Structure-Based Optimization and Discovery of M3258, a Specific Inhibitor of the Immunoproteasome Subunit LMP7 (β5i)

Article abstract of DOI:10.1021/acs.jmedchem.1c00604

Proteasomes are broadly expressed key components of the ubiquitin-dependent protein degradation pathway containing catalytically active subunits (β1, β2, and β5). LMP7 (β5i) is a subunit of the immunoproteasome, an inducible isoform that is predominantly expressed in hematopoietic cells. Clinically effective pan-proteasome inhibitors for the treatment of multiple myeloma (MM) nonselectively target LMP7 and other subunits of the constitutive proteasome and immunoproteasome with comparable potency, which can limit the therapeutic applicability of these drugs. Here, we describe the discovery and structure-based hit optimization of novel amido boronic acids, which selectively inhibit LMP7 while sparing all other subunits. The exploitation of structural differences between the proteasome subunits culminated in the identification of the highly potent, exquisitely selective, and orally available LMP7 inhibitor 50 (M3258). Based on the strong antitumor activity observed with M3258 in MM models and a favorable preclinical data package, a phase I clinical trial was initiated in relapsed/refractory MM patients.

Full text of DOI:10.1021/acs.jmedchem.1c00604

pubs.acs.org/jmc
Article
Structure-Based Optimization and Discovery of M3258, a Specific
Inhibitor of the Immunoproteasome Subunit LMP7 (β5i)
Markus Klein,* Michael Busch, Manja Friese-Hamim, Stefano Crosignani, Thomas Fuchss,
Djordje Musil, Felix Rohdich, Michael P. Sanderson, Jeyaprakashnarayanan Seenisamy,
Gina Walter-Bausch, Ugo Zanelli, Philip Hewitt, Christina Esdar, and Oliver Schadt
Cite This: J. Med. Chem. 2021, 64, 10230−10245
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Proteasomes are broadly expressed key components of the
ubiquitin-dependent protein degradation pathway containing catalytically
active subunits (β1, β2, and β5). LMP7 (β5i) is a subunit of the
immunoproteasome, an inducible isoform that is predominantly expressed
in hematopoietic cells. Clinically effective pan-proteasome inhibitors for
the treatment of multiple myeloma (MM) nonselectively target LMP7 and
other subunits of the constitutive proteasome and immunoproteasome
with comparable potency, which can limit the therapeutic applicability of
these drugs. Here, we describe the discovery and structure-based hit
optimization of novel amido boronic acids, which selectively inhibit LMP7
while sparing all other subunits. The exploitation of structural differences
between the proteasome subunits culminated in the identification of the
highly potent, exquisitely selective, and orally available LMP7 inhibitor 50 (M3258). Based on the strong antitumor activity
observed with M3258 in MM models and a favorable preclinical data package, a phase I clinical trial was initiated in relapsed/
refractory MM patients.
LMP7, play an essential role in restoring homeostasis in cells
under elevated proteotoxic or oxidative stress.9
INTRODUCTION
■
Proteasomes are large multicatalytic complexes that are central
components of the cellular machinery for maintenance of
protein homeostasis.1−3 The barrel-shaped 20S core particle of
proteasomes consists of 28 subunits assembled in four rings of
seven subunits each. The core particle of the broadly expressed
constitutive proteasome (cP) contains the distinct proteolytic
subunits β1 (PSMB6), β2 (PSMB7), and β5 (PSMB5), which
possess caspase-like, trypsin-like, and chymotrypsin-like
proteolytic activity, respectively. The differential substrate
preferences of each subunit enable the processing of diverse
ubiquitinated proteins in cells. A distinct core proteasome
particle called the immunoproteasome (iP) contains the
unique proteolytic subunits β1i (LMP2 and PSMB9), β2i
(MECL1 and PSMB10), and β5i (LMP7 and PSMB8) with
chymotrypsin-like, trypsin-like, and chymotrypsin-like activity,
respectively. The iP is predominantly expressed in cells of
hematolymphoid origin4−7 and can be induced in other cell
types by exposure to inflammatory stimuli such as IFNγ or
TNFα.8
The essentiality of proteasomes for the viability of multiple
myeloma (MM) cells has been underpinned by the approval
and wide application of the proteasome inhibitors bortezomib,
carfilzomib, and ixazomib. These drugs interfere with the
activity of multiple cP and iP proteolytic subunits.10 While this
nonselective mechanism delivers robust clinical efficacy in
MM, it is also associated with diverse toxicities including
thrombocytopenia, neutropenia, and cardiotoxicity,10 which
can lead to dose reductions, less frequent regimens, or
cessation of treatment and thus limit the therapeutic potential
of these drugs.11−14 The restricted expression and unique
functional features of iP-specific subunits have led to
considerable interest in the potential of selective iP inhibitors
in diverse disease settings.3,12,15 Recent reports have suggested
that therapeutic efficacy in preclinical models of inflammation
and autoimmunity requires dual inhibition of the iP subunits
LMP2 and LMP7.15−18 However, in the case of preclinical
Received: April 1, 2021
Published: July 6, 2021
Aside from its essential function in the maintenance of
protein homeostasis, the iP also degrades pathogenic proteins
and generates peptidic fragments, which are more efficiently
loaded on the class I major histocompatibility complex (MHC
I) for antigen presentation compared to cP-derived peptides.2
Furthermore, the iP-specific proteolytic subunits, in particular
https://doi.org/10.1021/acs.jmedchem.1c00604
© 2021 American Chemical Society
J. Med. Chem. 2021, 64, 10230−10245
10230

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Table 1. Overview of Subunit-Specific In Vitro Biochemical IC50 Values of Previously Published Reference Compounds (β5i =
a
LMP7)
a
b
Assay data represent the mean of a minimum of three determinations including the standard deviation. Single measurement.
cancer models, the therapeutic effects of inhibition of
individual iP proteolytic subunits or multiple subunits remain
less well characterized, with most preclinical reports to date
only describing the activity of dual LMP2/7 inhibitors or
compounds with only partial selectivity against the cP subunit
β5.19−25
turned out to be less selective, inhibiting not only β1i, LMP7,
and β5 but also β2i and β2 proteolytic activity. Each of these
inhibitors contains an isobutyl group, which occupies the S1
pocket adjacent to the catalytically active site as a common
motif.
Although the iP and cP possess overlapping substrate
specificities, it is well established that the iP has a stronger
preference for hydrophobic amino acids like tryptophan
binding to S1 and is more effective in cleaving such
substrates.28 In contrast, the cP displays a greater preference
for cleavage following smaller and polar amino acid residues.
Crystallographic investigations from Huber et al. suggested
that LMP7 has a larger S1 pocket than β5 due to a Met45
conformation that allows access of sterically demanding groups
to the binding pocket.29 We speculated that these differences
could possibly be exploited to generate more selective iP
inhibitors. Furthermore, rationally designed LMP7-specific
substrates, like (Ac-ANW)2R110 or Ac-ANW-AMC30 con-
taining a tryptophan to fill the S1 pocket, underline the ability
of the S1 pocket to accommodate large bicyclic moieties. Very
recently, the reported structures of the cP and chimeric
immuno-constitutive proteasome variants in yeast and the
electron microscopy report of the human iP31 have been
complemented by the crystal structure of an α-aminoboronic
acid derivative in complex with the bona fide human iP,17 in
which a large 2,4-dimethylphenyl moiety fills the S1 pocket.
Cocrystal Structure of Compound 4 with the iP. Since
this information was unknown at the time at which we initiated
our drug discovery program some years ago, our initial efforts
focused on generating insights into the human iP structure. We
were able to generate a cocrystal structure of the human iP
with compound 4 (Table 1), which is a bortezomib derivative
synthesized similar to reported methods.32 Instead of an
isobutyl group, 4 contains a more sterically demanding 3-ethyl-
substituted benzyl moiety (Figure 1). Although 4 displayed
The work published in this paper is based on the hypothesis
that selective LMP7 inhibition could potentially achieve
antitumor activity in B cell-derived malignancies.3,12,13
Although in recent years substantial progress has been made
in the design of selective LMP7 inhibitors, to our knowledge,
no compound has been published to date, which has properties
enabling oral administration, in order to evaluate this
hypothesis. We have generated highly potent LMP7 inhibitors
that demonstrate exquisite selectivity against all other
proteolytic subunits of the iP and cP. Compound 50
(M3258) also demonstrated an attractive overall profile with
regard to physicochemical and drug metabolism and
pharmacokinetic (DMPK) properties and delivered robust in
vivo efficacy and target inhibition in an MM xenograft model.
These data supported the recent initiation of a phase I clinical
trial of M3258 in relapsed/refractory MM patients
(NCT04075721).
RESULTS AND DISCUSSION
■
As a starting point, we evaluated the subunit inhibition profiles
of the clinically approved proteasome inhibitors bortezomib,
ixazomib, and carfilzomib (Table 1). The inhibition of each
proteolytic subunit of iP (β1i, β2i, and LMP7) and cP (β1, β2,
and β5) was determined using fluorescence intensity assays
applying specific substrates that undergo cleavage by the
individual proteasome subunits. According to our data and in
line with previous reports,26,27 bortezomib and ixazomib
displayed comparable subunit specificity with the highest
activity against β1i, LMP7, and β5. In contrast, carfilzomib
10231
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Figure 1. Human iP (S20 core particle) X-ray structure in complex with compound 4.
Figure 2. Binding modes of 4 in β1i, β2i, and LMP7. The orientation of the ethyl-phenyl moiety in β1i is flipped by roughly 180° compared to that
in β2i and LMP7. The hydrogen bonding network of compound 4 with the catalytic subunits is mostly built in an antiparallel β-sheet manner,
interacting with the protein main chain in the substrate binding channel. The pyrazine group of 4 binds into the S3 pockets of the catalytic subunits,
adopting several ring orientations, depending on the subunit type and on the neighboring protein chains.
Scheme 1. Strategy to Increase LMP7 Specificity
different potencies against β1i, β2i, and LMP7, in the crystal
structure, all subunits were found to be occupied, likely as a
result of the high concentration used during cocrystallization
(Figure 2). In each case, the boron atom covalently interacted
with the nucleophilic oxygen lone pair of Thr1 and led to the
formation of a tetrahedrally coordinated boronate adduct as
expected. In the LMP7 subunit, the large 3-ethylphenyl group,
replacing the isobutyl group of bortezomib, nicely filled the S1
pocket.
Compound 4 showed ∼24-fold biochemical selectivity for
LMP7 over β5, contrasting to the ∼2-fold selectivity measured
with bortezomib (Table 1). Compound 4 contains an identical
dipeptide residue, which adopts the known antiparallel β-sheet
conformation.33 The amide bond next to the active site is
essential for activity and interacts with Ser21 and Gly47 in
LMP7 through two strong hydrogen bonds, while the second
amide group forms two hydrogen bonds with the peptidic
backbone of the LMP7 subunit (Ala49 and Ser21). Similar
10232
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
a
Scheme 2. Synthetic Access to α-Amino Boronic Acid Building Blocks
a
Reagents and conditions: (i) B2pin2, Pd(dppf), KOAc, dioxane, 85 °C, 16 h ; (ii) (+)-pinanediole, diethyl ether, room temperature, 16 h; (iii)
Matteson homologation: (1) DCM, BuLi, −100 to −78 °C and (2) ZnCl2, −78 to room temperature; (iv) LiHMDS, −78 °C to room temperature,
18 h; (v) HCl in ether, −10 °C to room temperature, 2 h.
a
Scheme 3. General Synthesis of Pinanediol Boronates; Separation and Deprotection
a
Reagents and conditions: (i) R2COOH, HATU, DMF, DIPEA, room temperature, 2 h; (ii) chiral SFC separation; (iii) isobutylboronic acid, 2 N
HCl, MeOH, pentane.
interactions were identified in β2i (Thr21 replaces Ser21), β1i,
and in the β5 subunit of the cP. Despite the significant changes
in S1 pocket selectivity, the H-bond network formed by the
dipeptide residue of 4 did not result in a pronounced potency
difference between subunits.
We therefore hypothesized that LMP7 selectivity could be
gained by weakening this hydrogen bonding network and
replacing the second amide group (Scheme 1).
In our initial approaches, we replaced R2 with highly
substituted sulfonamides or tertiary amides.34 However, a
balance between LMP7 specificity, good metabolic stability,
and sufficient permeability could not be achieved. In order to
improve the physicochemical profile and lower tPSA, we
moved our focus to R2 groups containing a reduced number of
H donors and acceptors to fill the S2/S3 pockets. Additionally,
we aimed to combine these with large R1 groups to allow for an
optimized fit into the S1 pocket.
Synthesis of LMP7 Inhibitors. Since the commercial
accessibility of protected α-amino-boronates containing
various R-groups is limited, we synthesized these building
blocks (Scheme 2) to systematically explore the SAR of the S1
pocket of the iP. Starting from aliphatic, benzylic, or bicyclic
methylene bromides 5, like 3-(bromomethyl)benzofuran, Pd-
catalyzed borylation applying established protocols35 led to the
corresponding pinacol boronates 6 in good yields. In order to
enable the stereoselective introduction of the chiral α-amino
substituent, the pinacol boronates were converted into
(+)-pinanediol boronates 7 as a chiral directing group using
a transesterification protocol. Matteson homologation36
applying (dichloromethyl)-lithium at low temperatures (−78
to −100 °C) and anhydrous ZnCl2 as a catalyst led to
diastereoselective formation of α-S-chloro boronates 8 usually
with good stereoselectivity based on GC-analysis. Nucleophilic
displacement of the chlorine substituent with N-lithiohex-
amethyldisilazane followed by cleavage of the silyl groups using
HCl or trifluoroacetic acid yielded the desired R-configured
derivatives 10, in most cases as solid salts, which could be
stored at −20 °C without degradation or epimerization.
These α-aminoboronic acid building blocks were then
coupled with acids using standard coupling reagents such as
HATU to obtain the corresponding amides 12 in good yields
(Scheme 3). At this stage, the stereochemical purity could also
be readily determined by chiral supercritical fluid chromatog-
raphy (SFC). In order to further improve the stereochemical
purity of the desired R-diastereomers and to investigate the
activity of the S-diastereomers, a chiral separation of both
diastereomers was performed on a preparative scale for
selected examples. However, the S-configured diastereomer
proved to be significantly less active on LMP7 than the
corresponding R-diastereomer. These observations confirm
those previously reported by Zhu et al.32 In all cases
investigated, the R-configured diastereomers were between
10233
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
a
Table 2. Structures and SAR of Compounds 15−21, Summarizing Substituent Effects on R1 Groups (β5i = LMP7)
a
All assay data represent the geometric mean including standard deviation of a minimum of two independent experiments in duplicate.
10- and 100-fold more active on LMP7 and minor amounts of
S-isomers did not impair the SAR evaluations. Consequently,
mixtures containing low amounts of S-diastereomers could be
readily used in biochemical experiments for determination of
IC50 values. In our hands, the deprotection of the
(+)-pinanediol boronates 13 was most effectively achieved
by a biphasic transesterification method applying isobutylbor-
onic acid37 in order to obtain the free boronic acid derivatives
14 after lyophilization as amorphous powders. Crystalline
trimers comparable to those previously reported with
bortezomib38 could be finally obtained by recrystallization.
SAR of R1 (Tables 2 and 3). We began our SAR
exploration of the S1 pocket using compounds containing a
benzyl group as R2 (see Table 2), in which a bending induced
by the methylene linker allowed the phenyl moiety to interact
with Cys48 in the LMP7 subunit via van der Waals
interactions. Despite not being the best choice in terms of
potency, benzyl resulted in LMP7/β5 selectivity gains superior
to other R2 groups with longer linkers like ethylene or
propylene (data not shown). Direct attachment of phenyl to
the amide resulted in a loss of potency.
The first compound of this series (Table 2, 15) harboring an
isobutyl group as R1 yielded only modest IC50 values of 1000
and 15,000 nM against human LMP7 and β5, respectively,
indicating that the high activity of bortezomib or ixazomib
toward LMP7 is mainly driven by the H-bond interactions of
the dipeptidyl moiety in S2/S3. Changing R1 to benzyl (16)
led to a 6-fold improvement in potency for LMP7, whereas
activity for β5 remained unchanged. Further optimization of
the substitution pattern of the phenyl group revealed that small
nonpolar substituents like methyl or chloro in para and meta
positions further enhanced potency toward LMP7 and resulted
in biochemical IC50 values below 100 nM (17, 19, and 20),
likely as a result of enhanced lipophilic interactions with the S1
pocket of LMP7. In terms of selectivity, the best ratios were
observed for compounds containing either 3,4-dimethyl or 2,4-
dimethyl substitution patterns (ratio: >150 for 18 and 19).
While the larger LMP7 S1 pockets nicely accommodated these
more sterically demanding moieties, they were not compatible
with the more sterically restricted β5 S1 pocket. This resulted
in improved LMP7 selectivity for these examples. Similar
trends were also observed in proteolytic cleavage assays in
A549 cells (cellular β5 assay, Table 1) and human peripheral
blood mononuclear cells (PBMCs; cellular LMP7 assay, Table
1) applying specific luminogenic substrates.30 Although cellular
selectivity ratios were generally smaller compared to those
observed in biochemical assays, significant selectivity was
confirmed for 18 and 19. As mentioned above, inverting the
configuration at the stereocenter reduced both LMP7 activity
and selectivity in biochemical and cellular assays, as
exemplified by compound 19 versus 21.
Docking studies with 4 employing our crystal structure
suggested that bicyclic aromatic structures may overlay well
with the selectivity-maximizing 3,4-dimethylphenyl moiety.
Indeed, replacement of the 3,4-dimethyl phenyl in R1 by 2-
benzofuranyl (22, Table 3) gave comparable results to 19.
Introduction of a regioisomeric 3-benzo-furanyl group further
enhanced LMP7 potency by 10-fold into the single digit nM
IC50 range and simultaneously improved biochemical selectiv-
ity versus β5 to >400-fold (23). Notably, 23 also led to a
robust increase in LMP7 cellular inhibition into the low two-
digit nM range. Encouraged by these results, various other
10234
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
a
Table 3. Structures and SAR of Compounds 22−32 Containing Bicyclic R1 Groups (β5i = LMP7)
a
All assay data represent the geometric mean including standard deviation of a minimum of two independent experiments in duplicates.
bicyclics like 2-naphthalene, 3-naphthalene, 3-benzothiophene,
3-N-methylindole, or 3-N-methylindazole (not reported here)
were investigated but none of these outperformed the 3-
benzofuranyl group in terms of potency and selectivity.
The favorable physicochemical profile (MW, 323 g/mol;
tPSA, 83Å2; log P, 3.5; log D, 1.0; kinetic solubility, >200 μM)
and high metabolic stability in mouse and human liver
microsomes qualified 23 for examination of its in vivo PK
properties. After i.v. application in mice, 23 showed acceptable
clearance (CL), volume of distribution (Vss), and half-life
(t1/2) values of 0.18 L/h/kg, 2.25 L/kg, and 0.37 h,
respectively. The oral bioavailability of 23 was 37% when
applied at 0.5 mg/kg. Although benzofuranes are commonly
considered to be metabolically labile,39 in the case of 23, the
presence of the nearby polar boronic acid likely prevented
oxidative metabolism, which is usually observed for these kinds
of electron-rich heterobicycles.
We next examined whether modifications by small
substituents could further enhance the selectivity of com-
pounds containing 3-benzo-furanyl. While substituents such as
10235
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
chloro were not allowed at the 2-position (exemplified by 24),
F, Cl, Me, or OMe at position 7 (25−28) or 6 and 7 (29) was
tolerated but had no significant positive effect on either activity
or selectivity. In another approach, we investigated the
influence of partial saturation of the oxo heterobicycle and
tested both stereoisomers after chiral separation. The S-
stereoisomer (31) failed to fit into the LMP7 S1 pocket and
was only weakly active. In contrast, R-stereoisomer 30 served
as a potent inhibitor of LMP7 activity and showed increased
selectivity against β5. Replacing 7-chloro with 7-methyl led to
an exquisitely selective compound (32) within this subseries,
as demonstrated in both biochemical and cellular assays. Of
note, this improvement in selectivity was only accompanied by
a relatively modest reduction of cellular LMP7 potency
compared to 23.
Table 4. Structures and SAR of Compounds 23 and 33−41
Containing Heterocyclic and Cyano-Substituted R2 Groups
a
(β5i = LMP7)
SAR of R2 (Tables 4 and 5). Having identified potent and
selective R1 residues, we next shifted our focus toward the
optimization of the R2 group. We maintained the synthetically
tractable 3-benzo-furanyl group as R1 and replaced the phenyl
moiety of R2 with various N-containing mono- or bicyclics
(Tables 4 and 5). These retained attractive biochemical
potency as exemplified by 33−37. However, unsurprisingly,
these modifications reduced biochemical and cellular selectiv-
ity toward LMP7, likely as a result of enhanced interactions by
the additional H-bond acceptors with the backbone of both the
β5 and LMP7 subunits. Intensive efforts to improve LMP7
specificity by introduction of substituents (e.g., halogen,
alkoxy, amides, etc.) at the phenyl ring of R2 had little effect
compared to 23. One of the few exceptions was compound 38
carrying an ortho-nitrile group. Combination of this R2 group
with (3S)-2,3-dihydro-1-benzofuran-3-yl (39) or its 7-chloro
(40) or 7-methyl (41) derivatives as R1 led to derivatives with
impressive biochemical β5 splits (>1000-fold) as well as
improved selectivity (>400-fold) in cells. However, despite
possessing a good overall physicochemical profile (e.g.,
compound 39: MW, 350 g/mol; tPSA, 103 Å2; log P, 1.75;
log D, 1.3; kinetic solubility, >200 μM) and acceptable
metabolic stability in mouse (37 μL/min/mg protein) and
human liver microsomes (<10 μL/min/mg protein), an in vivo
mouse PK study with 39 showed a slightly higher clearance
following i.v. application (1.97 L/h/kg) than for compound 23
and a half-life of 0.71 h, while the volume of distribution (Vss)
was 1.88 L/kg. Bioavailability was found to be 13% when
applied at 0.5 mg/kg.
a
All assay data represent the geometric mean including standard
deviation of a minimum of two independent experiments in
duplicates.
Moving back to the 3-benzo-furanyl group as R1, we decided
to redesign R2 by switching from flat aromatic groups to three-
dimensional saturated moieties containing H-bond acceptors
to potentially enable more specific interactions. Surprisingly,
the introduction of α- or β-ethers already led to slight
improvements in both biochemical LMP7 potency and
selectivity (43 and 44), which contrasted to alkyl moieties
like the acetyl derivative 42. Based on our crystal structure, this
observation can be rationalized by formation of a new H bond
between the backbone NH of Ala49 and the ether oxygen
(Figure 3A).
Similar effects on potency and selectivity were observed
when oxygen was incorporated into the 2- or 3-position of five-
or six-membered rings as exemplified by compounds 45−48,
while the corresponding purely carbocyclic cyclo-pentyl or
hexyl derivatives (data not shown) were only weakly active on
LMP7. Based on the docking poses of 47 and 48, a clear
discrimination of the preferred stereochemistry was not
predictable and indeed both compounds exhibited similar
IC50 values in the biochemical LMP7 assay. Productive
hydrogen bonding requires orientation of the tetrahydrofuran
oxygen toward the peptide bond (formed by Ala49-Cys48).
However, this also required the furan system to adopt an
energetically unfavorable conformation. Freezing of this
bioactive conformation was achieved by incorporation into a
bicyclic system like the 7-oxabicyclo[2.2.1]heptane system,
which is only rarely used in drug design. Indeed, this approach
led to the identification of compounds 49 and 50, which both
exhibit an exo-stereochemistry in the bicyclic system.
Compound 50 combined high biochemical potency and
selectivity, which translated well into cellular assays. In
contrast, the endo isomers 51 and 52, which do not allow
optimal positioning of the ether oxygen, were only moderately
active.
The anticipated H-bond formation with the peptidic
backbone in the S2 pocket could be confirmed by the
generation of an additional cocrystal structure of 50 with the
20S iP, which is shown in Figure 3.40−42 As a result of the
10236
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
studies, which indicated clearance values of 15 and 6.2 μL/
min/106 cells, respectively. The incorporation of the alpha-
carbon as a bridgehead into a bicyclic system likely prevents its
metabolism either due to steric reasons or because the
constrained pyramidal conformation does not allow a sufficient
stabilization of the radical intermediate that is usually formed
during CYP-mediated metabolism.
Table 5. Structures and SAR of Compounds 42−52
Containing Linear, Cyclic, and Bicyclic Ethers as R2 Groups
a
(β5i = LMP7)
In addition to its high biochemical (IC50 = 3.6 nM) and
cellular (IC50 = 3.4 nM) potency against the LMP7 subunit,
compound 50 was metabolically stable and displayed a
favorable physicochemical profile (MW, 329 g/mol; tPSA, 92
Å2; log P, 2.3; log D, 1.1; kinetic solubility, >200 μM) and
acceptable permeability (Papp,AB: 8.2 × 10−6 cm/s, ER: 2.8).
Biochemical assays with each iP and cP proteolytic subunit
confirmed the exquisite LMP7 selectivity of 50, exemplified by
the lack of inhibition of β1, β2, β1i, and β2i up to the highest
tested concentration of 30 μM and the greatly reduced
inhibition of β5 (IC50 = 2500 nM; Table 7). Jump dilution
experiments revealed a prolonged target occupancy in a range
of several hours.41,42 Based on these attractive features, 50 was
chosen for extended in vitro selectivity profiling, including a
selection of proteases (SI, Table S2) and other safety relevant
targets (SI, Table S3). No potent off-target activity in the
nanomolar range was observed, and only a few proteases and
5-HT2A were found to be inhibited at micromolar
concentrations. In parallel, cell viability was assessed in
HepG2 cells and rat and human hepatocytes by ATP depletion
(Cell Titer Glo assay at 48 or 72 h). In contrast to bortezomib
1, compound 50 consistently showed a weak cytotoxicity with
IC50 values of 48 μM in HepG2 cells, 64 μM in rat
hepatocytes, and 91 μM in human primary hepatocytes (SI,
Tables S4−S6). Consequently, the cellular split between
significant target modulation and initial cytotoxicity is around
10,000-fold. The exquisite in vitro profile justified the in-depth
in vivo characterization of compound 50 in PK and
pharmacodynamic (PD) experiments.
In Vivo Profiling of Compound 50. After i.v. application
in mice, 50 revealed overall favorable PK characteristics. Values
for clearance, volume of distribution, and half-life were
determined to be 0.19 L/h/kg (CL), 0.47 L/kg (Vss), and
1.71 h (t1/2), respectively. The oral bioavailability of 50 was
∼35% when applied at 10 mg/kg (vehicle: 0.25% Methocel/
0.25% Tween 20 in PBS). The dose-dependent in vivo
antitumor activity and in vivo inhibition of tumor LMP7 by
compound 50 were first evaluated in mice xenografted with the
human MM cell line U266B1. When applied daily per os at 1
mg/kg, 50 achieved a significant and strong antitumor activity
exemplified by sustained tumor regression (Figure 4A).
A lower daily oral dose of 0.3 mg/kg still resulted in
significant tumor growth inhibition albeit less pronounced,
while 0.1 mg/kg did not significantly affect tumor growth. Each
dosing regimen was well tolerated in mice (Figure 4B).
U266B1 tumor samples were taken from mice 1 and 6 h
following the final application of the vehicle or 50 at 0.3 and
0.1 mg/kg for examination of LMP7 activity as a PD readout
(Figure 4C). Consistent with the antitumor effects described
above, the 0.3 mg/kg dose of 50 led to a significant reduction
of tumor LMP7 activity at 1 and 6 h compared to the control
group, while the 0.1 mg/kg group was without significant effect
at any time point. The regression of U266B1 tumors under
treatment with 50 at 1 mg/kg precluded PD assessments in
this experiment. As such, a separate PK/PD experiment was
performed using a single oral application of 50 at 1 and 0.3
a
All assay data represent the geometric mean including standard
deviation of a minimum of two independent experiments in
duplicates.
bicyclic system, conformational changes are highly constrained
and prevent interaction between the ether oxygen and the
peptidic backbone of the other iP and cP subunits.
Cyclic ethers like THF are known to be prone to CYP-
mediated oxidation in the alpha-position43 and are usually not
regarded as desirable structural motifs in drug design. THF
derivatives have been even used as prodrugs44 due to their
rapid oxidative cleavage. Indeed, our metabolic studies
suggested an elevated oxidative metabolism, which was more
pronounced for 45 and 46 and surprisingly low for 47 and 48.
Notably, the bicyclic ether motif in compounds 49 and 50
was metabolically beneficial despite its slightly higher lip-
ophilicity (Table 6). Improved stability of bicyclic ether amides
toward metabolism has also been reported recently for CXCR7
modulators.45 Additionally, the metabolic stability of com-
pound 50 was confirmed in mouse and human hepatocyte
10237
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Figure 3. Interactions of 44 and 50. (A) Anticipated H-bond formation for 44. H bonds are represented as cylindrical spheres. (B) Crystal
structure of 50 (PDB ID: 7AWE). Compound 50 is shown in the LMP7 subunit. (C) Ligand interactions of 50 generated using the MOE program.
Relevant interactions are indicated by dashed lines. (D) Graphical illustration of the surface of the LMP7 pocket. Green and magenta indicate
lipophilic and polar surfaces, respectively.
Table 6. CLint Data and log P of Compounds 45−50
compound
45
46
47
48
49
50
liver microsome CLint (m/h) [μL/min/mg protein]
hepatocyte CLint (m/h) [μl/min/106 cells]
log P
48/-
90.7/-
2.10
34/-
99.7/-
1.90
16/-
-/-
2.10
19/-
26.7/-
1.85
16/<10
15.9/6.6
2.25
14/<10
∼15/6.2
2.34
a
Table 7. Overview of Subunit Specific In Vitro Biochemical IC50 Values of Compound 50
a
Assay data represent the mean of a minimum of three determinations including the standard deviation if applicable.
mg/kg to U266B1 tumor-bearing mice (Figure 5). In this
experiment, the single treatment of 50 at 1 mg/kg significantly
reduced LMP7 activity compared to the vehicle control group
for up to 14 h, with LMP7 activity returning to baseline by 24
h. Consistent with the PD effects described above for repeated
application of 50 at 0.3 mg/kg, a single treatment at this same
dose suppressed LMP7 activity at 1 and 6 h, yet it was without
effects by 14 h. A higher dose of 1 mg/kg led to a dose-
proportional increase in 50 exposure, which was associated
with more prolonged suppression of tumor LMP7 activity.
10238
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
program, based on rational drug design and structural insights,
culminated in the identification of the highly potent and
exquisitely selective amido boronic acid-based LMP7 inhibitor
50 (M3258).41 The high potency and selectivity of M3258
were achieved via the combination of the amido boronic acid
scaffold with a 3-benzo-furanyl moiety as the R1 group. This
allowed optimal interaction with the large and lipophilic S1
pocket of LMP7 while fitting less efficiently into other iP and
cP proteolytic subunits. Furthermore, the (1S,2R,4R)-7-
oxabicyclo[2.2.1]heptane-2-carboxylic amide R2 group of
M3258 is positioned such that the ether-oxygen forms a
highly specific H-bond interaction with the peptide backbone
of LMP7 but not with other iP or cP subunits.
The attractive overall profile of M3258 enabled its
application in preclinical in vivo models of MM. These studies
provided convincing preclinical evidence that LMP7 inhibition
could be an attractive pharmacological strategy in this disease.
Furthermore, our findings suggest that the spectrum of iP
proteolytic subunit inhibition required to achieve efficacy in
preclinical MM models contrasts starkly to that in models of
autoimmunity and inflammation.16−18 Similar to previous
reports using other selective LMP7 inhibitors,17,46 we have
recently reported that M3258 was not cytotoxic toward diverse
primary human cell types including PBMCs.42 To our
knowledge, the in vivo activity of these other LMP7 inhibitors
in MM models has not yet been reported but would be of high
interest for future studies to further understand the value of
LMP7 inhibition in MM. In conclusion, the findings with
M3258 reported here, and those from nonclinical safety studies
(Sloot et al., manuscript in preparation), supported the recent
initiation of a phase I clinical trial of M3258 in relapsed/
refractory MM patients (ClinicalTrials.gov identifier:
NCT04075721).
EXPERIMENTAL SECTION
■
Chemistry. General Information. All reactions were carried out
under a nitrogen atmosphere or in sealed vials unless noted otherwise.
Commercial reagents and dry solvents were used as purchased
without additional purification. Reactions were magnetically stirred.
Nuclear magnetic resonance (NMR) spectra were recorded on Bruker
1
NMR Spectrometers operating at 400, 500, or 700 MHz for H and
Figure 4. In vivo efficacy study with 50 in the MM model U266B1.
(A) Tumor growth inhibition upon once daily oral treatment with 0.1,
0.3, and 1 mg/kg of 50. (B) Body weight of U266B1 tumor-bearing
mice treated with 50. (C) LMP7 activity was determined in tumor
lysates by employing the LMP7-specific fluorescent (Ac-ANW)2R110
cleavage assay at 1 and 6 h after the last treatment. Inhibition of
LMP7 in treated samples was calculated relative to vehicle-treated
tumors. ***P < 0.001.
101 or 176 Hz for 13C, respectively, and are referenced to tetramethyl
silane as the internal standard. Compounds reported in the
publication have a purity of >95% unless noted otherwise. NMR
data were processed using MestreNova software and recorded as
1
follows: H NMR: chemical shift (δ, ppm), multiplicity (s, singlet; d,
doublet; t, triplet; q, quartet; m, multiplet), coupling constant (Hz),
and integration; 13C NMR: chemical shift (δ, ppm). In the case of
boronic acid samples, which usually contain varying ratios of
monomer/trimer, D2O was added to sharpen the NMR peaks.
High-resolution mass spectra (HRMS) were recorded on a Bruker
Daltonik maXis mass spectrometer. GC−MS spectra of boronic acids
have been acquired after transformation into the corresponding
boronic acid ethylene glycol esters using a Waters GCT Premier (ion
source temperature: 230 °C, EI ionization at 70 eV). Thin-layer
chromatography (TLC) was performed on Merck Silica gel 60 F254
plates and visualized with UV light.
LCMS Analysis. For monitoring of reactions and purity assessment,
the following devices and methods have been used: UPLC-MS
method: Waters Acquity UPLC; column CORTECS C18 (1.6 μm,
50−2.1 mm) flow: 0.9 mL/min, buffer A: H2O + 0.05% HCOOH;
buffer B: MeCN + 0.04% HCOOH; T: 40 °C, 0−1.0 min 2% →
100% B; 1.0−1.3 min 100% B. HPLC method: Elite La Chrom;
column: Waters XBridge C8 (3.5 μm 50 × 4.6 mm); flow: 2 mL/min;
215 nm; buffer A: 0.05% TFA/H2O; buffer B: 0.04% TFA/ACN;
0.0−0.2 min 5% buffer B; 0.2−8.1 min 5% → 100% buffer B; 8.1−
Together, these data suggest that the elevated antitumor
activity observed with the 1 mg/kg dose of 50, compared to
reduced dose levels, is likely explained by a longer duration of
tumor LMP7 inhibition.
CONCLUSIONS
■
The iP subunit LMP7 has been implicated in the pathogenesis
of diverse disease settings including autoimmunity, inflamma-
tion, and selected malignancies. However, up until very
recently, pharmacological assessment of the role of LMP7 in
preclinical models of these diseases has been hampered by the
unavailability of LMP7 inhibitors exhibiting high target
selectivity and DMPK and physicochemical properties
enabling in vivo application. Our LMP7 inhibitor discovery
10239
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
Figure 5. PK/PD correlation after treatment with 0.3 and 1.0 mg/kg qd p.o. 50, respectively. Plasma samples were taken at 1, 6, 14, 24, and 48 h
for PK analysis. LMP7 activity was determined in tumor lysates by employing the LMP7-specific fluorescent (Ac-ANW)2R110 cleavage assay.
Inhibition of LMP7 in treated samples was calculated relative to vehicle-treated tumors. Total exposure of 50 in the plasma of mice was measured
as described in the Experimental Section.
a
Scheme 4. Synthesis of α-Aminoboronic Acid Intermediate 58
a
Reagents and conditions: (i) B2pin2, Pd(PPh3)4, KOAc, dioxane, 85 °C, 16 h ; (ii) (+)-pinanediole, diethyl ether, room temperature, 16 h; (iii)
Matteson homologation: (1) DCM, BuLi, −100 to −78 °C; (2) ZnCl2, −78 °C to room temperature; (iv) LiHMDS, −78 °C to room temperature,
18 h; (v) HCl in ether, −10 °C to room temperature, 2 h.
a
Scheme 5. Synthesis of Intermediate 59 and of 50 (M3258)
a
Reagents and conditions: (i) (R)-1-phenylethanol, DMAP, EDCI, DCM, 0 °C; (ii) chiral separation; (iii) H2, Pd/C, THF; (iv) 59, HATU, DMF,
DIPEA, room temperature, 2 h; (v) isobutylboronic acid, 2 N HCl, MeOH, pentane, room temperature, overnight.
10240
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
10.0 min 100% → 5% buffer B. Chromatography purifications were
performed on a Teledyne Isco Combiflash system utilizing Redisept
columns using a mobile phase composed of either ethyl acetate/
heptane/ or dichloromethane/methanol.
The synthesis of 50 (M3258) is shown in Schemes 4 and 5. The
synthesis of key intermediate 58 is exemplified as a typical procedure
for the preparation of a protected α-amino-boronate building block.
Preparation of 4−52 followed similar methods using commercial
starting materials. The descriptions of general methods, character-
izations, and spectra of all final compounds are available in the
Supporting Information.
(+)-pinanediol ester (7.3 g, 20.3 mmol) in 40 mL of anhydrous THF
was added lithium bis(trimethylsilyl)amide (1 M in THF, 25.5 mL,
25.5 mmol). The mixture was allowed to reach room temperature,
stirred for 18 h, and concentrated to dryness. To the resulting residue
heptane was added, and then the precipitated solid was filtered off.
The filtrate was concentrated to give the crude title compound (6.7 g,
1
68%). H NMR (400 MHz, CDCl3): δ 7.60−7.59 (m, 1H), 7.50−
7.45 (m, 2H), 7.28−7.24 (m, 2H), 4.31 (dd, J = 1.56, 8.70 Hz, 1H),
3.18−3.14 (m, 1H), 2.92−2.90 (m, 1H), 2.75−2.72 (m, 1H), 2.34−
2.30 (m, 1H), 2.15−2.14 (m, 1H), 2.03 (t, J = 5.68 Hz, 1H), 1.88−
1.80 (m, 2H), 1.39 (s, 3H), 1.30 (s, 3H), 1.01 (d, J = 10.88 Hz, 1H),
0.84 (s, 3H), 0.09 (s, 18H).
2-(Benzofuran-3-ylmethyl)-4,4,5,5-tetramethyl-1,3,2-dioxaboro-
lane (54). To a solution of 3-(bromomethyl)benzofuran (7.1 g, 33.8
mmol) in degassed 1,4-dioxane (70 mL) were added bis(pinacolato)-
diboron (10.3 g, 40.5 mmol), potassium carbonate (13.9 g, 101.0
mmol), and tetrakis(triphenylphosphine)palladium(0) (1.9 g, 1.7
mmol), and the mixture was stirred at 85 °C for 16 h. The reaction
mixture was cooled to room temperature and filtered through a celite
bed. The filtrate was concentrated, and the crude was purified by flash
column chromatography on silica gel, eluting with 2−5% of ethyl
acetate in petroleum ether to afford the title compound (6.1 g, 69%)
[(1R)-1-Amino-2-(Benzofuran-3-Ylmethyl)] Boronic Acid (+)-Pi-
nanediol Ester Hydrochloride (58). To a stirred, cooled (−10 °C)
solution of [(1R)-1-[bis(trimethylsilyl)amino]-2-(benzofuran-3-
ylmethyl)]boronic acid (+)-pinanediol ester (6.7 g, 13.9 mmol) in
MTBE (30 mL) under nitrogen a solution of hydrochloride acid in
ethyl acetate (2.50 eq.) was added dropwise. The reaction mixture
was stirred at room temperature for 3 h, resulting in a precipitate. The
reaction mixture was evaporated to dryness, and the obtained solid
was triturated with MTBE and filtered. The filtered solid was washed
with cold MTBE and dried under vacuum to afford the title
1
as yellow oil. H NMR (400 MHz, CDCl3): δ 7.57−7.52 (m, 2H),
1
compound (3.76 g, white solid, 72%). H NMR (400 MHz, DMSO-
7.46−7.44 (m, 1H), 7.30−7.21 (m, 2H), 2.23 (s, 2H), 1.29 (s, 12H).
13C NMR (101 MHz, DMSO): δ 154.4, 141.4, 128.7, 124.0, 122.1,
d6): δ 7.66 (s, 1H), 7.61−7.60 (m, 1H), 7.47−7.45 (m, 1H), 7.29−
7.20 (m, 2H), 4.30−4.28 (m, 1H), 3.27−3.16 (m, 3H), 2.25−2.13
(m, 3H), 1.94 (t, J = 5.56 Hz, 1H), 1.86−1.81 (m, 2H), 1.25 (s, 6H),
1.01 (d, J = 8.00 Hz, 1H), 0.75 (s, 3H). 13C NMR (126 MHz,
DMSO-d6): δ 154.6, 143.5, 127.3, 124.4, 122.5, 119.7, 115.1, 111.3,
86.8, 77.5, 50.5, 38.6, 37.7, 35.6, 34.4, 27.9, 26.7, 25.6, 23.5, 23.1.
HRMS: calcd for C20H24BNO3 [M − 2H]+: 337.1849; found:
337.1849.
119.7, 116.1, 111.1, 83.3, 24.5, 5.6 HRMS: calcd for C15H19BO3 M =
258.1427; found M = 258.1426.
2-(Benzofuran-3-ylmethyl) Boronic Acid (+)-Pinanediol Ester
(55). To a solution of 2-(benzofuran-3-ylmethyl)-4,4,5,5-tetramethyl-
1,3,2-dioxaborolane (6.1 g, 23.6 mmol) in diethyl ether (60 mL) was
added (1S,2S,3R,5S)-(+)-pinanediol (6.0 g, 35.4 mmol), and the clear
solution was stirred at room temperature for 12 h. The reaction
mixture was washed twice with water and brine and dried over
anhydrous sodium sulfate. The solvent was evaporated under vacuum,
and the remaining oil was purified by flash column chromatography
on silica gel, eluting with 5% of ethyl acetate in petroleum ether, to
(1S,2R,4R)-7-Oxa-bicyclo[2.2.1]heptane-2-carboxylic Acid (R)-1-
Phenyl-ethyl Ester (60a). To a solution of rac 7-oxa-bicyclo[2.2.1]-
heptane-2-carboxylic acid (4.68 g; 31.3 mmol, racemic) in dry
dichloromethane (100 mL) under an atmosphere of argon (R)-1-
phenyl-ethanol (4.62 mL; 37.5 mmol), 4-(dimethylamino)pyridine
for synthesis (DMAP) (3.82 g; 31.3 mmol), and (3-dimethylamino-
propyl)-ethyl-carbodiimide hydrochloride (EDCI) (6.73 g; 34.4
mmol) were added under stirring at 0 °C. Subsequently, the clear
reaction solution was stirred overnight at room temperature. After
completion of the ester formation, the reaction was quenched by
adding sat. NH4Cl(aq) solution and the mixture was extracted twice
with CH2Cl2. The organic layer was washed thrice with sat.
NaHCO3(aq) and brine, dried over Na2SO4, filtrated, and evaporated
to dryness. The crude product was purified by flash chromatography
(silica gel; n-heptane/ethyl acetate, 0−30% ethyl acetate) to obtain
7.50 g (30.4 mmol, yield: 97.3%) of a colorless oil (HPLC: 100%
pure, mixture of diastereomers). The mixture of diastereomers was
separated by preparative, chiral HPLC (Chiralcel OD-H; n-heptane/
2-propanol, 95/5; 220 nm) to obtain (1R,2S,4S)-7-oxa-bicyclo[2.2.1]-
heptane-2-carboxylic acid (R)-1-phenyl-ethyl ester (3.22 g, colorless
oil, yield: 41.8%, chiral HPLC 100%) and (1S,2R,4R)-7-oxa-
bicyclo[2.2.1]heptane-2-carboxylic acid (R)-1-phenyl-ethyl ester
(3.14 g, oil, yield: 40.7%, chiral HPLC 100%). HRMS: calcd for
C15H18O3: 246.1256; found: 246.1256.
(1S,2R,4R)-7-Oxabicyclo[2.2.1]heptane-2-carboxylic Acid (59).
To a solution of (1S,2R,4R)-7-oxa-bicyclo[2.2.1]heptane-2-carboxylic
acid (R)-1-phenyl-ethyl ester (46.74 g; 182.75 mmol; 1.00 equiv) in
THF (233.70 mL), palladium on carbon (10% w/w) (1.94 g; 1.83
mmol; 0.01 equiv) was added. The reaction mixture is hydrogenated
under a H2 atmosphere at 50 °C and 5 bar pressure for 16 h. After
completion of the hydrogenation, the reaction mixture was filtered
through celite, and the filtrate was evaporated to dryness and taken up
in pentane. The organic layer was extracted thrice with water.
Subsequently, the water layer was lyophilized to obtain (1S,2R,4R)-7-
oxabicyclo[2.2.1]heptane-2-carboxylic acid (22.62 g; 159.1 mmol;
yield: 87.1%) as a colorless solid. TLC: chloroform/methanol (9.5/
0.5), Rf: 0.5. 1H NMR 400 MHz, DMSO-d6: 12.16 (s, 1H), 4.66 (d, J
= 4.4 Hz, 1H), 4.54 (t, J = 4.4 Hz, 1H), 2.57 (d, J = 35.2 Hz, 1H),
1.91−1.86 (m, 1H), 1.65−1.37 (m, 4H), 1.34−1.33 (m, 1H). HRMS
1
afford the title compound (6.3 g, 82%) as a solid. H NMR (400
MHz, CDCl3): δ 7.58−7.56 (m, 1H), 7.55−7.53 (m, 1H), 7.46−7.44
(m, 1H), 7.28−7.23 (m, 2H), 4.33 (dd, J = 1.88, 8.76 Hz, 1H), 2.34−
2.32 (m, 1H), 2.28 (s, 2H), 2.22−2.21 (m, 1H), 2.08 (t, J = 5.88 Hz,
1H), 1.42 (s, 3H), 1.29 (s, 3H), 1.13 (d, J = 10.92 Hz, 1H), 0.85 (s,
3H). 13C NMR (176 MHz, DMSO): δ 154.4, 141.4, 128.7, 124.0,
122.1, 119.8, 116.2, 111.1, 85.5, 77.0, 50.7, 38.9, 37.7, 35.0, 28.3, 26.8,
25.9, 23.6, 5.1. GCMS: m/z: 310. HRMS: calcd for C19H23BO3: M =
310.1740; found M = 310.1752.
[(1S)-1-Chloro-2-(benzofuran-3-ylmethyl)] Boronic Acid (+)-Pi-
nanediol Ester (56). To a cooled (−95 °C) mixture of dichloro-
methane (6.3 mL, 60.9 mmol) and anhydrous THF (36 mL) was
added n-butyl lithium (1.6 M in hexanes, 14.0 mL, 22.3 mmol) over
20 min. After stirring for 20 min. at −95 °C, a solution of 2-
(benzofuran-3-ylmethyl) boronic acid (+)-pinanediol ester (6.3 g,
20.3 mmol) in anhydrous THF (22 mL) was added over 20 min.
Then, a solution of anhydrous zinc chloride (0.5 M in THF, 36.5 mL,
18.2 mmol) was added at −95 °C for 30 min while maintaining the
inner temperature between −95 and − 100 °C. The mixture was
allowed to reach room temperature and stirred for 18 h, and the
solvent was removed under vacuum. To the resulting oil was added
diethyl ether and saturated ammonium chloride. The organic layer
was dried over anhydrous sodium sulfate and concentrated in vacuo
(residue: 7.3 g, 99%). 1H NMR (400 MHz, DMSO-d6): δ 7.60−7.57
(m, 2H), 7.49−7.47 (m, 1H), 7.31−7.25 (m, 2H), 4.36−4.34 (m,
1H), 3.31−3.29 (m, 1H), 3.24−3.22 (m, 1H), 2.35−2.31 (m, 1H),
2.14−2.12 (m, 1H), 2.06 (t, J = 5.84 Hz, 1H), 1.90−1.86 (m, 2H),
1.42 (s, 3H), 1.04 (d, J = 11.04 Hz, 1H), 0.85 (s, 3H). 13C NMR (176
MHz, DMSO): δ 154.4, 143.2, 127.3, 124.3, 122.4, 119.9, 116.8,
111.1, 86.1, 77.4, 50.6, 42.0, 38.6, 37.7, 34.6, 27.9, 27.8, 26.6, 25.4,
23.4. GCMS: m/z: 358.2. HRMS: calcd for C20H24BClO3: 358.1507;
found: 358.1501.
[(1R)-1-[Bis(trimethylsilyl)amino]-2-(benzofuran-3-ylmethyl)]
Boronic Acid (+)-Pinanediol Ester (57). To a cooled (−78 °C)
solution of [(1S)-1-chloro-2-(benzofuran-3-ylmethyl)]boronic acid
10241
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
calcd for C7H10O3: 142.0630; found: 142.0629. Optical rotation:
[α]2D0 = + 31.9° (ethanol, 20.16 mg/10 mL).
the inhibitory activity of compounds: Ac-nLPnLD-AMC at 50 μmol/
L for β1, (Ac-PAL)2R110 (Biomol) at 80 μmol/L for LMP2, Ac-
RLR-AMC at 20 μmol/L for β2 and MECL-1, and Suc-LLVY-AMC
at either 40 μmol/L for LMP7 or at 50 μmol/L for β5.30 Fluorescence
was measured using an Envision 2104 plate reader (PerkinElmer)
immediately following substrate addition and again after 1 h of
incubation. Excitation and emission settings were used in accordance
with the instructions of the provider of each peptide substrate. The
inhibitory activity of compounds was ascertained by calculating the
difference in fluorescence at each time point. IC50 values for each
compound were calculated by nonlinear regression analysis,
normalized to DMSO controls, using Genedata Screener (Genedata).
Cellular Assays. Like in the biochemical assays, the cellular
potencies of compounds were determined by measuring the
proteolytic activities of proteasome subunits toward specific
substrates. PBMCs (AccuCell Human PBMC) were used to assess
the activity of cellular LMP7 (β5i). A549 cells (ATCC no. CRL-185)
were used to assess cellular β5 activity. A549 cells have previously
been shown to dominantly express the cP.48 Both cell lines were
applied at a final concentration of 3 × 105 cells/ml. The assays were
conducted in tissue culture plates (384 wells, white, PS) from Greiner.
Compound dilution series were prepared in DMSO. The frozen cells
were diluted to a stock concentration of 4.5 × 105 cells/ml in
DMEM/10% FCS. Cell suspension (22.5 μL) was added in each well
of the test plate. After an incubation (2 h, 5% CO2), 50 nL of
compound solution or vehicle (DMSO) was added to each well. After
shaking for 15 s at 900 rpm, the plates were incubated for 2 h at 7 °C
(5% CO2). Then, 12.5 μL of Proteasome-Glo detection solution was
then added (prepared according to the manufacturer’s instructions).
For the cellular LMP7 (β5i) assay, the Proteasome-Glo custom kit
from Promega (Ac-ANW-luciferin substrate) was used. For the β5
assay, the Proteasome-Glo Chymotrypsin-Like Cell-Based Assay kit
from Promega (Suc-LLVY-luciferin substrate) was used. After shaking
for 15 s at 900 rpm and 30 min of incubation at room temperature, a
luminometric readout was performed. These data served as the basis
for the calculation of IC50 values as described above for the
biochemical assays.
(1S,2R,4R)-7-Oxa-bicyclo[2.2.1]heptane-2-carboxylic Acid [(R)-2-
(Benzofuran-3-yl)-1-((1S,2S,6R,8S)-2,9,9-trimethyl-3,5-dioxa-4-
bora-tricyclo[6.1.1.02,6]dec-4-yl)-ethyl]-amide (61). To a solution of
(1S,2R,4R)-7-oxa-bicyclo[2.2.1]heptane-2-carboxylic acid (1.87 g;
13.18 mmol), HATU (4.62 g; 14.37 mmol), and 4-methyl-
morpholine (3.29 mL; 29.94 mmol) in 70 mL of dry DMF was
added under ice cooling and an argon atmosphere (R)-2-(benzofuran-
3-yl)-1-((1S,2S,6R,8S)-2,9,9-trimethyl-3,5-dioxa-4-boratricyclo-
[6.1.1.02,6]dec-4-yl)-ethylamine hydrochloride (4.50 g; 11.98 mmol).
The yellow solution was stirred for 2.5 h at room temperature. The
reaction mixture was poured into 500 mL of ice-cooled, saturated
NaHCO3 solution and stirred for 15 min to give a precipitate, which
was collected by vacuum filtration and washed with water. The
obtained solid was triturated with acetonitrile, diluted with MTB-
ether, and sucked off to yield (1S,2R,4R)-7-oxa-bicyclo[2.2.1]-
heptane-2-carboxylic acid [(R)-2-(benzofuran-3-yl)-1-
((1S,2S,6R,8S)-2,9,9-trimethyl-3,5-dioxa-4-bora-tricyclo[6.1.1.02,6
]
dec-4-yl)-ethyl]-amide (3.26 g, yield: 58.8%) as white solid (purity:
100%). 1H NMR (700 MHz, DMSO-d6): δ 9.23−9.21 (m, 1H),
7.74−7.73 (m, 1H), 7.58−7.56 (m, 1H), 7.54−7.52 (m, 1H), 7.30−
7.27 (m, 1H), 7.26−7.23 (m, 1H), 4.60−4.57 (m, 2H), 4.00 (dd, J =
8.6, 2.3 Hz, 1H), 2.82−2.78 (m, 1H), 2.78−2.75 (m, 1H), 2.71 (dd, J
= 9.1, 5.0 Hz, 1H), 2.66−2.62 (m, 1H), 2.17−2.12 (m, 1H), 1.91−
1.86 (m, 1H), 1.84−1.80 (m, 1H), 1.77 (t, J = 5.6 Hz, 1H), 1.73−
1.69 (m, 2H), 1.61−1.58 (m, 1H), 1.58−1.51 (m, 2H), 1.51−1.47
(m, 1H), 1.45−1.42 (m, 1H), 1.26 (d, J = 9.9 Hz, 1H), 1.20 (s, 3H),
1.19 (s, 3H), 0.79 (s, 3H). 13C NMR (176 MHz, DMSO-d6): δ 177.8,
154.5, 142.4, 128.1, 123.9, 122.2, 119.6, 118.3, 111.1, 81.8, 78.2, 75.3,
75.2, 52.0, 44.3, 41.7, 39.9, 37.5, 36.4, 34.1, 29.3, 29.2, 28.6, 27.1,
25.8, 25.1, 23.9. LCMS method A: (M + H) 464.2; Rt: 2.57 min.
[(1R)-2-(1-Benzofuran-3-yl)-1-{[(1S,2R,4R)-7-oxabicyclo[2.2.1]-
heptan-2-yl]formamido} ethyl]boronic Acid (50). To a two-phase
system of (1S,2R,4R)-7-oxa-bicyclo[2.2.1]heptane-2-carboxylic acid
[(R)-2-benzofuran-3-yl-1-((1S,2S,6R,8S)-2,9,9-trimethyl-3,5-dioxa-4-
bora-tricyclo[6.1.1.02,6]dec-4-yl)-ethyl]-amide (ee = 97%, 3.45 mmol;
1.60 g) in 150 mL of n-pentane and 50 mL methanol were added
isobutylboronic acid (13.81 mmol; 1.41 g) and 1 N hydrochloric acid
(15.54 mmol; 15.54 mL) at 0 °C. The reaction was stirred at room
temperature overnight. The pentane phase was discarded, and the
methanolic phase was washed with pentane (3×, 80 mL). The
methanolic phase was concentrated (bath temp. below 30 °C) in
vacuo, diluted with ice water, and alkalized with 1 N NaOH (pH 11−
12). This basic solution was extracted with DCM (3 × 80 mL). The
aqueous phase was acidified with 1 N HCl (pH 2) and extracted with
DCM (5 × 80 mL) again. The combined organic phase was dried
over Na2SO4, filtrated, and evaporated. The residue was dissolved in
acetonitrile/water and lyophilized to give 0.697 g (yield: 61.3%) of
Efficacy Testing In Vivo (All Animal Experiments Performed in
the Manuscript were Conducted in Compliance with Institutional
Guidelines). The human MM cell line U266B1 was obtained from
ATCC (TIB-196). A suspension (100 μL) of 5 million cells in
phosphate-buffered saline (PBS) mixed with 1:1 Matrigel (Becton
Dickinson) was subcutaneously injected into H2d Rag2 female mice.
Once tumors reached a mean tumor volume of 179 mm3, animals
were assigned to treatment groups (n = 10) and compound 50 was
administered at doses of 0.1, 0.3, and 1 mg/kg daily per os
(formulation: 0.5% Methocel Premium K4M (Colorcon) and 0.25%
Tween 20 in PBS). Mean tumor volume and standard error of the
mean (SEM) were measured. Statistical analyses were performed
using repeated measures analysis of covariance.
1
the title compound as white powder. H NMR (500 MHz, DMSO-
d6/D2O): δ 7.61 (s, 1H), 7.59 (d, J = 7.7 Hz, 1H), 7.48 (d, J = 8.1 Hz,
1H), 7.29−7.25 (m, 1H), 7.24−7.19 (m, 1H), 4.48−4.45 (m, 1H),
4.42−4.40 (m, 1H), 3.12−3.08 (m, 1H), 2.84 (dd, J = 14.9, 5.9 Hz,
1H), 2.73 (dd, J = 14.9, 8.3 Hz, 1H), 2.45 (dd, J = 9.1, 4.9 Hz, 1H),
1.76−1.71 (m, 1H), 1.60 (dd, J = 11.9, 9.1 Hz, 1H), 1.52−1.44 (m,
2H), 1.43−1.34 (m, 2H). 13C NMR (176 MHz, DMSO-d6): δ 174.6,
155.0, 142.6, 128.6, 124.7, 122.9, 120.4, 118.9, 111.7, 79.4, 75.9, 47.7,
40.7, 34.8, 29.8, 29.3, 24.9. GC−MS: calcd as ethylene glycol ester
C19H22BNO5 M = 355.1591; found: 355.1568.
Biological Assays. Biochemical Activity Testing. Human
immunoproteasomes, purified as described previously47 and used at
0.25 nmol/L, or purified human constitutive proteasomes (Boston
Biochem, used at 1.25 nmol/L) were preincubated for 2 h at 25 °C in
384-well plates with compounds or the vehicle dimethyl sulfoxide
(DMSO) in an assay buffer containing 20 mmol/L Tris (pH 7.5),
0.03% sodium dodecyl sulfate (SDS), and 1 mmol/L ethyl-
enediaminetetraacetic acid (EDTA). The following fluorogenic
peptidic substrates (from Bachem Holding, unless stated otherwise),
which undergo preferential processing by specific proteasome
subunits, were added at the indicated final concentrations to assess
In Vivo PD. For PD analyses of xenograft tumors, 50−100 mg of
tissue was lysed, and lysates were incubated with 10 μM LMP7
substrate (Ac-ANW)2R110 (Biomol) for 60 min at 37 °C.
Fluorescence (excitation: 485 nm, emission: 535 nm) was measured,
and percent of LMP7 inhibition was calculated relative to vehicle-
treated controls.
In Vivo PK. For plasma sampling, blood was collected from mice via
an axillary cut and transferred into tubes containing sodium heparin as
the anticoagulant, mixed well, and then centrifuged at 12,000g for 3
min in a refrigerated centrifuge at 4 °C. Plasma was collected and
transferred into Eppendorf tubes. Samples were stabilized by the
addition of 1 μL of 22% formic acid per 10 μL of plasma and briefly
mixed. The samples were stored on ice and protected from light
before freezing at −20 °C until further PK analysis by UPLC-MS/MS.
X-ray Crystallography. A previously described method was used
for the purification of human immunoproteasomes47 using the
following modifications. Ion exchange chromatography on a DAE-
650 M and CHT ceramic hydroxyapatite was done to isolate
immunoproteasomes. Ammonium sulfate was then added to the
pooled fractions to a final concentration of 1.7 mol/L. Chromatog-
10242
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
raphy was then performed on Butyl-Sepharose followed by
purification using a Superdex200 16/60 gel filtration column
equilibrated in 50 mmol/L 4-(2-hydroxyethyl)-1-piperazineethane-
sulfonic acid (HEPES) (pH 7.6), 100 mmol/L NaCl and 1 mmol/L
dithiothreitol (DTT). For crystallization trials, fractions containing
the human immunoproteasomes were concentrated to 8−10 mg/mL
by ultrafiltration and stored at −80 °C. Fractions obtained at each
chromatographic step were tested for proteolytic activity using the
immunoproteasome peptidic substrate cleavage assays described
above, and enzymatically active fractions were pooled. Hanging-
drop vapor diffusion from a 1:0.5 mixture of protein (6 mg/mL
immunoproteasomes in 50 mmol/L HEPES (pH 7.6), 100 mmol/L
NaCl, and 1 mmol/L DTT) and reservoir solution (0.2 mol/L
sodium thiocyanate, 32−41% 2,4-methyl-pentanediol) was used to
grow human immunoproteasome crystals. The quality of crystals was
improved using cyclic temperature gradients between 15 and 18 °C.
The immunoproteasome−cocrystal complex was formed by soaking
immunoproteasome apo crystals in 10 mmol/L compound 4. The
immunoproteasome apo structure was solved by molecular replace-
ment using Phaser (CCP4)49 using the bovine constitutive
proteasome as a starting model.50 The structure of the immunopro-
teasome−cocrystal complex was subsequently solved by molecular
replacement using the structure of apo immunoproteasome. COOT51
was used for model building. REFMAC5 (CCP4) and BUSTER52
were used for refinement.
Author Contributions
The manuscript was written by M.K., O.S., and M.P.S. with
contributions from all authors. All authors approved the
manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge the support and contributions of the
following people: Andreas Becker, Andrée Blaukat, Julie
DeMartino, Hugues Dolgos, Elise Drouin, Samer El Bawab,
Ralf Emmerich, Stefanie Gaus, Claude Gimmi, Samantha
Goodstal, Philipp Haselmayer, Hendrik Hollmann, Frank
Jährling, Dilafruz Juraeva, Mirek Jurzak, Mireille Krier, Daniel
Kuhn, Long Li, Floriane Lignet, Laura Liu, Jianguo Ma, Pia
Sanfelice, Melanie Schwarz, Willem Sloot, Sofia Stinchi, Heinz
Thoma, Klaus Urbahns, Anja Victor, Ping Yu, and Lars zur
Brugge. We also acknowledge the contribution of Proteros
̈
Biostructures GmbH (Germany), Lead Discovery Center
GmbH (Germany), Pharmaron Beijing Co., Ltd. (China),
and Syngene International Ltd. (India). The authors would like
to thank Claudia Mahr, Heike Naumann, Andrea C. Delp, and
Simone Jost for their synthesis support and generation of
analytical data.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.jmedchem.1c00604.
■
sı
ABBREVIATIONS
■
cP,, constitutive proteasome; ER, efflux ratio; iP, immunopro-
teasome; IFNγ, interferon γ; LLOQ, lower limit of
quantification; LMP2, low-molecular-mass polypeptide-2;
MECL-1, multicatalytic endopeptidase complex-like 1;
LMP7, low-molecular-mass polypeptide-7; MTBE, methyl
tert-butyl ether; HATU, 1-[bis(dimethylamino)-methylene]-
1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophos-
phate; PBMC, peripheral blood mononuclear cell; PD,
pharmacodynamic; PK, pharmacokinetic; SAR, structure
activity relationship; THF, tetrahydrofuran; TNFα, tumor
necrosis factor α
General procedures for compound preparation; charac-
terization of all final products by1H NMR,13C NMR,
HRMS, and HPLC; data collection and refinement
statistics for the cocrystal structure of compound 4 in
complex with the human 20S iP; extended biochemical
profiling; and in vitro safety profiling (PDF)
Molecular formula strings (CSV)
AUTHOR INFORMATION
Corresponding Author
Markus Klein − Merck KGaA, Darmstadt 64293, Germany;
orcid.org/0000-0003-3675-2637; Phone: +49 6151
727472; Email: markus.b.klein@merckgroup.com
■
REFERENCES
■
(1) Micale, N.; Scarbaci, K.; Troiano, V.; Ettari, R.; Grasso, S.;
̀
Zappala, M. Peptide-based proteasome inhibitors in anticancer drug
design. Med. Res. Rev. 2014, 34, 1001−1069.
Authors
(2) Groettrup, M.; Kirk, C. J.; Basler, M. Proteasomes in immune
cells: more than peptide producers? Nat. Rev. Immunol. 2010, 10, 73−
78.
Michael Busch − Merck KGaA, Darmstadt 64293, Germany
Manja Friese-Hamim − Merck KGaA, Darmstadt 64293,
Germany
Stefano Crosignani − Merck KGaA, Darmstadt 64293,
Germany
(3) Miller, Z.; Lee, W.; Kim, K. B. The immunoproteasome as a
therapeutic target for hematological malignancies. Curr. Cancer Drug
Targets 2014, 14, 537−548.
(4) Altun, M.; Galardy, P. J.; Shringarpure, R.; Hideshima, T.;
LeBlanc, R.; Anderson, K. C.; Ploegh, H. L.; Kessler, B. M. Effects of
PS-341 on the activity and composition of proteasomes in multiple
myeloma cells. Cancer Res. 2005, 65, 7896−7901.
(5) Eleuteri, A. M.; Angeletti, M.; Lupidi, G.; Tacconi, R.; Bini, L.;
Fioretti, E. Isolation and characterization of bovine thymus multi-
catalytic proteinase complex. Protein Expression Purif. 2000, 18, 160−
168.
(6) Noda, C.; Tanahashi, N.; Shimbara, N.; Hendil, K. B.; Tanaka,
K. Tissue distribution of constitutive proteasomes, immunoprotea-
somes, and PA28 in rats. Biochem. Biophys. Res. Commun. 2000, 277,
348−354.
(7) Singh, A. V.; Bandi, M.; Aujay, M. A.; Kirk, C. J.; Hark, D. E.;
Raje, N.; Chauhan, D.; Anderson, K. C. PR-924, a selective inhibitor
of the immunoproteasome subunit LMP-7, blocks multiple myeloma
cell growth both in vitro and in vivo. Br. J. Haematol. 2011, 152, 155−
163.
Thomas Fuchss − Merck KGaA, Darmstadt 64293, Germany
Djordje Musil − Merck KGaA, Darmstadt 64293, Germany
Felix Rohdich − Merck KGaA, Darmstadt 64293, Germany
Michael P. Sanderson − Merck KGaA, Darmstadt 64293,
Germany
Jeyaprakashnarayanan Seenisamy − Syngene International
Limited, Bangalore 560 099, India
Gina Walter-Bausch − Merck KGaA, Darmstadt 64293,
Germany
Ugo Zanelli − Merck KGaA, Darmstadt 64293, Germany
Philip Hewitt − Merck KGaA, Darmstadt 64293, Germany
Christina Esdar − Merck KGaA, Darmstadt 64293, Germany
Oliver Schadt − Merck KGaA, Darmstadt 64293, Germany
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jmedchem.1c00604
10243
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
(8) Hallermalm, K.; Seki, K.; Wei, C.; Castelli, C.; Rivoltini, L.;
Kiessling, R.; Levitskaya, J. Tumor necrosis factor-alpha induces
coordinated changes in major histocompatibility class I presentation
pathway, resulting in increased stability of class I complexes at the cell
surface. Blood 2001, 98, 1108−1115.
proteasome subunit assembly overcomes bortezomib resistance in
human hematological cell lines. J. Hematol. Oncol. 2014, 7, 7.
(24) Singh, P. K.; Fan, H.; Jiang, X.; Shi, L.; Nathan, C. F.; Lin, G.
Immunoproteasome β5i-selective dipeptidomimetic inhibitors. Chem-
MedChem 2016, 11, 2127−2131.
(25) Yeo, I. J.; Lee, M. J.; Baek, A.; Miller, Z.; Bhattarai, D.; Baek, Y.
M.; Jeong, H. J.; Kim, Y. K.; Kim, D. E.; Hong, J. T.; Kim, K. B. A dual
inhibitor of the proteasome catalytic subunits LMP2 and Y attenuates
disease progression in mouse models of Alzheimer’s disease. Sci. Rep.
2019, 9, 18393.
(26) Demo, S. D.; Kirk, C. J.; Aujay, M. A.; Buchholz, T. J.; Dajee,
M.; Ho, M. N.; Jiang, J.; Laidig, G. J.; Lewis, E. R.; Parlati, F.; Shenk,
K. D.; Smyth, M. S.; Sun, C. M.; Vallone, M. K.; Woo, T. M.;
Molineaux, C. J.; Bennett, M. K. Antitumor activity of PR-171, a novel
irreversible inhibitor of the proteasome. Cancer Res. 2007, 67, 6383−
6391.
(27) Dubiella, C.; Baur, R.; Cui, H.; Huber, E. M.; Groll, M.
Selective inhibition of the immunoproteasome by structure-based
targeting of a non-catalytic cysteine. Angew. Chem., Int. Ed. 2015, 54,
15888−15891.
(28) Kisselev, A. F.; van der Linden, W. A.; Overkleeft, H. S.
Proteasome inhibitors: an expanding army attacking a unique target.
Chem. Biol. 2012, 19, 99−115.
(29) Huber, E. M.; Heinemeyer, W.; de Bruin, G.; Overkleeft, H. S.;
Groll, M. A humanized yeast proteasome identifies unique binding
modes of inhibitors for the immunosubunit β5i. EMBO J 2016, 35,
2602−2613.
(30) Blackburn, C.; Gigstad, K. M.; Hales, P.; Garcia, K.; Jones, M.;
Bruzzese, F. J.; Barrett, C.; Liu, J. X.; Soucy, T. A.; Sappal, D. S.;
Bump, N.; Olhava, E. J.; Fleming, P.; Dick, L. R.; Tsu, C.; Sintchak,
M. D.; Blank, J. L. Characterization of a new series of non-covalent
proteasome inhibitors with exquisite potency and selectivity for the
20S β5-subunit. Biochem. J. 2010, 430, 461−476.
(31) Santos, R. d. L. A.; Bai, L.; Singh, P. K.; Murakami, N.; Fan, H.;
Zhan, W.; Zhu, Y.; Jiang, X.; Zhang, K.; Assker, J. P.; Nathan, C. F.;
Li, H.; Azzi, J.; Lin, G. Structure of human immunoproteasome with a
reversible and noncompetitive inhibitor that selectively inhibits
activated lymphocytes. Nat. Commun. 2017, 8, 1692.
(32) Zhu, Y.; Zhao, X.; Zhu, X.; Wu, G.; Li, Y.; Ma, Y.; Yuan, Y.;
Yang, J.; Hu, Y.; Ai, L.; Gao, Q. Design, synthesis, biological
evaluation, and structure-activity relationship (SAR) discussion of
dipeptidyl boronate proteasome inhibitors, part I: comprehensive
understanding of the SAR of alpha-amino acid boronates. J. Med.
Chem. 2009, 52, 4192−4199.
(33) Groll, M.; Berkers, C. R.; Ploegh, H. L.; Ovaa, H. Crystal
structure of the boronic acid-based proteasome inhibitor bortezomib
in complex with the yeast 20S proteasome. Structure 2006, 14, 451−
456.
(34) Klein, M.; Schadt, O.; Haselmayer, P.; Busch, M. Boronic acid
derivatives. WO2,016,050,356. 2016.
(35) Ishiyama, T.; Murata, M.; Miyaura, N. Palladium(0)-catalyzed
cross-coupling reaction of alkoxydiboron with haloarenes: a direct
procedure for arylboronic esters. J. Org. Chem. 1995, 60, 7508−7510.
(36) Matteson, D. S.; Sadhu, K. M.; Lienhard, G. E. R-1-Acetamido-
2-phenylethaneboronic acid. A specific transition-state analog for
chymotrypsin. J. Am. Chem. Soc. 1981, 103, 5241−5242.
(37) Coutts, S. J.; Adams, J.; Krolikowski, D.; Snow, R. J. Two
efficient methods for the cleavage of pinanediol boronate esters
yielding the free boronic acids. Tetrahedron Lett. 1994, 35, 5109−
5112.
(9) Angeles, A.; Fung, G.; Luo, H. Immune and non-immune
functions of the immunoproteasome. Front. Biosci. 2012, 17, 1904−
1916.
(10) Kubiczkova, L.; Pour, L.; Sedlarikova, L.; Hajek, R.; Sevcikova,
S. Proteasome inhibitors - molecular basis and current perspectives in
multiple myeloma. J. Cell. Mol. Med. 2014, 18, 947−961.
(11) Cengiz Seval, G.; Beksac, M. The safety of bortezomib for the
treatment of multiple myeloma. Expert Opin. Drug Saf. 2018, 17,
953−962.
̀
(12) Ettari, R.; Zappala, M.; Grasso, S.; Musolino, C.; Innao, V.;
Allegra, A. Immunoproteasome-selective and non-selective inhibitors:
a promising approach for the treatment of multiple myeloma.
Pharmacol. Ther. 2018, 182, 176−192.
(13) Kuhn, D. J.; Orlowski, R. Z. The immunoproteasome as a target
in hematologic malignancies. Semin. Hematol. 2012, 49, 258−262.
(14) Schlafer, D.; Shah, K. S.; Panjic, E. H.; Lonial, S. Safety of
proteasome inhibitors for treatment of multiple myeloma. Expert
Opin. Drug Saf. 2017, 16, 167−183.
(15) Basler, M.; Groettrup, M. Recent insights how combined
inhibition of immuno/proteasome subunits enables therapeutic
efficacy. Genes Immun. 2020, 21, 273−287.
(16) Basler, M.; Lindstrom, M. M.; LaStant, J. J.; Bradshaw, J. M.;
Owens, T. D.; Schmidt, C.; Maurits, E.; Tsu, C.; Overkleeft, H. S.;
Kirk, C. J.; Langrish, C. L.; Groettrup, M. Co-inhibition of
immunoproteasome subunits LMP2 and LMP7 is required to block
autoimmunity. EMBO Rep. 2018, 19, No. e46512.
(17) Ladi, E.; Everett, C.; Stivala, C. E.; Daniels, B. E.; Durk, M. R.;
Harris, S. F.; Huestis, M. P.; Purkey, H. E.; Staben, S. T.; Augustin,
M.; Blaesse, M.; Steinbacher, S.; Eidenschenk, C.; Pappu, R.; Siu, M.
Design and evaluation of highly selective human immunoproteasome
inhibitors reveal a compensatory process that preserves immune cell
viability. J. Med. Chem. 2019, 62, 7032−7041.
(18) Johnson, H. W. B.; Lowe, E.; Anderl, J. L.; Fan, A.; Muchamuel,
T.; Bowers, S.; Moebius, D. C.; Kirk, C.; McMinn, D. L. Required
immunoproteasome subunit inhibition profile for anti-inflammatory
efficacy and clinical candidate KZR-616 ((2 S,3 R)- N-(( S)-3-
(Cyclopent-1-en-1-yl)-1-(( R)-2-methyloxiran-2-yl)-1-oxopropan-2-
y l ) - 3 - h y d r o x y - 3 - ( 4 - m e t h o x y p h e n y l ) - 2 - ( ( S ) - 2 - ( 2 -
morpholinoacetamido)propanamido)propenamide). J. Med. Chem.
2018, 61, 11127−11143.
(19) Downey-Kopyscinski, S.; Daily, E. W.; Gautier, M.; Bhatt, A.;
Florea, B. I.; Mitsiades, C. S.; Richardson, P. G.; Driessen, C.;
Overkleeft, H. S.; Kisselev, A. F. An inhibitor of proteasome β2 sites
sensitizes myeloma cells to immunoproteasome inhibitors. Blood Adv.
2018, 2, 2443−2451.
̀
(20) Ettari, R.; Pallio, G.; Pizzino, G.; Irrera, N.; Zappala, M.;
Maiorana, S.; Di Chio, C.; Altavilla, D.; Squadrito, F.; Bitto, A. Non-
covalent immunoproteasome inhibitors induce cell cycle arrest in
multiple myeloma MM.1R cells. J. Enzyme Inhib. Med. Chem. 2019,
34, 1307−1313.
(21) Koerner, J.; Brunner, T.; Groettrup, M. Inhibition and
deficiency of the immunoproteasome subunit LMP7 suppress the
development and progression of colorectal carcinoma in mice.
Oncotarget 2017, 8, 50873−50888.
(22) Niewerth, D.; Franke, N. E.; Jansen, G.; Assaraf, Y. G.; van
Meerloo, J.; Kirk, C. J.; Degenhardt, J.; Anderl, J.; Schimmer, A. D.;
Zweegman, S.; de Haas, V.; Horton, T. M.; Kaspers, G. J. L.; Cloos, J.
Higher ratio immune versus constitutive proteasome level as novel
indicator of sensitivity of pediatric acute leukemia cells to proteasome
inhibitors. Haematologica 2013, 98, 1896−1904.
(23) Niewerth, D.; Kaspers, G. J.; Assaraf, Y. G.; van Meerloo, J.;
Kirk, C. J.; Anderl, J.; Blank, J. L.; van de Ven, P. M.; Zweegman, S.;
Jansen, G.; Cloos, J. Interferon-γ-induced upregulation of immuno-
(38) Bross, P. F.; Kane, R.; Farrell, A. T.; Abraham, S.; Benson, K.;
Brower, M. E.; Bradley, S.; Gobburu, J. V.; Goheer, A.; Lee, S. L.;
Leighton, J.; Liang, C. Y.; Lostritto, R. T.; McGuinn, W. D.; Morse, D.
E.; Rahman, A.; Rosario, L. A.; Verbois, S. L.; Williams, G.; Wang, Y.
C.; Pazdur, R. Approval summary for bortezomib for injection in the
treatment of multiple myeloma. Clin. Cancer Res. 2004, 10, 3954−
3964.
(39) Welter, J.; Kavanagh, P.; Meyer, M. R.; Maurer, H. H.
Benzofuran analogues of amphetamine and methamphetamine:
10244
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245

Journal of Medicinal Chemistry
pubs.acs.org/jmc
Article
studies on the metabolism and toxicological analysis of 5-APB and 5-
MAPB in urine and plasma using GC-MS and LC-(HR)-MSn
techniques. Anal. Bioanal. Chem. 2015, 407, 1371−1388.
(40) Klein, M.; Busch, M.; Esdar, C.; Friese-Hamim, M.; Krier, M.;
Musil, D.; Rohdich, F.; Sanderson, M.; Walter, G.; Schadt, O.; Zanelli,
U.; Ma, J. Abstract LB-054: Discovery and profiling of M3258, a
potent and selective LMP7 inhibitor demonstrating high efficacy in
multiple myeloma models. Cancer Res. 2019, 79, LB-054.
(41) Sanderson, M.; Busch, M.; Esdar, C.; Friese-Hamim, M.; Krier,
M.; Ma, J.; Musil, D.; Rohdich, F.; Sloot, W.; Walter, G.; Zanelli, U.;
Schadt, O.; Klein, M. Abstract DDT02-01: First-time disclosure of
M3258: a selective inhibitor of the immunoproteasome subunit
LMP7 with potential for improved therapeutic utility in multiple
myeloma compared to pan-proteasome inhibitors. Cancer Res. 2019,
79, DDT02−01.
(42) Sanderson, M. P.; Friese-Hamim, M.; Walter-Bausch, G.;
Busch, M.; Gaus, S.; Musil, D.; Rohdich, F.; Zanelli, U.; Downey-
Kopyscinski, S. L.; Mitsiades, C. S.; Schadt, O.; Klein, M.; Esdar, C.
M3258 is a selective inhibitor of the immunoproteasome subunit
LMP7 (β5i) delivering efficacy in multiple myeloma models. Mol.
Cancer Ther. 2021, DOI: 10.1158/1535-7163.MCT-21-0005.
(43) Couper, F. J.; Marinetti, L. J. γ-hydroxybutyrate (GHB)
effects on human performance and behavior. Forensic Sci. Rev. 2002,
14, 101−121.
(44) Engel, D.; Nudelman, A.; Tarasenko, N.; Levovich, I.;
Makarovsky, I.; Sochotnikov, S.; Tarasenko, I.; Rephaeli, A. Novel
prodrugs of tegafur that display improved anticancer activity and
antiangiogenic properties. J. Med. Chem. 2008, 51, 314−323.
(45) Menhaji-Klotz, E.; Hesp, K. D.; Londregan, A. T.; Kalgutkar, A.
S.; Piotrowski, D. W.; Boehm, M.; Song, K.; Ryder, T.; Beaumont, K.;
Jones, R. M.; Atkinson, K.; Brown, J. A.; Litchfield, J.; Xiao, J.;
Canterbury, D. P.; Burford, K.; Thuma, B. A.; Limberakis, C.; Jiao,
W.; Bagley, S. W.; Agarwal, S.; Crowell, D.; Pazdziorko, S.; Ward, J.;
Price, D. A.; Clerin, V. Discovery of a novel small-molecule modulator
of C-X-C chemokine receptor type 7 as a treatment for cardiac
fibrosis. J. Med. Chem. 2018, 61, 3685−3696.
(46) Sula Karreci, E.; Fan, H.; Uehara, M.; Mihali, A. B.; Singh, P.
K.; Kurdi, A. T.; Solhjou, Z.; Riella, L. V.; Ghobrial, I.; Laragione, T.;
Routray, S.; Assaker, J. P.; Wang, R.; Sukenick, G.; Shi, L.; Barrat, F.
J.; Nathan, C. F.; Lin, G.; Azzi, J. Brief treatment with a highly
selective immunoproteasome inhibitor promotes long-term cardiac
allograft acceptance in mice. Proc. Natl. Acad. Sci. U. S. A. 2016, 113,
E8425−E8432.
(47) Dechavanne, V.; Vilbois, F.; Glez, L.; Antonsson, B. Purification
and separation of the 20S immunoproteasome from the constitutive
proteasome and identification of the subunits by LC-MS. Protein
Expression Purif. 2013, 87, 100−110.
(48) Parlati, F.; Lee, S. J.; Aujay, M.; Suzuki, E.; Levitsky, K.; Lorens,
J. B.; Micklem, D. R.; Ruurs, P.; Sylvain, C.; Lu, Y.; Shenk, K. D.;
Bennett, M. K. Carfilzomib can induce tumor cell death through
selective inhibition of the chymotrypsin-like activity of the
proteasome. Blood 2009, 114, 3439−3447.
(49) Collaborative Computational Project, N. The CCP4 suite:
programs for protein crystallography. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 1994, 50, 760−763.
(50) Unno, M.; Mizushima, T.; Morimoto, Y.; Tomisugi, Y.; Tanaka,
K.; Yasuoka, N.; Tsukihara, T. The structure of the mammalian 20S
proteasome at 2.75 Å resolution. Structure 2002, 10, 609−618.
(51) Emsley, P.; Cowtan, K. Coot: model-building tools for
molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004,
60, 2126−2132.
(52) Bricogne, G.; Blanc, E.; Brandl, M.; Flensburg, C.; Keller, P.;
Paciorek, W.; Roversi, P.; Sharff, A.; Smart, O. S.; Vonrhein, C.;
Womack, T. O. BUSTER version 2.11.7 Global Phasing Ltd:
Cambridge, United Kingdom.
10245
https://doi.org/10.1021/acs.jmedchem.1c00604
J. Med. Chem. 2021, 64, 10230−10245