Rapid assembly of a library of lipophilic iminosugars via the thiol-ene reaction yields promising pharmacological chaperones for the treatment of Gaucher disease

Article abstract of DOI:10.1021/jm201633y

A highly divergent route to lipophilic iminosugars that utilizes the thiol-ene reaction was developed to enable the rapid synthesis of a collection of 16 dideoxyiminoxylitols bearing various different lipophilic substituents. Enzyme kinetic analyses revea

Full text of DOI:10.1021/jm201633y

Article
pubs.acs.org/jmc
Rapid Assembly of a Library of Lipophilic Iminosugars via the
Thiol−Ene Reaction Yields Promising Pharmacological Chaperones
for the Treatment of Gaucher Disease
Ethan D. Goddard-Borger,† Michael B. Tropak,‡ Sayuri Yonekawa,‡ Christina Tysoe,† Don J. Mahuran,‡
and Stephen G. Withers*,†
†Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia, V6T 1Z1, Canada
‡Division of Genetics and Genome Biology, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, M5G 1X8, Canada
S
* Supporting Information
ABSTRACT: A highly divergent route to lipophilic iminosugars that utilizes
the thiol−ene reaction was developed to enable the rapid synthesis of a
collection of 16 dideoxyiminoxylitols bearing various different lipophilic
substituents. Enzyme kinetic analyses revealed that a number of these
products are potent, low-nanomolar inhibitors of human glucocere-
brosidase that stabilize the enzyme to thermal denaturation by up to
20 K. Cell based assays conducted on Gaucher disease patient derived fibroblasts demonstrated that administration of the com-
pounds can increase lysosomal glucocerebrosidase activity levels by therapeutically relevant amounts, as much as 3.2-fold in cells
homozygous for the p.N370S mutation and 1.4-fold in cells homozygous for the p.L444P mutation. Several compounds elicited
this increase in enzyme activity over a relatively wide dosage range. The data assembled here illustrate how the lipophilic moiety
common to many glucocerebrosidase inhibitors might be used to optimize a lead compound’s ability to chaperone the protein in
cellulo. The flexibility of this synthetic strategy makes it an attractive approach to the rapid optimization of glycosidase inhibitor
potency and pharmacokinetic behavior.
effective in type I GD patients, this therapy is expensive and
does not address the neuronopathologies of types II and III GD
owing to the inability of the enzyme to cross the blood−brain
barrier. The other therapy approved for GD is substrate reduc-
tion therapy (SRT).6,7 This approach reduces the amount of
GC in a patient’s lysosomes by inhibiting glucosylceramide synthase
(GCS), the enzyme that creates this ubiquitous lipid. GCS inhibitors
currently approved for this therapy have poor selectivity and
exhibit various off-target effects,8,9 although more selective in-
hibitors and better therapeutic protocols continue to be
developed.2,10 Regardless, one could make the case that this
concept does not address the root cause of the disease state
(insufficient GBA activity) and that alternative therapies that
remedy this dearth of activity may have greater therapeutic
potential than SRT.
There are over 200 known GD-associated missense muta-
tions in GBA, and they do not appear to be localized to any one
region of the enzyme.11 Most of these mutations are remote to
the enzyme’s active site, suggesting that the majority affects
GBA activity indirectly, most likely by deleteriously affecting
folding of the enzyme. There is evidence to suggest that this
is the case for the most common GD-associated mutation
(p.N370S).12,13 Such mutant enzymes have difficulty obtaining
and retaining their native fold within the lumen of the endo-
plasmic reticulum (ER), which means that a greater proportion
INTRODUCTION
■
The lysosomal storage disorders (LSDs) are a collection of rare
genetic diseases characterized by the accumulation of various
metabolites within the lysosomes of the cell. These disorders
are caused by mutations in genes encoding various lysosomal
proteins, which leads to the impaired catabolism of the
metabolite(s) in question.1,2 The most prevalent of the LSDs is
Gaucher disease (GD), which occurs at a frequency of around 1
in 40 000 births, although it is more prevalent in some ethnic
groups.3 There are three clinical types of GD, the symptoms of
which have been summarized well in a review.3 Although it is
worth reiterating that while all GD types may present with
hepatosplenomegaly, pancytopenia, and bone abnormalities, it
is only types II and III that feature neuronopathologies. These
problems are ultimately caused by insufficient glucocerebrosi-
dase (GBA) activity within the lysosomes of GD patients, which
is a direct result of mutations in the encoding gene GBA1. GBA
is a retaining β-glucosidase from glycoside hydrolase (GH) family
30 (http://www.cazy.org) that hydrolyzes the glycosidic bond of
β-D-glucosylceramide (GC).4 It is the accumulation of GC that
ultimately leads to the various pathologies of GD, which may in
part result from perturbed intracellular calcium homeostasis in
compartments such as the endoplasmic reticulum.5
Two therapeutic approaches are currently approved for the
treatment of GD. The first and most common treatment is
enzyme replacement therapy (ERT), which involves the direct
administration of recombinantly expressed human GBA to
supplement the patient’s levels of GBA activity.3 While this is
Received: December 2, 2011
Published: February 23, 2012
© 2012 American Chemical Society
2737
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
of the enzyme undergoes ER-associated degradation (ERAD)
rather than reaching maturity and being trafficked to the
lysosome.14 The variability in GBA activity in normal individuals is
quite high,15,16 although type I GD patients typically have activity
values that are less than 20% of the mean enzyme activity found
in healthy individuals.15,17 If one could increase the flux of the
GBA mutant to the lysosomes within the cells of a GD patient,
then it might be possible to ameliorate the disease. This goal is
the basis of enzyme enhancement therapy (EET), a third, as yet
clinically unproven, approach to the treatment of GD.18−20 The
EET concept proposes that small molecules could be used to
bind and stabilize correctly folded mutant enzymes within the
ER, thereby elevating the steady-state concentration of folded
enzyme within this organelle and increasing the amount of
enzyme trafficked to lysosomes. Importantly, once in the acidic
and substrate-rich environment of the lysosome, the mutant
enzyme should be sufficiently stable and catalytically active to
perform its catabolic duties with or without the small molecule
bound. Such molecules are referred to as “pharmacological
chaperones” (PCs), for obvious reasons.
While it is true that a small-molecule chaperone of GBA
might stabilize the enzyme through allosteric interactions alone,
the vast majority of GBA chaperones identified to date do so
through active site interactions.20 This is not to say that occupa-
tion of the enzyme’s active site is necessarily the best way to
stabilize a given GBA mutant but rather reflects the fact that
high affinity, active-site-binding ligands can be rationally designed
and screened for with relative ease. A typical GBA competitive
inhibitor possesses two important features: (1) a D-glucopyranose-
mimicking moiety to occupy the sugar binding pocket of GBA
and (2) a lipophilic fragment to approximate the ceramide
aglycone of the natural substrate.2 This general design philos-
ophy has led to many known glucose-like glycosidase inhibitors
being modified with simple alkyl groups to generate some quite
potent GBA inhibitors, some of which are capable of chaperon-
ing the enzyme in various cell-based assays (Figure 1).
given that the lipophilic portion of such compounds is likely to
be the most tolerant of structural variation, a significant contrib-
utor to inhibitor potency (courtesy of the hydrophobic effect),
and a strong determinant of a given compound’s pharmacokinetic
properties. Perhaps one reason that few have explored different
lipophilic moieties at this position is that the synthesis and
evaluation of such compounds as PCs using established pro-
tocols would be a laborious process. Indeed, it would be partic-
ularly arduous to prepare alkyl-chain-modified collections of the
more potent glucocerebrosidase inhibitors/chaperones like the
alkylated isofagomines 1 and dideoxyiminoxylitols 2, the syn-
theses of which involve Grignard reactions that restrict divergence
and limit the chemical diversity of the lipophilic moiety. A highly
divergent method for the construction of different lipophilic
iminosugars based on compound 1 or 2 would provide an
excellent opportunity to explore the biological consequences of
this lipophilic appendage on the ability of a compound to
chaperone GBA.
It was imagined that the thiol−ene reaction might be used to
realize such a goal: the anti-Markovnikov addition of a given
thiol to a deprotected, alkene-containing iminosugar would provide
a new lipophilic iminosugar, ready for testing, in a single chem-
ical step. In this scenario, a single sulfur atom would serve as the
only apparent ligature between the sugar-mimicking and lipo-
philic fragments. This low profile “scar” of ligation was con-
sidered preferable to those left by other methods like, for
example, the comparatively bulky triazole of an azide−alkyne
cycloaddition.29 If need be, the fully carbon-linked versions of the
best candidates could later be synthesized and reasonably
expected to possess similar affinities for GBA.
This concept was explored on the vinyl derivative of the
dideoxyiminoxylitols 2: the allylamine 3 (Figure 2). This
particular class of compounds was chosen because simple alkyl
chain derivatives of 2 have proven to be potent and selective
GBA inhibitors and competent pharmacological chaperones.28
Figure 2. The thiol−ene reaction offers a highly divergent route to
lipophilic iminosugars.
RESULTS AND DISCUSSION
■
Synthesis. The synthesis of allylamine 3 was accomplished
in seven steps from D-xylose (Scheme 1). The first three steps
of the sequence involved preparation of glycosylamine 4 using
established protocols.31 This product was treated with
vinylmagnesium bromide to provide the alkene 5 in good
yield as the major diastereomeric product. The configuration of
the new stereocenter would later be unequivocally assigned using
single crystal X-ray diffraction experiments on the ultimate
product, allylamine 3 (Figure 3B). Dehydrative ring closure of
compound 5 was effected with methanesulfonyl chloride to
provide the piperidine 6. Attempts at global deprotection of
piperidine 6 in a single step using various reducing metals in
liquid ammonia were unsuccessful; in each case the N-(4-
methoxybenzyl) group remained intact while the alkene was
rapidly reduced. Thus, the N-(4-methoxybenzyl) moiety was
first removed by the action of ceric ammonium nitrate and
water to provide amine 7, albeit in modest yield. A subsequent
Figure 1. Structures of some GBA pharmacological chaperones.20−28
While the quest for a good GBA pharmacological chaperone
(PC) has seen significant effort invested into evaluating various
glucose-mimicking moieties,20−28 considerably less attention
has been devoted to exploring how the structure of the lipophilic
fragment might influence the properties of the chaperone.29,30
For the most part, only simple straight-chain alkyl groups have
been used as ceramide surrogates. This is somewhat surprising,
2738
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
vinyl group. Following this result, a superior deprotection strat-
egy was devised whereby the N-(4-methoxybenzyl) group of 6
was initially substituted for a benzyloxycarbonyl group, by the
action of benzyl chloroformate, to provide the carbamate 8.
This transformation provided consistently high yields and a
product that was easier to purify and more stable than the
amine 7. A Birch reduction of the carbamate 8, again in the pre-
sence of excess allylamine, provided the desired product 3 in
high yield. A small sample of the carbamate 8 was also subjected
to hydrogenolysis to return an authentic sample of the
iminosugar 9, a byproduct of the Birch reductions.
Scheme 1. Synthesis of the vinyl-dideoxyiminoxylitol 3
Within the literature, a number of different methods are
commonly used to initiate the radical thiol−ene reaction.32−34
No single technique appears to work well in a general manner,
and some substrate combinations simply provide poor yields
under all conditions trialed. As a result, the radical thiol−ene
reaction cannot really be considered a true “click” reaction,
although it is certainly still a chemical transformation of great
utility.35,36 Taking this into consideration, a number of thermal
and photochemical radical initiation methods were investigated
for the addition of thiols to alkene 3 (as per Figure 2). Good to
moderate yields of adduct could be isolated when the reaction
was promoted by AIBN in degassed methanol at reflux (Table 1),
although the reaction proved to be particularly sensitive to
the purity of both substrates. The thiol−ene adducts 10−25
were thereby synthesized using a range of different thiols to
give products containing aromatic groups, amides, ureas, and
straight, branched, and cyclic aliphatic chains of various length
and size.
Others have noted that some dideoxyiminoxylitols appear
to adopt an unusual 2C5 conformation in solution, as evidenced
1
by H NMR spectroscopy, where all of the piperidine ring’s
hydroxyl groups are oriented axially.28 The 1H NMR spectra of
the hydrochlorides of compounds 3 and 9−25 suggest that
2
Birch reduction of 7 provided the desired product 3, although it
was necessary to perform this reaction in the presence of a vast
excess of sacrificial allylamine to minimize reduction of the
these piperidinium ions also adopt the C5 conformation in
solution. However, at a higher pH, once deprotonated, the
5
piperidine reverts to the C2 conformation (Figure 3A). This
pH-dependent change in conformation is consistent with the
behavior of other β-hydroxypiperidines.37
2
The C5 conformation of the piperidinium ion was also ob-
served in the solid state, as revealed by an X-ray diffraction ex-
periment on a single crystal of 3·HCl (Figure 3B). This crystal
structure also enabled stereochemical assignment of the chiral
center that was created in the conversion of glycosylamine 4
into alkene 5 (Scheme 1).
It was imagined that this pH-dependent conformational
change might be exploited to modulate the relative affinity of
the iminosugars for the enzyme in the ER (pH ≈ 7.0) and
lysosome (pH ≈ 5.5), since it seemed unlikely that the enzyme
active site would accommodate a ligand with axial hydroxyl
groups. To test this idea, the pKa values of a subset of iminosugars
from Table 1 were measured and the IC50 values of each
compound against GBA determined at pH 7.0 and pH 5.5
(Table S1). Unfortunately, very little difference in inhibition of
GBA was observed between these pH values, and because
inhibition pH profiles are a composite of the pKa values for free
enzyme, free inhibitor, and the enzyme−inhibitor complex, it is
perilous to attempt to rationalize this result.38−40
Inhibition of, Specificity for, and Thermal Stability
Enhancement of GBA. The inhibition constants of
compounds 3 and 9−25, as well as the parent compound
dideoxyiminoxylitol (2, R = H), against GBA were determined
at pH 5.5 (Table 2). All of these compounds were found to
1
Figure 3. (A) H NMR spectra of 9 reveal different conformations at
high and low pH. (B) Drawing of 3·HCl with thermal ellipsoids shown
at the 50% probability level.
2739
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
a
the Ki determined for compound 12 (1.8 nM) and that pre-
viously reported for its sulfur-to-methylene homologue (2, R =
n-nonyl; 2.2 nM)28 are essentially the same. This suggests that
the sulfur ligature behaves as a simple methylene surrogate on
enzyme.
Table 1. Thiol−Ene Reactions on Alkene 3
Several trends are apparent from the data in Table 2. First,
modification of the dideoxyiminoxylitol with small groups, such
as the vinyl and ethyl groups of compounds 3 and 9, does not
significantly alter the parent iminosugar’s affinity for GBA.
However, further elaboration resulted in significant improve-
ments in binding, as evidenced in the simplest case of com-
pound 10, which is a 15-fold better inhibitor than the vinyl
derivative 3 from which it was derived. A comparison of compounds
10−13, which all contain straight-chain aliphatic modifications,
revealed that extension of this lipophilic moiety by just two
methylene units improves inhibition by around an order of
magnitude each time. Further, it appears that addition of a
single branching methyl group can account for a 2- to-3-fold
improvement in Ki, as observed for 10/14, 11/16, and 14/15.
Collectively, these data illustrate in a quantitative fashion how
simple modifications of the lipophilic moiety can be used to
tune the potency of GBA inhibitors.
Compounds 21−25, which contain amide or urea moieties,
are inherently less lipophilic than the other compounds prepared
here, and yet they still retain a good affinity for GBA. It is note-
worthy that compounds 22 and 23, which both contain aryl
groups, were significantly better inhibitors than those amide-
containing inhibitors without aryl groups.
The specificity of this collection of inhibitors for GBA over
other cellular glycosidases was subsequently evaluated. Remarkably,
none of the compounds inhibited human α-glucosidase, α-
galactosidase, or β-galactosidase even at concentrations as high
as 200 μM, with the one exception being compound 13, which
inhibited human neutral cytosolic β-glucosidase41 with an IC50
of 2.0 μM. This exquisite specificity validated these chaperone
candidates as target-specific.
Compounds shown to act as pharmacological chaperones for
GBA (or other lysosomal enzymes for that matter) also stabilize
the enzyme against thermal denaturation. Since the ability of a
pharmacological chaperone to stabilize GBA is most relevant in
the ER, the melting temperature of this protein was determined
at pH 7.0 in the absence and presence of saturating concentra-
tions of each inhibitor. The difference between the melting
temperatures of free and inhibitor-bound GBA for each com-
pound is presented in Table 2. An inverse linear relationship
was observed between the Ki of each compound and the in-
crease in melting temperature (Figure 4). Thus, the parent
iminosugar dideoxyiminoxylitol 2, R = H, with the highest Ki
(2.8 μM) increases the melting temperature of GBA by just
3.6 K, whereas compound 13 with the lowest Ki (0.4 nM)
increases the melting temperature of GBA by 20 K.
Cell-Based Assay for Pharmacological Chaperone
Activity. GD patient fibroblasts homozygous for the
p.N370S or p.L444P mutation were treated with each of the
compounds in Table 2 to determine if any could behave as
pharmacological chaperones for GBA in cellulo. The parent
iminosugar dideoxyiminoxylitol 2, R = H, failed to improve
GBA activity at the concentrations evaluated for either cell line.
The simple vinyl and ethyl derivatives 3 and 9, all of which are
hydrophilic, low micromolar inhibitors of GBA, showed
minimal levels of enzyme enhancement (Figure 5). All of the
other compounds were capable of increasing GBA activity at
one concentration or another in fibroblasts homozygous for the
a
Alkene 3 (0.1 mmol), thiol (0.5 mmol), and AIBN (20 μmol) were
refluxed in MeOH for 24 h.
inhibit GBA in a competitive manner. Every thiol−ene product
had a submicromolar Ki, with one, compound 13, having a sub-
nanomolar Ki that makes it among the tightest-binding com-
petitive GBA inhibitors reported to date. It is worth noting that
2740
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
Table 2. Comparison of Kinetic, Biophysical, and Biological Parameters for the Dideoxyiminoxylitol Derivatives
c
AC50 (nM)
max. increase
a
b
subclass
parent
compd
Ki (nM)
ΔTm (K)
p.N370S
24000
p.L444P
44620
p.N370S
1.5
p.L444P
1.1
2,R=H
2800 400
2700 400
1100 200
180 20
3.6
linear
3
9
3.4
4.4
1800
1700
7000
1.4
2.1
1.3
1.2
30000
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
7
600
18
300
12
2.5
2.4
2.1
1.6
2.3
2.3
2.0
2.2
2.2
2.1
2.2
2.7
2.3
2.5
2.7
3.2
1.3
1.3
1.3
1.2
1.4
1.2
1.2
1.4
1.3
1.1
1.1
1.3
1.1
1.3
1.3
1.3
18
2
11
15
20
7.8
8.4
14
11
13
13
14
8.2
11
11
10
7.6
1.8 0.4
0.4 0.2
76 12
2.7
1.0
38
0.6
0.6
2.2
25
branched
cyclic
29
1
49
4.4 0.6
1.2
5.3
1.3
1.6
2.8
2700
90
0.3
1.2
0.6
2.6
2.1
2900
64
6
1
1
4
19
2
3.0 0.7
200 20
amide/urea
12
30
2
2
190
800
4500
63
250 20
500
1500
450 50
a
b
c
Determined at pH 5.5 using wild type GBA. All compounds evaluated at 100 μM and pH 7.0 using wild type GBA. Concentration at which 50%
of maximum increase in intracellular GBA activity was achieved in Gaucher patient fibroblasts.
pound, the GBA activity in the cell lysate decreased (Figure 6)
while, with the exception of compound 13, the activity of lysosomal
β-N-acetyl hexosaminidase remained unchanged at each compound
concentration. To determine if these effects were due to residual
inhibitor in the cell lysate or toxicity of the compounds, the viability
of GD fibroblasts grown in the presence of compounds 2 (R = H),
3, and 9−25 at three different concentrations (10, 33, and
100 μM) was determined using a resazurin-based spectropho-
tometric assay. The results (Figure S1) indicate that only com-
pounds 12 and 13 elicit a significant reduction in cell viability at
the concentrations tested. Compound 12 exhibits toxicity at
100 μM, while compound 13 is toxic at concentrations as low
as 10 μM. It is noteworthy that these two compounds are toxic
at concentrations much greater than the concentrations at
which they provide an enhancement in GBA activity.
The data in Table 2 also reveal a strong linear correlation
between Ki and the concentration at which a half maximal
increase in GBA activity (AC50) is observed (Figure 7). It is of
note that even though compounds 21−25 are among the least
lipophilic compounds in the library and have only intermediate
Ki values, they still managed to promote the greatest enhance-
ment in GBA activity for both cell lines. It may be that these
compounds strike the right balance of bioavailability and inhib-
itory activity to maximize enzyme enhancement. It could also
be that the amide moiety is processed within the lysosome,
decreasing the affinity of the resultant ligand or aiding in clearance
of the inhibitor from this organelle. Certainly these amide-con-
taining derivatives are promising candidates with great potential
for further structural optimization. Another promising candi-
date is compound 14, which appears to provide relatively good
improvements in GBA activity for both mutant cell lines over a
wide range of concentrations, alluding to the possibility of a
broad therapeutic dosage window. Together these compounds
provide ample opportunities for further structural optimization,
Figure 4. Relationship between Ki and ΔTm.
p.N370S or p.L444P mutations in GBA (Figure 5, Table 2). Far
greater improvements in GBA activity were observed for cells
harboring the p.N370S mutation than for those with the
p.L444P mutation, as has been seen with all the pharmaco-
logical chaperones of GBA evaluated to date. Interestingly,
while the most lipophilic compounds with very low Ki values
like 12 and 13 were able to chaperone at the lowest concentra-
tions, it was the amide/urea-containing compounds 21−25 that
elicited the most impressive improvements in GBA activity,
managing a 2.3- to 3.2-fold increase in GBA activity in the
p.N370S cell line and a 1.3-fold increase in the p.L444P cell
line. Thus, in addition to effects that the lipophilic groups have
on the inhibitory potency of the compounds these pendent groups
also influence the chaperoning efficacy of drugs, possibly through
effects on their bioavailability.
A subset of four compounds (13, 14, 20, and 23) were
selected from each of the structural subclasses on the basis of
their appended ligature (linear, branched, cyclic, and amide/
urea) and were evaluated over a wider range of concentrations.
This revealed that above a certain concentration for each com-
2741
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
Figure 5. Increased GBA activity in compound-treated fibroblasts derived from a GD patient homozygous for (A) the p.N370S GBA mutation and
(B) the p.L444P GBA mutation. Patient fibroblasts were grown in the presence of each of the compounds for 5 days. GBA activity in the lysates from
treated cells was normalized to activity in untreated cells. IFG is isofagomine.
Figure 7. Plot examining the correlation of AC50 and Ki.
substituent can have a pronounced effect on the iminosugar’s
Ki. This is in turn closely related to the ability of the compound
to stabilize the target enzyme against thermal denaturation
and the minimal concentration at which it chaperones GBA in
cellulo. However, the magnitude of increase in GBA activity
elicited by a given chaperone appears to be independent of its
Ki and likely related to subtle, difficult-to-predict pharmacoki-
netic effects that are also largely determined by the compound’s
lipophilic appendage. The present collection of compounds
contains chaperones that possess various desirable properties,
including activity at low concentrations, large therapeutic dosage
Figure 6. GBA activity in GD patient fibroblasts homozygous for the
p.N370S GBA mutation when treated with representative compounds
from each subclass. Patient fibroblasts were grown in the presence of
windows, and very large increases in GBA activity. The methods
described here will enable further structural optimization of
the pharmacological chaperone leads identified in this work for
the treatment of Gaucher disease. Additionally, this synthetic
approach may be applicable to the identification and optimi-
zation of inhibitors of other glycosidases for use as pharmacological
chaperones to treat other LSDs.
each of the compounds for 5 days. GBA activity in lysates from treated
cells was normalized to activity in untreated cells.
which will assist in the fine-tuning of their pharmacokinetic
properties. Such second-generation compounds may provide
still greater increases in GBA activity or attain similar increases
at a lower dose.
EXPERIMENTAL SECTION
CONCLUSIONS
■
■
General Materials and Methods. All reagents for synthesis were
obtained from Sigma-Aldrich (U.S.) and were of reagent grade. All
reagents were used without further purification except for the com-
mercially obtained thiols, which were redistilled just prior to use. Thin
layer chromatography (TLC) was performed on Merck silica gel
60 F254 plates. Flash chromatography was performed with Merck silica
By adoption of the thiol−ene reaction as a means to ligate
various lipophilic molecules to a common iminosugar, an effi-
cient and divergent method to prepare collections of phar-
macological chaperone candidates has been developed, allowing
the systematic exploration of the consequences of modifying
the hydrophobic moiety therein. Results obtained with the
collection of lipophilic dideoxyiminoxylitol GBA inhibitors pre-
pared here illustrate how subtle modifications of the lipophilic
1
gel 60 (230−400 mesh). H and 13C NMR spectra (referenced using
the residual solvent peak or using an internal MeOH standard in the
case of 13C for D2O) were recorded on a Bruker AV-300 or AV-400
2742
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
instrument. High-resolution mass spectra of all compounds were
obtained in the mass spectrometry laboratory of the Department of
Chemistry at University of British Columbia, Canada. Pure GBA was
obtained as Cerezyme from Genzyme (U.S.). Purified human cytosolic
neutral β-glucosidase was kindly provided by Dr. N. Juge (Institue of
Food Research, Norwich, U.K.).41 A lysosomal enriched concanavalin
A fraction from human placenta was used as a source of human
β-galactosidase, α-galactosidase, and α-glucosidase.42 Concanavalin A
conjugated to agarose beads was obtained from GE Healthcare (U.S.).
2,4-Dinitrophenyl β-D-glucopyranoside was prepared according to the
procedure of Sharma et al.,43 while the fluorogenic enzyme substrates
(4-methylumbelliferyl β-D-galactopyranoside, 4-methylumbelliferyl
α-D-galactopyranoside, 4-methylumbelliferyl β-D-glucopyranoside, 4-
methylumbelliferyl α-D-glucopyranoside, and 4-methylumbelliferyl
N-acetyl-β-D-glucosaminide) were obtained commercially from
Sigma-Aldrich (U.S.) or Toronto Research Chemicals (Canada).
NanoOrange was obtained from Invitrogen (U.S.). Gaucher fibroblast
cell lines homozygous for the p.N370S and p.L444P mutations in GBA
were obtained from the cell line repositories at the Toronto Hospital
for Sick Kids and the Coriell Institute for Medical Research, respec-
tively. The purity of compounds 19−25 was determined to be >95%
by analytical HPLC, while all other compounds were judged to be
>95% pure by elemental microanalysis.
Enzymology. All GBA kinetics were performed at 37 °C using a
buffer at pH 5.5 or pH 7.0 that contained 20 mM citric acid, 50 mM
Na2HPO4, 1.0 mM tetrasodium EDTA, 0.25% v/v Triton X-100, and
0.25% w/v taurocholic acid. To determine the Ki of each inhibitor for
GBA, initial reaction rates were measured at three substrate con-
centrations for each of a range of inhibitor concentrations (typically
seven concentrations bracketing the Ki ultimately determined such that
0.2Ki < [I] < 5Ki). These data were fit to a competitive inhibition
model using nonlinear regression analysis, as performed by the GraFit
5.0.13 program, to provide the Ki. Dixon and Lineweaver−Burke plots
of each data set validated the use of a competitive inhibition model.
For inhibitors with Ki values greater than 20 nM, this value was deter-
mined using a continuous UV spectrophotometric assay where the
final GBA concentration was 4.0 nM and 2,4-dinitrophenyl β-D-
glucopyranoside served as the substrate. Reactions were initiated by
the addition of GBA, and the time-dependent increase in absorption at
400 nm, as determined using a Varian Cary 4000 or Varian Cary 300
UV−vis spectrophotometer, gave the initial reaction rate. This increase
in absorbance was linear for all measurements over a period of 3 min.
For inhibitors with Ki values less than 20 nM, this value was deter-
mined using a discontinuous fluorimetric assay where the final GBA
concentration was 10 pM and 4-methylumbelliferyl β-D-glucopyrano-
side served as the substrate. Reactions were initiated by the addition of
GBA, and after 4, 10, and 15 min, 150 μL aliquots of the reaction
mixture were removed and diluted into a cuvette containing 450 μL of
glycine buffer (1.0 M, pH 10.8). The fluorescence of each sample
(λex = 365 nm, λem = 460 nm) was determined using a Varian Cary
Eclipse fluorimeter. The three time points were fit by linear regression
to provide the initial reaction rate. In each case, an excellent linear fit
was obtained over this 15 min interval. The inhibitory activities of
compounds against selected human lysosomal enzymes (α-glucosidase,
α-galactosidase, and β-galactosidase) were evaluated at a single, high
concentration of 200 μM, in triplicate. All reactions were performed
using 100 mM citrate−phosphate buffer at pH 4.5 supplemented with
0.025% w/v human serum albumin. The activities of human lysosomal
α-glucosidase, α-galactosidase, and β-galactosidase in a lysosomal
enzyme enriched ConA-bound fraction from human placenta were
monitored using an equal volume of the fluorogenic substrates [4-
methylumbelliferyl α-D-glucopyranoside (3.0 mM), 4-methylumbelli-
feryl α-D-galactopyranoside (1.0 mM), and 4-methylumbelliferyl β-D-
galactopyranoside (0.5 mM), respectively] at 37 °C. All reactions were
terminated by the addition of a 4-fold excess of 100 mM 2-amino-2-
methylpropanol (MAP), and fluorescence was measured using a
Molecular Devices M2 fluorimeter (λex = 365 nm, λem = 460 nm).
Inhibitory activities against purified cytosolic neutral human β-
glucosidase were determined by incubating a 3-fold dilution series of
the compound in the presence of the enzyme for 15 min at room
temperature, followed by the addition of an equal volume of substrate
(10 mM 4-methylumbelliferyl β-D-glucopyranoside) in 100 mM
citrate−phosphate buffer at pH 7.0. Each mixture was incubated at
37 °C for 30 min. The reaction was terminated by the addition of a
4-fold excess of 100 mM MAP and the fluorescence determined as
described above. IC50 values were deduced by nonlinear fitting of the
data to the appropriate model using Prism Graphpad, version 5.2.
Differential Scanning Fluorimetry. The melting temperatures of
human GBA (1 μg) in the presence of compounds 2 (R = H), 3, and
9−25 (200 μM) were determined by scanning differential fluorimetry,
using NanoOrange (1/100 dilution as per the manufacturer's protocol)
as the fluorescent reporter, in accordance with established protocols.44
Cell-Based Assay for GBA Chaperone Activity. Patient
fibroblasts homozygous for the p.N370S mutation or p.L444P
mutation were grown in 96-well tissue culture plates (10−20000
cells/well) in α-MEM medium supplemented with 10% v/v fetal
calf serum at 37 °C in a CO2 humidified incubator. Chaperoning
efficiencies of the compounds were evaluated by preparing a series of
3-fold dilutions starting at 10 mM in water. Compounds were directly
added to the media at a 1/100 dilution. Following incubation of the cells
for 5 days in the compound-supplemented medium, lysates in 100 mM
citrate−phosphate buffer at pH 5.5 containing 0.4% v/v Triton X-100
and 0.2% w/v taurocholic acid were prepared as described.45 GBA activity
was measured using 10 mM 4-methylumbelliferyl β-D-glucopyranoside,
and for compound treated cells, this value was normalized against the
GBA activity measured in mock-treated control cells (n = 3). Similar
activity measurements were also performed for lysosomal hexosaminidase
(Hex) using 3.2 mM 4-methylumbelliferyl N-acetyl-β-D-glucosaminide as
the substrate.
Thiol−Ene Reaction Protocol. A solution of AIBN (3.3 mg,
20 μmol) in degassed MeOH (1.0 mL) was added dropwise over 8 h
to a solution of thiol (0.50 mmol) and alkene 3 (16 mg, 0.10 mmol) in
degassed MeOH (2.0 mL) refluxing under an atmosphere of N2. The
solution was refluxed for a further 16 h, concentrated to dryness, and
then subjected to flash chromatography. The product was dissolved in
HCl (2.0 mL, 0.1 M) and the solution concentrated to dryness. The
residue was dissolved in H2O and passed through a Waters tC18 Sep-
Pak (2 g, H2O/MeOH, 1:0 to 1:4). Fractions containing product were
combined and lyophilized.
ASSOCIATED CONTENT
* Supporting Information
■
S
pKa and IC50 data for selected iminosugars, solution and refine-
ment data for the single crystal X-ray diffraction experiment,
cell viability assay data, experimental and characterization data,
1H and 13C NMR spectra for all novel compounds, and
crystallographic information file in txt format. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Telephone: (604) 822-3402. E-mail: withers@chem.ubc.ca.
■
Notes
The authors declare the following competing financial
interest(s): Patent protection has been sought by the Hospital
for Sick Children (Toronto) and the University of British
Columbia.
ACKNOWLEDGMENTS
■
This work was funded by the Canadian Institutes of Health
Research through a TEAM grant. E.D.G.-B thanks the Canadian
Institutes of Health Research for a postdoctoral fellowship. S.G.W.
acknowledges support from the Canada Research Chairs Program,
the Canadian Foundation for Innovation, and the B.C. Knowledge
Development Fund. Jennifer Dhammi is acknowledged for
assistance in the laboratory.
2743
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
Gaucher disease: physicokinetic and immunologic studies of the
residual enzyme in cultured fibroblasts from non-neuronopathic and
neuronopathic patients. Am. J. Med. Genet. 1985, 21, 529−549.
(18) Parenti, G. Treating lysosomal storage diseases with
pharmacological chaperones: from concept to clinics. EMBO Mol.
Med. 2009, 1, 268−279.
(19) Valenzano, K. J.; Khanna, R.; Powe, A. C. Jr.; Boyd, R.; Lee, G.;
Flanagan, J. J.; Benjamin, E. R. Identification and characterization of
pharmacological chaperones to correct enzyme deficiencies in
lysosomal storage disorders. Assay Drug Dev. Technol. 2011, 9, 213−
235.
ABBREVIATIONS USED
■
LSD, lysosomal storage disorder; GD, Gaucher disease; GBA,
glucosidase, beta, acid (glucocerebrosidase); GH, glycoside
hydrolase (glycosidase); GC, β-D-glucopyranosyl ceramide;
ERT, enzyme replacement therapy; SRT, substrate reduction
therapy; GCS, glucosylceramide synthase; ER, endoplasmic
reticulum; ERAD, endoplasmic reticulum associated degrada-
tion; EET, enzyme enhancement therapy; PC, pharmacological
chaperone; IFG, isofagomine.
(20) Benito, J. M.; Garcia Fernandez, J. M.; Mellet, C. O.
Pharmacological chaperone therapy for Gaucher disease: a patent
review. Expert Opin. Ther. Pat. 2011, 21, 885−903.
REFERENCES
■
(1) Futerman, A. H.; van Meer, G. The cell biology of lysosomal
storage disorders. Nat. Rev. Mol. Cell Biol. 2004, 5, 554−565.
(2) Wennekes, T.; van den Berg, R. J.; Boot, R. G.; van der Marel,
G. A.; Overkleeft, H. S.; Aerts, J. M. Glycosphingolipids: nature,
function, and pharmacological modulation. Angew. Chem., Int. Ed. Engl.
2009, 48, 8848−8869.
(3) Grabowski, G. A. Phenotype, diagnosis, and treatment of
Gaucher’s disease. Lancet 2008, 372, 1263−1271.
(4) Vocadlo, D. J.; Davies, G. J. Mechanistic insights into glycosidase
chemistry. Curr. Opin. Chem. Biol. 2008, 12, 539−555.
(5) Vitner, E. B.; Platt, F. M.; Futerman, A. H. Common and
uncommon pathogenic cascades in lysosomal storage diseases. J. Biol.
Chem. 2010, 285, 20423−20427.
(6) Pastores, G. M.; Giraldo, P.; Cherin, P.; Mehta, A. Goal-oriented
́
therapy with miglustat in Gaucher disease. Curr. Med. Res. Opin. 2009,
25, 23−37.
(7) Butters, T. D.; Dwek, R. A.; Platt, F. M. Imino sugar inhibitors for
treating the lysosomal glycosphingolipidoses. Glycobiology 2005, 15,
43R−52R.
(8) Belmatoug, N.; Burlina, A.; Giraldo, P.; Hendriksz, C. J.; Kuter,
D. J.; Mengel, E.; Pastores, G. M. Gastrointestinal disturbances and
their management in miglustat-treated patients. J. Inherited Metab. Dis.
2011, 34, 991−1001.
(9) Porto, C.; Cardone, M.; Fontana, F.; Rossi, B.; Tuzzi, M. R.;
Tarallo, A.; Barone, M. V.; Andria, G.; Parenti, G. The pharmacological
chaperone N-butyldeoxynojirimycin enhances enzyme replacement
therapy in Pompe disease fibroblasts. Mol. Ther. 2009, 17, 964−971.
(10) Marshall, J.; McEachern, K. A.; Chuang, W. L.; Hutto, E.; Siegel,
C. S.; Shayman, J. A.; Grabowski, G. A.; Scheule, R. K.; Copeland, D. P.;
Cheng, S. H. Improved management of lysosomal glucosylceramide
levels in a mouse model of type 1 Gaucher disease using enzyme and
substrate reduction therapy. J. Inherited Metab. Dis. 2010, 33, 281−289.
(11) Hruska, K. S.; LaMarca, M. E.; Scott, C. R.; Sidransky, E.
Gaucher disease: mutation and polymorphism spectrum in the
glucocerebrosidase gene (GBA). Hum. Mutat. 2008, 29, 567−583.
(12) Steet, R. A.; Chung, S.; Wustman, B.; Powe, A.; Do, H.;
Kornfeld, S. A. The iminosugar isofagomine increases the activity of
N370S mutant acid beta-glucosidase in Gaucher fibroblasts by several
mechanisms. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 13813−13818.
(13) Wei, R. R.; Hughes, H.; Boucher, S.; Bird, J. J.; Guziewicz, N.;
van Patten, S. M.; Qiu, H.; Pan, C. Q.; Edmunds, T. X-ray and
biochemical analysis of N370S mutant human acid β-glucosidase.
J. Biol. Chem. 2011, 286, 299−308.
(21) Butters, T. D. Gaucher disease. Curr. Opin. Chem. Biol. 2007, 11,
412−418.
(22) Chang, H.-H.; Asano, N.; Ishii, S.; Ichikawa, Y.; Fan, J.-Q.
Hydrophilic iminosugar active-site-specific chaperones increase
residual glucocerebrosidase activity in fibroblasts from Gaucher
patients. FEBS J. 2006, 273, 4082−4092.
(23) Luan, Z.; Higaki, K.; Aguilar-Moncayo, M.; Li, L.; Ninomiya, H.;
Nanba, E.; Ohno, K.; García-Moreno, M. I.; Ortiz Mellet, C.; García
Fernan
́
dez, J. M.; Suzuki, Y. A fluorescent sp2-iminosugar with
pharmacological chaperone activity for Gaucher disease: synthesis and
intracellular distribution studies. ChemBioChem 2010, 11, 2453−2464.
(24) Wang, G.-N.; Reinkensmeier, G.; Zhang, S.-W.; Zhou, J.; Zhang,
L.-R.; Zhang, L.-H.; Butters, T. D.; Ye, X.-S. Rational design and
synthesis of highly potent pharmacological chaperones for treatment
of N370S mutant Gaucher disease. J. Med. Chem. 2009, 52, 3146−
3149.
(25) Zhu, X.; Sheth, K. A.; Li, S.; Chang, H.-H.; Fan, J.-Q. Rational
design and synthesis of highly potent β-glucocerebrosidase inhibitors.
Angew. Chem., Int. Ed. 2005, 44, 7450−7453.
(26) Trapero, A.; Alfonso, I.; Butters, T. D.; Llebaria, A.
Polyhydroxylated bicyclic isoureas and guanidines are potent
glucocerebrosidase inhibitors and nanomolar enzyme activity
enhancers in Gaucher cells. J. Am. Chem. Soc. 2011, 133, 5474−5484.
(27) Oulaïdi, F.; Front-Deschamps, S.; Gallienne, E.; Lesellier, E.;
Ikeda, K.; Asano, N.; Compain, P.; Martin, O. R. Second-generation
iminoxylitol-based pharmacological chaperones for the treatment of
Gaucher disease. ChemMedChem 2011, 6, 353−361.
(28) Compain, P.; Martin, O. R.; Boucheron, C.; Godin, G.; Yu, L.;
Ikeda, K.; Asano, N. Design and synthesis of highly potent and
selective pharmacological chaperones for the treatment of Gaucher’s
disease. ChemBioChem 2006, 7, 1356−1359.
(29) Diot, J. D.; Moreno, I. G.; Twigg, G.; Mellet, C. O.; Haupt,
K.; Butters, T. D.; Kovensky, J.; Gouin, S. G. Amphiphilic
1-deoxynojirimycin derivatives through click strategies for chemical
chaperoning in N370S Gaucher cells. J. Org. Chem. 2011, 76, 7757−
7768.
(30) Ardes-Guisot, N.; Alonzi, D. S.; Reinkensmeier, G.; Butters,
T. D.; Norez, C.; Becq, F.; Shimada, Y.; Nakagawa, S.; Kato, A.;
Bleriot, Y.; Sollogoube, M.; Vauzeilles, B. Selection of the biological
́
activity of DNJ neoglycoconjugates through click length variation of
the side chain. Org. Biomol. Chem. 2011, 9, 5373−5388.
(31) Wennekes, T.; van den Berg, R. J. B. H. N.; Boltje, T. J.;
Donker-Koopman, W. E.; Kuijper, B.; van der Marel, G. A.; Strijland,
A.; Verhagen, C. P.; Aerts, J. M. F. G.; Overkleeft, H. S. Synthesis and
evaluation of lipophilic aza-C-glycosides as inhibitors of glucosylcer-
amide metabolism. Eur. J. Org. Chem. 2010, 1258−1283.
(32) Uygun, M.; Tasdelen, M. A.; Yagci, Y. Influence of type of
initiation on thiol−ene “click” chemistry. Macromol. Chem. Phys. 2010,
211, 103−110.
(33) Griesbaum, K. Problems and possibilities of the free-radical
addition of thiols to unsaturated compounds. Angew. Chem., Int. Ed.
Engl. 1970, 9, 273−287.
(34) Wittrock, S.; Becker, Y.; H., K. Synthetic vaccines of tumor-
associated glycopeptide antigens by immune-compatible thioether
(14) Ron, I.; Horowitz, M. ER retention and degradation as the
molecular basis underlying Gaucher disease heterogeneity. Hum. Mol.
Genet. 2005, 14, 2387−2398.
(15) Legini, E.; Orsini, J. J.; Hung, C.; Martin, M.; Showers, A.;
Scarpa, M.; Zhang, X. K.; Keutzer, J.; Muhl, A.; Bodamer, O. A.
Analysis of glucocerebrosidase activity in dry blood spots using tandem
mass spectrometry. Clin. Chim. Acta 2011, 412, 343−346.
(16) Li, Y.; Scott, C. R.; Chamoles, N. A.; Ghavami, A.; Pinto, B. M.;
Turecek, F.; Gelb, M. H. Direct multiplex assay of lysosomal enzymes
in dried blood spots for newborn screening. Clin. Chem. 2004, 50,
1785−1796.
(17) Grabowski, G. A.; Goldblatt, J.; Dinur, T.; Kruse, J.;
Svennerholm, L.; Gatt, S.; Desnick, R. J. Genetic heterogeneity in
2744
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745

Journal of Medicinal Chemistry
Article
linkage to bovine serum albumin. Angew. Chem., Int. Ed. 2007, 46,
5226−5230.
(35) Fiore, M.; Marra, A.; Dondoni, A. Photoinduced thiol−ene
coupling as a click ligation tool for thiodisaccharide synthesis. J. Org.
Chem. 2009, 74, 4422−4425.
(36) Koo, S. P. S; Stamenovic, M. M.; Prasath, R. A.; Inglis, A. J.;
Du Prez, F. E.; Barner-Kowollik, C.; van Camp, W.; Junkers, T.
Limitations of radical thiol−ene reactions for polymer−polymer
conjugation. J. Polym. Sci., Part A: Polym. Chem. 2010, 48, 1699−1713.
(37) Asensio, J. L.; Canada, F. J.; García-Herrero, A.; Murillo, M. T.;
̃
Fernan
́
dez-Mayoralas, A.; Johns, B. A.; Kozak, J.; Zhu, Z.; Johnson, C. R.;
Jimenez-Barbero, J. Conformational behavior of aza-C-glycosides:
́
experimental demonstration of the relative role of the exo-anomeric
effect and 1,3-type interactions in controlling the conformation of regular
glycosides. J. Am. Chem. Soc. 1999, 121, 11318−11329.
(38) van Bueren, A. L.; Popat, S. D.; Lin, C.-H.; Davies, G. J.
Structural and thermodynamic analyses of α-L-fucosidase inhibitors.
ChemBioChem 2010, 11, 1971−1974.
(39) Gloster, T. M.; Meloncelli, P. J.; Stick, R. V.; Zechel, D.; Vasella,
A.; Davies, G. J. Glycosidase inhibition: an assessment of the binding
of 18 putative transition-state mimics. J. Am. Chem. Soc. 2007, 129,
2345−2354.
(40) Knowles, J. R. The intrinsic pKa-values of functional groups in
enzymes: improper deductions from the pH-dependence of steady-
state parameters. Crit. Rev. Biochem. 1976, 4, 165−173.
(41) Tribolo, S.; Berrin, J. G.; Kroon, P. A.; Czjzek, M.; Juge, N. The
crystal structure of human cytosolic β-glucosidase unravels the sub-
strate aglycone specificity of a family 1 glycoside hydrolase. J. Mol. Biol.
2007, 370, 964−975.
(42) Mahuran, D.; Lowden, J. A. The subunit and polypeptide
structure of hexosaminidases from human placenta. Can. J. Biochem.
1980, 58, 287−294.
́
(43) Sharma, S. K.; Corrales, G.; Penades, S. Single step
stereoselective synthesis of unprotected 2,4-dinitrophenyl glycosides.
Tetrahedron Lett. 1995, 36, 5627−5630.
(44) Kornhaber, G. J.; Tropak, M. B.; Maegawa, G. H.; Tuske, S. J.;
Coales, S. J.; Mahuran, D. J.; Hamuro, Y. Isofagomine induced
stabilization of glucocerebrosidase. ChemBioChem 2008, 9, 2643−
2649.
(45) Hill, T.; Tropak, M. B.; Mahuran, D.; Withers, S. G. Synthesis,
kinetic evaluation and cell-based analysis of C-alkylated isofagomines
as chaperones of β-glucocerebrosidase. ChemBioChem 2011, 12,
2151−2154.
2745
dx.doi.org/10.1021/jm201633y | J. Med. Chem. 2012, 55, 2737−2745