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.
3 Espérance Moine a,*, Manel Boukhallat a, David Cia c, Nathalie Jacquemot c, Laurent Guillou b,
4 Thierry Durand a, Joseph Vercauteren a, Philippe Brabet b and Céline Crauste a,*.
5
6
7
8
9
a 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
de Médecine et de Pharmacie, Clermont-Ferrand, 63000, France
10
11
12 * Correspondence: celine.crauste@umontpellier.fr; Tel.: +33-4-11-75-95-66,
13
esperance.moine@umontpellier.fr; Tel.: +33-4-11-75-95-66
14 Keywords
15 Lipophenol, PUFA, carbonyl stress, antioxidant, macular degeneration, structure-activity
16 relationship.
17
18 Abbreviations
19
A2E, N-retinylidene-N-retinylethanolamine; ABCA4, ATP binding cassette subfamily A
20 member 4; AGEs, advanced glycation end products; AMD, age-related macular degeneration;

21 ARPE-19, adult retinal pigment epithelial cell line-19; ASAP, atmospheric solids analysis probe;
22 atRAL, all-trans-retinal; BRB, blood retina barrier; BSMR, based on starting material recovery;
23 COS, carbonyl and oxidative stresses; DCC, dicyclohexylcarbodiimide; DCFDA, 2’,7’-
24 dichlorofluorescin diacetate; DCM, dichloromethane; DHA, docosahexaenoic acid; DMAP,
25 dimethylaminopyridine; DMEM/F12, Dulbecco’s Modified Eagle’s Medium (DMEM)/Ham
26 F12; DMF, dimethylformamide; DMSO, dimethylsulfoxide; DNA, deoxyribonucleic acid; EC50,
27 efficiency concentration 50; EGCG, epigallocatechin-3-O-gallate; EPA, eicosapentaenoic acid;
28 ESI, electrospray ionization; FBS, fetal bovine serum; 4-HHE, 4-hydroxyhexenal; 4-HNE, 4-
29 hydroxynonenal; HRMS, high resolution mass spectroscopy; LA, linoleic acid; Mp, melting
30 point; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; NMR, nuclear
31 magnetic resonance; Nrf2/Keap1, Nuclear factor (erythroid-derived 2)-like 2/Kelch-like ECH-
32 associated protein 1; PUFA, poly-unsaturated fatty acid; RCS, reactive carbonyl species; Rf,
33 retardation factor; ROS, reactive oxygen species; RPE, retinal pigment epithelium; TFA,
34 trifluoroacetic acid; THF, tetrahydrofuran; TIPS, triisopropylsilyl; TLC, thin layer
35 chromatography; UPLC, ultra-performance liquid chromatography; VEGF, vascular endothelial
36 growth factor.
37
38 Abstract
39
Dry age-related macular degeneration and Stargardt disease undergo a known toxic
40 mechanism caused by carbonyl and oxidative stresses (COS). This is responsible for
41 accumulation in the retinal pigment epithelium (RPE) of A2E, a main toxic pyridinium bis-
42 retinoid lipofuscin component. Previous studies have shown that carbonyl stress in retinal cells
43 could be reduced by an alkyl-phloroglucinol-DHA conjugate (lipophenol). Here, we performed a

44 rational design of different families of lipophenols to conserve anti-carbonyl stress activities and
45 improve antioxidant properties. Five synthetic pathways leading to alkyl-(poly)phenol
46 derivatives, with phloroglucinol, resveratrol, catechin and quercetin as the main backbone, linked
47 to poly-unsaturated fatty acid, are presented. These lipophenols were evaluated in ARPE-19 cell
48 line for their anti-COS properties and a structure-activity relationship study is proposed.
49 Protection of ARPE-19 cells against A2E toxicity was assessed for the four best candidates.
50 Finally, interesting anti-COS properties of the most promising quercetin lipophenol were
51 confirmed in primary RPE cells.
52
53
54
Introduction
Oxidative stress, resulting from an overproduction of reactive oxygen species (ROS) within
55 cells or in the extracellular matrix, highly damages key cellular proteins, lipids and DNA.
56 Reactive carbonyl species (RCS), such as sugars and osones, endogenous aldehydes or
57 metabolites derived from lipid oxidation, are involved in glycation and cross-linking reactions
58 and thus, affect cellular viability leading to tissue injury. These two carbonyl and oxidative stress
59 (COS) mechanisms play a crucial role in aging-associated pathologies, like age-related macular
60 degeneration (AMD), or in some inherited forms of macular degeneration, such as Stargardt
61 disease [1–3]. AMD is one of the primary causes of central and irreversible visual loss among
62 the elderly in occidental countries, and there is no treatment currently available to stop retinal
63 degeneration, especially in the prevalent dry form (80-85% of cases). COS mechanisms are
64 responsible for the accumulation in retinal pigment epithelium (RPE) of a toxic bis-retinoid
65 conjugate called A2E (its photoisomers and its oxidized metabolites). Pathologic A2E
66 biosynthesis occurs when all-trans-retinal (atRAL), rather than undergoing reduction to retinol in

67 the RPE, accumulates abnormally in photoreceptors. This accumulation can be due to age
68 (AMD) or loss of function of the ABCA4 transporter (Stargardt disease) [2]. Two molecules of
69 this reactive aldehyde (RCS) in excess, condense with one molecule of phosphatidyl
70 ethanolamine (carbonyl stress step) into an unstable dihydropyridinium bis-retinoid.
71 Transformation of the dihydropyridinium bis-retinoid into A2E then arises during the oxidative
72 step leading to a more stable pyridinium form. A2E and its various oxidative metabolites are the
73 major constituent of lipofuscin, a marker of AMD. The massive accumulation of lipofuscin in the
74 RPE (following phagocytosis of photoreceptor outer segments) is cytotoxic and causes
75 progressive RPE cell death and subsequent photoreceptor degeneration [4]. A2E cytotoxicity is
76 explained by several mechanisms: loss of membrane integrity due to its amphiphilic properties
77 during membrane integration; increase of oxidative stress by the generation of singlet oxygen
78 during exposition to blue light; metabolization in A2E-oxidized metabolites resulting in DNA
79 lesions [5–7]. Limitation of A2E biosynthesis (by clearance of atRAL) and oxidation is therefore
80 an attractive target to slow the progression of macular degeneration [8].
81
Based on epidemiology studies, natural plant antioxidants, such as (poly)phenols, secondary
82 metabolites that protect plants from several aggressions, are efficient at protecting animal
83 organisms against oxidative stress. Some (poly)phenol structures have been reported to protect a
84 variety of retinal cell types from oxidative stress-induced cell death [9–11]. Such activity may be
85 related to their capacity to scavenge directly excess ROS, or to stimulate the enzymatic
86 antioxidant defenses of the organism through Nrf2/Keap1 signaling pathway[12,13]. Moreover,
87 recent literature addressed the efficiency of (poly)phenols to inhibit Advanced Glycation End
88 product formation (AGEs), which result from both carbonyl and oxidative stresses [14,15], and
89 to act also as anti-carbonyl stress derivatives by trapping toxic RCS (glyoxal, acrolein, 4-

90 hydroxynonenal (4-HNE)) [16–18].
91
Unfortunately, while showing interesting in vitro protection, the limited bioavailability of most
92 (poly)phenols (weak drug absorption, high systemic metabolization, restricted cell
93 penetration…) negatively influences their in vivo potency and thus their development as drug
94 candidates. Increasing the lipophilicity of those (poly)phenols may be a useful way to improve i)
95 their protective effect on lipid membranes or lipid derivatives (such as A2E), ii) their absorption
96 [19], and iii) their formulation for in vivo administration [20–22]. In our previous work, we
97 synthesized lipophenols (or phenolipids), polyunsaturated fatty acid (PUFA) linked to alkyl-
98 phloroglucinol (LEAD A/B, Figure 1), that were designed to reduce carbonyl stress associated
99 with retinal dystrophies [23,24]. The main objective of those lipophenols was to avoid/slow the
100 pace of A2E formation by scavenging atRAL (Figure 1). To increase lipophilicity and cellular
101 permeability of starting (poly)phenol, PUFAs were preferred to saturated ones, due to their
102 observed benefits in an AMD clinical trial [25]. Both isopropyl-phloroglucinols linked to linoleic
103 acid (C18:2 n-6; LA - LEAD A, Figure 1) and docosahexaenoic acid (C22:6 n-3; DHA - LEAD
104 B, Figure 1) showed promising protection of RPE cells against atRAL toxicity. A isopropyl
105 functional group was rationally selected to increase the nucleophilic properties of the aromatic
106 cycle involved in the trapping of aldehyde function of atRAL and leading to the formation of a
107 stable non-toxic chromene adduct (Figure 1) [23,26]. Both isopropyl and PUFA parts were
108 proven to be indispensable in a phloroglucinol series (LEAD A and LEAD B) to ensure cell
109 protection against carbonyl stress in cellular assays [23,24]. The DHA analogue (LEAD B) was
110 selected for further in vivo evaluations due to several benefits of this omega-3 lipid: i) the high
111 proportion of DHA in the membrane of photoreceptor outer segments, ii) its ability to reach the
112 retina through PUFA transports [27]; iii) recent data showing that dietary supplementation with

113 high doses of DHA/EPA significantly improves the visual acuity of AMD patients [25]; and iv)
114 the report that specific oxidative metabolites of DHA (Neuroprotectin D1) protect RPE cells
115 against oxidative stress [28]. Intravenous administration of LEAD B allowed photoreceptor
116 protection against acute light-induced degeneration in a mouse model used for the development
117 of novel therapeutics for Stargardt disease (Abca4-/- mice) [29,30]. Despite interesting
118 preliminary in vivo protection observed using LEAD B, this phloroglucinol derivative lacks
119 efficiency to reduce ROS production and cellular oxidative stress. In the present study, we report
120 new potent alkyl-lipophenols that act on both carbonyl and oxidative stresses, to reduce atRAL
121 toxicity, and oxidation associated with A2E formation, metabolism and toxicity, critical steps
122 involved in photoreceptor degeneration.
123
Several possibilities have been proposed to increase antioxidant activities while preserving
124 anti-carbonyl stress properties. First, using the same (poly)phenol backbone, PUFA was linked
125 directly to the aromatic cycle to free one phenolic function of the phloroglucinol involved in the
126 antioxidant activity (C-phloroglucinol series, Figure 1). Second, new series of lipophenols were
127 also developed from natural (poly)phenols, different from phloroglucinol, meeting the criterion
128 that the starting (poly)phenol contained an isopropyl-resorcinol framework mimicking the
129 phloroglucinol backbone, as this was indispensable for the protection against carbonyl stress
130 [23,24]. More effective natural antioxidants, such as resveratrol stilbenoid (true vinylogous
131 analogue of phloroglucinol), or flavonoids, such as quercetin or catechin, all of which were
132 reported to trap aldehyde function in cell free assays [18,31,32] and to protect ARPE-19 cells
133 against A2E photo-oxidation damage [33–35], have been proposed as new anti-COS alkyl-
134 lipophenol series (Figure 1). Nine novel alkyl-lipophenol-PUFAs have been synthesized using
135 original chemical strategies. Toxicity assays, protection against a lethal concentration of atRAL

136 and reduction of H2O2 induced ROS production, were used to first screen these molecules in
137 ARPE-19 cells. The necessity of the isopropyl and the PUFA parts for carbonyl stress protection
138 has been highlighted with all new (poly)phenol backbones (resveratrol, catechin and quercetin).
139 The ability to preserve antioxidant properties, despite reduction of free phenolic functions
140 compared to native (poly)phenols, has been studied. Best DHA-lipophenol conjugates offering
141 both oxidative and carbonyl stress protection, were compared with LEAD B, for their ability to
142 reduce A2E toxicity during photo-oxidation process in RPE. According to the comparison of
143 both carbonyl and oxidative stress protection, a new alkyl-quercetin-PUFA conjugate has been
144 highlighted. Cell protection was finally validated with this lipophenol in primary RPE cells, as
145 primary cultures are likely to reflect in vivo cell morphology and function more accurately.
146
LEAD phloroglucinol series
atRAL scavenging by phloroglucinol [26]
OH
OPUFA
O
atRAL
O
OH
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
147
148
Figure 1. Chemical structures of developed lipophenol series.
149
150
Results and Discussion
151 1. Synthesis and anti-COS evaluation of C-phloroglucinol derivatives

152
The first chemical changes considered to increase antioxidant capacity concerned the link of
153 the PUFA moiety. Indeed, releasing one more phenolic function on the phloroglucinol (P)
154 backbone should allow better ROS scavenging properties. This modification was investigated
155 through a C-alkylation on the phloroglucinol aromatic ring, by the introduction of a
156 hydroxylated-alkyl spacer, in order to perform PUFA functionalization. The synthesis of the two
157 C-linker analogues (8a/b) of LEAD A (P-OiP-OLA) and LEAD B (P-OiP-ODHA) is described
158 in Scheme 1.
159
160
Scheme 1. Chemical pathway leading to C-phloroglucinol derivativesa
161
162
aReagents and conditions: (a) Zn(CN)2, ZnCl2, Et2O-HCl (2N), rt, 16 h and H2O, reflux, 20
163 min or POCl3, EtOAc, DMF, rt, 86 h, 64%; (b) Ph3PCHCOOMe, DCM, rt, 18 h, 59%; (c) H2,
164 Pd/C, EtOAc, rt, 4 h, 71%; (d) TIPS-OTf, NEt3, DCM, rt, 2 h, 35%; (e) LiAlH4 (1M), Et2O, 0
165 °C, 4 h30, 77%; (f) PUFA, DCC, DMAP, DCM, rt to 50 °C, 7 h, 7a: 66% and 7b: 28%; (g)
166 NEt3/3HF, THF, rt, 19 h, 8a: 39% and 8b: 30%.
167

168
Starting from mono-isopropyl phloroglucinol 1 [23], a Gattermann-Koch formylation [36] lead
169 to salicylic aldehydes 2a/2b (80/20) in 64% combined yield. The next Wittig reaction was
170 performed on the mixture of 2a/2b for a two-carbon homologation, and the resulting two α,β-
171 unsaturated esters 3a/3b were reduced under hydrogen in the presence of Pd/C catalyst which
172 after purification lead to the isolation of isomer 4a. Phenolic functions were then protected by
173 silylated protecting groups (compound 5). The alcohol 6 was obtained after the ester reduction
174 and was coupled either with LA or DHA using classical Steglich coupling conditions. Finally,
175 LA and DHA desired conjugates (respectively 8a and 8b) were obtained after deprotection of
176 silylated phenolic function in the presence of NEt3/3HF.
177
In vitro evaluations of the C-phloroglucinol derivatives were performed on the ARPE-19 cell
178 line. All compounds were tested at various concentrations in the range of 0-160 µM for
179 cytotoxicity assessment, and in the range of 0-80 µM for anti-COS evaluation. Regarding the
180 new C-phloroglucinol derivatives (8a and 8b), our main concerns were to maintain anti-carbonyl
181 stress activity, despite the reduction of one nucleophilic C-alkylation site (on the aromatic cycle)
182 compared to LEAD A/B structures, and to increase antioxidant properties by releasing a phenolic
183 function. As shown in Figure 2, there is no apparent cytotoxicity due to the modification of the
184 lipid position.
185

100
80
60
40
20
0
**
**
*** ***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
10 µM
40 µM
80 µM
120 µM
160 µM
)
)
)
)
A
a
B
b
8
8
(
D
(
D
A
A
A
L
A
E
E
L
H
L
(
C
(
D
-
A
C
-
A
P
i
L
H
P
i
O
O
-
D
-
O
P
P
O
-
i
-
P
O
P
i
-
P
O
-
P
186
187 Figure 2. Evaluation of cytotoxicity of C-phloroglucinol derivatives; P-OiP-OLA (LEAD A)
188 [24], P-OiP-CLA (8a), P-OiP-ODHA (LEAD B) [23] and P-OiP-CDHA (8b). Results are
189 expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell viability
190 (MTT) after incubation of phloroglucinol derivatives (0-160 µM). The data are expressed as the
191 percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-
192 treated control cells.
193
194
Anti-carbonyl stress activity of C-phloroglucinol derivatives is evaluated by comparing cell
195 survival in presence of a toxic concentration of carbonyl stressor atRAL, and is represented in
196 Figure 3A. The two new C-phloroglucinol derivatives (8a and 8b) displayed interesting
197 protective properties, increasing cell viability (at 80 µM) of + 25% for LA derivative (8a) and +
198 28% for DHA derivative (8b), compared to non-treated and atRAL-exposed cells. No statistical
199 differences in cell protection were observed under these conditions between P-OiP-OLA (LEAD
200 A) versus P-OiP-CLA (8a) and between P-OiP-ODHA (LEAD B) versus P-OiP-CDHA (8b), at
201 equivalent concentrations. Modification of the PUFA position on the phloroglucinol backbone
202 did not affect the anti-carbonyl stress activity of the derivative: the PUFA moiety can be
203 introduced indifferently on a phenolic function or with a spacer on the aromatic core. This result

204 is in agreement with the literature, as phloretine, a natural phloroglucinol analogue with a
205 substituent on the aromatic ring, is also able to trap toxic aldehydes, such as acrolein, a toxic
206 aldehyde found in cigarette smoke and involved in AMD via induction of oxidative damage in
207 RPE cells, or 4-HNE, derived from lipid peroxidation [31].
208
(A)
(B)
60
***
***
***
120
100
80
60
40
20
0
***
***
50
40
30
20
10
0
***
***
***
°
°°°
###
**
)
)
)
)
a
b
A
B
8
8
(
(
D
D
)
)
)
A
A
P
A
A
)
a
b
L
B
E
E
H
A
8
8
L
C
L
(
(
(
D
-
D
(
D
C
P
i
A
A
A
A
-
A
A
L
E
H
E
L
P
O
H
i
L
C
-
L
(
D
(
-
O
-
D
O
C
-
P
-
i
A
O
A
P
-
i
P
P
L
H
O
-
i
P
O
i
D
O
-
O
-
P
-
O
O
P
-
P
-
P
i
P
P
i
O
-
O
P
10 µM
40 µM
80 µM
-
P
10 µM
40 µM
80 µM
209 Figure 3. In vitro anti-COS evaluation of C-phloroglucinol derivatives. Comparison of P-OiP-
210 OLA (LEAD A) [24], P-OiP-CLA (8a), P-OiP-ODHA (LEAD B) [23] and P-OiP-CDHA (8b)
211 activities. Results are expressed in mean ± SEM and are from n = 3-5 independent experiments.
212 (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after incubation of phloroglucinol
213 derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as the percentage of non-
214 treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus
215 non-treated and exposed to atRAL cells. (B) Antioxidant assay: representation of ROS
216 production (DCFDA probe) after incubation of phloroglucinol derivatives (0-80 µM) and H2O2
217 (600 µM) in ARPE-19 cells. The data are expressed as the percentage of non-treated and
218 exposed to H2O2 cells. All conditions have a p-value < 0.001 versus non-treated and exposed to
219 H2O2 control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus P-OiP-OLA (LEAD A) at the
220 same concentration. ° p < 0.05, °° p < 0.01, °°° p < 0.001, versus P-OiP-ODHA (LEAD B) at the
221 same concentration.
222

223
The main objective of the synthesis of compounds 8a and 8b was to increase antioxidant
224 properties of the LEADs A and B in cell experiments. As expected, 8a and 8b logically showed
225 weaker antioxidant activity compared to phloroglucinol (data not shown). Compared to untreated
226 cells, C-phloroglucinol analogues were able to decrease H2O2 induced ROS production by 41%
227 for LA derivative (8a) and 46% for DHA derivative (8b) at 80 µM (Figure 3B). Thus, a
228 significant improvement in antioxidant activity was observed using C-phloroglucinols (8a/8b)
229 that show a dose-dependent effect compared to LEAD A and B. Direct hydroxyl radical
230 scavenging activity of phloroglucinol has been previously reported [26,37], and the liberation of
231 one phenolic function explains the increased antioxidant efficacy of C-phloroglucinol derivatives
232 (8a and 8b) compared to LEAD A and LEAD B, that reduce ROS production by only 25% and
233 30% at 80 µM, respectively.
234
235 2. Synthesis and anti-COS evaluation of resveratrol lipophenol analogues
236
Resveratrol (R),
a
major active phytoalexin from the stilbene family (3,4’,5-
237 trihydroxystilbene), is mainly found in Vitis vinifera (Vitaceae) stalks and in the roots of
238 Fallopia japonica var. japonica (Polygonaceae). Resveratrol has a wide range of biological
239 activities including anti-bacterial and anti-fungal properties [38], antioxidant activity [10,39],
240 and anti-inflammatory properties [40]. Furthermore, previous evaluations on RPE cell assays
241 highlighted the benefit of resveratrol in reducing VEGF effect and increasing natural antioxidant
242 enzymatic and molecular defenses [40]. Resveratrol should not only be considered as a simple
243 ROS scavenger. The work of Vlachogianni et al. shows better radical scavenging activity of 4’-
244 acetylated resveratrol compared to 3-acetylated analogues [41]; we decided therefore to link the
245 PUFA at the 4’ position of resveratrol. The isopropyl moiety was introduced at the 5 position in

246 order to keep the alkyl-resorcinol moiety, which was proven to be responsible for high anti-
247 carbonyl stress activity on phloroglucinol series [23]. In order to confirm the necessity of alkyl as
248 well as PUFA moieties on the resveratrol backbone, as we did previously with phloroglucinol
249 [23,24], we produced in addition to alkyl resveratrol-PUFA (15a/b), an isopropyl-resveratrol
250 derivative not linked to PUFA (R-5OiP 13, Scheme 2) and derivatives linked to PUFAs without
251 alkyl function named resveratrol-4’LA (R-4’LA) and resveratrol-4’DHA (R-4’DHA), both
252 synthesized as previously described [42,43].
253 Scheme 2. Synthesis and chemical structures of resveratrol derivativesa
OH
OH
OH
3
a
b
A
HO
5
HO
O
B
4'
OH
OBu
OBu
resveratrol
R-com
10
9
c, d
OH
OTIPS
OTIPS
g
f
O
O
O
OPUFA
OPUFA
OH
12
e
15a: PUFA = LA
15b: PUFA = DHA
14a: PUFA = LA
14b: PUFA = DHA
OH
OH
OH
HO
HO
O
ODHA
OLA
OH
R-4'DHA
R-4'LA
13
254
255
aReagents and conditions: (a) 2-methyl-butan-2-ol, vinyl butyrate, Novozyme 435 (CALB), 40
256 °C, 8 days, 52%; (b) diisopropylsulfate, K2CO3, acetone, rt, 24 h, 22% (67% BSMR); (c) TIPS-
257 OTf, DIPEA, THF, rt, 6 h, 63%; (d) NH3/MeOH, DCM, 0 to 5 °C, 24 h, 80%; (e) NEt3/3HF,
258 THF, rt, 23 h, 67%; (f) PUFA, DCC, DMAP, DCM, rt, 14a: 35 min, 83% and 14b: 2 h, 46%; (g)
259 NEt3/3HF, THF, rt, 15a: 19 h, 67% and 15b: 20 h, 64%.
260

261
A tedious step to access the desired alkylated resveratrols 15a/b, was the isopropylation of the
262 5 position of resveratrol as the 4’ position is the more reactive. Utilization of the supported lipase
263 Novozyme 435 (Immobilized lipase from Candida antarctica) was efficient and selective to
264 protect the 4’ position of resveratrol, using vinyl butyrate (52%). Alkylated compound 10 was
265 then obtained using diisopropylsulfate and K2CO3 at room temperature for 24 h [44]. Successive
266 protection of the resulting phenol with TIPS groups, and deprotection of the butyrate in position
267 4’ using ammonia solution in MeOH, gave access to compound 12, which could be esterified by
268 LA or DHA through classical Steglish conditions, or desilylated to give the desired isopropylated
269 resveratrol (R-5OiP, 13). Final desilylation of the esterified compounds (14a and 14b) gave
270 access to the desired LA- and DHA-isopropylated resveratrol derivatives 15a and 15b using mild
271 NEt3/3HF reagent.
272
First, cytotoxicity of resveratrol derivatives on ARPE-19 cell line was evaluated and is
273 represented in Figure 4. No dose dependent toxicity was observed for LA or DHA alkyl-
274 resveratrol 15a/b up to 160 µM. Introduction of only the isopropyl moiety on resveratrol (R-
275 5OiP, compound 13), led to highly toxic effects from 80 µM. Interestingly, association of LA or
276 DHA and isopropyl moieties (compounds 15a and 15b) resulted in absent or weak toxicity up to
277 160 µM, suggesting a stability of the lipophenol ester during the toxicity assay. Indeed, if the
278 ester link was easily cleaved, the release of the remaining R-5OiP (13) would have induced high
279 toxicity levels.
280

***
***
***
120
100
80
60
40
20
0
10 µM
40 µM
80 µM
120 µM
160 µM
***
***
***
***
*** ***
*** ***
***
281
282
283 Figure 4. Evaluation of cytotoxicity of resveratrol derivatives; natural resveratrol (R-com), R-
284 4’LA [42], R-4’DHA [43], R-5OiP (13), R-5OiP-4’LA (15a) and R-5OiP-4’DHA (15b). Results
285 are expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell
286 viability (MTT) after incubation of resveratrol derivatives (0-160 µM). The data are expressed as
287 the percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-
288 treated control cells.
289
290
In order to study the impact of the isopropyl, as well as the PUFA moieties, for anti-carbonyl
291 stress activity, commercial resveratrol (R-com), R-5OiP (13), R-4’LA and R-4’DHA [42,43]
292 were evaluated and compared to the new alkyl lipophenols 15a and 15b (Figure 5A).
293 Derivatives bearing both alkyl and PUFA substituents, presented the best protective effects under
294 carbonyl stress: 29% and 35% increase in viability using 80 µM of R-5OiP-4’LA (15a) and R-
295 5OiP-4’DHA (15b), respectively. By contrast, no significant cell protection was observed for the
296 four other derivatives (R-com, R-5OiP (13), R-4’LA and R-4’DHA), whose chemical structures
297 presented isopropyl, or PUFA substituent, or no additional substituent. These interesting results
298 confirm the importance of the O-isopropyl-resorcinol group, as well as the necessity of a PUFA
299 substituent for the anti-carbonyl stress activity, as was observed for the phloroglucinol LEAD
300 series [23]. In agreement with the toxicity results, the protective effect of the esterified

301 compounds (15a and 15b) and the loss in cell viability observed using R-5OiP (13), confirm the
302 stability of the ester bound of these lipophenols during the cell protection assay.
(B)
(A)
ns
100
80
60
40
20
0
***
*
***
***
***
***
60
40
20
0
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
)
)
)
A
A
m
)
)
a
)
3
b
L
'
A
A
R
o
m
H
a
3
b
5
1
5
L
'
o
c
H
4
(
5
1
5
D
'
-
1
1
(
-
(
c
4
(
1
1
D
'
-
-
4
(
(
R
R
P
5
-
i
4
P
5
R
R
A
A
i
-
R
A
A
O
L
'
H
R
O
L
'
H
5
-
4
5
D
'
4
-
-
D
'
-
R
4
P
4
-
i
P
-
i
P
i
O
P
O
i
-
O
-
O
5
-
R
5
R
-
R
R
10 µM
40 µM
80 µM
10 µM
40 µM
80 µM
303 Figure 5. In vitro anti-COS evaluation of resveratrol derivatives. Comparison of natural
304 resveratrol (R-com), R-4’LA [42], R-4’DHA [43], R-5OiP (13), R-5OiP-4’LA (15a) and R-
305 5OiP-4’DHA (15b) activities. Results are expressed in mean ± SEM and are from n = 3-5
306 independent experiments. (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after
307 incubation of resveratrol derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as
308 the percentage of non-treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01,
309 *** p < 0.001, versus non-treated and exposed to atRAL cells. (B) Antioxidant assay:
310 representation of ROS production (DCFDA probe) after incubation of resveratrol derivatives (0-
311 80 µM) and H2O2 (600 µM) in ARPE-19 cells. The data are expressed as the percentage of non-
312 treated and exposed to H2O2 cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-treated and
313 exposed to H2O2 cells.
314
315
The anti-carbonyl potential of resveratrol has been described by Wang et al. who demonstrated
316 the formation of an adduct by co-incubation of resveratrol with acrolein at equimolar
317 concentration [18], and by Shen et al. with methyl glyoxal [45]. Resveratrol was able to scavenge
318 toxic aldehydes and formed an heterocyclic ring at the C-2 and C-3 positions through
319 nucleophilic addition. Resveratrol protection of acrolein-treated cells was also attributed to a

320 direct stimulatory action on mitochondrial bioenergetics [46]. Here, in our cellular assay, natural
321 resveratrol did not display sufficient anti-carbonyl stress activity to counteract toxic effects of
322 high doses of atRAL. However, increasing the nucleophilicity of resorcinol backbone with an
323 isopropyl group, and increasing lipophilicity of this polyphenol by addition of a PUFA,
324 prevented aldehyde toxicity.
325
Regarding the antioxidant profiles of resveratrol derivatives, shown in Figure 5B, 40 µM
326 commercial resveratrol (R-com) was able to reach the oxidative status of cells that did not
327 received the oxidant stressor H2O2. Logically, the introduction of a fatty acid moiety (LA or
328 DHA) and/or an alkyl residue masking the phenolic function, led to a loss of antioxidant capacity
329 compared to R-com. However, both the studied alkyl-lipophenols R-5OiP-4’LA (15a) and R-
330 5OiP-4’DHA (15b) were still potent antioxidant derivatives, as, at 80 µM, they decreased ROS
331 production by up to 47% and 37%, respectively. The reduction of the accessibility of two
332 phenolic functions on the resveratrol backbone still allowed 15a and 15b to be potent
333 antioxidants able to reduce oxidative status in a dose-dependent manner.
334
Thus, both resveratrol alkyl-lipophenols 15a and 15b exhibited an interesting protection
335 against the double COS in ARPE-19 cells.
336
337 3. Synthesis and anti-COS evaluation of flavonoid lipophenol analogues
338
339 worked on two natural flavonoid derivatives, quercetin and catechin (Figure 1).
340 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
341 quantities in many fruit and vegetables, including apple, cranberry, red onion, asparagus,
342 spinach, walnuts and coriander. Over the past decades, quercetin has gained research interest due

343 to its numerous pharmacological activities. In addition to its antioxidant properties, quercetin
344 was proven to have anti-inflammatory, antidiabetic, anticancer, cardiovascular, hepato-
345 protective, antiplatelet, antibacterial and neuroprotective properties [49]. Regarding its
346 antioxidant action, quercetin acts by scavenging free radicals, and also increases antioxidant
347 enzymes, such as glutathione peroxidase, superoxide dismutase and catalase [50]. (+)-Catechin
348 (C) is a flavanol abundant in berries, chocolate, cacao and green tea that makes a significant
349 contribution to total dietary antioxidant intake. Catechin is the monomer of proanthocyanidin B2,
350 whose protection against A2E photo-oxidation-induced apoptosis has been shown in ARPE-19
351 cells [35]. Compared to quercetin, catechin should have increased nucleophilic properties
352 coming from the A ring, due to the absence of the carbonyl function link to this aromatic cycle.
353
Distinctive chemical structures related with flavonoid antioxidant activities have been
354 established including hydroxyl groups of the A-ring (resorcinol), ortho-dihydroxy arrangement
355 in the B-ring (catechol), and in the case of quercetin, C2-C3 unsaturated bond combined with C-
356 4 carbonyl group in the C-ring [48]. Moreover, the position instead of total number of hydroxyl
357 groups, considerably influences the efficiency of antioxidant activity [51]. The B-ring hydroxyl
358 structure is the utmost significant actor of scavenging oxygen free radicals [52]. In addition,
359 according to the study of Hong et al. [53], radical scavenging properties of catechin seem to be
360 less affected by acylation at the 3 or 7 position compared to acylation of the catechol moiety.
361 Thereby, in order to conserve the best antioxidant properties, the B-ring, corresponding to the
362 catechol moiety, was left free of any substituent, and position 3 was selected to introduce the
363 PUFA moiety on both catechin and quercetin. Finally, to preserve the alkylated resorcinol group
364 needed for the anti-carbonyl stress activity in both phloroglucinol and resveratrol series, two
365 positions of the A-ring, that mimic the resorcinol moiety of the LEADs A/B, were selected to

366 introduce the isopropyl; both positions 5 and 7 were alkylated in order to evaluate the most
367 favorable for carbonyl stress protection. As for resveratrol derivatives, lipophilic catechin and
368 quercetin bearing only the PUFA moiety (C-3PUFA and Q-3PUFA) [42], and alkyl-flavonoids
369 bearing only the isopropyl moiety (C-5OiP 18b, C-7OiP 18a, Q-5OiP 23 and Q-7OiP 35), were
370 produced to evaluate their impact on biological properties.
371
372 3.1. Catechin lipophenols synthesis and evaluation
373
Synthesis of alkyl-catechin derivatives C-7OiP (18a) and C-5OiP (18b), as well as the alkyl-
374 lipophenol derivatives C-3LA-7OiP (19a) and C-3LA-5OiP (19b), are presented in Scheme 3.
375
376 Scheme 3. Synthesis and chemical structures of (+)-catechin derivativesa
Ph
Ph
Ph
Ph
OH
3'
O
O
O
O
OH
4'
B
7
HO
O
R1O
O
HO
O
C
a
b
A
OH
OH
3
OH
5
OH
OH
OR2
(+)-catechin
C-com
16
17a: R1 = iP, R2 = H
17b: R1 = H, R2 = iP
c
OH
OH
OH
OH
OH
OH
HO
O
R1O
O
R1O
O
d
OLA
OLA
OH
OH
OR2
OR2
18a: R1 = iP, R2 = H
18b: R1 = H, R2 = iP
19a: R1 = iP, R2 = H
19b: R1 = H, R2 = iP
C-3LA
377
378
aReagents and conditions: (a) Ph2CCl2, K2CO3, MeCN, rt, 23 h, 33%; (b) diisopropylsulfate,
379 K2CO3, acetone, 40 °C, 70 h, 37% separable mixture of 17a (19%) and 17b (18%); (c) H2,
380 Pd(OH)2, THF/EtOH, rt, 18a : 42 h, 56% and 18b : 64 h, 72%; (d) LACOCl, TFA, 1,4-dioxan, rt,
381 68 h, 1.5% for 19a and 19b.
382

383
The first step of the synthesis was the protection of the catechol moiety of catechin by a
384 diphenyldioxole in acetonitrile [54]. The alkylation step allowed access to both isomers 17a and
385 17b with 1/1 proportion in 37% yield. The two isomers were separated by column
386 chromatography and engaged separately for the end of the synthesis. Deprotection of the
387 catechol was performed by hydrogenation using palladium hydroxide and led to the desired
388 alkylated catechin derivatives: C-7OiP (compound 18a) and C-5OiP (compound 18b). The
389 lipophenols were then obtained using TFA and freshly prepared linoleyl chloride, as described
390 by Uesato et al. [55]. Compounds C-3LA-7OiP (19a) and C-3LA-5OiP (19b) were isolated in
391 sufficient quantities to be evaluated in vitro.
392
Commercial catechin (C-com), as well as the two alkylated catechins C-5OiP (18b) and C-
393 7OiP (18a), displayed no toxicity on ARPE-19 cells up to 160 µM (Figure 6). However, C-3LA
394 [42] and C-3LA-5OiP (19b) displayed an important cell death above 120 µM. Toxicity was even
395 higher for C-3LA-7OiP (19a) with no observed cell viability at 80 µM. In contrast with
396 resveratrol derivatives, alkylated catechins 18a and 18b displayed no toxicity, but association of
397 PUFA and alkyl on catechin backbone (19a and 19b), resulted in high toxicity in the ARPE-19
398 cell line.
399

***
***
***
**
120
100
80
60
40
20
0
10 µM
40 µM
80 µM
120 µM
160 µM
***
***
***
***
***
***
***
***
)
)
)
)
m
A
a
a
b
b
o
L
9
8
c
9
8
3
1
-
-
1
1
1
(
(
(
(
C
C
P
i
P
i
P
i
P
i
O
O
O
O
7
7
5
5
-
-
-
-
A
C
A
C
L
L
3
3
-
-
C
C
400
401 Figure 6. Evaluation of cytotoxicity of (+)-catechin derivatives; natural (+)-catechin (C-com), C-
402 3LA [42], C-5OiP (18b), C-7OiP (18a), C-3LA-5OiP (19b) and C-3LA-7OiP (19a). Results are
403 expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell viability
404 (MTT) after incubation of catechin derivatives (0-160 µM). The data are expressed as the
405 percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-
406 treated control cells.
407
408
Interestingly, despite the absence of cell protection using C-com, C-3LA was able to display a
409 mild protection against carbonyl stressor, with dose-dependent effects. The catechin ring, in
410 contrast to resveratrol or phloroglucinol ring, should be nucleophilic enough to show a protective
411 effect against carbonyl stressor, without the presence of isopropyl function on the resorcinol
412 moiety (responsible for inductive effect). The direct trapping of RCS, such as methylglyoxal (an
413 intermediate reactive carbonyl of AGE formation) by natural catechin was described in cell-free
414 experiments by Peng et al. [56]. Wang confirmed this observation in 2010, by identifying
415 adducts produced by co-incubation of catechin with glyoxal, methylglyoxal and acrolein [32].
416 Zhu et al. also demonstrated trapping of lipid-derived α,β-unsaturated aldehydes, which have
417 been implicated as causative agents in the development of carbonyl stress-associated pathologies
418 (i.e. 4-HNE and acrolein), by several (poly)phenols including catechin [31]. However, most of

419 those adducts have been observed only under simulated physiological conditions but rarely in
420 actual cellular media. As expected, in our work, the addition of isopropyl, preferentially at the 7
421 position (C-3LA-7OiP (19a), Figure 7A), increased cellular protection when tested at 40 µM
422 (below the toxic concentration of 19a). The position of the isopropyl group seemed to influence
423 anti-carbonyl stress activity, as well as the cytotoxic profile of the derivatives. Alkylation at the 7
424 position favor aldehyde trapping in both C8 and C6 positions, whereas alkylation at the 5
425 position orients the formation of adduct only with C6. Both sites were reported in the literature to
426 react with aldehyde, however, some work performed on the activity of epigallocatechin-3-O-
427 gallate (EGCG) led only to the identification of the C8 adduct [57]. Unexpectedly, the most
428 active lipophenol against carbonyl stress in this series, C-3LA-7OiP (19a), was also the most
429 toxic compound.
430
(B)
(A)
100
***
40
###
80
60
40
20
0
###
30
20
10
0
***
###
###
###
***
***
###
###
***
**
)
)
)
)
)
)
)
)
m
a
A
a
m
b
A
b
a
a
b
b
o
L
9
o
8
9
L
8
9
8
8
9
c
-
c
3
-
1
(
3
1
(
1
1
1
1
1
(
1
-
(
-
(
(
(
(
C
C
C
C
P
i
P
P
i
P
P
i
P
i
P
C
P
i
i
i
i
O
O
O
O
O
O
O
O
5
-
7
7
7
5
7
5
-
5
-
-
-
-
-
-
A
C
A
A
A
C
C
C
L
L
L
L
3
3
3
3
-
-
-
-
C
C
C
80 µM
10 µM
40 µM
10 µM
40 µM
80 µM
431 Figure 7. In vitro anti-COS evaluation of (+)-catechin derivatives. Comparison of natural (+)-
432 catechin (C-com), C-3LA [42], C-5OiP (18b), C-7OiP (18a), C-3LA-5OiP (19b) and C-3LA-
433 7OiP (19a) activities. Results are expressed in mean ± SEM and are from n = 3-5 independent

434 experiments. (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after incubation of
435 catechin derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as the percentage of
436 non-treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p < 0.001,
437 versus non-treated and exposed to atRAL cells. (B) Antioxidant assay: representation of ROS
438 production (DCFDA probe) after incubation of catechin derivatives (0-80 µM) and H2O2 (600
439 µM) in ARPE-19 cells. The data are expressed as the percentage of non-treated and exposed to
440 H2O2 cells. All conditions have a p-value < 0.001 versus non-treated and exposed to H2O2
441 control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus non-treated and non-exposed to H2O2
442 cells.
443
444
As expected, C-com, and the two alkylated derivatives C-5OiP (18b) and C-7OiP (18a),
445 displayed the best antioxidant activity with equivalent ROS production levels compared to
446 control cells non-exposed to H2O2, as shown in Figure 7B. With two substituents reducing free
447 phenolic functions, C-3LA-5OiP (19b) and C-3LA-7OiP (19a) displayed dose-dependent effects
448 on ROS produced by H2O2 treatment, with a decrease by up to 60% and 67% in ROS production,
449 respectively, at 80 µM.
450
Despite these interesting antioxidant properties, C-3LA-5OiP (19b) lacked potent anti-
451 carbonyl stress activity (only 10% increase in viability observed at 80 µM) and C-3LA-7OiP
452 (19a) showed high toxicity (no survival from 80 µM), making these derivatives less interesting
453 anti-COS candidates compared to the resveratrol series.
454
455 3.2. Quercetin lipophenols synthesis and evaluation
456
Lipophilic quercetins Q-3LA and Q-3DHA were produced as previously described [42] and
457 two original pathways were developed to access the lipophenol derivatives of quercetin-5OiP
458 (29) and quercetin-7OiP (39a/b) (Scheme 4 and 5).
459
460 3.2.1. Quercetin-5OiP derivatives

461
A first chemical strategy was developed in order to access the lipophenol derivative Q-3LA-
462 5OiP (29) and the PUFA-free analogue Q-5OiP (23) (Scheme 4).
463
464 Scheme 4. Synthesis and chemical structures of quercetin-5OiP derivativesa
OH
OH
HO
O
OH
O
O
23
d
OBn
OH
3'
OBn
OBn
OH
4'
OBn
B
7
BnO
O
HO
O
C
BnO
O
O
a, b
c
A
OH
O
O
3
5
O
O
OH
O
O
O
O
O
O
OH
OH
22
O
O
OH
OH
OH OH
OH OH
OH
OH
OH
OH
e
rutin
21
OBn
OTIPS
OTIPS
OTIPS
OTIPS
OBn
BnO
O
h
f, g
TIPSO
O
TIPSO
O
OAc
OH
OAc
O
O
O
O
O
O
24
26
27
i
OH
OH
OTIPS
OTIPS
OH
OH
HO
O
HO
O
TIPSO
O
j
OLA
OLA
OLA
O
O
OH
O
O
O
29
Q-3LA
28
465
466
aReagents and conditions: (a) BnBr, K2CO3, DMF, 40 °C, 4 h, 61%; (b) 2-bromopropane,
467 K2CO3, DMF, 80 °C, 18 h, 88%; (c) HCl, EtOH, 70 °C, 5 h, 98%; (d) H2, Pd/C, THF/EtOH, rt,
468 21 h, 69%; (e) pyridine, Ac2O, rt, 20 h, 100%; (f) H2, Pd/C, THF/EtOH, rt, 15 h, 100%; (g)
469 TIPS-OTf, NEt3, THF, rt, 4 h, 62%; (h) NH3/MeOH, DCM, 0 °C, 3 h, 88%; (i) LA, DCC,
470 DMAP, DCM, rt, 23 h, 54%; (j) NEt3/3HF, THF, rt, 15 min, 72%.
471

472
In order to access quercetin-5OiP derivatives, we started from commercially available rutin
473 using the diholoside rutinoside as protecting group for the phenolic function at the 3 position.
474 The first step was the protection of the phenols in positions 7, 3’ and 4’ with benzyl groups.
475 Alkylation of the phenol in position 5 was performed with 2-bromopropane by heating at 80 °C
476 in dry DMF (88%). Cleavage of rutinoside in acidic conditions gave compound 22, which was
477 either de-benzylated to access desired Q-5OiP (23), or acetylated in position 3 to pursue the
478 lipophenol synthesis. Benzyl groups were removed by hydrogenation to be replaced by silyl
479 protecting groups, which can be easily deprotected without damaging PUFA moiety leading to
480 compound 26. Acetate in position 3 was then cleaved using ammonia solution in MeOH at 0 °C
481 to allow esterification with fatty acid. Final deprotection of TIPS group gave the desired Q-3LA-
482 5OiP (29).
483
For comparison purposes, commercial quercetin (Q-com) and Q-3LA [42] properties were also
484 evaluated in vitro. Under the conditions tested, Q-com was found to be toxic for the ARPE-19
485 cell line from 80 µM (Figure 8). Introduction of the isopropyl moiety (Q-5OiP, 23) reduced the
486 toxicity with mild mortality starting at 160 µM. A similar reduction was observed with PUFA
487 introduction, as Q-3LA displayed no toxicity up to 160 µM, and was found to increase cell
488 viability (156% cell viability at 80 µM). However, the alkyl-lipophenol derivative Q-3LA-5OiP
489 (29) was very toxic, with no survival at 80 µM.
490

***
***
160
140
120
100
80
***
***
***
***
**
**
*
**
*
*
10 µM
40 µM
80 µM
120 µM
160 µM
***
***
***
***
60
40
20
***
***
***
0
)
)
3
m
A
9
o
L
2
2
c
(
3
(
-
-
P
i
P
i
Q
Q
O
O
5
5
-
-
A
Q
L
3
-
Q
491
492 Figure 8. Evaluation of cytotoxicity of quercetin-5OiP derivatives; natural quercetin (Q-com),
493 Q-3LA [42], Q-5OiP (23) and Q-3LA-5OiP (29). Results are expressed in mean ± SEM and are
494 from at n = 3-5 independent experiments. ARPE-19 cell viability (MTT) after incubation of
495 quercetin-5OiP derivatives (0-160 µM). The data are expressed as the percentage of non-treated
496 control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-treated control cells.
497
498
Regarding anti-carbonyl stress properties, here again the presence of isopropyl, as well as the
499 PUFA moiety, seemed decisive for cell protection. As shown in Figure 9A, Q-com, Q-3LA and
500 Q-5OiP (23) did not display any protective effect against atRAL toxicity, whereas Q-3LA-5OiP
501 (29) protected cells with 50% increase in cell survival at 80 µM. Even if toxic at 80 µM, a high
502 anti-carbonyl stress activity was observed at this concentration and can be explained by the
503 differences between cytotoxicity and anti-carbonyl stress activity protocols (different incubation
504 times).
505

(B)
(A)
***
***
80
60
40
20
0
100
80
60
40
20
0
###
###
###
###
###
**
### ###
###
###
)
)
)
)
A
3
9
m
A
9
m
3
L
2
2
L
2
o
o
2
(
(
(
3
(
c
3
c
-
-
-
-
P
P
P
i
i
P
Q
i
Q
Q
i
Q
O
O
O
O
5
5
5
5
-
-
-
-
A
A
Q
Q
L
L
3
3
-
-
Q
Q
10 µM
40 µM
80 µM
10 µM
40 µM
80 µM
506 Figure 9. In vitro anti-COS evaluation of quercetin-5OiP derivatives. Comparison of natural
507 quercetin (Q-com), Q-3LA [42], Q-5OiP (23) and Q-3LA-5OiP (29) activities. Results are
508 expressed as mean ± SEM and are from n = 3-5 independent experiments. (A) Anti-carbonyl
509 stress assay: ARPE-19 cell viability (MTT) after incubation of quercetin-5OiP derivatives (0-80
510 µM) and atRAL (15 µM). The data are expressed as the percentage of non-treated and non-
511 exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p < 0.001, versus non-treated and
512 exposed to atRAL cells. (B) Antioxidant assay: representation of ROS production (DCFDA
513 probe) after incubation of quercetin-5OiP derivatives (0-80 µM) and H2O2 (600 µM) in ARPE-
514 19 cells. The data are expressed as the percentage of non-treated and exposed to H2O2 cells. All
515 conditions have a p-value < 0.001 versus non-treated and exposed to H2O2 control cells. # p <
516 0.05, ## p < 0.01, ### p < 0.001, versus non-treated and non-exposed to H2O2 cells.
517
518
Finally, antioxidant capacity was evaluated and is reported in Figure 9B. The Di-substituted
519 lipophenol derivative Q-3LA-5OiP (29) was efficient to reduce H2O2-induced ROS production
520 by 53% at 80 µM. Here again, a dose-dependent antioxidant potency was still observable despite
521 the reduction of two free phenolic functions of quercetin. A similar protection profile was
522 observed for the isopropyl-free analogue Q-3LA, which suggests that position 5 is not primordial
523 for ROS scavenging activity. This was also confirmed by comparing Q-com and Q-5OiP (23)
524 cell treatment, which led to low ROS levels.

525
526
To conclude, in vitro assessments indicate that in this series, alkylated resorcinol and PUFA
527 moieties are necessary for anti-carbonyl stress activity, as observed for the phloroglucinol [23]
528 and resveratrol series. Derivative Q-3LA-5OiP (29) displays interesting anti-COS properties,
529 however, its high cytotoxicity on ARPE-19 has to be considered for further evaluations.
530
531 3.2.2. Quercetin-7OiP derivatives
532
A second chemical strategy, presented in Scheme 5, was developed for the synthesis of alkyl
533 quercetin Q-7OiP (35) and the two lipophenol derivatives Q-3LA-7OiP (39a) and Q-3DHA-
534 7OiP (39b).
535
536 Scheme 5. Synthesis and chemical structures of quercetin-7OiP derivativesa

Ph
Ph
Ph
Ph
Ph
Ph
OH
3'
O
O
O
O
O
O
OH
4'
B
7
a
b, c
d
HO
O
C
HO
O
HO
O
O
O
A
OH
OH
OAc
OAc
3
5
OH
O
OH
O
OAc O
OAc O
quercetin
Q-com
30
32
33
e
OTIPS
OTIPS
OTIPS
OTIPS
OH
OTIPS
OTIPS
OH
O
O
O
O
O
O
O
O
i
h
g
OH
OAc
OAc
OPUFA
OH
O
OAc O
OAc O
OH
O
38a: PUFA = LA
38b: PUFA = DHA
37
36
34
j
f
OH
OH
OH
OH
OH
OH
O
O
HO
O
O
O
O
OPUFA
ODHA
OH
OH
O
OH
OH
O
39a: PUFA = LA
39b: PUFA = DHA
Q-3DHA
35
537
538
aReagents and conditions: (a) Ph2CCl2, Ph2O, 175 °C, 2 h, 89%; (b) pyridine, Ac2O, rt, 20 h,
539 100%; (c) PhSH, imidazole, NMP, 0 to 5 °C, 5 h, 100%; (d) diisopropylsulfate, K2CO3, acetone,
540 rt, 22 h, 91%; (e) H2, Pd(OH)2, THF/EtOH, rt, 20 h, 47%; (f) NH3/MeOH, 0 °C, 35 min, 46%;
541 (g) TIPS-OTf, NEt3, THF, rt, 10 min, 71%; (h) NH3/MeOH, DCM, 0 °C, 1 h, 100%; (i) PUFA,
542 DCC, DMAP, DCM, rt, 5 h, 38a: 86% and 38b: 85%; (j) NEt3/3HF, THF, rt, 30 min, 92% for
543 39a and 39b.
544
545
For the synthesis of Q-7OiP derivatives, the catechol of the commercial quercetin was first
546 protected by diphenyldioxole, followed by the protection of the remaining phenolic functions
547 with acetate moieties. Selective deprotection of phenol in position 7 was then performed
548 according to Li et al., using thiophenol and imidazole in N-methylmorpholine in quantitative
549 yield [58]. Compound 32 was then alkylated with diisopropylsulfate leading to derivative 33 in
550 91% yield. Hydrogenation of diphenyldioxole with palladium hydroxide led to compound 34,
551 which can be either deacetylated to access desired Q-7OiP (35) or protected with silyl ethers to
552 access protected derivative 36. Deacetylation of 36 with diluted ammonia solution in MeOH at

553 0°C allowed access to compound 37, which could be selectively esterified in position 3 by PUFA
554 (86% for LA and 85% for DHA derivatives). Resulting compounds underwent a final TIPS
555 deprotection leading to the desired lipophenol derivatives Q-3LA-7OiP (39a) and Q-3DHA-7OiP
556 (39b).
557
Toxicity and activity of quercetin-7OiP lipophenols Q-3LA-7OiP (39a) and Q-3DHA-7OiP
558 (39b), were compared to Q-com, Q-3LA [42], Q-3DHA [42] and Q-7OiP (35). Toxicity profile
559 (Figure 10) shows that Q-7OiP (35) was highly toxic even at low concentrations with no
560 survival observed at 40 µM. The position of the alkyl moiety appeared critical, as Q-5OiP (23,
561 Figure 8) did not display such high toxicity. Reduction of Q-com toxicity was observed by
562 introduction of the PUFA moieties, as Q-3DHA and Q-3LA increased ARPE-19 cells viability
563 from 10 to 80 µM until toxicity appeared. Alkyl-lipophenol derivatives Q-3LA-7OiP (39a) and
564 Q-3DHA-7OiP (39b) did not display any toxicity up to 160 µM, and Q-3LA-7OiP (39a) even
565 increased cell survival at 160 µM (156% cell viability).
566
***
***
***
***
***
160
140
120
100
80
60
40
20
***
***
***
***
***
***
***
10 µM
40 µM
80 µM
120 µM
160 µM
**
**
***
***
***
***
***
***
***
***
***
0
)
)
)
5
m
A
A
H
a
b
o
L
9
9
3
c
3
(
-
3
3
(
D
-
(
3
Q
P
i
Q
-
P
i
P
i
Q
O
O
7
O
-
7
7
-
-
Q
A
A
L
H
3
D
-
3
Q
-
Q
567
568 Figure 10. Evaluation of cytotoxicity of quercetin-7OiP derivatives; natural quercetin (Q-com),
569 Q-3LA [42], Q-3DHA [42], Q-7OiP (35), Q-3LA-7OiP (39a) and Q-3DHA-7OiP (39b). Results
570 are expressed in mean ± SEM and are from n = 3-5 independent experiments. ARPE-19 cell

571 viability (MTT) after incubation of quercetin-7OiP derivatives (0-160 µM). The data are
572 expressed as the percentage of non-treated control cells. * p < 0.05, ** p < 0.01, *** p < 0.001,
573 versus non-treated control cells.
574
575
As also observed for the resveratrol and catechin series, in our cellular assay the natural
576 quercetin alone did not protect against carbonyl stress toxicity (Figure 11A), whereas it has been
577 reported the formation of adduct by co-incubation of quercetin with several reactive aldehydes
578 (i.e. glyoxal, methylglyoxal and acrolein) in cell-free assays [32]. As expected regarding toxicity
579 profile, the PUFA-free quercetin Q-7OiP (35) is not protective against carbonyl stress and
580 presents high toxicity. Globally, the addition of the isopropyl on this lipophenol derivatives led
581 to an increase of cell protection against RCS, however, the impact of the isopropyl seemed less
582 important than in the resveratrol, phloroglucinol or quercetin-5OiP lipophenol series.
583 Comparison between the effect of Q-7OiP (35) and the alkyl-lipophenols Q-3LA-7OiP (39a) and
584 Q-3DHA-7OiP (39b) (increased viability by 38% and 24%, respectively, at 80µM) confirmed
585 the importance of the PUFA part to confer high cellular protection against atRAL toxicity using
586 alkyl-(poly)phenol (already observed in other series). Lipid peroxidation of PUFA produces
587 reactive aldehydes (4-HNE or 4-hydroxyhexenal (4-HHE)) [59]. Exposure to excessive 4-HNE
588 or 4-HHE can cause cytotoxicity and is implied in the detrimental pathogenesis of a number of
589 degenerative diseases [60]. However, such lipid peroxidation metabolites are also signaling
590 molecules able to induce gene expression of antioxidant and detoxifying aldehyde enzymes, by
591 activation of the Nrf2 pathway [61]. In ARPE-19, Johansson et al. reported the ability of DHA
592 treatment to induce cellular antioxidant responses, by Nrf2 pathway activation, and to stimulate
593 autophagy [62]. In addition, some works reported also the anti-glycation properties of PUFAs
594 [63,64]. As glycation reaction is caused by reactive aldehydes, anti-glycation properties can be
595 related to anti-carbonyl stress activity. Additional studies reported the cytoprotective effect of

596 PUFAs, by increasing S-phase cell promotion or lipid metabolism [65]. In the retina for example,
597 DHA was shown to protect photoreceptors from oxidative stress by preserving mitochondrial
598 membrane integrity [66]. In most of the (poly)phenol series studied in this work, the introduction
599 of a PUFA part and an alkyl moiety on the (poly)phenol was a prerequisite to provide important
600 cellular protection against atRAL toxicity. This may be due to improved cell penetration due to
601 an increased lipophilicity, and/or a synergic effect of the alkyl-(poly)phenol and the PUFA
602 moiety.
603
(B)
(A)
***
***
###
100
60
***
###
###
###
###
###
80
60
40
20
0
***
***
***
###
###
***
###
40
20
0
###
###
###
#
#
***
***
)
)
)
)
)
)
A
A
-
m
A
L
A
m
b
5
a
a
5
b
L
o
H
o
H
9
3
9
9
3
9
3
c
3
-
(
(
D
c
-
3
-
3
D
3
3
(
-
(
(
(
3
3
Q
P
i
-
P
Q
Q
-
Q
i
P
P
i
P
P
i
i
Q
i
Q
O
O
O
7
7
O
7
O
O
-
7
7
7
-
-
-
-
Q
Q
A
A
A
A
L
H
L
H
3
3
D
-
-
D
3
3
Q
-
Q
-
Q
Q
10 µM
40 µM
80 µM
10 µM
40 µM
80 µM
604 Figure 11. In vitro anti-COS evaluation of quercetin-7OiP derivatives. Comparison of natural
605 quercetin (Q-com), Q-3LA [42], Q-3DHA [42], Q-7OiP (35), Q-3LA-7OiP (39a) and Q-3DHA-
606 7OiP (39b) activities. Results are expressed as mean ± SEM and are from n = 3-5 independent
607 experiments. (A) Anti-carbonyl stress assay: ARPE-19 cell viability (MTT) after incubation of
608 quercetin-7OiP derivatives (0-80 µM) and atRAL (15 µM). The data are expressed as the
609 percentage of non-treated and non-exposed to atRAL control cells. * p < 0.05, ** p < 0.01, *** p
610 < 0.001, versus non-treated and exposed to atRAL cells. (B) Antioxidant assay: representation of
611 ROS production (DCFDA probe) after incubation of quercetin-7OiP derivatives (0-80 µM) and
612 H2O2 (600 µM) in ARPE-19 cells. The data are expressed as the percentage of non-treated and
613 exposed to H2O2 cells. All conditions have a p-value < 0.001 versus non-treated and exposed to

614 H2O2 control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus non-treated and non-exposed to
615 H2O2 cells.
616
617
Antioxidant capacities of the derivatives are reported in Figure 11B. As expected, lower
618 antioxidant profile of the two lipophenols Q-3LA-7OiP (39a) and Q-3DHA-7OiP (39b),
619 compared to natural quercetin, can be explained by introduction of two substituents on phenolic
620 positions. However, Q-3LA-7OiP (39a) and Q-3DHA-7OiP (39b) displayed interesting dose-
621 dependent reduction of ROS close to the alkyl-free lipophenols Q-3DHA and Q-3LA. They
622 decreased by up to 48% and 38% of H2O2 induced ROS production, respectively, when
623 introduced at 80 µM.
624
The two di-substituted candidates Q-3LA-7OiP (39a) and Q-3DHA-7OiP (39b) are potent
625 anti-COS derivatives for further evaluations, as they displayed interesting dose-dependent
626 protection on ARPE-19 cell line and induced no toxicity up to 160 µM.
627
628 4. Selection of best anti-COS candidates in ARPE-19 cells
629
After systematic evaluation of our lipophenol derivatives, only the most promising candidates
630 were selected for further in vitro evaluations. Catechin and quercetin-5OiP lipophenols have
631 been rejected as they display important cytotoxicity in the ARPE-19 cell line. Regarding the
632 close protection provided by either LA or DHA lipophenols, omega-3 DHA lipophenols were
633 selected in view of in vivo evaluation. Indeed, many beneficial effects of DHA have been
634 reported in relation to retinal affections [67], such as anti-inflammatory, neuroprotective effects
635 [28,68] and antioxidant properties [66]. Visual processing deficits have been improved with
636 DHA supplementation in some clinical studies [25,67]. Moreover, as predominant PUFA of the
637 photoreceptor membrane, DHA is transported by several specific pathways across the Blood

638 Retina Barrier (BRB) [69]. These properties of DHA are favorable, as increasing the lipophilicity
639 of (poly)phenols using this PUFA, can help the derivatives through retinal cell membranes and
640 across the RPE barrier. In order to select the best candidate for pharmacological development,
641 we compared anti-COS profiles of the most promising lipophenols. According to SAR study of
642 anti-COS assays, selected derivatives bear an isopropyl moiety, as well as DHA lipophilic
643 function.
644
645 4.1. Anti-COS comparative assays
646
All anti-COS activities of promising alkyl-DHA lipophenols have been summarized at the
647 same concentration (80 µM) that displayed the best protective effects, and are represented in
648 Figure 12. The range of anti-carbonyl stress activity is similar for three of the four derivatives:
649 R-5OiP-4’DHA (15b) is the only compound that displayed a significantly higher protective
650 effect against atRAL toxicity compared to P-OiP-ODHA (LEAD B), P-OiP-CDHA (8b) or Q-
651 3DHA-7OiP (39b), as shown in Figure 12A.
652

(B)
(A)
##
##
80
#
100
°°°
ó
60
40
20
0
°°°
80
60
40
20
0
ó
)
)
)
)
)
)
)
)
b
b
b
b
b
B
b
B
8
8
9
5
9
5
(
(
3
D
1
3
(
D
-
1
(
(
(
A
-
A
A
A
P
P
H
A
P
i
E
A
E
H
i
L
L
H
D
D
H
(
(
O
O
D
'
C
D
C
7
7
'
-
-
-
A
A
4
4
P
-
P
A
A
H
H
i
i
P
i
H
H
D
D
i
O
O
-
-
D
-
D
O
O
-
O
O
-
P
P
3
5
5
3
-
-
-
-
P
P
i
i
Q
R
R
Q
O
O
-
-
P
P
653 Figure 12. Comparison of anti-COS activities of selected DHA alkyl-lipophenols: P-OiP-ODHA
654 (LEAD B) [23], P-OiP-CDHA (8b), R-5OiP-4’DHA (15b) and Q-3DHA-7OiP (39b). Results
655 are expressed as mean ± SEM and are from n = 3-5 independent experiments. (A) Anti-carbonyl
656 stress assay: ARPE-19 cell viability (MTT) after incubation of lipophenol derivatives (80 µM)
657 and atRAL (15 µM). The data are expressed as the percentage of non-treated and non-exposed to
658 atRAL control cells. All conditions have a p-value < 0.001 versus non-treated and exposed to
659 atRAL control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus R-5OiP-4’DHA (15b). (B)
660 Antioxidant assay: representation of ROS production (DCFDA probe) after incubation of
661 lipophenol derivatives (80 µM) and H2O2 (600 µM) in ARPE-19 cells. The data are expressed as
662 the percentage of non-treated and exposed to H2O2 cells. All conditions have a p-value < 0.001
663 versus non-treated and exposed to H2O2 control cells and versus non-treated and non-exposed to
664 H2O2 control cells. ° p < 0.05, °° p < 0.01, °°° p < 0.001, versus P-OiP-ODHA (LEAD B). ó p <
665 0.05, óó p < 0.01, óóó p < 0.001, versus R-5OiP-4’DHA (15b).
666
667
The main objective of the (poly)phenol backbone modification was to overcome the low
668 antioxidant activity of P-OiP-ODHA (LEAD B). This goal was achieved, as two of the three new
669 lipophenol derivatives displayed a significant better efficiency against H2O2-induced ROS
670 production when compared to the LEAD B: 16 and 17% reduction in ROS production for P-OiP-
671 CDHA (8b) and Q-3DHA-7OiP (39b), respectively, as shown in Figure 12B. No significant
672 difference in antioxidant evaluation was observed between P-OiP-ODHA (LEAD B) and R-

673 5OiP-4’DHA (15b). Lower efficacy of P-OiP-ODHA (LEAD B) and R-5OiP-4’DHA (15b)
674 could be explained by comparing their number of free phenolic positions as those lipophenols
675 only have one free phenolic function to scavenge ROS. Papuc et al. detailed structure-activity
676 relationship of (poly)phenol antioxidant activity and highlighted that not only the number of free
677 phenols is important for radical scavenging activity but also their position [70]. Indeed,
678 flavonoids with a catechol moiety (B-ring) are the most effective radical scavengers due to the
679 degree of stability conferred by the catechol structure participating in electron delocalization and
680 in the chelation of metals involved in ROS generation.
681
682 4.2. Protection against photo-induced A2E toxicity
683
Lipofuscin, a fluorescent lysosomal pigment composed of several lipophilic molecules (bis-
684 retinoids), is associated with age-related pathophysiological processes in the RPE. The best-
685 studied bis-retinoid and the first component of lipofuscin to be identified is A2E. Accumulation
686 and photo-oxidation of the di-retinal conjugate A2E in the RPE through COS mechanisms are
687 known to be one of the critical causes of AMD [5–7,71]. Therefore, the reduction of lipid
688 oxidation is a promising approach to prevent the progression of AMD [72]. RPE cell death was
689 observed by photo-oxidation of A2E due to generation of singlet oxygen and superoxide radicals
690 [73]. A2E can be degraded in epoxide and aldehyde derivatives leading also to carbonyl stress
691 [7,74]. This is why we did not only evaluate the protective effects against atRAL toxicity and
692 ROS scavenging properties herein. For an in-depth lipophenol comparison, a more specific
693 cellular assay is the evaluation of the protective effects of DHA alkyl-lipophenols against photo-
694 oxidized A2E toxicity, which more closely resembles AMD cytotoxicity.

695
Evaluation of survival of the RPE cell line ARPE-19 after incubation of lipophenols with A2E
696 and photo-oxidation by intensive blue light was performed. As represented in Figure 13, the best
697 protective effect was obtained for Q-3DHA-7OiP (39b) which increased cell viability by 50% at
698 80 µM, whereas P-OiP-ODHA (LEAD B) and P-OiP-CDHA (8b) only improved survival by
699 16% and 27%, respectively. R-5OiP-4’DHA (15b) was also an interesting candidate, as it
700 increased cell viability by 43% at 80 µM.
701
80
°
###
60
40
20
0
Compound
A2E EC50 (µM)
144.60 ± 28.1
63.99 ± 9.73
44.87 ± 6.07
15.38 ± 2.53
##
P-OiP-ODHA (LEAD B)
P-OiP-CDHA (8b)
R-5OiP-4’DHA (15b)
Q-3DHA-7OiP (39b)
)
)
)
)
b
b
B
b
8
5
9
(
D
P
1
3
(
(
A
i
A
E
A
D
P
i
H
L
H
D
(
O
D
C
'
7
-
A
-
4
-
H
A
P
D
H
i
O
-
O
O
-
P
3
5
-
-
P
i
Q
R
O
-
P
702 Figure 13. DHA-lipophenols protection against photo-oxidized A2E toxicity. Comparison of P-
703 OiP-ODHA (LEAD B) [23], P-OiP-CDHA (8b), R-5OiP-4’DHA (15b) and Q-3DHA-7OiP
704 (39b) protection. Results are expressed as mean ± SEM and are from n = 3-5 independent
705 experiments. ARPE-19 cell viability (MTT) after incubation of lipophenol derivatives (80 µM)
706 and toxic concentration of photo-oxidized A2E (20 µM, blue light exposure 30 min). The data
707 are expressed as the percentage of non-treated and non-exposed to A2E control cells. EC50 are
708 calculated for all lipophenols using dose reponses curves (0-80 µM) and GraphPad prism
709 software. All conditions have a p-value < 0.001 versus non-treated and exposed to toxic photo-
710 oxidized A2E control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus P-OiP-CDHA (8b). ° p
711 < 0.05, °° p < 0.01, °°° p < 0.001, versus R-5OiP-4’DHA (15b).
712

713
All DHA-lipophenols are able to protect cells against photo-oxidized A2E toxicity. In order to
714 better compare the four selected alkyl-lipophenols, dose-dependent responses from 10 µM to 80
715 µM were performed to calculate Efficiency Concentration 50 (EC50), the concentration of
716 lipophenol needed to protect 50% of the cells from photo-oxidized A2E toxicity. P-OiP-ODHA
717 (LEAD B) was less potent than the other derivatives with an EC50 of 145 µM. This can be
718 explained by its lower ROS scavenging ability. P-OiP-CDHA (8b) was significantly more
719 protective than P-OiP-ODHA (LEAD B) with an EC50 of 64 µM. This result was in accordance
720 with the anti-COS study, as both lipophenols displayed the same anti-carbonyl stress activity but
721 P-OiP-CDHA (8b) was more efficient at scavenging ROS than its lead P-OiP-ODHA (LEAD B).
722 Surprisingly, R-5OiP-4’DHA (15b) was more potent than P-OiP-CDHA (8b), protecting up to
723 59% of cells exposed to photo-oxidized A2E at 80 µM (EC50 = 45 µM), although it was less
724 efficient at scavenging ROS generated by H2O2 in the first antioxidant evaluation. However,
725 anti-carbonyl stress activity (higher for R-5OiP-4’DHA (15b) than P-OiP-CDHA (8b)) likely
726 participated in the protection of cells against photo-oxidized A2E. The most efficient lipophenol
727 against A2E toxicity is Q-3DHA-7OiP (39b) with an EC50 of 15 µM only and a maximal
728 viability of 67% of cells at 80 µM. Indeed, this lipophenol, which displayed an equivalent anti-
729 carbonyl stress activity as the others, was the most efficient in scavenging ROS and was
730 therefore the most protective in the assays. Comparison of natural (poly)phenols (i.e.
731 phloroglucinol, resveratrol and quercetin) in the A2E assay showed that quercetin was itself
732 more protective than phloroglucinol and resveratrol against photo-induced toxicity of A2E (data
733 not shown).

734
Among the three in vitro cell assays performed in this study and the toxicity profile, Q-3DHA-
735 7OiP (39b) has proven to be the most promising anti-COS lipophenol and R-5OiP-4’DHA (15b)
736 an interesting fallback solution.
737
738 5. Validation of Q-3DHA-7OiP anti-COS properties in primary rat RPE cells
739
Currently, the ARPE-19 cell line and primary cultures are both sources of RPE cells for in
740 vitro models used in fundamental and applied research, including the development of new
741 approaches for ophthalmological disorders [75]. The advantage of cell lines is that they maintain
742 their characteristics over a number of passages and have longer survival times, compared to
743 primary cultures. Moreover, ARPE-19 can be plated at constant cell density throughout the
744 study, whereas primary cultures exhibit more cell density variability [24]. This cell density
745 depends on the cytotoxicity of atRAL and H2O2 [24,76]. Nevertheless, even if ARPE-19 cells are
746 a valuable model for human RPE cells, it seems important to validate the anti-COS effect of the
747 compounds on primary cells to be closer to the physiological protection in vivo [11]. However,
748 the production of primary RPE cells is tedious and time consuming work because it depends on
749 the number of eyeballs available [26,77]. Therefore, only the most promising lipophenol
750 derivative Q-3DHA-7OiP (39b) was evaluated in primary rat RPE cells for its protective action
751 against atRAL toxicity and ROS production. A dose-dependent increase in cell viability was
752 observed for cultures incubated with atRAL and Q-3DHA-7OiP (39b) and is represented in
753 Figure 14A. Treatment of primary RPE cells with atRAL (25 µM) caused a significant decrease
754 in cell viability (26% of cell viability for untreated control cells), whereas treatment with atRAL
755 and Q-3DHA-7OiP (40, 80 and 160 µM) significantly improved cell viability by 15%, 22% and
756 39%, respectively). Interestingly, the stress condition (25 µM of atRAL), as well as treatment

757 with up to 80 µM Q-3DHA-7OiP, showed very similar effects in primary RPE and ARPE-19
758 cells, suggesting that both stressor and protector use a common mechanism of action. By
759 contrast, exposure of primary RPE cells to H2O2 at 600 µM for 4 h increased intracellular ROS
760 levels by 12-fold compared to the untreated cells (Ctrl - H2O2). Moreover, cells treated with Q-
761 3DHA-7OiP (40, 80 and 160 µM) markedly reduced ROS by 81%, 91% and 92%, respectively,
762 in comparison to the non-treated cells, exposed to H2O2 (Figure 14B).
763
(A)
(B)
###
80
100
80
60
40
20
0
###
60
40
20
0
#
Q-3DHA-7OiP (µM)
Q-3DHA-7OiP (µM)
764 Figure 14. Anti-COS activity of Q-3DHA-7OiP (39b) on primary rat RPE cells. Results are
765 expressed as mean ± SEM and are from n = 3-5 independent experiments. (A) Anti-carbonyl
766 stress assay: cell viability (MTT) after incubation of Q-3DHA-7OiP (39b) (0-160 µM) and
767 atRAL (25 µM). The data are expressed as the percentage of non-treated and non-exposed to
768 atRAL control cells. All conditions have a p-value < 0.001 versus non-treated and exposed to
769 atRAL control cells. # p < 0.05, ## p < 0.01, ### p < 0.001, versus 80 µM of lipophenol. (B)
770 Antioxidant assay: representation of ROS production (DCFDA probe) after incubation of Q-
771 3DHA-7OiP (39b) (0-160 µM) and H2O2 (600 µM) in primary cells. The data are expressed as
772 the percentage of non-treated and exposed to H2O2 cells. All conditions have a p-value < 0.001
773 versus non-treated and exposed to H2O2 control cells. # p < 0.05, ## p < 0.01, ### p < 0.001,
774 versus non-treated and non-exposed to H2O2 cells.
775
The protective effects of Q-3DHA-7OiP (39b) on primary rat RPE cells fully validate the
776 results obtained on ARPE-19 cells. This is particularly true with respect to the carbonyl stress

777 generated by atRAL, as the primary RPE showed a high protection effectiveness of Q-3DHA-
778 7OiP. This demonstrates the robustness of the results obtained on the cell line and suggests that
779 these data can be more easily extrapolated to in vivo assays. Moreover, Q-3DHA-7OiP (39b) is
780 now a lead for future investigations. The antioxidant activity is itself remarkable because it is
781 much more effective in the primary RPE compared to ARPE-19. Such a difference has already
782 been reported in the past with polyphenols, such as quercetin, especially in the context of
783 oxidative stress caused by H2O2 [76]. The authors reported that many of the flavonoids were
784 more effective at protecting primary RPE compared with ARPE-19. On this basis, it can be
785 speculated that quercetin derivative 39b can protect retinal cell types through different
786 mechanisms, including direct scavenging of ROS, anti-apoptotic activity, and phase -2 induction
787 [78]. It was also shown that the most effective compounds are more hydrophobic than quercetin,
788 indicating that they should pass through cell membranes and accumulate intracellularly [79]. The
789 DHA combined with quercetin should be consistent with an increase in hydrophobicity and cell
790 bioavailability of the compound.
791
792
793
Conclusions
Among the research performed to develop pharmacological treatment for macular
794 degeneration, most of the molecules have been discarded in the past because of toxicity issues.
795 We propose here, a pharmacological approach based on a natural product ((poly)phenolic
796 compound linked to omega-3 derivatives). Both parts of the molecule have already proven their
797 beneficial effect in numerous studies as dietary complements: as (poly)phenols and omega-3
798 PUFAs are already present in our alimentation, the risk of toxicity compared to other synthetic

799 drugs may be considerably reduced, as well as the unwanted side effects. Naturally, in vivo
800 toxicity of lipophenols and their potential atRAL-adduct will have to be studied in future studies.
801
Starting from the phloroglucinol backbone, the present work evaluated the modification of the
802 (poly)phenol core to achieve protection against both carbonyl and oxidative stresses in RPE
803 cells. New synthetic routes were developed to access four original lipophenol series based on
804 phloroglucinol, resveratrol, catechin and quercetin (poly)phenols. Despite cell-free assays
805 proving aldehyde scavenging of natural (poly)phenols, our work highlights that, in biological
806 media, the protection against carbonyl stress (produced by atRAL toxicity) requires additional
807 chemical modifications of natural (poly)phenol to promote cellular protection. Both isopropyl-
808 resorcinol and PUFA moieties were essential for promising anti-carbonyl stress activity in RPE
809 cells. Substitution of two phenolic functions lower antioxidant capacity, compared to native
810 (poly)phenols, as reported in literature [22], however di-substituted alkyl-lipophenols still
811 provide sufficient antioxidant activities to reduce intracellular ROS. Despite significant literature
812 relating direct RCS trapping using (poly)phenols in cell-free assays, there is limited information
813 on this direct interaction in complex biological systems. This work highlights the importance of
814 cellular assays to validate anti-carbonyl stress potency of (poly)phenol conjugates: chemicals
815 able to scavenge aldehyde in cell-free assays are not necessarily active in biological media,
816 because of a lack of bioavailability or cell penetration, but also because of the importance of the
817 pH in RCS scavenging (increased in basic conditions) [80,81]. It is therefore relevant to consider
818 the pH dependency of the aldehyde trapping by (poly)phenol, as pH can vary from one cell
819 compartment to another. The exact mechanism of alkyl-lipophenol protection against atRAL
820 toxicity is currently under investigation and may be an association of 1) direct aldehyde
821 scavenging, chemically or enzymatically catalyzed [82]; 2) activation of aldehyde detoxification

822 enzymes, as the alkyl-phloroglucinol lipophenol LEAD B is able to activate Nrf2-Keap1
823 pathway [24] (also reported to activate aldehyde dehydrogenases and glutathione S-tranferases
824 gene expression involved in aldehyde detoxification), and 3) reduction of ROS levels and
825 oxidative stress induced by atRAL cell treatment [24,26,83].
826
Taken together, the different cellular assays performed on the ARPE-19 cell line led to the
827 discovery of an optimal DHA-quercetin anti-COS lipophenol (Q-3DHA-7OiP, 39b) showing
828 optimized antioxidant properties compared to the previous phloroglucinol lead (LEAD B), and
829 high protection against atRAL toxicity. Its anti-COS properties have also been validated in
830 primary RPE cells. To conclude, Q-3DHA-7OiP was the most powerful lipophenol to suppress
831 photo-oxidative toxicity initiated in RPE cells by A2E. Evaluation of photoreceptor protection
832 against acute light-induced degeneration in the Abca4-/- mouse model (involving both carbonyl
833 and oxidative stresses), will be reported in due course using the best lipophenol candidate, Q-
834 3DHA-7OiP.
835
836 Experimental Section
837
838
1. Chemical synthesis
Experimental Procedures. The detailed discussion and the experimental procedures of
839 chemical/enzymatic synthesis of all intermediates and final lipophenols are described in the
840 supporting information (sections S1 to S7), as well as their full analytic characterization (1H and
841 13C NMR, HRMS analysis, Rf, melting point). Known compounds were prepared according to
842 previously described procedures: mono-isopropyl-phloroglucinol (1) and P-OiP-ODHA (LEAD
843 B) [23], P-OiP-OLA (LEAD A) [24], R-4’LA and R-4’DHA [43], Q-3LA, Q-3DHA and C-3LA
844 [42].

845
846
847
2. Biological evaluations
Chemicals. All lipophenols were dissolved in dimethylsulfoxide (DMSO) to prepare a stock
848 solution at 80 mM. Hydrogen peroxide solution (H2O2, 30 wt. % in H2O), all-trans retinal
849 (atRAL) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) were
850 purchased from Sigma-Aldrich. N-retinylidene-N-retinylethanolamine (A2E) was synthesized as
851 previously described by Parish et al. in 1998 [84]. 2’,7’-dichlorofluorescin diacetate (DCFDA)
852 was purchased from Sigma-Aldrich (Saint-Quentin, France) and dissolved in DMSO to prepare
853 stock solution at 20 mM. All stock solutions of lipophenols, atRAL, A2E and probe were stored
854 at -20 °C in the dark.
855
Cell Culture. ARPE-19 cells were obtained from ATCC (USA) and were grown in Dulbecco’s
856 Modified Eagle’s Medium (DMEM)/Ham F12 (GIBCO) containing 10% v/v fetal bovine serum
857 (FBS) and 1% v/v penicillin/streptomycin under atmospheric humidified air (95%) / CO2 (5%) at
858 37 °C. For experimental cell seeding and sub-culturing, the cells were dissociated with 0.25%
859 trypsin-EDTA, resuspended in the culture medium and then plated at 1–3 × 105 cells/mL. ARPE-
860 19 cells were cultured and used up to 15 passages. Primary RPE cells were established from
861 Long-Evans newborn rats according to the procedure described previously [26,77]. Briefly, after
862 enucleation of the eyes, intact RPE sheets were separated from the choroid and dissociated in a
863 trypsin-EDTA solution (GIBCO) to obtain a suspension of single cells. RPE cells, cultured in 96-
864 well plates, reached 80–85% of confluence after 3 days in atmospheric humidified air (95%) /
865 CO2 (5%) at 37 °C, and were used for cell assays without passaging.
866
Cell Viability. Cell viability was determined in ARPE-19 and primary rat RPE cells by MTT
867 colorimetric assay. The cells were incubated for 2 h with MTT reagent (0.5 mg/mL). During this

868 incubation time, mitochondrial dehydrogenases of the living cells reduce the MTT to insoluble
869 purple formazan, which was then dissolved with DMSO to form a colored solution. The
870 absorbance of supernatants, which was proportional to the number of living cells, was measured
871 at 570 nm and 655 nm using a microplate reader (BioRad 550, USA or CLARIOstar Plus, BMG
872 Labtech). The absorbance of the compounds tested does not interfere with the absorbance at 570
873 and 655 nm. The percentage of the viable cells was calculated as [(OD570 sample−OD655
874 sample)/(OD570 control−OD655 control)]×100%.
875
Cytotoxicity of Lipophenols. ARPE-19 cells were plated into 96-well plates (4 × 104
876 cells/well) and cultured for 24 h to reach confluence before lipophenol treatment. The cell
877 cultures were treated with serum free medium containing the lipophenols at different
878 concentrations (0–160 µM) for 24 h. Control cells were incubated with DMSO (0.2%). The
879 viability of the cells was determined using MTT colorimetric assay, as described above, and
880 expressed as a percentage of viable cells normalized with control conditions in the absence of
881 lipophenols.
882
Protection of Lipophenols against atRAL Toxicity. ARPE-19 cells were plated into 96-well
883 plates (4 × 104 cells/well) and cultured for 24 h to reach confluence before lipophenol treatment.
884 The cell cultures were treated with serum free DMEM/F12 medium containing lipophenols at
885 different concentrations (0–80 µM) for 1 h. Then atRAL was added to a final concentration of 15
886 µM for 4 h before rinsing with medium. Control cells were incubated with DMSO (0.2%) ±
887 atRAL. The cells incubated at 37 °C and viability was determined 16–20 h later using a MTT
888 colorimetric assay. For primary rat RPE cultures, cells were treated with atRAL (25 µM) in the
889 presence of lipophenol (40–160 µM) for 4 h before cell viability determination. Results are

890 expressed in percentage of viable cells normalized with control conditions in the absence of
891 lipophenol and stressor.
892
Impact of Lipophenols on ROS Level. ROS level was measured in ARPE-19 and primary rat
893 RPE cells using dichlorofluorescein diacetate (DCFDA) reagent. The cell permeant reagent
894 DCFDA is deacetylated by cellular esterases to dichlorofluorescein (DCFH2), which can be
.
895 oxidized by several radical reactive species (peroxyl, alkoxyl, NO2 , carbonate, HO., …) into the
896 fluorophore 2’,7’–dichlorofluorescein (DCF) [85]. Intensity of fluorescence was measured
897 during DCFDA oxidation by radical species to calculate level of radical reactive species. ARPE-
898 19 cells were plated into black, optically clear flat bottom 96-well plates (4 × 104 cells/well) and
899 cultured for 24 h to reach confluence before the drug treatment. The cell cultures were incubated
900 with 2 µM of DCFDA for 45 min in DMEM/F12 medium without phenol red + 1% FBS. The
901 cells were rinsed and incubated with the medium containing lipophenols at different
902 concentrations (0–80 µM) for 1 h. Then, H2O2 was added to a final concentration of 600 µM for
903 4 h. Primary rat RPE cells were seeded on white, opaque-bottomed 96-well plates. On day 3,
904 cells were incubated for 45 minutes at 37 °C in 1X Buffer containing 25 µM of DCFDA. The
905 cells were then treated with 600 µM of H2O2 in the presence of lipophenol (40–160 µM) for 4 h
906 at 37 °C. For both ARPE-19 and primary cells, DCF production was measured by fluorescence
907 spectroscopy with excitation wavelength at 485 nm and emission wavelength at 535 nm. The
908 fluorescence of the compounds tested does not interfere with DCFDA signal. Control cells were
909 incubated with DMSO (0.2%) ± DCFDA ± H2O2. The percentage of ROS produced was
910 calculated as [(fluorescence of sample)/(fluorescence of control)]×100%. The results are
911 expressed in percentage of ROS produced normalized with control conditions in the absence of
912 lipophenol and presence of H2O2.

913
Protection of Lipophenols against Photo-Oxidized A2E Toxicity. ARPE-19 cells were
914 plated into 96-well plates (4 × 104 cells/well) and cultured for 24 h to reach confluence before
915 lipophenol treatment. The cell cultures were treated with serum free DMEM/F12 medium
916 without phenol red containing lipophenols at different concentrations (0–80 µM) for 1 h. Then
917 A2E was added to a final concentration of 20 µM for 6 h before rinsing with medium. Control
918 cells were incubated with DMSO (0.2%) ± A2E. The cells were exposed to intense blue light
919 (4600 LUX) for 30 min to induce phototoxicity of A2E and incubated at 37 °C. The cell viability
920 was determined 16–20 h later using a MTT colorimetric assay. Results are expressed as a
921 percentage of viable cells normalized with control conditions in the absence of lipophenols and
922 stressor. When a dose-dependent efficiency was observed, EC50 was calculated.
923
Statistical Analysis. The data are presented as means ± SEM determined from at least three
924 independent experiments. In each experiment, all conditions were done at least in quadruplicate.
925 Statistical analyses were performed by Oneway ANOVA test using Newman-Keuls’s post-hoc
926 for Gaussian distributions (the normality of distributions was evaluated with a Shapiro-Wilk test)
927 and differences with p-values < 0.05 were considered as statistically significant. EC50 were
928 calculated using GraphPad Prism version 5.03 and non-linear regression.
929
930 Associated Content
931 Supporting Information
932 Full discussion on chemical synthesis of C-phloroglucinol derivatives (Section S1);
933 Characterization of the silylated chromane orthoester byproduct of compound 5 (Section S2);
934 Full discussion on chemical synthesis of resveratrol derivatives (Section S3); Full discussion on
935 chemical synthesis of (+)-catechin derivatives (Section S4); Full discussion on chemical
936 synthesis of quercetin-5OiP derivatives (Section S5); Full discussion on chemical synthesis of

937 quercetin-7OiP derivatives (Section S6); Experimental procedure and full analysis
1
938 characterization of all intermediates and final compounds (2a to 39b, Section S7); H and 13C
939 NMR spectra for all intermediates and final compounds (2a to 39b, Section S8).
940
941 Author Contributions
942 All authors contributed to the writing of the manuscript. All authors gave approval to the final
943 version of the manuscript. PB and CC contributed equally to this work.
944 Funding Sources
945 This research was funded by University of Montpellier, SATT AxLR (181088BU) and ANR
946 (LipoPheRet - ANR-18-CE18-0017). Inserm and CNRS are thanked for their support.
947
948 Acknowledgments
949 The authors would like to thank Dr. Gaetan Drouin and Pr. Vincent Rioux for providing DHA-
950 ethyl ester necessary for lipophenol synthesis, and Dr. Vasiliki Kalatzis for English correction of
951 the manuscript.
952
953
954
References
[1] J.Z. Nowak, Oxidative stress, polyunsaturated fatty acids-derived oxidation products and
955 bisretinoids as potential inducers of CNS diseases: focus on age-related macular degeneration,
956 Pharmacol. Rep. 65 (2013) 288–304.

957
[2] J.R. Sparrow, E. Gregory-Roberts, K. Yamamoto, A. Blonska, S.K. Ghosh, K. Ueda, J.
958 Zhou, The Bisretinoids of Retinal Pigment Epithelium, Prog. Retin. Eye Res. 31 (2012) 121–
959 135.
960
[3] J.T. Handa, M. Cano, L. Wang, S. Datta, T. Liu, Lipids, oxidized lipids, oxidation-
961 specific epitopes, and Age-related Macular Degeneration, Biochim. Biophys. Acta 1862 (2017)
962 430–440.
963
[4] A.V. Cideciyan, T.S. Aleman, M. Swider, S.B. Schwartz, J.D. Steinberg, A.J. Brucker,
964 A.M. Maguire, J. Bennett, E.M. Stone, S.G. Jacobson, Mutations in ABCA4 result in
965 accumulation of lipofuscin before slowing of the retinoid cycle: a reappraisal of the human
966 disease sequence, Hum Mol Genet. 13 (2004) 525–534.
967
[5] J.R. Sparrow, N. Fishkin, J. Zhou, B. Cai, Y.P. Jang, S. Krane, Y. Itagaki, K. Nakanishi,
968 A2E, a byproduct of the visual cycle, Vision Res. 43 (2003) 2983–2990.
969
[6] S.R. Kim, S. Jockusch, Y. Itagaki, N.J. Turro, J.R. Sparrow, Mechanisms involved in
970 A2E oxidation, Exp. Eye Res. 86 (2008) 975–982.
971
[7] Y. Wu, E. Yanase, X. Feng, M.M. Siegel, J.R. Sparrow, Structural characterization of
972 bisretinoid A2E photocleavage products and implications for age-related macular degeneration,
973 Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 7275–7280.
974
[8] A.E. Sears, P.S. Bernstein, A.V. Cideciyan, C. Hoyng, P.C. Issa, K. Palczewski, P.J.
975 Rosenfeld, S. Sadda, U. Schraermeyer, J.R. Sparrow, I. Washington, H.P.N. Scholl, Towards
976 Treatment of Stargardt Disease: Workshop Organized and Sponsored by the Foundation Fighting
977 Blindness, Transl. Vis. Sci. Technol. 6 (2017) 1–32.

978
[9] D. Kook, A.H. Wolf, A.L. Yu, A.S. Neubauer, S.G. Priglinger, A. Kampik, U.C. Welge-
979 Lussen, The protective effect of quercetin against oxidative stress in the human RPE in vitro,
980 Invest Ophthalmol. Vis. Sci. 49 (2008) 1712–1720.
981
[10] R.E. King, K.D. Kent, J.A. Bomser, Resveratrol reduces oxidation and proliferation of
982 human retinal pigment epithelial cells via extracellular signal-regulated kinase inhibition, Chem.
983 Biol. Interact. 151 (2005) 143–149.
984
[11] W. Kalt, A. Hanneken, P. Milbury, F. Tremblay, Recent research on polyphenolics in
985 vision and eye health, J. Agric. Food Chem. 58 (2010) 4001–4007.
986
[12] S. Quideau, D. Deffieux, C. Douat-Casassus, L. Pouységu, Plant Polyphenols: Chemical
987 Properties, Biological Activities, and Synthesis, Angew. Chem. Int. Ed. 50 (2011) 586–621.
988
[13] H.J. Forman, K.J.A. Davies, F. Ursini, How do nutritional antioxidants really work:
989 Nucleophilic tone and para-hormesis versus free radical scavenging in vivo, Free Radic. Biol.
990 Med. 66 (2014) 24–35.
991
[14] Y. Xie, X. Chen, Structures required of polyphenols for inhibiting advanced glycation
992 end products formation, Curr. Drug Metab. 14 (2013) 414–431.
993
[15] I. Sadowska-Bartosz, G. Bartosz, Prevention of Protein Glycation by Natural
994 Compounds, Molecules. 20 (2015) 3309–3334.
995
[16] C.Y. Lo, W.T. Hsiao, X.Y. Chen, Efficiency of trapping methylglyoxal by phenols and
996 phenolic acids, J. Food Sci. 76 (2011) 90–96.

997
[17] X. Shao, H. Chen, Y. Zhu, R. Sedighi, C.-T. Ho, S. Sang, Essential Structural
998 Requirements and Additive Effects for Flavonoids to Scavenge Methylglyoxal, J. Agric. Food
999 Chem. 62 (2014) 3202–3210.
1000
[18] W. Wang, Y. Qi, J.R. Rocca, P.J. Sarnoski, A. Jia, L. Gu, Scavenging of Toxic Acrolein
1001 by Resveratrol and Hesperetin and Identification of Adducts, J. Agric. Food Chem. 63 (2015)
1002 9488–9495.
1003
[19] F. Yin, X. Wang, Y. Hu, H. Xie, X. Liu, L. Qin, J. Zhang, D. Zhou, F. Shahidi,
1004 Evaluation of Absorption and Plasma Pharmacokinetics of Tyrosol Acyl Esters in Rats, J. Agric.
1005 Food Chem. 68 (2020) 1248–1256.
1006
[20] O.M. Feeney, M.F. Crum, C.L. McEvoy, N.L. Trevaskis, H.D. Williams, C.W. Pouton,
1007 W.N. Charman, C.A.S. Bergström, C.J.H. Porter, 50 years of oral lipid-based formulations:
1008 Provenance, progress and future perspectives, Adv. Drug Deliv. Rev. 101 (2016) 167–194.
1009
[21] C.W. Pouton, Lipid formulations for oral administration of drugs: non-emulsifying, self-
1010 emulsifying and ‘self-microemulsifying’ drug delivery systems, Eur. J. Pharm. Sci. 11 (2000)
1011 S93–S98.
1012
[22] C. Crauste, M. Rosell, T. Durand, J. Vercauteren, Omega-3 polyunsaturated lipophenols,
1013 how and why?, Biochimie. 120 (2016) 62–74.
1014
[23] C. Crauste, C. Vigor, P. Brabet, M. Picq, M. Lagarde, C. Hamel, T. Durand, J.
1015 Vercauteren, Synthesis and Evaluation of Polyunsaturated Fatty Acid–Phenol Conjugates as
1016 Anti-Carbonyl-Stress Lipophenols, Eur. J. Org. Chem. (2014) 4548–4561.

1017
[24] A. Cubizolle, D. Cia, E. Moine, N. Jacquemot, L. Guillou, M. Rosell, C. Angebault
1018 Prouteau, G. Lenaers, I. Meunier, J. Vercauteren, T. Durand, C. Crauste, P. Brabet, Isopropyl-
1019 phloroglucinol-DHA protects outer retinal cells against lethal dose of all-trans-retinal, J. Cell.
1020 Mol. Med. 24 (2020) 5057–5069.
1021
[25] T. Georgiou, A. Neokleous, D. Nicolaou, B. Sears, Pilot study for treating dry age-related
1022 macular degeneration (AMD) with high-dose omega-3 fatty acids, PharmaNutrition. 2 (2014) 8–
1023 11.
1024
[26] D. Cia, A. Cubizolle, C. Crauste, N. Jacquemot, L. Guillou, C. Vigor, C. Angebault, C.P.
1025 Hamel, J. Vercauteren, P. Brabet, Phloroglucinol protects retinal pigment epithelium and
1026 photoreceptor against all trans retinal–induced toxicity and inhibits A2E formation, J. Cell.
1027 Mol. Med. 20 (2016) 1651–1663.
1028
[27] M. Suh, A.A. Wierzbicki, E. Lien, M.T. Clandinin, Relationship between dietary supply
1029 of long-chain fatty acids and membrane composition of long- and very long chain essential fatty
1030 acids in developing rat photoreceptors, Lipids. 31 (1996) 61–64.
1031
[28] P.K. Mukherjee, V.L. Marcheselli, C.N. Serhan, N.G. Bazan, Neuroprotectin D1: a
1032 docosahexaenoic acid-derived docosatriene protects human retinal pigment epithelial cells from
1033 oxidative stress, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 8491–8496.
1034
[29] P. Brabet, A. Cubizolle, C. Crauste, L. Guillou, A.L. Bonnefont, J. Vercauteren, T.
1035 Durand, C.P. Hamel, A phloroglucinol-DHA derivative protects against light-induced retinal
1036 degeneration in an Abca4-deficient mouse model, Invest. Ophthalmol. Vis. Sci. 58 (2017) 2031.

1037
[30] N. Taveau, A. Cubizolle, L. Guillou, N. Pinquier, E. Moine, D. Cia, J. Vercauteren, T.
1038 Durand, C. Crauste, P. Brabet, Preclinical pharmacology of a lipophenol in a mouse model of
1039 light-induced retinopathies., Exp. Mol. Med. 52 (2020) 1090–1101.
1040
[31] Q. Zhu, Z.-P. Zheng, K.-W. Cheng, J.-J. Wu, S. Zhang, Y.S. Tang, K.-H. Sze, J. Chen, F.
1041 Chen, M. Wang, Natural polyphenols as direct trapping agents of lipid peroxidation-derived
1042 acrolein and 4-hydroxy-trans-2-nonenal, Chem. Res. Toxicol. 22 (2009) 1721–1727.
1043
[32] W. Wang, Antiglycation effects and reactive carbonyl trapping capacity of berry and
1044 grape phytochemicals, Thesis of master of sciences of University of Florida. (2010) 1–129.
1045
[33] A. Alaimo, M.C. Di Santo, A.P. Domínguez Rubio, G. Chaufan, G. García Liñares, O.E.
1046 Pérez, Toxic effects of A2E in human ARPE-19 cells were prevented by resveratrol: a potential
1047 nutritional bioactive for age-related macular degeneration treatment, Arch. Toxicol. 94 (2020)
1048 553–572.
1049
[34] Y. Wang, H.J. Kim, J.R. Sparrow, Quercetin and cyanidin-3-glucoside protect against
1050 photooxidation and photodegradation of A2E in retinal pigment epithelial cells, Exp. Eye Res.
1051 160 (2017) 45–55.
1052
[35] W. Li, Y. Jiang, T. Sun, X. Yao, X. Sun, Supplementation of procyanidins B2 attenuates
1053 photooxidation-induced apoptosis in ARPE-19 cells, Int. J. Food Sci. Nutr. 67 (2016) 650–659.
1054
[36] O. Gia, A. Anselmo, M.T. Conconi, C. Antonello, E. Uriarte, S. Caffieri, 4‘-Methyl
1055 Derivatives of 5-MOP and 5-MOA:ꢀ Synthesis, Photoreactivity, and Photobiological Activity, J.
1056 Med. Chem. 39 (1996) 4489–4496.

1057
[37] M.J. So, E.J. Cho, Phloroglucinol Attenuates Free Radical-induced Oxidative Stress,
1058 Prev. Nutr. Food Sci. 19 (2014) 129–135.
1059
[38] M. Vestergaard, H. Ingmer, Antibacterial and antifungal properties of resveratrol, Int. J.
1060 Antimicrob. Agents. 53 (2019) 716–723.
1061
[39] J.-H. Kang, S.-Y. Choung, Protective effects of resveratrol and its analogs on age-related
1062 macular degeneration in vitro, Arch. Pharm. Res. 39 (2016) 1703–1715.
1063
[40] A. Lançon, R. Frazzi, N. Latruffe, Anti-Oxidant, Anti-Inflammatory and Anti-
1064 Angiogenic Properties of Resveratrol in Ocular Diseases, Molecules. 21 (2016) 1–8.
1065
[41] I.C. Vlachogianni, E. Fragopoulou, I.K. Kostakis, S. Antonopoulou, In vitro assessment
1066 of antioxidant activity of tyrosol, resveratrol and their acetylated derivatives, Food Chem. 177
1067 (2015) 165–173.
1068
[42] E. Moine, P. Brabet, L. Guillou, T. Durand, J. Vercauteren, C. Crauste, New Lipophenol
1069 Antioxidants Reduce Oxidative Damage in Retina Pigment Epithelial Cells, Antioxidants. 7
1070 (2018) 1–18.
1071
[43] A. Shamseddin, C. Crauste, E. Durand, P. Villeneuve, G. Dubois, T. Pavlickova, T.
1072 Durand, J. Vercauteren, F. Veas, Resveratrol-Linoleate protects from exacerbated endothelial
1073 permeability via a drastic inhibition of the MMP-9 activity, Biosci. Rep. (2018) 1–13.
1074
[44] C. Cren-Olivé, S. Lebrun, C. Rolando, An efficient synthesis of the four mono
1075 methylated isomers of (+)-catechin including the major metabolites and of some dimethylated
1076 and trimethylated analogues through selective protection of the catechol ring, J. Chem. Soc.
1077 Perkin 1. (2002) 821–830.

1078
[45] Y. Shen, Z. Xu, Z. Sheng, Ability of resveratrol to inhibit advanced glycation end product
1079 formation and carbohydrate-hydrolyzing enzyme activity, and to conjugate methylglyoxal, Food
1080 Chem. 216 (2017) 153–160.
1081
[46] S.-J. Sheu, N.-C. Liu, C.-C. Ou, Y.-S. Bee, S.-C. Chen, H.-C. Lin, J.Y.H. Chan,
1082 Resveratrol Stimulates Mitochondrial Bioenergetics to Protect Retinal Pigment Epithelial Cells
1083 From Oxidative Damage, Invest. Ophthalmol. Vis. Sci. 54 (2013) 6426–6438.
1084
[47] K.E. Heim, A.R. Tagliaferro, D.J. Bobilya, Flavonoid antioxidants: chemistry,
1085 metabolism and structure-activity relationships, J. Nutr. Biochem. 13 (2002) 572–584.
1086
[48] S.D.S. Banjarnahor, N. Artanti, Antioxidant properties of flavonoids, Med. J. Indones. 23
1087 (2015) 239–244.
1088
[49] R. Khursheed, S.K. Singh, S. Wadhwa, M. Gulati, A. Awasthi, Enhancing the potential
1089 preclinical and clinical benefits of quercetin through novel drug delivery systems, Drug Discov.
1090 Today. 25 (2020) 209-222.
1091
[50] H. Kalantari, H. Foruozandeh, M.J. Khodayar, A. Siahpoosh, N. Saki, P. Kheradmand,
1092 Antioxidant and hepatoprotective effects of Capparis spinosa L. fractions and Quercetin on tert-
1093 butyl hydroperoxide- induced acute liver damage in mice, J. Tradit. Complement. Med. 8 (2018)
1094 120–127.
1095
[51] A.J. Dugas, J. Castañeda-Acosta, G.C. Bonin, K.L. Price, N.H. Fischer, G.W. Winston,
1096 Evaluation of the Total Peroxyl Radical-Scavenging Capacity of Flavonoids:ꢀ Structure−Activity
1097 Relationships, J. Nat. Prod. 63 (2000) 327–331.

1098
[52] K.L. Wolfe, R.H. Liu, Structure−Activity Relationships of Flavonoids in the Cellular
1099 Antioxidant Activity Assay, J. Agric. Food Chem. 56 (2008) 8404–8411.
1100
[53] S. Hong, S. Liu, Targeted acylation for all the hydroxyls of (+)-catechin and evaluation of
1101 their individual contribution to radical scavenging activity, Food Chem. 197 (2016) 415–421.
1102
[54] N.I. Howard, M.V.B. Dias, F. Peyrot, L. Chen, M.F. Schmidt, T.L. Blundell, C. Abell,
1103 Design and Structural Analysis of Aromatic Inhibitors of Type II Dehydroquinase from
1104 Mycobacterium tuberculosis, ChemMedChem. 10 (2015) 116–133.
1105
[55] S. Uesato, K. Taniuchi, A. Kumagai, Y. Nagaoka, R. Seto, Y. Hara, H. Tokuda, H.
1106 Nishino, Inhibitory effects of 3-O-acyl-(+)-catechins on Epstein-Barr virus activation, Chem.
1107 Pharm. Bull. 51 (2003) 1448–1450.
1108
[56] X. Peng, K.-W. Cheng, J. Ma, B. Chen, C.-T. Ho, C. Lo, F. Chen, M. Wang, Cinnamon
1109 Bark Proanthocyanidins as Reactive Carbonyl Scavengers To Prevent the Formation of
1110 Advanced Glycation Endproducts, J. Agric. Food Chem. 56 (2008) 1907–1911.
1111
[57] C.-Y. Lo, S. Li, D. Tan, M.-H. Pan, S. Sang, C.-T. Ho, Trapping reactions of reactive
1112 carbonyl species with tea polyphenols in simulated physiological conditions, Mol. Nutr. Food
1113 Res. 50 (2006) 1118–1128.
1114
[58] M. Li, X. Han, B. Yu, Facile Synthesis of Flavonoid 7- O -Glycosides, J. Org. Chem. 68
1115 (2003) 6842–6845.
1116
[59] B. Yang, R. Li, T. Woo, J.D. Browning, H. Song, Z. Gu, J. Cui, J.C. Lee, K.L. Fritsche,
1117 D.Q. Beversdorf, G.Y. Sun, C.M. Greenlief, Maternal Dietary Docosahexaenoic Acid Alters

1118 Lipid Peroxidation Products and (n-3)/(n-6) Fatty Acid Balance in Offspring Mice, Metabolites.
1119 9 (2019) 1–14.
1120
[60] Y. Huang, W. Li, A.-N.T. Kong, Anti-oxidative stress regulator NF-E2-related factor 2
1121 mediates the adaptive induction of antioxidant and detoxifying enzymes by lipid peroxidation
1122 metabolite 4-hydroxynonenal, Cell Biosci. 2 (2012) 1–7.
1123
[61] A. Ishikado, K. Morino, Y. Nishio, F. Nakagawa, A. Mukose, Y. Sono, N. Yoshioka, K.
1124 Kondo, O. Sekine, T. Yoshizaki, S. Ugi, T. Uzu, H. Kawai, T. Makino, T. Okamura, M.
1125 Yamamoto, A. Kashiwagi, H. Maegawa, 4-Hydroxy hexenal derived from docosahexaenoic acid
1126 protects endothelial cells via Nrf2 activation, PLoS One. 8 (2013) 1–13.
1127
[62] I. Johansson, V.T. Monsen, K. Pettersen, J. Mildenberger, K. Misund, K. Kaarniranta, S.
1128 Schønberg, G. Bjørkøy, The marine n-3 PUFA DHA evokes cytoprotection against oxidative
1129 stress and protein misfolding by inducing autophagy and NFE2L2 in human retinal pigment
1130 epithelial cells, Autophagy. 11 (2015) 1636–1651.
1131
[63] Z. Sun, X. Peng, J. Liu, K.-W. Fan, M. Wang, F. Chen, Inhibitory effects of microalgal
1132 extracts on the formation of advanced glycation endproducts (AGEs), Food Chem. 120 (2010)
1133 261–267.
1134
[64] M. Odjakova, E. Popova, M.A. Sharif, R. Mironova, Plant-Derived Agents with Anti-
1135 Glycation Activity, in: S. Petrescu (Eds.), Glycosylation, IntechOpen Publishing, chapter
1136 10,2012: pp. 223-256.

1137
[65] S.A. Belal, A.S. Sivakumar, D.R. Kang, S. Cho, H.S. Choe, K.S. Shim, Modulatory
1138 effect of linoleic and oleic acid on cell proliferation and lipid metabolism gene expressions in
1139 primary bovine satellite cells, Anim. Cells Syst. 22 (2018) 324–333.
1140
[66] N.P. Rotstein, L.E. Politi, O.L. German, R. Girotti, Protective Effect of Docosahexaenoic
1141 Acid on Oxidative Stress-Induced Apoptosis of Retina Photoreceptors, Invest. Ophthalmol. Vis.
1142 Sci. 44 (2003) 2252–2259.
1143
[67] J.P. SanGiovanni, E.Y. Chew, The role of omega-3 long-chain polyunsaturated fatty
1144 acids in health and disease of the retina, Prog. Retin. Eye Res. 24 (2005) 87–138.
1145
[68] E.E. Martínez Leo, R.A. Rojas Herrera, M.R. Segura Campos, Protective Effect of
1146 Omega 3 Fatty Acids EPA and DHA in the Neurodegenerative Disease, in: J.-M. Mérillon, K.G.
1147 Ramawat (Eds.), Bioact. Mol. Food, Springer International Publishing, Cham, 2018: pp. 1–17.
1148
[69] M. Tachikawa, S. Akanuma, T. Imai, S. Okayasu, T. Tomohiro, Y. Hatanaka, K. Hosoya,
1149 Multiple Cellular Transport and Binding Processes of Unesterified Docosahexaenoic Acid in
1150 Outer Blood–Retinal Barrier Retinal Pigment Epithelial Cells, Biol. Pharm. Bull. 41 (2018)
1151 1384–1392.
1152
[70] C. Papuc, G.V. Goran, C.N. Predescu, V. Nicorescu, G. Stefan, Plant Polyphenols as
1153 Antioxidant and Antibacterial Agents for Shelf-Life Extension of Meat and Meat Products:
1154 Classification, Structures, Sources, and Action Mechanisms, Compr. Rev. Food Sci. Food Saf.
1155 16 (2017) 1243–1268.

1156
[71] J.R. Sparrow, H.R. Vollmer-Snarr, J. Zhou, Y.P. Jang, S. Jockusch, Y. Itagaki, K.
1157 Nakanishi, A2E-epoxides Damage DNA in Retinal Pigment Epithelial Cells, J. Biol. Chem. 278
1158 (2003) 18207–18213.
1159
[72] M. Wrona, M. Różanowska, T. Sarna, Zeaxanthin in combination with ascorbic acid or α-
1160 tocopherol protects ARPE-19 cells against photosensitized peroxidation of lipids, Free Radic.
1161 Biol. Med. 36 (2004) 1094–1101.
1162
[73] V.M. Parmar, T. Parmar, E. Arai, L. Perusek, A. Maeda, A2E-associated cell death and
1163 inflammation in retinal pigmented epithelial cells from human induced pluripotent stem cells,
1164 Stem Cell Res. 27 (2018) 95–104.
1165
[74] S. Ben Shabat, Y. Itagaki, S. Jockusch, J.R. Sparrow, N.J. Turro, K. Nakanishi,
1166 Formation of a Nonaoxirane from A2E, a Lipofuscin Fluorophore related to Macular
1167 Degeneration, and Evidence of Singlet Oxygen Involvement, Angew. Chem. Int. Ed. 41 (2002)
1168 814–817.
1169
[75] A.V. Kuznetsova, A.M. Kurinov, M.A. Aleksandrova, Cell Models to Study Regulation
1170 of Cell Transformation in Pathologies of Retinal Pigment Epithelium, J. Ophthalmol. (2014) 1–
1171 18.
1172
[76] A. Hanneken, F.-F. Lin, J. Johnson, P. Maher, Flavonoids Protect Human Retinal
1173 Pigment Epithelial Cells from Oxidative-Stress–Induced Death, Invest. Ophthalmol. Vis. Sci. 47
1174 (2006) 3164–3177.

1175
[77] D. Cia, J. Vergnaud-Gauduchon, N. Jacquemot, M. Doly, Epigallocatechin Gallate
1176 (EGCG) Prevents H O -Induced Oxidative Stress in Primary Rat Retinal Pigment Epithelial
2
2
1177 Cells, Curr. Eye Res. 39 (2014) 944–952.
1178
[78] E. Pawlowska, J. Szczepanska, A. Koskela, K. Kaarniranta, J. Blasiak, Dietary
1179 Polyphenols in Age-Related Macular Degeneration: Protection against Oxidative Stress and
1180 Beyond, Oxid. Med. Cell. Longev. 2019 (2019) 1–13.
1181
[79] K. Ishige, D. Schubert, Y. Sagara, Flavonoids protect neuronal cells from oxidative stress
1182 by three distinct mechanisms, Free Radic. Biol. Med. 30 (2001) 433–446.
1183
[80] H. Zhu, M.M. Poojary, M.L. Andersen, M.N. Lund, Effect of pH on the reaction between
1184 naringenin and methylglyoxal: A kinetic study, Food Chem. 298 (2019) 1–9.
1185
[81] S. Sang, X. Shao, N. Bai, C.-Y. Lo, C.S. Yang, C.-T. Ho, Tea Polyphenol (−)-
1186 Epigallocatechin-3-Gallate: A New Trapping Agent of Reactive Dicarbonyl Species, Chem. Res.
1187 Toxicol. 20 (2007) 1862–1870.
1188
[82] B. Szwergold, Reactions between methylglyoxal and its scavengers in-vivo appear to be
1189 catalyzed enzymatically, Med. Hypotheses. 109 (2017) 153–155.
1190
[83] Y. Chen, K. Okano, T. Maeda, V. Chauhan, M. Golczak, A. Maeda, K. Palczewski,
1191 Mechanism of All- trans -retinal Toxicity with Implications for Stargardt Disease and Age-
1192 related Macular Degeneration, J. Biol. Chem. 287 (2012) 5059–5069.
1193
[84] C.A. Parish, M. Hashimoto, K. Nakanishi, J. Dillon, J. Sparrow, Isolation and one-step
1194 preparation of A2E and iso-A2E, fluorophores from human retinal pigment epithelium, Proc.
1195 Natl. Acad. Sci. U. S. A. 95 (1998) 14609–14613.

1196
[85] B. Halliwell, M. Whiteman, Measuring reactive species and oxidative damage in vivo
1197 and in cell culture: how should you do it and what do the results mean?, Br. J. Pharmacol. 142
1198 (2004) 231–255.
1199

Highlights
•
•
•
•
•
Carbonyl and oxidative stresses play a crucial role in macular degeneration.
All-trans-retinal accumulates abnormally in AMD causing toxic A2E formation.
Lipophenol derivatives are polyphenols functionalized with PUFA.
Quercetin lipophenol is a potent photo-oxidative toxicity suppressor in RPE cells.
Lipophenols protect against toxicity induced by carbonyl and oxidative stresses.