2 Covalent JAK3 Inhibitors
As mentioned before, the Janus kinases feature a highly conserved ATP pocket
impeding the design of specific compounds. The very limited differences in this
binding site are highlighted in the structure-derived sequence alignment depicted
in Fig. 4. Accordingly, one of the few distinguishing features that could enable the
rational design of selective JAK3 inhibitors is a cysteine at position 909 in the solvent
exposed front region, which is amenable to covalent targeting. This residue is located
in the αD-1 position (nomenclature according to ref. [33]) seven amino acids after the
gatekeeper residue. Remarkably, there are ten other kinases (BLK, BMX, BTK,
EGFR, HER2, HER4, ITK, MAP2K7, TEC, and TXK) known to feature an equivalently positioned cysteine [33]. The ligandability of this residue has long been proven
for cysteine 797 in the EGFR kinase domain, for which the first covalent inhibitors had
already been developed at the end of the last century [34]. Notably, cysteines in the
αD-1 position are also targeted by all the currently approved covalent kinase inhibitors, i.e., the EGFR/HER-family inhibitors afatinib (1), neratinib (2), and osimertinib
(3) and the BTK inhibitors ibrutinib (4) and acalabrutinib (5; see Fig. 2a). Addressing
Cys909 in JAK3 has led to the phase II clinical candidate PF-06651600 [35], which is
under investigation for the treatment of rheumatoid arthritis, Crohn’s disease, ulcerative colitis, and alopecia areata (vide infra). Efforts which have enabled the development of this and other covalent JAK3 inhibitors are summarized in the following
sections. For general consideration on the design and development of covalent kinase
inhibits, please consult the respective chapter in this book. It should also be noted at
this point that IC 50 values of covalent inhibitors are time-dependent and therefore
difficult to compare. Thus, the kinetics of covalent inactivation described the secondorder rate constant k inact /K I should be used to compare the overall efficiency of
covalent inhibitors (see the aforementioned chapter on covalent kinase inhibitors for
details). However, since no kinetic data was provided in most of the cited studies, the
bulk of the discussion in this chapter is still based on IC 50 data.
The first putatively covalent JAK3 inhibitors were described in the year 2000 by
researchers from AstraZeneca [37]. They investigated naphthyl(β-aminoethyl)
ketones and analogous Mannich bases (Fig. 5), which can form reactive
vinylketones upon elimination of the β-amino substituent. The covalent modification
of JAK3 was not scrutinized in this study. Nevertheless, a covalent mechanism of
action is very likely since these relatively simple compounds, although being devoid
of a classical hinge-binding motif, are potent JAK3 inhibitors (pIC 50 values up to
7.1) while possessing only weak JAK1 inhibitory potency. In support of this
assumption, the isolated vinyl 2-naphthyl ketone 25 possessed a similar inhibitory
potency (pIC 50 ¼ 7.1) as the most potent compound 24, while the stabilized
derivatives 27 and 28, which cannot undergo elimination (retro-aza-Michael addition), remained completely inactive.
A similar mode of action can be assumed for the macrocyclic JAK3 inhibitors
NC1153 (29) and EP009 (30, Fig. 6), which were described in 2005 and 2014,
respectively [38, 39]. However, neither was a medicinal chemistry optimization of
these compounds disclosed nor was the covalent mode of action experimentally
Covalent Janus Kinase 3 Inhibitors
231
As mentioned before, the Janus kinases feature a highly conserved ATP pocket
impeding the design of specific compounds. The very limited differences in this
binding site are highlighted in the structure-derived sequence alignment depicted
in Fig. 4. Accordingly, one of the few distinguishing features that could enable the
rational design of selective JAK3 inhibitors is a cysteine at position 909 in the solvent
exposed front region, which is amenable to covalent targeting. This residue is located
in the αD-1 position (nomenclature according to ref. [33]) seven amino acids after the
gatekeeper residue. Remarkably, there are ten other kinases (BLK, BMX, BTK,
EGFR, HER2, HER4, ITK, MAP2K7, TEC, and TXK) known to feature an equivalently positioned cysteine [33]. The ligandability of this residue has long been proven
for cysteine 797 in the EGFR kinase domain, for which the first covalent inhibitors had
already been developed at the end of the last century [34]. Notably, cysteines in the
αD-1 position are also targeted by all the currently approved covalent kinase inhibitors, i.e., the EGFR/HER-family inhibitors afatinib (1), neratinib (2), and osimertinib
(3) and the BTK inhibitors ibrutinib (4) and acalabrutinib (5; see Fig. 2a). Addressing
Cys909 in JAK3 has led to the phase II clinical candidate PF-06651600 [35], which is
under investigation for the treatment of rheumatoid arthritis, Crohn’s disease, ulcerative colitis, and alopecia areata (vide infra). Efforts which have enabled the development of this and other covalent JAK3 inhibitors are summarized in the following
sections. For general consideration on the design and development of covalent kinase
inhibits, please consult the respective chapter in this book. It should also be noted at
this point that IC 50 values of covalent inhibitors are time-dependent and therefore
difficult to compare. Thus, the kinetics of covalent inactivation described the secondorder rate constant k inact /K I should be used to compare the overall efficiency of
covalent inhibitors (see the aforementioned chapter on covalent kinase inhibitors for
details). However, since no kinetic data was provided in most of the cited studies, the
bulk of the discussion in this chapter is still based on IC 50 data.
The first putatively covalent JAK3 inhibitors were described in the year 2000 by
researchers from AstraZeneca [37]. They investigated naphthyl(β-aminoethyl)
ketones and analogous Mannich bases (Fig. 5), which can form reactive
vinylketones upon elimination of the β-amino substituent. The covalent modification
of JAK3 was not scrutinized in this study. Nevertheless, a covalent mechanism of
action is very likely since these relatively simple compounds, although being devoid
of a classical hinge-binding motif, are potent JAK3 inhibitors (pIC 50 values up to
7.1) while possessing only weak JAK1 inhibitory potency. In support of this
assumption, the isolated vinyl 2-naphthyl ketone 25 possessed a similar inhibitory
potency (pIC 50 ¼ 7.1) as the most potent compound 24, while the stabilized
derivatives 27 and 28, which cannot undergo elimination (retro-aza-Michael addition), remained completely inactive.
A similar mode of action can be assumed for the macrocyclic JAK3 inhibitors
NC1153 (29) and EP009 (30, Fig. 6), which were described in 2005 and 2014,
respectively [38, 39]. However, neither was a medicinal chemistry optimization of
these compounds disclosed nor was the covalent mode of action experimentally
Covalent Janus Kinase 3 Inhibitors
231