Synthetic route and process optimization proposal of Rucaparib a PARP inhibitor (PF-01367338 or AG-014699)

Target

Original synthesis procedure and route

Multkilogram Scale-Up of a Reductive Alkylation Route to a Novel PARP Inhibitor

Chemical Research and Development, Pfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom,

Organic Process Research and Development 2012, Vol 16, 1897−1904

http://pubs.acs.org/doi/abs/10.1021%2Fop200238p

Details of the publication is here: http://newdrugapprovals.org/2015/06/09/ag-014699-rucaparib/

Optimization

The publication indicate a poor global yield (2.9%) even if the process is optimized, also the chosen indolization method seems to be difficult, metal catalyzed reaction employing boronic acids which are now recognized as mutagenic, reduction with cyanoborohydride and numerous steps.

The “ideal” route i have chosen involves a Fisher indolization. Few steps are tricky due to some functional group fragility, but by designing the process correctly, some side reactions could be avoided. Even if undesirable reaction occurs significantly, i have indicated some alternatives which, unfortunately, involve a cost.

The key point of this route is the indolization, i have evaluated the feasibility by some lectures and by using the software Hulis (Huckel theory), to roughly make a comparison between the hydrazone intermediate and intermediates seen in the literature, especially the trifluoroacetyl one. Starting at 4, it is really more an exploration, that’s why i have indicated some alternatives.

Lastly, it has 9 steps instead of 12, with 5 isolations.

Optimized route and process

para-Bromo-Benzaldehyde protection (1)

This starting material is 200 $/kg cheaper than the protected one by the diethylketal. Depending the protecting group resistance versus the Fisher indol reaction conditions (described later which takes in consideration the fragility of this PG), there is a cheap method to protect the aldehyde by ethanol or ethylene glycol with Zeolite H-ZMS-5 as acid catalyst (tested on benzaldehyde with methanol).

Conditions: ethanol with a loop taking the distillate by a pump, pass through a column packed with anhydrous Na₂SO₄  and re-injected into the reactor ; or ethylene glycol + other solvent not absorbed by Zeolite which make an azeotrope with water ; for both methods heat for water removal.

Reference: 
H-ZSM-5-CatalysedShape-Selective Acetalisation of Aldehydes, M. V. Joshi  and  C.S. Narasimhan, Journal Of Catalysis 128,  63--68  (1991).

Acylation (2)

For this step, i have chosen an acylation through a Grignard alone or followed by a transmetallation to give an organocuprate. There is several methods involving Br/Mg exchange with iPrMgBr.LiCl followed by the exchange with CuCN.LiCl, or a Grignard complex with the bis[2-(N,N-dimethylamino) ethyl] ether which is less reactive and could avoid a potential alkylation. Finally there is the iron-catalyzed cross-coupling of Grignard with acid chloride, it potentially has my preference, but could has a patent issue (the patent should be checked in details).

Ideally, and if it works, it could be a Br/Mg exchange with iPrMgCl-bis[2-(N,N-dimethylamino) ethyl] ether complex, and make the acylation. It is un-reactive versus halo-aryl, so most probably not reactive versus halo-alkyl. The doubt of the feasibility is with the alkyl acyl chloride, because of the +I effect of the alkyl chain on the acyl. On the contrary, the organo cuprate works with an alkyl acid chloride, but react too with alkyl iodide, it should be checked if it is un-reactive versus a Cl-C bond.

Solvent: THF, temperature: room temp

References: 
Addition ofGrignard Reagents to Aryl Acid Chlorides: An Efficient synthesis of ArylKetones, Xiao-jun Wang,* Li Zhang, Xiufeng Sun, Yibo Xu, Dhileepkumar Krishnamurthy,and Chris H. Senanayake, Organic Letters, 2005, Vol. 7, N° 25, p5593-5595.

Noncryogenic I/Br-MgExchange of Aromatic Halides Bearing Sensitive Functional Groups Usingi-PrMgCl-Bis[2-(N,N-dimethylamino)ethyl] Ether Complexes, X.-J. Wang, X. Sun,L. Zhang, Y. Xu, D. Krishnamurthy, C. H. Senanayake, Org. Lett., 2006, 8,305-307.

A LiCl-Mediated Br/MgExchange Reaction for the Preparation of Functionalized Aryl- andHeteroarylmagnesium Compounds from Organic Bromides, A. Krasovski, P. Knochel,Angew. Chem. Int. Ed., 2004, 43, 3333-3336.

Demonstrating that alkyl acyl chloride also works with organocuprate: Efficient Synthesis ofFunctionalized Organozinc Compounds by the Direct Insertion of Zinc intoOrganic Iodides and Bromides, A. Krasovski, V. Malakhov, A. Gavryushin, P.Knochel, Angew. Chem. Int. Ed., 2006, 45, 6040-6044.

SelectiveIron-Catalyzed Cross-Coupling Reactions of Grignard Reagents with EnolTriflates, Acid Chlorides, and Dichloroarenes, Bodo Scheiper , MelanieBonnekessel , Helga Krause , and Alois Fürstner *, J. Org. Chem., 2004, 69(11), pp 3943–3949.

Aromatic Nucleophilic Substitution (3)

By using DMSO at 90°C during 48h and the appropriate amount of hydrazine (6 eq ?), the conditions lead to the 3-fluoro-5-cyanophenylhydrazine. Solvent and conditions could probably be optimized through 3 DoE (solvent screening, temperature screening and hydrazine quantity screening).

References: 
Discoveryof Novel Heteroarylazoles That Are Metabotropic Glutamate Subtype 5 ReceptorAntagonists with Anxiolytic Activity, Jeffrey Roppe ,† Nicholas D. Smith ,†Dehua Huang , Lida Tehrani , Bowei Wang , Jeffrey Anderson , Jesse Brodkin ,Janice Chung , Xiaohui Jiang , Christopher King , Benito Munoz , Mark A. Varney, Petpiboon Prasit , and Nicholas D. P. Cosford *, J. Med. Chem., 2004, 47(19), pp 4645–4648.

Aromatic Nucleophilic SubstitutionsUsing X-Cube Flash™ Continuous Flow Reactor, ThalesNano Nanotechnology Inc(reference therein).

Fisher Indol synthesis (4)

The neuralgic point in this route is the Fisher indol synthesis. There are two major difficulties: the diethylketal which is sensible to acid and water, and the halo-alkyl moiety which could react with the amine intermediate to afford a 6 members ring, and lastly, the alkyl chain could make a too high stabilization by +I effect of the ene-hydrazine intermediate which could lower the yield or prevent the cyclization (see reference). On the contrary, the substituent on the aromatic ring makes the cyclization easier by their electro-withdrawing characteristics. If there is a too high stabilization, by PG removal, the aldehyde could make also an electro-withdrawing effect.

The diethylketal is a major problem only during the hydrazone formation with the formation of water. The halo-alkyl is a more important one during the cyclization.

To resolve the ketal problem, an anhydrous mineral salt could be used as MgSO₄ or Na₂SO₄ to trap the water immediately, a catalytic amount of AcOH and a gentle heating during the hydrazone formation. An intermediate filtration must be done to eliminate the salts before the cyclization which use Amberlite A15. Lastly, the reaction maybe couldn’t be conducted in an alcohol due to the potential imidate ester formation during the cyclization. If despite the precautions, the diethylketal is removed, use 1,3-dioxolane as PG which is more resistant (more expensive than the diketal if bought).

To avoid the alkylation, there are two solutions, acid catalyzed indolization with 1 eq or more of strong acid, and N-acylation which is also driving the cyclization.

For the reaction conditions, i would try THF or other polar water-miscible aprotic solvent, AcOH in a catalytic amount to avoid the PG removal during the hydrazone formation and a gentle heating to also halt to this intermediate. I will try before ethanol (or IPA my preference) as solvent considering the PG and indolization conditions seen in the literature, but i am afraid about the alcoholysis (in a small amount) of the nitrile because of the cyclization conditions (see Pinner alcoholysis). Once the first step is complete, filter out the salts, add Amberlit A15 (to keep the ammonium out of the liquid phase - see reference) and heat. If N-alkylation is observed, use the TFAA strategy (see reference) which also drive the cyclization.

Lastly, Fisher indolization could be conducted with zeolite, an alternative which could possibly avoid N-alkylation of the amine intermediate. This last solution has my preference if it works (see reference).

References: 
Thecatalytic asymmetric Fisher indolization and beyond, Thesis, Steffen Müller,2012, p83-92 for Amberlite and p26 for TFAA (interesting and well redacted).

Thermal cyclization ofN-trifluoracetyl enehydrazines under mild conditions: A novel entry into theFischer indole synthesis, O. Miyata, Y. Kimura, K. Muroya, H. Hiramatsu, T.Naito, Tetrahedron Lett., 1999, 40, 3601-3604.

Aryl Hydrazide beyondas Surrogate of Aryl Hydrazine in the Fischer Indolization: The Synthesis ofN-Cbz-indoles, N-Cbz-carbazoles, and N,N'-Bis-Cbz-pyrrolo[2,3-f]indoles, I.-K.Park, S.-E. Suh, B.-Y. Lim, C.-G. Cho, Org. Lett., 2009, 11, 5454-5456.

Why Do Some Fischer Indolizations Fail ?, Nihan Çelebi-Ölçüm, Ben W. Boal, Alexander D. Huters, Neil K. Garg*, and K. N. Houk*, J Am Chem Soc. 2011 April 20; 133(15): 5752–5755.

Zeolites as catalystin the Fisher indole synthesis. Enhanced regioselectivity for unsymmetricalketone substrates. Michal P. Prochazka, Lars Eklund and Rolf Carlson*, ActaChem. Scand., 44, 1990, 610-613.

Hydrolysis: PG removal of the aldehyde and nitrile hydrolysis followed by amide N-alkylation (5 and 6)

Add water to the precedent mixture, easy PG removal if this is an acyclic diethylketal, more difficult if this is the 1,3-dioxolane. Once the PG removal is complete, filter out the amberlite A15 and add the Amberlite A26 for the nitrile hydrolysis and N-alkylation of the amide, NaI and LiX LiI could help.

A particular attention should be taken about the N of indol during the alkylation, since a 6 members ring could be formed. If it occur in a significant manner and cannot be avoided, consider protect N before the N-amide alkylation or earlier in the route (consider also change some reaction orders if Cbz is used which also seems to drive the indolization, see earlier reference).

Update: Since the C-3 carbon is not Sp3, the alkylation of the N from indole does not occur (the terminal carbon bearing the Cl is too far).

Reference: 
AmberlystA26 OH as a Recyclable Catalyst for Hydration of Nitriles and Water-BasedSynthesis of 4(1H)-Quinazolinones from 2-Aminobenzonitrile and CarbonylCompounds, Fatemeh Tamaddon,* Farzaneh Pouramini, Synlett 2014, 25, 1127–1131.

Reductive amination (Leuckart reaction conditions) (7)

The more known condition (by some “chemist”) is N-methyl formamide with formic acid. I will add ammonium formate or MgCl₂ if the reaction is too slow, but there are some risks of N-formyl benzamide derivative or traces of Mg²⁺ which must be removed. Also the required temperature could afford impurities degradation.

The alternative could be methylamine in ethanol with triethylorthoformate to remove water, formic acid as reductive agent and ethanol or IPA as solvent. If the imine doesn’t tolerate the alcohol in a too large quantity, use methylamine in THF and acetonitrile as solvent.

Reference: 
A synthesisof abemaciclib utilizing a Leuckart–Wallach reaction, Michael O. Frederick,Douglas P. Kjell, Tetrahedron Letters Volume 56, Issue 7, Pages 949–951.

PF-01367338 PARP inhibitor Rucaparib (8)

The N deformylation could be tried directly with camphor sulfonic acid which has a pKa of 1.2 in IPA/water, if a Leuckart with N-methyl formamide is tried, else the treatment designed in the publication. If IPA is used in the previous step, directly isolate the CSA salt from the mixture.

The product must be of course recrystallized.

Costing

This is only a price comparison of majors starting materials, since the process is subject to modifications.

Initial route (12 steps):

5-Fluoro-2-methylbenzoic acid (molbase): 550$/kg (84.78$/mol)

Phthalimidoacetaldehyde diethyl acetal (molbase): 2369$/kg (623.73$/mol)

4-Formylphenylboronic acid (molbase) : 350 $/kg (52.48$/mol)

Total: 3269 $/kg (760.99$/mol)

Optimized route (9 steps):

4-Bromobenzaldehyde (molbase): 101$/kg (18.69$/mol)

5-Chlorovaleryl chloride (molbase): 141 $/kg (21.86$/mol)

3,5-Difluorobenzonitrile (molbase): 150 $/kg (20.87$/mol)

Total: 392 $/kg (61.42$/mol)

Disclaimer:
This is some personal works on paper only, i have no responsibility in any way if somebody would try this route and has all sort of troubles, including but not limited to: injuries and money loss. This is for experienced chemists only, and tests must be conducted in a suitable lab only.

But if my work is used to synthesize the targeted molecule described here, please, send a word, even if it fails, chemistry is always an experimental science. This will make me pleased, thank you.

© David Le Borgne, 2015, specialist in chemical process development and optimization.