Discovery of a crystalline sulforaphane analog with good solid-state stability and engagement of the Nrf2 pathway in vitro and in vivo

Article abstract of DOI:10.1016/j.bmc.2018.12.026

The antioxidant natural product sulforaphane (SFN) is an oil with poor aqueous and thermal stability. Recent work with SFN has sought to optimize methods of formulation for oral and topical administration. Herein we report the design of new analogs of SFN with the goal of improving stability and drug-like properties. Lead compounds were selected based on potency in a cellular screen and physicochemical properties. Among these, 12 had good aqueous solubility, permeability and long-term solid-state stability at 23 °C. Compound 12 also displayed comparable or better efficacy in cellular assays relative to SFN and had in vivo activity in a mouse cigarette smoke challenge model of acute oxidative stress.

Full text of DOI:10.1016/j.bmc.2018.12.026

Accepted Manuscript
Discovery of a crystalline sulforaphane analog with good solid-state stability
and engagement of the Nrf2 pathway in vitro and in vivo
Jeffrey Boehm, Roderick Davis, Claudia E. Murar, Tindy Li, Brent McCleland,
Shuping Dong, Hongxing Yan, Jeffrey Kerns, Christopher J. Moody, Anthony
J. Wilson, Alan P. Graves, Mary Mentzer, Hongwei Qi, John Yonchuk, Jen-
Pyng Kou, Joseph Foley, Yolanda Sanchez, Patricia L. Podolin, Brian
Bolognese, Catherine Booth-Genthe, Marc Galop, Lawrence Wolfe, Robin
Carr, James F. Callahan
PII:
S0968-0896(18)31630-4
https://doi.org/10.1016/j.bmc.2018.12.026
BMC 14675
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
18 September 2018
6 December 2018
15 December 2018
Please cite this article as: Boehm, J., Davis, R., Murar, C.E., Li, T., McCleland, B., Dong, S., Yan, H., Kerns, J.,
Moody, C.J., Wilson, A.J., Graves, A.P., Mentzer, M., Qi, H., Yonchuk, J., Kou, J-P., Foley, J., Sanchez, Y., Podolin,
P.L., Bolognese, B., Booth-Genthe, C., Galop, M., Wolfe, L., Carr, R., Callahan, J.F., Discovery of a crystalline
sulforaphane analog with good solid-state stability and engagement of the Nrf2 pathway in vitro and in vivo,
Bioorganic & Medicinal Chemistry (2018), doi: https://doi.org/10.1016/j.bmc.2018.12.026
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Discovery of a crystalline sulforaphane analog with good solid-state stability
and engagement of the Nrf2 pathway in vitro and in vivo.
Respiratory Stress and Repair Discovery Performance Unit, GSK, Upper Providence,
PA 19426, USA
Leave this area blank for abstract info.
2

Bioorganic & Medicinal Chemistry
Discovery of a crystalline sulforaphane analog with good solid-state stability and
engagement of the Nrf2 pathway in vitro and in vivo.
Jeffrey Boehma*, Roderick Davisa, Claudia E. Murara, Tindy Lia, Brent McClelanda, Shuping Donga,
Hongxing Yana, Jeffrey Kernsa, Christopher J. Moodyc, Anthony J. Wilsonc, Alan P. Gravesb, Mary
Mentzerb, Hongwei Qib, John Yonchuka, Jen-Pyng Koua, Joseph Foleya, Yolanda Sancheza, Patricia L
Podolina, Brian Bolognesea, Catherine Booth-Genthea, Marc Galopb, Lawrence Wolfeb, Robin Carra, and
James F Callahana
a Respiratory Stress and Repair DPU, GlaxoSmithKline, Upper Providence, PA 19426, USA
b Platform Technology and Science, GlaxoSmithKline, Upper Merion, PA 19406 and Upper Providence, PA 19426, USA
c School of Chemistry, University of Nottingham, Nottingham NG7 2RD, U.K.
ARTICLE INFO
AB STR AC T.
Article history:
Received
Received in revised form
Accepted
The antioxidant natural product sulforaphane (SFN) is an oil with poor aqueous and thermal
stability. Recent work with SFN has sought to optimize methods of formulation for oral and
topical administration. Herein we report the design of new analogs of SFN with the goal of
improving stability and drug-like properties. Lead compounds were selected based on potency
in a cellular screen and physicochemical properties. Among these, 12 had good aqueous
solubility, permeability and long-term solid-state stability at 23 °C. Compound 12 also
displayed comparable or better efficacy in cellular assays relative to SFN and had in vivo
activity in a mouse cigarette smoke challenge model of acute oxidative stress.
Available online
Keywords:
Nrf2
KEAP1
Sulforaphane
Isothiocyanate
COPD
2009 Elsevier Ltd. All rights reserved.
 Corresponding author. Tel.: +1-484 923-3596; fax:
+1-610-917-4621; e-mail: jeffrey.c.boehm@gsk.com
.
1. Introduction
Among these different electrophilic Nrf2 activators, natural R-
sulforaphane (SFN) 15, is derived by the enzymatic action of
One of the pathophysiological drivers of Chronic Obstructive
Pulmonary Disease (COPD) is the imbalance between
biochemical oxidants and antioxidants resulting in oxidative
stress that leads to inflammation, tissue damage and emphysema.
Levels of expression of antioxidant genes are regulated by Nrf2
(NF-E2-related factor 2) and the associated sequestering protein
Kelch-like ECH-associated protein 1 (KEAP1). Activation of the
myrosinase (β-thioglucosidase) on glucoraphanin found in
cruciferous vegetables (Scheme 1). Chewing damages the plant
tissue which releases compartmentalized myrosinase and
glucoraphanin. Myrosinase is also found in the bacterial
microflora of the human gastrointestinal tract.6 Thus, beneficial
antioxidant effects have been attributed to cruciferous
vegetables.6a,7 Efforts have been made to activate the Nrf2
pathway by administration of both extracts from cruciferous
vegetables, as well as synthetic natural or racemic SFN. SFN
itself is an oil with poor aqueous stability and therefore the goal
of developing synthetic SFN as a drug involves the development
of a suitable formulation for oral or topical administration.8,9,10,11
1
Nrf2 pathway is a potential target for the treatment of COPD.
Cytoplasmic Nrf2 binds to the dimeric protein, KEAP1, that
targets Nrf2 for Cullin-3 (Cul3) mediated ubiquitination and
subsequent proteasome degradation. Under oxidative stress
conditions, the KEAP1/Cul3-dependent ubiquitination of Nrf2 is
disrupted and Nrf2 accumulates in the cytoplasm of the cell and
subsequently translocates to the nucleus leading to the induction
of antioxidant and cytoprotective gene expression.2,3
Analogous to the physiological effect of oxidative stress,
inactivating KEAP1/Cul3/Nrf2-mediated Nrf2 degradation by
reversible-covalent modification of reactive cysteine residues in
KEAP1, e.g., Cys151, increases the amount of Nrf2 reaching the
nucleus. This is the reported mechanism of Nrf2 activation by
many natural and synthetic electrophiles.2,4
Scheme 1. Sulforaphane is biosynthetically derived from the naturally
occurring glucosinolate, glucoraphanin.
SFN has several additional disadvantages hindering its
development as a drug, including relatively low potency in
3

functional assays which measure the induction of antioxidant
proteins and detoxifying enzymes downstream of Nrf2 activation
in cellular systems. Furthermore, the fact that natural SFN is a
chiral sulfoxide adds complexity to the chemical synthesis of the
natural isomer.12 However, as a lead compound for optimization,
SFN does have a low molecular weight (177) and is a polar
chloride.13c This route was used to effectively prepare multigram
quantities of highly purified 12.
molecule (clogP
= 0.15), leaving many possibilities for
modification to improve potency and physicochemical properties.
As part of our ongoing efforts to discover novel approaches to
the treatment of COPD, we investigated analogs related to SFN
with the goal of identifying a molecule with antioxidant-inducing
effects comparable or superior to SFN and with enhanced
capacity to be readily formulated for oral administration.
2. Results and Discussion
Scheme 3. Scalable synthetic route to high purity 12. Conditions: (a)
Boc2O, Et3N, DCM, 88%; (b) MeNH2, THF, quant.; (c) EtOH, 97%; (d)
MeNH2∙HCl, EtOH, 80 °C; (e) HCl/dioxane, 55 °C, 95%; (f) CS2, TsCl, 50-
80%; (g) SFC chromatographic purification; (h) crystallization
acetone/hexanes, 85-90%
Chemistry
The described isothiocyanates and dithiocarbamates were
prepared from primary amine precursors starting with readily
available n-alkyl diamines with the desired linker. A typical
synthesis is depicted in Scheme 2. Analogs with a urea such as 5
were synthesized from the mono-Boc protected diamines by first
preparing the p-nitrophenol carbamate 3 with p-nitrophenyl
chloroformate and then reaction of the activated carbamate with
an amine afforded mixed ureas such as 4 after Boc cleavage.
There are many methods of isothiocyanate synthesis in the
chemical literature and these methods can be optimized for
specific structures.13 In the present example, conversion to the
isothiocyanate was effected by reaction with CS2 and then
desulfurization with hydrogen peroxide affords isothiocyanate
Compound design and lead identification
The starting point for SAR assessment of the sulforaphane
analogs was the general structure depicted in Chart 1 where a
polar functionality is connected via an alkyl linker to a reactive
group.14, 15
Chart 1. The relationship of the analogs to sulforaphane.
5.13b
Additional methods of isothiocyanate synthesis are
described for other analogs in the Supplementary Materials.
Formation of dithiocarbamates such as 6 were achieved by
reaction of the isothiocyanates with the appropriate thiol in
Efforts to replace the isothiocyanate electrophile
Initially, analogs of SFN, where the isothiocyanate was
replaced by other potentially reactive functionality (Y in the
general structure in Chart 1), were assessed. These analogs
retain a methyl sulfoxide or a closely related methyl sulfone as
depicted in Table 1. Our primary method for compound
evaluation was a functional assay screen using normal human
bronchial epithelial (NHBE) cells with a 10 µM concentration of
the analog being evaluated. The level of expression of the Nrf2
pathway dependent gene heme oxygenase-1 (HO-1) was
measured after a 24 h incubation with the analog and was
expressed as fold-change relative to 10 µM of SFN determined in
the same assay Of note, in NHBE cells treatment with SFN
resulted in approximately a 4-fold induction in HO-1 gene
expression over control cells in basal state (i.e., DMSO treated)
at a 10 µM concentration used in the screening assay (data not
shown; see Biological Methods in the Supplementary
Materials for additional assay details).
dichloromethane.
Scheme 2. General synthetic route for isonitrile and dithiocarbamates.
Conditions: (a) p-NO2-PhOCOCl, Et3N, DCM, 0 to 23 °C, 57%; (b) 7-NH2-
quinoxalin-6-amine, Et3N; (c) HCl/dioxane, 57%; (d) CS2, Et3N; (e) H2O2,
44%; (f) thiol, solvent, 44%
Multigram quantities of the lead compound 12 were prepared
according to the synthetic sequence depicted in Scheme 3. This
synthesis is six steps from the HBr salt of 3-bromopropyl amine
that is commercially available, as is the subsequently formed Boc
protected intermediate 7, and the mono-protected diamine 8.
Preparation of the unsymmetrical squareamide takes advantage of
the lowered reactivity of the mono-ethoxy squareamide 9
compared with diethoxysquarate. The most challenging step was
the large scale isothiocyanate synthesis that was developed into
an efficient and scaleable procedure using CS2 and tosyl
Among the reactive groups in Table 1, the only replacement
for the isothiocyanate that was comparably effective compared to
SFN was the dithiocarbamate 22. A sulfoxythiocarbamate
analog 21 related to a set of compounds previously reported to be
significantly less electrophilic compared to SFN was about 4-fold
less potent in the cellular screen than SFN.16 Based on the results
summarized in Table 1, an additional 43 dithiocarbamates (DTC)
and 61 isothiocyanates (ITC) were prepared to assess the effects
of varying the replacements for the polar sulfoxide (X in the
general structure in Chart 1). 17
4

Table 1. 1 Induction of HO-1 gene expression in NHBE cells treated with diverse electrophiles, compared to (±)-SFN (1).a,b,c
5

Table 2. – Electrostatic potentials of polar replacements for the sulfoxide.a
a
Quantum mechanical calculations were performed in Jaguar from Schrödinger using density functional theory (DFT) with the 6-31G** basis set to assess the
electrostatic potentials of these various groups. 18
Design of polar functionality for isothiocyanate and
dithiocarbamate SFN analogs seeking improved properties for
formulation and oral administration
either the ketone or sulfone and is electronically closer in
character to the sulfoxide. Based on these calculations several
analogs were prepared as direct SFN analogs and the potency in
the NHBE cellular screen was assessed (Table 3). In the case of
the squareamide 12, the change in the polar feature had a
noticeable impact on the cellular results. However, though the
differences in this assay could be based on affinity for KEAP1,
which as proposed could be related to the electronic properties of
the polar group, the SAR is based on a cellular screen. Thus, the
results may also be affected by other properties such as aqueous
solubility and membrane permeability.
The isothiocyanates and dithiocarbamates were designed with
the goal of maximizing potency while retaining properties
consistent with the desirable physicochemical properties required
for oral administration. Some of the physicochemical features
sought included crystallinity, permeability, and aqueous
solubility. Given the low molecular weight of the lead compound
(SFN), inclusion of amides, ureas and carbamates with aromatic
substitutions was considered a viable strategy to improve
crystallinity without causing unacceptably low levels of aqueous
solubility. Therefore, many of the analogs synthesized and
evaluated were from these classes.
SAR and mechanistic assessment of a set of isothiocyanates
and dithiocarbamates
Analysis of the data for all the isothiocyanates and
dithiocarbamates with polar group variations in the NHBE
cellular screen vs the Cul3/KEAP1 TR-FRET assay17 is depicted
in Figure 1. Figure 1 clearly shows a low correlation between
the cellular potency relative to SFN (log scale) of the new
isothiocyanate (ITC) and dithiocarbamate (DTC) analogs and the
Cul3/KEAP1 TR-FRET assay pIC50 measuring the binding of full
length Cul3 and full length KEAP1.18 Furthermore, in the NHBE
cellular screen using SFN as a control, HO-1 fold-change in gene
expression has
Additionally, calculations of the electrostatic character of
polar group replacements for the sulfoxide were undertaken to
gain insight into structural features that would improve ligand
binding to KEAP1. Among the analogs proposed based on this
design strategy were a ketone and the novel squareamide
replacement of the sulfoxide as depicted in Table 2.
The calculations depicted in Table 2 show the squareamide
has a more negative electrostatic potential when compared to
6

Table 3. – Induction of HO-1 gene expression in NHBE cells by
sulfoxide mimics designed by electrostatic comparison of the
polar group.
much narrower range (1.5 log units) and many of the analogs
(including SFN) were below the limits of the Cul3/KEAP1 TR-
FRET assay (pIC50 < 4). Likewise, a TR-FRET assay measuring
binding of full length Nrf2 to full length KEAP1 detects no
measurable activity (pIC50 < 4) for many of the analogs from this
set of compounds.17 These data suggest that the biochemical
assays may not be sensitive enough to measure the effects of SFN
and the other analogs binding to KEAP1, but it is also possible
that the analogs can inhibit the degradation of Nrf2 without
disrupting the KEAP1-Cul3-Nrf2 complex19 or that the analogs
are affecting HO-1 expression independent of binding to the
KEAP1-Cul3-Nrf2 complex.
a
broad window (~ 0.04-100 fold-change), while the
Cul3/KEAP1 TR-FRET assay gives meaningful results over a
7

Figure 1. Fold-change of HO-1 gene expression in NHBE cells (average values depicted on log scale) vs. KEAP1-Cul3 TR-FRET
binding data (pIC50) colored by isothiocyanates (ITC) (green) and dithiocarbamates (DTC) (blue).
Given the lack of significant potency of many of the
molecules in the KEAP1/Cul3 TR-FRET biochemical assay, and
the fact that the SAR is based on a functional screen for increased
expression of the HO-1 gene (as well as other gene products, data
not shown), we sought additional support for the hypothesis that
the mechanism of action of these analogs is through covalent
binding to KEAP1. It has been reported that SFN binds
incubation. The mass spectra show the addition products to both
acetylated and non-acetylated forms of the protein construct. In
contrast, incubation with the C151S mutant construct yielded
only one addition product for both SFN and 12 (Figure 2).
These data demonstrate that the isothiocyanate analogs bind to
KEAP1 covalently through the BTB domain as previously
reported for SFN and supports the hypothesis that disruption of
the KEAP1-Nrf2-Cul3 complex is not required to increase Nrf2
cellular concentration19 and thus to increase HO-1 gene
expression as shown in the cellular assay screen. Therefore,
compounds potent in the cellular screen were selected for further
assessment utilizing additional cellular systems.
covalently to multiple Cys residues in the BTB and IVR domains
20,21
of KEAP1 including Cys151 in the BTB domain.2,4,19,
In
order to confirm that these analogs bind KEAP1, the interaction
of SFN and 12 with both wild type His-Keap1 (35-178) and
mutant H6huKeap1 (35-182) C151S construct was evaluated by
LC-MS. These experiments demonstrated binding of SFN and
12 to two Cys residues in the WT KEAP1 construct after a 1 h
8

Figure 2. Comparison of the LC-MS data for WT and mutant C151S protein with SFN and 12 after incubation for 1 h; A. SFN with
HisKeap1 (35-178); B. 12 at 1 h with HisKeap1 (35-178); C. SFN at 1 h with H6huKeap1 (35-182) C151S; D. 12 at 1 h with
H6huKeap1 (35-182) C151S.
+
Table 4. Representative comparison of isothiocyanate its
corresponding dithiocarbamate.
The structure-activity relationships (SAR) for the
isothiocyanates and dithiocarbamates using the cellular screen
demonstrated dependence of potency on the linker length (n = 4
appears optimal, see Table 1) as well as the polar functional
group (X in Table 1). For the dithiocarbamates, with an
additional point of structural diversity at the dithiocarbamate, the
SAR is also influenced by the substitution pattern on the
dithiocarbamate. In many cases, dithiocarbamates are more
potent inducing the expression of the HO-1 gene than the
corresponding isothiocyanates when screened at a 10 μM
concentration. This trend is illustrated in Figure 3 and an
example of the relative potencies of a parent isothiocyanate and
corresponding dithiocarbamate is shown in Table 4.
9

Figure 3. Induction of HO-1 gene expression by compounds screened at 10 μM (expressed relative to 10 μM SFN) for the parent
isothiocyanates (ITC) (green) and corresponding dithiocarbamates (DTC) (blue) analogs.
Figure 4. Induction of HO-1 gene expression by compounds screened at 10 μM (expressed relative to 10 μM SFN) vs kinetic solubility
at pH 7.4 colored by isothiocyanates (ITC) (green) and dithiocarbamates (DTC) (blue).17, 23
Solubility issues with the in vivo vehicle arose with many of
the analogs selected for further evaluation based on potency in
the cellular screen. The high throughput kinetic solubility
(utilizing chemiluminescent nitrogen detection (CLND)) of ten of
the most potent dithiocarbamates and three of the most potent
isothiocyanates from the cellular screen was evaluated (Figure
4).23
The three compounds with good kinetic solubility at pH 7.4
(compounds highlighted in the box in Figure 4) are
10

representative of the structures from the two chemical classes.
Some of the physicochemical characteristics of these compounds
equilibrium solubility from solid of compound 12 at pH 6.5 in
fasted state simulated intestinal fluid (FaSSIF),25 is relatively
are summarized in Table 5.
Furthermore, the measured
high,
with
a
value
of
8
mg/mL.
11

Table 5. Novel analogs of SFN with good kinetic solubility and efficacy (screened at 10 μM) in the NHBE cellular screen (represented
as induction of HO-1 gene expression relative to 10 μM SFN).
To evaluate cellular potency in relation to physicochemical
properties, we utilized the logarithm of the fold-change in
expression level of the HO-1 gene (relative to SFN) as a relative
measure of potency. Based on the concept of ligand efficiency26,
we utilized the log of the fold-change data from the screening
assay divided by the number of heavy atoms (HA) (multiplied by
100 for better visualization) to obtain a measure of potency in the
cellular screen relative to molecular size. Furthermore, when this
ratio is plotted against the chrom log D at pH 7.4 (an
experimentally determined measure of lipophilicity)27, the
comparison affords a depiction where fold-change in cellular
activity adjusted for molecular size is compared to lipophilicity,
similar to lipophilic ligand efficiency (LLE) (Figure 5). 26,28 The
compounds with higher potency in the cellular screen relative to
molecular size and those with lower chrom log D would be of
interest as potentially developable oral Nrf2 pathway activators.
Figure 5. Log [(induction of HO-1 gene expression relative to SFN at 10 μM/number of heavy atoms(HA)) x100] vs chrom log D27 at
pH 7.4. Colored by aromatic ring count; blue, 0; green, 1; red, 2.
12

Identification of 12 as a lead compound with improved
properties for formulation and oral administration
regulated genes, as the HO-1 gene can be regulated by several
additional transcription factors including the repressor Bach1.36
A
potential mechanistic explanation for this observation can be
attributed to the differential activity of 12 and SFN with respect
to the transcriptional repressor Bach1 as determined in a cellular
fragment complementation assay measuring the interaction of
Bach1 and MafK. In this assay, 12 was a more potent inhibitor
than SFN (pIC50 = 5.5 and 4.8, respectively) of the interaction of
Bach1-MafK, an interaction required for the binding of the
repressor and possibly contributing to increased HO-1
upregulation noted in the cellular screen (see Biological Methods
in the Supplementary Materials for additional details).
By the above criteria, 12 emerged as the most interesting
compounds in the series (Figure 5). Compound 12 also held
additional advantages such as stability as a crystalline solid (>3
years at 23 °C),29 excellent FaSSIF solubility (measured from
crystalline solid), moderate topological polar surface area and
low molecular weight (Table 5). Further, the tertiary squaramide
incorporated in 12 mimics similar polar tertiary urea and amide
analogs. For instance, since dialkyl tertiary squaramides such as
12 can only act as H-bond acceptors and not H-bond donors30
they likely mimic the difference in permeability between tertiary
and secondary amides.31 The MDR1-MDCK permeability of 12
was considered good with Papp = 12.5x10-6 cm/s (AB) and
Papp = 7.67x10-6 cm/s (BA).32
Finally, 12 was evaluated in vivo utilizing a murine cigarette
smoke challenge model that incorporates oxidative stress induced
by cigarette smoke and which is widely used to study
inflammatory responses in the airways (See Supplementary
Materials section for model details). In this model, 12 dosed
orally 24 h prior to the cigarette smoke challenge inhibited the
influx of neutrophils in a dose dependent manner (Figure 7).
Furthermore, the inhibitory effect of 12 on neutrophilic
infiltration was comparable to the effect of SFN (Figure 7).
Similar observations were made for both 12 and SFN with
respect to inhibition of total and mononuclear cell infiltration into
the airways (data not shown). Altogether, these data demonstrate
that the ability of 12 to activate the Nrf2 pathway in human
bronchial epithelial cells translates into dose-dependent inhibition
of lung inflammation in an in vivo model of pulmonary oxidative
stress, which might be relevant in diseases presenting with high
neutrophilic-driven inflammation and oxidative stress such as
COPD.37
Compound 12 was further evaluated relative to known Nrf2
activators in a related cellular assay that measures increase of
activity of an enzymatic protein, NAD(P)H dehydrogenase
[quinone] 1 (NQO1), downstream of Nrf2 activation. This assay
utilizes an immortalized human bronchial epithelial (BEAS2B)
cell line measuring increases in levels of NQO1 protein utilizing
an NQO1 enzyme assay.33 The results of this comparison to SFN
and bardoxolone methyl (CDDO-Me)34,35 are depicted in Figure
6. Compound 12 increases the levels of NQO1 (measured by
NQO1 activity) in a dose dependent manner with a potency
similar to SFN. In addition, although bardoxolone methyl is
considerably more potent than 12 and SFN, all three compounds
display similar efficacy in the assay. Interestingly, 12 and SFN
display differential upregulation of HO-1 in the NHBE cellular
screen. HO-1 has a complex promoter compared to other ARE-
Figure 6. The comparison of analog 12, SFN, and bardoxolone methyl in a cellular assay measuring increases of NQO1 protein
utilizing a NQO1 enzymatic assay. Bardoxolone methyl (CDDO-Me): red; 12: blue; (±) SFN: black. The corresponding pEC50 values
are: 5.68 (SFN), 8.06 (CDDO-Me) and 5.78 (12).
13

Figure 7. Inhibition of neutrophil infiltration by SFN and 12 in a cigarette smoke challenge model in mice.
Table 6. Profile of SFN and 12 and their GSH and NAC adducts in BEAS2B cell assay determining levels of NQO1 (measured by
NQO1 enzyme activity).
Compound 12 can
be isolated as a crystalline solid that displays good aqueous
When dosed orally (2 mg/kg) in a rat pharmacokinetic study,
12 displayed high systemic clearance although the major
metabolites were shown to be the glutathione (GSH) 28 and N-
solubility and long term solid phase stability. Compound 12 has
acetylcysteine (NAC) 29 adducts of 12 (data not shown).38
demonstrated efficacy for activation of the Nrf2 pathway in
Taking these metabolites into consideration, oral administration
BEAS2B cells that also translates to a dose-dependent inhibition
does provide a high combined systemic exposure of 12 and its
of lung inflammation in an in vivo model of pulmonary oxidative
adducts. Using a subsequently developed high throughput
stress. In conclusion, 12 is an orally bioavailable, efficacious,
version of the BEAS2B/NQO1 assay described previously, the
crystalline and more easily formulated analog of SFN that has
GSH and NAC adducts of both 12 and SFN demonstrated a
been shown to activate the Nrf2 pathway both in vitro and in vivo
similar potency compared to the corresponding isothiocyanate
and can function as a readily prepared tool for evaluation of the
containing parent molecules (Table 6). Thus, the activity of the
Nrf2 signaling pathway.
adducts is consistent with the hypothesis that these isothiocyanate
analogs exist in a reversible equilibrium with thiocarbonates of
GSH and NAC in vivo.
In addition, the use of a tertiary alkyl squaramide as a
replacement for amides or ureas to increase permeability,
aqueous solubility, and crystallinity, may be generally applicable
to other medicinal chemistry target series.
Summary
Analogs of SFN were prepared and evaluated in a cellular
screen that compared their ability to activate the Nrf2 pathway as
measured by their ability to upregulate HO-1 gene expression
relative to SFN. The more potent examples were then assessed
for physicochemical properties desirable for their formulation
and administration as oral drugs. Among the most promising
analogs, the dialkyl tertiary squaramide 12 was shown to be a
covalent Nrf2 activator that binds to the BTB domain of KEAP1.
All studies were conducted in accordance with the GSK
Policy on the Care, Welfare and Treatment of Laboratory
Animals and were reviewed the Institutional Animal Care and
Use Committee either at GSK or by the ethical review process at
the institution where the work was performed.
14

Substituents at the Sulfinyl Sulfur: Asymmetric Synthesis and
Biological Activities. J. Org. Chem. 2009 74, 6002 – 6009.
13. (a) Mukerjee, A. K.; Ashare, R. Isothiocyanates in the Chemistry
of Heterocycles. Chemistry Reviews 1991 91(1) 1-24. (b) Li, G.;
Tajima, H.; Ohtani, T. An Improved Procedure for the
Preparation of Isothiocyanates from Primary Amines by Using
Hydrogen Peroxide as the Dehydrosulfurization Reagent. J.
Org. Chem. 1997, 62, 4539-4540. (c) Wong, R.; Dolman, S. J.
Isothiocyanates from Tosyl Chloride Mediated Decomposition of
in Situ Generated Dithiocarbamic Acid Salts. J. Org. Chem.
2007, 72, 3969-3971. (d) Kim, S.; Yi, K.Y. 1,1’-Thiocarbonyldi-
2,2’-pyridone a new useful reagent for functional-group
conversions under essentially neutral conditions. J. Org.Chem.
1986 2613-2615.
14. Posner, G.H.; Cho, C-G; Green, J.V.; Zhang, Y.; Talalay, P.
Design and Synthesis of Bifunctional Isothiocyanate Analogs of
Su1foraphane: Correlation between Structure and Potency as
Inducers of Anticarcinogenic Detoxication Enzymes. J. Med.
Chem. 1994 37, 170-176
15. Wilson, A. J.; Kerns, J.K.; Callahan, J.F.; Moody, C.J. Keap
Calm, and Carry on Covalently J. Med Chem. 2013;56(19),7463-
76 and references therein.
16. Electrophilic tuning of the chemoprotective natural product
sulforaphane. Ahn, Y-H; Hwang, H.; Liu, H.; Wang, X.J.;
Zhankg, Y.; Stephenson, K.K.; Boronina, T.N.; Cole, R.N.;
Dinkova-Kostova, A.T.; Talalay, P.; Cole, P.A. PNAS 2010
107(21) 95(0-9595.
17. A listing of the structures of these 104 compounds with the
levels of HO-1 mRNA induced relative to SFN as well as TR-
FRET data for inhibition of Cul3 binding to Keap1 are in a table
attached to the Supplementary Materials. Except where noted on
the Supplementary Materials experimentals, these compounds
had good LC-MS data with LC purity ≥ 95% and 1H NMR data
consistent with the structures.
18. Poore, D. D., Hofmann, G., Wolfe III, L. A., Qi, H., Jiang, M.,
Fischer, M., Wu, Z., Sweitzer, T. D., Chakravorty, S., Donovan,
B., Li, H. Development of a High-Throughput Cul3-Keap1
Time-Resolved Fluorescence Resonance Energy Transfer (TR-
FRET) Assay for Identifying Nrf2 Activators. SLAS Discovery,
2018, 1-15. https://doi.org/10.1177/2472555218807698.
19. Baird, L., Dinkova-Kostova, A. T. Diffusion dynamics of the
Keap1-Cullin3 interaction in single live cells. Biochem. Biophys.
Res. Comm. 2013, 433 (1), 58-65.
20. Bochevarov, A.D.; Harder, E.; Hughes, T.F.; Greenwood, J.R.;
Braden, D.A.; Philipp, D.M.; Rinaldo, D.; Halls, M.D.; Zhang,
J.; Friesner, R.A. Jaguar A high-performance quantum
chemistry software program with strengths in life and materials
sciences. Int. J. Quantum Chem., 2013, 113(18), 2110-2142.
21. Cleasby A., Yon J., Day P.J., Richardson C., Tickle I.J.,
Williams P.J., Callahan J.F., Carr R., Concha N., Kerns J.K., Qi
H., Sweitzer T., Ward P., Davies T.G. Structure of the BTB
Domain of Keap1 and Its Interaction with the Triterpenoid
Antagonist CDDO. PLOSONE
Acknowledgements
The authors would like to thank Nicole Goodwin for
providing scientific review and helpful suggestions.
GlaxoSmithKline provided funding to C.J. Moody and A.J.
Wilson.
References and notes
1. Boutten, A.; Goven, D.; Artaud-Macari, E.; Boczkowski, J.;
Bonay, M. NRF2 targeting: a promising therapeutic strategy in
chronic obstructive pulmonary disease. Trends in Molecular
Medicine 2011 17(7), 363 -371.
2. (a) Eggler, A.L.; Gay, K.A.; Mesecar;A.D; Molecular
mechanisms of natural products in chemoprevention: Induction
of cytoprotective enzymes by Nrf2. Mol. Nutr. Food Res. 2008,
52, S84 –S94; (b) Baird, L.; Dinkova-Kostova, A. T.; The
cytoprotective role of the Keap1-Nrf2 pathway Arch Toxicol
2011 85, 241 – 272; (c) Lu, M.C.; Ji, J.A.; Jiang, Z.Y.; You
Q.D.; The Keap1–Nrf2–ARE Pathway As a Potential Preventive
and Therapeutic Target: An Update. Med. Res. Rev., 2016, 36,
924–963.
3. An alternative to small molecule activation of the Nrf2 pathway
through covalent interaction with nucleophilic cysteines on the
Keap1 protein is through direct, non-covalent, competition with
Nrf2 binding to the KEAP1 Kelch domain. For example see:
Davies,T.G.; Wixted W.E., Coyle J.E., Griffiths-Jones C., Hearn
K., McMenamin R., Norton D., J. Rich S.J., Richardson C.,
Saxty G., Willems H.M.G., Woolford A.J-A., Cottom J.E., Kou
J-P., J.G., Feldser H.G., Sanchez Y., Foley J.P., Bolognese B.J.,
Logan G., Podolin P.L., Yan H., Callahan J.F., Heightman T.D.,
Kerns J.K., Monoacidic Inhibitors of the Kelch-like ECH-
Associated Protein 1: Nuclear Factor Erythroid 2-Related Factor
2 (KEAP1:NRF2) Protein–Protein Interaction with High Cell
Potency Identified by Fragment-Based Discovery. J. Med.
Chem. 2016 59(8) 3991–4006.
4. Zhang, D.D.; Lo, S-C.; Cross, J.V.; Templeton, D.J.; Hannick,
M. Keap1 Is a Redox-Regulated Substrate Adaptor Protein for a
Cul3-Dependent Ubiquition Ligase Complex. Mol. And Cell.
Biol. 2004 24(24) 10941 – 10953.
5. In this paper where the chiral, natural isomer of sulforaphane is
referred to, it will be referred to as “natural SFN”. Racemic
sulforaphane was used for testing, as a synthetic intermediate for
conversion to some analogs and for comparison with analogs.
The racemic sulforaphane will be denoted as SFN or 1 without
any further designation.
6. (a) Dinkova-Kostova1, A.T.; Kostov, R.V. Glucosinolates and
isothiocyanates in health and disease. Trends in Molecular
Medicine 2012 18(6) 337-347; (b) Wittstock, U.; Halkier, B.A.
Glucosinolate research in the Arabidopsis era. Trends in Plant
Science 2002 7(6) 263-270.
https://doi.org/10.1371/journal.pone.0098896
7. Fahey, J.W.; Talalay, P. Antioxidant Functions of Sulforaphane:
a Potent Inducer of Phase II Detoxication. Enzymes Food and
Chemical Toxicology 1999 37 973-979.
8. Egner, P. A.; Chen, J-G.; Zarth, A. T.; Ng, D. K.; Wang, J-B.;
Kensler, K. H.; Jacobson, L. P.; Munoz, A.; Johnson, J. L.;
Groopman, J. D. Rapid and Sustainable Detoxication of
Airborne Pollutants by Broccoli Sprout Beverage: Results of a
Randomized Clinical Trial in China. Cancer Prevention
Research 2014, 7(8), 813-823.
9. Doa, D.P.; Paib, S.B.; Rizvib, S.A.A.; D’Souza, M.J.
Development of sulforaphane-encapsulated microspheres for
cancer epigenetic therapy. International Journal of
Pharmaceutics 2010 386, 114–121.
22. Hu, C.; Eggler, A.L.; Mesecar, A.D.; van Breemen R.B.;
Modification of Keap1 Cysteine Residues by Sulforaphane.
Chem. Res. Toxicol. 2011 24, 515–521.
23. Bhattachar, S.N.; Wesley, J.A.; Seadeek, C. Evaluation of the
chemiluminescent nitrogen detector for solubility determinations
to support drug discovery. Journal of Pharmaceutical and
Biomedical Analysis 2006 41 152–157.
24. Calculated as described in Ertl P., Rohde B, Selzer P. Fast
calculation of molecular polar surface area as a sum of fragment-
based contributions and its application to the prediction of drug
transport properties. J Med Chem. 2000 43(20), 3714-3717,
using software available through Chemaxon
https://chemaxon.com/
10. Danafar H.; Sharafi A.; Manjili H.K.; Andalib S. Sulforaphane
delivery using mPEG–PCL co-polymer nanoparticles to breast
cancer cells. Pharmaceutical Development and Technology
2017 22(5), 642-651.
11. U.S. Patent US7879822 (Stabilized Sulforaphane A method of
stabilizing sulforaphane is provided. The method includes
contacting sulforaphane, or an analog thereof, and a cyclodextrin
to form a complex between the sulforaphane, or analog thereof,
and the cyclodextrin.). U.S. Patent US20130143963
http://evgen.com/
25. Augustijns, P.; Wuyts, B.; Hens, B.; Annaert, P.; Butler, J.;
Brouwers, J. A review of drug solubility in human intestinal
fluids: Implications for the prediction of oral absorption.
European Journal of Pharmaceutical Sciences 2014 57 322–332.
And references therein.
26. Shultz, M.D. Setting expectations in molecular optimizations:
Strengths and limitations of commonly used composite
parameters. Biorg. and Med. Chem. Let. 2013 23, 5980-5991.
27. Young, R.J.; Green, D.V.S.; Luscombe, C.N.; Hill, A.P. Getting
physical in drug discovery II: the impact of chromatographic
hydrophobicity measurements and aromaticity. Drug Discovery
Today 2011 16( 17/18 ) 822 - 830.
12. Khiar, N.; Werner, S.; Mallouk,S.; Lieder,F.; Alcudia, A.;
Fernandez,I. Enantiopure Sulforaphane Analogues with Various
15

28. Hopkins, A.L.; Keserű, G.M.; Leeson; P.L., Rees; D.C.,
Reynolds; C.H. The Role of Ligand Efficiency Measures in
Drug Discovery. Nature Rev. Drug Disc. 2014 13(2): 105-121.
29. The stability was determined by checking LC-MS and 1H NMR
data after > 3 years at 23 °C in a desiccator and noting no change
in the spectral data over this time.
38. Perez, H.L.; Junnotula, V.; Knecht, D.; Nie, H.; Sanchez, Y.;
Boehm, J.C.; Booth-Genthe, C.; Yan, H.; Davis, R.; Callahan,
J.F. Analytical approaches for quantification of a Nrf2 pathway
activator: overcoming bioanalytical challenges to support a
toxicity study. Analyst 2014, 139, 1902-1912.
30. Storer, R.I.; Aciro, C.; Jones, L.H. Squaramides: physical
properties, synthesis and applications. Chem. Soc. Rev. 2011 40,
2330–2346.
31. Conradi, R.A.; Hilgers, A.R.; Ho, N.F.H.; Burton, P.S. The
Influence of Peptide Structure on Transport Across Caco-2 Cells.
II. Peptide Bond Modification Which Results in Improved
Permeability. Pharm. Res 1992 9, 435 – 439.
Supplementary Data
Supplementary data associated with this article can be found,
the online version, at
https://doi.org/10.1016/j.bmc.2018.01.005.
in
32. (a) Volpe, D.A. Drug-permeability and transporter assays in
Caco-2 and MDCK cell lines. Future Med. Chem. 2011 3(16),
2063–2077; (b) Thiel-Demby, V.E.; Humphreys, J.E.; St. John
Williams, L.A.; Ellens, H.M.; Shah, N.; Ayrton, A.D.; Polli, J.W.
Biopharmaceutics Classification System: Validation and
Learnings of an in Vitro Permeability Assay. Molecular
Pharmaceutics 2008 6, 11–18.
33. Dinkova-Kostova, A. T.; Liby, K.T.; Stephenson, K.K.;
Holtzclaw, W.D.; Xiangqun Gao, X.; Suh, N.; Charlotte
Williams, C.; Risingsong, R.; Tadashi Honda, T.; Gordon W.
Gribble, G.W.; Sporn, M.B., Talalay, T. Extremely potent
triterpenoid inducers of the phase 2 response: Correlations of
protection against oxidant and inflammatory stress. Proc. Natl.
Acad. Sci. U.S.A. 2005 102 (12) 4584–4589.
Experimental procedures for all the chemical structures
appearing in the Schemes and tables and biological methods and
a
table containing the listing of the structures of the
isothiocyanates and dithiocarbamates included in Figures 1, 3, 4
and 5 with the levels of induced HO-1 gene expression (screened
at 10 μM) relative to 10 μM SFN as well as TR-FRET data for
inhibition of Cul3 and NRF2 binding to KEAP1 are included as
Supplemental Materials. All compound structures are supported
by LC-MS data with LC purity ≥ 95% and 1H NMR data.
Abbreviations Used
34. Liby, K. T.; Sporn, M. B. Synthetic oleanane triterpenoids:
multifunctional drugs with a broad range of applications for
prevention and treatment of chronic disease. Pharmacol. Rev.
2012, 64, 972−1003.
35. Sporn, M. B.; Liby, K. T.; Yore, M. M.; Fu, L.; Lopchuk, J. M.;
Gribble, G. W. New synthetic triterpenoids: potent agents for
prevention and treatment of tissue injury caused by inflammatory
and oxidative stress. J. Nat. Prod. 2011, 74, 537−545.
36. Zhou Y, Wu H, Zhao M, Chang C, Lu Q. The Bach Family of
Transcription Factors: A Comprehensive Review. Clin Rev
Allergy Immunol. 2016, 50(3), 345-56.
SFN, sulforaphane; KEAP1, Kelch-like ECH-associated
protein 1; GSH, glutathione; NAC, N-acetyl cysteine; NHBE,
Normal Human Bronchial Epithelial cells; NQO1, NAD(P)-
H:quinone acceptor oxidoreductase 1; Nrf2, Nuclear factor
erythroid 2−related factor 2; BEAS2B; Bronchial epithelium +
Adenovirus 12-SV40 virus hybrid (Ad12SV40), an immortalized
human bronchial epithelial cell line; COPD, chronic obstructive
pulmonary disease.
37. Barnes, P.J. Inflammatory mechanisms in patients with chronic
obstructive pulmonary disease. J Allergy Clin Immunol. 2016,
138(1), 16-27.
.
16

Table 1. Induction of HO-1 gene expression in NHBE cells
treated with diverse electrophiles, compared to (±)-SFN (1).a,b,c
X = polar group
Y = reactive group
n = 4-6
fold-change
Compound
Number
fold-change
relative to
SFN
Compound
Number
X
Y
n
relative to
SFN
X
Y
n
(±)-1
4
1
18
4
0.14
13
14
3
5
0.47
0.35
19
20
5
5
0.11
0.13
15
6
0.15
21
22
4
4
0.27
0.93
16
17
5
5
0.33
0.18
aCell assay using normal human bronchial epithelial (NHBE)
cells with a 10 µM concentration of the test article. The level of
expression of the Nrf2 pathway gene, heme oxygenase-1 (HO-1),
is normalized and expressed as a fold-change relative to 10 µM
b
of (±)-SFN (1). Schemes and experimental procedures for the
preparation of compounds 13-22 are in the Supplementary
c
Materials. Racemic SFN was used for analog synthesis and
evaluation
17

Table 2. Electrostatic potentials of polar replacements for the
sulfoxide.a
Squaramide
Ketone
Sulfone
Sulfoxide
-43.554.1
-40.332.4
-48.629.3
-36.819.4
-49.0
54.1
aQuantum mechanical calculations were performed in Jaguar
from Schrödinger using density functional theory (DFT) with
the 6-31G** basis set to assess the electrostatic potentials of
these various groups.18
18

Table 3. – Induction of HO-1 gene expression in NHBE cells by
sulfoxide mimics designed by electrostatic comparison of the
polar group.
a
1
0
fol
d-
Co
mp
oun
d
Nu
mb
er
ch
an
ge
µ
M
Structure rel
c
o
n
c
e
n
t
ati
ve
to
SF
Na
(±)-
1
r
a
t
1
i
o
n
24
.9
12
23
o
f
t
h
e
0.
9
t
0.
7
e
s
t
24
a
a
r
t
C
e
l
i
l
c
l
e
.
T
h
e
a
s
s
a
y
u
s
l
i
n
g
e
v
e
l
N
H
B
E
o
f
e
x
p
r
e
s
s
i
c
e
l
l
s
w
i
o
n
t
h
o
19

f
µ
H
O
-
M
o
f
1
i
(
s
±
)
n
o
r
m
a
l
-
S
F
N
(
1
)
.
i
z
e
d
a
n
d
e
x
p
r
e
s
s
e
d
a
s
a
f
o
l
d
-
c
h
a
n
g
e
r
e
l
a
t
i
v
e
t
o
1
0
20

Table 4. Representative
comparison of
isothiocyanate its
corresponding
dithiocarbamate
fo
ld
-
c
h
a
n
g
Co
m
po
un
d
N
u
m
be
r
e
Structure
r
el
at
iv
e
to
S
F
N
a
(±
)-1
1
1
3.
2
5
6
1
0
1.
8
aCell assay using NHBE
cells with a 10 µM
concentration of the test
article. The level of
expression of HO-1 is
normalized and expressed
as a fold-change relative
to 10 µM of (±)-SFN (1).
21

Table 5. Novel analogs of SFN with good kinetic solubility and
efficacy (screened at 10 μM) in the NHBE cellular screen
(represented as induction of HO-1 gene expression relative to 10
μM SFN.
fold-change
relative to
SFNa
kinetic
solubility
(µM)
Compound KEAP1/Cul3
Chrom
logD25
Structure
MW
tpsab
Number
FRET pIC50
5.1
102
432.6
82
1.52
≥ 400
6
4.7
4.3
28
25
248.3
253.3
41
53
4.59
2.39
≥ 418
≥ 394
25
12
aCell assay using NHBE cells with a 10 µM concentration of the
test article. The level of expression of HO-1 is normalized and
expressed as a fold-change relative to 10 µM of (±)-SFN (1).
btpsa = topological polar surface area.22
22

Table 6. Profile of SFN and 12 and their GSH and NAC adducts
in BEAS2B cell assay determining levels of NQO1 (measured by
NQO1 enzyme activity).
Com
poun
d
Num
ber
Com
poun
d
Num
ber
p
E
C5
p
E
C5
Structure
Structure
0
0
5.
8
5.
8
(±)-1
12
5.
8
5.
8
26
27
28
29
5.
8
5.
8
Activity was determined in BEAS2B cells by determining
increase in NQO1 activity (as a measure of NQO1 protein level)
after incubation of compounds for 48 h.; NAC = N-acetyl
cysteine and GSH = glutathione.
23