Selective chemical rescue of a thyroid-hormone-receptor mutant, TRβ(H435Y), identified in pituitary carcinoma and resistance to thyroid hormone

Article abstract of DOI:10.1002/anie.200801742

(Figure Presented) Ligand to the rescue! Thyroid hormone receptor (TR) plays a critical role in development and homeostasis. The mutant TRβ(H435Y) was identified in both cancer and resistance to thyroid hormone. A designed synthetic ligand recovered the cellular reporter-gene activity of TRβ(H435Y) with a 5850-fold improvement in selectivity for the mutant relative to that of the natural hormone T3 (shown in a modeled structure with TRβ-(H435Y)).

Full text of DOI:10.1002/anie.200801742

Communications
DOI: 10.1002/anie.200801742
Ligand Design
Selective Chemical Rescue of a Thyroid-Hormone-Receptor Mutant,
TRb(H435Y), Identified in Pituitary Carcinoma and Resistance to
Thyroid Hormone**
A. Quamrul Hassan and John T. Koh*
The thyroid hormone receptors (TRs) are ligand-dependent
transcriptional regulators that control critical genes in devel-
opment and homeostasis in response to triiodothyronine
(T3).[1] As an important regulator of differentiation, TRb has
been shown to be mutated in a high percentage of certain
cancer types, including kidney, pituitary, liver, and thyroid
cancer.[2,3] These spontaneous TRb mutations cause the
reduction or loss of TR function in a similar way to germline
TRb mutants associated with inheritable genetic disease
resistance to thyroid hormone (RTH).[4] Paradoxically, RTH
patients do not appear to be predisposed to these forms of
cancer, although, in a few cases, identical TRb mutants have
been identified in cancer and RTH.
As part of our studies exploring applications of chemical
rescue by small-molecule complementation, we previously
examined how mutations to the TRb “His-Phe switch” motif,
which mediates ligand-dependent transactivation response,
can dramatically impair receptor function.[5] Herein, we
describe a new strategy to rescue a naturally occurring TRb
mutant, His435!Tyr, by reorienting hydrogen-bonding inter-
actions at the ligand–receptor interface. As TRb(H435Y) has
been found in both RTH and pituitary carcinoma, our results
serve perhaps as the first example of chemical rescue that
targets a mutant protein involved in multiple disease states.
Upon ligand binding, TR undergoes a conformational
change that involves the repositioning of helix 12 to form a
coactivator-binding interface (Figure 1a).[6] For most nuclear
receptors, the hormone does not make direct contact with
helix 12, but rather interacts with residues on helix 11. These
residues make contacts with helix 12 through a His-Trp or a
His-Phe switch, which transduces ligand binding into a
transcription response.[5,7] Mutations to the His-Phe switch
of TRb have been associated with dramatic (320– > 5000-
fold) reductions in ligand potency.[2,8,9]
of the thyroid-hormone-receptor agonist GC-1 have greater
potency and efficacy with TRb(H435A) than the natural
hormone T3; however, these analogues are ineffective in
rescuing the activity of the His435 mutants TRb(H435Y) and
TRb(H435L), which are known to be associated with RTH
and cancer.[5] Whereas TRb(His435L) is inactive at all T3
concentrations tested (ꢀ 5 mm), T3 is a full agonist (100%
efficacy) for TRb(H435Y), although it is 390 times less potent
with this mutant than with wild-type (wt) TRb. These results
suggest that TRb(H435Y) retains its intrinsic ability to
mediate ligand-dependent transcription response but requires
extreme supraphysiological concentrations of T3 that would
not be tolerated in vivo owing to the overstimulation of wild-
type TRs. As in other studies in which the thyroid hormone
receptor was targeted, the delicate balance of TR activity
within the hypothalamic-pituitary-thyroid axis emphasizes
the need for a ligand with subtype selectivity.[4,10–12]
Molecular modeling of TRb(H435Y) suggested that the
Tyr435 side chain is still able to engage Phe459 through aryl–
aryl interactions (Figure 2, right). Although the phenol
hydroxy group of tyrosine is capable of forming a hydrogen
bond, it is not appropriately positioned to interact with
receptor-bound T3. We reasoned that appropriately designed
hormone analogues may be able to rescue potency to
TRb(H435Y) selectively by restoring hydrogen-bonding/
aryl–aryl interactions of the His-Phe switch through the
creation of a novel Tyr-Phe switch. This strategy presented a
unique challenge, as the side chain of tyrosine is considerably
longer than that of histidine; therefore, it was necessary to
introduce a hydrogen-bonding group while making the over-
all ligand structure smaller. As an initial approach, we
reasoned that the outer phenyl ring of T3 could be replaced
by a pyridyl ring (Figure 1b, right). For ease of synthesis and
product stability, we chose to make analogues of the halogen-
free thyromimetic GC-1[13] rather than analogues of T3 itself.
We could then vary the alkyl substituent at the 3’-position
with the aim of optimizing the hydrogen-bond geometry and
hydrophobic contacts of the 3’ substituent (Scheme 1). As a
control, we also synthesized a “phenyl” analogue of GC-1,
QH9, in which the phenol hydroxy group has been replaced
by a hydrogen atom. Pyridyl analogues were derived from the
corresponding 2-substituted 4-cyanopyridines by the nucleo-
philic addition to 4-cyanopyridine of alkyl radicals generated
by silver-promoted radical decarboxylation of the corre-
sponding carboxylic acids. This method provided efficient
access to the 2-alkyl pyridine series of ligands (see Scheme S1
in the Supporting Information).[14]
The high-resolution crystal structures of T3-bound TRb
and TRa suggest that His435 forms a hydrogen bond with the
4’-OH group of T3 and participates simultaneously in aryl–
aryl interactions with the Phe459 residue of helix 12 (Figure 2,
left). We demonstrated previously that 4’-alkoxy derivatives
[*] A. Q. Hassan, J. T. Koh
Department of Chemistry and Biochemistry, University of Delaware
Newark, DE 19716 (USA)
E-mail: johnkoh@udel.edu
[**] We thank the National Institutes of Health (R01 DK54257) for
financial support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200801742.
7280
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7280 –7283

Angewandte
Chemie
Figure 1. a) The conformational change to helix 12 (H12) upon ligand binding leads to the recruitment of transcriptional coactivators (CoA).
b) The His-Phe switch involves residues on helix 11 (H11) and H12. c) The His435!Tyr mutation disrupts the normal His-Phe switch. The new
residue on H11 may be complemented by appropriate heterocyclic ligands.
Figure 2. Left: TRb(wt)/T3 cocrystal structure (PDBID: 1BSX). Right:
Modeled structure of T3 in TRb(H435Y).
All monosubstituted pyridine analogues, except QH15,
had potencies at or below 1 mm for the mutant TRb(H435Y)
(Figure 3). Although many factors influence potency, the
Scheme 1. Structure of GC-1 and QH9–QH18.
pyridine analogues QH10–QH14 all show a preference for the
mutant TRb(H435Y) over TRb(wt). In contrast, GC-1 and T3
are strongly biased (by a factor of 225 and 390, respectively)
for the wild type. Thus, the substitution of the phenol ring for
a pyridine moiety appeared to be a successful strategy to
complement the His!Tyr mutation (Figure 3b). One ana-
logue, QH13 (EC50 = 151 nm), is twice as potent as the parent
compound GC-1 and more potent than T3 with the mutant
TRbH435Y (Figure 3c).
Although QH13 is a weaker agonist with TRb(H435Y)
than T3 is with TRb(wt), QH13 shows 15-fold selectivity for
the mutant. This result corresponds to a 5850-fold improve-
ment in selectivity for the mutant relative to that of T3
(Table 1). Furthermore, QH13 shows very weak activity
towards TRa(wt) (EC50 > 10 mm, data not shown). QH10,
which has a 3’-isobutyl substituent, was also selective (by a
Angew. Chem. Int. Ed. 2008, 47, 7280 –7283
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7281

Communications
most potent and efficacious analogue QH13, like GC-1, has
an isopropyl group.
Compounds QH18 and QH17 were prepared in an
attempt to create more room within the binding site for the
larger Tyr side chain of the mutant. QH18 is related to GC-1
through the exchange of the phenol ring for an imidazole ring
to complement the corresponding replacement of the imida-
zole ring of histidine with a phenol moiety in the His!Tyr
mutation. Unfortunately, these compounds have poor poten-
cies, and accurate EC50 values could not be determined;
however, they did show partial activity at the highest
concentrations tested. The high-affinity binding of T3 and
GC-1 is due in part to the extensive hydrophobic contacts of
the receptor with the outer phenol ring of the ligand.
Therefore, it is not clear if the lower potency is due to a
geometric misalignment of hydrogen-bonding partners or to
the weaker hydrophobic contacts of the receptor with the
smaller and more polar imidazole ring.
Overall, chemical complementation of the His435!Tyr
mutant by changing the presentation of hydrogen-bonding
groups on the ligand is a successful strategy for rescuing
TRb(H435Y) function with a modest change in ligand
potency and a dramatic change in receptor selectivity relative
to those of the natural ligand T3 or GC-1. With the caveat of
having tested only a limited set of ligands, we speculate that
potency and efficacy are intrinsically more difficult to rescue
from receptor mutations that encroach into the ligand-
binding pocket, as the strategy of making a smaller ligand
to accommodate a smaller binding site leads invariably to a
diminishment in the ligand–receptor interactions needed to
maintain high potency.
Figure 3. Cellular reporter-gene response to designed analogues.
RLU=relative light units.
The role of TR mutations in cancer and other diseases
remains poorly understood. Whereas the dominant negative
actions of TRb mutants in RTH have been well documented,
the identification of identical somatic mutations in cancer
presents a unique paradox that may ultimately help reveal a
role for TR in carcinogenesis or cancer progression, or as a
potential therapeutic target. We are currently investigating
QH13 and related analogues for their ability to rescue the
tumor-suppressor properties of TR in in vitro cancer models.
Table 1: Potencies and efficacies of natural and synthetic ligands for TRb
and TRb(H435Y) on the direct repeat 4 (DR4) promoter.
Compound
TRb(wt)
EC50 [nm]
TRb(H435Y)
EC50 [nm]
Selectivity
wt/mutant
(efficacy [%])
(efficacy [%])
T3
GC-1
QH9
0.51 (100)
1.3 (93)
1400
199ꢁ84 (94)
293ꢁ23 (56)
920ꢁ43
1/390
1/225
1.5
6.8
4.1
15
2
–
–
–
–
QH10
QH11
QH13
QH14
QH15
QH16
QH17
QH18
2877 (75)
3194 (30)
2259 (65)
1018 (85)
1142 (105)
1860 (35)
n.d. (50)[a]
n.d. (35)[a]
421ꢁ25 (100)
780ꢁ33 (105)
151ꢁ20 (115)
501ꢁ32 (115)
n.d. (60)[a]
Received: April 14, 2008
Published online: August 7, 2008
Keywords: hormones · ligand design · medicinal chemistry ·
n.d. (50)[a]
.
receptors · thyroid
n.d. (40)[a]
n.d. (60)[a]
[a] Efficacy at a concentration of 10 mm (n.d.=notdetermined).
[1] P. M. Yen, Physiol. Rev. 2001, 81, 1097.
[2] S. Ando, N. J. Sarlis, E. H. Oldfield, P. M. Yen, J. Clin.
Endocrinol. Metab. 2001, 86, 5572.
[3] I. H. Chan, M. L. Privalsky, Oncogene 2006, 25, 3576.
[4] P. M. Yen, Trends Endocrinol. Metab. 2003, 14, 327.
[5] A. Q. Hassan, J. T. Koh, J. Am. Chem. Soc. 2006, 128, 8868.
[6] A. Aranda, A. Pascual, Physiol. Rev. 2001, 81, 1269.
[7] S. Williams, R. K. Bledsoe, J. L. Collins, S. Boggs, M. H.
Lambert, A. B. Miller, J. Moore, D. D. McKee, L. Moore, J.
Nichols, D. Parks, M. Watson, B. Wisely, T. M. Willson, J. Biol.
Chem. 2003, 278, 27138.
factor of 6.8), but was slightly less potent (EC50 = 421 nm;
Figure 3c). Although nearly isosteric with the pyridyl ana-
logue QH13, the phenyl analogue QH9 is six times less potent
and shows almost no (1.5-fold) selectivity for the mutant over
the wild type. Thus, the hydrogen-bonding group appears to
be critical for potency and selectivity. The nature of the
substituent at the 3’-position of the pyridine analogues has a
significant effect on their selectivity and efficacy, whereby the
[8] G. J. Kahaly, C. H. Matthews, S. Mohr-Kahaly, C. A. Richards,
V. K. K. Chatterjee, J. Clin. Endocrinol. Metab. 2002, 87, 204.
7282
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7280 –7283

Angewandte
Chemie
[9] Y. Nomura, T. Nagaya, H. Tsukaguchi, J. Takamatsu, H. Seo,
Endocrinology 1996, 137, 4082.
[12] H. F. Ye, K. E. OꢀReilly, J. T. Koh, J. Am. Chem. Soc. 2001, 123,
1521.
[10] J. D. Baxter, P. Webb, G. Grover, T. S. Scanlan, Trends Endo-
crinol. Metab. 2004, 15, 154.
[13] G. Chiellini, N. H. Nguyen, H. A. I. Yoshihara, T. S. Scanlan,
Bioorg. Med. Chem. Lett. 2000, 10, 2607.
[11] Y. H. Shi, H. F. Ye, K. H. Link, M. C. Putnam, I. Hubner, S.
Dowdell, J. T. Koh, Biochemistry 2005, 44, 4612.
[14] E. Baciocchi, E. Muraglia, G. Sleiter, J. Org. Chem. 1992, 57,
6817.
Angew. Chem. Int. Ed. 2008, 47, 7280 –7283
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7283