secondary amino functionality in 10 under acidic conditions in an aqueous HCl/EtOAc mixture with Et 3 SiH as cation scavenger [25], the obtained hydrochloride salt 14 was directly coupled with the DNA-binding subunit TMI-CO 2 H (15) to give 16 in 43 % yield over two steps. Then, the benzyl ether moiety in 16 was cleaved by transfer hydrogenolysis with an aqueous ammonium formate solution and palladium on charcoal as the catalyst [26] to yield phenol 3b which was subsequently coupled with isocyanate 13 to afford carbamate 17 in a very good yield of 83 % over two steps. The benzyl ester in 17 was cleaved again by using transfer hydrogenolytic conditions to give 18 in 90 % yield. This reaction had to be carefully monitored by TLC as the carbamate was sensitive to these conditions. Finally, carboxylic acid 18 was treated with HOSu/EDC·HCl and the resulting active ester 19 directly coupled with the fully unprotected tetrapeptide 20 [27] to yield carbamate prodrug 2 in 57 % (83 % based on recovered starting material) over two steps.

In vitro cytotoxicity tests
The in vitro cytotoxicity assays were carried out in duplicate with CCK-B/gastrin-receptor positive cells of the human pancreatic cell line MIA PaCa-2 and CCK-B/gastrin-receptor negative cells of the human bronchial carcinoma cell line A549 as control in six multiwell plates with concentrations of 10 2 , 10 3 and 10 4 cells per cavity. Incubation with various concentrations of the seco-drug 3b and the prodrug 2 was performed in ultraculture medium (Table 1).

Table 1.
In vitro cytotoxicity of prodrug 2 and of seco-drug 3b against CCK-B/gastrinreceptor positive cells of the human pancreatic cell line MIA PaCa-2 and CCK-B/gastrinreceptor negative human bronchial carcinoma cells (A549). Cells were exposed to various concentrations of the test substance for 24 h at 37 °C; after 10 days of incubation following the exposure to the substance, clone formation was compared to an untreated control assay and the relative colony-forming rate was determined. IC 50 is the drug concentration required for 50% growth inhibition of target cells. 0.31 0.14 Prodrug 2 shows the same cytotoxicity as its corresponding seco-drug 3b in the cell culture assays using the CCK-B/gastrin-receptor positive cell line (MIA PaCa) and the CCK-B/gastrin-receptor negative cell line (A549). Thus, the obtained IC 50 -values are almost identical in these four experiments. This indicates that prodrug 2 seems not to be stable under the used cell culture conditions. In fact, HPLC-MS-measurements revealed a decomposition of prodrug 2 under loss of the targeting pentagastrin moiety thereby forming the corresponding seco-drug 3b.

Compound
We suppose that the unstability of the carbamate moiety can be traced back to the hydrogen atom at its nitrogen which in turn is part of the β-alanine moiety of pentagastrin. We therefore plan to replace the hydrogen by a carbon moiety though it is not known whether such a modification of the pentagastrin would interfere with the binding of the conjugate to the corresponding CCK-B/gastrinreceptor.

β-Alanine benzyl ester hydrochloride (12):
A magnetically stirred suspension of β-alanine (11) (1.00 g, 11.2 mmol) in benzylic alcohol (56.0 mL, 58.0 g, 539 mmol) was treated dropwise over a period of 10 min with trimethylsilylchloride (3.6 mL, 3.0 g, 28 mmol) and stirring continued for a further 15 h at 20 °C. The resulting clear solution was poured into Et 2 O (600 mL), the precipitate collected by filtration and washed with Et 2 O (100 mL). Drying of this material under reduced pressure gave hydrochloride 12 (

3-Isocyano-propionic acid benzyl ester (13):
A magnetically stirred suspension of hydrochloride 12 (2.13 g, 9.88 mmol) in toluene (15 mL) was treated with triphosgene (2.93 g, 9.88 mmol) and heated to reflux for 7.5 h (end of HCl-evolution). The resulting solution was concentrated under reduced pressure to afford isocyanate 13 (1.99 g, 98 %) as yellow liquid which was used for the next reaction without further purification.