1. Introduction.

Fluorine is the most bountiful halogen in the Earth’s crust, ranking 13^th in abundance of all the elements. Yet organically bound fluorine appears to be rare in Nature; virtually, it doesn’t possess a footprint in biochemical evolution. [1] Thus being one of the most xenobiotic class of naturally occurring compounds, how did the interest in fluorinated motifs grow into placing them as the quintessence of modern drug discovery processes? Several reviews agree that it all began in the mid-1950s with the introduction to the market of fludrocortisone 1, the earliest fluorine-containing pharmaceutical (Figure 1). [2]

Before 1950, the concept of introducing fluorine into biologically active compounds had seemed implausible. After all, if Nature doesn’t do it, why should we? However, due to its unique properties and the growing knowledge acquired over the last decades, fluorine-scanning and editing have become standard procedures in the design of new drugs. [2a]

The substitution of fluorine for hydrogen or the addition of fluorinated groups is a field of medicinal chemistry. It harnesses the physicochemical properties, reactivity, and biological activities of these fluorine-containing organic molecules to overcome shortcomings in current drug design. Membrane permeability, lipophilicity, clearance rates, microsomal metabolism, or even the conformation or the potency of a molecule represent some of the current issues that require solutions. It is due to the electronic properties and relatively small size of fluorine that makes it suitable for bioisostere construction, to enhance the in vitro and in vivo properties of a drug.

2. Bioisosteres

The design of bioisosteres requires a deep understanding of the physicochemical properties of the element to be replaced as well as the binding site of the molecule. For instance, fluorine is approximately 20% larger than hydrogen, and the length of a C–F bond is slightly larger than that of a C=O bond. Being the most electronegative element on the periodic table, the dipole moment of the C–F bond is larger and in opposite direction of a C–H but slightly less than that of a C=O moiety. Unlike the other members of the halogen family, fluorine does not engage in strong halogen bonding interactions and is considered to be nonpolarizable. For these reasons, fluorine has played a prominent role as a bioisostere in drug design to date. It has been used as a substitute for lone pairs of electrons, the hydrogen atom, the methyl group, or even as a functional mimetic of the carbonyl group. [2b] Let’s see some examples.

tert-Butyl replacement for metabolically stable substrates.

Optimization of the metabolic stability of lead molecules is of critical importance in the drug discovery process. For instance, susceptibility to metabolism is a common issue with the tert-butyl group. During a lead optimization effort, the authors found a pocket in a receptor with a strong preference for a tert-butyl group. However, it presented high clearance in rats and microsomal incubations. Thus, the effects of structural modification of 2 were compared in an in vitro study in rat and human liver microsomes (RLM and HLM, table 1). The tert-butyl moiety was replaced for a series of compounds. While the more polar substituents achieved some success, the CF₃-substituted cyclopropyl unit emerged as the optimal lipophilic substituent to enhance the metabolic stability of the substrate, in view of the replacement being accommodated in a hydrophobic binding pocket. [3]

R
RLM t1/2 (min)	30	125	>400
HLM t1/2 (min)	51	202	150

Hydrogen atom replacement for solubility enhancement.

Because of its strong electronegativity, fluorine is a powerful tool for modulating the pKa of proximal functionality. The next example refers to a molecule developed by Bristol Myers Squibb for the treatment of migraine. It turns out that replacement of a hydrogen atom ortho to an anilide N–H by fluorine, 3,(Scheme 1) allowed an over 30-fold increase of the measured aqueous solubility of this CGRP receptor antagonist. Accordingly, the fluorine atom polarizes the adjacent N–H bond promoting enhanced solvation of the molecule. [4]

Trifluoromethyl insertion for improved agonist selectivity.

PPI-4955, 4, is the prodrug of a selective sphingosine-1-phosphate receptor-1 (S1P1) agonist, and it is used as an immunosuppressant (Scheme 2). Screening of different structural surrogates demonstrated that the incorporation of the trifluoromethyl group on the aromatic ring improved the agonist selectivity for the S1P1 receptor over the S1P3 substantially. [5]

Clearly, the introduction of fludrocortisone 1 to the market heralded the advent of a new class of drugs during the last century. In this century, we have witnessed the interest in finding novel and practical applications of fluorine as a bioisostere in drug design skyrocket––perhaps, stimulated by the constant development of state-of-the-art synthetic methodologies that deliver new and, until now, inaccessible fluorinated motifs.

Synthetic methodologies to introduce fluorine.

Introducing fluorinated substituents into conformationally restricted rings such as cyclobutane is of special interest in modern medicinal chemistry; the physicochemical properties of the corresponding derivatives are important to ADMET parameters. While the presence of fluoroalkyl-substituted cyclobutanes in the literature is notorious, 1,2-disubstituted cyclobutane derivatives bearing a fluorinated motif are practically omitted, until now. Recently, Grygorenko and co-workers reported the synthesis of these carboxylic acids and primary amines, (Scheme 3) giving access to a new and ready-available array of building blocks. [6]

To incorporate the fluoroalkyl groups into the targeted molecules, they made use of classical approaches such as deoxyfluorination with SF₄ (Scheme 4a), cyanotrifluoromethylation (Scheme 4b), TMAF-mediated nucleophilic substitution (Scheme 4c), and deoxofluorination with DAST (Scheme 4d).

There are many other routes to install a fluoroalkyl group onto a molecule: decarboxylative trifluoromethylation of carboxylic acids, the addition of the Ruppert–Parkash type reagents (TMSCF₃, TMSC₂F₅), fluoride-mediated nucleophilic substitution. Some modern techniques to the direct and selective introduction of small perfluoroalkyl groups involve the metal-catalyzed cross-coupling of organohalides and the appropriate perfluoroalkyl reagent. However, and considering the growing interest of fluorinated molecules for their potential use in drugs (i.e., the need of reducing costs and increasing efficiency), there is still room for improvement, which may come from the hand of basic research.

The cross-coupling of organohalides with perfluoroalkyl reagents employing nickel or palladium is hindered by the high kinetic barrier to reductive elimination. On the other hand, while copper complexes undergo facile reductive elimination, the oxidative addition of Cu(I) with organohalides is somehow limited. Trying to overcome this issue, MacMillan and co-workers developed a method for the efficient perfluoroalkylation of a variety of aryl and alkyl bromides that allowed the late-stage functionalization of several drug analogs (Scheme 5). [7]

Selvita also contributes to this area developing novel photodecarboxylative approaches for introduction of fluorinated substituents in drug molecules. Our research is funded by the industrial grant received from Małopolska Centre for Entrepreneurship (RPMP.01.02.01-IP.01-12-038/16). The results will be disclosed in due course as a separate publication and meanwhile a patent application has been filed.

As previously mentioned, the development of new protocols that give access to ever more intricate fluorinated motifs stimulates the incorporation of fluorine into a drug molecule. Or maybe, it is the growing interest of medicinal chemists in incorporating these fluoroalkylated motifs on drugs that encourage the research of the new methodologies. Either way, this field of medicinal chemistry is on-trend, and we expect this feedback-loop to bring tremendous improvements over the next decades.

• Author:

Lluis Llorens-Palomo, Ph.D.
Scientist III
Chemistry Department, Selvita

References:

[1]. D. B. Harper, D. O’Hagan, Nat. Prod. Rep., 1994, 11, 123–133.
[2]. (a) H. Mei, J. Han, S. White, D. J. Graham, K. Izawa, T. Sato, S. Fustero, N. A. Meanwell, V. A. Soloshonok, Chem. Eur. J. 2020, 26, 11349. (b) N.A. Meanwell, J. Med. Chem. 2018, 61, 5822−5880. (c) M. C. Walker, M. C. Y. Chang, Chem. Soc. Rev., 2014, 43, 6527–6536.
[3]. D. Barnes-Seeman, M. Jain, L. Bell, S. Ferreira, S. Cohen, X.-H. Chen, J. Amin, B. Snodgrass, P. Hatsis, ACS Med. Chem. Lett., 2013, 4, 514−516.
[4]. A. P. Degnan, P. V. Chaturvedula, C. M. Conway, D. A. Cook, C. D. Davis, R. Denton, X. Han, R. Macci, N. R. Mathias, P. Moench, S. S. Pin, S. X. Ren, R. Schartman, L. J. Signor, G. Thalody, K. A. Widmann, C. Xu, J. E. Macor, G. M. Dubowchik, J. Med. Chem., 2008, 51, 4858−4861.
[5]. G. Evindar, S. G. Bernier, E. Doyle, M. J. Kavarana, A. L. Satz, J. Lorusso, H. S. Blanchette, A. K. Saha, G. Hannig, B. A. Morgan, W. F. Westlin, Bioorg. Med. Chem. Lett., 2010, 20, 2520–2524.
[6] O. P. Demchuk, O. V. Hryshchuk, B. V. Vashchenko, S. A. Trofymchuk, K. P. Melnykov, A. Skreminskiy, D. M. Volochnyuk, O. O. Grygorenko, Eur. J. Org. Chem. 10.1002/ejoc.202001345.
[7] X. Zhao, D. W. C. MacMillan, J. Am. Chem. Soc., 2020, 142, 46, 19480–19486.