New lipophenols prevent carbonyl and oxidative stresses involved in macular degeneration

Article abstract of DOI:10.1016/j.freeradbiomed.2020.10.316

Dry age-related macular degeneration and Stargardt disease undergo a known toxic mechanism caused by carbonyl and oxidative stresses (COS). This is responsible for accumulation in the retinal pigment epithelium (RPE) of A2E, a main toxic pyridinium bis-retinoid lipofuscin component. Previous studies have shown that carbonyl stress in retinal cells could be reduced by an alkyl-phloroglucinol-DHA conjugate (lipophenol). Here, we performed a rational design of different families of lipophenols to conserve anti-carbonyl stress activities and improve antioxidant properties. Five synthetic pathways leading to alkyl-(poly)phenol derivatives, with phloroglucinol, resveratrol, catechin and quercetin as the main backbone, linked to poly-unsaturated fatty acid, are presented. These lipophenols were evaluated in ARPE-19 cell line for their anti-COS properties and a structure-activity relationship study is proposed. Protection of ARPE-19 cells against A2E toxicity was assessed for the four best candidates. Finally, interesting anti-COS properties of the most promising quercetin lipophenol were confirmed in primary RPE cells.

Full text of DOI:10.1016/j.freeradbiomed.2020.10.316

Journal Pre-proof
New lipophenols prevent carbonyl and oxidative stresses involved in macular
degeneration
Espérance Moine, Manel Boukhallat, David Cia, Nathalie Jacquemot, Laurent Guillou,
Thierry Durand, Joseph Vercauteren, Philippe Brabet, Céline Crauste
PII:
S0891-5849(20)31590-2
DOI:
https://doi.org/10.1016/j.freeradbiomed.2020.10.316
FRB 14896
Reference:
To appear in: Free Radical Biology and Medicine
Received Date: 24 September 2020
Revised Date: 22 October 2020
Accepted Date: 23 October 2020
Please cite this article as: E. Moine, M. Boukhallat, D. Cia, N. Jacquemot, L. Guillou, T. Durand,
J. Vercauteren, P. Brabet, C. Crauste, New lipophenols prevent carbonyl and oxidative stresses
involved in macular degeneration, Free Radical Biology and Medicine, https://doi.org/10.1016/
j.freeradbiomed.2020.10.316.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. 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.
©
2020 Elsevier Inc. All rights reserved.

1
2
New lipophenols prevent carbonyl and oxidative
stresses involved in macular degeneration.
a,
a
c
c
b
3
4
Espérance Moine *, Manel Boukhallat , David Cia , Nathalie Jacquemot , Laurent Guillou ,
a
a
b
a,
Thierry Durand , Joseph Vercauteren , Philippe Brabet and Céline Crauste *.
a
5
6
7
8
9
Institut des Biomolécules Max Mousseron (IBMM), Université de Montpellier, CNRS,
ENSCM, Montpellier, 34093, France
b
Institut des Neurosciences de Montpellier, INSERM U1051, Université de Montpellier,
Montpellier, 34091, France
c
Laboratoire de Biophysique Neurosensorielle, UMR INSERM 1107, Facultés
10
de Médecine et de Pharmacie, Clermont-Ferrand, 63000, France
1
1
1
2 * Correspondence: celine.crauste@umontpellier.fr; Tel.: +33-4-11-75-95-66,
1
1
1
1
1
1
1
2
3
esperance.moine@umontpellier.fr; Tel.: +33-4-11-75-95-66
4 Keywords
5 Lipophenol, PUFA, carbonyl stress, antioxidant, macular degeneration, structure-activity
6 relationship.
7
8 Abbreviations
9
A2E, N-retinylidene-N-retinylethanolamine; ABCA4, ATP binding cassette subfamily A
0 member 4; AGEs, advanced glycation end products; AMD, age-related macular degeneration;

2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
1 ARPE-19, adult retinal pigment epithelial cell line-19; ASAP, atmospheric solids analysis probe;
2 atRAL, all-trans-retinal; BRB, blood retina barrier; BSMR, based on starting material recovery;
3 COS, carbonyl and oxidative stresses; DCC, dicyclohexylcarbodiimide; DCFDA, 2’,7’-
4 dichlorofluorescin diacetate; DCM, dichloromethane; DHA, docosahexaenoic acid; DMAP,
5 dimethylaminopyridine; DMEM/F12, Dulbecco’s Modified Eagle’s Medium (DMEM)/Ham
6 F12; DMF, dimethylformamide; DMSO, dimethylsulfoxide; DNA, deoxyribonucleic acid; EC ,
5
0
7 efficiency concentration 50; EGCG, epigallocatechin-3-O-gallate; EPA, eicosapentaenoic acid;
8 ESI, electrospray ionization; FBS, fetal bovine serum; 4-HHE, 4-hydroxyhexenal; 4-HNE, 4-
9 hydroxynonenal; HRMS, high resolution mass spectroscopy; LA, linoleic acid; Mp, melting
0 point; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; NMR, nuclear
1 magnetic resonance; Nrf2/Keap1, Nuclear factor (erythroid-derived 2)-like 2/Kelch-like ECH-
2 associated protein 1; PUFA, poly-unsaturated fatty acid; RCS, reactive carbonyl species; Rf,
3 retardation factor; ROS, reactive oxygen species; RPE, retinal pigment epithelium; TFA,
4 trifluoroacetic acid; THF, tetrahydrofuran; TIPS, triisopropylsilyl; TLC, thin layer
5 chromatography; UPLC, ultra-performance liquid chromatography; VEGF, vascular endothelial
6 growth factor.
7
8 Abstract
9
Dry age-related macular degeneration and Stargardt disease undergo a known toxic
0 mechanism caused by carbonyl and oxidative stresses (COS). This is responsible for
1 accumulation in the retinal pigment epithelium (RPE) of A2E, a main toxic pyridinium bis-
2 retinoid lipofuscin component. Previous studies have shown that carbonyl stress in retinal cells
3 could be reduced by an alkyl-phloroglucinol-DHA conjugate (lipophenol). Here, we performed a

4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
4 rational design of different families of lipophenols to conserve anti-carbonyl stress activities and
5 improve antioxidant properties. Five synthetic pathways leading to alkyl-(poly)phenol
6 derivatives, with phloroglucinol, resveratrol, catechin and quercetin as the main backbone, linked
7 to poly-unsaturated fatty acid, are presented. These lipophenols were evaluated in ARPE-19 cell
8 line for their anti-COS properties and a structure-activity relationship study is proposed.
9 Protection of ARPE-19 cells against A2E toxicity was assessed for the four best candidates.
0 Finally, interesting anti-COS properties of the most promising quercetin lipophenol were
1 confirmed in primary RPE cells.
2
3
4
Introduction
Oxidative stress, resulting from an overproduction of reactive oxygen species (ROS) within
5 cells or in the extracellular matrix, highly damages key cellular proteins, lipids and DNA.
6 Reactive carbonyl species (RCS), such as sugars and osones, endogenous aldehydes or
7 metabolites derived from lipid oxidation, are involved in glycation and cross-linking reactions
8 and thus, affect cellular viability leading to tissue injury. These two carbonyl and oxidative stress
9 (COS) mechanisms play a crucial role in aging-associated pathologies, like age-related macular
0 degeneration (AMD), or in some inherited forms of macular degeneration, such as Stargardt
1 disease [1–3]. AMD is one of the primary causes of central and irreversible visual loss among
2 the elderly in occidental countries, and there is no treatment currently available to stop retinal
3 degeneration, especially in the prevalent dry form (80-85% of cases). COS mechanisms are
4 responsible for the accumulation in retinal pigment epithelium (RPE) of a toxic bis-retinoid
5 conjugate called A2E (its photoisomers and its oxidized metabolites). Pathologic A2E
6 biosynthesis occurs when all-trans-retinal (atRAL), rather than undergoing reduction to retinol in

6
6
6
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
8
8
8
8
8
7 the RPE, accumulates abnormally in photoreceptors. This accumulation can be due to age
8 (AMD) or loss of function of the ABCA4 transporter (Stargardt disease) [2]. Two molecules of
9 this reactive aldehyde (RCS) in excess, condense with one molecule of phosphatidyl
0 ethanolamine (carbonyl stress step) into an unstable dihydropyridinium bis-retinoid.
1 Transformation of the dihydropyridinium bis-retinoid into A2E then arises during the oxidative
2 step leading to a more stable pyridinium form. A2E and its various oxidative metabolites are the
3 major constituent of lipofuscin, a marker of AMD. The massive accumulation of lipofuscin in the
4 RPE (following phagocytosis of photoreceptor outer segments) is cytotoxic and causes
5 progressive RPE cell death and subsequent photoreceptor degeneration [4]. A2E cytotoxicity is
6 explained by several mechanisms: loss of membrane integrity due to its amphiphilic properties
7 during membrane integration; increase of oxidative stress by the generation of singlet oxygen
8 during exposition to blue light; metabolization in A2E-oxidized metabolites resulting in DNA
9 lesions [5–7]. Limitation of A2E biosynthesis (by clearance of atRAL) and oxidation is therefore
0 an attractive target to slow the progression of macular degeneration [8].
1
Based on epidemiology studies, natural plant antioxidants, such as (poly)phenols, secondary
2 metabolites that protect plants from several aggressions, are efficient at protecting animal
3 organisms against oxidative stress. Some (poly)phenol structures have been reported to protect a
4 variety of retinal cell types from oxidative stress-induced cell death [9–11]. Such activity may be
5 related to their capacity to scavenge directly excess ROS, or to stimulate the enzymatic
6 antioxidant defenses of the organism through Nrf2/Keap1 signaling pathway[12,13]. Moreover,
7 recent literature addressed the efficiency of (poly)phenols to inhibit Advanced Glycation End
8 product formation (AGEs), which result from both carbonyl and oxidative stresses [14,15], and
9 to act also as anti-carbonyl stress derivatives by trapping toxic RCS (glyoxal, acrolein, 4-

90 hydroxynonenal (4-HNE)) [16–18].
9
9
9
9
9
9
9
9
9
1
Unfortunately, while showing interesting in vitro protection, the limited bioavailability of most
2 (poly)phenols (weak drug absorption, high systemic metabolization, restricted cell
3 penetration…) negatively influences their in vivo potency and thus their development as drug
4 candidates. Increasing the lipophilicity of those (poly)phenols may be a useful way to improve i)
5 their protective effect on lipid membranes or lipid derivatives (such as A2E), ii) their absorption
6 [19], and iii) their formulation for in vivo administration [20–22]. In our previous work, we
7 synthesized lipophenols (or phenolipids), polyunsaturated fatty acid (PUFA) linked to alkyl-
8 phloroglucinol (LEAD A/B, Figure 1), that were designed to reduce carbonyl stress associated
9 with retinal dystrophies [23,24]. The main objective of those lipophenols was to avoid/slow the
1
1
1
1
1
1
1
1
1
1
1
1
1
00 pace of A2E formation by scavenging atRAL (Figure 1). To increase lipophilicity and cellular
01 permeability of starting (poly)phenol, PUFAs were preferred to saturated ones, due to their
02 observed benefits in an AMD clinical trial [25]. Both isopropyl-phloroglucinols linked to linoleic
03 acid (C18:2 n-6; LA - LEAD A, Figure 1) and docosahexaenoic acid (C22:6 n-3; DHA - LEAD
04 B, Figure 1) showed promising protection of RPE cells against atRAL toxicity. A isopropyl
05 functional group was rationally selected to increase the nucleophilic properties of the aromatic
06 cycle involved in the trapping of aldehyde function of atRAL and leading to the formation of a
07 stable non-toxic chromene adduct (Figure 1) [23,26]. Both isopropyl and PUFA parts were
08 proven to be indispensable in a phloroglucinol series (LEAD A and LEAD B) to ensure cell
09 protection against carbonyl stress in cellular assays [23,24]. The DHA analogue (LEAD B) was
10 selected for further in vivo evaluations due to several benefits of this omega-3 lipid: i) the high
11 proportion of DHA in the membrane of photoreceptor outer segments, ii) its ability to reach the
12 retina through PUFA transports [27]; iii) recent data showing that dietary supplementation with

1
1
1
1
1
1
1
1
1
1
13 high doses of DHA/EPA significantly improves the visual acuity of AMD patients [25]; and iv)
14 the report that specific oxidative metabolites of DHA (Neuroprotectin D1) protect RPE cells
15 against oxidative stress [28]. Intravenous administration of LEAD B allowed photoreceptor
16 protection against acute light-induced degeneration in a mouse model used for the development
-
/-
17 of novel therapeutics for Stargardt disease (Abca4 mice) [29,30]. Despite interesting
18 preliminary in vivo protection observed using LEAD B, this phloroglucinol derivative lacks
19 efficiency to reduce ROS production and cellular oxidative stress. In the present study, we report
20 new potent alkyl-lipophenols that act on both carbonyl and oxidative stresses, to reduce atRAL
21 toxicity, and oxidation associated with A2E formation, metabolism and toxicity, critical steps
22 involved in photoreceptor degeneration.
1
1
1
1
1
1
1
1
1
1
1
1
1
23
Several possibilities have been proposed to increase antioxidant activities while preserving
24 anti-carbonyl stress properties. First, using the same (poly)phenol backbone, PUFA was linked
25 directly to the aromatic cycle to free one phenolic function of the phloroglucinol involved in the
26 antioxidant activity (C-phloroglucinol series, Figure 1). Second, new series of lipophenols were
27 also developed from natural (poly)phenols, different from phloroglucinol, meeting the criterion
28 that the starting (poly)phenol contained an isopropyl-resorcinol framework mimicking the
29 phloroglucinol backbone, as this was indispensable for the protection against carbonyl stress
30 [23,24]. More effective natural antioxidants, such as resveratrol stilbenoid (true vinylogous
31 analogue of phloroglucinol), or flavonoids, such as quercetin or catechin, all of which were
32 reported to trap aldehyde function in cell free assays [18,31,32] and to protect ARPE-19 cells
33 against A2E photo-oxidation damage [33–35], have been proposed as new anti-COS alkyl-
34 lipophenol series (Figure 1). Nine novel alkyl-lipophenol-PUFAs have been synthesized using
35 original chemical strategies. Toxicity assays, protection against a lethal concentration of atRAL

1
1
1
1
1
1
1
1
1
1
1
2 2
36 and reduction of H O induced ROS production, were used to first screen these molecules in
37 ARPE-19 cells. The necessity of the isopropyl and the PUFA parts for carbonyl stress protection
38 has been highlighted with all new (poly)phenol backbones (resveratrol, catechin and quercetin).
39 The ability to preserve antioxidant properties, despite reduction of free phenolic functions
40 compared to native (poly)phenols, has been studied. Best DHA-lipophenol conjugates offering
41 both oxidative and carbonyl stress protection, were compared with LEAD B, for their ability to
42 reduce A2E toxicity during photo-oxidation process in RPE. According to the comparison of
43 both carbonyl and oxidative stress protection, a new alkyl-quercetin-PUFA conjugate has been
44 highlighted. Cell protection was finally validated with this lipophenol in primary RPE cells, as
45 primary cultures are likely to reflect in vivo cell morphology and function more accurately.
46
LEAD phloroglucinol series
atRAL scavenging by phloroglucinol [26]
OH
OPUFA
O
atRAL
OH
O
OH
HO
O
Chromene
HO
OH
LEAD A: PUFA = linoleic acid (LA)
LEAD B: PUFA = docosahexahenoic acid (DHA)
New lipophenol series
OH
OH
OH
OH
OH
OH
*
*
HO
O
HO
O
O
OPUFA
HO
O
O
OPUFA
OPUFA
O
*
O
*
OPUFA
C-phloroglucinol series
resveratrol series
catechin series
quercetin series
*
: interchangeable positions for isopropyl function
1
1
47
48
Figure 1. Chemical structures of developed lipophenol series.
1
1
1
49
50
Results and Discussion
51 1. Synthesis and anti-COS evaluation of C-phloroglucinol derivatives

1
1
1
1
1
1
1
1
1
52
The first chemical changes considered to increase antioxidant capacity concerned the link of
53 the PUFA moiety. Indeed, releasing one more phenolic function on the phloroglucinol (P)
54 backbone should allow better ROS scavenging properties. This modification was investigated
55 through a C-alkylation on the phloroglucinol aromatic ring, by the introduction of a
56 hydroxylated-alkyl spacer, in order to perform PUFA functionalization. The synthesis of the two
57 C-linker analogues (8a/b) of LEAD A (P-OiP-OLA) and LEAD B (P-OiP-ODHA) is described
58 in Scheme 1.
59
a
60
Scheme 1. Chemical pathway leading to C-phloroglucinol derivatives
1
61
62
a
1
1
1
1
1
Reagents and conditions: (a) Zn(CN)
, EtOAc, DMF, rt, 86 h, 64%; (b) Ph
64 Pd/C, EtOAc, rt, 4 h, 71%; (d) TIPS-OTf, NEt , DCM, rt, 2 h, 35%; (e) LiAlH
2
, ZnCl
2
, Et
PCHCOOMe, DCM, rt, 18 h, 59%; (c) H
(1M), Et O, 0
2 2
O-HCl (2N), rt, 16 h and H O, reflux, 20
63 min or POCl
3
3
2
,
3
4
2
65 °C, 4 h30, 77%; (f) PUFA, DCC, DMAP, DCM, rt to 50 °C, 7 h, 7a: 66% and 7b: 28%; (g)
66 NEt /3HF, THF, rt, 19 h, 8a: 39% and 8b: 30%.
3
1
67

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
68
Starting from mono-isopropyl phloroglucinol 1 [23], a Gattermann-Koch formylation [36] lead
69 to salicylic aldehydes 2a/2b (80/20) in 64% combined yield. The next Wittig reaction was
70 performed on the mixture of 2a/2b for a two-carbon homologation, and the resulting two α,β-
71 unsaturated esters 3a/3b were reduced under hydrogen in the presence of Pd/C catalyst which
72 after purification lead to the isolation of isomer 4a. Phenolic functions were then protected by
73 silylated protecting groups (compound 5). The alcohol 6 was obtained after the ester reduction
74 and was coupled either with LA or DHA using classical Steglich coupling conditions. Finally,
75 LA and DHA desired conjugates (respectively 8a and 8b) were obtained after deprotection of
76 silylated phenolic function in the presence of NEt /3HF.
3
77
In vitro evaluations of the C-phloroglucinol derivatives were performed on the ARPE-19 cell
78 line. All compounds were tested at various concentrations in the range of 0-160 µM for
79 cytotoxicity assessment, and in the range of 0-80 µM for anti-COS evaluation. Regarding the
80 new C-phloroglucinol derivatives (8a and 8b), our main concerns were to maintain anti-carbonyl
81 stress activity, despite the reduction of one nucleophilic C-alkylation site (on the aromatic cycle)
82 compared to LEAD A/B structures, and to increase antioxidant properties by releasing a phenolic
83 function. As shown in Figure 2, there is no apparent cytotoxicity due to the modification of the
84 lipid position.
85

1
00
**
*
*
*
**
*** ***
***
*** ***
*
**
*** *** ** ** **
*
**
***
***
*
**
***
***
10 µM
*
**
8
6
4
2
0
0
0
0
4
8
1
1
0 µM
0 µM
20 µM
60 µM
0
)
)
)
)
A
a
B
b
D
8
8
(
(
D
A
A
E
A
A
E
L
L
H
(
L
C
(
D
-
A
C
-
i
P
A
L
H
O
O
iP
-
-
D
iP
P
O
-
P-O
O
i
P
-
P
O
-
P
1
86
1
1
1
1
1
1
87 Figure 2. Evaluation of cytotoxicity of C-phloroglucinol derivatives; P-OiP-OLA (LEAD A)
88 [24], P-OiP-CLA (8a), P-OiP-ODHA (LEAD B) [23] and P-OiP-CDHA (8b). Results are
89 expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell viability
90 (MTT) after incubation of phloroglucinol derivatives (0-160 µM). The data are expressed as the
91 percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-
92 treated control cells.
1
1
1
1
1
1
1
2
2
2
2
93
94
Anti-carbonyl stress activity of C-phloroglucinol derivatives is evaluated by comparing cell
95 survival in presence of a toxic concentration of carbonyl stressor atRAL, and is represented in
96 Figure 3A. The two new C-phloroglucinol derivatives (8a and 8b) displayed interesting
97 protective properties, increasing cell viability (at 80 µM) of + 25% for LA derivative (8a) and +
98 28% for DHA derivative (8b), compared to non-treated and atRAL-exposed cells. No statistical
99 differences in cell protection were observed under these conditions between P-OiP-OLA (LEAD
00 A) versus P-OiP-CLA (8a) and between P-OiP-ODHA (LEAD B) versus P-OiP-CDHA (8b), at
01 equivalent concentrations. Modification of the PUFA position on the phloroglucinol backbone
02 did not affect the anti-carbonyl stress activity of the derivative: the PUFA moiety can be
03 introduced indifferently on a phenolic function or with a spacer on the aromatic core. This result

2
2
2
2
2
04 is in agreement with the literature, as phloretine, a natural phloroglucinol analogue with a
05 substituent on the aromatic ring, is also able to trap toxic aldehydes, such as acrolein, a toxic
06 aldehyde found in cigarette smoke and involved in AMD via induction of oxidative damage in
07 RPE cells, or 4-HNE, derived from lipid peroxidation [31].
08
(
A)
(
B)
6
0
0
0
0
0
0
0
*
**
*
**
***
120
*
**
*
**
5
4
3
2
1
***
***
*
**
°
°
1
00
###
°°
*
*
80
60
40
20
0
)
)
)
)
a
b
A
B
-
8
8
(
(
D
D
A
P
A
A
A
)
)
)
)
(8a
B
b
E
L
E
H
A
8
(
L
C
L
D
(
-
(
D
D
P
C
A
A
A
A
A
i
A
E
L
(LE
H
L
O
H
P
-
i
L
C
-
D
O
D
(
-
O
C
-
-
P
P
-O
A
i
A
i
P
P
i
L
O
H
O
iP
-
D
-
-O
P
O
-
P
-O
P
O
-
i
P
P
P
i
O
-
O
1
0 µM
40 µM
80 µM
P
-
P
1
0 µM
40 µM
80 µM
2
2
2
2
2
2
2
2
2
2
2
2
2
09 Figure 3. In vitro anti-COS evaluation of C-phloroglucinol derivatives. Comparison of P-OiP-
10 OLA (LEAD A) [24], P-OiP-CLA (8a), P-OiP-ODHA (LEAD B) [23] and P-OiP-CDHA (8b)
11 activities. Results are expressed in mean ± SEM and are from n = 3-5 independent experiments.
12 (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after incubation of phloroglucinol
13 derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as the percentage of non-
14 treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus
15 non-treated and exposed to atRAL cells. (B) Antioxidant assay: representation of ROS
16 production (DCFDA probe) after incubation of phloroglucinol derivatives (0-80 µM) and H O
2
2
17 (600 µM) in ARPE-19 cells. The data are expressed as the percentage of non-treated and
18 exposed to H O cells. All conditions have a p-value < 0.001 versus non-treated and exposed to
2
2
2 2
19 H O control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus P-OiP-OLA (LEAD A) at the
20 same concentration. ° p < 0.05, °° p < 0.01, °°° p < 0.001, versus P-OiP-ODHA (LEAD B) at the
21 same concentration.
2
22

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
23
The main objective of the synthesis of compounds 8a and 8b was to increase antioxidant
24 properties of the LEADs A and B in cell experiments. As expected, 8a and 8b logically showed
25 weaker antioxidant activity compared to phloroglucinol (data not shown). Compared to untreated
26 cells, C-phloroglucinol analogues were able to decrease H O induced ROS production by 41%
2
2
27 for LA derivative (8a) and 46% for DHA derivative (8b) at 80 µM (Figure 3B). Thus, a
28 significant improvement in antioxidant activity was observed using C-phloroglucinols (8a/8b)
29 that show a dose-dependent effect compared to LEAD A and B. Direct hydroxyl radical
30 scavenging activity of phloroglucinol has been previously reported [26,37], and the liberation of
31 one phenolic function explains the increased antioxidant efficacy of C-phloroglucinol derivatives
32 (8a and 8b) compared to LEAD A and LEAD B, that reduce ROS production by only 25% and
33 30% at 80 µM, respectively.
34
35 2. Synthesis and anti-COS evaluation of resveratrol lipophenol analogues
36
Resveratrol (R), a major active phytoalexin from the stilbene family (3,4’,5-
37 trihydroxystilbene), is mainly found in Vitis vinifera (Vitaceae) stalks and in the roots of
38 Fallopia japonica var. japonica (Polygonaceae). Resveratrol has a wide range of biological
39 activities including anti-bacterial and anti-fungal properties [38], antioxidant activity [10,39],
40 and anti-inflammatory properties [40]. Furthermore, previous evaluations on RPE cell assays
41 highlighted the benefit of resveratrol in reducing VEGF effect and increasing natural antioxidant
42 enzymatic and molecular defenses [40]. Resveratrol should not only be considered as a simple
43 ROS scavenger. The work of Vlachogianni et al. shows better radical scavenging activity of 4’-
44 acetylated resveratrol compared to 3-acetylated analogues [41]; we decided therefore to link the
45 PUFA at the 4’ position of resveratrol. The isopropyl moiety was introduced at the 5 position in

2
2
2
2
2
2
2
46 order to keep the alkyl-resorcinol moiety, which was proven to be responsible for high anti-
47 carbonyl stress activity on phloroglucinol series [23]. In order to confirm the necessity of alkyl as
48 well as PUFA moieties on the resveratrol backbone, as we did previously with phloroglucinol
49 [23,24], we produced in addition to alkyl resveratrol-PUFA (15a/b), an isopropyl-resveratrol
50 derivative not linked to PUFA (R-5OiP 13, Scheme 2) and derivatives linked to PUFAs without
51 alkyl function named resveratrol-4’LA (R-4’LA) and resveratrol-4’DHA (R-4’DHA), both
52 synthesized as previously described [42,43].
a
2
53 Scheme 2. Synthesis and chemical structures of resveratrol derivatives
OH
OH
OH
3
a
b
A
HO
5
HO
O
B
4'
OH
OBu
OBu
resveratrol
R-com
9
10
c, d
OH
OTIPS
OTIPS
g
f
O
O
O
OPUFA
OPUFA
OH
12
1
5a: PUFA = LA
14a: PUFA = LA
14b: PUFA = DHA
1
5b: PUFA = DHA
e
OH
OH
OH
HO
HO
O
ODHA
OLA
OH
R-4'DHA
R-4'LA
13
2
54
55
a
2
2
2
2
2
Reagents and conditions: (a) 2-methyl-butan-2-ol, vinyl butyrate, Novozyme 435 (CALB), 40
56 °C, 8 days, 52%; (b) diisopropylsulfate, K CO , acetone, rt, 24 h, 22% (67% BSMR); (c) TIPS-
57 OTf, DIPEA, THF, rt, 6 h, 63%; (d) NH /MeOH, DCM, 0 to 5 °C, 24 h, 80%; (e) NEt /3HF,
58 THF, rt, 23 h, 67%; (f) PUFA, DCC, DMAP, DCM, rt, 14a: 35 min, 83% and 14b: 2 h, 46%; (g)
2
3
3
3
59 NEt /3HF, THF, rt, 15a: 19 h, 67% and 15b: 20 h, 64%.
3
2
60

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
61
A tedious step to access the desired alkylated resveratrols 15a/b, was the isopropylation of the
62 5 position of resveratrol as the 4’ position is the more reactive. Utilization of the supported lipase
63 Novozyme 435 (Immobilized lipase from Candida antarctica) was efficient and selective to
64 protect the 4’ position of resveratrol, using vinyl butyrate (52%). Alkylated compound 10 was
65 then obtained using diisopropylsulfate and K CO at room temperature for 24 h [44]. Successive
2
3
66 protection of the resulting phenol with TIPS groups, and deprotection of the butyrate in position
67 4’ using ammonia solution in MeOH, gave access to compound 12, which could be esterified by
68 LA or DHA through classical Steglish conditions, or desilylated to give the desired isopropylated
69 resveratrol (R-5OiP, 13). Final desilylation of the esterified compounds (14a and 14b) gave
70 access to the desired LA- and DHA-isopropylated resveratrol derivatives 15a and 15b using mild
71 NEt /3HF reagent.
3
72
First, cytotoxicity of resveratrol derivatives on ARPE-19 cell line was evaluated and is
73 represented in Figure 4. No dose dependent toxicity was observed for LA or DHA alkyl-
74 resveratrol 15a/b up to 160 µM. Introduction of only the isopropyl moiety on resveratrol (R-
75 5OiP, compound 13), led to highly toxic effects from 80 µM. Interestingly, association of LA or
76 DHA and isopropyl moieties (compounds 15a and 15b) resulted in absent or weak toxicity up to
77 160 µM, suggesting a stability of the lipophenol ester during the toxicity assay. Indeed, if the
78 ester link was easily cleaved, the release of the remaining R-5OiP (13) would have induced high
79 toxicity levels.
80

*
**
*
**
***
1
20
00
1
10 µM
40 µM
80 µM
120 µM
160 µM
***
***
*
** ***
*
** ***
80
60
40
20
0
*
** ***
*
**
2
2
81
82
2
2
2
2
2
2
83 Figure 4. Evaluation of cytotoxicity of resveratrol derivatives; natural resveratrol (R-com), R-
84 4’LA [42], R-4’DHA [43], R-5OiP (13), R-5OiP-4’LA (15a) and R-5OiP-4’DHA (15b). Results
85 are expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell
86 viability (MTT) after incubation of resveratrol derivatives (0-160 µM). The data are expressed as
87 the percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-
88 treated control cells.
2
2
2
2
2
2
2
2
2
2
2
3
89
90
In order to study the impact of the isopropyl, as well as the PUFA moieties, for anti-carbonyl
91 stress activity, commercial resveratrol (R-com), R-5OiP (13), R-4’LA and R-4’DHA [42,43]
92 were evaluated and compared to the new alkyl lipophenols 15a and 15b (Figure 5A).
93 Derivatives bearing both alkyl and PUFA substituents, presented the best protective effects under
94 carbonyl stress: 29% and 35% increase in viability using 80 µM of R-5OiP-4’LA (15a) and R-
95 5OiP-4’DHA (15b), respectively. By contrast, no significant cell protection was observed for the
96 four other derivatives (R-com, R-5OiP (13), R-4’LA and R-4’DHA), whose chemical structures
97 presented isopropyl, or PUFA substituent, or no additional substituent. These interesting results
98 confirm the importance of the O-isopropyl-resorcinol group, as well as the necessity of a PUFA
99 substituent for the anti-carbonyl stress activity, as was observed for the phloroglucinol LEAD
00 series [23]. In agreement with the toxicity results, the protective effect of the esterified

3
3
01 compounds (15a and 15b) and the loss in cell viability observed using R-5OiP (13), confirm the
02 stability of the ester bound of these lipophenols during the cell protection assay.
(B)
(A)
ns
*
**
100
*
*
**
***
6
0
0
0
0
***
**
***
*
**
8
6
4
2
0
0
0
0
0
*
**
*
***
*
**
***
*
**
***
***
4
2
*
** ***
***
*
**
***
***
***
A
A
)
)
)
m
a
)
)
)
'
L
3
b
m
A
A
R
3
a
b
o
H
5
'L
H
5
c
4
1
(
5
o
1
5
'
D
(
1
1
-
-
(
1
1
(
c
4
-4'
D
-
-
(
(
R
4
-
R
iP
R
R
iP
A
A
A
A
R
O
'
L
R
O
L
5
H
'
H
4
-
5
4
-
'D
-4
-
-
'
D
R
iP
iP
4
-
iP
-5O
iP
O
5
-
O
-5O
R
R
5
-
R
R
1
0 µM
40 µM
80 µM
10 µM
40 µM
80 µM
3
3
3
3
3
3
3
3
3
3
3
03 Figure 5. In vitro anti-COS evaluation of resveratrol derivatives. Comparison of natural
04 resveratrol (R-com), R-4’LA [42], R-4’DHA [43], R-5OiP (13), R-5OiP-4’LA (15a) and R-
05 5OiP-4’DHA (15b) activities. Results are expressed in mean ± SEM and are from n = 3-5
06 independent experiments. (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after
07 incubation of resveratrol derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as
08 the percentage of non-treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01,
09 *** p < 0.001, versus non-treated and exposed to atRAL cells. (B) Antioxidant assay:
10 representation of ROS production (DCFDA probe) after incubation of resveratrol derivatives (0-
11 80 µM) and H O (600 µM) in ARPE-19 cells. The data are expressed as the percentage of non-
2
2
12 treated and exposed to H O cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-treated and
2
2
2 2
13 exposed to H O cells.
3
3
3
3
3
3
14
15
The anti-carbonyl potential of resveratrol has been described by Wang et al. who demonstrated
16 the formation of an adduct by co-incubation of resveratrol with acrolein at equimolar
17 concentration [18], and by Shen et al. with methyl glyoxal [45]. Resveratrol was able to scavenge
18 toxic aldehydes and formed an heterocyclic ring at the C-2 and C-3 positions through
19 nucleophilic addition. Resveratrol protection of acrolein-treated cells was also attributed to a

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
20 direct stimulatory action on mitochondrial bioenergetics [46]. Here, in our cellular assay, natural
21 resveratrol did not display sufficient anti-carbonyl stress activity to counteract toxic effects of
22 high doses of atRAL. However, increasing the nucleophilicity of resorcinol backbone with an
23 isopropyl group, and increasing lipophilicity of this polyphenol by addition of a PUFA,
24 prevented aldehyde toxicity.
25
26 commercial resveratrol (R-com) was able to reach the oxidative status of cells that did not
27 received the oxidant stressor H . Logically, the introduction of a fatty acid moiety (LA or
Regarding the antioxidant profiles of resveratrol derivatives, shown in Figure 5B, 40 µM
2
O
2
28 DHA) and/or an alkyl residue masking the phenolic function, led to a loss of antioxidant capacity
29 compared to R-com. However, both the studied alkyl-lipophenols R-5OiP-4’LA (15a) and R-
30 5OiP-4’DHA (15b) were still potent antioxidant derivatives, as, at 80 µM, they decreased ROS
31 production by up to 47% and 37%, respectively. The reduction of the accessibility of two
32 phenolic functions on the resveratrol backbone still allowed 15a and 15b to be potent
33 antioxidants able to reduce oxidative status in a dose-dependent manner.
34
Thus, both resveratrol alkyl-lipophenols 15a and 15b exhibited an interesting protection
35 against the double COS in ARPE-19 cells.
36
37 3. Synthesis and anti-COS evaluation of flavonoid lipophenol analogues
38
39 worked on two natural flavonoid derivatives, quercetin and catechin (Figure 1).
40 Quercetin (3,3’,4’,5,7-pentahydroxyflavone, Q) is a plant flavonol, which is present in various
Flavonoids are widely known as potent antioxidants by direct or indirect action [13,47,48]. We
41 quantities in many fruit and vegetables, including apple, cranberry, red onion, asparagus,
42 spinach, walnuts and coriander. Over the past decades, quercetin has gained research interest due

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
43 to its numerous pharmacological activities. In addition to its antioxidant properties, quercetin
44 was proven to have anti-inflammatory, antidiabetic, anticancer, cardiovascular, hepato-
45 protective, antiplatelet, antibacterial and neuroprotective properties [49]. Regarding its
46 antioxidant action, quercetin acts by scavenging free radicals, and also increases antioxidant
47 enzymes, such as glutathione peroxidase, superoxide dismutase and catalase [50]. (+)-Catechin
48 (C) is a flavanol abundant in berries, chocolate, cacao and green tea that makes a significant
49 contribution to total dietary antioxidant intake. Catechin is the monomer of proanthocyanidin B2,
50 whose protection against A2E photo-oxidation-induced apoptosis has been shown in ARPE-19
51 cells [35]. Compared to quercetin, catechin should have increased nucleophilic properties
52 coming from the A ring, due to the absence of the carbonyl function link to this aromatic cycle.
53
Distinctive chemical structures related with flavonoid antioxidant activities have been
54 established including hydroxyl groups of the A-ring (resorcinol), ortho-dihydroxy arrangement
55 in the B-ring (catechol), and in the case of quercetin, C2-C3 unsaturated bond combined with C-
56 4 carbonyl group in the C-ring [48]. Moreover, the position instead of total number of hydroxyl
57 groups, considerably influences the efficiency of antioxidant activity [51]. The B-ring hydroxyl
58 structure is the utmost significant actor of scavenging oxygen free radicals [52]. In addition,
59 according to the study of Hong et al. [53], radical scavenging properties of catechin seem to be
60 less affected by acylation at the 3 or 7 position compared to acylation of the catechol moiety.
61 Thereby, in order to conserve the best antioxidant properties, the B-ring, corresponding to the
62 catechol moiety, was left free of any substituent, and position 3 was selected to introduce the
63 PUFA moiety on both catechin and quercetin. Finally, to preserve the alkylated resorcinol group
64 needed for the anti-carbonyl stress activity in both phloroglucinol and resveratrol series, two
65 positions of the A-ring, that mimic the resorcinol moiety of the LEADs A/B, were selected to

3
3
3
3
3
3
3
3
3
66 introduce the isopropyl; both positions 5 and 7 were alkylated in order to evaluate the most
67 favorable for carbonyl stress protection. As for resveratrol derivatives, lipophilic catechin and
68 quercetin bearing only the PUFA moiety (C-3PUFA and Q-3PUFA) [42], and alkyl-flavonoids
69 bearing only the isopropyl moiety (C-5OiP 18b, C-7OiP 18a, Q-5OiP 23 and Q-7OiP 35), were
70 produced to evaluate their impact on biological properties.
71
72 3.1. Catechin lipophenols synthesis and evaluation
73
Synthesis of alkyl-catechin derivatives C-7OiP (18a) and C-5OiP (18b), as well as the alkyl-
74 lipophenol derivatives C-3LA-7OiP (19a) and C-3LA-5OiP (19b), are presented in Scheme 3.
3
3
75
a
76 Scheme 3. Synthesis and chemical structures of (+)-catechin derivatives
Ph
Ph
Ph
Ph
OH
O
O
3
'
O
O
OH
'
4
B
HO
7
O
C
a
HO
O
b
R
1
O
O
A
3
OH
OH
OH
5
OH
OH
OR
2
(
+)-catechin
16
17a: R = iP, R = H
17b: R1 = H, R2 = iP
1 2
C-com
c
OH
OH
OH
OH
OH
OH
HO
O
R O
O
R O
O
1
1
d
OLA
OLA
OH
OH
OR
2
OR
2
C-3LA
19a: R
9b: R
1
= iP, R
2
= H
= iP
18a: R
18b: R
1
= iP, R
= H, R
2
2
= H
= iP
1
1
= H, R
2
1
3
77
a
3
3
3
3
78
79 K
Reagents and conditions: (a) Ph CCl , K CO , MeCN, rt, 23 h, 33%; (b) diisopropylsulfate,
2 2 2 3
2
CO
3
, acetone, 40 °C, 70 h, 37% separable mixture of 17a (19%) and 17b (18%); (c) H
2
,
80 Pd(OH)
81 68 h, 1.5% for 19a and 19b.
2
, THF/EtOH, rt, 18a : 42 h, 56% and 18b : 64 h, 72%; (d) LACOCl, TFA, 1,4-dioxan, rt,
3
82

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
83
The first step of the synthesis was the protection of the catechol moiety of catechin by a
84 diphenyldioxole in acetonitrile [54]. The alkylation step allowed access to both isomers 17a and
85 17b with 1/1 proportion in 37% yield. The two isomers were separated by column
86 chromatography and engaged separately for the end of the synthesis. Deprotection of the
87 catechol was performed by hydrogenation using palladium hydroxide and led to the desired
88 alkylated catechin derivatives: C-7OiP (compound 18a) and C-5OiP (compound 18b). The
89 lipophenols were then obtained using TFA and freshly prepared linoleyl chloride, as described
90 by Uesato et al. [55]. Compounds C-3LA-7OiP (19a) and C-3LA-5OiP (19b) were isolated in
91 sufficient quantities to be evaluated in vitro.
92
Commercial catechin (C-com), as well as the two alkylated catechins C-5OiP (18b) and C-
93 7OiP (18a), displayed no toxicity on ARPE-19 cells up to 160 µM (Figure 6). However, C-3LA
94 [42] and C-3LA-5OiP (19b) displayed an important cell death above 120 µM. Toxicity was even
95 higher for C-3LA-7OiP (19a) with no observed cell viability at 80 µM. In contrast with
96 resveratrol derivatives, alkylated catechins 18a and 18b displayed no toxicity, but association of
97 PUFA and alkyl on catechin backbone (19a and 19b), resulted in high toxicity in the ARPE-19
98 cell line.
99

*
**
***
*
**
**
120
1
00
10 µM
40 µM
80 µM
120 µM
*
**
*
**
8
0
***
*
**
6
0
*
**
4
0
***
**
*
1
60 µM
2
0
***
0
m
A
)
)
)
a
)
o
L
b
a
b
c
3
8
8
9
9
-
-
1
1
(
1
(
1
C
C
(
(
iP
i
P
i
P
iP
-7O
O
O
O
5
5
-
7
-
-
A
C
C
A
-3L
L
3
-
C
C
4
00
4
4
4
4
4
4
01 Figure 6. Evaluation of cytotoxicity of (+)-catechin derivatives; natural (+)-catechin (C-com), C-
02 3LA [42], C-5OiP (18b), C-7OiP (18a), C-3LA-5OiP (19b) and C-3LA-7OiP (19a). Results are
03 expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell viability
04 (MTT) after incubation of catechin derivatives (0-160 µM). The data are expressed as the
05 percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-
06 treated control cells.
4
4
4
4
4
4
4
4
4
4
4
4
07
08
Interestingly, despite the absence of cell protection using C-com, C-3LA was able to display a
09 mild protection against carbonyl stressor, with dose-dependent effects. The catechin ring, in
10 contrast to resveratrol or phloroglucinol ring, should be nucleophilic enough to show a protective
11 effect against carbonyl stressor, without the presence of isopropyl function on the resorcinol
12 moiety (responsible for inductive effect). The direct trapping of RCS, such as methylglyoxal (an
13 intermediate reactive carbonyl of AGE formation) by natural catechin was described in cell-free
14 experiments by Peng et al. [56]. Wang confirmed this observation in 2010, by identifying
15 adducts produced by co-incubation of catechin with glyoxal, methylglyoxal and acrolein [32].
16 Zhu et al. also demonstrated trapping of lipid-derived α,β-unsaturated aldehydes, which have
17 been implicated as causative agents in the development of carbonyl stress-associated pathologies
18 (i.e. 4-HNE and acrolein), by several (poly)phenols including catechin [31]. However, most of

4
4
4
4
4
4
4
4
4
4
4
4
19 those adducts have been observed only under simulated physiological conditions but rarely in
20 actual cellular media. As expected, in our work, the addition of isopropyl, preferentially at the 7
21 position (C-3LA-7OiP (19a), Figure 7A), increased cellular protection when tested at 40 µM
22 (below the toxic concentration of 19a). The position of the isopropyl group seemed to influence
23 anti-carbonyl stress activity, as well as the cytotoxic profile of the derivatives. Alkylation at the 7
24 position favor aldehyde trapping in both C8 and C6 positions, whereas alkylation at the 5
25 position orients the formation of adduct only with C6. Both sites were reported in the literature to
26 react with aldehyde, however, some work performed on the activity of epigallocatechin-3-O-
27 gallate (EGCG) led only to the identification of the C8 adduct [57]. Unexpectedly, the most
28 active lipophenol against carbonyl stress in this series, C-3LA-7OiP (19a), was also the most
29 toxic compound.
30
(A)
(B)
*
**
100
4
0
0
0
0
0
#
##
8
6
4
2
0
0
0
0
0
###
3
2
1
***
#
##
###
###
*
**
***
###
*
*
***
###
)
)
)
)
)
)
)
m
)
a
a
m
A
a
a
A
b
b
b
b
-co
L
9
o
L
9
8
8
9
8
8
9
3
c
3
1
1
1
1
(
-
-
1
1
1
1
(
-
(
(
(
(
(
(
C
C
C
C
iP
iP
C
i
P
iP
iP
iP
iP
iP
O
O
O
O
O
-7O
-5O
-7O
5
7
5
7
5
-
-
-
-
-
C
C
A
A
C
C
A
A
L
L
-3L
-3L
3
3
-
-
C
C
C
1
0 µM
40 µM
80 µM
10 µM
40 µM
80 µM
4
4
4
31 Figure 7. In vitro anti-COS evaluation of (+)-catechin derivatives. Comparison of natural (+)-
32 catechin (C-com), C-3LA [42], C-5OiP (18b), C-7OiP (18a), C-3LA-5OiP (19b) and C-3LA-
33 7OiP (19a) activities. Results are expressed in mean ± SEM and are from n = 3-5 independent

4
4
4
4
4
4
4
4
4
34 experiments. (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after incubation of
35 catechin derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as the percentage of
36 non-treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p < 0.001,
37 versus non-treated and exposed to atRAL cells. (B) Antioxidant assay: representation of ROS
38 production (DCFDA probe) after incubation of catechin derivatives (0-80 µM) and H O (600
2
2
39 µM) in ARPE-19 cells. The data are expressed as the percentage of non-treated and exposed to
40 H O cells. All conditions have a p-value < 0.001 versus non-treated and exposed to H O
2
2
2
2
2
41 control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus non-treated and non-exposed to H O
2
42 cells.
4
4
43
44
As expected, C-com, and the two alkylated derivatives C-5OiP (18b) and C-7OiP (18a),
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
45 displayed the best antioxidant activity with equivalent ROS production levels compared to
46 control cells non-exposed to H O , as shown in Figure 7B. With two substituents reducing free
2
2
47 phenolic functions, C-3LA-5OiP (19b) and C-3LA-7OiP (19a) displayed dose-dependent effects
48 on ROS produced by H O treatment, with a decrease by up to 60% and 67% in ROS production,
2
2
49 respectively, at 80 µM.
50 Despite these interesting antioxidant properties, C-3LA-5OiP (19b) lacked potent anti-
51 carbonyl stress activity (only 10% increase in viability observed at 80 µM) and C-3LA-7OiP
52 (19a) showed high toxicity (no survival from 80 µM), making these derivatives less interesting
53 anti-COS candidates compared to the resveratrol series.
54
55 3.2. Quercetin lipophenols synthesis and evaluation
56
Lipophilic quercetins Q-3LA and Q-3DHA were produced as previously described [42] and
57 two original pathways were developed to access the lipophenol derivatives of quercetin-5OiP
58 (29) and quercetin-7OiP (39a/b) (Scheme 4 and 5).
59
60 3.2.1. Quercetin-5OiP derivatives

4
4
4
4
61
A first chemical strategy was developed in order to access the lipophenol derivative Q-3LA-
62 5OiP (29) and the PUFA-free analogue Q-5OiP (23) (Scheme 4).
63
a
64 Scheme 4. Synthesis and chemical structures of quercetin-5OiP derivatives
OH
OH
HO
O
O
OH
O
23
d
OH
OBn
OBn
3
'
OH
'
OBn
OBn
4
B
HO
7
O
C
BnO
O
O
BnO
O
O
a, b
c
A
3
O
O
OH
5
OH
O
O
O
O
OH
O
O
OH
O
22
O
O
OH
OH
OH OH
OH OH
OH
OH
OH
OH
e
rutin
21
OTIPS
OTIPS
OTIPS
OTIPS
OBn
OBn
TIPSO
O
O
h
TIPSO
O
f, g
BnO
O
OH
OAc
OAc
O
O
O
O
O
2
7
26
24
i
OTIPS
OTIPS
OH
OH
OH
OH
TIPSO
O
O
HO
O
HO
O
O
j
OLA
OLA
OLA
O
O
O
OH
2
8
29
Q-3LA
4
65
66
a
4
4
4
4
4
Reagents and conditions: (a) BnBr, K CO , DMF, 40 °C, 4 h, 61%; (b) 2-bromopropane,
2 3
67 K CO , DMF, 80 °C, 18 h, 88%; (c) HCl, EtOH, 70 °C, 5 h, 98%; (d) H , Pd/C, THF/EtOH, rt,
68 21 h, 69%; (e) pyridine, Ac O, rt, 20 h, 100%; (f) H , Pd/C, THF/EtOH, rt, 15 h, 100%; (g)
69 TIPS-OTf, NEt , THF, rt, 4 h, 62%; (h) NH /MeOH, DCM, 0 °C, 3 h, 88%; (i) LA, DCC,
2
3
2
2
2
3
3
70 DMAP, DCM, rt, 23 h, 54%; (j) NEt /3HF, THF, rt, 15 min, 72%.
3
4
71

4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
72
In order to access quercetin-5OiP derivatives, we started from commercially available rutin
73 using the diholoside rutinoside as protecting group for the phenolic function at the 3 position.
74 The first step was the protection of the phenols in positions 7, 3’ and 4’ with benzyl groups.
75 Alkylation of the phenol in position 5 was performed with 2-bromopropane by heating at 80 °C
76 in dry DMF (88%). Cleavage of rutinoside in acidic conditions gave compound 22, which was
77 either de-benzylated to access desired Q-5OiP (23), or acetylated in position 3 to pursue the
78 lipophenol synthesis. Benzyl groups were removed by hydrogenation to be replaced by silyl
79 protecting groups, which can be easily deprotected without damaging PUFA moiety leading to
80 compound 26. Acetate in position 3 was then cleaved using ammonia solution in MeOH at 0 °C
81 to allow esterification with fatty acid. Final deprotection of TIPS group gave the desired Q-3LA-
82 5OiP (29).
83
For comparison purposes, commercial quercetin (Q-com) and Q-3LA [42] properties were also
84 evaluated in vitro. Under the conditions tested, Q-com was found to be toxic for the ARPE-19
85 cell line from 80 µM (Figure 8). Introduction of the isopropyl moiety (Q-5OiP, 23) reduced the
86 toxicity with mild mortality starting at 160 µM. A similar reduction was observed with PUFA
87 introduction, as Q-3LA displayed no toxicity up to 160 µM, and was found to increase cell
88 viability (156% cell viability at 80 µM). However, the alkyl-lipophenol derivative Q-3LA-5OiP
89 (29) was very toxic, with no survival at 80 µM.
90

*
**
1
1
1
1
8
6
60
40
20
00
0
0
*** ***
*
**
***
*
**
*
**
**
**
*
*
10 µM
40 µM
*
**
*
**
*
8
1
0 µM
**
***
20 µM
4
20
0
*
** ***
***
160 µM
0
)
)
m
A
9
(23
o
L
2
c
3
-
(
-
Q
Q
iP
iP
O
-5O
5
-
A
Q
L
3
-
Q
4
91
4
4
4
4
4
92 Figure 8. Evaluation of cytotoxicity of quercetin-5OiP derivatives; natural quercetin (Q-com),
93 Q-3LA [42], Q-5OiP (23) and Q-3LA-5OiP (29). Results are expressed in mean ± SEM and are
94 from at n = 3-5 independent experiments. ARPE-19 cell viability (MTT) after incubation of
95 quercetin-5OiP derivatives (0-160 µM). The data are expressed as the percentage of non-treated
96 control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-treated control cells.
4
4
4
5
5
5
5
5
5
97
98
Regarding anti-carbonyl stress properties, here again the presence of isopropyl, as well as the
99 PUFA moiety, seemed decisive for cell protection. As shown in Figure 9A, Q-com, Q-3LA and
00 Q-5OiP (23) did not display any protective effect against atRAL toxicity, whereas Q-3LA-5OiP
01 (29) protected cells with 50% increase in cell survival at 80 µM. Even if toxic at 80 µM, a high
02 anti-carbonyl stress activity was observed at this concentration and can be explained by the
03 differences between cytotoxicity and anti-carbonyl stress activity protocols (different incubation
04 times).
05

(A)
(B)
*
**
8
6
4
2
0
0
0
0
0
100
*
**
###
80
###
#
##
###
###
*
*
60
#
## ###
###
#
##
4
0
0
0
2
)
)
)
)
m
A
3
(29
m
A
3
9
o
L
2
o
L
2
2
(
c
3
(
c
3
-
(
-
-
-
P
iP
P
i
Q
Q
i
Q
Q
i
P
O
O
O
O
5
5
5
5
-
-
-
-
Q
A
Q
A
L
L
3
3
-
-
Q
Q
1
0 µM
40 µM
80 µM
10 µM
40 µM
80 µM
5
5
5
5
5
5
5
5
5
5
5
06 Figure 9. In vitro anti-COS evaluation of quercetin-5OiP derivatives. Comparison of natural
07 quercetin (Q-com), Q-3LA [42], Q-5OiP (23) and Q-3LA-5OiP (29) activities. Results are
08 expressed as mean ± SEM and are from n = 3-5 independent experiments. (A) Anti-carbonyl
09 stress assay: ARPE-19 cell viability (MTT) after incubation of quercetin-5OiP derivatives (0-80
10 µM) and atRAL (15 µM). The data are expressed as the percentage of non-treated and non-
11 exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-treated and
12 exposed to atRAL cells. (B) Antioxidant assay: representation of ROS production (DCFDA
13 probe) after incubation of quercetin-5OiP derivatives (0-80 µM) and H O (600 µM) in ARPE-
2
2
14 19 cells. The data are expressed as the percentage of non-treated and exposed to H O cells. All
2
2
15 conditions have a p-value < 0.001 versus non-treated and exposed to H
16 0.05, ## p < 0.01, ### p < 0.001, versus non-treated and non-exposed to H
2
O
2
control cells. # p <
2
cells.
2
O
5
5
17
18
Finally, antioxidant capacity was evaluated and is reported in Figure 9B. The Di-substituted
5
5
5
5
5
5
19 lipophenol derivative Q-3LA-5OiP (29) was efficient to reduce H O -induced ROS production
2
2
20 by 53% at 80 µM. Here again, a dose-dependent antioxidant potency was still observable despite
21 the reduction of two free phenolic functions of quercetin. A similar protection profile was
22 observed for the isopropyl-free analogue Q-3LA, which suggests that position 5 is not primordial
23 for ROS scavenging activity. This was also confirmed by comparing Q-com and Q-5OiP (23)
24 cell treatment, which led to low ROS levels.

5
5
5
5
5
5
5
5
5
5
25
26
To conclude, in vitro assessments indicate that in this series, alkylated resorcinol and PUFA
27 moieties are necessary for anti-carbonyl stress activity, as observed for the phloroglucinol [23]
28 and resveratrol series. Derivative Q-3LA-5OiP (29) displays interesting anti-COS properties,
29 however, its high cytotoxicity on ARPE-19 has to be considered for further evaluations.
30
31 3.2.2. Quercetin-7OiP derivatives
32
A second chemical strategy, presented in Scheme 5, was developed for the synthesis of alkyl
33 quercetin Q-7OiP (35) and the two lipophenol derivatives Q-3LA-7OiP (39a) and Q-3DHA-
34 7OiP (39b).
5
5
35
a
36 Scheme 5. Synthesis and chemical structures of quercetin-7OiP derivatives