Molecular modelling and synthesis of spiroimidazolidine-2,4-diones with dual activities as hypoglycemic agents and selective inhibitors of aldose reductase

Article abstract of DOI:10.1016/j.bioorg.2018.04.028

Novel derivatives of spiroimidazolidinedione were synthesized and evaluated as hypoglycemic agents through binding to sulfonylurea receptor 1 (SUR1) in pancreatic beta-cells. Their selectivity index was calculated against both aldehyde reductase (ALR1) an

Full text of DOI:10.1016/j.bioorg.2018.04.028

Accepted Manuscript
Molecular modelling and Synthesis of Spiroimidazolidine-2,4-diones with Dual
Activities as Hypoglycemic Agents and Selective Inhibitors of Aldose Reduc-
tase
Manar G. Salem, Yasmine M. Abdel Aziz, Marwa Elewa, Hosam A. Elshihawy,
Mohamed M. Said
PII:
S0045-2068(18)30281-5
DOI:
https://doi.org/10.1016/j.bioorg.2018.04.028
YBIOO 2351
Reference:
Bioorganic Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
22 March 2018
30 April 2018
30 April 2018
Please cite this article as: M.G. Salem, Y.M. Abdel Aziz, M. Elewa, H.A. Elshihawy, M.M. Said, Molecular
modelling and Synthesis of Spiroimidazolidine-2,4-diones with Dual Activities as Hypoglycemic Agents and
Selective Inhibitors of Aldose Reductase, Bioorganic Chemistry (2018), doi: https://doi.org/10.1016/j.bioorg.
2018.04.028
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.

Molecular modelling and Synthesis of
Spiroimidazolidine-2,4-diones with Dual Activities
as Hypoglycemic Agents and Selective Inhibitors of
Aldose Reductase
Manar G. Salem, Yasmine M. Abdel Aziz, Marwa Elewa*, Hosam A. Elshihawy and Mohamed
M. Said
Pharmaceutical Organic Chemistry Department, Faculty of Pharmacy, Suez Canal University,
P.O. 41522, Ismailia, Egypt
Corresponding Author
* marwa_elewa@pharm.suez.edu.eg, phone: +201226688211, fax: +20643230741.
1

ABSTRACT
Novel derivatives of spiro-imidazolidinedione were synthesized and evaluated as hypoglycemic
agents through binding to sulfonylurea receptor 1 (SUR1) in pancreatic beta-cells. Their
selectivity index was calculated against both aldehyde reductase (ALR1) and aldose reductase
(ALR2). Aldehyde reductase is a key enzyme in the polyol pathway that is involved in the
etiology of the secondary diabetic complications. All structures were confirmed by microanalysis
1
13
and by IR, H-NMR, C-NMR and EI-MS spectroscopy. The investigated compounds were
subjected to molecular docking and an in silico prediction study to determine their free energy of
binding (∆G) values and predict their physicochemical properties and drug-likeness scores.
Compound 1'-(5-chlorothiophene-2-ylsulfonyl)spiro[cyclohexane-1,5'-imidazolidine]-2',4'-dione
showed IC50 0.47 µM and 79 % reduction in blood glucose level with a selectivity index 127 for
ALR2.
KEYWORDS
Spiro-imidazolidinedione, Hypoglycaemic agents, Aldose reductase, In Silico Prediction, Mixed
pharmacophore.
2

1. INTRODUCTION
Diabetes mellitus is a metabolic disease characterized by a chronic increase in blood glucose
level. This is mainly attributed to impaired insulin secretion by pancreatic β cells and insulin
resistance in peripheral target tissues.1,2 Sulfonylureas are common in the design of oral
hypoglycemic agents such as tolazamide 1 and glibenclamide 2.3,4 These derivatives play a
crucial role in the improvement of insulin secretion and the reduction of insulin resistance. They
exert their action via several mechanisms including pancreatic ones through increasing insulin
secretion from pancreatic beta-cells and the others are extra-pancreatic by improving insulin
resistance in peripheral target tissues.5,6 Despite numerous trials to control hyperglycemia, most
diabetic patients still suffer from one or more of the long-term diabetic complications including
retinopathy, nephropathy, neuropathy and cataract.7,8 The enlarged glucose flux during the polyol
pathway, which occurs in hyperglycemic conditions, is a potential factor in the onset and
progression of such chronic complications (Figure 1).9,10 Aldose reductase enzyme (ALR2,
E.C.1.1.1.21) is the first and key enzyme involved in the polyol pathway. The main role of ALR2
is to catalyze the NADPH-dependent reduction of glucose to sorbitol. Aldose reductase
inhibitors (ARIs), therefore, could be considered as a pharmacologically direct management for
diabetic complications.11
3

Figure 1. A schematic interaction diagram elucidating the metabolism of glucose to sorbitol in
polyol pathway by aldose reductase (AR) under hyperglycemic conditions.11 AGE: Advanced
glycated end product; AR: Aldose reductase; GSH: Glutathione; SDH: Sorbitol dehydrogenase.
Various compounds have been designed to inhibit aldose reductase (AR).12 These compounds
can be classified into two main categories, the first category comprises those containing a
carboxylic acid moiety, for example, 3-thiazolidineacetic acid derivative, which has been
reported to be a potent ARI with IC50 9 nM.13 The second category includes those based on a
cyclic imide in the form of spirohydantoin such as sorbinil 3.11 The carboxylic acid inhibitors
were found to be less potent than the spirohydantoin inhibitors in vivo.14 This might be due to the
higher ability of the carboxylic acid derivatives to ionize at physiological pH. Subsequently,
these derivatives could not easily cross the biological membranes. Moreover, the selectivity of a
specific aldose reductase (ALR) is crucial to avoid the undesirable side effects of non-selective
4

inhibitors. Thus, the medication of choice should discriminate between aldehyde reductase
(ALR1, EC 1.1.1.2) and aldose reductase (ALR2, EC 1.1.1.21), since, ALR1 is responsible for
an important detoxification mechanism in the liver.15,16 However, the structure of ALR1 has 65 %
homology with ALR2, however some differences in their respective active sites provide the basis
for their selective inhibition.17,18 The present work has aimed to design and synthesize a single
medication that can control the plasma glucose level and also reduce adjunctive secondary
diabetic complications via applying a mixed pharmacophore theory (Figure 2).19,20
A literature survey showed that, in 2013, Iqbal et al reported the design and synthesis of 3'-(4-
chlorophenylsulfonyl)spiro[cycloheptane-1,5'-imidazolidine]-2',4'-dione 4 with an IC50 value of
1.8 μM against ALR2.21 In 2015, they also found that 3'-(4-chlorophenylsulfonyl)spiro[fluorene-
9,5'-imidazolidine]-2',4'-dione 5 was 53-fold more selective toward ALR2 (with an IC50 value of
0.89 µM) compared with ALR1.5 In 2017, Andleeb et al reported the design and synthesis of
(Z)-4-((3-(2methoxyphenyl)-4-oxo-2-thioxothiazolidin-5-ylidene)methyl)phenyl-4-
chlorobenzene-sulfonate 6 as an aldose reductase inhibitor with an IC50 value of 0.46 µM and 93-
fold selectivity for ALR2 over ALR1 with no apparent potential for hypoglycemic activity.14
Herein we assumed the optimization of the structure of these valuable candidates; compounds 4
and 5. Therefore, this study has been focused on modifying spirohydantoin derivatives based on
the enhancement of the binding affinity of the newly synthesized compound and improvement of
its physicochemical properties according to Lipinski's "rule of five""Ro5".22-24 This was applied
through two main strategies. Strategy one was the design of various hydrophobic moieties
attached to the spirohydantoin core particularly the flexible cyclohexanonyl ring and the rigid
fluorenonyl one. Strategy two was studying different aryl and heteroaryl sulfonyl moieties and
exploring their biological influence.
5

Figure 2. Mixed pharmacophore theory. The red box highlights the minimum requirements for
enhanced hypoglycemic activity by sulfonylureas. The green box highlights minimal
pharmacophoric requirements for an aldose reductase inhibitor. The violet rings represent the
points of diversity.
6

2. RESULTS AND DISCUSSION
2.1. Chemistry
Spiro-imidazolidinedione derivatives II and VI were prepared as previously reported.5,25 The
synthetic procedures involved the reaction of 9-fluorenone I (scheme 1, series A) and
cyclohexanone V (scheme 2, series B) with ammonium carbonate (NH4CO3) and potassium
cyanide (KCN) in 50 % aq. Ethanol.25 3-Arylsulfonyl derivatives IIIa and IIIb (series A) and
VIIa–d (series B), respectively, were prepared via the coupling of compounds II and VI with
various arylsulfonyl chlorides (ArSO2Cl) in triethylamine (Et3N) and dimethylaminopyridine
(DMAP) in CH2Cl2.26 Compounds IIIa, IIIb and VIIa–d were further converted to their 1-
arylsulfonyl analogues IVa, IVb and VIIIa–d, respectively using sodium hydride (NaH) under
dry conditions (schemes 1 and 2).27
7

Scheme 1. Synthesis of spiro [fluorene-9,5'-imidazolidine]-2',4'-dione derivatives (Series A).
Reagents and conditions: (i) NH4CO3, KCN, 55 °C and aq. ethanol; (ii) ArSO2Cl, Et3N and
DMAP; (iii) NaH.
8

Scheme 2. Synthesis of spiro[cyclohexane- 1,5'-imidazolidine]-2',4'-dione derivatives (Series B).
Reagents and conditions: (i) NH4CO3, KCN, 55 °C and aq. ethanol; (ii) ArSO2Cl, Et3N and
DMAP; (iii) NaH.
9

The structures of the newly synthesized compounds were confirmed by microanalyses and
1
13
spectral data (IR, H-NMR, C-NMR and EI-MS), which showed full agreement with their
structures. The IR spectra showed bands for anti-symmetric and symmetric O=S=O absorptions
in the range of 1350–1322 cm-1 and 1175–1140 cm-1, respectively. The presence of a band for the
N-H stretching in the range of 3400–3200 cm-1 also indicated the successful coupling between
the aryl sulfonyl chloride derivatives and spirohydantoin compounds resulting in the synthesis of
3-arylsulfonylspiroimidazolidine-2,4-diones, IIIa and IIIb (series A) and VIIa–d (series B). The
1H-NMR spectra of compounds IIIa, IIIb and VIIa–d displayed singlets at 5.70-6.00 ppm. This
peak referred to only one imidazolidinedione NH group and confirmed the successful coupling
of the other NH moiety. The characteristic multiplets assigned to fluorenonyl protons appeared
downfield at 7.15-8.08 ppm and hexanonyl protons upfield at 1.39-2.08 ppm. Compounds IIIa
and IVa showed doublets assigned to thienyl protons at 6.41-7.04 ppm. Moreover, compounds
IIIb and IVb displayed doublets assigned to the phenyl protons at 8.09-8.20 ppm. The 13C-NMR
spectra showed signals for spiro[fluorene-9,5'-imidazolidine]-2',4'-dione, phenyl, thienyl and
spiro[cyclohexan-1,5'-imidazolidine]-2',4'-dione rings at the expected regions. Additionally, the
mass spectra of compounds IIIa, IIIb, IVa and IVb (series A) and for compounds VIIa, VIIb,
VIIIa and VIIIb (series B) showed the presence of molecular ion peaks M+ and their isotopes
1
M++2, characteristic for sulfur and halogen-containing compounds. The H-NMR spectra of
compounds IVa and IVb and compounds VIIIa–d confirmed the successful rearrangement by
the appearance of the downfield singlets at 12.16-12.55 ppm assigned to the NH group at
position 3. This was accompanied by the disappearance of the upfield singlets corresponding to
the NH group at position 1 observed in the spectra of compounds IIIa, IIIb and compounds
VIIa–d, respectively. Regarding series B, compounds VIIb–d and VIIIb–d displayed doublets
1
assigned to the phenyl protons at 7.49-7.96 ppm. Furthermore, the H-NMR spectra of
10

compounds VIIc and VIIIc are characterized by the appearance of singlets at 2.30 and 2.19 ppm,
respectively, assigned to the protons of the methyl group. Concerning compounds VIId and
1
VIIId, their H-NMR spectra are characterized by the appearance of singlets at 3.90 and 3.80
ppm, respectively, referred to the protons of the methoxy groups.
2.2. Biological evaluation
All newly synthesized compounds were evaluated for in vivo antidiabetic activity and the
inhibitory activity against ALR1 (EC 1.1.1.2) and ALR2 (EC 1.1.1.21). The selectivity indexes
of the new compounds were determined. The results revealed that the newly synthesized
compounds probably serve as a single medication with a mixed pharmacophore not only to treat
diabetes but also to avoid diabetic complications.
2.2.1. In vivo hypoglycemic activity
In vivo study was performed on BALB/c mice after induction of diabetes and the blood
glucose level was measured through collecting the blood samples at 0 h followed by 2, 4, 6 and 8
h.5 The majority of the synthesized compounds of series A and B were found effective
antidiabetic agents (Table 1). Concerning series A, compounds IIIb and IVb displayed the
highest activities with 82 and 74 % reduction in blood glucose levels, respectively. They were
more potent than glibenclamide 2 (63 % reduction in blood glucose level) and the reported active
compound 3'-(4-chlorophenylsulfonyl)spiro[fluorene-9,5'-imidazolidine]-2',4'-dione 5 (72 %
reduction in blood glucose level). Both of the compounds contain a halogen atom (F) at the
aromatic ring (phenyl). The comparison between compounds IIIa and IVa revelled that the
activity increased to a great extent (from 52 % to 70 %) after the rearrangement step. Concerning
series B, the most active compound was VIIIa with 80 % reduction in blood glucose level. It
was more potent than glibenclamide 2 (63 %) and the reported active compound 3'-(4-
chlorophenylsulfonyl)spiro[cycloheptane-1,5'-imidazolidine]-2',4'-dione 4 (68 %). Compound
11

VIIIc was also effective in reducing blood glucose level (69 %) and was more potent than the
standard drug. Regarding compound VIIIc, the activity was not much affected by replacing the
halogen atom with methyl group. This might be attributed to the moderate size and lipophilicity
of the methyl group compared with the halogen atom. Compound VIId was ineffective in
lowering plasma glucose level. This negative effect might be attributed to the relatively lower
lipophilic property of the methoxy group.
Table 1. Blood glucose concentration (mg dL-1) in diabetic mice treated with series A and B
compounds.
Compounds Glucose Glucose
Glucose
mg/dl
Glucose Glucose Inhibition
after 8 h
mg/dl
basal
mg/dl
mg/dl
mg/dl
(%)
After 2h
After 4h
After 6h After 8h
Series A
IIIa
324±4.57 255±3.78
204±5.08
187±4.62 155±2.47 52.1
385
5.69
± 314±3.59
284 ± 3.78 214
2.74
± 71 ± 1.51 81.6
IIIb
421±3.28 315±5.57
362±7.49 225±2.67
235±8.36
184±5.82
179±5.11 125±8.45 70.4
134±3.69 93 ± 2.43 74.3
IVa
IVb
Series B
VIIa
VIIb
VIIc
451±4.56 352±4.98
398±8.23 317±7.93
366±4.77 276±8.41
321±6.75 275±7.55
414±7.89 278±6.34
356±8.89 269±9.37
274±5.43
282±5.57
221±3.86
194±9.36
192±5.12
228±7.23
204±1.87 154±4.78 65.9
221±4.91 143±5.33 64.1
198±6.78 141±2.24 61.4
219±2.61 254±8.45 ------
125±2.76 83 ± 3.26 79.9
206±6.98 156±5.36 56.2
VIId
VIIIa
VIIIb
12

391±8.55 284±8.34
408±7.93 364±7.58
217±5.86
261±9.22
181±7.33 121±2.44 69.1
213±4.71 161±5.89 60.5
VIIIc
VIIId
4a
67.6d
128±8.95
395
287 ±4.63 280 ±2.23 143
±6.81
c
±3.53
5b
245.0±4. 232.0±4.7 160.4±5.8 78.8±2.6 68.0±1.0 72.24
34
2
3
3
0
Glibencla-
mide
374±8.13 312±9.64
254±4.78
203±6.55 149±7.75 62.8
a: The reported results of compound 4 by Z. Iqbal et al.21
b: The reported results of compound 5 by Z. Iqbal et al.5
c: Plasma glucose concentration (mg/dl-1) After 7 h. by Z. Iqbal et al.21
d: Inhibition after 7 h (%).
Values represent mean ± SEM.
2.2.2. ALR1 and ALR2 activity
Inhibitory activities of the newly synthesized compounds were screened in vitro against ALR1
(the positive control was valproic acid)5 and ALR2 (the positive control was sorbinil 3)11 based
on estimating the decrease in the UV absorption by NADPH at 340 nm using UV
spectrophotometer.28 Most of the new compounds were found more potent towards ALR2 than
ALR1 (Table 2). Among the compounds of series A, compound VIIIa was the most potent
inhibitor of ALR2 with an IC50 value 0.47 µM. Compound VIIIa was more potent than the
standard drug sorbinil with an IC50 value 3.14 µM and was also found more potent than the
reported
compounds
3'-(4-chlorophenylsulfonyl)spiro[cycloheptane-1,5'-imidazolidine]-
2',4'-dione 4 and 3'-(4-chlorophenylsulfonyl)spiro[fluorene-9,5'-imidazolidine]-2',4'-dione 5 with
IC50 values 5.59 µM and 0.89 µM, respectively. A slight decrease in the activity was observed
when the thienyl group was replaced with a phenyl group in compound IVb with an IC50 value
13

1.62 µM. This reduction in the activity might be attributed to the relatively bigger size of the
phenyl group compared to the thienyl group, in addition to the bigger size and the relative
rigidity of the fluorenonyl moiety. The comparison between the newly synthesized compounds
before and after rearrangements revealed that the inhibitory activity was almost better after
rearrangement step. It has been shown that the arylsulfonyl group at position 1 of hydantoins
affects greatly the physicochemical properties and intestinal absorption of the drugs, and they are
more active than those with arylsulfonyl substituent at position 3.29 Similarly, compounds IIIa,
VIIa, VIId and VIIId exhibited a reasonable inhibition values to ALR1 and were more potent
than valproic acid. This probably due to the relatively higher lipophilic characters of
arylsulfonylspiroimidazolidine-2,4-diones if compared with 2-propylpentanoic acid moiety in the
standard valproic acid. The selectivity for the synthesized inhibitors was calculated using their
IC50 values against ALR1 versus ALR2. In series B, a remarkable selectivity was observed for
compound VIIIa. The selectivity index showed that compound VIIIa was 127- fold more
selective for ALR2 over ALR1. Compound VIIIa exhibited the maximum selectivity compared
with the reported active compounds 4, 5, and 6 (Figure 2). This remarkable selectivity towards
ALR2 is necessary to minimize toxicity and side effects. Therefore, compound VIIIa was the
exception with selectivity for ALR2 over ALR1 ranging from 10.8 VIIIc to 0.98 IIIb.
Compound VIIIa could be considered a potent and selective inhibitor with minimal side effects.
14

Table 2. In vitro ALR1 and ALR2 activities of the new compounds of series A and B.
Compounds
ALR1
µM
ALR2
µM
S.E
IC50[ALR1]/
IC50[ALR2]
IC50
IC50
IC50 ± SEM (µM)a/(% Inhibition)b
Series A
12.9 ± 0.30
10.8 ±0.05
5.54 ± 0.89
3.75 ± 0.04
11.1±2.07
4.91±0.02
3.44
0.98
1.12
IIIa
IIIb
IVa
17.1± 0.98
1.62 ± 0.11
10.5
IVb
Series B
VIIa
9.16 ±0.07
16.7 ± 0.61
13.6 ± 0.43
21.4 ± 0.06
4.78 ± 0.13
7.76±0.08
8.72 ± 0.06
2.53 ± 0.03
1.91
2.15
1.55
8.45
VIIb
VIIc
VIId
60.1 ± 0.72
0.47 ± 0.02
127
VIIIa
17.72 ±0.63 7.68 ± 0.82
2.31
10.8
1.18
10.9
VIIIb
VIIIc
VIIId
421
38.2 ± 1.42
9.81 ±0.46
61.1 ± 0.85
3.53 ± 0.06
8.31 ± 0.26
5.59 ± 0.07
15

55
47.9
0.89 ± 0.01
53.82
614
43.55 %
-
0.468 ± 0.003 93.1
Sorbinilc
Valproic acidd
3.14 ± 0.02
-
-
-
57.4 ± 0.89
a: n = 3.
b: % Inhibition at 0.6 µM inhibitor concentration
c: Reported IC50 of 3.42 µM by Rakowitz et al.30
d: Reported IC50 of 56.1 ± 2.7 µM by Stefek et al.28
2.3. Molecular docking study
In an attempt to rationalize the obtained in vitro biological results and obtain insights into the
possible binding modes of compounds IVa, VIId and VIIIa together with the reference drug,
sorbinil 3 within the active site of ALR2. The X-ray crystal structures of ALR2 complexed with
NADPH (PDB ID, 4QR6-ALR2 and 3H4G-ALR1) co-crystalized with {2-[(1,3-benzothiazol-2-
ylmethyl)carbamoyl]-5-fluorophenoxy}acetic acid (37V) and (2S,4S)-2-aminoformyl-6-fluoro-
spiro[chroman-4,4'-imidazolidine]-2',5'-dione (FID), respectively were download from the
Protein Data Bank (PDB). The molecular docking simulations were performed by MOE dock
(MOE2016. 8)31 using Triangle matcher as placement scheme and London dG as a scoring
function. The binding affinity between the ligand and the target is better when the value of ∆G is
lower. The saved pose for the ligand-target complex of each compound was subjected to detailed
2D and 3D analysis for its interactions with the target. Molecular docking simulation of
compounds IVa, VIId and VIIIa and sorbinil into ALR2 revealed that all the ligands were
configured into the active site with several molecular interactions probably responsible for the
16

observed affinity on target enzyme ALR2 (Figures 3, 4 and 5). Binding scores and the type of
interaction together with pharmacological activity are listed in (Table 3).
The energy data showed a direct correlation between the binding free energy (∆G) values of the
target compounds and their inhibitory activities32 with ALR2. All the docked compounds showed
either hydrogen bond acceptor or donor interaction with Cys298 in the active site except
compound IVa. The structural moieties participating in such interactions involved C=O, NH or
CH. Initially, the docking studies for compounds IVa, VIId and VIIIa on ALR2 revealed that
the most active compound was VIIIa with an IC50 value 0.47 μM and with the lowest ∆G value -
7.21 Kcal/mol. Its structure revealed that it has the chloro atom attached to the thienyl group at
position 5 rendering it more lipophilic. The thienyl ring along with the cyclohexanonyl moiety
might render the compound more flexible and smaller than its analogues with the relatively rigid
and bigger fluorenonyl moiety. This comparative flexibility could increase the affinity of the
compound VIIIa towards the active site of ALR2. All these features might result in an increased
interaction with the active site of ALR2. Compound VIIIa made several hydrophobic and H-
bonding contacts with key amino acid residues inside the active site. This active site included a
hydrophobic portion that was lined with Trp20, Ala299, Trp111, Phe122, Trp219 and Leu300
and a hydrophilic portion contained Thr113, Cys298, Try48, His110, Tyr309 and cyc303.
Compound VIIIa established strong binding affinity through a hydrophobic π-πinteraction with
the phenyl ring of Trp111 and H-πinteraction with Thr113. The structural moiety participating in
π-πinteractions or H- πinteractions involved the five membered ring of compound VIIIa. In
addition, it showed a hydrogen bond donor interaction with Cys298 (Figure 3). The smaller size
of both the thienyl and cyclohexanonyl moieties as well as the flexibility of its structure could
increase the compound affinity towards the active site of ALR2. Therefore, all the previous
features could rationalize its selective inhibition of ALR2 (Table 3 and Figure 3).
17

Figure 3. 2D ligand interaction of compound VIIIa in the inhibitor active site of ALR2.
18

Table 3. Inhibitory activity (μM, IC50), Selectivity index (SI), Binding scores and binding interaction of compounds IVa, VIId and
Compounds Pharmacological activity
ALR1 ALR2
5.54 ±0.89 4.91±0.02
ALR2
S (∆G)
-6.304
ALR1
S(∆G)
SI
Binding Interaction
Binding Interaction
H-accept. Trp114-C=O
H-accept. His113-C=O
H-donor Ile261-Cl
1.12
H-donor Asp216-S
-6.1241
IVa
H-accept.Lys262-C=O
H-accept.Ser263-C=O
H-accept. Thr265-C=O
H-accept. Arg268-C=O
pi-H Arg268-phenyl
H-accept. Tyr48-C=O
H-accept. Cys298-C=O
H-accept. His110-C=O
H-donor Cys298-C=O
H-donor Thr113-Cl
21.4 ±0.06 2.53 ± 0.03 8.45
-5.110
-7.206
-3.251
-4.5874
-5.8008
VIId
H-accept. Tyr50-C=O
H-accept. Trp114-C=O
H-pi Tyr-210-NH
60.1
0.72
± 0.47 ± 0.02 127
VIIIa
Sorbinil
H-accept. Tyr50-C=O
H-pi Tyr-22-C
pi-pi Trp111-5-mem. ring
H-accept. Trp111-C=O
H-pi Trp-20-C
H-accept. His113-C=O
3.14 ± 0.02
19

VIIIa and reference drug sorbinil with aldose reductase targets (ALR2, ALR1).
S.I = IC50[ALR1]/ IC50[ALR2]
20

The structures of compound VIIIa and sorbinil were not superimposed in the active site of
ALR2 enzyme (Figure 4). Their root mean square deviations (RMSD) values were 1.54 Å and
0.69 Å, respectively. This means that these compounds were not able to generate similar docking
poses. Compound VIIIa fits plausibly in the active site of ALR2 with a binding score -7.21.
Investigation of its docking results revealed an optimum binding mode with the key amino acids.
It displayed one major hydrogen bonding interaction with the key ALR2 residue Cys298. It was
noticed that the thiophene ring was extended in the hydrophobic pocket created by Leu300,
Alaa299 and Trp111. It was stabilized by a π-πinteraction between the thiophene ring and the
phenyl ring of Trp111. These binding features played a crucial role in stabilizing the three
dimensional structure of the inhibitor-enzyme complex. On the other side, it was noticed that
sorbinil did not exploit the pocket created by Leu300, Alaa299 and Trp111. Its core structure
was shifted slightly outside the binding domain of ALR2. It showed a higher binding score -3.25,
reflecting its lower binding affinity when compared to compound VIIIa. This probably explains
the superior potency of compound VIIIa.
Compound VIId displayed the highest ∆G value -5.11 Kcal/mol. This might illustrate its low
activity and reduced stability inside the active pocket of ALR2 (Table 3 and Figure 5).
Compound IVa showed more hydrogen bond interactions with several amino acids of ALR2,
which could be due to its higher Polar Surface Area TPSA 83.549 Å and rigidity. Compound
IVa was found to bind in a different site (it might be an allosteric site) (Table 3 and Figure 6).
21

Figure 4. 3D Comparable binding interaction of compound VIIIa and sorbinil in the inhibitor
active site of ALR2.
Figure 5. 2D Ligand interaction of compound VIId in the inhibitor active site of ALR2.
22

Figure 6. 2D Ligand interaction of IVa in the inhibitor active site of ALR2.
The compounds were then docked in the active site of ALR1 and showed either one or two
hydrogen bond acceptor interaction between Tyr50 or Trp114 in the active site and the C=O
moiety. Docking results of the investigated compounds revealed that compound IVa showed the
lowest ∆G scores -6.12 Kcal/mol for ALR1 binding. The presence of chloro atom on the thienyl
ring (H-donor Ile261-Cl) contributed to its good binding with ALR1 (Table 3 and Figure 7).
23

Figure 7. 2D Ligand interaction of compound IVa with the active site residues of ALR1.
As a result, the investigation of the binding values of compounds IVa, VIId and VIIIa on both
targets ALR1 and ALR2, revealed that the designed compounds had inhibitory activities against
both targets with greater selectivity towards ALR2 in case of VIId and VIIIa.
2.4. In silico prediction of physicochemical properties and drug-likeness properties
In the development of drugs intended for oral use, good drug absorption and appropriate drug
delivery are very important issues.33 Poor pharmacokinetics is responsible for the failure of 30 %
of oral drugs.34 Thus, in silico prediction of oral bioavailability is very important in the early
stages of drug discovery in order to select the most promising compounds for further
optimization. Moreover, its importance in the latter stages is to identify candidates for further
clinical development.35 Predictions of physicochemical properties and drug-likeness scores for
the investigated compounds are presented in (Table 4). Candidate drugs that conform to the
Lipinski's "rule of five""Ro5" tend to have lower attrition rates during clinical trials and hence
24

have an increased chance of reaching the market.22-24 The rule states that those most “drug-like”
molecules must have log P (logarithm of the octanol/water partition coefficient) ≤ 5, molecular
weight ≤ 500 Daltons, number of hydrogen bond donors ≤ 5 and number of hydrogen bond
acceptors ≤ 10. Molecules violating more than one of these rules may have problems with oral
bioavailability. The rule is called "Rule of 5", because the border values are 5, 500, 5, and 2X5.
Veber et al. offered further modifications in Ro5.36 According to Veber, it is revealed that for
passing oral bioavailability criteria, the number of rotatable bond (nrotb) should be ≤10. This number
(nrotb) is important for conformational changes and ultimately for binding with the amino acids
in the active site of the targets.22 Molecular or topological polar surface area TPSA is a very useful
parameter for the prediction of drug transport properties. TPSA values should be ≤ 140 Å2 for
prevailing drug absorption through intestine and ≤ 90 Å2 for molecule to penetrate the blood brain
barrier (standard values).37,38 The physicochemical properties of the tested compounds were
calculated using MolSoft software39,40 and all compounds under investigation showed good
permeability and absorption. They possessed 1-2 hydrogen bond donors and 3-5 hydrogen bond
acceptors (Table 4). In addition, all the compounds were showed good permeability through cell
membranes (log P ranges from 2.12 – 4.81). Furthermore, the titled compounds in general had 3-
4 rotatable bonds, so they fulfilled the criterion and exhibited moderate conformational
flexibility for binding with both enzymes except compound IVa which showed some rigidity
(nrobt = 1). In addition, compound IVa obtained relatively higher molecular weight, PSA and
molecular volume MV, this made compound IVa approximately more rigid and affected
ultimately its binding with the target and subsequently it might shift to an allosteric site.
Moreover, all the tested compounds had zero violation and obeyed Lipinski's rule. Besides, the
compounds had PSA values within acceptable range 78.50 – 92.78. Therefore, all the compounds
obtained the ability to penetrate the blood brain barrier (except compound VIIId, TPSA > 90 Å2)
25

and facilitate drug absorption through intestine. Accordingly, they are considered as drug delivery
candidates. PSA were used to calculate the percentage of absorption (ABS%) using the following
equation: ABS% = 109 – 0.345 PSA.41 Generally, the tested compounds demonstrated the (ABS%)
range of 76.99 – 81.91 which indicated their good oral bioavailability. Furthermore, the tested
compounds were found to fulfil the requirements of solubility of more than 0.0001mg/L42 and
could be considered as drug candidates for oral absorption. Compound VIId and its analogues
VIIIa exhibited higher solubility characters in comparison to compound IVa. The drug-likeness
score (Table 4) is a combined effect of physicochemical properties, pharmacokinetics and
pharmacodynamics of a compound and is represented by a numerical value. Drug-likeness model
score of compound VIIIa was 0.41 (Figure 8). Compounds that show positive values should be
considered as drug-like candidates. The green color means non-drug like behavior and those fall
under blue color area are considered as drug-like. Those compounds having negative or zero value
should not be considered as drug like. Compound VIIIa showed positive value so it should be
considered as drug-like.
Furthermore, the higher PSA for compound VIId was 92.78 Å versus 83.55 Å for compounds
IVa and VIIIa together with the relative increase in the number of rotatable bond (nrotb =3)
(Table 4) might increase its binding affinity towards the active site of ALR1. Finally, the
investigated compounds showed reasonable drug-likeness scores and physicochemical characters
along with non-violation of Lipinski’s rule of five and acceptable pharmacokinetics
(permeability, intestinal absorption). Accordingly, these compounds are suitable to be drug-like
candidates.
26

MWta MVa
HBAb HBDb nrotba PSAa
Solubilityb nviolationsa
mg/L
Compounds
Log
pa
ABS
%
Drug-
likeness
(A3)
(Ao)
(Ao)
(A2)
scoreb
(D)
430.89 320.51 4.02 5
338.38 283.00 2.12 5
348.83 260.70 2.81 5
236.20 189.16 0.86 3
1
1
1
2
1
3
2
0
83.55 80.17 0.01
92.78 76.99 35.05
83.55 80.17 19.40
67.43 85.73 140.11
0.0
0.0
0.0
0.0
-0.81
0.28
0.41
-0.09
IVa
VIId
VIIIa
Sorbinil
Table 4. Molecular properties and drug-likeness.
MWt: Molecular weight,
MV: Molecular Volume,
Log p: Log partition coefficient,
HBA: Hydrogen Bond Acceptor,
HBD: Hydrogen Bond Donor,
nrotb: number of rotatable bond,
PAS: Polar Surface Area
ABS%: Percentage of absorption
nviolations: number of violations
a: Values calculated using molinspiration 2016.10
b: values calculated using MolSoft
27

Figure 8. Drug-likeness plot of compound VIIIa (MolSoft). The green color means non-drug
like behavior and those fall under blue color area are considered as drug-like. Those compounds
having negative or zero value should not be considered as drug like. Drug-likeness model score
of compound VIIIa was 0.41. It showed positive value so it should be considered as drug-like.
3. CONCLUSION
Based on the scaffold of the previously reported aldose reductase inhibitors with hydantoin
core, a novel series of sulfonylureas were designed and tested for both ALR2 inhibition and
hypoglycemic activity. Spiroimidazolidine-2,4-dione derivatives preserve all the minimum
requirements of the proposed pharmacophoric models to enhance both activities. All compounds
(except VIId) reduced the blood glucose level and all inhibited aldose reductase activity.
Biological studies have led to the identification of a novel molecule with a remarkable selectivity
towards ALR2 versus ALR1. The investigation showed that the most active compound is VIIIa
with IC50 0.47 µM with selectivity index 127-fold. Docking studies gave a partial explanation for
28

the binding mode of these novel compounds in both ALR1 and ALR2 active sites. The biological
and computational data provided valuable information for the further development of this very
promising lead compound.
4. EXPERIMENTAL
4.1. Chemistry
The commercial chemicals and solvents were reagent grade and used without further
1
13
purification. H- and C-NMR Spectra were measured on a Varian Mercury VX-300 NMR
1
13
spectrometer or Jeol LA (400 MHz for H-NMR, 100 MHz for C-NMR). Chemical shifts were
reported in parts per million (ppm) using tetramethylsilane (TMS) as an internal standard.
Electron impact mass spectra were recorded on Shimadzu GCMS-QP 5050A gas chromatograph
mass spectrometer (70 eV). Elemental analysis was performed at the Microanalytical Center of
Cairo University, Egypt. Melting points of the compounds were determined in open capillaries
using Gallenkemp melting point apparatus and are uncorrected. All the reactions were monitored
through thin layer chromatography (TLC) using Merck 0.2 mm pre-coated DF-Aluminium sheets
60F254. IR spectra were recorded on a Shimadzu FT-IR 8101 PC IR spectrophotometer (KBr
pellets). Values were represented in cm-1. Compounds II and VI were prepared as previously
described.5,25
4.1.1. General method for the preparation of 3'-aryl sulfonyl derivatives of
spiro[fluorene-9, 5'-imidazolidine]-2',4'-dione (IIIa and IIIb)
Arylsulfonyl chloride (5.8 mmol) was added dropwise to a stirred mixture of compound II (4.8
mmol), triethylamine (4.8 mmol) and catalytic amount of DMAP in DCM (10 mL), and stirred
for 15 h. The reaction mixture was neutralized with 1N aquous HCl and extracted with
dichloromethane (4 x 25 mL). After evaporation of the solvent, the crude product was
29