Synthesis of pyrrolocarbazoles with N -substituted alkynyl-, alkylcyano-and alkylhydroxyl-groups

Due to their involvement in almost all stages of cellular life, kinase biomolecular catalysts have been linked to cancer development and, thus, remain attractive drug targets for cancer therapeutics. 6-(3 ꞌ -Hydroxypropyl)-, 6-(2 ꞌ -hydroxyethyl)-, 6-(2 ꞌ -propynyl)- and 6-(3 ꞌ -propanenitrile)-pyrrolo[3,4-c ]carbazole-1,3(2 H ,6 H )-diones were synthesized as potential small molecule EGFR kinase inhibitors. The pyrrolocarbazole compounds were synthesized by way of a Diels-Alder approach involving N -alkylated 2-vinyl-1 H -indole and maleimide as starting materials followed by aromatization with MnO 2 .


Introduction
Over the past three decades, extensive research efforts have contributed to rapid developments in the field of oncology.Despite the apparent progress, cancer continues to be a worldwide leading cause of death.The need to develop new and less toxic treatments against a disease with the rather frustrating ability to remodel itself as drug-resistant variants is, thus, as important as ever.
The reversible phosphorylation of proteins is arguably one of the most general regulatory strategies adopted by eukaryotic cells and represents a key step in many crucial cellular processes.In this regard, protein kinases are enzymes that promote phosphorylation, i.e., the transfer of a phosphate group from ATP to a substrate protein.Due to the central involvement of kinases in almost all stages of cellular life (including growth factor signaling, cell cycle control, apoptosis and angiogenesis), these biocatalysts have been linked to cancer development and thus remain attractive drug targets for cancer therapeutics.The history concerning the development of kinase inhibitors has enjoyed much success; however, fundamental challenges, such as the lack of efficiency, drug resistance due to key amino acid mutations and inhibitor selectivity, persist.The development of effective long-term cancer treatments, including those which involve kinase inhibition, thus remains a pursuit of many researchers.
One of the most commonly selected kinase families targeted for the development of cancer therapeutics has been the receptor tyrosine kinases (RTKs), which include the epidermal growth factor receptors (EGFRs), the vascular endothelial growth factor receptors (VEGFRs) and the platelet-derived growth factor receptors (PDGFRs). 1Many of these cell-surface receptors are known to be mutated or overexpressed in cancer systems, which makes them attractive candidates as targets.For this particular project, we decided to focus on the EGFR family which consists of EGFR, human EGRF-related 2 (HER2) and the kinase-impaired HER3 and HER4. 2 EGFR itself has been the target of many successful small-molecule drugs, including erlotinib, gefitinib, afatinib, the more recent osimertinib, and the more experimental brigatinib and icotinib.Even for the more recent compounds, the development of drug-resistant cancer cells is a serious limitation (see for instance, the exon 20 C797S mutation experienced by the 3 rd generation inhibitor, osimertinib). 3n terms of finding inspiration for new scaffolds which might provide the basis for kinase inhibitors with different and, hopefully, favourable characteristics, Nature continues to be one of the best sources of ideas. 4ith this in mind, it was soon realized that staurosporine 1 is in fact a natural, potent kinase inhibitor, initially isolated from the bacterium Streptomyces staurosporeus, and has widely served as a structural muse for the design of protein kinase inhibitors with the overall aim of improved specificity and selectivity.][7][8][9][10] Staurosporine-inspired drug candidates have been in vogue, see for instance the staurosporin-inspired midostaurin 2 (Rydapt  , in clinic as a tyrosine kinase 3 (FLT3) inhibitor), CEP-2563 3 (phase 1) and endotecarin 4 (phase 3) depicted in Figure 1. 10 The bisaryl maleimide derivative enzastaurin 5 could be considered an "open" form of staurosporine, but, unfortunately, it failed its phase III lymphoma clinical trial.Structurally simplified staurosporine-inspired pyrrolocarbazoles have also been considered as possible kinase inhibitors -examples include Chk 1 inhibitors 6 and PARP 1 inhibitors 7 -and these simplified staurosporine motifs were the basis for the design of new potential inhibitors described in this work.It should be noted that the pyrrolocarbazole core, apart from its ubiquitous role in natural products, 11 has seen frequent prior application in medicinal chemistry, 12 with particular emphasis as kinase inhibitors.] Similar scaffolds have also been identified as PARP-1 inhibitors. 24n this project we envisaged the development of potential kinase inhibitors that would selectively suppress EGFR, an important therapeutic target (for other collaborative studies from our group involving this kinase see [25][26][27] ).Exploiting the attractive features demonstrated by the natural product staurosporine, the design considered was based on a staurosporine scaffold.It was our intention that the pyrrolocarbazole scaffold 8 could act as a driving portion and present a suitable platform to incorporate potentially electrophilic warheads (R) at a proper trajectory (Figure 2).Notably, the scaffold displaying the warhead at a particular distance and orientation, could result in a covalent interaction to cysteine 797 in the kinase-active site, a strategy utilized before in our research. 26,28n terms of the synthetic strategy towards the desired substituted scaffolds, a Diels-Alder-oxidative aromatization approach was utilized.(For an excellent overview of the many synthetic approaches to the carbazole scaffold, please refer to the review by Knölker and co-workers 11 ).Approach A involved generating the pyrrolo [3,4-c]carbazole-1,3(2H,6H)-dione first, followed by selective acylation/alkylation of the carbazole nitrogen atom as shown in Figure 2.Alternatively, approach B would generate the desired compounds with the desired N-functionalizations already in place on the 2-vinylindole precursor 9b.It should be noted that the Diels-Alder/aromatization strategy has been effectively utilized before to efficiently deliver substituted pyrrolocarbazoles (see the examples listed in the following references [13][14][15][17][18][21][22]24 and the following examples which include related modifications [29][30][31][32] with respect to the 2-vinylindole motif 33 ).

Results and Discussion
The initial strategy focused on the synthesis of the known pyrrolo [3,4-c]carbazole-1,3(2H,6H)-dione scaffold 15 in order to use this compound in a divergent approach to obtain a small library of alkylated N-carbazole derivatives.To this end, commercially available ethyl 1H-indole-2-carboxylate 10 was converted into 1Hindole-2-carbaldehyde 12, via alcohol 11, through a reduction (LiAlH 4 )-oxidation (MnO 2 ) sequence.A Wittig reaction (MePPh 3 Br with nBuLi) involving carbaldehyde 12 readily afforded 2-vinyl-1H-indole 13, which gratifyingly underwent a Diels-Alder reaction as a neat mixture at 170 C with maleimide to afford the fused indole 14a in quantitative yield.This compound was then oxidized into the fully-aromatized substituted carbazole scaffold 15 with DDQ in DMSO at 50 C in 50% yield (Scheme 1).
Initial attempts to react 15 with acryloyl chloride in DMF at 0 C to obtain N-acylated 16, and even after heating to 80 C under N 2 for 48 h, did not indicate any evidence of N-substituted product 16 formation by TLC.In addition, treatment of 15 with NaH in DMF at RT under N 2 afforded a purple solution which turned a yellow color upon the drop-wise addition of acryloyl chloride with stirring at RT for 4 days, but still gave no new products (TLC).To address the regioselectivity issue between the two nitrogen atoms in 15, the known N-Boc-protected maleimide 34 was reacted with diene 13 at 70 C for 30 min to afford the expected adduct 14b in 59% yield.Attempts to react this latter adduct with acryloyl chloride, in DMF containing DIPEA (-5 to 60 C for 2 days), only led to the cleavage of the Boc-protecting group.Aromatization of 14b with either DDQ or MnO 2 also resulted in the Boc group being cleaved.The unsuccessful work involving the Boc-protected compounds is not described further.An alternaive method at accessing N-substituted carbazoles involved the initial introduction of substituents on the nitrogen atom of the indole ring by the synthesis of N-alkylated 2-vinylindoles.These could, subsequently, be utilized as the starter dienes in the critical Diels-Alder cyclization step.To this end, ethyl and methyl 1H-indole-2-carboxylates 10 and 18, respectively (the latter readily obtained from carboxylic acid 17), were dissolved in DMF to which NaH was added, followed by (2-bromoethoxy)(tertbutyl)dimethylsilane (19)  35 or (3-bromopropoxy)(tert-butyl)dimethylsilane (20) 36 to afford 22 and 21, respectively, in reasonable yields (Scheme 2).Reduction of 22 and 21 was readily achieved by the use of LiAlH 4 in THF at 0 C to afford 23a and 23b in excellent yields of 97% and 88%, respectively.It should be noted, however, that the reaction temperature needed to remain below room temperature in order to retain the ethoxy and propoxy silyl groups; at temperatures above room temperature these groups were cleaved.
MnO 2 was found to be the best oxidizing agent for the conversion of 23a and 23b into substituted 1Hindole-2-carbaldehydes 24a and 24b in respectable yields of 83% and 99%, respectively.Prior activation of the MnO 2 was necessary and achieved by placing a beaker of MnO 2 in an oven at 120 °C for 24 h.The oxidant, after the oxidation procedure, was readily removed by filtration and it was found that the aldehyde products were sufficiently pure to be used for conversion into the respective vinyl analogues without further chromatographic purification.
The Wittig protocol for conversion of the aldehydes into their corresponding vinyl analogues involved the initial generation of the methylene ylide, by treatment of MePPh 3 Br with nBuLi in dry THF at 0 C, followed by the drop-wise addition of carbaldehydes 24a and 24b to afford the two alkylated 2-vinyl-1H-indoles 25a and 25b in yields of 79% and 55% respectively.The key Diels-Alder reaction to produce the additional 6-membered ring of the desired pyrrolocarbazole scaffold involved the cycloaddition of compounds 25a and 25b with maleimide.The reagents were heated together as a neat mixture and, after melting, they reacted to produce a solid adduct, which, in both cases, was purified chromatographically to afford the desired products 26a and 26b in reasonable yields of 78% and 91%, respectively, based on recovered starting material (brsm) (Scheme 3).
The target pyrrolocarbazoles 29a and 29b were prepared by a protocol involving the same steps, but in the opposite order, since it was found that this alternative sequence produced the best overall yields.Thus for pyrrolocarbazole 29a, adduct 26a was firstly treated with TBAF at 0 C to remove the TBDMS protecting group to produce 27 in 86% yield, followed by the MnO 2 oxidation in dioxane under reflux to form the desired aromatized product 29a in a yield of 53%.On the other hand, 29b was readily obtained by first oxidizing 26b with MnO 2 in refluxing dioxane to form the aromatized carbazole 28 in 63% yield, followed by cleavage of the TBDMS protecting group with TBAF in THF at RT for 30 min, to afford the longer chain product 29b in 65% yield.
The propargyl group was introduced on the nitogen atom of the previously synthesized 2-vinyl-1Hindole ( 13) by a nucleophilic substitution reaction employing Cs 2 CO 3 as a base, as per a literature procedure, 37 to afford 30a in a moderate (45%) yield.We found the best way to introduce the cyanoethyl group on the nitrogen atom of indole 13, in order to obtain 30b, involved a Michael addition between the 2-vinyl-1H-indole (13) and acrylonitrile in the presence of DBU in acetonitrile, at RT for 8 h, which gave the desired product 30b in 84% yield [based on recovered starting material (brsm)], as shown in Scheme 4.

Scheme 3. Synthesis of the N-alkylated pyrrolocarbazoles 29a and 29b.
A Diels-Alder reaction between compound 30a and maleimide was successfully achieved in the same manner as described previously, but at a slightly lower reaction temperature of 160 C.It should be noted that cycloadduct 31a was obtained in an acceptable yield of 73% (brsm) and, thus, in sufficient amounts to carry on with the syntheses.For the cycloaddition between diene 30b and maleimide, it was found that the addition of a Lewis acid, SnCl 2 , produced the best results.Even under the best conditions available, however, unreacted starting material was still present which, fortunately, could easily be separated by chromatography.In this latter case, the cycloadduct 31b was obtained in an acceptable yield of 71% (calculated brsm).
Finally, aromatization of the cyclohexene ring of 31a and 31b, using an excess of MnO 2 in refluxing dioxane, afforded the desired substituted pyrrolocarbazoles 32a and 32b in yields of 72% and 30%, respectively.It should be noted that pyrrolocarbazoles 32a and 32b were found to be quite insoluble in most laboratory solvents which precluded their absolute purification by chromatography.Consequently, a trituration protocol for 32a and a recrystallization from DMF for 32b were employed to obtain the final compounds in sufficient purity.The recrystallization from DMF, unfortunately, afforded 32b in a rather low yield.

Conclusions
A set of four carbazole N-substituted pyrrolocarbazoles bearing the 6-(3ꞌ-hydroxypropyl)-, 6-(2ꞌ-hydroxyethyl)-, 6-(2ꞌ-propynyl)-and 6-(3ꞌ-propanenitrile)-fragments on their nitrogen atoms was prepared for evaluation as potential EGFR kinase inhibitors using Diels Alder cycloadditions and MnO 2 oxidative aromatizations to good effect.These compounds will be utilized as part of screening libraries for the identification of potential lead compounds in future high-throughput screenings.

Experimental Section
General.Purification of solvents and reagents: Ethyl acetate and hexane used for chromatographic purposes were distilled by means of conventional distillation procedures.Solvents used for reaction purposes were dried over an appropriate drying agent and distilled under nitrogen gas.Tetrahydrofuran, 1,4-dioxane, and diethyl ether were distilled from sodium wire using benzophenone as an indicator.Dimethylformamide and acetonitrile were distilled from calcium hydride.
Chromatography: Separation of compounds by column chromatography was performed using Merck silica gel (particle size 0.063-0.200mm).Thin layer chromatography was performed using Merck silica gel 60 F 254 coated on aluminium sheets.Compounds on TLC plates were viewed under UV light.Spectroscopic and physical data: 1 H and 13 C NMR spectra were recorded on a Bruker ADVANCE 300 or Varian Gemini-300 spectrometer ( 1 H NMR at 300 MHz and 13 C at 75 MHz).A Varian VXR-400 machine ( 1 H NMR at 400 MHz and 13 C NMR at 101 MHz) or 600 MHz Varian Unity Inova ( 1 H NMR at 600 MHz and 13 C NMR at 151 MHz) were also utilized.Spectra were recorded in deuterated chloroform (CDCl 3 ) and DMSO (DMSO-d 6 ) as indicated.Infra-red spectra were recorded using a Bruker Tensor 27 spectrometer.Melting points were measured using a Stuart SMP10 melting point machine.Mass spectra were recorded on a Thermo Electron DFS Magnetic Sector Mass Spectrometer (E.I. mode).Other general procedures: Most reactions were carried out under nitrogen or argon and reaction vessels were dried in an oven.Removal of solvent in vacuo refers to removal of the solvent using a rotary evaporator followed by removal of trace amounts of solvent using a high vacuum pump.

2-Vinyl-1H-indole (13).
Methyltriphenylphosphonium bromide (MePPh 3 Br) (7.39 g, 20.7 mmol, 6 equiv.)and dry THF (140 mL) were added to a two-neck round-bottom flask under N 2 in an ice bath and nBuLi (13.5 mL, 19.0 mmol, 5.5 equiv.) was added drop-wise to the solution at 0 o C.During this time, the color of the solution changed from white to a deep yellow.The ice bath was removed and the temperature increased to 30 ºC.The solution was stirred for 30 min at this temperature and then cooled to 0 C.Carbaldehyde 12 (0.500 g, 3.45 mmol) in THF (20 mL) was then added drop-wise to the methylenetriphenylphosphorane solution at 0 C.The ice bath was removed and the mixture stirred overnight at RT under N 2 after which diethylether (80 mL) was added and the reaction mixture was washed with H 2 O (2 × 80 mL).The aqueous layer was collected and further extracted with Et 2 O (2 × 50 mL).The organic layers were combined, washed with brine (200 mL), dried (MgSO 4 ) and the residue purified by chromatography (EtOAc/Hexane, 5:95) to afford 13 as an off-white solid (0.361 g, 74%), the 1 H NMR spectrum of which compared well to that in the literature. 35

Methyl 1H-indole-2-carboxylate (18)
. Into a two-neck round-bottom flask fitted with a condenser was placed methanol (130 mL) followed by commercially-available indole-2-carboxylic acid 17 (5.00g, 31.0 mmol).The solution was saturated with HCl gas followed by heating under reflux for 24 h.The reaction mixture was treated with saturated aqueous sodium bicarbonate until effervescence ceased, concentrated in vacuo, and the residue extracted with EtOAc (3  100 mL).The organic phases were combined and washed with brine.
Hexane was added until a slight cloudiness persisted, and the solution was cooled in an ice bath to precipitate 18 as a white powder (4.04 g, 74%). 1

Figure 1 .
Figure 1.Staurosporine 1, and examples of other important indolo-and pyrrolocarbazoles (see following references for reviews on clinically relevant staurosporin analogues 9 and other relevant carbazoles 10 ).

Figure 2 .
Figure 2. Retrosynthetic analysis to afford pyrrolocarbazole skeletons 8 containing a group on the nitrogen atom of the resultant carbazole (Route B was eventually the successful one).