Synthesis and In Vitro Evaluation of a Set of 6-Deoxy-6-thio-carboranyl d -Glucoconjugates Shed Light on the Substrate Specificity of the GLUT1 Transporter

Abstract

Glucose- and sodium-dependent glucose transporters (GLUTs and SGLTs) play vital roles in human biology. Of the 14 GLUTs and 12 SGLTs, the GLUT1 transporter has gained the most widespread recognition because GLUT1 is overexpressed in several cancers and is a clinically valid therapeutic target. We have been pursuing a GLUT1-targeting approach in boron neutron capture therapy (BNCT). Here, we report on surprising findings encountered with a set of 6-deoxy-6-thio-carboranyl d-glucoconjugates. In more detail, we show that even subtle structural changes in the carborane cluster, and the linker, may significantly reduce the delivery capacity of GLUT1-based boron carriers. In addition to providing new insights on the substrate specificity of this important transporter, we reach a fresh perspective on the boundaries within which a GLUT1-targeting approach in BNCT can be further refined.

Full text

Synthesis and In Vitro Evaluation of a Set of 6‑Deoxy-6-thio-
carboranyl D‑Glucoconjugates Shed Light on the Substrate
Specificity of the GLUT1 Transporter
Jelena Matovic,
#
Juulia Järvinen,
#
Iris K. Sokka,
#
Philipp Stockmann, Martin Kellert, Surachet Imlimthan,
Mirkka Sarparanta, Mikael P. Johansson, Evamarie Hey-Hawkins, Jarkko Rautio, and Filip S. Ekholm*
Cite This: ACS Omega 2022, 7, 30376−30388
Read Online
ACCESS Metrics & More Article Recommendations *
sı Supporting Information
ABSTRACT: Glucose- and sodium-dependent glucose transporters (GLUTs and SGLTs) play vital
roles in human biology. Of the 14 GLUTs and 12 SGLTs, the GLUT1 transporter has gained the
most widespread recognition because GLUT1 is overexpressed in several cancers and is a clinically
valid therapeutic target. We have been pursuing a GLUT1-targeting approach in boron neutron
capture therapy (BNCT). Here, we report on surprising findings encountered with a set of 6-deoxy-6-
thio-carboranyl D-glucoconjugates. In more detail, we show that even subtle structural changes in the
carborane cluster, and the linker, may significantly reduce the delivery capacity of GLUT1-based
boron carriers. In addition to providing new insights on the substrate specificity of this important transporter, we reach a fresh
perspective on the boundaries within which a GLUT1-targeting approach in BNCT can be further refined.
1. INTRODUCTION
Glucose- and sodium-dependent glucose transporters (GLUTs
and SGLTs) play a central role in human biology.
1,2
They are
responsible for transporting the vital energy source D-glucose
across the plasma membrane, which is essential for the
sustenance of living cells. Therefore, glucose transporters are
expressed on all cells. This makes them appealing targets from
a molecular biology perspective. These transporters can be
harnessed for multiple purposes and have been considered as
promising therapeutic targets for prodrug strategies for a
considerable timespan.
3−5
While this is true for most
transporters belonging to these families, the GLUT1 trans-
porter has become the central target in the cancer research
field, as it is overexpressed on a variety of cancers.
6
In addition,
the impaired D-glucose metabolism and increased D-glucose
demand observed in cancer cells further increase the appeal of
this approach.
7,8
Indeed, the potential embedded in a GLUT1-
targeting approach within the cancer research field is already
clear as this targeting strategy is in widespread clinical use as a
tumor imaging technique (2-deoxy-2-fluoro-D-glucose (FDG)
in combination with positron emission tomography imaging).
9
Due to the existing sound foundations, pursuing a GLUT1-
targeting strategy within a boron neutron capture therapy
(BNCT) frame represents an interesting possibility.
In a BNCT context, the transportation speed and
capabilities of GLUT1 can be harnessed to deliver boron-10
atoms across the plasma membrane into the cancer cells. Once
the intracellular boron concentration reaches 20−35 μg/g of
tumor and a high tumor-to-blood/tumor-to-normal tissue ratio
is observed (above 3:1 but ideally more than 10:1), the cancer
cells can in theory be selectively eradicated in a separate step
by irradiation with an external thermal neutron beam at the
tumor site (see Figure 1).
10−12
Recently, we have been
exploring the premises of a GLUT1-targeting strategy for
BNCT in detail, thereby complementing the earlier work
conducted in the field.
13,14
Through our medicinal chemistry
approach, we have been able to identify the most promising
boron cluster attachment site in D-glucose and our
comprehensive in vitro assessment proved that on the
molecular biology level, the GLUT1-targeting strategy is able
to outperform the LAT1-targeting and passive transport
strategies in current clinical use (boronophenylalanine
(BPA)
15
and sodium borocaptate (BSH)
16
). Therefore, the
continued refinement of the GLUT1-targeting strategy is
warranted. Herein, we set out to shorten the synthetic routes
to these types of boron delivery agents by switching the boron
cluster conjugation strategy and attachment points while
continuing to build a deeper overall understanding of the
biochemical foundations of this approach. A representative
library of 6-deoxy-6-thio-carboranyl D-gluconjugates were
synthesized and characterized in detail by a wide palette of
NMR spectroscopic techniques and HRMS before being
subjected to a preliminary in vitro assessment featuring
cytotoxicity, experimental and computational GLUT1 anity,
and cellular uptake studies. Notable dierences to our earlier
work were observed.
13,14
In more detail, subtle structural
Received: June 11, 2022
Accepted: August 4, 2022
Published: August 17, 2022
Article
http://pubs.acs.org/journal/acsodf
© 2022 The Authors. Published by
American Chemical Society 30376
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388

modifications in the carborane and linker structures were
found to significantly diminish the delivery capacity of these
boron carriers. These results help identify the boundaries
within which the GLUT1-targeting approach to BNCT can be
further refined and aid in the interpretation of earlier
observations noted in the field.
17−20
2. RESULTS AND DISCUSSION
2.1. Synthesis and Characterization. The first report on
the synthesis of carbohydrate-based delivery agents for BNCT
was reported by Hawthorne et al. in 1988.
21
Since then, a vast
number of glycoconjugates has been synthesized for this
purpose.
19
Not only have dierent targeting strategies been
pursued,
14,22,23
the carbohydrates have at times likewise been
employed in order to modify the physicochemical properties of
other delivery agents.
23,24
One of the pursued targeting
strategies, which holds tremendous potential, is the GLUT1-
targeting approach. Unfortunately, insights on the biochemical
foundations and boundaries of this approach are limited, as
detailed studies focusing on these central aspects have only
recently begun to emerge.
13,14,25
In our recent work, we
synthesized and studied the toxicity, GLUT1-anity, and
cellular uptake profiles of the complete positional isomer
library of ortho-carboranylmethyl-bearing D-glucoconju-
gates.
13,14
We revealed that the optimal attachment site for a
boron cluster in the carbohydrate core is position 6. In the
current study, we set out to shorten the synthetic routes to
GLUT1-based boron carriers by targeting D-glucoconjugates
bearing an S-linked carboranyl substituent at position 6. We
envisioned that the nonionic nature of the boron cluster
selected, in combination with the elongated bond lengths to
sulfur, would allow us to preserve the functional basis of these
delivery agents. In addition, we decided to map the eects of
the attachment site and configuration of the boron cluster by
preparing a representative set containing both ortho- and meta-
carboranes (1,2-dicarba-closo-dodecaborane (-oCb) and 1,7-
dicarba-closo-dodecaborane (-mCb), respectively), connected
to sulfur through either a carbon or a boron atom (1-Cb and 9-
Cb, respectively). From a synthesis perspective, the S-linked
carboranyl cluster is commonly installed through an SN2-
displacement reaction regardless of the attachment site or
structure of the carborane.
26−28
Therefore, we envisioned that
a protected 6-deoxy-6-iodo glucopyranoside building block
could serve as a suitable carbohydrate-based electrophile. The
synthetic routes to the six end products 1−4,10, and 11 are
summarized in Scheme 1 and will be discussed in more detail
next.
While the carbohydrate starting material 5is commercially
available, it can likewise be prepared from methyl α-D-
glucopyranoside in two synthetic steps.
29,30
We opted to
Figure 1. Top: Chemical structures of the four 6-deoxy-6-thio-carboranyl D-glucoconjugates synthesized and assessed and indication of symbols
used in the figure below (boron atoms in pink and carbon atoms in gray in the boron cluster). Bottom: Principles of BNCT. The 10B delivery
agents enter a cancer cell through GLUT1 (Step 1), and the cell is then irradiated with a precise thermal neutron beam (Step 2). 10B captures the
thermal neutrons briefly forming 11B*. Excited 11B*quickly undergoes a fission reaction producing 4He and 7Li. 4He nuclei have a destructive eect
on the cell (Step 3).
Scheme 1. Synthetic Routes to 1−4, 10, and 11
a
a
Reaction conditions: (i) (1) PPh3, imidazole, I2, toluene, 80 °C, 1 h; (2) Ac2O:pyridine 1:1, rt, 17 h, 63%; (ii) corresponding mercapto-carborane,
N,N-diisopropylethylamine (DIPEA) or K2CO3, acetone, 60 °C, 16−24 h, 55% (6, R = 1-oCb), 98% (7, R = 1-mCb), 80% (8, R = 9-oCb), 90% (9,
R = 9-mCb); (iii) 1−4 N HCl, 110−120 °C, 4−7 h, 35% (1, R = 1-oCb), 43% (2, R = 1-mCb), 75% (3, R = 9-oCb), 74% (4, R = 9-mCb); (iv) 1 N
HCl, 85−105 °C, 2−4 h, 84% (10, R = 1-mCb), 74% (11, R = 9-mCb).
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30377

utilize this methyl glucoside as a starting material since this
would allow both the synthesis of a pure anomer as well as the
more important hemiacetal through the same synthetic
sequence. The SN2-displacement reaction was performed
according to literature protocols with four dierent carboranyl
thiols (1-oCb, 1-mCb, 9-oCb, and 9-mCb) under slightly basic
conditions in acetone.
26,27
The isolated yields of 1−4were in
the 55−98% range. While the reaction conditions were not
optimized, the yields are nevertheless competitive as, e.g.,
installing a carboranyl species through a coupling reaction
between decaborane and an alkyne usually results in maximum
yields of 65%.
31
With the protected glucoconjugates at hand, we decided to
evaluate whether the deprotection conditions could be tailored
to access either the fully deprotected hemiacetal or its methyl
glucoside. The most commonly employed deacetylation
conditions, i.e., Zemplen conditions,
32
would not be suitable
for this task as certain carboranyl species are labile to basic
conditions.
33
Therefore, we decided to explore the possibility
of using acidic conditions instead. Aqueous hydrochloric acid
solutions at elevated temperatures have been reported to work
well for this purpose.
34
In our early trials, the solubility of the
protected glucoconjugates was low under the employed
conditions and therefore elevated temperatures were a
necessity.
Simultaneously, it should be mentioned that we have
observed a dierence in the aqueous solubility between the
carbon- (1−2) and boron-linked carboranes (3−4), with the
boron-linked species displaying an enhanced solubility
compared to their carbon-linked counterparts. Regardless of
these factors, we found that the deprotection outcome could
be successfully altered by employing a 1−4 M solution of HCl
while varying the reaction temperature and time. In more
detail, employing temperatures in the 85−105 °C range and
reaction times up to 4 h resulted in the formation of the
deacetylated methyl glucopyranosides. However, increasing the
temperature to the 115−120 °C range and extending the
reaction time by a few hours resulted in the hydrolysis of the
methyl glycoside. While all four fully deprotected hemiacetals
could be obtained through this protocol, notable dierences in
the yields were observed. Substantially higher yields were
obtained with clusters linked through a boron atom (73−75%)
than for their carbon-linked counterparts (35−43%). We
evaluated a number of dierent acids and conditions
(trifluoroacetic acid (TFA), HCl, dierent co-solvents, etc.);
however, only minor dierences were noted in the isolated
yields. Thus, no further emphasis was put on exploring the
underlying reasons for these observations as short and
accessible synthetic routes to the new types of GLUT1-based
delivery agents had been developed.
During the synthetic work, NMR spectroscopic (nuclear
magnetic resonance, variety of 1D/2D-techniques) and mass
spectrometric techniques (high-resolution mass spectrometry,
HRMS) were utilized to ascertain the molecular structures of
the compounds and quantitative NMR (qNMR) was
employed in determining the purity of the final products 1−
4prior to their in vitro assessment. The NMR spectra were
fully assigned using 1D-TOCSY (total correlation spectrosco-
py), 1H, 13C{1H}, 11B{1H}, COSY (correlation spectroscopy),
ed-HSQC (edited heternonuclear single quantum coherence),
and HMBC (heteronuclear multiple bond correlation). The
1H NMR spectra were further simulated with the ChemAdder
software.
35,36
We use 1as an example to shortly convey the
process used. Highlights from the NMR spectroscopic
characterization are provided in Figure 2. Due to mutarotation,
1exists as a mixture of anomers. Using ed-HSQC, the C-1
atoms, at 98.3 and 94.0 ppm, and the corresponding H-1
atoms, at 5.04 and 4.45 ppm, could easily be identified. The β
anomer, at 4.45 ppm, has a distinctly higher coupling constant
(J1,2 = 7.8 Hz) due to the larger dihedral angle between the
axial H-1 and H-2 protons. The αanomer, at 5.04 ppm, has a
smaller coupling constant of (J1,2 = 3.7 Hz) corresponding to
Figure 2. Highlights from the NMR spectroscopic characterization of compound 1. Top: qNMR spectrum of 1(R = 1-oCb) using maleic acid as
the internal standard. Bottom: 5.0−3.0 ppm region of the 1H NMR spectrum of compound 1showcasing the accuracy of the spectral simulation
(simulated spectrum at the top, measured spectrum at the bottom).
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30378

an axial-equatorial relationship between H-1 and H-2. Through
the conventional use of COSY, ed-HSQC, and HMBC, most
of the signals can be identified; however, in severely
overlapping regions of the 1H NMR spectrum, this becomes
increasingly challenging. To overcome these issues, we did
spectral simulation work with the ChemAdder software and
were able to obtain accurate coupling constants and chemical
shifts for all signals, including those in crowded areas of the
spectrum (see Figure 2). In addition, we used the ChemAdder
software to determine the anomeric ratios, which were found
to be similar for compounds 1−4(α:β56:44 for 1, 49:51 for 2,
48:52 for 3, and 55:45 for 4). Finally, qNMR was utilized to
ascertain that the purity of the compounds (1−4) submitted
for the in vitro assessment studies exceeded 95%. The protocol
described by Pauli et al.
37
was employed, and maleic acid was
used as an internal standard.
2.2. Molecular Recognition Studies. From the drug
development perspective, it is central to understand the
molecular interactions between the ligand and the transporter.
For example, in a BNCT context, the selectivity and success of
the treatments are directly related to the preferential uptake of
boron-10 enriched delivery agents by cancer cells. In our case,
the GLUT1 transporter, which is overexpressed on various
cancers, is responsible for the uptake process. The clinical
validity and therapeutic utility of the approach stem from the
“Warburg eect”.
7,8
Cancer cells have an impaired D-glucose
metabolism while simultaneously displaying a high prolifer-
ation rate and energy demand. Together, these factors
contribute to an increased D-glucose uptake in cancer cells
compared to healthy cells and oer a basis for pursuing a
GLUT1-targeting strategy in BNCT. Based on the available
tumor imaging data obtained through the use of FDG,
38
head
and neck cancers represent a promising target for our
approach. Therefore, we employed the human squamous
carcinoma cell line CAL 27 as a suitable model for the in vitro
assessment of GLUT1-targeting delivery agents. In addition to
being a relevant head and neck cancer cell line of human
origin, the GLUT1-transporter has been indicated to play a key
role in the aberrant growth of CAL 27 cells.
39,40
In our
previous work, we validated the GLUT1-function, quantified
the GLUT1-expression, and developed a cis-inhibition assay for
determining the relative anities of GLUT1-targeting delivery
agents to GLUT1 in the CAL 27 cell line.
13,14
To allow
comparison of the results of the present study to our earlier
ones, we used the same protocols herein. The fully deprotected
glucoconjugates 1−4were subjected to the cis-inhibition assay,
i.e., they were forced to compete for the transporter against
radiolabeled [14C]-D-glucose in order to assess their targeting
capabilities from a functional point of view. A high GLUT1
anity, i.e., a low IC50 value, is characteristic of delivery agents
that are able to target the transporter in the presence of the
natural substrate D-glucose in a biological milieu. The
experimentally determined IC50 values were 126.9 μM for 1;
120.6 μM for 2; 982.3 μM for 3; 2128 μM for 4; and >1000
μM for the control D-glucose (Figure 3). Large dierences in
the IC50 values were thus noted. While the evaluated library is
limited, a clear trend pointing toward the importance of the
cluster conjugation site was revealed. The carbon-linked
species 1−2displayed GLUT1 anities that were only slightly
lower than those previously reported for the O-carboranyl-
methyl-bearing positional isomer library.
13,14
These two
delivery agents display acceptable GLUT1-targeting capabil-
ities. The boron-linked species 3−4on the other hand
displayed significantly lower anities to the GLUT1-trans-
Figure 3. The anity curves and the calculated IC50 values obtained through the cis-inhibition assays are displayed for each of the four
glucoconjugates 1,2,3, and 4.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30379

porter, indicating that the targeting capabilities are not
sucient for use as boron delivery agents in BNCT. While
the trend is clear, the results are surprising. The question that
arises is as follows: which underlying factors contribute to
these notable deviations uncovered through the experimental
cis-inhibition assay?
In an attempt to provide an answer to this intriguing
question, we set out to inspect the molecular interactions
between GLUT1 and the delivery agents through a docking
study. We used our previously benchmarked computational
model.
13,14
Our model is derived from XylE, a D-xylose-proton
symporter, for which both the inside open (PDB ID 4QIQ)
41
and outside open (PDB ID 6N3I)
42
crystal structures are
available. This transporter shares a structural similarity with the
GLUT1−4 proteins (29% sequence identity and 49%
similarity),
43
the carbohydrate-binding domain is well
preserved, and by virtual mutation of a single amino acid
residue, Gln-415 to Asn-415, a structurally similar binding
pocket to that in GLUT1 can be constructed.
44,45
The
carbohydrate delivery agents exist as a mixture of anomers;
however, the individual anomers were initially modeled
separately in the docking assay and then the overall mean
binding energies were calculated to fit with the experimentally
determined anomeric ratios. While the mean binding energies
were calculated for both the outside and inside open
conformations, the outside open conformation is more
important when forming a tie to the experimental anity
data. Based on the docking study, the species containing
carbon-linked clusters interact more strongly with GLUT1
than the species that contain boron-linked clusters. The
computational results support the experimental ones, although
the dierences in mean binding energies are not as striking as
the experimentally determined values would suggest. They do
fall within the experimental margin of error, i.e., 1 kcal/mol. In
our earlier work,
13,14
the computational model has been found
to be a reliable indicator of observed GLUT1 anity and thus
great care needs to be practiced when interpreting the overall
results obtained. Two possibilities for the slight discrepancy
come to mind: (1) the limitations of the current docking
model, which does not take molecular dynamic factors into
account on a sucient level, and/or, (2) the boron-linked
clusters display deviating overall properties in a biological
milieu compared to the carbon-linked cluster and these
hinder/limit their ability to compete for the transporter.
Based on the results generated, it is too early to draw definite
conclusions and more work with a larger substrate library will
be required to map the molecular level basis of the
observations noted. In order to gather more pieces of the
puzzle, we continued our in vitro assessment by evaluating the
cytotoxicity and uptake profiles of compounds 1−4.
2.3. Cytotoxicity and Boron Delivery Capacity. With
the observations noted in our molecular recognition studies,
we proceeded with the in vitro assessment. We chose to focus
on two topics that are of interest from a potential translational
medicine perspective: cytotoxicity and boron delivery capacity.
These studies were performed in the CAL 27 cell line in order
for the results to be comparable to our earlier work.
13,14
The
importance of being able to compare results across series of
molecules cannot be suciently stressed when aiming to
understand the biochemical premises of the GLUT1-approach.
On a general level, the delivery agents need to display low
cytotoxicity and a high boron delivery capacity; otherwise, they
will not be suitable for clinical use.
In the cytotoxicity assays, a series containing the
glucoconjugates 1−4, BSH (as a representative of a clinically
approved boron delivery agent), non-ionic surfactant triton X
(as a positive control for cytotoxicity), and the cell culture
media (as a negative control) were screened. The concen-
tration range (5−250 μM) and time points (6 and 24 h) were
selected based on our previous experiences from working with
the GLUT1-targeting approach.
13,14
A commercial CellTiter-
Glo assay was used to quantify cell viability based on the ATP
production of the live cells after treatments. The results are
summarized in Figure 4. It is worth mentioning that the
substantial dierences in the experimental GLUT1 anities
noted do not influence the cytotoxicity profiles of these
delivery agents. Regardless of time points and concentrations
employed, the investigated glucoconjugates appear to be less
toxic than the clinically employed boron carrier BSH. This
entails that the toxicity of these agents would not hamper their
use in BNCT, although it should be mentioned that a complete
view of the toxicity profiles is not possible to obtain through in
vitro studies alone. A notable dierence in the cytotoxicity
between the 6 h and 24 h time points was observed. At the 24
h time point, the glucoconjugates 1−4were considerably more
toxic than the corresponding glucoconjugates bearing an O-
linked ortho-carboranylmethyl substituent at position 6.
14
Therefore, the O-linked ortho-carboranylmethyl substituent
seems to be better tolerated than the S-linked carboranyl
substituent in the GLUT1-targeting delivery agents. On a
general level, the toxicity profiles are likely to be connected to a
certain extent to the uptake profiles. Therefore, as a last
measure, we sought to assess the boron delivery capacity of the
synthesized glucoconjugates 1−4. In these studies, the
GLUT1-targeting strategy was further compared to the
clinically employed delivery agents BPA and BSH, which
Figure 4. Results from the cytotoxicity studies of 1−4and BSH across the 5−250 μM concentration range in the CAL 27 cell line. The incubation
times were 6 h (A) and 24 h (B). The cell culture medium was used as the negative control and 1% Triton X-100 as the positive control. The
statistical significance was determined through an unpaired Student’s t-test where the significance was set at *p< 0.05, **p< 0.01, and ***p<
0.001.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30380

enter cancer cells either through a LAT1-mediated uptake
(BPA) or a passive diusion through the cell membrane
(BSH).
The uptake studies were performed according to our
recently reported protocol.
13,14
It should be noted that the
protocol employed does not dierentiate the intracellular
boron concentration from that arising from delivery agents
trapped on the cell membrane. This, however, is not a concern,
as the cell-killing eect caused by the thermal neutron beam in
actual BNCT treatments has a destructive radius reminiscent
to the diameter of a single cell (5−9μm in biological tissue).
46
In the uptake studies, the concentration range 10−400 μM was
screened. The incubation times 5, 30, and 120 min were
chosen due the optimal performance of [14C]-D-glucose under
these conditions. Since inductively-coupled plasma mass
spectrometry (ICP-MS) is somewhat insensitive when it
comes to quantification of the boron content, thoroughly
washed cells from four wells were combined and digested
before being subjected to the analysis. The boron quantifica-
tion results are presented in Figure 5. The new set of
glucoconjugates was found to display comparable boron
delivery capacity to the delivery agents in current clinical use
(BPA and BSH) with the exception of glucoconjugate 2, which
has a fivefold higher delivery capacity compared to the rest of
the delivery agents screened. However, when these results are
compared to the ones obtained with our previously reported
hit molecule 6-O-carboranylmethyl D-glucose,
14
the S-linked
carboranyl species seem to be at a significant disadvantage as
their boron delivery capacity is roughly 5−100 times lower.
However, we noticed that the Kmvalues were consistently
higher with the S-linked carboranyl species compared to the O-
linked carboranylmethyl species.
14
The higher Kmvalue may
imply that there are additional transport mechanisms involved
in the uptake of the S-series delivery agents. More work is
needed to ascertain these factors. Regardless, the dierent Km
values may likewise indicate that a greater concentration,
which may no longer be clinically relevant, would be required
for the S-series species to reach Vmax.
When trying to form a connection between the GLUT1-
anity and the cellular uptake profiles, it is important to note
that the anity results convey the competitive edge of the
glucoconjugates over the natural substrate in terms of targeting
GLUT1, whereas the uptake profiles showcase the ability of
the glucoconjugates to remain attached to or within the cells.
Therefore, we did not expect the anity results to correlate
directly with the uptake profiles. The large deviations in the
uptake profiles, especially in light of our previous results, are
nevertheless very interesting. Based on the results, glucoconju-
gate 1seems to act as a GLUT1-antagonist while
glucoconjugate 2still displays acceptable properties from the
boron delivery perspective. In more detail, both agents display
high binding anity to GLUT1; however, only glucoconjugate
2enters the cells through the transporter at an acceptable rate.
These findings highlight that great diligence should be
practiced when switching atoms in the carbohydrate core
and may explain, for example, why the previously prepared
BSH-glucose conjugates have been discarded at an early
development stage.
20
Altogether, our findings provide new
insights on the substrate specificity of the important GLUT1-
transporter and will aid in the rational design of therapeutic
approaches focusing on this target.
3. CONCLUSIONS
We are currently developing a GLUT1-targeting strategy for
BNCT of head and neck cancers. The first phase of our
medicinal chemistry approach has been devoted to exploring
the biochemical foundations in detail. Recently, we reported
on new types of GLUT1-based delivery agents that were able
to outshine the clinically employed boron carriers in a detailed
in vitro assessment.
13,14
Herein, we attempted to shorten the
synthetic routes to such delivery agents while simultaneously
preserving their functional basis. We developed short and
accessible synthetic methods for the construction of a set of 6-
deoxy-6-thio-carboranyl D-glucoconjugates and characterized
the products and all intermediates by a wide palette of NMR
spectroscopic techniques and mass spectrometry. Four
glucoconjugates were subjected to a preliminary in vitro
assessment including molecular recognition, cytotoxicity, and
cellular uptake studies. The molecular recognition studies
revealed that the atom involved in cluster conjugation has a
marked eect on the GLUT1 anity; species connected to a
carbon atom have a significantly higher anity than species
connected to a boron atom. While qualitatively agreeing with
the experimental observations, the molecular basis of these
observations could not be completely uncovered with our
current computational model. More work in this area including
considerations on dynamics and more accurate descriptions of
intermolecular interactions is needed to fully understand the
underlying factors at play.
All four glucoconjugates displayed acceptable cytotoxicity
profiles and their boron delivery capacity was found to be in
the same range as those of the clinically employed agents BPA
Figure 5. Results from cellular uptake studies performed in the CAL 27 cell line. The following substrates were included: 1(circle), 2(diamond), 3
(triangle), 4(inverted triangle), BPA (square), and BSH (asterisk) across the 10−400 μM concentration range. The incubation times were 5, 30,
and 120 min. (n= 3 at all three time points). The following Michaelis−Menten kinetic parameters were obtained (Vmax is given as μg B/mg
protein; Kmis given as μM) at 5 min incubation time, 1:Vmax = 1.34; Km= 1827.8, 2:Vmax = 6.60; Km= 1813, and 3:Vmax = 0.03; Km= 109.7. At 30
min incubation time, 1:Vmax = 0.83; Km= 555.6, 2:Vmax = 1.80; Km= 236.9. At 120 min incubation time, 2:Vmax = 2.19; Km= 207.6.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30381

and BSH, apart from glucoconjugate 2, which was found to be
roughly five times more ecient. However, the boron delivery
capacity was found to be significantly lower compared to our
previous molecular libraries,
13,14
and especially, glucoconjugate
1seemed to display a GLUT1 antagonist behavior rather than
act as a promising delivery agent for BNCT. In addition to
providing knowledge that may help explain earlier observations
noted in the field,
17−20
our results provide understanding on
the boundaries within which this approach to BNCT can be
further developed. The new insights on the substrate specificity
of the GLUT1 transporter are important for the development
of other therapeutic approaches centered on this molecular
biology target. Overall, the results point toward the dire need
of continued studies on the biochemical foundations of
GLUT1-targeting strategies with more diverse substrate
libraries, a topic that we are currently in full pursuit of.
4. EXPERIMENTAL SECTION
4.1. Synthesis and Structural Characterization Data.
All starting materials and reagents were commercially
purchased and used without further purification. Solvents
used in reactions were purified with the VAC vacuum solvent
purification system and further dried over 4 Å molecular sieves.
Reactions were carried out under inert conditions using either
an argon or nitrogen atmosphere. For all NMR experiments, a
Bruker Avance III spectrometer was used (operating at 1H:
500.13 MHz, 13C: 125.76 MHz, 11B: 160.46 MHz) and the
probe temperature was kept at 25 °C. All intermediates as well
as final products were fully characterized using 1D (1H, 1D-
TOCSY, 13C{1H}, and 11B{1H}) and 2D (COSY, ed-HSQC,
and HMBC) NMR experiments with pulse sequences provided
by the instrument manufacturer. Spectral simulations were
performed with the ChemAdder software in order to obtain
precise chemical shifts and coupling constants. The coupling
constants are reported in Hz and provided when first
encountered. Coupling patterns are given as s (singlet), d
(doublet), dd (doublet of a doublet), etc. Chemical shifts are
expressed on the δscale (in ppm). The following reference
signals are employed: TMS (tetramethylsilane), residual
chloroform, methanol, or 15% BF3in CDCl3. HRMS were
obtained with a Bruker Micro Q-TOF with electrospray
ionization operated in positive mode. The purity of substrates
1−4was determined to be >95% by qNMR. TLC was
performed on aluminum sheets precoated with silica gel 60
F254 (Merck), and spots were uncovered by spraying with
conc. H2SO4:MeOH (1:5) solution followed by heating.
Compounds were purified by flash chromatography using
silica gel 40 as the stationary phase.
4.1.1. Substrate-Specific Analytical Data. 4.1.1.1. Methyl
2,3,4-Tri-O-acetyl-6-deoxy-6-iodo-α-D-glucopyranoside (5).
Synthesized over two steps, starting from methyl α-D-
glucopyranoside (5.00 g, 25.8 mmol, 1.00 equiv), which was
dissolved in toluene (100 mL) and heated to 80 °C. PPh3
(8.35 g, 31.8 mmol, 1.2 equiv) and imidazole (6.51 g, 95.6
mmol, 3.7 equiv) were added to the reaction mixture and left
to stir for 10 min. I2(9.96 g, 39.2 mmol, 1.5 equiv) was then
added to the reaction mixture over the course of 0.5 h after
which the mixture was left to reflux for 1 h. The reaction
mixture was then brought to room temperature, and the
product was extracted with water (3 ×100 mL). The crude
product was dried, then dissolved in Ac2O:pyridine 1:1 (100
mL), and left to stir overnight. The solvents were then
removed under reduced pressure, and the crude product was
purified by column chromatography (EtOAc:Hex 1:1).
Compound 5was obtained as a white solid (6.95 g, 16.1
mmol, 63%). TLC: Rf: 0.60 (EtOAc:Hex 1:1).
1H NMR (500.13 MHz, CDCl3, 25 °C): δ= 5.47 (dd, 1H,
J2,3 = 10.3, J3,4 = 9.3 Hz, H-3), 4.96 (d, 1H, J1,2 = 3.7 Hz, H-1),
4.89 (dd, 1H, H-2), 4.87 (dd, 1H, J4,5 = 9.8 Hz, H-4), 3.79
(ddd, 1H, J5,6a = 2.5, J5,6b = 8.3 Hz, H-5), 3.48 (s, 3H, 1-
OCH3), 3.30 (dd, 1H, J6a,6b =−10.9 Hz, H-6a), 3.14 (dd, 1H,
H-6b), 2.08 (s, 3H, 2-OCOCH3), 2.06 (s, 3H, 4-OCOCH3),
and 2.01 (s, 3H, 3-OCOCH3) ppm.
13C{1H} NMR (125.76 MHz, CDCl3, 25 °C): δ= 170.2 (2-
OCOCH3), 170.1 (3-OCOCH3), 169.8 (4-OCOCH3), 96.8
(C-1), 72.6 (C-4), 71.0 (C-2), 69.8 (C-3), 68.8 (C-5), 55.9 (1-
OCH3), 20.8 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3),
and 3.7 (C-6) ppm.
HRMS m/z: calcd for C13H19INaO8[M + Na]+, 453.0022;
found, 453.0125.
4.1.1.2. Methyl 2,3,4-Tri-O-acetyl-6-deoxy-6-thio-(1,2-di-
carba-closo-dodecaboran-1-yl)-α-D-glucopyranoside (6).
Compound 5(0.66 g, 1.53 mmol, 1.00 equiv) and 1-
(mercapto)-1,2-dicarba-closo-dodecaborane(12) (0.41 g, 2.33
mmol, 1.50 equiv) were placed in an oven-dried Schlenk flask
under an inert atmosphere. The mixture was suspended in dry
acetone (25 mL), and K2CO3(0.72 g, 5.21 mmol, 3.41 equiv)
was added. The reaction mixture was allowed to stir for 24 h at
60 °C. After cooling to room temperature, the pH was adjusted
to 7 using 1 M HCl. Acetone was removed under reduced
pressure, and the resulting aqueous solution was extracted with
ethyl acetate (3 ×30 mL). The combined organic layers were
washed with water and brine (30 mL) and dried over
magnesium sulfate. The crude product was purified by column
chromatography (EtOAc:Hex 2:3). Compound 6was obtained
as a colorless oil (0.40 g, 0.84 mmol, 55%). TLC: Rf: 0.65
(EtOAc:Hex 2:3).
1H NMR (500.13 MHz, CDCl3, 25 °C) δ= 5.45 (dd, 1H,
J2,3 = 10.3, J3,4 = 9.3 Hz, H-3), 4.92 (dd, 1H, J4,5 = 9.9 Hz, H-
4), 4.88 (d, 1H, J1,2 = 3.6 Hz, H-1), 4.82 (dd, 1H, H-2), 3.95
(ddd, 1H, J5,6a = 2.8, J5,6b = 8.2 Hz, H-5), 3.78 (br s, 1H,
cluster-CH), 3.40 (s, 3H, 1-OCH3), 3.13 (dd, 1H, J6a,6b =
−13.5 Hz, H-6a), 3.01 (dd, 1H, H-6b), 2.09 (s, 3H, 4-
OCOCH3), 2.06 (s, 3H, 2-OCOCH3), 2.00 (s, 3H, 3-
OCOCH3), and 3.15−1.54 (br m, 10H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CDCl3, 25 °C) δ= 170.3,
170.1, 170.0 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3),
96.8 (C-1), 74.5 (cluster-C), 71.5 (C-4), 70.8 (C-2), 69.8 (C-
3), 68.8 (cluster-CH), 68.2 (C-5), 55.8 (1-OCH3), 38.5 (C-6),
and 20.8 and 20.7 (2-OCOCH3, 3-OCOCH3and 4-
OCOCH3) ppm.
11B{1H} NMR (160.46 MHz, CDCl3, 25 °C) δ=−1.6,
−4.9, −8.7, −10.5, −12.6 ppm.
HRMS: m/zcalcd. for B10C15H30NaO8S [M + Na]+:
503.2490, found: 503.2494.
4.1.1.3. Methyl 2,3,4-Tri-O-acetyl-6-deoxy-6-thio-(1,7-di-
carba-closo-dodecaboran-1-yl)-α-D-glucopyranoside (7).
Compound 5(0.65 g, 1.51 mmol, 1.00 equiv) and 1-
(mercapto)-1,7-dicarba-closo-dodecaborane(12) (0.40 g, 2.27
mmol, 1.50 equiv) were placed in an oven-dried Schlenk flask
under an inert atmosphere. The mixture was suspended in dry
acetone (25 mL), and K2CO3(0.63 g, 4.54 mmol, 3.00 equiv)
was added. The reaction mixture was allowed to stir for 24 h at
60 °C. After cooling to room temperature, the pH was adjusted
to 7 using 1 M HCl. Acetone was removed under reduced
pressure, and the resulting aqueous solution was extracted with
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30382

ethyl acetate (3 ×30 mL). The combined organic layers were
washed with water and brine (30 mL) and dried over
magnesium sulfate. The crude product was purified by column
chromatography (EtOAc:Hex 1:2). Compound 7was obtained
as a colorless foam (0.72 g, 1.48 mmol, 98%). TLC: Rf: 0.61
(EtOAc:Hex 2:3).
1H NMR (500.13 MHz, CDCl3, 25 °C) δ= 5.44 (dd, 1H,
J2,3 = 10.2, J3,4 = 9.3 Hz, H-3), 4.88 (d, 1H, J1,2 = 3.6 Hz, H-1),
4.86 (dd, 1H, J4,5 = 9.9 Hz, H-4), 4.83 (dd, 1H, H-2), 3.89
(ddd, 1H, J5,6a = 2.6, J5,6b = 9.1 Hz, H-5), 3.41 (s, 3H, 1-
OCH3), 2.99 (br s, 1H, cluster-CH), 2.89 (dd, 1H, J6a,6b =
−13.6 Hz, H-6a), 2.83 (dd, 1H, H-6b), 2.08 (s, 3H, 4-
OCOCH3), 2.07 (s, 3H, 2-OCOCH3), 2.00 (s, 3H, 3-
OCOCH3), and 3.15−1.54 (br m, 10H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CDCl3, 25 °C) δ= 170.3,
170.1, 170.0 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3),
96.6 (C-1), 74.2 (cluster-C), 71.9 (C-4), 70.9 (C-2), 69.9 (C-
3), 68.2 (C-5), 55.7 (1-OCH3), 55.7 (cluster-CH), 37.8 (C-6),
20.9 and 20.8 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3)
ppm.
11B{1H} NMR (160.46 MHz, CDCl3, 25 °C) δ=−3.6,
−10.4, −13.3, and −14.5 ppm.
HRMS: m/zcalcd. for B10C15H30NaO8S [M + Na]+:
503.2490, found: 503.2504.
4.1.1.4. Methyl 2,3,4-Tri-O-acetyl-6-deoxy-6-thio-(1,2-di-
carba-closo-dodecaboran-9-yl)-α-D-glucopyranoside (8).
Compound 5(0.58 g, 1.34 mmol, 1.00 equiv) and 9-
(mercapto)-1,2-dicarba-closo-dodecaborane(12) (0.26 g, 1.48
mmol, 1.10 equiv) were placed in an oven-dried Schlenk flask
under an inert atmosphere. The mixture was suspended in dry
acetone (25 mL), and K2CO3(0.41 g, 2.95 mmol, 2.20 equiv)
was added. The reaction mixture was allowed to stir overnight
at 60 °C. After cooling to room temperature, 10 mL of
saturated aqueous ammonium chloride solution was added and
the pH was adjusted to 7 using 1 M HCl. Acetone was
removed under reduced pressure, and the resulting aqueous
solution was extracted with ethyl acetate (3 ×30 mL). The
combined organic layers were washed with water and brine (30
mL) and dried over magnesium sulfate. The crude product was
purified by column chromatography (EtOAc:Hex 1:4).
Compound 8was obtained as a colorless foam (0.51 g, 1.08
mmol, 80%). TLC: Rf: 0.35 (EtOAc:Hex 2:3).
1H NMR (500.13 MHz, CDCl3, 25 °C) δ= 5.43 (dd, 1H,
J2,3 = 10.2, J3,4 = 9.3 Hz, H-3), 4.91 (d, 1H, , J1,2 = 3.7 Hz, H-
1), 4.87 (dd, 1H, H-2), 4.86 (dd, 1H, J4,5 = 10.0 Hz, H-4), 3.84
(ddd, 1H, J5,6a = 2.4, J5,6b = 9.4 Hz, H-5), 3.58 and 3.47 (each
br s, each 1H, cluster-CH), 3.45 (s, 3H, 1-OCH3), 2.63 (dd,
1H, J6a,6b =−13.6 Hz, H-6a), 2.56 (dd, 1H, H-6b), 2.07 (s, 3H,
2-OCOCH3), 2.04 (s, 3H, 4-OCOCH3), 2.00 (s, 3H, 3-
OCOCH3), and 3.08−1.51 (br m, 9H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CDCl3, 25 °C) δ= 170.3,
170.3, 170.0 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3),
96.4 (C-1), 72.3 (C-4), 71.3 (C-2), 70.4 (C-3), 69.6 (C-5),
55.4 (1-OCH3), 53.0 and 47.8 (both cluster-CH), 33.7 (C-6),
and 20.9 and 20.8 (2-OCOCH3, 3-OCOCH3and 4-
OCOCH3) ppm.
11B{1H} NMR (160.46 MHz, CDCl3, 25 °C) δ= 7.1, −2.5,
−8.8, −14.5, and −15.5 ppm.
HRMS: m/zcalcd. for B10C15H30NaO8S [M + Na]+:
503.2490, found: 503.2458.
4.1.1.5. Methyl 2,3,4-Tri-O-acetyl-6-deoxy-6-thio-(1,7-di-
carba-closo-dodecaboran-9-yl)-α-D-glucopyranoside (9).
Compound 5(0.40 g, 0.93 mmol, 1.00 equiv) and 9-
(mercapto)-1,7-dicarba-closo-dodecaborane(12) (0.25 g, 1.39
mmol, 1.50 equiv) were placed in an oven-dried Schlenk flask
under an inert atmosphere. The mixture was suspended in dry
acetone (25 mL), and DIPEA (0.47 mL, 2,79 mmol, 3,00
equiv) was added. The reaction was allowed to stir overnight at
60 °C. After cooling to room temperature, 10 mL of saturated
aqueous ammonium chloride solution was added, and the pH
was adjusted to 7 using 1 M HCl. Acetone was removed under
reduced pressure, and the resulting aqueous solution was
extracted with ethyl acetate (3 ×30 mL). The combined
organic layers were washed with water and brine (30 mL) and
dried over magnesium sulfate. The crude product was purified
by column chromatography (EtOAc:Hex 1:1). Compound 9
was obtained as a colorless foam (0.40 g, 0.84 mmol, 90%).
TLC: Rf: 0.46 (EtOAc:Hex 2:3).
1H NMR (500.13 MHz, CDCl3, 25 °C) δ= 5.45 (dd, 1H,
J2,3 = 10.3, J3,4 = 9.3 Hz, H-3), 4.94 (d, 1H, J1,2 = 3.7 Hz, H-1),
4.90 (dd, 1H, J4,5 = 10.0 Hz, H-4), 4.88 (dd, 1H, H-2), 3.90
(ddd, 1H, J5,6a = 2.4, J5,6b = 9.3 Hz, H-5), 3.47 (s, 3H, 1-
OCH3), 2.97 (br s, 2H, cluster-CH), 2.73 (dd, 1H, J6a,6b =
−13.6 Hz, H-6a), 2.65 (dd, 1H, H-6b), 2.07 (s, 3H, 4-
OCOCH3), 2.04 (s, 3H, 2-OCOCH3), 2.00 (s, 3H, 3-
COCH3), and 3.15−1.54 (br m, 9H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CDCl3, 25 °C) δ= 170.3,
170.2, 170.0 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3), 96.
5 (C-1), 72.3 (C-4), 71.2 (s, C-2), 70.4 (s, C-3), 69.6 (s, C-5),
55.5 (1-OCH3), 54.2 (both cluster-CH), 34.1 (C-6), 20.9 and
20.8 (2-OCOCH3, 3-OCOCH3and 4-OCOCH3) ppm.
11B{1H} NMR (160.46 MHz, CDCl3, 25 °C) δ= 0.2, −6.5,
−10.0, −13.2, −13.9, −17.6, and −20.5 ppm.
HRMS: m/zcalcd. for B10C15H30NaO8S [M + Na]+:
503.2490, found: 503.2466.
4.1.1.6. 6-Deoxy-6-thio-(1,2-dicarba-closo-dodecaboran-
1-yl)-D-glucopyranose (1). Compound 6(0.043 g, 0.090
mmol) was dissolved in 2 M HCl (3 mL) and stirred at 90 →
115 °C. After 7 h, the reaction mixture was brought to room
temperature, cooled with an ice bath, and neutralized by the
addition of Na2CO3. The reaction mixture was concentrated,
and the residue was dissolved in MeOH (10 mL) and stirred
for 15 min, after which the formed solid was removed by
filtration. The filtrate was concentrated, and the crude product
was purified by column chromatography (DCM:MeOH 7:1).
Compound 1was obtained as a white solid (0.009 g, 0.026
mmol, 35%, α/β56:44). TLC: Rf: 0.34 (DCM:MeOH 7:1).
αanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
5.02 (d, 1H, J1,2 = 3.7 Hz, H-1), 4.66 (br s, 1H, cluster-CH),
3.92 (ddd, 1H, J4,5 = 9.7, J5,6a = 2.9, J5,6b = 8.2 Hz, H-5), 3.62
(dd, 1H, J2,3 = 9.6, J3,4 = 8.9 Hz, H-3), 3.44 (dd, 1H, J6a,6b =
−12.6 Hz, H-6a), 3.32 (dd, 1H, H-2), 3.14 (dd, 1H, H-4), 3.10
(dd, 1H, H-6b), and 3.05−1.50 (br m, 10H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 94.0 (C-
1), 77.0 (cluster-C), 74.8 (C-4), 74.6 (C-3), 73.7 (C-2), 71.2
(C-5), 70.1 (cluster-CH), and 40.4 (C-6) ppm.
βanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
4.73 (br s, 1H, cluster-CH), 4.43 (d, 1H, J1,2 = 7.8 Hz, H-1),
3.46 (dd, 1H, J5,6a = 2.8, J6a,6b =−13.2 Hz, H-6a), 3.40 (ddd,
1H, J4,5 = 9.5, J5,6b = 8.2 Hz, H-5), 3.30 (dd, 1H, J2,3 = 9.4, J3,4
= 8.9 Hz, H-3), 3.17 (dd, 1H, H-4), 3.11 (dd, 1H, H-6b), 3.10
(dd, 1H, H-2), and 3.05−1.50 (br m, 10H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 98.3 (C-
1), 77.7 (C-3), 76.8 (cluster-C), 76.2 (C-2), 76.0 (C-5), 74.4
(C-4), 69.9 (cluster-CH), and 40.1 (C-6) ppm.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30383

11B{1H} NMR (160.46 MHz, CD3OD, 25 °C): δ=−1.1,
−4.5, −8.5, and −11.7 ppm.
HRMS m/z: calcd for C8H22B10NaO5S [M + Na]+,
363.2016; found, 363.2002.
4.1.1.7. 6-Deoxy-6-thio-(1,7-dicarba-closo-dodecaboran-
1-yl)-D-glucopyranose (2). Compound 7(0.045 g, 0.094
mmol) was dissolved in 4 M HCl (3 mL) and stirred at 90 →
110 °C. After 4 h, the reaction mixture was brought to room
temperature, then cooled with an ice bath, and neutralized by
the addition of Na2CO3. The reaction mixture was
concentrated, and the residue was dissolved in MeOH (10
mL) and stirred for 15 min, after which the formed solid was
removed by filtration. The filtrate was concentrated, and the
crude product was purified by column chromatography
(DCM:MeOH 7:1). Compound 2was obtained as a white
solid (0.014 g, 0.040 mmol, 43%, α/β49:51). TLC: Rf: 0.33
(DCM:MeOH 7:1).
αanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
5.01 (d, 1H, J1,2 = 3.7 Hz, H-1), 3.89 (ddd, 1H, J4,5 = 9.7, J5,6a =
2.5, J5,6b = 8.6 Hz, H-5), 3.61 (dd, 1H, J2,3 = 9.6, J3,4 = 9.0 Hz,
H-3), 3.60 (br s, 1H, cluster-CH), 3.31 (dd, 1H, H-2), 3.29
(dd, 1H, J6a,6b =−12.4 Hz, H-6a), 3.10 (dd, 1H, H-4), 2.89
(dd, 1H, H-6b), and 2.83−1.48 (br m, 10H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 93.9 (C-
1), 75.0 (C-4), 74.6 (C-3), 73.8 (C-2), 73.0 (cluster-C), 71.1
(C-5), 57.6 (cluster-CH), and 39.8 (C-6) ppm.
βanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
4.41 (d, 1H, J1,2 = 7.9 Hz, H-1), 3.60 (br s, 1H, cluster-CH),
3.35 (ddd, 1H, J4,5 = 9.7, J5,6a = 2.3, J5,6b = 8.7 Hz, H-5), 3.32
(dd, 1H, J6a,6b =−12.7 Hz, H-6a), 3.29 (dd, 1H, J2,3 = 9.5, J3,4
= 8.6 Hz, H-3), 3.13 (dd, 1H, H-4), 3.09 (dd, 1H, H-2), 2.91
(dd, 1H, H-6b), and 2.83−1.48 (br m, 10H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 98.3 (C-
1), 77.7 (C-3), 76.2 (C-2), 76.1 (C-5), 74.7 (C-4), 73.0
(cluster-C), 57.6 (cluster-CH), and 39.6 (C-6) ppm.
11B{1H} NMR (160.46 MHz, CD3OD, 25 °C): δ=−2.8,
−9.4, −10.1, −12.7, and −13.7 ppm.
HRMS m/z: calcd for C8H22B10NaO5S [M + Na]+,
363.2016; found, 363.1998.
4.1.1.8. 6-Deoxy-6-thio-(1,2-dicarba-closo-dodecaboran-
9-yl)-D-glucopyranose (3). Compound 8(0.089 g, 0.185
mmol) was dissolved in 1 M HCl (5 mL) and stirred at 90 →
120 °C. After 7 h, the reaction mixture was brought to room
temperature, then cooled with an ice bath, and neutralized by
the addition of Na2CO3. The reaction mixture was
concentrated, and the residue was dissolved in MeOH (15
mL) and stirred for 15 min, after which the formed solid was
removed by filtration. The filtrate was concentrated, and the
crude product was purified by column chromatography
(DCM:MeOH 7:1). Compound 3was obtained as a white
solid (0.047 g, 0.138 mmol, 75%, α/β48:52). TLC: Rf: 0.30
(DCM:MeOH 7:1).
αanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
5.04 (d, 1H, J1,2 = 3.8 Hz, H-1), 4.29 (br s, 2H, cluster-CH),
3.83 (ddd, 1H, J4,5 = 9.7, J5,6a = 2.7, J5,6b = 8.2 Hz, H-5), 3.61
(dd, 1H, J2,3 = 9.6, J3,4 = 8.9 Hz, H-3), 3.32 (dd, 1H, H-2), 3.13
(dd, 1H, H-4), 2.97 (dd, 1H, J6a,6b =−12.8 Hz, H-6a), 2.57
(dd, 1H, H-6b), and 2.94−1.46 (br m, 9H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 93.8 (C-
1), 74.9 (C-4), 74.7 (C-3), 73.8 (C-2), 72.5 (C-5), 55.7 and
50.0 (both cluster-CH), and 35.5 (C-6) ppm.
βanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
4.42 (d, 1H, J1,2 = 7.8 Hz, H-1), 4.37 (br s, 2H, cluster-CH),
3.29 (dd, 1H, J2,3 = 9.4, J3,4 = 9.3 Hz, H-3), 3.27 (ddd, 1H, J4,5
= 9.3, J5,6a = 2.4, J5,6b = 8.8 Hz, H-5), 3.12 (dd, 1H, H-4), 3.10
(dd, 1H, H-2), 3.01 (dd, 1H, J6a,6b =−13.2 Hz, H-6a), 2.53
(dd, 1H, H-6b), and 2.94−1.46 (br m, 9H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 98.1 (C-
1), 77.9 (C-3), 77.8 (C-5), 76.3 (C-2), 75.1 (C-4), 55.7 and
50.0 (both cluster-CH), and 35.3 (C-6) ppm.
11B{1H} NMR (160.46 MHz, CD3OD, 25 °C): δ= 7.4,
−2.3, −8.3, −13.2, −13.9, and −14.6 ppm.
HRMS m/z: calcd for C8H22B10NaO5S [M + Na]+,
363.2016; found, 363.2035.
4.1.1.9. 6-Deoxy-6-thio-(1,7-dicarba-closo-dodecaboran-
9-yl)-D-glucopyranose (4). Compound 9(0.099 g, 0.206
mmol) was dissolved in 1 M HCl (6 mL) and stirred at 80 →
115 °C. After 7 h, the reaction mixture was brought to room
temperature, cooled with an ice bath, and neutralized by the
addition of Na2CO3. The reaction mixture was concentrated,
and the residue was dissolved in MeOH (15 mL) and stirred
for 15 min, after which the formed solid was removed by
filtration. The filtrate was concentrated, and the crude product
was purified by column chromatography (DCM:MeOH 7:1).
Compound 4was obtained as a white solid (0.052 g, 0.153
mmol, 74%, α/β55:45). TLC: Rf: 0.32 (DCM:MeOH 7:1).
αanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
5.07 (d, 1H, J1,2 = 3.8 Hz, H-1), 3.89 (ddd, 1H, J4,5 = 9.7, J5,6a =
2.7, J5,6b = 8.2 Hz, H-5), 3.63 (dd, 1H, J2,3 = 9.6, J3,4 = 8.9 Hz,
H-3), 3.55 (br s, 2H, cluster-CH), 3.35 (dd, 1H, H-2), 3.17
(dd, 1H, H-4), 3.08 (dd, 1H, J6a,6b =−12.7 Hz, H-6a), 2.68
(dd, 1H, H-6b), and 3.04−1.50 (br m, 9H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 93.8 (C-
1), 75.0 (C-4), 74.7 (C-3), 73.8 (C-2), 72.5 (C-5), 56.0 (both
cluster-CH), and 35.9 (C-6) ppm.
βanomer: 1H NMR (500.13 MHz, CD3OD, 25 °C): δ=
4.46 (d, 1H, J1,2 = 7.8 Hz, H-1), 3.55 (br s, 2H, cluster-CH),
3.33 (ddd, 1H, J4,5 = 9.0, J5,6a = 2.6, J5,6b = 9.8 Hz, H-5), 3.31
(dd, 1H, J2,3 = 9.2, J3,4 = 9.4 Hz, H-3), 3.16 (dd, 1H, H-4), 3.13
(dd, 1H, H-2), 3.12 (dd, 1H, J6a,6b =−12.9 Hz, H-6a), 2.65
(dd, 1H, H-6b), and 3.04−1.50 (br m, 9H, cluster-BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 98.2 (C-
1), 77.9 (C-3), 77.8 (C-5), 76.3 (C-2), 74.9 (C-4), 56.0 (both
cluster-CH), and 35.7 (C-6) ppm.
11B{1H} NMR (160.46 MHz, CD3OD, 25 °C): δ= 1.0,
−6.1, −9.4, −12.4, −13.4, −16.8, and −19.8 ppm.
HRMS m/z: calcd for B10C8H22NaO5S [M + Na]+,
363.2016; found, 363.1976.
4.1.1.10. Methyl 6-Deoxy-6-thio-(1,7-dicarba-closo-do-
decaboran-1-yl)-D-glucopyranoside (10). Compound 7
(0.049 g, 0.105 mmol) was dissolved in 1 M HCl (3 mL)
and stirred at 105 °C. After 2 h, the reaction mixture was
brought to room temperature and neutralized with the
addition of Na2CO3at 0 °C. The reaction mixture was
concentrated, and the residue was dissolved in MeOH (10
mL) and stirred for 15 min, after which the formed solid was
removed by filtration. The filtrate was concentrated, and the
crude product was purified by column chromatography
(DCM:MeOH 7:1). Compound 10 was obtained as a clear
oil (0.036 g, 0.088 mmol, 84%). TLC: Rf: 0.68 (DCM/MeOH
7:1).
1H NMR (500.13 MHz, CD3OD, 25 °C): δ= 4.59 (d, 1H,
J1,2 = 3.8 Hz, H-1), 3.63 (br s, 1H, cluster-CH), 3.59 (ddd, 1H,
J4,5 = 9.7, J5,6a = 2.4, J5,6b = 9.4 Hz, H-5), 3.55 (dd, 1H, J2,3 =
9.7, J3,4 = 8.9 Hz, H-3), 3.38 (s, 3H, 1-OCH3), 3.36 (dd, 1H,
H-2), 3.31 (dd, 1H, J6a,6b =−12.9 Hz, H-6a), 3.09 (dd, 1H, H-
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30384

4), 2.87 (dd, 1H, H-6b), and 2.81−1.48 (br m, 10H, cluster-
BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 101.2
(C-1), 75.0 (C-4), 74.9 (C-3), 74.2 (cluster-C), 73.5 (C-2),
71.8 (C-5), 57.6 (cluster-CH), 55.7 (1-OCH3), and 39.6 (C-6)
ppm.
11B{1H} NMR (160.46 MHz, CD3OD, 25 °C): δ=−2.2,
−8.8, −9.5, −12.1, and −13.1 ppm.
HRMS m/z: calcd for B10C9H24NaO5S [M + Na]+,
377.2173; found, 377.2105.
4.1.1.11. Methyl 6-Deoxy-6-thio-(1,7-dicarba-closo-do-
decaboran-9-yl)-D-glucopyranoside (11). Compound 9
(0.052 g, 0.112 mmol) was dissolved in 1 M HCl (5 mL)
and stirred at 85 °C. After 4 h, the reaction mixture was
brought to room temperature and neutralized with the
addition of Na2CO3at 0 °C. The reaction mixture was
concentrated, and the residue was dissolved in MeOH (10
mL) and stirred for 15 min, after which the formed solid was
removed by filtration. The filtrate was concentrated, and the
crude product was purified by column chromatography
(DCM:MeOH 7:1). Compound 11 was obtained as a clear
oil (0.039 g, 0.083 mmol, 74%). TLC: Rf: 0.41 (DCM/MeOH
9:1).
1H NMR (500.13 MHz, CD3OD, 25 °C): δ= 4.63 (d, 1H,
J1,2 = 3.8 Hz, H-1), 3.60 (ddd, 1H, J4,5 = 9.7, J5,6a = 2.0, J5,6b =
9.1 Hz, H-5), 3.57 (dd, 1H, J2,3 = 9.7, J3,4 = 8.9 Hz, H-3), 3.56
(br s, 2H, cluster-CH), 3.44 (s, 3H, 1-OCH3), 3.39 (dd, 1H,
H-2), 3.11 (dd, 1H, H-4), 3.10 (dd, 1H, J6a,6b =−13.2 Hz, H-
6a), 2.60 (dd, 1H, H-6b), and 3.05−1.46 (br m, 9H, cluster-
BH) ppm.
13C{1H} NMR (125.76 MHz, CD3OD, 25 °C): δ= 100.9
(C-1), 75.2 (C-4), 75.1 (C-3), 73.7 (C-2), 73.3 (C-5), 56.1
(both cluster-CH), 55.6 (1-OCH3), and 35.7 (C-6) ppm.
11B{1H} NMR (160.46 MHz, CD3OD, 25 °C): δ= 1.1,
−6.0, −9.4, −12.4, −13.3, −16.7, and −19.7 ppm.
HRMS m/z: calcd for B10C9H24NaO5S [M + Na]+,
377.2173; found, 377.2139.
4.2. In Vitro Assessment Protocols. The CAL 27
(ATCC CRL-2095) cells used in all assays were acquired
from ATCC (Manassas, VA, USA) and cultured in Dulbecco’s
Modified Eagle Medium (DMEM) supplemented with L-
glutamine (2.0 mM), heat-inactivated fetal bovine serum
(10%), and penicillin (50 U/mL)-streptomycin (50 μg/mL) at
37 °C with 5% CO2and 95% relative humidity.
4.2.1. Anity Studies. The GLUT1 anity assay was
performed according to our previously developed protocol.
13,14
The cells were incubated at room temperature for 5 min with
1−4, BSH, and BPA across the 0.1−1800 μM concentration
range. In addition, the solutions contained 1.8 μM (0.1 mCi/
mL) of [14C]-D-glucose in glucose-free HBSS (Hanks’
balanced salt solution) (250 μL). Ice-cold buer was used to
quench the reactions. Afterward, the cells were washed twice.
Lysis was carried out with 0.1 M NaOH (250 μL) before
further mixing with emulsifier safe cocktail (1.0 mL)
(PerkinElmer, Waltham, MA, USA). Liquid scintillation
counter (MicroBeta2counter, PerkinElmer, Waltham, MA,
USA) was used to assess the radioactivity, and the IC50 values
were determined by nonlinear regression analysis.
4.2.2. Cellular Uptake Studies. The cells were first pre-
incubated. The substrates were then added in dierent
concentration (10−400 μM in 250 μL of HBSS) and
incubation was continued for 5, 30, and 120 min at room
temperature. Ice-cold buer was used to quench the reaction at
the set time points. Washing and lysis was carried out as
mentioned above. A combined sample was prepared in an
Eppendorf tube by addition of cell lysate from four wells. The
combined sample was centrifuged at +4 °C. From the
supernatant of every sample, 800 μL was taken and digested
in 1.0 mL of HNO3(TraceMetal grade, Fisher Chemical) for
24 h. Milli-Q water (USF Elga Purelab Ultra) was added until
the sample volume was 10 mL, and the boron contents were
assessed in triplicates by ICP-MS. More details on the ICP
method can be found in our earlier work.
13,14
PerkinElmer
Syngistix Data Analysis Software was used for data analysis and
GraphPad Prism v.5.03 software (GraphPad Software, San
Diego, Ca, USA) for statistical analysis.
4.2.3. Cytotoxicity Studies. The CAL 27 cells were seeded
on an opaque-walled 96-well plate at a density of 5000 cells per
well and incubated overnight. The cell culture media
containing dierent substrates at concentrations of 5, 25, 50,
125, and 250 μM were incubated with the cells for 6 and 24 h
at 37 °C, 5% CO2atmosphere, and 95% relative humidity. At
designated time points, the test media were disposed followed
by washing of the cells two times with 1 ×DPBS (pH 7.4). For
the cell viability assay, 1×HBSS and CellTiter-Glo reagent (50
μL each) were added to each well and left for cell lysis at
ambient temperature in the dark. The Synergy H1 Hybrid
multimode microplate reader (BioTek, Winooski, VT, USA)
was used to measure the luminescence signal of viable cells.
After luminescent reading, the Pierce colorimetric bicincho-
ninic acid (BCA) protein assay (Thermo Fisher Scientific,
Waltham, MA, USA) was used to quantify the total protein
content of every sample. Briefly, 25 μL of the sample in each
well was taken to a new transparent 96-well plate. Then, 200
μL of BCA working reagent was added and the plates were
wrapped with aluminum foil and incubated at 37 °C for 30
min. The absorbance was recorded at 562 nm using the
microplate reader. The protein content in each sample was
calculated using the BSA standard curve (concentration range:
0−2000 μg/mL). The protein content was used for the
normalization of the cell viability data. All assays were carried
out in triplicate.
4.3. Computational and Docking Studies. Starting
geometries for the studied molecules were built and these were
first optimized with xtb (version 6.3.2) followed by conforma-
tional sampling with CREST (Conformer-Rotamer Ensemble
Sampling Tool) using the GFN2-xTB method.
47
The re-
strained electrostatic potential (RESP) partial charges were
calculated for the CREST best conformation of each
molecule.
48
Since the conformation of the molecule aects
the calculation of the partial charges, we wanted to first
perform a small docking protocol to get a conformation that
represents docked/bound structures well. A docking study was
conducted, with 300 independent search runs. From this, the
cluster with the lowest energy featuring the majority of
structures was chosen as the representing one, and the lowest
energy conformer was chosen for further use. Since
AutoDock
49,50
merges polar hydrogens for the docking, those
hydrogens were added back to the ligand molecules in the next
step of the geometry optimization. The RESP partial charge
calculation was conducted for these, and the final results were
obtained from a docking study featuring 2000 independent
search runs for each ligand. A local minimum was identified on
DFT level by optimizing the geometries with dispersion-
corrected hybrid Tao−Perdew−Scuseria−Staroverov func-
tional TPSSh-D3(BJ).
51−53
The def2-TZVPP basis set was
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30385

used.
54
To compute the partial charges of the atoms, the RESP
protocol was used.
48
For RESP charge calculations, each ligand
was split into two parts (one part consisting of the carborane,
the sulfur atom, and the 6th position of the glucose backbone,
and the other part comprising the carbohydrate core).
Hydrogen atoms at the same carbon atom were treated as
equal. Turbomole 7.4.1 was used for the optimization of the
geometries,
55,56
and NWChem 6.8.1 was used for RESP
calculations.
57
AutoDock 4.2.6 was used in molecular docking studies.
49,50
In the carborane part, all rotatable bonds were rendered
inactive, i.e., set as nonrotatable. Further, the torsional degrees
of freedom for the carborane were set to 7 (torsdof 7). The
XylE inward open 4QIQ
41
and outward-open 6N3I
42
PDB
structures were used as the base. PyMOL was used to mutate
these structures by replacing Gln-415 with Asn-415. The
rotamer, in which the lowest number of clashes between
neighboring amino acids was reported, was selected. The
preparation of the transporter was performed in the following
way: the ligand and small molecules present (Zn for 4QIQ)
were removed, hydrogens were added and then merged, and
Gasteiger partial charges were calculated. A grid size of 46 ×56
×60 was selected with a 0.375 value for grid spacing. For the
binding site to be covered by the grid box, the midpoint of the
protein cavity was set as the center of the grid. The protein was
kept in a rigid state during the docking studies while the
torsional angles in the ligand were altered. Two separate
dockings were performed for each ligand: the initial one with
300 independent search runs and the second one with 2000.
The following parameters were employed: a maximum of 2.5
million energy evaluations and a maximum of 27,000
generations with a population size of 150. Further, the
Lamarckian genetic algorithm (LGA) was used in the default
setting, i.e., employing a 0.02 mutation rate and a 0.8 crossover
rate, with the top individual moving onto the next generation.
Last, ranking/clustering of conformations was done with a
cluster RMS 2.0 Å.
The missing boron parameters in AutoDock were added to
the parameter file as follows: R 2.285, Rii 4.57, epsilon 0.179,
vol 49.9744. Remaining parameters were given analogous
values as for carbon. The chosen parameters are based on the
work of Oda et al.,
58
as reproduced by Couto et al.
59
The
parameter set employed was thus atom_par B 4.57 0.179
49.9744−0.00143 0.0 0.0 0 −1−1 0 #Boron for Carborane.
■ASSOCIATED CONTENT
*
sı Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acsomega.2c03646.
All NMR spectra (1H, 13C, and 11B) of synthesized
compounds as well as the overall mean binding energy
(MBE) of each glucoconjugate (PDF)
■AUTHOR INFORMATION
Corresponding Author
Filip S. Ekholm −Department of Chemistry, University of
Helsinki, FI-00014 Helsinki, Finland; orcid.org/0000-
0002-4461-2215; Email: filip.ekholm@helsinki.fi
Authors
Jelena Matovic−Department of Chemistry, University of
Helsinki, FI-00014 Helsinki, Finland; orcid.org/0000-
0001-6529-2671
Juulia Järvinen −School of Pharmacy, University of Eastern
Finland, FI-70211 Kuopio, Finland
Iris K. Sokka −Department of Chemistry, University of
Helsinki, FI-00014 Helsinki, Finland; orcid.org/0000-
0002-5148-4987
Philipp Stockmann −Institute of Inorganic Chemistry, Leipzig
University, D-04103 Leipzig, Germany
Martin Kellert −Institute of Inorganic Chemistry, Leipzig
University, D-04103 Leipzig, Germany
Surachet Imlimthan −Department of Chemistry, University of
Helsinki, FI-00014 Helsinki, Finland; orcid.org/0000-
0003-2520-2146
Mirkka Sarparanta −Department of Chemistry, University of
Helsinki, FI-00014 Helsinki, Finland; orcid.org/0000-
0002-2956-4366
Mikael P. Johansson −Department of Chemistry, University
of Helsinki, FI-00014 Helsinki, Finland; Helsinki Institute of
Sustainability Science, HELSUS, FI-00014 Helsinki, Finland;
CSC −IT Center for Science Ltd., FI-02101 Espoo, Finland;
orcid.org/0000-0002-9793-8235
Evamarie Hey-Hawkins −Institute of Inorganic Chemistry,
Leipzig University, D-04103 Leipzig, Germany;
orcid.org/0000-0003-4267-0603
Jarkko Rautio −School of Pharmacy, University of Eastern
Finland, FI-70211 Kuopio, Finland
Complete contact information is available at:
https://pubs.acs.org/10.1021/acsomega.2c03646
Author Contributions
#
J.M., J.J., and I.K.S. contributed equally.
Notes
The authors declare no competing financial interest.
■ACKNOWLEDGMENTS
The authors would like to thank the Jane and Aatos Erkko
Foundation, the Academy of Finland (#308329, #341106,
#318422, and #320102), the Chemistry and Molecular
Sciences Graduate School, the Finnish Pharmaceutical Society,
the Cancer Foundation, the Ruth and Nils-Erik Stenbäck
Foundation, the University of Helsinki Research Funds, Orion
Research Foundation, and the DAAD (#57567358) for
financial support. The authors are further grateful to
CSC−IT Center for Science for providing access to computa-
tional resources.
■REFERENCES
(1) Thorens, B.; Mueckler, M. Glucose Transporters in the 21st
Century. Am. J. Physiol.: Endocrinol. Metab. 2010,298, E141−E145.
(2) Wright, E. M.; Loo, D. D. F.; Hirayama, B. A. Biology of Human
Sodium Glucose Transporters. Physiol. Rev. 2011,91, 733−794.
(3) Gynther, M.; Ropponen, J.; Laine, K.; Leppänen, J.; Haapakoski,
P.; Peura, L.; Järvinen, T.; Rautio, J. Glucose Promoiety Enables
Glucose Transporter Mediated Brain Uptake of Ketoprofen and
Indomethacin Prodrugs in Rats. J. Med. Chem. 2009,52, 3348−3353.
(4) Rautio, J.; Laine, K.; Gynther, M.; Savolainen, J. Prodrug
Approaches for CNS Delivery. AAPS J. 2008,10, 92−102.
(5) Patching, S. G. Glucose Transporters at the Blood-Brain Barrier:
Function, Regulation and Gateways for Drug Delivery. Mol. Neurobiol.
2017,54, 1046−1077.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30386

(6) Szablewski, L. Expression of Glucose Transporters in Cancers.
Biochim. Biophys. Acta, Rev. Cancer 2013,1835, 164−169.
(7) Warburg, O. On the Origin of Cancer Cells. Science 1956,123,
309−314.
(8) Vander Heiden, M. G.; Cantley, L. C.; Thompson, C. B.
Understanding the Warburg Effect: The Metabolic Requirements of
Cell Proliferation. Science 2009,324, 1029−1033.
(9) Gallamini, A.; Zwarthoed, C.; Borra, A. Positron Emission
Tomography (PET) in Oncology. Cancers 2014,6, 1821−1889.
(10) Barth, R. F.; Zhang, Z.; Liu, T. A Realistic Appraisal of Boron
Neutron Capture Therapy as a Cancer Treatment Modality. Cancer
Commun. 2018,38, 36.
(11) Goodman, J. H.; Yang, W.; Barth, R. F.; Gao, Z.; Boesel, C. P.;
Staubus, A. E.; Gupta, N.; Gahbauer, R. A.; Adams, D. M.; Gibson, C.
R.; Ferketich, A. K.; Moeschberger, M. L.; Soloway, A. H.; Carpenter,
D. E.; Albertson, B. J.; Bauer, W. F.; Zhang, M. Z.; Wang, C. C. Boron
Neutron Capture Therapy of Brain Tumors: Biodistribution,
Pharmacokinetics, and Radiation Dosimetry of Sodium Borocaptate
in Patients with Gliomas. Neurosurgery 2000,47, 608−622.
(12) Barth, R. F.; Grecula, J. C. Boron Neutron Capture Therapy at
the Crossroads - Where Do We Go from Here? Appl. Radiat. Isot.
2020,160, No. 109029.
(13) Matovic, J.; Järvinen, J.; Sokka, I. K.; Imlimthan, S.; Raitanen, J.
E.; Montaser, A.; Maaheimo, H.; Huttunen, K. M.; Peräniemi, S.;
Airaksinen, A. J.; Sarparanta, M.; Johansson, M. P.; Rautio, J.; Ekholm,
F. S. Exploring the Biochemical Foundations of a Successful GLUT1-
Targeting Strategy to BNCT: Chemical Synthesis and in Vitro
Evaluation of the Entire Positional Isomer Library of Ortho-
Carboranylmethyl-Bearing Glucoconjugates. Mol. Pharmaceutics
2021,18, 285−304.
(14) Matovic, J.; Järvinen, J.; Bland, H. C.; Sokka, I. K.; Imlimthan,
S.; Ferrando, R. M.; Huttunen, K. M.; Timonen, J.; Peräniemi, S.;
Aitio, O.; Airaksinen, A. J.; Sarparanta, M.; Johansson, M. P.; Rautio,
J.; Ekholm, F. S. Addressing the Biochemical Foundations of a
Glucose-Based “Trojan Horse”-Strategy to Boron Neutron Capture
Therapy: From Chemical Synthesis to In Vitro Assessment. Mol.
Pharmaceutics 2020,17, 3885−3899.
(15) Nemoto, H.; Cai, J.; Iwamoto, S.; Yamamoto, Y. Synthesis and
Biological Properties of Water-Soluble p-Boronophenylalanine
Derivatives. Relationship between Water Solubility, Cytotoxicity,
and Cellular Uptake. J. Med. Chem. 1995,38, 1673−1678.
(16) Wittig, A.; Collette, L.; Appelman, K.; Buhrmann, S.; Jäckel, M.
C.; Jöckel, K.-H.; Schmid, K. W.; Ortmann, U.; Moss, R.; Sauerwein,
W. A. G. EORTC Trial 11001: Distribution of Two 10B-Compounds
in Patients with Squamous Cell Carcinoma of Head and Neck, a
Translational Research/Phase 1 Trial. J. Cell. Mol. Med. 2009,13,
1653−1665.
(17) Imperio, D.; Panza, L. Sweet Boron: Boron-Containing Sugar
Derivatives as Potential Agents for Boron Neutron Capture Therapy.
Symmetry 2022,14, 182.
(18) Tietze, L. F.; Bothe, U.; Griesbach, U.; Nakaichi, M.;
Hasegawa, T.; Nakamura, H.; Yamamoto, Y. Ortho-Carboranyl
Glycosides for the Treatment of Cancer by Boron Neutron Capture
Therapy. Bioorg. Med. Chem. 2001,9, 1747−1752.
(19) Satapathy, R.; Dash, B. P.; Mahanta, C. S.; Swain, B. R.; Jena, B.
B.; Hosmane, N. S. Glycoconjugates of Polyhedral Boron Clusters. J.
Organomet. Chem. 2015,798, 13−23.
(20) Soloway, A. H.; Tjarks, W.; Barnum, B. A.; Rong, F.-G.; Barth,
R. F.; Codogni, I. M.; Wilson, J. G. The Chemistry of Neutron
Capture Therapy. Chem. Rev. 1998,98, 1515−1562.
(21) Maurer, J. L.; Serino, A. J.; Hawthorne, M. F. Hydrophilically
Augmented Glycosyl Carborane Derivatives for Incorporation in
Antibody Conjugation Reagents. Organometallics 1988,7, 2519−
2524.
(22) Giovenzana, G. B.; Lay, L.; Monti, D.; Palmisano, G.; Panza, L.
Synthesis of Carboranyl Derivatives of Alkynyl Glycosides as Potential
BNCT Agents. Tetrahedron 1999,55, 14123−14136.
(23) Stadlbauer, S.; Lönnecke, P.; Welzel, P.; Hey-Hawkins, E. Bis-
Carbaborane-Bridged Bis-Glycophosphonates as Boron-Rich Delivery
Agents for BNCT. Eur. J. Org. Chem. 2010,2010, 3129−3139.
(24) Mori, Y.; Suzuki, A.; Yoshino, K.; Kakihana, H. Complex
Formation of P-Boronophenylalanine With Some Monosaccharides.
Pigm. Cell Res. 1989,2, 273−277.
(25) Patra, M.; Johnstone, T. C.; Suntharalingam, K.; Lippard, S. J. A
Potent Glucose-Platinum Conjugate Exploits Glucose Transporters
and Preferentially Accumulates in Cancer Cells. Angew. Chem., Int. Ed.
2016,55, 2550−2554.
(26) Vinas, C.; Benakki, R.; Teixidor, F.; Casabo, J. Dimethoxy-
ethane as a Solvent for the Synthesis of C-Monosubstituted o-
Carborane Derivatives. Inorg. Chem. 1995,34, 3844−3845.
(27) Zakharkin, L. I.; Pisareva, I. v. A New Simple Method For The
Production And Some Conversions Of B[Sbnd]S Bondcontaining o-
and m-carboranyl. Phosphorus Sulfur Relat. Elem. 1984,20, 357−370.
(28) Kellert, M.; Worm, D. J.; Hoppenz, P.; Sárosi, M. B.; Lönnecke,
P.; Riedl, B.; Koebberling, J.; Beck-Sickinger, A. G.; Hey-Hawkins, E.
Modular Triazine-Based Carborane-Containing Carboxylic Acids −
Synthesis and Characterisation of Potential Boron Neutron Capture
Therapy Agents Made of Readily Accessible Building Blocks. Dalton
Trans. 2019,48, 10834−10844.
(29) Garegg, P. J.; Samuelsson, B. Novel Reagent System for
Converting a Hydroxy-Group into an Iodo-Group in Carbohydrates
with Inversion of Configuration. J. Chem. Soc., Chem. Commun. 1979,
22, 978−980.
(30) Abayakoon, P.; Epa, R.; Petricevic, M.; Bengt, C.; Mui, J. W.-Y.;
van der Peet, P. L.; Zhang, Y.; Lingford, J. P.; White, J. M.; Goddard-
Borger, E. D.; Williams, S. J. Comprehensive Synthesis of Substrates,
Intermediates, and Products of the Sulfoglycolytic Embden−Meyerh-
off−Parnas Pathway. J. Org. Chem. 2019,84, 2901−2910.
(31) Tietze, L. F.; Bothe, U.; Griesbach, U.; Nakaichi, M.;
Hasegawa, T.; Nakamura, H.; Yamamoto, Y. Carboranyl Bisglycosides
for the Treatment of Cancer by Boron Neutron Capture Therapy.
ChemBioChem 2001,2, 326−334.
(32) Zemplén, G.; Gerecs, A.; Hadácsy, I. Uber Die Verseifung
Acetylierter Kohlenhydrate. Ber. Dtsch. Chem. Ges. 1936,69, 1827−
1829.
(33) Valliant, J. F.; Guenther, K. J.; King, A. S.; Morel, P.; Schaffer,
P.; Sogbein, O. O.; Stephenson, K. A. The Medicinal Chemistry of
Carboranes. Coord. Chem. Rev. 2002,232, 173−230.
(34) Fokt, I.; Szymanski, S.; Skora, S.; Cybulski, M.; Madden, T.;
Priebe, W. D-Glucose- and d-Mannose-Based Antimetabolites. Part 2.
Facile Synthesis of 2-Deoxy-2-Halo-d-Glucoses and -d-Mannoses.
Carbohydr. Res. 2009,344, 1464−1473.
(35) Laatikainen, R.; Tiainen, M.; Korhonen, S.-P.; Niemitz, M.
Computerized Analysis of High-Resolution Solution-State Spectra.
eMagRes 2011,DOI: 10.1002/9780470034590.emrstm1226.
(36) Laatikainen, R.; Laatikainen, P.; Hakalehto, E. Quantitative
quantum mechanical nmr analysis: the superior tool for analysis of
biofluids. In Proceedings of The 1st International Electronic Conference
on Metabolomics; MDPI: Basel, Switzerland, 2016; p C005.
(37) Pauli, G. F.; Chen, S.-N.; Simmler, C.; Lankin, D. C.; Gödecke,
T.; Jaki, B. U.; Friesen, J. B.; McAlpine, J. B.; Napolitano, J. G.
Importance of Purity Evaluation and the Potential of Quantitative 1H
NMR as a Purity Assay. J. Med. Chem. 2014,57, 9220−9231.
(38) Tani, H.; Kurihara, H.; Hiroi, K.; Honda, N.; Yoshimoto, M.;
Kono, Y.; Murakami, R.; Kumita, S.; Arai, Y.; Itami, J. Correlation of
18F-BPA and 18F-FDG Uptake in Head and Neck Cancers.
Radiother. Oncol. 2014,113, 193−197.
(39) Wang, Y.-D.; Li, S.-J.; Liao, J.-X. Inhibition of Glucose
Transporter 1 (GLUT1) Chemosensitized Head and Neck Cancer
Cells to Cisplatin. Technol. Cancer Res. Treat. 2013,12, 525−535.
(40) Li, S.; Yang, X.; Wang, P.; Ran, X. The Effects of GLUT1 on
the Survival of Head and Neck Squamous Cell Carcinoma. Cell.
Physiol. Biochem. 2013,32, 624−634.
(41) Wisedchaisri, G.; Park, M.; Iadanza, M. G.; Zheng, H.; Gonen,
T. Proton-Coupled Sugar Transport in the Prototypical Major
Facilitator Superfamily Protein XylE. Nat. Commun. 2014,5, 4521.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30387

(42) Jiang, X.; Wu, J.; Ke, M.; Zhang, S.; Yuan, Y.; Lin, J. Y.; Yan, N.
Engineered XylE as a Tool for Mechanistic Investigation and Ligand
Discovery of the Glucose Transporters GLUTs. Cell Discovery 2019,
5, 14.
(43) Sun, L.; Zeng, X.; Yan, C.; Sun, X.; Gong, X.; Rao, Y.; Yan, N.
Crystal Structure of a Bacterial Homologue of Glucose Transporters
GLUT1-4. Nature 2012,490, 361−366.
(44) Park, M. S. Molecular Dynamics Simulations of the Human
Glucose Transporter GLUT1. PLoS One 2015,10, No. e0125361.
(45) Deng, D.; Sun, P.; Yan, C.; Ke, M.; Jiang, X.; Xiong, L.; Ren,
W.; Hirata, K.; Yamamoto, M.; Fan, S.; Yan, N. Molecular Basis of
Ligand Recognition and Transport by Glucose Transporters. Nature
2015,526, 391−396.
(46) Barth, R. F.; Coderre, J. A.; Vicente, M. G. H.; Blue, T. E.
Boron Neutron Capture Therapy of Cancer: Current Status and
Future Prospects. Clin. Cancer Res. 2005,11, 3987−4002.
(47) Pracht, P.; Bohle, F.; Grimme, S. Automated Exploration of the
Low-Energy Chemical Space with Fast Quantum Chemical Methods.
Phys. Chem. Chem. Phys. 2020,22, 7169−7192.
(48) Bayly, C. I.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A Well-
Behaved Electrostatic Potential Based Method Using Charge
Restraints for Deriving Atomic Charges: The RESP Model. J. Phys.
Chem. 1993,97, 10269−10280.
(49) Huey, R.; Morris, G. M.; Olson, A. J.; Goodsell, D. S. A
Semiempirical Free Energy Force Field with Charge-Based Desolva-
tion. J. Comput. Chem. 2007,28, 1145−1152.
(50) Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew,
R. K.; Goodsell, D. S.; Olson, A. J. AutoDock4 and AutoDockTools4:
Automated Docking with Selective Receptor Flexibility. J. Comput.
Chem. 2009,30, 2785−2791.
(51) Staroverov, V. N.; Scuseria, G. E.; Tao, J.; Perdew, J. P.
Comparative Assessment of a New Nonempirical Density Functional:
Molecules and Hydrogen-Bonded Complexes. J. Chem. Phys. 2003,
119, 12129−12137.
(52) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and
Accurate Ab Initio Parametrization of Density Functional Dispersion
Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010,
132, 154104.
(53) Becke, A. D.; Johnson, E. R. A Density-Functional Model of the
Dispersion Interaction. J. Chem. Phys. 2005,123, 154101.
(54) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence,
Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn:
Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005,7,
3297−3305.
(55) Ahlrichs, R.; Bär, M.; Häser, M.; Horn, H.; Kölmel, C.
Electronic Structure Calculations on Workstation Computers: The
Program System Turbomole. Chem. Phys. Lett. 1989,162, 165−169.
(56) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Auxiliary
Basis Sets for Main Row Atoms and Transition Metals and Their Use
to Approximate Coulomb Potentials. Theor. Chem. Acc. 1997,97,
119−124.
(57) Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma,
T. P.; van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T.
L.; de Jong, W. A. NWChem: A Comprehensive and Scalable Open-
Source Solution for Large Scale Molecular Simulations. Comput. Phys.
Commun. 2010,181, 1477−1489.
(58) Oda, A.; Ohta, K.; Endo, Y.; Fukuyoshi, S. Determination of
AMBER Force Field Parameters for Carborane Moiety Using
Quantum Chemical Calculations. Proceedings of the Symposium on
Chemoinformatics 2014, 2014, P02.
(59) Couto, M.; García, M. F.; Alamón, C.; Cabrera, M.; Cabral, P.;
Merlino, A.; Teixidor, F.; Cerecetto, H.; Vinas, C. Discovery of Potent
EGFR Inhibitors through the Incorporation of a 3D-Aromatic-Boron-
Rich-Cluster into the 4-Anilinoquinazoline Scaffold: Potential Drugs
for Glioma Treatment. Chem. −Eur. J. 2018,24, 3122−3126.
ACS Omega http://pubs.acs.org/journal/acsodf Article
https://doi.org/10.1021/acsomega.2c03646
ACS Omega 2022, 7, 30376−30388
30388