In search for multi-target ligands as potential agents for diabetes mellitus and its complications—a structure-activity relationship study on inhibitors of aldose reductase and protein tyrosine phosphatase 1b

molecules

Article

In Search for Multi-Target Ligands as Potential Agents for

Diabetes Mellitus and Its Complications—A Structure-Activity

Relationship Study on Inhibitors of Aldose Reductase and

Protein Tyrosine Phosphatase 1B

Rosaria Ottanà ¹, Paolo Paoli ², Mario Cappiello ³, Trung Ngoc Nguyen ⁴, Ilenia Adornato ¹,

Antonella Del Corso ³, Massimo Genovese ², Ilaria Nesi ², Roberta Moschini ³, Alexandra Naß ⁴,

Gerhard Wolber ⁴and Rosanna Maccari ^1,

*

1

Department of Chemical, Biological, Pharmaceutical and Environmental Sciences, University of Messina,

Viale Palatucci, Polo Universitario Annunziata, 98168 Messina, Italy; rottana@unime.it (R.O.);

ilenia_adornato@hotmail.it (I.A.)

Department of Scienze Biomediche Sperimentali e Cliniche, Sezione di Scienze Biochimiche, University of

Firenze, Viale Morgagni 50, 50134 Firenze, Italy; paolo.paoli@uniﬁ.it (P.P.);

massimo.genovese@student.unisi.it (M.G.); i.nesi2@student.unisi.it (I.N.)

Department of Biology, Biochemistry Unit, University of Pisa, Via S. Zeno, 51, 56123 Pisa, Italy;

mario.cappiello@unipi.it (M.C.); antonella.delcorso@unipi.it (A.D.C.); roberta.moschini@unipi.it (R.M.)

Molecular Design Lab, Institute of Pharmacy, Freie Universität Berlin, Königin-Luisestr. 2 + 4, 14195 Berlin,

Germany; gockyy@zedat.fu-berlin.de (T.N.N.); alexandra.nass@stenon.io (A.N.);

gerhard.wolber@fu-berlin.de (G.W.)

2

3

4

ꢀꢁꢂꢀꢃꢄꢅꢆꢇ

ꢀꢁꢂꢃꢄꢅꢆ

Citation: Ottanà, R.; Paoli, P.;

Cappiello, M.; Nguyen, T.N.;

*

Correspondence: rmaccari@unime.it; Tel.: +39-090-6766406

Adornato, I.; Del Corso, A.; Genovese,

M.; Nesi, I.; Moschini, R.; Naß, A.;

et al. In Search for Multi-Target

Ligands as Potential Agents for

Diabetes Mellitus and Its

Abstract: Diabetes mellitus (DM) is a complex disease which currently affects more than 460 million

people and is one of the leading cause of death worldwide. Its development implies numerous

metabolic dysfunctions and the onset of hyperglycaemia-induced chronic complications. Multiple

ligands can be rationally designed for the treatment of multifactorial diseases, such as DM, with the

precise aim of simultaneously controlling multiple pathogenic mechanisms related to the disease and

providing a more effective and safer therapeutic treatment compared to combinations of selective

drugs. Starting from our previous ﬁndings that highlighted the possibility to target both aldose

reductase (AR) and protein tyrosine phosphatase 1B (PTP1B), two enzymes strictly implicated

in the development of DM and its complications, we synthesised 3-(5-arylidene-4-oxothiazolidin-

3-yl)propanoic acids and analogous 2-butenoic acid derivatives, with the aim of balancing the

effectiveness of dual AR/PTP1B inhibitors which we had identiﬁed as designed multiple ligands

(DMLs). Out of the tested compounds, 4f exhibited well-balanced AR/PTP1B inhibitory effects at low

micromolar concentrations, along with interesting insulin-sensitizing activity in murine C2C12 cell

cultures. The SARs here highlighted along with their rationalization by in silico docking experiments

into both target enzymes provide further insights into this class of inhibitors for their development

as potential DML antidiabetic candidates.

Complications—A Structure-Activity

Relationship Study on Inhibitors of

Aldose Reductase and Protein

Tyrosine Phosphatase 1B. Molecules

2021, 26, 330. https://doi.org/

10.3390/molecules26020330

Academic Editor: George Kokotos

Received: 18 December 2020

Accepted: 7 January 2021

Published: 10 January 2021

Publisher’s Note: MDPI stays neu-

tral with regard to jurisdictional clai-

ms in published maps and institutio-

nal afﬁliations.

Keywords: multi-target ligands; diabetes mellitus; aldose reductase; protein tyrosine phosphatase

1B; 4-thiazolidinones; molecular docking

censee MDPI, Basel, Switzerland.

This article is an open access article

distributed under the terms and con-

ditions of the Creative Commons At-

tribution (CC BY) license (https://

creativecommons.org/licenses/by/

4.0/).

1. Introduction

In the last decade, the design and discovery of multi-target-directed ligands has

emerged as a modern promising strategy to develop new drugs aimed to the treatment

of multifactorial diseases, such as diabetes, cancer, infectious diseases, cardiovascular

and neurodegenerative disorders. Therapies based on a single-target-directed agent may

be insufﬁcient for the management of these complex pathologies, thus combinations of

different drugs generally represent therapeutic regimens necessary in order to achieve a

Molecules 2021, 26, 330. https://doi.org/10.3390/molecules26020330

https://ww w .mdpi.com/journal/molecules

Molecules 2021, 26, 330

2 of 32

higher efﬁcacy. However, combination therapies may involve pharmacokinetics, toxicity or

patient compliance problems.

Therefore, recently multi-targeted ligands have attracted great interest in medicinal

chemistry as a valuable alternative to combinations of single-targeted drugs. In particu-

lar, designed multiple ligands (DMLs) are compounds rationally designed to modulate

two or more selected targets simultaneously and, thus, to achieve pharmacological effects

favourable for the efﬁcient management of multifactorial pathologies. Growing evidence in-

dicates that DMLs may be endowed with enhanced efﬁcacy and improved safety compared

to drug combinations, especially in long-term therapies [1–4].

A drug design strategy frequently used to obtain DMLs starts from the knowledge

of the distinct pharmacophores of agents active on each of the individual targets and

consists in the integration of these pharmacophores into a single structure to obtain hybrid

molecules. Hybridization should be carried out in such a way that the pharmacophoric

elements inserted in a resulting DML maintain the ability to interact with speciﬁc sites of

the different selected targets, thus producing simultaneously the desired effects. In order

to obtain molecules endowed with appropriate drug-like properties, it is often preferred

to partially overlap the different pharmacophores, by using or merging them. However,

by applying the different possible approaches for the design of DMLs, it is quite likely

to achieve lead compounds endowed with high potency towards only one of the desired

targets, with less effect on the others; therefore, it is often necessary to carry out a balancing

process, using structural modiﬁcations of the lead compound, aimed to modulate and

balance the activity towards the selected targets until to reach an optimal ratio [1,4].

Recently we have reported an initial investigation on 4-thiazolidinone derivatives

active as dual inhibitors of aldose reductase and protein tyrosine phosphatase 1B, in the

context of our search for new candidates for the treatment of type 2 diabetes mellitus

and its complications [5]. Diabetes mellitus (DM) is a complex chronic metabolic disease

which represents a global severe health problem. It has been estimated that currently

more than 460 million people are affected by DM worldwide and the prevalence of this

disease is expected to increase to about 700 million by 2045 [6,7]. Moreover, it is estimated

that about 10% of the total deaths in the world are linked to DM. Glycaemic control

is the main goal in DM care and oral antihyperglycaemic drugs or insulin therapy are

commonly used for this purpose. However, when not adequately treated, diabetic patients

may be chronically exposed to unnatural glucose ﬂuctuations, which induce cell damage

and trigger a low-grade systemic inﬂammation, ultimately causing the onset of severe

long-term complications, such as micro/macroangiopathies, neuropathy, nephropathy

and cardiovascular dysfunctions. Therefore, to obtain an effective glycaemic control

and prevent chronic complications, the standard therapeutic treatment of DM is often

represented by drug cocktails. Consequently, it is conceivable that different key pathogenic

events involved in DM could be advantageously addressed by DMLs [8,9].

On these bases, in the context of a search for new DMLs as potential antidiabetic agents,

we selected as targets aldose reductase (AR) and protein tyrosine phosphatase 1B (PTP1B),

two enzymes differently involved in the development of DM and its complications.

AR (E.C. 1.1.1.21) is an aldo-keto reductase capable to catalyse the reduced nicoti-

namide adenine dinucleotide phosphate (NADPH)-dependent reduction of a variety of

aldehydes, including aldoses. Under hyperglycaemic conditions, such as in DM, an in-

creased ﬂux of glucose through the polyol pathway occurs. AR catalyses the ﬁrst and

rate-limiting step of this metabolic pathway, by reducing glucose to sorbitol which is

then oxidized to fructose by sorbitol dehydrogenase. The increased glucose metabolism

through the polyol pathway causes osmotic imbalance, decreases antioxidant cellular

defence and promotes glycation and inﬂammatory events, thus leading to tissue dam-

age which is responsible for the development of diabetic chronic complications [10,11].

AR is also capable to reduce reactive unsaturated aldehydes, derived from oxidative

stress-induced lipid peroxidation and their glutathione-conjugates, such as 4-hydroxy-

2,3-nonenal (HNE), thus participating in the detoxiﬁcation processes of toxic aldehydes.

Molecules 2021, 26, 330

3 of 32

However, the AR-catalysed reduction of the adduct glutathione-HNE (GS-HNE) gener-

ates 3-glutathionyl-1,4-dihydroxynonane (GS-DHN) which is a potent pro-inﬂammatory

molecule; inﬂammatory events triggered by GS-DHN are strictly related to the develop-

ment of cellular and vascular damage responsible for diabetic complications. On these

bases, AR inhibitors are actively searched to prevent and control the onset and progression

of hyperglycaemia-induced pathologies associated to DM and also as potential agents for

other inﬂammatory diseases [11–13].

Numerous evidences have undoubtedly demonstrated that PTP1B (E.C. 3.1.3.48) is

involved in the development of insulin-resistance, which is a characteristic condition in

type 2 DM (T2DM). This protein tyrosine phosphatase acts as a major negative regulator

of insulin signalling, by dephosphorylating speciﬁc phosphotyrosine residues of the acti-

vated insulin receptor and insulin receptor substrate proteins and thus attenuating insulin

signalling pathways. PTP1B has been shown to be overexpressed in T2DM and in other

pathologies associated with insulin resistance, such as obesity [14–16]. The deregulation of

this phosphatase is also strictly linked to a chronic low grade inﬂammation [17], related to

insulin resistance and DM. The inhibition or genetic ablation of PTP1B can enhance insulin

receptor phosphorylation and improve cellular response to insulin, thus counteracting

insulin resistance and increasing glucose uptake into cells [18,19]. Therefore, PTP1B is

considered a validated target to develop inhibitors as drug candidates for the treatment of

T2DM [20–22].

On these bases, both PTP1B and AR can be assumed as attractive molecular targets

for the design of DMLs capable to control simultaneously different cellular mechanisms

underlying the development of both T2DM and its complications. However, this research

is still in its infancy and, so far, only few dual ALR2/PTP1B inhibitors have been reported,

among which some natural compounds such as berberine [23,24].

Although these two enzymes belong to different families of proteins, the knowledge of

their active sites and the structure-activity relationships (SARs) of inhibitors of each of these

targets highlights several shared structural features that enable the design of dual inhibitors

by merging the respective pharmacophores. In particular, our previous studies on different

4-thiazolidinone derivatives, which were investigated as AR or PTP1B inhibitors (Figure 1)

,

evidenced that the pharmacophores of both classes of inhibitors comprise a polar head

in position 3 of the heterocyclic core, including an acidic or H-bond acceptor group and

a lipophilic portion in position 5, containing one or more suitably substituted aromatic

moieties [25–38].

Therefore, starting from a knowledge-based approach, recently we merged (5-arylidene-

4-oxothiazolidin-3-yl)acetic acid derivatives, which are endowed with excellent AR in-

hibitory activity and 4-[(5-arylidene-4-oxothiazolidin-3-yl)methyl]benzoic acids, which

we identiﬁed as potent PTP1B inhibitors, to obtain a series of new 4-thiazolidinone

derivatives (

1

,

2, Figure 1) that were evaluated as dual inhibitors of both human AR

and PTP1B enzymes [

5

]. Out of them, we identiﬁed two analogues, that is, (4-oxo-5-

{[3-(2-phenylethoxy)pheny]lmethylidene}-2-thioxothiazolidin-3-yl)acetic acid (2e) and the

corresponding 2,4-thiazolidinedione counterpart 1e (Figure 1), which exhibited potent AR

inhibitory effects along with appreciable PTP1B inhibitory capability. Kinetic and in silico

docking studies clearly indicated that these compounds behave as reversible inhibitors of

both human AR and PTP1B, displaying an uncompetitive and non-competitive mechanism

of action, respectively [5].

The opportunity to design inhibitors capable to bind non-catalytic regions of both AR

and PTP1B is a promising tool for the discovery of new DMLs in this context. Therefore, in

order to gain further insights into structural requirements for dual AR/PTP1B inhibition,

starting from the SARs acquired from compounds

3-(5-arylidene-4-oxothiazolidin-3-yl)propanoic acids 3a

2,4-dioxothiazolidin-3-yl)-2-butenoic acids 5a (Figure 1). In these newly-synthesised

1–

2

, we designed and synthesised

–

f

and 4a and 4-(5-arylidene-

–f

–

e

compounds, we modiﬁed the carboxylic chain on N-3, with the aim of balancing the

inhibitory effects against the two target enzymes. The elongation of this carboxylic chain

Molecules 2021, 26, 330

4 of 32

might be beneﬁcial to ensure further effective interactions with both target enzymes, as

suggested by our previous molecular docking investigation.

Figure 1. Structures of 4-thiazolidinone derivatives active as inhibitors of aldose reductase (AR) and/or protein tyrosine

phosphatase 1B (PTP1B). ^(a)Reference [5].

Moreover, a previously conducted shape-based analogue search on known active

co-crystallized ligands and de novo designed PTP1B inhibitors yielded over 700 commer-

cially available hits, which were docked into PTP1B catalytic site and the most promising

ligands were chosen for in vitro evaluation. Out of fourteen selected virtual hits, 3-{[5-(4-

benzyloxyphenyl)methylidene]-2,4-dioxothiazolidin-3-yl}propanoic acid showed a moder-

ate capability to inhibit human PTP1B. Therefore, this compound was assumed as a starting

point for hit-to-lead optimization.

The insertion of the 2-butenoic chain of compounds 5a–e was also suggested by our

previous investigations on both AR and PTP1B inhibitors [26]. In addition, preliminary

assays with 4-[(5-arylidene-4-oxothiazolidin-3-yl)methyl]benzoic acids (Figure 1) against

bovine lens AR had revealed no inhibition, probably because the bulky benzoic substituent

on the N-3 might prevent the interaction with the AR active site. Now, we decided to assess

if 2-butenoic acid derivatives are useful to achieve both PTP1B and AR inhibition since the

2-butenoic residue could be assumed as a simpler open mimetic of the benzoic acid residue

capable to interact effectively with PTP1B and, at the same time, could bind to AR.

In the 5-arylidene moiety of compounds 3–5, the structural features, such as two

differently linked aromatic rings that had proven to be effective for dual AR/PTP1B

inhibition [5] were maintained (Table 1).

Molecules 2021, 26, 330

5 of 32

a

Table 1. Inhibitory activities of compounds

3

–

5

against human AR and human PTP1B, expressed as IC₅₀

.

AR

IC₅₀(µM)

PTP1B

IC₅₀(µM)

X

R

Ar

3a

3b

3c

3d

3e

3f

O

S

(CH₂)₂COOH

3-OC₆H₅-C₆H₄

4-OC₆H₅-C₆H₄

11.9 ± 0.9

43.8 ± 7.1

14.3 ± 1.0

35.7 ± 3.0

27.9 ± 3.1

50.2 ± 4.6

2.2 ± 0.1

7.6 ± 0.6

3.8 ± 0.1

8.4 ± 0.7

2.3 ± 0.1

5.3 ± 0.4

3.9 ± 0.2

84% at 10 µM

88% at 5 µM

3.7 ± 0.2

79% at 50 µM

56% at 50 µM

76% at 50 µM

77% at 50 µM

64% at 50 µM

46% at 50 µM

34.1 ± 0.5

29.5 ± 0.4

42.8 ± 0.7

34.9 ± 0.7

55.5 ± 0.8

12.7 ± 0.3

42.1 ± 0.3

39.7 ± 0.1

34.8 ± 0.5

40.3 ± 0.5

37.1 ± 0.4

(CH₂)₂COOH

3-OCH₂C₆H₅-C₆H₄

4-OCH₂C₆H₅-C₆H₄

3-OCH₂CH₂C₆H₅-C₆H₄

4-OCH₂CH₂C₆H₅-C₆H₄

3-OC₆H₅-C₆H₄

(CH₂)₂COOH

4a

4b

4c

4d

4e

4f

(CH₂)₂COOH

S

(CH₂)₂COOH

4-OC₆H₅-C₆H₄

S

(CH₂)₂COOH

3-OCH₂C₆H₅-C₆H₄

4-OCH₂C₆H₅-C₆H₄

3-OCH₂CH₂C₆H₅-C₆H₄

4-OCH₂CH₂C₆H₅-C₆H₄

3-OC₆H₅-C₆H₄

S

(CH₂)₂COOH

S

(CH₂)₂COOH

S

(CH₂)₂COOH

5a

5b

5c

5d

5e

O

CH₂CH=CHCOOH

4-OC₆H₅-C₆H₄

4-C₆H₅-C₆H₄

1-naphthyl

2-naphthyl

86% at 10 µM

Epalrestat

Vanadate

0.102 ± 0.005

0.4 ± 0.01

a

IC₅₀

(µ

M) or % enzyme residual activity at the indicated concentration. Values are expressed as the mean

±

S.E.M (see methods for details).

Calculated physicochemical and pharmacokinetic parameters of the synthesised com-

pounds indicated good drug likeness for these molecules and suggested that high oral

bioavailability could be expected for all of them (see Supplementary Table S1).

In addition, although 2-thioxothiazolidinones 4a are present in the literature [39

3–5

–d

–41],

to our knowledge none of them had ever been evaluated as AR or PTP1B inhibitor so far.

2. Results and Discussion

2.1. Chemistry

3-(5-Arylidene-2,4-dioxothiazolidin-3-yl)propanoic acids (3a

2,4-dioxothiazolidin-3-yl)-2-butenoic acids (5a ) were prepared through multi-step pro-

cedures starting from 5-arylidene-2,4-thiazolidinediones 6a (Scheme 1). Precursor com-

pounds were prepared with good yields by the Knoevenagel condensation of commercial

–f) and 4-(5-arylidene-

–e

–i

6

2,4-thiazolidinedione with appropriate aldehydes, in reﬂuxing ethanol and in the presence

of piperidine as a base, according to a procedure that we had previously reported [26,37].

Subsequently, the reaction of the appropriate compound 6 with 3-chloropropanoic acid

in reﬂuxing acetone and in the presence of potassium carbonate, followed by a work-up

in acidic medium and recrystallization from methanol, provided pure 3-(5-arylidene-2,4-

dioxothiazolidin-3-yl)propanoic acids (3a–f) in high yields (Scheme 1).

Molecules 2021, 26, 330

6 of 32

The reaction of precursor 2,4-thiazolidinediones

6

with methyl 4-bromocrotonate

in reﬂuxing acetone provided 4-(5-arylidene-2,4-dioxothiazolidin-3-yl)-2-butenoic acid

methyl esters, which were hydrolysed in an acidic medium to generate the corresponding

acids 5a–e (Scheme 1).

The synthetic procedure to obtain 3-(5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)propanoic

acids (4a

sised by the cyclization of

in the presence of sodium hydroxide, in aqueous solution. Subsequently, the desired

3-(5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)propanoic acids (4a ) were obtained by the

Knoevenagel condensation of 3-(4-oxo-2-thioxothiazolidin-3-yl)propanoic acid ( ) with the

appropriate aldehyde, in reﬂuxing acetic acid in the presence of sodium acetate (Scheme 2).

–f) included the preparation of precursor

7 (Scheme 2 ). Compound 7 was synthe-

β-alanine with carbon disulphide and sodium bromoacetate,

–f

7

Scheme 1. Synthesis of 3-(5-arylidene-2,4-dioxothiazolidin-3-yl)propanoic acids 3a–f and 4-(5-arylidene-2,4-dioxothiazolidin

-3-yl)-2-butenoic acids 5a–e.

Molecules 2021, 26, 330

7 of 32

Scheme 2. Synthesis of 3-(5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)propanoic acids 4a–f.

The structures of compounds

¹H and ¹³C-NMR spectroscopy data (see Materials and Methods and Supplementary Figures

S1–S28). NMR spectroscopy revealed that all synthesised 5-arylidene derivatives were

3–5 were unambiguously assigned using analytical and

3–5

obtained only as Z isomers, analogously to previously investigated 5-arylidene-4-thiazolidinone

derivatives which had also been analysed through X-ray crystallography [25 ,42 ]. All ¹H-NMR

spectra showed only one singlet attributable to the resonance of the 5-methylidene group,

which appeared in the range 7.63–8.57 ppm.

¹H-NMR spectra of propanoic acids

3 and 4 exhibited two characteristic triplets at

2.57–2.63 ppm and 3.84–4.23 ppm (J = 7.2–7.5 Hz) due to the resonance of the methylene

protons of the propanoic chain on N-3. In the spectra of compounds 3e,f and 4e,f, two

additional triplets (with smaller coupling constant values of 6.8–6.9 Hz) are present at 3.05–

3.06 ppm and 4.25–4.29 ppm, originated from the resonance of the ethoxy chain included

in the 5-arylidene portion.

In ¹H-NMR spectra of compounds

5, the resonance of 2-butenoic chain gave rise to

characteristic signals, that is, a doublet at 4.45–4.47 attributable to NCH₂protons and two

multiplets in the range 5.87–6.83 ppm originated by CH=CH.

In ¹³C-NMR spectra, the singlets produced by the resonance of the carbonyl groups in

the range 165.8–173.1 ppm are diagnostic as well as, in the case of compounds

at 193.7–194.2 ppm attributable to the thiocarbonyl group.

4, the signal

2.2. AR and PTP1B Inhibition

The inhibitory effects of compounds 3–5 were assessed in vitro against both human

recombinant AR, by using L-idose as substrate and epalrestat as the reference drug and

human recombinant PTP1B, by using p-nitrophenyl phosphate as substrate and sodium

metavanadate as the reference drug. Table 1 reports the data resulting from these assays

for the two target enzymes.

Molecules 2021, 26, 330

8 of 32

Compounds 5a

–

c

had already been evaluated against bovine lens AR, using D,L-

glyceraldehyde as a substrate [26]; here, they were assayed against the human enzyme

with L-idose as a substrate to obtain comparable data useful for the SAR study.

All tested 3-(5-arylidene-4-oxothiazolidin-3-yl)propanoic acids 3a–f and 4a–f showed

good inhibitory properties against human AR, with IC₅₀values in the mid- and low-

micromolar range (Table 1). Their AR inhibitory potency proved to be markedly inﬂuenced

by the substituents in the positions 2 and 3 of the thiazolidinone scaffold and, to a lesser

extent, by the nature of the 5-arylidene moiety. 3-(5-Arylidene-4-oxo-2-thioxothiazolidin-

3-yl)propanoic acids

of them, 4a and 4e proved to be the most potent inhibitors of human AR, with IC₅₀

values of 2.2 M and 2.3 M, respectively. 2,4-Thiazolidinedione analogues showed

the appreciable capability to inhibit the enzyme, especially compounds 3a and 3c, which

displayed IC₅₀values slightly higher than 10 M (Table 1). On the whole, 2-thioxo-4-

thiazolidinone derivatives 4a proved to be from 4- to 12-fold more efﬁcient AR inhibitors

4 exhibited IC₅₀values in the low micromolar range (<10 µM); out

µ

3

µ

–

f

than the corresponding 2,4-thiazolidinones 3a–f (Table 1), as already observed for other

analogues [32].

However, compared with the parent acetic acids (compounds 1, 2 Figure 1), the

elongation of the carboxylic chain on N-3 from two to three carbon atoms appeared to be

responsible for a signiﬁcant decrease in AR inhibitory potency for both 3-(5-arylidene-4-

oxo-2-thioxothiazolidin-3-yl)propanoic acids

The IC₅₀values obtained for compounds 4a

than 100-fold less potent than acetic acid analogues

4

f

and their 2,4-thiazolidinedione analogues

revealed that they were from 40-fold to more

. Analogously, compounds 3e and 3f

3.

–

2

showed IC₅₀values two orders of magnitude higher than those of the corresponding acetic

acid derivatives 1 [5].

The inﬂuence exerted on the AR inhibitory potency by the nature of the moiety in

position 5 of the heterocyclic core did not appear to be marked; in fact, we observed

that the IC₅₀values fall within a relatively narrow range for each series

However, in both series and , compounds bearing a substituent in the meta position

of the 5-benzylidene ring (3a 3c 3e and 4a 4c 4e) exhibited better activity than the para-

substituted isomers (3b 3d 3f and 4b 4d 4f, respectively), conﬁrming a SAR that we had

observed in the previously investigated 4-thiazolidinones active as AR inhibitors [25 32].

Among 2-butenoic acid derivatives , compounds 5a and 5d exhibited signiﬁcant capability

to inhibit human AR with IC₅₀values of 3.9 M and 3.7 M, respectively, whereas the

other analogues (5b 5c 5e) produced scarce inhibition at concentrations up to 10

3 and 4 (Table 1).

3

4

,

–

5

µ

,

µM

(Table 1). It is worthwhile to notice that 5a had shown very scarce inhibitory activity

towards bovine lens AR, in the presence of glyceraldehyde as substrate, whereas 5b and 5c

were almost inactive towards both bovine and human enzymes with glyceraldehyde and

idose as substrate, respectively [26]. These data conﬁrmed previous observations that both

the enzymatic source and the substrate used are relevant factors affecting AR inhibitor

susceptibility [43,44].

The inhibitory activity towards human PTP1B of compounds

be lower than towards human AR. However, all 3-(5-arylidene-4-oxo-2-thioxothiazolidin-3-

yl)propanoic acids (4a ) and 4-(5-arylidene-2,4-dioxothiazolidin-3-yl)-2-butenoic acids (5a

3–5 was generally found to

–f

–e)

displayed interesting PTP1B inhibitory properties, with mid-micromolar IC₅₀values (Table 1 ).

Among them, 3-[4-oxo-{5-[4-(2-phenylethoxy)phenyl]methylidene}-2-thioxothiazolidin-3-yl]

propanoic acid (4f) was shown to be the most effective PTP1B inhibitor (IC₅₀= 12.7

compounds , we found that selected substituents in the para position of the 5-benzylidene

ring improved moderately the inhibitory effects, compared to the meta isomers.

In addition, it is worthwhile to notice that, in comparison with the previously investi-

gated acetic analogues , the elongation of the carboxylic chain on N-3 generally improved

the PTP1B inhibitory activity, as demonstrated by propanoic derivatives ; in fact, we found

that compounds 4a 4b 4d and 4f were from 2- to 4-fold more potent PTP1B inhibitors

than their acetic counterparts [ ]. On the other hand, compared with the corresponding

2-thioxo-4-thiazolidinones , the PTP1B inhibitory ability of 2,4-thiazolidinediones was

µM). In

4

2

4

,

5

4

3

Molecules 2021, 26, 330

9 of 32

shown to be signiﬁcantly lower (Table 1), highlighting that 3-(4-oxo-2-thioxothiazolidin-

3-yl)]propanoic moiety is more favourable for PTP1B inhibition. However, out of the

tested 2,4-thiazolidinedione derivatives, the replacement of the propanoic chain on N-3

(compounds

5b versus 3a, 3b) and allowed us to identify interesting PTP1B inhibitors.

Compounds 4a 4e and 4f, which exhibited appreciable dual AR/PTP1B inhibitory

3) with a 2-butenoic residue (compounds 5) provided a gain in potency (5a,

,

activity, were selected to be further studied for their kinetic behaviour on both target

enzymes and on cultured cells.

2.3. Kinetic Studies

Compounds 4a, 4e and 4f behaved as reversible inhibitors toward both AR and PTP1B

enzymes. In fact, in the case of AR more than 90% of enzyme activity was recovered after

removal of the inhibitor upon extensive dialysis (see Materials and Methods for details).

Similar results were obtained with PTP1B (Supplementary Figure S29).

The evaluation of apparent dissociation constants K_iand K’_ifor compounds 4a, 4e

and 4f (Table 2 and Supplementary Figures S30–S32) indicated 4e as the most potent

inhibitor toward AR. The characterization of the mechanism of action of the three selected

compounds revealed for 4a a mixed-type inhibition, with a slight preference toward the EIS

complex (K’_iapproximately 4-fold lower than K_i) (Figure S30). On the other hand, both 4e

(Figure S31) and 4f (Figure S32) resulted to behave essentially as uncompetitive inhibitors

(K_ivalues at least ten-fold higher concerning the corresponding K’i).

Table 2. Apparent dissociation constants for selected compounds ^a.

AR

PTP1B

0

K_i

(µM)

K _i

K_i

(µM)

K _i

Inhibitor

(µM)

68.9 ± 1.5

>12

4a

4e

4f

5.0 ± 0.57

>8

1.4 ± 0.022

0.66 ± 0.30

2.9 ± 0.12

10.6 ± 1.7

2.2 ± 0.3

0.9 ± 0.1

>30

2.3 ± 0.1

a

K_iand K’_irefer to the apparent dissociation constants for the ES and EIS complex, respectively

As far as concerned PTP1B, kinetic analyses revealed that both compounds 4a and

4f behave as mixed-type non-competitive inhibitors. Indeed, in both cases, experimental

points describing the PTP1B catalytic rate in the presence of increasing concentrations of

these compounds were ﬁtted by straight lines intersecting one each other in a point located

in the left quadrant (Supplementary Figures S33 and S35). Moreover, we observed that

both compounds caused an increase of the Km value (Figures S33 and S35B) and a decrease

of Vmax. (Figures S33 and S35C). However, kinetic behaviour of compound 4e appeared

different, as demonstrated by the fact that experimental points described straight lines

intersecting one each other’s in a point located on the Y-axis (Figure S34A). According to

these data, we observed that the Km value increased (Figure S34B), while Vmax did not

change with increasing concentration of compound 4e (Figure S34C); therefore, 4e behaved

as a competitive inhibitor

2.4. AR and PTP1B Docking Experiments

Docking studies of selected AR/PTP1B dual inhibitors 4a, 4e and 4f were carried out

to evaluate their binding mode with both target enzymes and rationalize the observed inhi-

bition mechanisms. In accordance with the kinetic studies that revealed 4a as a mixed AR

inhibitor, whereas 4e and 4f behaved as uncompetitive AR inhibitors, all three compounds

showed consistent docking poses with the AR-idose complex, whereas only 4a revealed

plausible docking poses when docked into the catalytic site of AR.

The shared propionic acid moiety of 4a

, 4e and 4f was located outside of the AR

catalytic site, in which L-idose is bound and showed ionic and hydrogen bonding in-

Molecules 2021, 26, 330

10 of 32

teractions with Arg217 and Lys221 (Figures 2–4). Hydrogen bonding interactions can

be observed for the carbonyl group in position 4 of the thiazolidinone scaffold with the

backbone of Ala299 and the bridging ether oxygen between the two phenyl rings with

Ser302. The thiocarbonyl group in position 2 was adjacent to Leu301 and pointed towards

the solvent. The 5-benzylidene ring of the compounds formed lipophilic contacts with

Trp219 and Leu301. As shown in our previous study [5], a lipophilic moiety in the meta

position of the 5-benzylidene ring is favourable over a para substitution due to the re-

quired bent that simultaneously allowed the ligands both to form hydrogen bonding with

Arg217 and Lys221 and to ﬁt the lipophilic pocket (Trp20, Val47, Trp79, Phe122) above

the idose-bound competitive binding site. The phenylethoxy moiety in the meta position

of the 5-benzylidene ring (4e, Figure 3) established these interactions, whereas the same

moiety in the para position (4f, Figure 4) was more exposed to the solvent and interacted

only with Phe122 and Leu124 located outside of the lipophilic pocket. The better ﬁt of 4e

compared to 4f is in agreement with the AR inhibitory potency trend (IC₅₀= 2.3 µM vs.

IC₅₀= 5.3 µM). The replacement of the phenylethoxy group in the meta position of the

5-benzylidene ring with a phenoxy one (compound 4a) resulted in a detrimental factor

for binding to the lipophilic pocket of the AR-idose complex, which is attributable to the

shorter linker between the two phenyl groups. However, in the case of 4a a better ﬁtting

of the shorter substituent into the AR narrow catalytic centre came out (Figure 5). This

may explain the comparable efﬁciency of 4a and 4e. On the other hand, 4e and 4f docked

into the catalytic centre revealed unfavorable docking poses where the terminal phenyl

group is evidently solvent exposed, thus conﬁrming the observed uncompetitive mode

of action. A recent analysis from Balestri et al. indicated rational binding modes to the

catalytic centre of AR for different substrates (L-idose, HNE and GSHNE) binding [45]. The

aldehyde moiety of all three substrates shows simultaneous hydrogen bonding to His110

and Tyr48. These interactions were established for 4a neither in our proposed binding

mode nor in the other generated docking poses. The reason for this is the rigidity and

size of 4a compared to AR substrates. The diphenylether moiety of 4a is located slightly

above the catalytic centre, embedded between lipophilic residues Trp20, Val47, Trp79 and

Phe122 (Figure 5). Interestingly, we observed two aromatic ring interactions, where the

thioxothiazolidinone moiety interacted with Trp219 and the 5-benzylidene ring with Trp20.

Furthermore, the thiocarbonyl group in position 2 of the thiazolidinone scaffold established

hydrogen bonding to Ser302 and the carbonyl group in position 4 to the backbone of

Leu300. These observations lead to the conclusion that 4a shows a different binding mode

compared to the AR substrates analysed by Balestri et al. [45] and that direct interaction

with Tyr48 and His110 might not be mandatory for competitive binding to the catalytic site

of AR.

PTP1B kinetic studies showed that 4a and 4f behave as mixed non-competitive in-

hibitors, whereas 4e behaves as a pure competitive inhibitor. Therefore, we performed

docking studies into both the catalytic binding pocket and the previously investigated

allosteric binding pocket of PTP1B [5].

Docking of 4a, 4e and 4f into the catalytic binding site of PTP1B (Figures 6–8) revealed

hydrogen bonding of the propionic acid moiety to Arg221 and the backbones of Phe182,

Ile219 and Gly220. Lipophilic contacts between the 5-benzylidene ring and Tyr46 and

Val49 were observed for all three compounds. The terminal phenyl ring of 4a is embedded

into the lipophilic environment consisting of Val49, Ile219 and Met258, whereas the longer

2-ethoxy linker of compounds 4e and 4f allowed the terminal phenyl ring to reach out

residues Arg24, Arg254 of the adjacent non-catalytic secondary aryl-phosphate binding

site [46]. Additionally, in the same binding site, docking poses of 4e were less consistent

compared to 4a and 4f, revealing binding poses where the terminal phenyl ring is adjacent

to Asp48 and more exposed to the solvent. Therefore, a longer linker on the 5-benzylidene

ring (4e and 4f) appeared to be not advantageous for binding at the catalytic region of

PTP1B, when in meta position, which could explain the lower inhibitory potency of 4e than

4f.

Molecules 2021, 26, 330

11 of 32

Figure 2. (Top): 3D depiction of the selected pose of 4a bound to the AR-idose complex. Blue surface color: Hydrophilic,

Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4a bound to the AR-idose complex.

Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines: Lipophilic contacts.

Docking of 4a

, 4e and 4f into the previously described allosteric binding site of

PTP1B [ ] (Figures 9–11) revealed for all three ligands comparable hydrogen bonding

5

interactions of the propionic acid moiety to Lys103, Arg105 and Arg169 and lipophilic

contacts of the 5-benzylidene ring to Pro210. The shorter linker between the two phenyl

rings of 4a did not allow the carboxyl group to bind to Lys103, Arg105 and Arg169 by

hydrogen bonding and, simultaneously, the terminal phenyl ring to reach the lipophilic

pocket consisting of Pro206 and surrounded by Arg79 and Ser80. The longer ethoxy

linker of 4f enabled both hydrogen bonding and lipophilic contacts at the same time.

This structural feature could explain the higher inhibitory potency of 4f (IC₅₀= 12.7

µM)

compared to 4a (IC₅₀= 34.1 M). Furthermore, the propionic acid in combination with a

µ

meta substitution of the 5-benzylidene ring of 4e made the 5-benzylidene ring more exposed

to the solvent, thus resulting in a lower afﬁnity for this allosteric site of 4e compared to 4a

and 4f. On the whole, these ﬁndings were in agreement with the kinetic study and could

further explain the difference in inhibitory potency between 4e and 4a, 4f.

Molecules 2021, 26, 330

12 of 32

Figure 3. (Top): 3D depiction of the selected pose of 4e bound to the AR-idose complex. Blue surface color: Hydrophilic,

Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein ligand interactions of 4e bound to the AR-idose complex.

Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines: Lipophilic contacts.

Figure 4. Cont.

Molecules 2021, 26, 330

13 of 32

Figure 4. (Top): 3D depiction of the selected pose of 4f bound to the AR-idose complex. Blue surface color: Hydrophilic,

Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein ligand interactions of 4f bound to the AR-idose complex.

Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines: Lipophilic contacts.

Figure 5. (Top): 3D depiction of the selected pose of 4a bound to the catalytic center AR. Blue surface color: Hydrophilic,

Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein ligand interactions of 4a bound to the catalytic center of

AR. Red circles: hydrogen bond acceptor, Yellow dotted lines: Lipophilic contacts, Blue donuts: aromatic ring interaction.

Molecules 2021, 26, 330

14 of 32

Figure 6.

(Top): 3D depiction of the selected pose of 4a bound to the catalytic center of PTP1B. Blue surface color:

Hydrophilic, Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4a bound to the

catalytic center of PTP1B. Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines: Lipophilic

contacts.

2.5. Ex Vivo Assays

We evaluated the effects of selected compound 4f on HepG2 and C2C12 cells. Cy-

totoxicity of the tested compound was determined using MTT assay. Data reported in

Figure 12 show that 4f was well tolerated by both human HepG2 and murine C2C12 cells,

as demonstrated by the fact that a reduction of cell viability was observed just when cells

(C2C12) were incubated with the highest concentration of compounds (50 µM).

The insulin mimetic/sensitizing activity of compound 4f was evaluated on both

HepG2 and C2C12 cells. Starved HepG2 and C2C12 cells were incubated in the presence of

compound 4f (25 µM) for 90 min and after stimulated or not with insulin (10 nM). Then,

cells extracts were analysed to evaluate phosphorylation levels of the insulin receptor.

Results of the experiments were showed in Figures 13 and 14.

Molecules 2021, 26, 330

15 of 32

Figure 7.

(Top): 3D depiction of the selected pose of 4e bound to the catalytic center of PTP1B. Blue surface color:

Hydrophilic, Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4e bound to the

catalytic center of PTP1B. Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines: Lipophilic

contacts.

It is interesting to note that compound 4f showed a different effect on liver and muscle

cells. As far as concerned HepG2 cells, we observed that the treatment with compound 4f

caused a marked reduction of basal phosphorylation levels of IR

after stimulation with insulin, IR phosphorylation levels increase in both control and

treated cells, although a relatively higher increase has been observed in the latter. In fact, we

calculated that the level of IR phosphorylation increases by 39% and 260%, respectively,

β (Figure 13B,C). However,

β

in the control and in the cells treated with compound 4f. This ﬁnding suggests that the

treatment with compound 4f can make the insulin receptor “more activatable” than that of

untreated cells. At the same time, we observed that treating C2C12 cells with compound

4f did not appreciably affect basal IR phosphorylation levels but signiﬁcantly improved

insulin activity. According to this hypothesis, we found that IRβ phosphorylation levels

resulted to be signiﬁcantly higher in treated cells than in those treated with insulin alone

(Figure 14B,C). Taken together, these ﬁndings suggest that compound 4f behaves as an

insulin-sensitizing agent in both the liver and muscle cells.

Molecules 2021, 26, 330

16 of 32

Figure 8. (Top): 3D depiction of the selected pose of 4f bound to the catalytic center of PTP1B. Blue surface color: Hydrophilic,

Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4f bound to the catalytic center of

PTP1B. Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines: Lipophilic contacts.

Molecules 2021, 26, 330

17 of 32

Figure 9. (Top): 3D depiction of the selected pose of 4a bound to the allosteric binding site of PTP1B. Blue surface color:

Hydrophilic, Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4a bound to the

allosteric binding site of PTP1B. Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines:

Lipophilic contacts.

Molecules 2021, 26, 330

18 of 32

Figure 10. (Top): 3D depiction of the selected pose of 4e bound to the allosteric binding site of PTP1B. Blue surface color:

Hydrophilic, Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4e bound to the

allosteric binding site of PTP1B. Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines:

Lipophilic contacts.

Molecules 2021, 26, 330

19 of 32

Figure 11. (Top): 3D depiction of the selected pose of 4f bound to the allosteric binding site of PTP1B. Blue surface color:

Hydrophilic, Yellow surface color: Lipophilic. (Bottom): 2D depiction of protein-ligand interactions of 4f bound to the

allosteric binding site of PTP1B. Red circles: hydrogen bond acceptor, Red stars: Anionic interaction, Yellow dotted lines:

Lipophilic contacts.

Molecules 2021, 26, 330

20 of 32

(A)

(B)

Figure 12. Viability assay. HepG2 cells (

A

) and C2C12 cells (

B) were incubated in the presence of

increasing concentrations (µM) of compounds 4f for 48 h. After this time, cells viability was assessed

using MTT assay. Data obtained were normalized respect to control test. Data showed in the ﬁgure

represent the mean ± S.E.M. (n = 4).

Molecules 2021, 26, 330

21 of 32

Figure 13. Phosphorylation levels of the insulin receptor in HepG2 cells. Cells were starved for 24 h and then treated for

90 min with compound 4f (ﬁnal concentration 25 M). After this time, cells were washed with PBS and stimulated with

µ

insulin (10 nM) for 30 min. Then, cells were washed with cold PBS, lysed and cell extracts analysed by western blot to

evaluate insulin receptor phosphorylation levels. The intensity of actin bands was used as a loading control. All tests were

carried out in triplicate (n = 3). (A): western blots. (B,C): quantiﬁcation of pIRβ/IRβ ratio.

Molecules 2021, 26, 330

22 of 32

Figure 14. Phosphorylation levels of the insulin receptor in C2C12 cells. Cells were starved for 24 h and then treated for

90 min with compound 4f (ﬁnal concentration 25 M). After this time, cells were washed with PBS and stimulated with

µ

insulin (10 nM) for 30 min. Then, cells were washed with cold PBS, lysed and cell extracts analysed by western blot to

evaluate insulin receptor phosphorylation levels. The intensity of actin bands was used as a loading control. Each test was

carried out in triplicate (n = 3). (A): western blots. (B,C): quantiﬁcation of pIRβ/IRβ ratio.

3. Materials and Methods

3.1. Chemistry

Melting points were recorded on a Koﬂer hot-stage apparatus and are uncorrected.

TLC controls were carried out on precoated silica gel plates (F 254 Merck). Rf values were

determined by using appropriate mixtures of diethyl ether/n-hexane as eluent. Combus-

tion analyses (C, H, N), determined by means of a C. Erba mod. 1106 elem. Analyzer,

1

were within

±

0.4% of the theoretical values. H and ¹³C-NMR spectra were recorded on a

Varian 500 MHz spectrometer operating at 499.74 and 125.73 MHz, respectively. Chemical

shifts are given in ppm and coupling constants are expressed in Hz. All the spectra were

δ

phased, baseline was corrected where necessary and CDCl₃or DMSO-d₆signals were used

as a reference for both ¹H and ¹³C spectra. All exchangeable protons were conﬁrmed by

addition of D₂O. Unless stated otherwise, all materials were obtained from commercial

suppliers and used without further puriﬁcation. 3-(2-Phenylethoxy)benzaldehyde and

Molecules 2021, 26, 330

23 of 32

4-(2-phenylethoxy)benzaldehyde were synthesised according to the procedure reported

in ref. 37. Compounds 3d, 4a, 4c, 4d, 4f and 7 are commercially available; however, we

synthesised them according to the procedure described below. The purity of synthetic

compounds was established as ≥ 95% by combustion analysis.

3.1.1. General Procedure for the Synthesis of 5-arylidene-2,4-dioxothiazolidinones 6a–i

A mixture of 2,4-thiazolidinedione (1.17 g, 10 mmol), the appropriate aldehyde

(10 mmol) and piperidine (0.68 g, 8 mmol) in ethanol (70 mL) was reﬂuxed for 24 h. The

crude mixture was poured into H₂O acidiﬁed with AcOH (pH 3–4) to give a crude solid

which was recrystallized from methanol to provide the corresponding pure 5-arylidene-2,4-

thiazolidinedione

6. Chemical-physical and spectroscopic data of compounds 6a–i were

reported in Refs. [26,37].

3.1.2. General Procedure for the Synthesis of 3-(5-arylidene-2,4-dioxothiazolidin-3-yl)

propanoic Acids 3a–f

A mixture of 5-arylidene-2,4-thiazolidinedione

6 (1.68 mmol), 3-chloropropanoic acid

(0.365 g, 3.36 mmol) and potassium carbonate (0.93 g, 6.73 mmol) in anhydrous acetone

(40 mL) was reﬂuxed for 24 h. The solvent was evaporated under reduced pressure. The

solid residue was poured into H₂O, acidiﬁed with HCl 6M (pH 3) and stirred until CO₂

disappearance. The solid was ﬁltered off, washed with H₂O and recrystallized from

methanol to provide pure compounds 3.

3-{(5Z)-[2,4-Dioxo-5-(3-phenoxyphenyl)methylidene]thiazolidin-3-yl}propanoic acid (3a). Yield

◦

66% (409.6 mg); m.p. 177–180 C; ¹H-NMR (DMSO-d₆):

δ 2.59 (t J = 7.5 Hz, 2H, CH₂COOH);

3.84 (t J = 7.5 Hz, 2H, NCH₂); 7.09 (m, 2H, arom); 7.13 (m, 1H, arom); 7.19–7.23 (m, 2H,

arom); 7.37 (m, 1H arom); 7.44 (m, 2H, arom), 7.54 (m, 1H, arom); 7.90 (s, 1H, methylidene);

12.45 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ 32.0 (CH₂), 38.0 (CH₂), 119.5 (CH), 119.9

(CH), 120.8 (C), 122.8 (C), 124.8 (CH), 125.4 (CH), 130.8 (CH), 131.6 (CH), 132.7 (CH), 135.4

(CH), 156.3 (C), 158.2 (C), 165.8 (C), 167.4 (C), 172.4 (C). Anal. (C₁₉H₁₅NO₅S) calcd: C 61.78;

H 4.09; N 3.97; found: 61.92; H 4.03; N 3.99.

3-{(5Z)-[2,4-Dioxo-5-(4-phenoxyphenyl)methylidene]thiazolidin-3-yl}propanoic acid (3b). Yield

◦

60% (372.3 mg); m.p. 198–202 C; ¹H-NMR (DMSO-d₆):

δ 2.60 (t J = 7.5 Hz, 2H, CH₂COOH);

3.86 (t J = 7.5 Hz, 2H, NCH₂); 7.11 (m, 4H, arom); 7.24 (m, 1H, arom); 7.45 (m, 2H, arom);

7.64 (m, 2H arom); 7.90 (s, 1H, methylidene); 12.45 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ

32.1 (CH₂), 38.0 (CH₂), 118.9 (CH), 120.1 (CH), 120.4 (C), 125.2 (CH), 128.2 (C), 130.9

(CH), 132.9 (CH), 133.0 (CH), 155.7 (C), 159.6 (C), 166.0 (C), 167.6 (C), 172.4 (C). Anal.

(C₁₉H₁₅NO₅S) calcd: C 61.78; H 4.09; N 3.97; found: C 61.97; H 3.98; N 4.01.

3-{(5Z)-[5-(3-Benzyloxyphenyl)_◦methylidene]-2,4-dioxothiazolidin-3-yl}propanoic acid (3c). Yield

25% (161.0 mg); m.p. 142–145 C; ¹H-NMR (DMSO-d₆):

δ 2.60 (t J = 7.5 Hz, 2H, CH₂COOH);

3.86 (t J = 7.5 Hz, 2H, NCH₂); 5.17 (s, 2H, OCH₂); 7.14–7.20 (m, 2H, arom); 7.24 (s, 1H, arom);

7.34 (m, 1H, arom); 7.40 (m, 2H arom); 7.44–7.48 (m, 3H, arom); 7.88 (s, 1H, methylidene);

12.45 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ 32.0 (CH₂), 38.0 (CH₂), 70.0 (CH₂), 116.7

(CH), 118.0 (C), 122.3 (C), 122.9 (CH), 128.3 (CH), 128.5 (CH), 129.1 (CH), 131.1 (CH), 133.3

(CH), 134.9 (CH), 137.2 (C), 159.3 (C), 165.9 (C), 167.6 (C), 172.4 (C). Anal. (C₂₀H₁₇NO₅S)

calcd: C 62.65; H 4.47; N 3.65; found: C 62.83; H 4.51; N 3.48.

3-{(5Z)-[5-(4-Benzyloxyphenyl)_◦methylidene]-2,4-dioxothiazolidin-3-yl}propanoic acid (3d). Yield

61% (392.9 mg); m.p. 195–199 C; ¹H-NMR (DMSO-d₆):

δ 2.58 (t J = 7.5 Hz, 2H, CH₂COOH);

3.85 (t J = 7.5 Hz, 2H, NCH₂); 5.19 (s, 2H, OCH₂); 7.18 (m, 2H, arom); 7.34 (m, 1H, arom);

7.40 (m, 2H, arom); 7.46 (m, 2H arom); 7.59 (m, 2H, arom); 7.88 (s, 1H, methylidene); 12.42

(s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ 32.2 (CH₂), 38.0 (CH₂), 70.1 (CH₂), 116.3 (CH),

118.7 (C), 126.2 (C), 128.4 (CH), 128.6 (CH), 129.1 (CH), 132.8 (CH), 133.4 (CH), 134.9 (CH),

137.0 (C), 160.8 (C), 166.1 (C), 167.7 (C), 172.6 (C). Anal. (C₂₀H₁₇NO₅S) calcd: C 62.65; H

4.47; N 3.65; found: C 62.78; H 4.56; N 3.55.

Molecules 2021, 26, 330

24 of 32

3-[(5Z)-2,4-Dioxo-{5-[3-(2-phenylethoxy)_◦phenyl]methylidene}thiazolidin-3-yl]propanoic acid (3e).

1

Yield 81% (540.8 mg); m.p. 130–133 C; H-NMR (DMSO-d₆):

δ 2.60 (t J = 7.5 Hz, 2H,

CH₂COOH); 3.06 (t J = 6.8 Hz, 2H, CH₂Ph); 3.86 (t J = 7.5 Hz, 2H, NCH₂); 4.25 (t J = 6.8 Hz

2H, OCH₂); 7.07 (m, 1H, arom); 7.16–7.24 (m, 3H, arom); 7.3–7.35 (m, 4H, arom); 7.44 (m,

1H arom); 7.89 (s, 1H, methylidene); 12.45 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

32.0

,

δ

(CH₂), 35.4 (CH₂), 38.0 (CH₂), 68.9 (CH₂), 116.6 (CH), 117.7 (C), 122.2 (C), 122.4 (CH), 126.9

(CH), 128.9 (CH), 129.5 (CH), 131.1 (CH), 133.4 (CH), 134.9 (CH), 138.8 (C), 159.4 (C), 165.9

(C), 167.6 (C), 172.4 (C). Anal. (C₂₁H₁₉NO₅S) calcd: C 63.46; H 4.82; N 3.52; found: C 63.66;

H 4.89; N 3.48

3-[(5Z)-2,4-Dioxo-{5-[4-(2-phenylethoxy)phenyl]methylidene}thiazolidin-3-yl]propanoic acid (3f).

◦

1

Yield 73% (487.4 mg); m.p. 160–163 C; H-NMR (DMSO-d₆):

δ 2.59 (t J = 7.2 Hz, 2H,

CH₂COOH); 3.06 (t J = 6.9 Hz, 2H, CH₂Ph); 3.85 (t J = 7.2 Hz, 2H, NCH₂); 4.28 (t J = 6.9 Hz,

2H, OCH₂); 7.11 (m, 2H, arom); 7.23 (m, 1H, arom); 7.30–7.34 (m, 4H, arom); 7.57 (m, 2H,

arom); 7.87 (s, 1H, methylidene); 12.44 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ 32.1 (CH₂),

35.3 (CH₂), 37.9 (CH₂), 69.0 (CH₂), 116.0 (CH), 118.5 (C), 126.0 (C), 126.9 (CH), 128.9 (CH),

129.5 (CH), 132.8 (CH), 133.5 (CH), 138.6 (C), 160.9 (C), 166.1 (C), 167.7 (C), 172.4 (C). Anal.

(C₂₁H₁₉NO₅S) calcd: C 63.46; H 4.82; N 3.52; found: C 63.55; H 4.80; N 3.54.

3.1.3. General Procedure for the Synthesis of 3-(5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)

propanoic Acids 4a–f

To a solution of β-alanine (1 g, 12.49 mmol) and NaOH (0.5 g, 12.49 mmol) in H₂O

(35 mL) CS₂(2.25 mL, 37.46 mmol) was added dropwise, in such a manner that the temper-

◦

ature of the reaction did not exceed 25 C and the mixture was stirred at room temperature

for 24 h. A solution of sodium bromoacetate, obtained by solubilizing bromoacetic acid

(1.73 g, 12.49 mmol) and NaOH (0.5 g, 12.49 mmol) in H₂O (45 mL), was added and the

mixture was stirred at room temperature for a further 24 h. The mixture was acidiﬁed

with HCl 6M until pH 3–4 and reﬂuxed for 24 h. The mixture was then cooled and poured

into H₂O, providing a pale yellow precipitate. Then, if necessary, pH was adjusted again

to 3–4 by the addition of HCl 6M and the solid was ﬁltered off. The aqueous solution

was extracted with ethyl acetate; the organic phase was washed with H₂O, dried with

anhydrous Na₂SO₄and evaporated under reduced pressure, providing a further amount of

3-(4-oxo-2-thioxothiazolidin-3-yl)propanoic acid (

were collected and washed with ethyl ether to give pure compound

mp 159–162 ^◦C; ¹H-NMR (CDCl₃):

2.77 (t J = 6 Hz, 2H, CH₂COOH), 4.01 (s, 2H, 5-CH₂)

7). The solid residue and the precipitate

7

. Yield 35% (897.2 mg)

;

δ

4.31 (t J = 6 Hz, 2H, NCH₂). Anal. (C₆H₇NO₃S₂) calcd: C 35.11; H 3.44; N 6.82; found: C

34.96, H 3.34, N 6.92.

A mixture of 3-(4-oxo-2-thioxothiazolidin-3-yl)propanoic acid (7) (0.25 g, 1.22 mmol)

glacial acetic acid (10 mL), sodium acetate (1.24 g, 9.13 mmol) and the appropriate alde-

hyde (1.22 mmol) was reﬂuxed for 3–4 h. The mixture was cooled and poured into H₂O,

providing a precipitate which was ﬁltered off, washed with H₂O and recrystallized from

methanol to give pure 3-[(5-arylidene-4-oxo-2-thioxothiazolidin-3-yl)]propanoic acids 4.

3-{(5Z)[4-Oxo-5-(3-phenoxyphenyl)methylidene-2-thioxothiazolidin-3-yl]}propanoic acid (4a). Yield

◦

46% (216.3 mg); m.p. 174–175 C; ¹H-NMR (DMSO-d₆):

δ 2.57 (t J = 7.5 Hz, 2H, CH₂COOH);

4.15 (t J = 7.5 Hz, 2H, NCH₂); 7.01–7.03 (m, 3H, arom); 7.08 (m, 1H, arom); 7.18 (m, 1H,

arom); 7.27 (m, 1H, arom); 7.38–7.41 (m, 2H, arom); 7.49 (m, 1H, arom); 7.63 (s, 1H, CH

methylidene). ¹³C-NMR (DMSO-d₆):

δ 31.9 (CH₂), 41.1 (NCH₂), 120.0 (CH), 120.4 (CH),

120.5 (C), 121.8 (CH), 124.3 (CH), 125.5 (CH), 126.5 (CH), 131.4 (CH), 132.4 (CH), 133.3 (CH),

135.7 (C), 156.7 (C), 158.8 (C), 167.7 (C), 173.1 (C), 194.0 (C). Anal. (C₁₉H₁₅NO₄S₂) calcd: C

59.20; H 3.92; N 3.63; found: C 59.06; H 4.01; N 3.60.

3-{(5Z)[4-Oxo-5-(4-phenoxyphenyl)methylidene-2-thioxothiazolidin-3-yl]}propanoic acid (4b). Yield

◦

46% (216.3 mg); m.p. 218–219 C; ¹H-NMR (DMSO-d₆):

δ

2.62 (t J = 7.5 Hz, 2H, CH₂COOH);

4.22 (t J = 7.5 Hz, 2H, NCH₂); 7.11 (m, 3H, arom); 7.24 (m, 2H, arom); 7.45 (m, 2H, arom);

7.64 (m, 2H, arom); 7.76 (s, 1H, CH methylidene). ¹³C-NMR (DMSO-d₆):

31.7 (CH₂), 40.8

δ

Molecules 2021, 26, 330

25 of 32

(NCH₂), 119.2 (CH), 120.8 (CH), 121.4 (C), 125.7 (CH), 128.4 (CH), 131.2 (C), 133.4 (CH),

133.9 (CH), 155.8 (C), 160.3 (C), 167.6 (C), 172.7 (C), 194.0 (C). Anal. (C₁₉H₁₅NO₄S₂) calcd:

C 59.20; H 3.92; N 3.63; found: C 59.25; H 4.00; N 3.57.

3-{(5Z)-5-[(3-Benzyloxyphenyl)methylid_◦ene]-4-oxo-2-thioxothiazolidin-3-yl}propanoic acid (4c).

1

Yield 62% (302.2 mg); m.p. 207–209 C; H-NMR (DMSO-d₆):

δ 2.62 (t J = 7.5 Hz, 2H,

CH₂COOH); 4.21 (t J = 7.5 Hz, 2H, NCH₂); 5.16 (s, 2H, OCH₂); 7.15–7.19 (m, 3H, arom); 7.33

(m, 1H, arom); 7.39 (m, 2H, arom); 7.44–7.47 (m, 3H, arom); 7.73 (s, 1H, CH methylidene).

¹³C-NMR (DMSO-d₆):

δ 31.8 (CH₂), 41.0 (CH₂), 70.4 (CH₂), 117.3 (CH), 119.0 (C), 123.8

(CH), 124.0 (C), 128.7 (CH), 129.1 (CH), 129.6 (CH), 131.8 (CH), 133.9 (CH), 135.2 (CH),

137.5 (C), 159.7 (C), 167.7 (C), 172.9 (C), 194.2 (C). Anal. (C₂₀H₁₇NO₄S₂) calcd: C 60.13; H

4.29; N 3.51; found: C 59.97; H 4.18; N 3.61.

3-{(5Z)-5-[(4-Benzyloxyphenyl)methylidene]-4-oxo-2-thioxothiazolidin-3-yl}propanoic acid (4d).

◦

1

Yield 49% (238.8 mg); m.p. 170–173 C; H-NMR (DMSO-d₆):

CH₂COOH); 4.21 (t J = 7.5 Hz, 2H, NCH₂); 5.20 (s, 2H, OCH₂); 7.19 (m, 2H, arom); 7.34–7.48

(m, 5H, arom); 7.62 (m, 2H, arom); 7.78 (s, 1H, CH methylidene). ¹³C-NMR (DMSO-d₆):

δ 2.61 (t J = 7.5 Hz, 2H,

δ

31.5 (CH₂), 40.6 (CH₂), 70.2 (CH₂), 116.5 (CH), 119.8 (C), 126.3 (C), 128.4 (CH), 128.6 (CH),

129.1 (CH), 133.5 (CH), 133.7 (CH), 137.0 (C), 161.2 (C), 167.3 (C), 172.4 (C), 193.7 (C). Anal.

(C₂₀H₁₇NO₄S₂) calcd: C 60.13; H 4.29; N 3.51; found: C 60.17; H 4.35; N 3.45.

3-[(5Z)-4-Oxo-{5-[3-(2-phenylethoxy)phenyl]methylidene}-2-thioxothiazolidin-3-yl]propanoic acid

(4e). Yield 26% (131.2 mg); m.p. 136–138 ^◦C; ¹H-NMR (DMSO-d₆): δ 2.63 (t J = 7.5 Hz, 2H,

CH₂COOH); 3.05 (t J= 6.8 Hz, 2H, CH₂Ph); 4.20–4.26 (m, 4H, NCH₂and OCH₂); 7.08 (m,

1H, arom); 7.15–7.18 (m, 2H, arom); 7.22 (m, 1H, arom); 7.29–7.34 (m, 4H, arom); 7.45 (m,

1H, arom); 7.76 (s, 1H, CH methylidene). ¹³C-NMR (DMSO-d₆):

δ 31.5 (CH₂), 35.5 (CH₂),

40.7 (CH₂), 69.1 (CH₂), 117.0 (CH), 118.3 (C), 123.2 (C), 123.4 (CH), 127.1 (CH), 129.1 (CH),

129.7 (CH), 131.4 (CH), 133.7 (CH), 135.0 (CH), 138.9 (C), 159.6 (C), 167.3 (C), 172.5 (C),

193.9 (C). Anal. (C₂₁H₁₉NO₄S₂) calcd: C 61.00; H 4.63; N 3.39; found: C 61.18; H 4.74; N

3.25.

3-[(5Z)-4-Oxo-{5-[4-(2-phenylethoxy)phenyl]methylidene}-2-thioxothiazolidin-3-yl]propanoic acid

(

4f). Yield 26% (131.2 mg); m.p. 171–174 ^◦C; ¹H-NMR (DMSO-d₆):

δ 2.63 (t J = 7.5 Hz, 2H,

CH₂COOH); 3.05 (t J= 6.8 Hz, 2H, CH₂Ph); 4.23 (t J = 7.5 Hz, 2H, NCH2); 4.29 (t J = 6.8 Hz,

2H, OCH₂); 7.12 (m, 2H, arom); 7.23 (m, 1H, arom); 7.30–7.34 (m, 4H, arom); 7.59 (m, 2H,

arom); 7.77 (s, 1H, CH methylidene); 12.51 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ 31.4

(CH₂), 35.3 (CH₂), 40.6 (CH₂), 69.1 (CH₂), 116.2 (CH), 119.6 (C), 126.1 (C), 126.9 (CH), 128.9

(CH), 129.5 (CH), 133.5 (CH), 133.8 (CH), 138.6 (C), 161.3 (C), 167.3 (C), 172.3 (C), 193.7 (C).

Anal. (C₂₁H₁₉NO₄S₂) calcd: C 61.00; H 4.63; N 3.39; found: C 60.92; H 4.58; N 3.51.

3.1.4. General Procedure for the Synthesis of 4-(5-arylidene-2,4-dioxothiazolidin-3-yl)-2

-butenoic Acids 5a–e

A mixture of 5-arylidene-2,4-thiazolidinedione 6 (2.5 mmol), methyl 4-bromocrotonate

(0.9 g, 5 mmol) and potassium carbonate (0.69 g, 5 mmol) in anhydrous acetone (50 mL)

was reﬂuxed for 24 h. After cooling, the inorganic salts were ﬁltered off and the solvent

was evaporated under reduced pressure. The solid residue was washed with H₂O and

recrystallized from ethanol providing pure 4-(5-arylidene-2,4-dioxothiazolidin-3-yl)-2-

butenoic acid methyl esters.

A mixture of the corresponding methyl ester (2.1 mmol), glacial acetic acid (8.5 mL)

and HCl 12M (2.1 mL) was reﬂuxed for 2 h. The reaction mixture was poured into H₂O and

the precipitate was ﬁltered off. The crude solid was washed with H₂O and recrystallized

from methanol to provide pure 2-butenoic acids 5. Chemical-physical and spectroscopic

data of compounds 5a–c were reported in [26].

4-{(5Z)-[5-(1-Naphtylmethylidene)-2,4-dioxothiazolidin-3-yl]}-2-butenoic acid (5d). Yield 26%

(220.6 mg); m.p. 170–172 ^◦C; ¹H-NMR (DMSO-d₆):

4.47 (d J = 5.0 Hz, 2H, NCH₂); 5.90

(d J = 15.5 Hz, 1H, CHCOOH); 6.83 (dt J = 15.5 and 5.0 Hz, 1H, CHCH₂); 7.62–7.72 (m,

δ

Molecules 2021, 26, 330

26 of 32

4H, arom); 8.03–8.15 (m, 3H, arom); 8.57 (s, 1H, CH methylidene), 12.52 (s, 1H, COOH).

¹³C-NMR (DMSO-d₆):

δ 42.4 (CH₂), 123.7 (CH), 124.1 (CH), 125.6 (CH), 126.2 (CH), 127.1

(CH), 127.5 (CH), 128.1 (CH), 129.5 (CH), 130.7 (CH), 130.8 (C), 131.5 (C), 133.8 (C), 141.2

(CH), 165.4 (C), 167.0 (C), 167.9 (C), 171.3 (C). Anal. (C₁₈H₁₃NO₄S) calcd: C 63.71; H 3.86;

N 4.13; found: C 63.59; H 3.76; N 4.22.

4-{(5Z)-[5-(2-Naphtylmethylidene)-2,4-dioxothiazolidin-3-yl]}-2-butenoic acid (5e). Yield 70%

◦

(593.9 mg); m.p. 235–238 C; ¹H-NMR (DMSO-d₆):

δ 4.45 (d J = 5.0 Hz, 2H, NCH₂); 5.87 (d

J = 16.0 Hz, 1H, CHCOOH); 6.81 (dt J = 16.0 and 5.0 Hz, 1H, CHCH₂); 7.59–7.65 (m, 2H,

arom); 7.71 (m, 1H, arom); 7.97 (m, 1H, arom); 8.04–8.08 (m, 3H, arom); 8.22 (s, 1H, CH

methylidene); 12.49 (s, 1H, COOH). ¹³C-NMR (DMSO-d₆):

δ 42.4 (CH₂), 122.1 (C), 123.6

(CH), 126.5 (CH), 127.8 (CH), 128.3 (CH), 128.7 (CH), 129.4 (CH), 129.6 (CH), 131.1 (CH),

131.8 (CH), 133.3 (C), 133.7 (C), 133.9 (C), 141.2 (CH), 165.8 (C), 167.0 (C), 167.7 (C). Anal.

(C₁₈H₁₃NO₄S) calcd: C 63.71; H 3.86; N 4.133; found: C 63.78; H 3.94; N 4.07.

3.2. Enzymatic Assays

3.2.1. AR Enzymatic Assay, Expression and Puriﬁcation

The human recombinant AR was expressed in BL21(DE3)pLysS E. coli cells and

puriﬁed to electrophoretic homogeneity as previously described [47]. The puriﬁed enzyme

(speciﬁc activity 5.0 U/mg) was stored at

−

80 ^◦C in 10 mM sodium phosphate buffer pH

7.0 containing 2 mM dithiothreitol and 30% (w/v) glycerol. AR activity was determined at

37 ^◦C as previously described [48] evaluating the decrease in absorbance at 340 nm linked

to NADPH oxidation. The standard assay mixture contained 0.25 M sodium phosphate

buffer pH 6.8, 0.18 mM NADPH, 2.4 M ammonium sulfate, 0.5 mM EDTA and 4.7 mM

GAL. One unit of enzyme activity is the amount that catalyzes the conversion of 1

substrate/min in the above assay conditions.

µmol of

3.2.2. Inhibition Studies on AR

The determination of IC₅₀(concentration of compound required to determine a 50%

inhibition of enzyme activity) values was performed in the above described assay condi-

tions using 2 mM L-idose as substrate. All compounds tested as aldose reductase inhibitors

were dissolved in DMSO and added to the assay mixture containing 8 mU of puriﬁed AR.

The reaction was started by addition of the substrate. The DMSO concentration in all the

assays was kept constant at 0.5% (v/v) in order to avoid effects on AR activity [49].

The IC₅₀values were determined by nonlinear regression analysis using Prism Graph-

Pad 6.0 ﬁtting experimental data to the following equation:

v_i

v₀

Max − Min

=

+ Min

(1)

ꢀ

ꢁ

slope

I

1 +

IC₅₀

In the equation, v_i/v₀, represents the ratio between the activity measured in the

presence of the inhibitor and the activity measured in the absence of inhibitor; Max and Min

represent the expected maximal and minimal value of the activity and were ﬁxed at 1 and

zero, respectively. Slope, which describes the steepness of curve in the transition region,

was ﬁxed at

the triplicate assay were analyzed.

−

1. For each compound, at least ﬁve different concentrations of inhibitor in

The kinetic analysis of compounds 4a, 4e and 4f was performed by measuring reaction

rates with different L-idose concentrations s (S) in the absence and in the presence of

different inhibitors concentrations. Data were analyzed by Linewever-Burk plots. The

apparent dissociation constants K_i’ (for the ESI complex) and K_i(for the EI complex) were

determined from secondary plots of 1/V_maxand K_M/V_maxas a function of the inhibitor

concentration, respectively.

In order to evaluate the reversibility of the inhibitory action, 8 mU of puriﬁed AR were

assayed in the presence of 5

µM of compounds 4a, 4e and 4f. In these conditions less than

15% of the enzyme activity was measured in the presence of 2 mM L-idose as substrate. The

Molecules 2021, 26, 330

27 of 32

mixture was then extensively dialysed on Amicon ultraﬁltration membrane (cut off 10 KDa

against 10 mM sodium phosphate buffer, pH 7.0. After dialysis, the enzyme activity was

again measured as above and compared to that of a mixture in the absence of inhibitor and

treated in the same conditions.

)

3.2.3. Enzymatic Assays with PTP1B

PTP1B activity was determined as follow. An aliquot of human recombinant PTP1B

was diluted in the assay buffer containing 0.075 M β-β-dimethylglutarate pH 7.0, 1 mM

EDTA, 0.1 mM DTT and p-nitrophenyl phosphate (pNPP) as substrate. After an appropriate

interval time, the reaction was stopped diluting assay solution with 2 mL of KOH 0.1 M.

The amount of p-nitrophenol released was determined measuring the absorbance of the

solution at 400 nm using a spectrophotometer and a 1-cm optical pathlength (ε_mMof

p-nitrophenol is 18).

The IC₅₀value was determined measuring the pNPP hydrolysis rate in the presence

of increasing inhibitor concentration. For each inhibitor, 15–16 different concentrations

were used. All tests were carried out in triplicate and data obtained were normalized

respect to control sample. Experimental points were ﬁtted using Equation (1), shows in the

Section 3.2.2.

3.2.4. Reversibility Assay (Dilution Test) with PTP1B

◦

An aliquot of the enzyme was incubated for 60 min at 37 C in the presence of saturat-

ing amount of PTP1B inhibitor. After this time, an aliquot of the enzyme was withdrawn

and diluted 400 folds in the assay buffer containing 2.5 mM of pNPP. After 15 min, the

reaction was stopped by adding 1 mL of KOH 0.2 M. The amount of pNP released was

calculating measuring the absorbance at 400 nm of solutions using a spectrophotometer.

Data obtained were normalized respect to control tests, which was carried out diluting

the enzyme with DMSO, the solvent used to dissolve all compounds. Data reported in the

ﬁgure represent the mean value +/− S.E.M. (n = 3).

3.2.5. Evaluation of Action Mechanism of Inhibitors

Kinetic analyses were carried out to dissect action mechanism of compound 4e and

4f. Brieﬂy, the main kinetic parameters, Km and Vmax, were calculated measuring the

initial hydrolysis rate of PTP1B in the presence of increasing pNPP concentration (seven

different concentrations in the 0.5–25 mM range). All assays were carried out in duplicate.

Data obtained were ﬁtted using the Michaelis-Menten equation. To evaluate the action

mechanism of inhibitors, we studied the dependence of Km and Vmax from inhibitors

concentration, using three different inhibitor concentrations, at least. Data obtained were

then analysed by double reciprocal plot (Lineweaver-Burk plot).

3.3. Ex Vivo Assay

3.3.1. Cell Cultures

Human liver cells (HepG2) and murine myoblasts (C2C12) were purchased by Sigma-

Aldrich, which obtained them from the European Collection of Authenticated Cell Cultures

(ECACC). HepG2 and C2C12 cells were grown in Dulbecco’s Modiﬁed Eagle’s medium

(DMEM) supplemented with 10% foetal bovine serum (FBS), 2 mM glutamine, 100 IU/mL

penicillin, 100 µg/mL streptomycin (Sigma-Aldrich, St. Louis, MO, USA), in a humidiﬁed

atmosphere and with 5% CO₂at 37 ^◦C. Differentiation of C2C12 was induced incubating

murine myoblasts in DMEM supplemented with 2% horse serum. C2C12 cells were plated

and grown until they reach 70% conﬂuence. Then, cells were incubated in the presence

of 2% horse serum for 96 h. Every 48 h, the medium was replaced with fresh medium.

After 96 h, the degree of differentiation was evaluated analysing cells with a contrast phase

microscope.

Molecules 2021, 26, 330

28 of 32

3.3.2. Cell Viability Assay

HepG2 and C2C12 cells were incubated in the presence of increasing concentrations

◦

of compounds 4e and 4f for 48 h at 37 C. After this time, cells were washed with PBS

and then incubated for 1 h with 0.5 mg/mL of MTT dissolved in complete medium. Then

cells were washed and lysed with 300 µL of DMSO. The absorbance of solutions were

evaluated using a microplate readers at 595 nm. All data were normalized respect to control

experiment. Data showed in the ﬁgures represent the mean values ± S.E.M. (n = 4).

3.3.3. Evaluation of Insulin Mimetic/Sensibilizing Activity of Inhibitors

HepG2 and differentiated C2C12 cells were starved for 24 h and then incubated in

the presence of compound 4f (25 µM, ﬁnal concentration) for 90 min. After this time, cells

were washed with PBS and then stimulated or not with 10 nM insulin for 30 min. Then,

cells were lysed and cellular proteins separated on a 12% SDS-PAGE. After electrophoresis,

cellular proteins were transferred on a PVDF membrane by western blot technique and

then probed against antibodies able to speciﬁcally recognize insulin receptor (I2033 Sigma

Aldrich) or phosphorylated form of IR (MABS65, Sigma-Aldrich). Sample loading was

evaluated probing PVDF membrane with anti-actin antibodies (ABT1485 Sigma Aldrich).

3.4. Molecular Docking

Crystal structure selection and preparation for PTP1B (PDB 1Q6T [50]) and the AR-

idose complex (based on PDB 3V36 [51]) was carried out as described in our previous

study [5]. Docking studies were performed into the catalytic site of PTP1B and AR, the

AR-idose complex and our previously described allosteric binding pocket of PTP1B [

5

]

with GOLD (version 5.2) [52], generating 25 docking poses per ligand with standard

settings using the scoring function CHEMPLP. In order to select a docking pose, the 25

docking poses were clustered based on LigandScouts [53,54] pharmacophore RDF-Code

(radial distribution function) similarity. Finally, all selected docking poses have been energy

minimized in Ligand Scout with the MMFF94 Force Field [55]. 3D protein ligand depictions

were rendered with MOE (version 2019) [56] and 3D pharmacophores were developed

using Ligand Scout.

Virtual Screening

Shape-based virtual screening was performed with ROCS (OpenEye, Scientiﬁc Soft-

ware, Santa Fe, NM, USA) [57] using default options. Docking of virtual screening hits

was carried out with the software package GOLD (version 5.2) [52] into the catalytic cen-

tre of PTP1B (PDB 1Q6T [50]), generating 10 docking poses per ligand with standard

settings as described above. The hit compound 3-{[5-(4-benzyloxyphenyl)methylidene]-

2,4-dioxothiazolidin-3-yl}propanoic acid was identiﬁed by using the co-crystallized ligand

(4-bromo-3-(carboxymethoxy)-5-[3-(cyclohexylamino)phenyl]thiophene-2-carboxylic acid)

from the PTP1B crystal structure 2QBS [58] as the template for the ROCS screening. The

compound was found screening the vendor space of Enamine HTS Collection (as of

06/2014).

4. Conclusions

The appreciable in vitro activity of compounds 4a, 4e and 4f as well as 5a and 5d

as dual AR/PTP1B inhibitors provided further insights into the possibility to develop

4-thiazolidinone derivatives as new potential multitarget antidiabetic candidates. Taking

into account that the partial inhibition of selected targets, in particular enzymes, may

be therapeutically more effective than the total control of a single target [2,3], these dual

inhibitors, along with previous analogues 1e and 2e, appear to be promising as lead

compounds, whose inhibitory activity towards both target enzymes might be further im-

proved. In particular, compound 4f exhibited the most interesting proﬁle, with balanced

AR/PTP1B inhibitory effectiveness at low micromolar concentrations, along with promis-

ing insulin-sensitizing activity in cultured cells. The IC₅₀values of 4f against both human

Molecules 2021, 26, 330

29 of 32

PTP1B and human AR resulted in being in the low micromolar range, whereas those of

the corresponding acetic acid analogue of series

2 differ of two orders of magnitude [5].

Therefore, although the elongation of the carboxylic chain on N-3 of the thiazolidinone

scaffold produced a decrease in AR inhibitory effectiveness, this structural feature was

responsible for a gain in potency against PTP1B, thus providing a more balanced inhibition

proﬁle towards the two target enzymes.

The SARs that emerged from our enzyme inhibition data, also considering our previous

findings, highlighted that the AR inhibitory potency of the investigated 4-thiazolidinones

is mainly inﬂuenced by the nature of the substituents in positions 2 and 3 of the hete-

rocyclic scaffold and only in a lesser extent by the nature of the 5-arylidene moiety. On

the other hand, this latter structural portion appeared to be critical to achieving good

PTP1B inhibition, considering that a more extended 5-arylidene portion, in particular

a (2-phenylethoxy)benzylidene moiety, showed to be more beneﬁcial. In addition, a

thiocarbonyl group in position 2 of the thiazolidinone scaffold generally proved to be

advantageous to achieve gain in potency towards both target enzymes. The elongation

of the carboxylic chain on N-3 in several cases improved inhibitory effectiveness towards

PTP1B; propanoic acid derivatives 4a, 4e and 4f as well as 2-butenoic acid derivatives 5a

and 5d exhibited good capability to inhibit both human PTP1B and AR. The data reported

here allowed us to extend the SARs of this class of DMLs, which were rationalized by

in silico docking experiments into both target enzymes. Therefore, these ﬁndings can be

of interest for the design of new 4-thiazolidinone derivatives endowed with improved

pharmacological proﬁles.

Supplementary Materials: The supplementary materials are available online.

Author Contributions: Conceptualization, R.O., R.M. (Rosanna Maccari), A.D.C., T.N.N., A.N. and

G.W.; Synthetic methodologies R.O., R.M. (Rosanna Maccari) and I.A.; Validation & Investigation,

R.O., R.M. (Rosanna Maccari), T.N.N., A.N., G.W., M.C. and R.M. (Roberta Moschini); Data Curation,

I.A., M.C., M.G., I.N., T.N.N. and A.N.; Writing-review editing R.O., R.M. (Rosanna Maccari), A.D.C.,

P.P., T.N.N. and G.W. All authors have read and agreed to the published version of the manuscript.

Funding: The APC was funded by MDPI.

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Data Availability Statement: The data presented in this study are available in this article and relative

supplementary material.

Acknowledgments: This work was supported by University of Messina, University of Florence and

University of Pisa.

Conﬂicts of Interest: The authors declare no conﬂict of interest.

Sample Availability: Samples of the compounds are not available from the authors.

References

1.

Morphy, R.; Rankovic, Z. Designed Multiple Ligands. An Emerging Drug Discovery Paradigm. J. Med. Chem. 2005, 48, 6523–6543.

[CrossRef]

2.

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. [CrossRef]

3.

4.

Hopkins, A.L. Network pharmacology: The next paradigm in drug discovery. Nat. Chem. Biol. 2008, 4, 682–690. [CrossRef]

Proschak, E.; Stark, H.; Merk, D. Polypharmacology by Design: A Medicinal Chemist’s Perspective on Multitargeting Compounds.

J. Med. Chem. 2019, 62, 420–444. [CrossRef]

5.

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 Med. Chem. Lett. 2018, 28,

3712–3720. [CrossRef] [PubMed]

6.

Cho, N.H.; Shaw, J.E.; Karuranga, S.; Huang, Y.; Fernandes, J.D.R.; Ohlrogge, A.; Malanda, B. IDF Diabetes Atlas: Global estimates

of diabetes prevalence for 2017 and projections for 2045. Diabetes Res. Clin. Pract. 2018, 138, 271–281. [CrossRef] [PubMed]

Molecules 2021, 26, 330

30 of 32

7.

Saeedi, P.; Petersohn, I.; Salpea, P.; Malanda, B.; Karuranga, S.; Unwin, N.; Colagiuri, S.; Guariguata, L.; Motala, A.A.; Ogurtsova,

K.; et al. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: Results from the

International Diabetes Federation Diabetes Atlas, 9th edition. Diabetes Res. Clin. Pract. 2019, 157, 107843. [CrossRef] [PubMed]

Gattrell, W.; Johnstone, C.; Patel, S.; Smith, C.S.; Scheel, A.; Schindler, M. Designed multiple ligands in metabolic disease research:

From concept to platform. Drug Discov. Today 2013, 18, 692–696. [CrossRef] [PubMed]

8.

9.

Dowarah, J.; Singh, V.P. Anti-diabetic drugs recent approaches and advancements. Bioorganic Med. Chem. 2020, 28, 115263.

[CrossRef] [PubMed]

10. Srivastava, S.K.; Ramana, K.V.; Bhatnagar, A. Role of Aldose Reductase and Oxidative Damage in Diabetes and the Consequent

Potential for Therapeutic Options. Endocr. Rev. 2005, 26, 380–392. [CrossRef]

11. 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. [CrossRef] [PubMed]

12. Ramana, K.V.; Srivastava, S.K. Aldose reductase: A novel therapeutic target for inﬂammatory pathologies. Int. J. Biochem. Cell

Biol. 2010, 42, 17–20. [CrossRef] [PubMed]

13. Frohnert, B.I.; Long, E.K.; Hahn, W.S.; Bernlohr, D.A. Glutathionylated Lipid Aldehydes Are Products of Adipocyte Oxidative

Stress and Activators of Macrophage Inﬂammation. Diabetes 2013, 63, 89–100. [CrossRef] [PubMed]

14. Di Paola, R.; Frittitta, L.; Miscio, G.; Bozzali, M.; Baratta, R.; Centra, M.; Spampinato, D.; Santagati, M.G.; Ercolino, T.; Cisternino,

0

C.; et al. A Variation in 3 UTR of hPTP1B Increases Speciﬁc Gene Expression and Associates with Insulin Resistance. Am. J. Hum.

Genet. 2002, 70, 806–812. [CrossRef] [PubMed]

15. Cheng, A.; Uetani, N.; Simoncic, P.D.; Chaubey, V.P.; Lee-Loy, A.; McGlade, C.J.; Kennedy, B.P.; Tremblay, M.L. Attenuation of

leptin action and regulation of obesity by protein tyrosine phosphatase 1B. Dev. Cell 2002, 2, 497–503. [CrossRef]

16. Zhang, Z.-Y.; Dodd, G.T.; Tiganis, T. Protein Tyrosine Phosphatases in Hypothalamic Insulin and Leptin Signaling. Trends

Pharmacol. Sci. 2015, 36, 661–674. [CrossRef]

17. Zabolotny, J.M.; Kim, Y.-B.; Welsh, L.A.; Kershaw, E.E.; Neel, B.G.; Kahn, B.B. Protein-tyrosine Phosphatase 1B Expression Is

Induced by Inﬂammation in Vivo. J. Biol. Chem. 2008, 283, 14230–14241. [CrossRef]

18. Elchebly, M. Increased Insulin Sensitivity and Obesity Resistance in Mice Lacking the Protein Tyrosine Phosphatase-1B Gene.

Science 1999, 283, 1544–1548. [CrossRef]

19. Klaman, L.D.; Boss, O.; Peroni, O.D.; Kim, J.K.; Martino, J.L.; Zabolotny, J.M.; Moghal, N.; Lubkin, M.; Kim, Y.-B.; Sharpe, A.H.;

et al. Increased Energy Expenditure, Decreased Adiposity and Tissue-Speciﬁc Insulin Sensitivity in Protein-Tyrosine Phosphatase

1B-Deﬁcient Mice. Mol. Cell. Biol. 2000, 20, 5479–5489. [CrossRef]

20. Combs, A.P. Recent Advances in the Discovery of Competitive Protein Tyrosine Phosphatase 1B Inhibitors for the Treatment of

Diabetes, Obesity and Cancer. J. Med. Chem. 2010, 53, 2333–2344. [CrossRef]

21. Lantz, K.A.; Hart, S.G.E.; Planey, S.L.; Roitman, M.F.; Ruiz-White, I.A.; Wolfe, H.R.; McLane, M.P. Inhibition of PTP1B by

Trodusquemine (MSI-1436) Causes Fat-speciﬁc Weight Loss in Diet-induced Obese Mice. Obesity 2010, 18, 1516–1523. [CrossRef]

[PubMed]

22. Hidalgo-Figueroa, S.; Estrada-Soto, S.; Ramírez-Espinosa, J.J.; Paoli, P.; Lori, G.; León-Rivera, I.; Navarrete-Vazquez, G. Synthesis

and evaluation of thiazolidine-2,4-dione/benzazole derivatives as inhibitors of protein tyrosine phosphatase 1B (PTP-1B):

Antihyperglycemic activity with molecular docking study. Biomed. Pharmacother. 2018, 107, 1302–1310. [CrossRef] [PubMed]

23. Singh, B.; Silakari, O.; Kaur, M.; Singh, B. Integrated pharmacophore and docking-based designing of dual inhibitors of aldose

reductase (ALR2) and protein tyrosine phosphatase 1B (PTP1B) as novel therapeutics for insulin-resistant diabetes and its

complications. J. Chemom. 2015, 29, 109–125. [CrossRef]

24. Kong, W.-J.; Vernierib, C.; Foiani, M.; Jiang, J. Berberine in the treatment of metabolism-related chronic diseases: A drug cloud

(dCloud) effect to target multifactorial disorders. Pharmacol. Ther. 2020, 209, 107496. [CrossRef] [PubMed]

25. Bruno, G.; Costantino, L.; Curinga, C.; Maccari, R.; Monforte, F.; Nicolò, F.; Ottanà, R.; Vigorita, M. Synthesis and aldose reductase

inhibitory activity of 5-arylidene-2,4-thiazolidinediones. Bioorganic Med. Chem. 2002, 10, 1077–1084. [CrossRef]

26. Maccari, R.; Ottanà, R.; Curinga, C.; Vigorita, M.G.; Rakowitz, D.; Steindl, T.; Langer, T. Structure–activity relationships and

molecular modelling of 5-arylidene-2,4-thiazolidinediones active as aldose reductase inhibitors. Bioorganic Med. Chem. 2005, 13,

2809–2823. [CrossRef] [PubMed]

27. Maccari, R.; Ottanà, R.; Ciurleo, R.; Vigorita, M.; Rakowitz, D.; Steindl, T.; Langer, T. Evaluation of in vitro aldose redutase

inhibitory activity of 5-arylidene-2,4-thiazolidinediones. Bioorganic Med. Chem. Lett. 2007, 17, 3886–3893. [CrossRef]

28. 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.

Bioorganic Med. Chem. 2008, 16, 5840–5852. [CrossRef]

29. 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. Bioorganic Med. Chem. 2010, 18, 4049–4055. [CrossRef]

30. Ottanà, R.; Maccari, R.; Giglio, M.; Del Corso, A.; Cappiello, M.; Mura, U.; Cosconati, S.; Marinelli, L.; Novellino, E.; Sartini, S.;

et al. 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. [CrossRef]

31. 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. Bioorganic Med. Chem. Lett. 2011, 21, 200–203. [CrossRef] [PubMed]

Article Doi

In search for multi-target ligands as potential agents for diabetes mellitus and its complications—a structure-activity relationship study on inhibitors of aldose reductase and protein tyrosine phosphatase 1b

DOI: 10.3390/molecules26020330

Source and publish data:

Authors:

Article abstract of DOI:10.3390/molecules26020330

Full text of DOI:10.3390/molecules26020330

Products guided by the article

R&D Labs maybe for 438190-31-9

Relevant to this article

Hot Product