An investigation on 4-thiazolidinone derivatives as dual inhibitors of aldose reductase and protein tyrosine phosphatase 1B, in the search for potential agents for the treatment of type 2 diabetes mellitus and its complications

Article abstract of DOI:10.1016/j.bmcl.2018.10.024

Designed multiple ligands (DMLs), developed to modulate simultaneously a number of selected targets involved in etiopathogenetic mechanisms of a multifactorial disease, such as diabetes mellitus (DM), are considered a promising alternative to combinations of drugs, when monotherapy results to be unsatisfactory. In this work, compounds 1–17 were synthesized and in vitro evaluated as DMLs directed to aldose reductase (AR) and protein tyrosine phosphatase 1B (PTP1B), two key enzymes involved in different events which are critical for the onset and progression of type 2 DM and related pathologies. Out of the tested 4-thiazolidinone derivatives, compounds 12 and 16, which exhibited potent AR inhibitory effects along with interesting inhibition of PTP1B, can be assumed as lead compounds to further optimize and balance the dual inhibitory profile. Moreover, several structural portions were identified as features that could be useful to achieve simultaneous inhibition of both human AR and PTP1B through binding to non-catalytic regions of both target enzymes.

Full text of DOI:10.1016/j.bmcl.2018.10.024

Accepted Manuscript
An investigation on 4-thiazolidinone derivatives as dual inhibitors of aldose
reductase and protein tyrosine phosphatase 1B, in the search for potential agents
for the treatment of type 2 diabetes mellitus and its complications
Rosanna Maccari, Antonella Del Corso, Paolo Paoli, Ilenia Adornato, Giulia
Lori, Francesco Balestri, Mario Cappiello, Alexandra Naß, Gerhard Wolber,
Rosaria Ottanà
PII:
S0960-894X(18)30825-4
https://doi.org/10.1016/j.bmcl.2018.10.024
BMCL 26083
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
9 August 2018
8 October 2018
15 October 2018
Please cite this article as: Maccari, R., Del Corso, A., Paoli, P., Adornato, I., Lori, G., Balestri, F., Cappiello, M.,
Naß, A., Wolber, G., Ottanà, R., An investigation on 4-thiazolidinone derivatives as dual inhibitors of aldose
reductase and protein tyrosine phosphatase 1B, in the search for potential agents for the treatment of type 2 diabetes
mellitus and its complications, Bioorganic & Medicinal Chemistry Letters (2018), doi: https://doi.org/10.1016/
j.bmcl.2018.10.024
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

An investigation on 4-thiazolidinone derivatives as dual inhibitors of aldose reductase and
protein tyrosine phosphatase 1B, in the search for potential agents for the treatment of
type 2 diabetes mellitus and its complications
Rosanna Maccari,a* Antonella Del Corso,b Paolo Paoli,c Ilenia Adornato,a Giulia Lori,c
Francesco Balestri,b Mario Cappiello,b Alexandra Naß,d Gerhard Wolber,d Rosaria Ottanàa
a Department of Chemical, Biological, Pharmaceutical and Environmental Sciences, University
of Messina, Polo Universitario Annunziata, Viale SS. Annunziata, 98168 Messina, Italy
b Department of Biology, Biochemistry Unit, University of Pisa, Via S. Zeno, 51, 56123 Pisa,
Italy
c
Department of Scienze Biomediche Sperimentali e Cliniche, Sezione di Scienze Biochimiche,
University of Firenze, Viale Morgagni 50, 50134 Firenze, Italy
d
Institute of Pharmacy, Computer-Aided Molecular Design, Freie Universitaet Berlin,
Koenigin-Luisestr. 2+4, 14195 Berlin, Germany
* Corresponding author – email: rmaccari@unime.it
Abstract – Designed multiple ligands (DMLs), developed to modulate simultaneously a
number of selected targets involved in etiopathogenetic mechanisms of a multifactorial disease,
such as diabetes mellitus (DM), are considered a promising alternative to combinations of
drugs, when monotherapy results to be unsatisfactory. In this work, compounds 1-17 were
synthesized and in vitro evaluated as DMLs directed to aldose reductase (AR) and protein
tyrosine phosphatase 1B (PTP1B), two key enzymes involved in different events which are
critical for the onset and progression of type 2 DM and related pathologies. Out of the tested 4-
thiazolidinone derivatives, compounds 12 and 16, which exhibited potent AR inhibitory effects
along with interesting inhibition of PTP1B, can be assumed as lead compounds to further
optimize and balance the dual inhibitory profile. Moreover, several structural portions were
identified as features that could be useful to achieve simultaneous inhibition of both human AR
and PTP1B through binding to non-catalytic regions of both target enzymes.
Keywords: diabetes mellitus; designed multiple ligands; 4-thiazolidinone derivatives; aldose
reductase; protein tyrosine phosphatase 1B.
1

Diabetes mellitus (DM) is a multifactorial chronic disease characterized by hyperglycaemia and
metabolic dysfunctions. According to WHO reports, currently DM affects more than 420
million people worldwide, being responsible for about 1.5 million deaths every year.1 The
number of people affected by DM is expected to rise to 592 million by 2035, but it might be
underestimated since, in the last years, previous estimates have been surpassed.2 The
prevalence of DM is growing, in both industrialized and developing countries, mainly as a
consequence of the dramatic incidence of type 2 DM (T2DM) which accounts for more than
90% cases of diabetes worldwide. Obesity, overweight, unbalanced diet and unhealthy
lifestyles are the main factors associated with increased risk of T2DM.
In diabetes, hyperglycaemia is a critical condition that originates from insulin deficiency
(particularly in type 1 DM) and/or reduced sensitivity of target tissues to the hormone (insulin-
resistance, which is a typical feature of T2DM). Hyperglycaemia triggers a number of
mechanisms and abnormal cellular signalling, including inflammatory response, increased
oxidative stress, lipid peroxidation, altered osmotic balance. On the whole, these
hyperglycaemia-induced cellular dysfunctions are responsible for vascular and nervous lesions
which underlie the development of serious chronic complications associated to DM, such as
neuropathy, nephropathy, retinopathy, atherosclerosis, increased risk of cardiovascular
pathologies.3
Despite the availability of different effective therapeutic agents, in diabetic patients the control
of hyperglycaemia and the prevention of chronic complications by drug monotherapy remain
often unsatisfactory. Therefore, an adequate therapeutic management of T2DM is often
achieved by means of combinations of drugs acting with different and generally
complementary mechanisms of action.
However, combination therapy may be associated with unwanted drug-drug interactions,
inappropriate pharmacokinetics, toxicity and scarce patient compliance. In order to overcome
the problems related to the polypharmacological approach, the development of “designed
multiple ligands” (DMLs) has been recently proposed as a promising strategy which could
provide novel drug candidates for the treatment of multifactorial diseases, such as T2DM.
DMLs are compounds that have been rationally designed to modulate two or more specific
targets which are involved in the etiopathology of a disease, thus possibly resulting in improved
efficacy and minor risks compared to both monotherapy and drug combinations.4-8
The design of DMLs as potential drugs poses substantial challenges to medicinal chemists,
particularly: i) the selection of suitable biological targets, which must be well-characterized and
share common structural features, ii) the optimization of the molecular properties in order to
obtain drug-like candidates, endowed with an appropriate balance between their multiple
2

actions. Although DMLs are often low-affinity ligands towards the selected targets, it has been
proposed that the even partial inhibition of selected multiple enzymes can result in effective
multi-target drugs.4,6
In the multifactorial etiopathology of T2DM as well as in the onset and progression of diabetic
complications several druggable targets are involved.
Among them, protein tyrosine phosphatase 1B (PTP1B, E.C. 3.1.3.48) is an enzyme crucially
implicated in the development of insulin-resistance, which is a characteristic condition in both
T2DM and obesity. This enzyme functions as a negative regulator of insulin action, by
dephosphorylating specific residues of phosphotyrosine (pTyr) of the activated insulin receptor
and thus interrupting the signalling pathways mediated by the hormone. PTP1B also
downregulates the signal of leptin, an adipocyte-derived hormone that controls food intake and
increases energy expenditure.9,10 PTP1B overexpression is strictly related to insulin-resistance
and it has been demonstrated that the inhibition or genetic ablation of this phosphatase can
improve glucose homeostasis, cellular sensitivity to both insulin and leptin, and resistance to
diet-induced obesity, without inducing hypoglycaemia or toxic effects.9-15
In the complex network of hyperglycaemia-induced dysfunctions that underlie the development
of long-term diabetic complications, the enzyme aldose reductase (AR, E.C. 1.1.1.21) plays
pivotal roles. Under hyperglycaemic conditions, this aldo-keto reductase catalyses the NADPH-
dependent reduction of an excess amount of glucose to sorbitol, which in turn is oxidized to
fructose by sorbitol dehydrogenase, in the polyol pathway. The increased consumption of
glucose through this pathway leads to osmotic and redox imbalances, increased oxidative
stress, and protein alterations.3,16,17 Moreover, AR can metabolize glutathione-conjugates of
reactive unsaturated aldehydes which originate from oxidative stress-induced lipid
peroxidation. These reactions represent a step in the overall detoxification pathway of toxic
aldehydes, that starts with the glutathione S-transferase-catalyzed adduct formation.18
However, 3-glutathionyl-1,4-dihydroxynonane, generated by the reductive action of AR on 3-
glutathionyl-4-hydroxynonanal, is a pro-inflammatory molecule which can trigger
inflammatory signalling, thus contributing significantly to tissue and vascular damage.17,19 It is
well-documented that the inhibition of AR can prevent or slow down the development of
chronic diabetic complications (particularly neuropathies, kidney failure, blindness,
atherosclerosis, and cardiovascular diseases) as well as can control DM-associated subclinical
tissue inflammation. On this basis, AR is considered an emerging target for therapeutic
interventions in different pathologies associated with increased inflammatory response, such as
diabetes and its complications.17,19-21
3

Because of their functional features, both PTP1B and AR can be assumed as molecular targets
of dual inhibitors designed as novel molecules capable to counteract simultaneously two
different coexisting pathogenic mechanisms which are crucial for the onset and progression of
T2DM and its chronic complications. To our knowledge, so far only a series of potential dual
AR/PTP1B inhibitors were designed, starting from a number of flavonoids and
naphthoquinones.22
Herein, starting from our structure-activity relationship (SAR) studies performed on numerous
4-thiazolidinone derivatives active as AR or PTP1B inhibitors,23-36 we report the evaluation of a
series of (5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)acetic acids and 2-oxo/phenylimino
analogues (1-17) (Table 1) as dual AR/PTP1B inhibitors. Our previous SAR studies23-36
highlighted that the pharmacophores of these inhibitors of AR and PTP1B share several
structural features that make possible the design of potential dual inhibitors: i) a polar portion,
including an acidic moiety or H-bond acceptor groups, which can strongly interact with the
positively charged region of the catalytic sites of both enzymes; ii) a lipophilic moiety,
generally containing an aromatic system, which can bind hydrophobic amino acid residues
lining the lipophilic specificity pocket of AR (such as Trp111, Thr113, Phe122, Ala299,
Leu300, Ser302) and the phosphate-binding secondary non-catalytic pocket of PTP1B (such as
Arg24, Ala27, Arg254, Met258, Gly259). It is worth pointing out that the 4-thiazolidinone
scaffold appears to be capable to maintain a proper orientation of these crucial portions to
effectively bind each enzyme and, at the same time, can establish useful interactions which may
contribute to stabilize the complex enzyme/inhibitor.23-36
Therefore, starting from a “knowledge-based” approach, we merged pharmacophoric elements
of inhibitors directed to each of the selected enzymes, maintaining the shared 4-thiazolidinone
core. In particular, we decided to insert on the N-3 of the thiazolidinone scaffold an acetic
chain, which had been shown to be critical to achieve strong AR inhibition23-30 and might also
be suitable to interact with the positively charged catalytic centre of PTP1B. Different
substituents were inserted in the 5-arylidene portion in order to modulate the inhibitory activity.
Moreover, it must be taken into consideration that the substitution pattern of the 5-arylidene
portion may influence the capability of inhibitors to interact with allosteric regions on the
enzyme surface, thus leading to different mechanisms of inhibition, as observed for 4-[(5-
arylidene-4-oxo-2-thioxothiazolidin-3-yl)methy]benzoic acids that we have recently reported as
allosteric PTP1B inhibitors.36
Previously, we had evaluated (5-arylidene-4-oxo-2-thioxothiazolidn-3-yl)acetic acids 1-9 as
AR inhibitors (ARIs) of the enzyme extracted from bovine lens.29,30 In this work, all
compounds 1-17 were tested against human recombinant AR and human recombinant PTP1B.
4

(5-Arylidene-4-oxo-2-thioxothiazolidin-3-yl)acetic acids (1-13, Table 1) were obtained in high
yields by the Knoevenagel condensation of (4-oxo-2-thioxothiazolidin-3-yl)acetic acid with
appropriate aldehydes, in refluxing glacial acetic acid and sodium acetate (Scheme 1).
2,4-Thiazolidinediones 14-16 (Table 1) were synthesised starting from the Knoevenagel
condensation of commercial 2,4-thiazolidinedione with the corresponding aldehydes, in
refluxing ethanol and in the presence of piperidine as a base, followed by N-alkylation with
bromoacetic acid (Scheme 2).
The synthesis of 2-phenylimino analogue 17 was performed according to a multistep procedure
starting from the reaction of phenyl isothiocyanate with glycine. The subsequent reaction
between (3-phenylthioureido)acetic acid and chloroacetyl chloride in refluxing ethanol
provided (4-oxo-2-phenyliminothiazolidin-3-yl)acetic acid, which was condensed with 3-
methoxy-4-benzyloxybenzaldehyde to obtain compound 17 (Scheme 3).
1
The structures of all compounds were unambiguously assigned by means of analytical and H
and 13C NMR spectroscopy data. NMR spectra highlighted that compounds 1-17 were obtained
only as Z isomers, in analogy to previously investigated 5-arylidene-4-thiazolidinone
derivatives which had been characterized by means of X-ray crystallography.23,37
Scheme 1
5

Scheme 2
6

Scheme 3
7

The in vitro inhibitory activity of compounds 1-17 was assessed both against the human
recombinant AR, by using L-idose as substrate38,39 and Epalrestat as reference drug, and against
human recombinant PTP1B, by using p-nitrophenyl phosphate as substrate and sodium
metavanadate as reference drug. Table 1 reports the IC50 values of the tested compounds for the
two target enzymes.
Concerning to AR inhibition, all tested 4-thiazolidinone derivatives exhibited excellent
inhibitory effects against the human enzyme, with IC50 values in the range 0.025 M-1.41 M
(Table 1), thus resulting in some cases more efficient than Epalrestat (IC50 = 0.102 M).
Moreover, except for compound 17, the IC50 values resulted of the same order of magnitude as
the enzyme concentration in the assay (67 nM), thus leading these compounds to be considered
as “tight binding inhibitors”. Relatively to compounds 1-9, these results appeared well
comparable with the inhibitory ability displayed against bovine lens AR, using D,L-
glyceraldehyde as substrate.29,30 Also in that case, in fact, taking into account of the higher
concentration of the enzyme in the assay (173 nM),29,30 due to the lower specific activity of the
bovine lens enzyme with respect to that of human AR, these powerful ARIs appeared to act as
“tight binding inhibitors”.
In agreement with previously observed SARs,29,30 (5-arylidene-4-oxo-2-thioxothiazolidin-3-
yl)acetic acids 11-13 were shown to be from 3- to almost 10-fold more potent inhibitors of
human AR than corresponding 2,4-thiazolidinediones 14-16 (Table 1). On the other hand, the
replacement of the thiocarbonyl or carbonyl group in position 2 of the thiazolidinone scaffold
of compounds 13 and 15 with a 2-phenylimino moiety (derivative 17, IC50 = 1.41 M)
produced a marked reduction of the AR inhibitory effect, leading to an IC50 value which was
35-fold and 4-fold higher than those of analogues 13 and 15, respectively (Table 1).
Among 2-thioxo-4-thiazolidinones 1-13 and 2,4-thiazolidinediones 14-16, the influence of the
5-arylidene moiety on the AR inhibitory effectiveness appeared moderate, since their IC50
values were included in a rather narrow submicromolar range. However, the presence of a
substituent in the meta position of the 5-benzylidene ring was generally more beneficial for the
inhibition of the target enzyme than the para-substitution (Table 1).
The inhibitory effectiveness of compounds 1-17 towards human PTP1B was significantly lower
when compared to AR inhibition (Table 1). However, the most potent PTP1B inhibitors 16 and
17 exhibited appreciable effectiveness, with IC50 values of 12.4 M and 10.6 M, respectively;
less effective compounds 1-4 and 11-14 displayed IC50 values ranging from 32.5 M
(compound 12) to 86.8 M (compound 2), whereas compounds 6-10 and 15 displayed scarce
activity, with IC50 values higher than 100 M (Table 1).
8

Table 1 – Inhibitory activities of compounds 1-17 against human PTP1B and human AR,
expressed as IC50 (M).
Compd.
X
Ar
PTP1B
AR
IC50 (M) a
IC50 (M) a
S
S
4-OC6H5-C6H4
3-OC6H5-C6H4
63.9 ± 2.0
86.8 ± 3.2
56.0 ± 1.0
43.1 ± 1.5
n.d.
0.060 ± 0.004
0.025 ± 0.002
0.052 ± 0.003
0.053 ± 0.004
0.078 ± 0.007
0.194 ± 0.011
0.064 ± 0.005
0.228 ± 0.012
0.139 ± 0.011
0.125 ± 0.009
0.104 ± 0.010
0.056 ± 0.006
0.040 ± 0.004
0.364 ± 0.049
0.323 ±0.033
0.276 ± 0.029
1.41 ± 0.126
0.102 ± 0.005
1
2
S
4-OCH2C6H5-C6H4
3
S
3-OCH2C6H5-C6H4
4
S
3-OCH3-C6H4
5
S
4-OCH2CONH2-C6H4
3-OCH2CONH2-C6H4
3-OCH3,4-OCH2CONH2-C6H3
4-OCH3,3-OCH2CONH2-C6H3
C6H5CH=CH
10 % b
6
S
716.5 ± 136
378 ± 23
679 ± 184
171 ± 27
49.4 ± 1.0
32.5 ± 1.1
54.1 ± 1.0
63.8 ± 2.6
151 ± 51
12.4 ± 0.8
10.6 ± 0.4
7
S
8
S
9
S
10
S
4-O(CH2)2C6H5-C6H4
3-O(CH2)2C6H5-C6H4
3-OCH3,4-OCH2C6H5-C6H3
4-O(CH2)2C6H5-C6H4
3-OCH3,4-OCH2C6H5-C6H3
3-O(CH2)2C6H5-C6H4
3-OCH3,4-OCH2C6H5-C6H3
11
S
12
S
13
O
O
O
=NPh
14
15
16
17
Epalrestat
Vanadate
0.4 ± 0.01
a Values are expressed as the mean ± S.E.M (see methods for details).
b Percent inhibition in the presence of compound 6 at the concentration of 300 M.
n.d. = not determined.
9

The presence of a more extended aromatic substituent on the 5-benzylidene ring was generally
related to enhanced inhibitory potency, whereas smaller and/or more polar substituents (such as
in compounds 6-10) were detrimental for PTP1B inhibition. Interestingly, as observed for AR
inhibition, the substitution on the meta position of the 5-benzylidene ring was generally more
beneficial for the PTP1B inhibitory effectiveness than the substitution in the para position
(compound 4 vs. 3, 12 vs. 11 and 16 vs. 14). This latter SAR appears to be in analogy with
SARs observed for certain 4-[(5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)methyl]benzoic acids
active as mixed non-competitive PTP1B inhibitors that we recently reported36; conversely, in
the case of 4-[(5-arylidene-2-arylimino/oxo-4-oxothiazolidin-3-yl)methyl]benzoic acid
derivatives that are active as competitive PTP1B inhibitors, the para-substitution on 5-
benzylidene ring was found to be more beneficial than the meta-substitution by providing
higher inhibition levels.31-36
It is worth noting that the effect of the substituent in position 2 of the 4-thiazolidinone scaffold
in modulating the PTP1B inhibitory activity could not be clearly defined, unlike what was
observed for AR inhibitory effects. For instance, 2,4-thiazolidinedione 16 was about 3-fold
more potent than its 2-thioxo analogue 12; conversely, 2-thioxo-4-thiazolidinone derivative 13
was about 3-fold more active than its 2,4-thiazolidinedione counterpart 15. In addition, the
replacement of the thiocarbonyl (compounds 13) or carbonyl group (compound 15) with a
phenylimino moiety (compound 17) provided an appreciable gain in potency toward PTP1B,
with a 5-fold and 14-fold decrease of the IC50 value, respectively.
Considering that, among the tested 4-thiazolidinones, compounds 12 and 16 displayed an
appreciable dual action in inhibiting both AR and PTP1B, they were further characterized for
their kinetic features on both target enzymes.
The kinetic characterization of these molecules had to take into account the remarkable
difference in their inhibitory potencies toward AR and PTP1B, thus requiring different
analytical approaches.
Firstly, additional tests were performed in order to verify if selected compounds 12 and 16 were
reversible or irreversible inhibitors. For PTP1B, appropriate aliquots of the enzyme were
incubated in the presence of saturating concentration of each inhibitor for 60 minutes at 37°C.
Then, the enzyme mixtures were diluted with assay buffer, and the residual enzyme activity
was determined. It was found that the recovery of enzyme activity was complete, suggesting
that compounds 12 and 16 behave as reversible PTP1B inhibitors. In the case of AR, being the
compounds tight binding inhibitors, active at concentrations comparable to those of the
enzyme, it was necessary to perform an extensive dialysis of the enzyme-inhibitor mixture. The
10

recovery of approximately 70 % of AR activity from a fully inhibited enzyme clearly indicated
that both compounds 12 and 16 also behaved as reversible inhibitors toward AR.
In order to determine the action mechanism of these reversible inhibitors, in the case of AR, the
effect of the inhibitors was evaluated at different substrate concentrations (Figures 1A and 2A).
When fitted by a nonlinear regression analysis to the Morrison equation (see eq. 2
Supplementary material), the apparent inhibition constants were determined. The fitting
by nonlinear regression analysis of
as a function of substrate concentration (Figures 1B
and 2B) into a general equation of tight binding non-competitive inhibition model (see eq. 3
Supplementary material) allowed the evaluation of the true inhibition constants and .
Thus, while values (asymptote/ordinate intercept) could be estimated being higher than 0.3
µM and 0.5 µM for compounds 12 and 16, respectively, at least fifty-fold lower values for
(abscissa asymptote) were determined (Table 2). This suggests that both the analysed
compounds can be considered as uncompetitive inhibitors.
In the case of PTP1B, reaction rate values at different inhibitor concentrations were analysed
by double reciprocal plots. Both compounds 12 and 16 displayed a pure non-competitive
mechanism of action (Figures 3A and 4A, respectively). A unique value of dissociation
constant for both EI and ESI emerged for each inhibitor (12.0 ± 1.2 µM and 13.1 ± 2.8 µM, for
compounds 12 and 16, respectively), when secondary plots of KM/Vmax versus the inhibitor
concentration were analysed (Figures 3B and 4B, respectively).
Being compound 17 the most active inhibitor for PTP1B, its mechanism of action was also
analysed. This compound, as occurred for compounds 12 and 16, also acted as reversible
inhibitor. The reaction rate measurements at different substrate concentrations and at
concentrations of compound 17 (Figure 5) revealed for this inhibitor a mixed non-competitive
mechanism of action. The Ki and K’i dissociation constants, evaluated from secondary plots (as
above) were 7.6 ± 0.5 µM and 40.6 ± 6.3 µM, respectively, suggesting a preferential binding of
compound 17 for the free enzyme.
11

Figure 1. Kinetic characterization of compound 12 as AR inhibitor. A) The activity of the
purified enzyme (7 mU), expressed as mM/min, was measured at the indicated concentrations
of the inhibitor in the presence of the following L-idose concentrations: (▲) 2.1 mM, (●) 3.1
mM, (▼) 4.1 mM, (■) 8.3 mM, (♦) 16.6 mM. Curves were plotted by non linear regression
analysis fitting the experimental data to Morrison equation (see Eq. 2 Supplementary material).
B) The apparent inhibition constants, Kiapp determined from Panel A for each substrate
concentration were plotted against substrate concentration and fitted by non linear regression
analysis to the equation 3 (see Supplementary material) relative to a general case of tight
binding non-competitive inhibition model.
12

Figure 2. Kinetic characterization of compound 16 as AR inhibitor. A) The activity of the
purified enzyme (7 mU), expressed as mM/min, was measured at the indicated concentrations
of the inhibitor in the presence of the following L-idose concentrations: (▲) 2.1 mM, (●) 3.1
mM, (▼) 4.1 mM, (■) 8.3 mM, (♦) 16.6 mM. Curves were plotted by non linear regression
analysis fitting the experimental data to Morrison equation (see Eq. 2 Supplementary material).
B) The apparent inhibition constants, Kiapp determined from Panel A for each substrate
concentration were plotted against substrate concentration and fitted by non linear regression
analysis to the equation 3 (see Supplementary material) relative to a general case of tight
binding non-competitive inhibition model.
13

Table 2 – Kinetic characterization of compounds 12 and 16 as dual AR/PTP1B inhibitors
Compound 12
Inhibition model
Ki (M)
Ki’(M)
Uncompetitive
> 0.3
0.0062
(0.0056– 0.0069)a
12.0 ± 1.2
AR
Non-competitive
12.0 ± 1.2
PTP1B
Compound 16
Ki (M)
Inhibition model
Ki’(M)
Uncompetitive
> 0.5
0.016
AR
(0.013 – 0.018)a
13.1 ± 3.9
Non-competitive
13.1 ± 3.9
PTP1B
a 95% confidence limit
The pure or mixed non-competitive mechanism of PTP1B inhibition produced by compounds
12, 16 and 17 appears to be consistent with the absence of the residue of the 4-methylbenzoic
acid that, in our previously reported 4-[(5-arylidene-4-oxothiazolidin-3-yl)methy]benzoic acid
derivatives,31-36 was generally shown to act as a phosphotyrosine-mimetic group with an
important role in the anchoring of these inhibitors into the PTP1B catalytic site. Considering
that preliminary assays (data not shown) had indicated that the 4-methylbenzoic acid residue
was detrimental for AR inhibitory effect, probably because it is not optimal to fit into the rigid
polar subsite of the AR catalytic pocket, it was replaced by the residue of acetic acid in
compounds 1-17. This led to PTP1B inhibitors, such as compounds 12, 16 and 17, that were
shown to be capable to bind an allosteric region rather than the catalytic site of the target
enzyme, as indicated by kinetic and in silico docking studies (see below).
14

20
15
10
5
A
-2
-1
0
1
2
1/[pNPP] mM-1
8
6
4
2
0
B
-20
-10
0
10
20
30
[Compound 12] M
Figure 3. Kinetic analysis of compounds 12 as PTP1B inhibitor. A) Double reciprocal plots;
the concentrations of compound 12 are: , 0 M, , 10 M, , 20 M, , 30 M. Data
reported in the figures represent the mean values  S.E.M. (n = 3). B) Secondary plots:
Km/Vmax versus inhibitor concentration.
15

16
12
8
A
4
-0.5
0.0
0.5
1.0
1.5
2.0
1/[pNPP] mM-1
6
5
4
3
2
1
B
-15 -12
-9
-6
-3
0
3
6
9
12
Compound 16 (M)
Figure 4. Kinetic analysis of compound 16 as PTP1B inhibitor. A) Double reciprocal plots of
compound 16; the concentrations of compound 16 are: , 0 M, , 3 M, , 6 M, , 9
M. Data reported in the figures represent the mean values  S.E.M. (n = 3). B) Secondary
plots: Km/Vmax versus inhibitor concentration.
16

14
12
10
8
A
A
6
4
2
-0.5
0.0
0.5
1.0
1.5
2.0
1/[pNPP] mM-1
B
1,2
0,9
0,6
0,3
-50
-40
-30
-20
-10
0
10
20
Compound 17 (M)
6
5
4
3
2
1
C
-8
-6
-4
-2
0
2
4
6
8
Compound 17 (M)
Figure 5. Kinetic analysis of compounds 17 as PTP1B inhibitor. A) Double reciprocal plots of
compound 17; the concentrations of compound 17 are: , 0 M, , 2 M, , 4 M, , 6
M. Data reported in the figures represent the mean values  S.E.M. (n = 3). B) and C)
Secondary plots: 1/Vmax, and Km/Vmax versus inhibitor concentration, respectively.
17

Docking studies were performed with compounds 12 and 16 in order to discover interaction
features which could be useful to achieve simultaneous inhibition of both AR and PTP1B
through binding to a region other than the catalytic site of the target enzymes.
The selected compounds were docked to the whole protein surface of PTP1B. In analogy to
previously published 4-thiazolidinone PTP1B inhibitors,36 compounds 12 and 16 were shown
to be able to fit a site connected to the catalytic loop via a -strand (Figure 6). Figure 7 depicts
the selected pose of compound 16 bound to this allosteric site. The terminal aromatic moiety of
compound 16 was found to be anchored in the lipophilic pocket of the allosteric region
surrounded by residues Pro206, Arg79 and Ser80. The meta position of the phenylethoxy
moiety allows the carboxylic group of the ligand to stretch out towards Arg105, Arg169 and
Lys103, which are distal to the PTP1B catalytic site, enabling several ionic and hydrogen
bonding contacts (Figure 7).
Figure 6. Representation of the second PTP1B binding site (yellow ligand shape) distal to the
catalytic cavity (catalytic loop in pink), but directly connected to it by a single beta strand
(violet).
18

Figure 7. Selected pose of compound 16 bound to a non-catalytic binding pocket of PTP1B.
Top: 3D depiction with protein surface coloured by hydrophilicity(cyan)/lipophilicity(yellow);
inhibitor in white ball-and-stick depiction. Bottom: 2D depiction with protein-ligand interaction
features: red circles, hydrogen bond acceptor; red stars, negative ionic interaction; yellow,
hydrophobic interaction.
19

Protein-ligand docking of compounds 12 and 16 was performed to the AR-idose complex, since
kinetic studies revealed that both compounds behave as uncompetitive AR inhibitors. So far no
allosteric sites have been published for AR; however some crystal structures derived with
different soaking conditions and ligand concentrations resulted in complexes with several
ligands on top of each other filling the pocket connected to the catalytic site (2FZB).40 Since no
crystal structure of AR was available with idose or the corresponding alcohol, a crystal
structure of AR with glyceraldehyde (3V36)41 was modified by inserting idose instead of the
former substrate.
Consistent poses were found for both compounds 12 and 16 (Figure 8). Both molecules are
located in the outer part of the catalytic pocket on top of the substrate idose, which is bound to
the catalytic site. The 2-phenylethoxy moiety is oriented towards idose and the catalytic
residues, by filling a hydrophobic pocket above them with the possibility of hydrophobic
interactions to Trp20 as well as Phe122, Trp79 and Val47. The hydrophilic part of the
molecules is pointed towards the solvent with ionic and hydrogen bonding interactions of the
carboxylic group with Lys221 and Arg217. The 4-carbonyl group of the thiazolidinone core
shows hydrogen bonding interactions with the backbone of Ala299. The carbonyl (or
thiocarbonyl) group in position 2 is faced towards the solvent, but surrounded by lipophilic
sidechains of the protein, especially of Leu301. The described environment leads to a better fit
of the bulkier 2-thioxo-4-thiazolidinone 12 compared to 2,4-thiazolidindione 16, in agreement
with the higher AR inhibitory potency of compound 12 compared to compound 16. Leu301
also interacts with the adjacent 5-benzylidene ring of the ligands. This phenyl ring is also close
to Trp219 with a feasible conformation that allows not only hydrophobic but also -stacking
interactions. Additionally, the bridging ether oxygen of the ligands could interact with Ser302
by means of weak hydrogen bonding interactions.
The described ligand binding poses are possible only for derivatives with meta-substituted 5-
benzyidene ring, since they require a bent ligand arrangement in order to simultaneously fit the
hydrophobic pocket above the substrate and establish ionic and hydrogen bonding interactions
to Lys221 and Arg217. For an optimal fit of the hydrophobic phenyl moiety to the hydrophobic
cavity a certain length of the spacer between the phenyl rings is also required and a spacer
length between three and four atoms appeared to be the most suitable for this purpose.
20

Figure 8. Selected pose of compound 16 bound to the AR-idose complex. Top: 3D depiction
with protein surface coloured by hydrophilicity(cyan)/lipophilicity(yellow); inhibitor in white,
substrate in dark grey. Bottom: 2D depiction with protein-ligand interaction features: red
circles, hydrogen bond acceptor; red stars, negative ionic interaction; yellow, hydrophobic
interaction.
21

Based on the conducted docking and SAR investigations, a main goal of this study was that
several ligand features were identified as important for simultaneous inhibiting both AR and
PTP1B enzymes through binding to non-catalytic regions. Firstly, both allosteric binding sites
require a bent ligand arrangement that is enabled by the 3-(2-phenylethoxy)benzylidene moiety
(in compounds 12 and 16) relative to the thiazolidinone core. Secondly, the central part of the
ligand, characterized in these compounds by the 5-benzylidene ring, needs to be relatively
lipophilic, but even more the distal part of the ligand, which is buried in the lipophilic pocket in
PTP1B and positioned in a lipophilic cleft on top of the substrate in AR. Furthermore, our
studies revealed that this terminal portion shows optimal fit to both binding sites of the two
target enzymes for a distance of approximately 8 Å to C5 of the central phenyl ring of the
ligand. Additionally, to ensure an effective binding, an acidic moiety is highly beneficial,
which should be able to bent out of the plane defined by the rest of the binding site to reach
basic amino acid residues present in both allosteric binding sites.
Our findings highlight that the chance of exploiting allosteric regions to effectively inhibit both
human AR and PTP1B offers promising opportunities for the design of new dual inhibitors of
these target enzymes.
Declarations of interest
None.
Acnowledgements
This work was supported by University of Messina, University of Florence and University of
Pisa and by a fund to P.P. from Italian Ministry of University and Research (Miur -
Finanziamento annuale individuale delle attività base di ricerca 2018).
The Authors thank Prof. Umberto Mura for critical reading of the enzymological section.
22

References
[1] http://www.who.int/diabetes
[2] Guariguata L, Whiting DR, Hambleton I, Beagley J, Linnenkamp U, Shaw JE. Global
estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res. Clin.
Pract. 2014; 103: 137-149.
[3] Brownlee M. Biochemistry and molecular cell biology of diabetic complications. Nature
2001; 414: 813-820.
[4] Csermely P, Agoston V, Pongor S. The efficiency of multi-target drugs: the network
approach might help drug design. Trends Pharmacol. Sci. 2005; 26: 178-182.
[5] Morphy R, Rankovic Z. Designed Multiple Ligands. An Emerging Drug Discovery
Paradigm. J. Med. Chem. 2005; 48: 6523-6543.
[6] Hopkins A. Network pharmacology: the next paradigm in drug discovery. Nature Chem.
Biol. 2008; 4: 682-690.
[7] Gattrell W, Johnstone C, Patel S, Sambrook Smith C, Scheel A, Schindler M. Designed
multiple ligands in metabolic disease research: from concept to platform. Drug Discov.
Today 2013; 18: 692-696.
[8] Costantino L, Barlocco D. Designing approaches to multitarget drugs. In: Handler H,
Buschmann H, eds. Drug selectivity: an evolving concept in medicinal chemistry. Wiley-
VCH, Weinheim, 2018.
[9] Zabolotny JM, Bence-Hanulec KK, Stricker-Krongrad A, Haj F, Wang Y, Minokoshi Y,
Kim YB, Elmquist JK, Tartaglia LA, Kahn BB, Neel BG. PTP1B regulates leptin signal
transduction in vivo. Dev. Cell 2002; 2: 489-495.
[10] Cheng A, Uetani N, Simoncic PD, Chaubey VP, Lee-Loy A, McGlade CJ, Kennedy BP,
Tremblay ML. Attenuation of leptin action and regulation of obesity by protein tyrosine
phosphatase 1B. Dev. Cell 2002; 2: 497-503.
[11] Elchebly M, Payette P, Michaliszyn E, Cromlish W, Collins S, Loy AL, Normandin D,
Cheng A, Himms-Hagen J, Chan CC, Ramachandran C, Gresser MJ, Tremblay ML,
Kennedy BP. Increased insulin sensitivity and obesity resistance in mice lacking the
protein tyrosine phosphatase-1B gene. Science 1999; 283: 1544-1548.
[12] Klaman LD, Boss O, Peroni OD, Kim JK, Martino JL, Zabolotny JM, Moghal N, Lubkin
M, Kim YB, Sharpe AH, Stricker-Krongrad A, Shulman GI, Neel BG, Kahn BB.
Increased energy expenditure, decreased adiposity, and tissue-specific insulin sensitivity
in protein-tyrosine phosphatase 1B–deficient mice. Mol. Cell. Biol. 2000; 20: 5479-5489.
[13] Di Paola R, Frittitta L, Miscio G, Bozzali M, Baratta R, Centra M, Spampinato D,
Santagati MG, Ercolino T, Cisternino C, Soccio T, Mastroianno S, Tassi V, Almgren P,
23

Pizzuti A, Vigneri R, Trischitta V. A variation in 3’ UTR of hPTP1B increases specific
gene expression and associates with insulin resistance. Am. J. Hum. Genet. 2002; 70:
806-812.
[14] Bence KK, Delibegovic M, Xue B, Gorgun CZ, Hotamisligil GS, Neel BG, Kahn BB.
Neuronal PTP1B regulates body weight, adiposity and leptin action. Nat. Med. 2006; 12:
917-924.
[15] Zhang ZY, Dodd GT, Tiganis T. Protein tyrosine phosphatases in hypothalamic insulin
and leptin signalling. Trends Pharmacol. Sci. 2015; 36: 661-674.
[16] Srivastava SK, Ramana KV, Bhatnagar A. Role of aldose reductase and oxidative
damage in diabetes and the consequent potential for therapeutic options. Endocr. Rev.
2005; 26: 380-392.
[17] Maccari R, Ottanà R. Targeting aldose reductase for the treatment of diabetes
complications and inflammatory diseases: new insights and future directions. J. Med.
Chem. 2015; 58: 2047-2067.
[18] Hubatsch I, Ridderström M, Mannervik B. Human glutathione transferase A4-4: an alpha
class enzyme with high catalytic efficiency in the conjugation of 4-hydroxynonenal and
other genotoxic products of lipid peroxidation. Biochem. J. 1998; 330: 175-179.
[19] Frohnert BI, Long EK, Hahn WS, Bernlohr DA. Glutathionylated lipid aldehydes are
products of adipocyte oxidative stress and activators of macrophage inflammation.
Diabetes 2014; 63: 89-100.
[20] Ramana KV, Srivastava SK. Aldose reductase: a novel therapeutic target for
inflammatory pathologies. Int. J. Biochem. Cell Biol. 2010; 42: 17-20.
[21] Chatzopoulou M, Pegklidou K, Papastavrou N, Demopoulos VJ. Development of aldose
reductase inhibitors for the treatment of inflammatory disorders. Expert Opin. Drug
Discov. 2013; 8: 1365-1380.
[22] Vyas B, Silakari O, Kaur M, Singh B. Integrated pharmacophore and docking-based
designing of dual inhibitors of aldose reductase (ALR2) and protein tyrosine phosphatase
(PTP1B) as novel therapeutics for insulin-resistant diabetes and its complications. J.
Chemometrics 2015; 29: 109-125.
[23] Bruno G, Costantino L, Curinga C, Maccari R, Monforte F, Nicolò F, Ottanà R, Vigorita
MG. Synthesis and aldose reductase inhibitory activity of 5-arylidene-2,4-
thiazolidinediones. Bioorg. Med. Chem. 2002; 10: 1077-1084.
[24] Maccari R, Ottanà R, Curinga C, Vigorita MG, Rakowitz D, Steindl T, Langer T.
Structure–activity relationships and molecular modelling of 5-arylidene-2,4-
24

thiazolidinediones active as aldose reductase inhibitors. Bioorg. Med. Chem. 2005; 13:
2809-2823.
[25] Maccari R, Ottanà R, Ciurleo R, Vigorita MG, Rakowitz D, Steindl T, Langer T.
Evaluation of in vitro aldose redutase inhibitory activity of 5-arylidene-2,4-
thiazolidinediones. Bioorg. Med. Chem. Lett. 2007; 17: 3886-3893.
[26] Maccari R, Ottanà R, Ciurleo R, Rakowitz D, Matuszczak B, Laggner C, Langer T.
Synthesis, induced-fit docking investigations, and in vitro aldose reductase inhibitory
activity of non-carboxylic acid containing 2,4-thiazolidinedione derivatives. Bioorg. Med.
Chem. 2008; 16: 5840-5852.
[27] Maccari R, Ciurleo R, Giglio M, Cappiello M, Moschini R, Del Corso A, Mura U,
Ottanà R. Identification of new non-carboxylic acid containing inhibitors of aldose
reductase. Bioorg. Med. Chem. 2010; 18: 4049-4055.
[28] Ottanà R, Maccari R, Giglio M, Del Corso A, Cappiello M, Mura U, Cosconati S,
Marinelli L, Novellino E, Sartini S, La Motta C, Da Settimo F. Identification of 5-
arylidene-4-thiazolidinone derivatives endowed with dual activity as aldose reductase
inhibitors and antioxidant agents for the treatment of diabetic complications. Eur. J. Med.
Chem. 2011; 46: 2797-2806.
[29] Maccari R, Del Corso A, Giglio M, Moschini R, Mura U, Ottanà R. In vitro evaluation of
5-arylidene-2-thioxo-4-thiazolidinones active as aldose reductase inhibitors. Bioorg. Med.
Chem. Lett. 2011; 21: 200-203.
[30] Maccari R, Vitale RM, Ottanà R, Rocchiccioli M, Marrazzo A, Cardile V, Graziano
ACE, Amodeo P, Mura U, Del Corso A. Structure-activity relationships and molecular
modelling of new 5-Arylidene-4-thiazolidinone derivatives as aldose reductase inhibitors
and potential anti-inflammatory agents. Eur. J. Med. Chem. 2014; 81: 1-14.
[31] Maccari R, Paoli P, Ottanà R, Jacomelli M, Ciurleo R, Manao G, Steindl T, Langer T,
Vigorita MG, Camici G. 5-Arylidene-2,4-thiazolidinediones as inhibitors of protein
tyrosine phosphatases. Bioorg. Med. Chem. 2007; 15: 5137-5149.
[32] Ottanà R, Maccari R, Ciurleo R, Paoli P, Jacomelli M, Manao G, Camici G, Laggner C,
Langer T. 5-Arylidene-2-phenylimino-4-thiazolidinones as PTP1B and LMW-PTP
inhibitors. Bioorg. Med. Chem. 2009; 17: 1928-1937.
[33] Maccari R, Ottanà R, Ciurleo R, Paoli P, Manao G, Camici G, Laggner C, Langer T.
Structure-based optimization of benzoic acids as inhibitors of protein tyrosine
phosphatase 1B and low molecular weight protein tyrosine phosphatase. ChemMedChem
2009; 4: 957-962.
25

[34] Ottanà R, Maccari R, Amuso S, Wolber G, Schuster D, Herdlinger S, Manao G, Camici
G, Paoli P. New 4-[(5-Arylidene-2-arylimino-4-oxo-3-thiazolidinyl)methyl]benzoic acids
active as protein tyrosine phosphatase inhibitors endowed with insulinomimetic effect on
mouse C2C12 skeletal muscle cells. Eur. J. Med. Chem. 2012; 50: 332-343.
[35] Ottanà R, Maccari R, Mortier J, Caselli A, Amuso S, Camici G, Rotondo A, Wolber G,
Paoli P. Synthesis, biological activity and structure-activity relationships of new benzoic
acid-based protein tyrosine phosphatase inhibitors endowed with insulinomimetic effects
in mouse C2C12 skeletal muscle cells. Eur. J. Med. Chem. 2014; 71: 112-127.
[36] Ottanà R, Paoli P, Naß A, Lori G, Cardile V, Adornato I, Rotondo A, Graziano ACE,
Wolber G, Maccari R. Discovery of 4-[(5-arylidene-4-oxothiazolidin-3-
yl)methyl]benzoic acid derivatives active as novel potent allosteric inhibitors of protein
tyrosine phosphatase 1B: In silico studies and in vitro evaluation as insulinomimetic and
anti-inflammatory agents. Eur. J. Med. Chem. 2017; 127: 840-858.
[37] Ottanà R, Maccari R, Barreca ML, Bruno G, Rotondo A, Rossi A, Chiricosta G, Di Paola
R, Sautebin L, Cuzzocrea S, Vigorita MG. 5-Arylidene-2-imino-4-thiazolidinones: design
and synthesis of novel antiinflammatory agents. Bioorg. Med. Chem. 2005; 13: 4243-
4252.
[38] Balestri F, Cappiello M, Moschini R, Rotondo R, Buggiani I, Pelosi P, Mura U, Del
Corso A. L-Idose: an attractive substrate alternative to D-glucose for measuring aldose
reductase activity. Biophys. Res. Commun. 2015; 456: 891-895.
[39] Balestri F, Cappiello M, Moschini R, Rotondo R, Abate A, Del Corso A, Mura U.
Modulation of aldose reductase activity by aldose hemiacetals. Biochim. Biophys. Acta
2015; 1850: 2329-2339.
[40] Steuber H, Zentgraf M, Gerlach C, Sotriffer CA, Heine A, Klebe G. Expect the
unexpected or caveat for drug designers: multiple structure determinations using aldose
reductase crystals treated under varying soaking and co-crystallisation conditions. J. Mol.
Biol. 2006; 363: 174-187.
[41] Zheng X, Zhang L, Chen W, Chen Y, Xie W, Hu X. Partial inhibition of aldose reductase
by nitazoxanide and its molecular basis. ChemMedChem 2012; 7: 1921-1923.
26

Graphical abstract
An investigation on 4-thiazolidinone derivatives as dual inhibitors of aldose reductase and protein
tyrosine phosphatase 1B, in the search for potential agents for the treatment of type 2 diabetes
mellitus and its complications
Rosanna Maccari, Antonella del Corso, Paolo Paoli, Ilenia Adornato, Giulia Lori, Francesco Balestri,
Mario Cappiello, Alexandra Naß, Gerhard Wolber, Rosaria Ottanà
27

An investigation on 4-thiazolidinone derivatives as dual inhibitors of aldose reductase and protein
tyrosine phosphatase 1B, in the search for potential agents for the treatment of type 2 diabetes
mellitus and its complications
Rosanna Maccari, Antonella del Corso, Paolo Paoli, Ilenia Adornato, Giulia Lori, Francesco Balestri,
Mario Cappiello, Alexandra Naß, Gerhard Wolber, Rosaria Ottanà
Highlights
- 4-Thiazolidinones were synthesised and tested as dual AR/PTP1B inhibitors.
- SAR, molecular docking and kinetic studies were carried out.
- Two compounds endowed with interesting AR/PTP1B inhibition profiles were identified.
- Features useful to achieve simultaneous inhibition of both AR and PTP1B emerged.
[42]
28