Targeted fluorination of a nonsteroidal anti-inflammatory drug to prolong metabolic half-life

Article abstract of DOI:10.1002/cmdc.201300490

In drug design, one way of improving metabolic stability is to introduce fluorine at a metabolically labile site. In the early stages of drug design, identification of such sites is challenging, and a rapid method of assessing the effect of fluorination on a putative drug's metabolic stability would be of clear benefit. One approach to this is to employ micro-organisms that are established as models of drug metabolism in parallel with the synthesis of fluorinated drug analogues. In this study, we have used the filamentous fungus Cunninghamella elegans to identify the metabolically labile site of the nonsteroidal anti-inflammatory drug flurbiprofen, to aid in the design of fluorinated derivatives that were subsequently synthesised. The effect of the additional fluorine substitution on cytochrome P450-catalysed oxidation was then determined via incubation with the fungus, and demonstrated that fluorine substitution at the 4′-position rendered the drug inactive to oxidative transformation, whereas substitution of fluorine at either 2′ or 3′ resulted in slower oxidation compared to the original drug. This approach to modulating the metabolic stability of a drug-like compound is widely applicable and can be used to address metabolic issues of otherwise good lead compounds in drug development. A metabolic stopper: By applying a chemical-microbiological approach to the design of drugs with enhanced metabolic stability, a series of fluorinated derivatives of the nonsteroidal anti-inflammatory drug (NSAID) flurbiprofen was synthesised that were more resistant to cytochrome P450-catalysed transformation than the original drug.

Full text of DOI:10.1002/cmdc.201300490

CHEMMEDCHEM
COMMUNICATIONS
DOI: 10.1002/cmdc.201300490
Targeted Fluorination of a Nonsteroidal Anti-inflammatory
Drug to Prolong Metabolic Half-Life
Maxwell J. Shaughnessy,[a] Antal Harsanyi,[a] Jingji Li,[b] Tara Bright,[b] Cormac D. Murphy,*[b]
and Graham Sandford*[a]
In drug design, one way of improving metabolic stability is to
introduce fluorine at a metabolically labile site. In the early
stages of drug design, identification of such sites is challeng-
ing, and a rapid method of assessing the effect of fluorination
on a putative drug’s metabolic stability would be of clear ben-
efit. One approach to this is to employ micro-organisms that
are established as models of drug metabolism in parallel with
the synthesis of fluorinated drug analogues. In this study, we
have used the filamentous fungus Cunninghamella elegans to
identify the metabolically labile site of the nonsteroidal anti-in-
flammatory drug flurbiprofen, to aid in the design of fluorinat-
ed derivatives that were subsequently synthesised. The effect
of the additional fluorine substitution on cytochrome P450-cat-
alysed oxidation was then determined via incubation with the
fungus, and demonstrated that fluorine substitution at the 4’-
position rendered the drug inactive to oxidative transforma-
tion, whereas substitution of fluorine at either 2’ or 3’ resulted
in slower oxidation compared to the original drug. This ap-
proach to modulating the metabolic stability of a drug-like
compound is widely applicable and can be used to address
metabolic issues of otherwise good lead compounds in drug
development.
electronegativity leads to significant electronic consequences.
One reason for incorporating fluorine into drug molecules is to
slow their metabolic processing though the cytochrome P450
(CYP)-mediated detoxification steps in the liver, thus extending
their potency.[2] This requires knowledge of where the metabol-
ically labile sites are in the drug compound, which is available
for drugs that have already undergone absorption, distribution,
metabolism, and excretion (ADME) studies, and also increasing-
ly for compounds that are at the very early stage of develop-
ment (hit-to-lead). One approach to fluorinating metabolically
labile sites is to activate the site via CYP-catalysed oxidation
enabling further functionalisation via deoxofluorination.[3] We
recently reported
a
straightforward chemical–microbial
method for identifying metabolically sensitive sites on drug-
like compounds, which can be fluorinated and then readily re-
assessed for biotransformation.[4] By employing a microorgan-
ism, such as the zygomycete fungus Cunninghamella elegans
or the actinomycete Streptomyces griseus, both of which ex-
press CYP and are known to metabolise drugs in a similar fash-
ion to mammals,[5] it was possible to identify the metabolically
labile sites of drug-like biphenyl molecules. Using Suzuki–
Miyaura coupling reactions with relevant fluorinated benzene
boronic acids, it was possible to generate fluorinated biphenyl
derivatives that, upon subsequent re-incubation with the
micro-organism, were stable to metabolic attack (Scheme 1).
Herein, we describe the next stage in the development of
our technique, by demonstrating its applicability to a drug cur-
rently in use. Flurbiprofen (1) [(R,S)-2-(2-fluoro-4-biphenyl)pro-
pionic acid] is a nonsteroidal anti-inflammatory drug used to
manage arthritic pain, is an active ingredient in some throat
lozenges, and the R-enantiomer is in clinical trials as a treat-
ment for prostate cancer.[6] When incubated with C. elegans
and S. griseus, flurbiprofen is transformed to the main human
metabolite, 4’-hydroxy-flurbiprofen (2), and to the minor me-
tabolites 3’,4’-dihydroxyflurbiprofen (3) and 3’-hydroxy-4’-me-
thoxy-flurbiprofen (4)[7] (Scheme 2); in C. elegans, a sulfated
conjugated metabolite is also generated. Therefore, we report
the synthesis of various fluorinated derivatives of flurbiprofen
5–9, which were subsequently incubated with C. elegans, so
that the effect of fluorine substitution on metabolically labile
sites could be evaluated.
The unique physicochemical properties of fluorine, such as its
size, electronegativity and strength of the carbon–fluorine
bond, make it an important element in the design of drugs
and other biologically active compounds. Consequently, fluo-
rine is present in approximately 20% of recently developed
pharmaceuticals, including the blockbuster drugs Prozac
(fluoxetine; (R,S)-N-methyl-3-phenyl-3-[4-(trifluoromethyl)phe-
noxy]propan-1-amine) and Lipitor (atorvastatin; (3R,5R)-7-[2-(4-
fluorophenyl)-3-phenyl-4-(phenylcarbamoyl)-5-propan-2-ylpyr-
rol-1-yl]-3,5-dihydroxyheptanoic acid), and 30% of agrochemi-
cals.[1] Owing to fluorine’s size, incorporation into biologically
active compounds has only minor steric effects, but its high
[a] M. J. Shaughnessy, A. Harsanyi, Prof. G. Sandford
Department of Chemistry, Durham University
South Road, Durham DH1 3LE (UK)
E-mail: graham.sandford@durham.ac.uk
A short series of flurbiprofen analogues (5–9) was synthes-
ised by a five-step strategy, adapting literature procedures[8]
(Scheme 3). Reaction of diethyl-2-methyl malonate ester with
2,4-difluoronitrobenzene (10) mediated by sodium hydride
proceeded in high yield to give desired product 11 (74%) aris-
ing from nucleophilic substitution of fluorine para to the nitro
[b] J. Li, T. Bright, Dr. C. D. Murphy
School of Biomolecular & Biomedical Science
Centre for Synthesis & Chemical Biology
University College Dublin, Ardmore House, Belfield, Dublin 4 (Ireland)
E-mail: cormac.d.murphy@ucd.ie
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300490.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 733 – 736 733

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
addition of the drug. After further 72 hours incuba-
tion, the culture was extracted with ethyl acetate and
the solvent removed under reduced pressure. The ox-
idation products were analysed by gas chromat-
ography–mass spectrometry (GC–MS) after silylation
with N-methyl-N-(trimethylsilyl) trifluoroacetamide
(MSTFA). Only the starting compounds 5, 8 and 9
were detected after incubation. The absence of trans-
formation was expected based on the knowledge
that CYP-catalysed hydroxylation occurs at the 4’-po-
sition of flurbiprofen, is consistent with our previous
observations concerning metabolism of model fluori-
nated biphenyl derivatives,[4] and confirms the effec-
tiveness of this methodology for identifying metabol-
ic labile sites for fluorinated drug design.
Scheme 1. A chemical–microbial method for developing metabolically stable fluorinated
drug-like molecules.[3] a) C. elegans, S. griseus; b) 1. Pd(PPh3)4, K2CO3, 2. NaOH, EtOH,
reflux.
2’-Fluoro- and 3’-fluoro-flurbiprofen were trans-
formed in varying degrees by the fungus (Figure 2).
3’-Fluoro-flurbiprofen (6) was almost entirely convert-
Scheme 2. Oxidative metabolites of flurbiprofen via incubation with C. ele-
gans and S. griseus.
Figure 1. X-ray crystallograpic structure of 7.
substituent. Small quantities of product arising from
competing ortho substitution were detected by
19F NMR, but the isomeric mixture was used in the
following stages without purification. Hydrogenation
of the nitro group of 11 to give corresponding ani-
line 12 was accomplished by reaction with hydrogen
over a palladium-on-carbon catalyst and subsequent
diazotisation–iodination gave key iodoaromatic inter-
mediate 13 in good yield. Palladium-catalysed
Suzuki–Miyaura coupling reactions between iodoaro-
matic 13 and a range of fluoroaryl boronic acids
(14a–f) gave corresponding biphenyl diesters 15a–f.
Subsequent decarboxylation and hydrolysis gave
a short series of the desired fluorinated flurbiprofen
analogues 5–9. Recrystallisation of the acid analogues
removed trace quantities of ortho-substituted prod-
ucts carried through the synthesis from the first
stage and, for example, the structure of 7 was con-
firmed by X-ray crystallography (Figure 1).
Fluorinated
flurbiprofen
analogues
5–9
(0.1 mgmLꢀ1) were incubated with C. elegans as we
described previously.[4] Briefly, the fungus was culti-
vated for 72 hours in 250 mL Erlenmeyer flasks con-
Scheme 3. Synthesis of fluorinated flurbiprofen analogues 5–9. Reagents and conditions:
a) NaH, DMF, RT, 30 min; b) H2 (3 bar), Pd/C, EtOH, 1 h; c) HCl, NaNO2, KI, H2O, 08C!RT,
24 h; d) Pd(PPh3)4, K2CO3, toluene/H2O, reflux, o/n; e) NaOH, EtOH, reflux, 6 h. * For yields
taining 45 mL sabouraud dextrose broth prior to the of 15a–f and 5–9, see the Supporting Information.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 733 – 736 734

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
products, all with the expected mass of correspond-
ing hydroxylated derivatives.
To identify whether any products are formed by in-
cubation of 6 and 7 with C. elegans and to aid oxida-
tion products identification,19F NMR analysis of the
extracts was conducted and, as expected, one fluori-
nated product was observed from the extracts recov-
ered from incubation with 6 and three products from
7 (Figure 3).
19F NMR substituent chemical shift (SCS) analysis
can be used to identify the hydroxylated products
arising from oxidation of 6 and 7 by C. elegans.
Peelen et al.[9] conducted extensive 19F NMR experi-
ments when investigating enzymatic hydroxylation of
fluorophenol and demonstrated that hydroxylations
ortho, meta and para to the fluorine resulted in sub-
stituent shifts from the starting substrate of ꢀ23,
+1.3 and ꢀ11 ppm, respectively, upon substitution
of hydrogen atom by a hydroxy group. Consequently,
simple calculations allow a reasonable prediction of
19F NMR chemical shifts for all possible hydroxylated
products (Figure 4).
Using the data in Figures 3 and 4, it is reasonable
to conclude that 3’-fluoro-flurbiprofen (6) is most
probably converted to the 4’-hydroxylated product
(10a), consistent with our earlier findings concerning
metabolism of model biphenyl derivatives at the 4’-
position,[4] although 10b is a possible product based
upon the 19F NMR shift. In contrast, 2’-fluoro-flurbi-
profen (7) gives three hydroxylated products by GC-
MS and 19F NMR analysis, the main product probably
being 11 b arising from hydroxylation of the 4’-posi-
tion consistent with earlier findings,[4] with minor
products 11 a and 11 c, in the approximate ratio 11 a/
11 b/11 c of 1:8:2 as determined by 19F NMR reso-
nance intensities. Again, the major product is most
probably hydroxylation at the least metabolically
stable 4’-position (11 b), although significant quanti-
ties of 11 a and 11 c are formed reflecting the stabilis-
ing effect of fluorine ortho and para to the sites of
oxidation.
To demonstrate that the fungus is an appropriate
model for metabolic stability experiments, rat liver
microsomes[10] were incubated with flurbiprofen de-
rivatives 5–8, plus NADPH. The same transformation
products were detected by GC-MS from 6 and 7; in
contrast, no transformation of 5 and 8 was observed
confirming that blocking the 4’-position with fluorine
prevented metabolism.
Figure 2. Total ion chromatograms from incubation of 6 and 7 with C. elegans: a) com-
pound 6 without fungus; b) biotransformation of 6 to hydroxylated product; c) com-
pound 7 without fungus; d) biotransformation of 7 to three hydroxylated products. All
of the extracts were derivatised with N-methyl-N-(trimethylsilyl) trifluoroacetamide
(MSTFA) prior to GC-MS analysis.
In summary, following our previously reported
strategy for identifying metabolically labile sites on
drug-like compounds for subsequent fluorination, we
have demonstrated that this strategy is effective for
generating fluorinated derivatives of known drugs
ed to a single hydroxylated product as observed by GC-MS
analysis, whereas 2’-fluoro-flurobiprofen (7) was incompletely
transformed (~30% starting material remained) yielding three
that are metabolically stabilised. Flurbiprofen (1) is known to
be initially oxidised at the 4’-position by both mammals and
micro-organisms. Analogues of flurbiprofen bearing fluorine
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 733 – 736 735

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
detected after rat liver microsomes were incubated
with 5–8, demonstrating the relevance of the micro-
bial model used here. It is often attractive to improve
drug efficacy by slowing their metabolism, thus 2’-
and 3’-fluoroflurbiprofen might be more useful deriv-
atives as their biotransformation is altered, but not
completely prevented, as with the 4’-fluorinated ana-
logue.
Experimental Section
Details of the synthetic procedures, characterisation of
compounds and microbiological methods can be found
in the Supplemental Information.
Figure 3. 19F NMR spectra of extracts from C. elegans cultures incubated with a) 3’-fluoro-
flurbiprofen (6) and b) 2’-fluoroflurbiprofen (7).
Acknowledgements
This work was supported in part by The Royal Society
through an International Joint Project grant, the
Durham University MChem Programme, and the University Col-
lege Dublin Research Demonstrator Scheme. The authors thank
Dr. D. S. Yufit (Durham University, UK) for X-ray crystallography.
Keywords: cross-coupling reactions · cytochrome P450s · drug
design · fluorine · drug metabolism
[1] a) C. Isanbor, D. O’Hagan, J. Fluorine Chem. 2006, 127, 303–319; b) K. L.
Kirk, J. Fluorine Chem. 2006, 127, 1013–1029; c) K. Muller, C. Faeh, F.
Diederich, Science 2007, 317, 1881–1886.
[2] S. Purser, P. R. Moore, S. Swallow, V. Gouverneur, Chem. Soc. Rev. 2008,
37, 320–330.
[3] a) A. Rentmeister, F. H. Arnold, R. Fasan, Nat. Chem. Biol. 2009, 5, 26–28;
b) K. Zhang, B. M. Shafer, M. D. Demars, H. A. Stern, R. Fasan, J. Am.
Chem. Soc. 2012, 134, 18695–18704.
[4] T. V. Bright, F. Dalton, V. L. Elder, C. D. Murphy, N. K. O’Connor, G. Sand-
ford, Org. Biomol. Chem. 2013, 11, 1135–1142.
[5] a) S. Asha, M. Vidyavathi, Biotechnol. Adv. 2009, 27, 16–29; b) D. C.
Lamb, M. R. Waterman, B. Zhao, Expert Opin. Drug Metab. Toxicol. 2013,
9, 1279–1294.
[6] a) L. Perioli, V. Ambrogi, S. Giovagnoli, M. Ricci, P. Blasi, C. Rossi, AAPS
PharmSciTech 2007, 8, E20–E27; b) E. J. Quann, F. Khwaja, K. H. Zavitz,
D. Djakiew, Cancer Res. 2007, 67, 3254–3262.
Figure 4. Calculated and observed 19F NMR shifts of hydroxylated flurbipro-
fen derivatives 10 and 11.
[7] a) J. Amadio, K. Gordon, C. D. Murphy, Appl. Environ. Microbiol. 2010, 76,
6299–6303; b) T. V. Bright, B. R. Clark, E. O’Brien, C. D. Murphy, J. Mol.
Catal. B 2011, 72, 116–121.
[8] I. Peretto, S. Radaelli, C. Parini, M. Zandi, L. F. Raveglia, G. Dondio, L.
Fontanella, P. Misiano, C. Bigogno, A. Rizzi, B. Riccardi, M. Biscaioli, S.
Marchetti, P. Puccini, S. Catinella, I. Rondelli, V. Cenacchi, P. T. Bolzoni, P.
Caruso, G. Villetti, F. Facchinetti, E. Del Giudice, N. Moretto, B. P. Imbim-
bo, J. Med. Chem. 2005, 48, 5705–5720.
atoms at the 2’-, 3’- and 4’-positions (5–7) were synthesised.
Upon incubation with C. elegans to determine if these modifi-
cations had any effect on biotransformation, the flurbiprofen
analogue that had an additional fluorine in the 4’-position (7)
was completely stable to metabolism. Analogues of the parent
drug with fluorine located at the 2’- and 3’-positions resulted
in hydroxylation to varying degrees upon incubation; 3’-fluoro-
flurbiprofen (6) was almost completely transformed to a hy-
droxylated product (10a), whereas 2’-fluoroflurbiprofen (7) was
transformed to a lesser extent, and to three monohydroxylated
products. Significantly, the same transformation products were
[9] S. Peelen, I. Rietjens, M. G. Boersma, J. Vervoort, Eur. J. Biochem. 1995,
227, 284–291.
[10] R. S. Obach, Drug Metab. Dispos. 1999, 27, 1350–1359.
Received: November 27, 2013
Published online on January 28, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 733 – 736 736