Novel lipid-mimetic prodrugs delivering active compounds to adipose tissue

DOI: 10.1016/j.ejmech.2017.04.034

Source and publish data:

European Journal of Medicinal Chemistry p. 77 - 88 (2017)

Update date:2022-08-10

Topics:

Authors:

Mattarei, Andrea

Rossa, Andrea

Bombardelli, Veronica

Azzolini, Michele

La Spina, Martina

Paradisi, Cristina

Zoratti, Mario

Biasutto, Lucia

Read Full Text PDF DownLoad Join now for total 90,000,000 free articles

Article abstract of DOI:10.1016/j.ejmech.2017.04.034

Obesity and associated pathologies are a dramatically growing problem. New therapies to prevent and/or cure them are strongly needed. Adipose tissue is a logical target for pharmacological intervention, since it is now recognized to exert an important endocrine function, secreting a variety of adipokines affecting, for example, adiposity and insulin resistance. This proof of principle work focuses on the development of novel lipid-mimetic prodrugs reaching fat deposits by the same lymphatic absorption route followed by dietary triglycerides. Pterostilbene, a natural phenolic compound with potential anti-obesity effects, was used as model “cargo”, attached via a carbamate group to an ω-aminodecanoate chain linked to either position 1 or position 2 of the glycerol moiety of synthetic triglycerides. The prodrugs underwent position-selective hydrolysis when challenged with pancreatic lipases in vitro. Pterostilbene-containing triglycerides as well as pterostilbene and its metabolites were present in the adipose tissue of mice fed an obesogenic diet containing one or the other of the derivatives. For the first time this approach is used to deliver an obesity antagonist to the adipose tissue. The results demonstrate the feasibility of delivering active compounds to adipose tissue by reversibly incorporating them into triglyceride-mimetic structures. Upon release in the target site these compounds are expected to exert their pharmacological activity precisely where needed.

Full text of DOI:10.1016/j.ejmech.2017.04.034

European Journal of Medicinal Chemistry 135 (2017) 77e88

Contents lists available at ScienceDirect

European Journal of Medicinal Chemistry

journal homepage: h ttp://www.elsevier.com/locate/ejmech

Research paper

Novel lipid-mimetic prodrugs delivering active compounds to adipose

tissue

, c

Andrea Mattarei ^{a 1}, Andrea Rossa ^{a c}, Veronica Bombardelli ^{a c}, Michele Azzolini ^b

Martina La Spina ^c, Cristina Paradisi ^a, Mario Zoratti ^{b c}, Lucia Biasutto ^b

, c, *

^aUniversity of Padova, Department of Chemical Sciences, Via F. Marzolo 1, 35131 Padova, Italy

^bCNR Neuroscience Institute, Viale G. Colombo 3, 35121 Padova, Italy

^cUniversity of Padova, Department of Biomedical Sciences, Viale G. Colombo 3, 35121 Padova, Italy

a r t i c l e i n f o

a b s t r a c t

Article history:

Obesity and associated pathologies are a dramatically growing problem. New therapies to prevent and/or

cure them are strongly needed. Adipose tissue is a logical target for pharmacological intervention, since it

is now recognized to exert an important endocrine function, secreting a variety of adipokines affecting,

for example, adiposity and insulin resistance. This proof of principle work focuses on the development of

novel lipid-mimetic prodrugs reaching fat deposits by the same lymphatic absorption route followed by

dietary triglycerides.

Received 30 January 2017

Received in revised form

20 March 2017

Accepted 11 April 2017

Available online 13 April 2017

Pterostilbene, a natural phenolic compound with potential anti-obesity effects, was used as model

Keywords:

Pterostilbene

“cargo”, attached via a carbamate group to an

u-aminodecanoate chain linked to either position 1 or

position 2 of the glycerol moiety of synthetic triglycerides. The prodrugs underwent position-selective

hydrolysis when challenged with pancreatic lipases in vitro. Pterostilbene-containing triglycerides as

well as pterostilbene and its metabolites were present in the adipose tissue of mice fed an obesogenic

diet containing one or the other of the derivatives.

Triglyceride

Lipid-mimetics

Adipose tissue targeting

For the ﬁrst time this approach is used to deliver an obesity antagonist to the adipose tissue. The

results demonstrate the feasibility of delivering active compounds to adipose tissue by reversibly

incorporating them into triglyceride-mimetic structures. Upon release in the target site these compounds

are expected to exert their pharmacological activity precisely where needed.

1. Introduction

other organs, including the immune system, liver, muscle and the

central nervous system [5]. For example, adiponectin levels

Obesity has been ﬁrmly correlated with an increased hazard for

a range of serious ailments [1,2]. The strongest link [3] may be that

between the overweight/obese status and the metabolic syndrome,

deﬁned as “a compilation of risk factors that predispose individuals

to the development of type 2 diabetes (T2D) and cardiovascular

disease (CVD)” [4]. White adipose tissue (WAT) secretes adipokines

(leptin, adiponectin and several others) which have major (patho)

physiological effects and mediate interactions between WAT and

decrease in adiposity, contributing to obesity-associated cancero-

genesis, CVD and T2D.

Upward violation of adipose tissue homeostasis leads to the

immigration of macrophages and to an increase of pro-

inﬂammatory cytokines such as TNFa, IL-1b, IL-6, as well as of

iNOS and ROS, which can act in paracrine/endocrine fashion lead-

ing to a chronic inﬂamed state linked to a host of pathologies [6].

Systemic oxidative stress [7] contributes heavily itself to the onset

and progression of obesity-associated health problems [8]. Thus,

developing effective tools against adiposity on one hand and its

consequences on the other are strong priorities of modern phar-

macology. The most successful treatment for obesity is an appro-

priate reduction of food intake, accompanied by exercise [9].

Exercise seems also to ameliorate the inﬂammatory status, inde-

pendently from weight loss [10], and to induce “browning” of white

* Corresponding author. CNR Neuroscience Institute c/o Dept. of Biomedical

Sciences, Viale G. Colombo 3, 35121 Padova, Italy.

** Corresponding author.

E-mail addresses: andrea.mattarei@unipd.it (A. Mattarei), lucia.biasutto@cnr.it

(L. Biasutto).

Present address: Department of Pharmaceutical & Pharmacological Sciences,

University of Padova, Via F. Marzolo 5, 35131 Padova, Italy.

http://dx.doi.org/10.1016/j.ejmech.2017.04.034

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

fat [11,12]. Brown fat (BAT) [11,13,14] is a mitochondria-rich type of

adipose tissue, abundant in early life, which provides - normally

when prompted by the sympathetic nervous system - non-

shivering thermogenesis, dissipating food-derived energy to pro-

duce heat in mitochondria “uncoupled” by the expression of

Uncoupling Protein 1 (UCP1). It can thus counteract obesity. An

important role in the complex regulation of BAT is attributed to

AMPK, expressed both in the controlling CNS structures and in

adipose tissue: reduction of AMPK activity in the hypothalamus or

upregulation of it in WAT and BAT increase “browning” and energy

dissipation [15]. AMPK signaling, activated by exercise, has to do

with insulin secretion, glucose transport, mitochondrial biogenesis,

fatty acid oxidation, inﬂammation. Development of anti-obesity

drugs has registered some failures, due to important side effects,

and has so far not had a signiﬁcant impact on population-wide

obesity [16]. A few drugs are currently available which have

shown some modest efﬁcacy and side effects ranging from con-

stipation to tumorigenesis (in experimental animals) [17]. They are

costly, some are classiﬁed as controlled substances because of the

possibility of abuse, all are off-limits in pregnancy and lactation.

The possibility of BAT recruitment by pharmacological intervention

is receiving much attention [13,18]. Organ-speciﬁc upregulation of

AMPK activity may provide a strategy (see above). Indeed metfor-

min, an AMPK activator and a major anti-diabetic drug, also induces

weight loss [19].

chylomicrons. These are preferentially taken up into the lacteals of

the lymphatic system rather than into the portal vein - thus

avoiding ﬁrst-pass metabolism - and eventually move to the

thoracic duct and into the blood. Plasma triglyceride levels are

regulated by the lipoprotein lipase (LPL) system [44], which hy-

drolyzes triglycerides to fatty acids for concomitant use or storage

by the underlying tissue.

As a proof of principle study, we have thus undertaken the

synthesis of lipid-mimetic prodrugs of pterostilbene to exploit the

efﬁcient and selective distribution of triglycerides, through the

lymphatic system, to adipose tissue (Fig. 1).

In our approach the APC is incorporated into a long-chain tri-

glyceride. Two classes of isomeric derivatives can be synthesized

and tested, characterized by connection of the APC either to posi-

tion 1 (Type-1) or 2 (Type-2) of the glycerol backbone (Fig. 2).

A similar strategy has been adopted in a few previous studies

[45e48]. To our knowledge, however, in no case the construct was

meant to target the adipose tissue.

The prodrugs we synthesized underwent position-selective

hydrolysis when challenged with Pancreatic Lipase in vitro.

Pterostilbene-containing triglycerides as well as pterostilbene and

its metabolites were present in the adipose tissue of mice fed an

obesogenic diet containing one or the other of the derivatives.

These results provide evidence that this approach can succeed in

delivering a cargo to the adipose tissue after oral administration.

Natural compounds may arguably be considered the best

currently available weapons in the anti-obesity, anti-metabolic

disorder arsenal [20e22]. For example berberine, genistein, ﬂavo-

noids, catechins and resveratrol indirectly activate AMPK [23].

Resveratrol [24e26] is just the most talked-about member of the

family of bioactive stilbene polyphenols, whose mechanisms of

action are thought to overlap to a considerable degree [27]. Upre-

gulation of UCP1 by resveratrol may account for the energy dissi-

pation that must take place for a “slimming” effect [28]. Its efﬁcacy

is limited however by its prompt and extensive modiﬁcation by

Phase II metabolism enzymes, for which it represents a ready-made

target [29]. Pterostilbene, i.e. 3,5-di-O-methylresveratrol [30e32],

having only one free hydroxyl, is less prone to metabolic conjuga-

tion, has a relatively high bioavailability [33,34] and a potentially

higher efﬁcacy than resveratrol [35,36]. Of relevance, the prodrug

approach to increasing absorption and reducing metabolic conju-

gation and elimination offers more promise with pterostilbene

than with resveratrol, due to the need to protect only one hydroxyl

instead of three [37].

2. Results and discussion

2.1. Synthesis of lipid-mimetic prodrugs of pterostilbene

Pterostilbene was reversibly linked to the triglyceride structure

through a carbamate bond. The choice was based on previous re-

sults, which showed that this group is resistant to hydrolysis by

digestive enzymes in the stomach and intestine, but releases the

active compound at convenient rates after passing into the blood

and tissues [49,50]. The length of the linker chain (n in Fig. 2) has

been chosen arbitrarily as nine methylene units but can be

modulated to optimize the performance of the derivatives. Details

of the syntheses are provided in Materials and Methods.

The ability of pterostilbene to antagonize the metabolic syn-

drome has been reported by several studies [38e40]. Like resver-

atrol, it can activate the Keap/Nrf2/ARE anti-oxidant pathway,

upregulate SIRT1, repress NF-kB activity, activate AMPK. It has been

reported to upregulate adiponectin in an in vitro 3T3-L1 adipocyte

cell model [41] and to downregulate instead the secretion of pro-

inﬂammatory cytokines upon interaction of these cells with RAW

264.7 macrophages [42].

It is clear that a strategy for the selective pharmacological tar-

geting of adipose tissue would represent a powerful tool in the ﬁght

against obesity and obesity-related problems. As discussed, the

issue is not just that of reducing fat deposits, but also that of

relieving chronic inﬂammation where needed.

Based on these premises, this work is focused on the develop-

ment of a tool to selectively convey a representative active phenolic

compound (APC), pterostilbene, to adipose tissue by imitating in-

testinal lipid absorption [43]. In the intestinal lumen dietary tri-

glycerides are degraded by pancreatic lipases (PL), which hydrolyze

the ester bonds selectively at position 1 in the glycerol backbone.

Free fatty acids and monoglycerides pass then into enterocytes,

where triglycerides are resynthesized and packed into

Fig. 1. Absorption mechanisms expected for an active phenolic compound (APC) and

for its triglyceride-mimetic prodrug. Only the Type-2 derivative is depicted.

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

Fig. 2. Molecular structures of resveratrol, pterostilbene and lipid-mimetic prodrugs of

active phenolic compounds.

2.2. Pancreatic lipase (PL) assays

Type-1 and Type-2 lipid-mimetic derivatives of pterostilbene

were subjected to PL-catalyzed hydrolysis to verify whether the

enzyme recognized them as “normal” triglycerides despite the

presence of the cargo molecule.

Since PL catalyzes the hydrolysis of triglycerides ester bonds

with a marked preference for the external positions (1 and 3 of the

glycerol backbone), one would expect to observe as main hydrolysis

processes those shown in Fig. 3. Type-1 is a chiral compound, but

structural differences distinguishing substituents are distant from

the chiral carbon (C-2) and the site of hydrolysis. It is therefore

unlikely that the two enantiomers undergo hydrolysis at detectably

different rates.

The kinetic results are reported in two graphs (Fig. 4) showing

the variation in time of the amount of Type-1 (Fig. 4A) and Type-2

(Fig. 4B) derivatives and of their hydrolysis products, detected and

quantiﬁed by HPLC/UV. In the chromatograms relative to the Type-

1 derivative, as expected, only three species were visible: the Type-

1 derivative itself, the product of partial hydrolysis 1, and the car-

boxylic acid 2. As shown in Fig. 4A, 2 formed at a rate comparable to

that of the disappearance of Type-1. The only other near UV-

absorbing molecule detected in the reaction solution, 1, formed

from the start and disappeared after reaching a maximum relative

concentration of less than 5%. Fig. 4B shows the concentration

proﬁles of four species: the Type-2 derivative, the products of

partial hydrolysis 3 and 4, and the carboxylic acid 2. The formation

of the latter in this case was slower than in the case of Type-1.

Moreover, the high concentration reached by 4 and its relative

stability under the reaction conditions is also consistent with the

known preference of PL enzymes for attack at the external chains

(positions 1 and 3 of the glycerol backbone), while the hydrolysis of

the ester bond in position 2 is considerably slower.

Fig. 3. Main hydrolysis process of Type-1 (panel A) and of Type-2 (panel B) derivatives

in the presence of PL. Only the molecules containing pterostilbene in their structure

are detectable by HPLC/UV (by their absorption, 300e320 nm).

of pterostilbene in the chain linked at that position. Since the levels

of 1 are very low and its generation is regulated by k₁, we can

hypothesize that its hydrolysis is very rapid. This hypothesis is

supported by kinetic analysis of the appearance of 2; ﬁtting the data

with a simple exponential equation (y ¼ 100 - y₀$e^-kt) yields a ki-

netic constant (2.88 0.07 h^ꢀ1) very similar to that describing the

disappearance of the Type-1 derivative. The ﬁtting of the experi-

mental data is shown in Fig. 4A. In the case of the Type-2 prodrug,

kinetic analysis of the data was performed by assuming that hy-

drolysis of the ester bonds occurs via consecutive losses of the three

acyl chains in pseudo ﬁrst-order processes (see Fig. 3B). The

experimental data were ﬁtted to estimate the hydrolysis rate con-

stants (k₁, k₂, k₃) using the set of equations utilized by Mattarei et al.

[51] for similar three-step hydrolysis processes. To simplify the

kinetic analysis, the rate constants of hydrolysis at position 2 of

Type-2 and 3 were considered to be negligible in comparison to k₁

and k₂. This assumption is justiﬁed a posteriori by the relative

values calculated for k₁, k₂and k₃(Fig. 4B). The rate constants

relative to the hydrolysis of the two external ester bonds (k₁and k₂)

have the same order of magnitude. k₂is larger indicating that hy-

drolysis of the second ester bond in the external positions of the

glycerol backbone is faster than hydrolysis of the ﬁrst. According to

the model, the rate constant for the hydrolysis of the central ester

bond (k₃) is considerably lower: approximately 1/13 of k₁and 1/40

of k₂respectively.

The disappearance of the Type-1 prodrug is adequately ﬁtted by

an exponential decay; th₀e derived rate constant k₁(3.24 0.09 h^ꢀ1

which is the sum of k₁and k₁⁰⁰, i.e. the two parallel hydrolysis

processes shown in Fig. 3A) is very similar to that obtained for the

Type-2 derivative (3.04 0.07 h^ꢀ1; see below), and suggests that

enzymatic hydrolysis at position 1 is not inﬂuenced by the presence

In summary, the results of these experiments conﬁrm that PL

recognizes these lipid-mimetic prodrugs and handles them as if

they were normal dietary lipids. Interestingly in neither case we

were able to detect, even after 23 h of reaction time, the formation

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

Fig. 4. Time-dependent relative abundance of Type-1 (panel A) and Type-2 (panel B) derivatives and UV-detectable species originated by their hydrolysis in the presence of PL.

of any pterostilbene, demonstrating the stability of the carbamate

bond toward PL.

region comprising the triglyceride peaks for two exemplary sam-

ples obtained from a mouse chronically fed Type-2 compound

(trace a) and from a mouse chronically fed Type-1 compound (trace

b). At least four non-background peaks are clearly present at

retention times between 5.8 and 6.5 min in trace a (Type-2). In

trace b (Type-1) the three peaks eluting at shorter retention times

are barely, if at all, visible. Other minor HPLC peaks with an ab-

sorption spectrum typical of pterostilbene-containing species and

eluting earlier than the triglycerides were also observed in most

samples and can be attributed to partial hydrolysis products. The

interpretation of the analytical data we propose is as follows: the

partial hydrolysis by pancreatic lipases in the intestinal lumen and

subsequent re-esteriﬁcation into triglycerides in the enterocytes

originates derivatives bearing fatty acid chains which may be

different from those present in the administered derivative. This

may account for the multiplicity of peaks observed in the triglyc-

eride region. The major peak remains that corresponding to the

administered prodrug, indicating that part of the latter actually

reaches the adipose tissue in its original form. Interestingly, the

peaks tentatively attributed to an exchange of the fatty acid chains

2.3. Tissue distribution

To assess if lipid mimetic derivatives actually reached the adi-

pose tissue and regenerated pterostilbene in vivo, we performed

tissue distribution experiments in mice, chronically administering

the derivatives for 2 weeks in a high-fat diet regimen (see Materials

and Methods for details).

Derivatives were found to reach in intact form the visceral and

subcutaneous adipose tissue (Fig. 5A); while statistical signiﬁcance

was not reached, in both cases the levels of the Type-2 compound

were higher than those of Type-1. Fig. 5B shows illustrative HPLC/

UV (320 nm) chromatograms of the extract from the visceral adi-

pose tissue of an animal chronically fed Type-2 derivative in high-

fat chow (trace a), of the extract from adipose tissue from an un-

treated mouse (control) spiked with Type-2 derivative (trace b) and

of an extract from control adipose tissue spiked with pterostilbene

(trace c). Panel C shows an expanded view of the chromatogram

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

Taken together, the data support the hypothesis that lipid-

mimetic derivatives are actually absorbed and handled in vivo as

normal dietary triglycerides.

Finally, we carried out the determination of pterostilbene and of

its main phase II metabolites, pterostilbene sulfate and glucuronide,

in major organs of mice receiving either Type-1 or Type-2 triglyc-

eride (Fig. 6A and B). Pterostilbene was present in the adipose tis-

sues in both cases. Among the tissues analyzed, liver and adipose

tissue turned out to have the highest levels of pterostilbene itself,

while undetectable or very low concentrations of pterostilbene

were found in the bloodstream (Fig. 6A and B).

To conclude, we have synthesized two variants of a triglyceride

incorporating an obesity-antagonizing natural compound. These

derivatives behave as expected for natural triglycerides in in vitro

lipolysis assays and upon oral administration in vivo, reaching the

adipose tissue with their “cargo”. They may therefore represent a

useful strategy to convey pharmacologically active molecules to

this tissue.

2.4. Innovation

This study demonstrates for the ﬁrst time that lipid-mimetic

Fig. 5. A) Concentration of Type-1 and -2 derivatives in visceral and subcutaneous

adipose tissue after a 2-week chronic treatment (see text for details). Mean values st.

dev., N ꢁ 4. B) HPLC/UV chromatograms (320 nm) of adipose tissue extracts from a

mouse treated with Type-2 derivative for 2 weeks (trace a), or from an untreated

animal, spiked with Type-2 derivative (trace b) or pterostilbene (trace c). C) Expanded

segments of HPLC/UV chromatograms (320 nm) produced by extracts from the visceral

adipose tissue of a mouse fed our Type-2 compound (trace a) and from the visceral

adipose tissue of a mouse fed Type-1 compound (trace b). Shown is the region con-

taining peaks attributed to triglycerides.

are much more evident in the case of Type-2 derivative: they may

represent roughly as much pterostilbene as the intact prodrug it-

self. In this case the ester bond connecting the pterostilbene-

containing chain to the glycerol backbone is resistant to lipase-

mediated hydrolysis: pterostilbene therefore remains attached

while the oleic acid linked to positions 1 and 3 is (in part) traded for

other (endogenous) fatty acids; the resulting triglycerides reach the

adipose tissue and can be detected thanks to the characteristic UV-

absorption by the pterostilbene moiety. In the case of the Type-1

derivative pterostilbene is linked to position 1. It is therefore lost

preferentially, and only in a minority of cases are the pterostilbene-

containing chains “recycled” and joined to a glycerol backbone,

eventually reaching the adipose tissue.

Fig. 6. Concentration of pterostilbene and metabolites in different organs of mice after

a 2-week chronic administration of Type-1 (panel A) or Type-2 (panel B) derivatives

(please see Materials and Methods for experimental details). Mean values st. dev.,

N ꢁ 5.

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

derivatives are actually able to target the adipose tissue, and they

are thus a useful strategy to convey pharmacologically active

molecules to this tissue.

mass spectrometer (G2445D SL) with ESI source. ESI-MS positive

and negative spectra of reaction intermediates and ﬁnal puriﬁed

products were obtained from solutions in methanol (MeOH) or

MeCN, eluting with MeOH or MeCN containing 0.1% formic acid

(only for positive mass spectra). ¹H and ¹³C NMR spectra were

recorded with a Bruker AV300 FT-NMR UltraShield operating at

300 MHz (for ¹H NMR) and 75 MHz (for ¹³C NMR). Chemical shifts

3. Materials and methods

3.1. Chemicals and instrumentation

(d) are given in parts per million (ppm) relative to the signal of the

solvent. The following abbreviations are used to indicate multi-

plicities: s, singlet; d, doublet; dd, doublet of doublets; t, triplet; m,

multiplet; br, broad signal.

Pterostilbene was obtained from Wonda Science (North Wal-

tham, MA, USA, cat. n. 25992). Lindlar catalyst was obtained from

Alfa Aesar (Karlsruhe, Germany, cat. n. 43172). Pterostilbene-4’-

sulfate was synthesized as described in Azzolini et al. [34]. HPLC

grade acetonitrile (MeCN) and THF were obtained from VWR and

used without further puriﬁcation. Other reagents and solvents

were purchased from Sigma Aldrich (Milan, Italy), and were used as

received: 1,10-decanediol (cat. n. D1203), periodic acid (cat. n.

375810), pyridinium chlorochromate (PCC; cat. n. 190144), oxalyl

chloride (cat. n. O8801), oleoyl chloride (cat. n. 367850), sodium

azide (cat. n. 199931), bis(4-nitrophenyl) carbonate (cat. n. 161691),

2,4,6-trimethylpyridine (cat. n. 27690), p-toluenesulfonic acid

(PTSA; cat. n. 402885), 2,2-dimethoxypropane (cat. n. D136808),

TRIS hydrochloride (cat. n. T3253), sodium periodate (cat. n.

363642), potassium phosphate monobasic (cat. n. P5655), pyridine

(cat. n. 270970), triethylamine (TEA; cat. n. T0886), 4-(dimethyla-

mino)pyridine (DMAP; cat. n. 107700), lithium aluminum hydride

solution (1.0 M in THF, cat. n. 212776), lipase from porcine pancreas

(Type VI-S, ꢁ20,000 units/mg protein, lyophilized powder, cat. n.

L0382), colipase from porcine pancreas (essentially salt-free,

lyophilized powder, cat. n. C3028). TLCs were run on silica gel

supported on plastic (Macherey-Nagel Polygram^®SIL G/UV254, sil-

ica thickness 0.2 mm) and visualized by UV detection or KMnO₄

oxidation. Flash chromatography was performed on silica gel

3.2. Synthesis of (Z)-3-((10-(((4-((E)-3,5-dimethoxystyryl)

phenoxy)carbonyl) amino)decanoyl)oxy) propane-1,2-diyl dioleate

(Type-1 prodrug)

The synthesis of Type-1 lipid mimetic derivative of pterostilbene

is outlined in Scheme 1. The ﬁrst step in the synthesis was the

esteriﬁcation of a commercially available isopropylidene protected

form of glycerol (5) with 10-chlorodecanoyl chloride (SI-1, see

Supporting Information S3) to obtain the 1-u-chloro substituted

ester intermediate (6). After the cleavage of the acetonide-

protecting group, the resulting glycol (7) was esteriﬁed using

oleoyl chloride to obtain the 1-

intermediate (8). This was treated with sodium azide in DMF to

obtain the corresponding 1- -azide (9), which was then reduced

u-chloro substituted triglyceride

using hydrogen in the presence of Lindlar catalyst and bis(4-

nitrophenyl)carbonate to achieve the lipid mimetic 4-nitrophenyl

activated urethane (10). This was ﬁnally transesteriﬁed with pter-

ostilbene to obtain the desired lipid mimetic Type-1 derivative. The

synthesis proceeds with good to excellent yields in isolated prod-

ucts for every step (see below).

(Macherey-Nagel

60,

230e400

mesh

granulometry

(0.063e0.040 mm)) under air pressure. Mass spectrometry ana-

lyses were performed with a 1100 Series Agilent Technologies

system, equipped with binary pump (G1312A) and MSD SL Trap

3.2.1. (2,2-dimethyl-1,3-dioxolan-4-yl)methyl 10-chlorodecanoate

(6)

10-chlorodecanoyl chloride (SI-1, 1.10 g, 4.89 mmol, 1.0 equiv.) in

Scheme 1. Synthesis of Type-1 lipid mimetic prodrugs of pterostilbene. Reagents and conditions: i) SI-1 (1.0 equiv.), 5 (1.8 equiv.), pyridine (2.8 equiv.), DCM, r.t. overnight, 90%

isolated yield; ii) 6, AcOH/H₂O 4:1 v/v, 50 ^ꢂC for 35 min, 73% isolated yield; iii) oleoyl chloride (2.8 equiv.), 7 (1.0 equiv.), pyridine (3.0 equiv.), DCM, r.t. overnight, 99% isolated yield;

iv) sodium azide (10.0 equiv.), 8 (1.0 equiv.), DMF, 60 ^ꢂC overnight, 96% isolated yield; v) bis(4-nitrophenyl)carbonate (1.5 equiv.), 2,4,6-trimethylpyridine (2.0 equiv.), Lindlar catalyst

(25% by wt.), 9 (1.0 equiv.), H₂, i-PrOH, 0 ^ꢂC for 3 h, 75% isolated yield; vi) pterostilbene (3.0 equiv.), DMAP (3.0 equiv.), 10 (1.0 equiv.), THF, 50 ^ꢂC for 18 h, 43% isolated yield.

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

DCM (5 mL) was added dropwise at 0 ^ꢂC to a solution of DL-1,2-

isopropylideneglycerol (5) (1.18 g, 8.94 mmol, 1.8 equiv.) and pyri-

dine (1.08 g, 13.7 mmol, 2.8 equiv.) in DCM (12 mL), and the

resulting solution was vigorously stirred at room temperature

overnight. The reaction mixture was then diluted with DCM

(50 mL) and washed with HCl 0.5 M (80 mL). The aqueous layer was

extracted with DCM (3 ꢃ 50 mL) and the combined organic layers

were dried over MgSO₄. The solvent was removed under reduced

pressure and the oily residue was puriﬁed by ﬂash chromatography

(DCM/acetone 92:8, R_f¼ 0.70) to afford 1.42 g, 4.41 mmol of 6 (90%

The reaction mixture was then diluted with EtOAc (60 mL) and

washed with brine/H₂O 1:1 (6 ꢃ 30 mL). The organic layer was

dried over MgSO₄, the solvent was removed under reduced pres-

sure and the oily residue was puriﬁed by ﬂash chromatography

(petroleum ether/acetone 95:5, R_f¼ 0.25) to afford 2.30 g,

2.82 mmol of 9 (96% yield) as a slightly yellow oil. ¹H NMR

(300 MHz, CDCl₃)

5.43e5.21 (m, 5H, 2 ꢃ -CH¼CH- and -CH₂-CH-

CH₂-), 4.29 (dd, J ¼ 11.9, 4.2 Hz, 2H, -CH₂-CH-CH₂-), 4.14 (dd,

J ¼ 11.9, 5.9 Hz, 2H, -CH₂-CH-CH₂-), 3.25 (t, J ¼ 6.9 Hz, 2H, -CH₂-N₃),

2.30 (t, J ¼ 7.5 Hz, 6H, 3 ꢃ -CH₂-COO-), 2.07e1.94 (m, 8H, 2 ꢃ -CH₂-

CH¼CH-CH₂-), 1.67e1.53 (m, 8H, -CH₂-CH₂-N₃and 3 ꢃ -CH₂-CH₂-

COO-), 1.38e1.22 (m, 50H, 25 ꢃ -CH₂-), 0.91e0.83 (m, 6H, 2 ꢃ -CH₃).

yield) as a slightly yellow oil. ¹H NMR (300 MHz, CDCl₃)

d 4.35e4.23

(m, 1H, -CH₂-CH-CH₂-), 4.20e4.00 (m, 3H, -CH₂-CH-CH₂-),

3.77e3.67 (m, 1H, -CH₂-CH-CH₂-), 3.51 (t, J ¼ 6.7 Hz, 2H, -CH₂-Cl),

2.33 (t, J ¼ 7.5 Hz, 2H, -CH₂-COO-), 1.82e1.68 (m, 2H, -CH₂-CH₂-Cl),

1.68e1.53 (m, 2H, -CH₂-CH₂-COO-), 1.47e1.19 (m, 16H, 2 ꢃ -CH₃and

¹³C NMR (75 MHz, CDCl₃)

d 173.4, 173.3, 172.9, 130.2, 130.1, 129.8,

129.8, 69.0, 62.2, 51.6, 34.3, 34.2, 34.1, 32.0, 29.9, 29.9, 29.8, 29.7,

29.5, 29.4, 29.3, 29.3, 29.3, 29.2, 29.2, 29.2, 29.0, 27.4, 27.3, 26.8,

25.0, 25.0, 24.9, 22.8, 14.2. ESI-MS (ESIþ): 834 m/z [M þ NH₄]^þ,

839 m/z [MþNa]^þ.

5 ꢃ -CH₂-). ¹³C NMR (75 MHz, CDCl₃)

d 173.7, 109.9, 73.8, 66.5, 64.6,

45.2, 34.2, 32.7, 29.3, 29.2, 29.2, 28.9, 26.9, 26.8, 25.5, 25.0. ESI-MS

(ESIþ): 343 m/z [MþNa]^þ, 359 m/z [MþK]^þ.

3.2.5. (Z)-3-((10-(((4-nitrophenoxy)carbonyl)amino)decanoyl)oxy)

propane-1,2-diyl dioleate (10)

3.2.2. 2,3-dihydroxypropyl 10-chlorodecanoate (7)

6 (1.36 g, 4.22 mmol) was dissolved in AcOH/H₂O 4:1 v/v (12 mL)

and stirred at 50 ^ꢂC for 35 min. The reaction mixture was then

diluted with EtOAc (80 mL) and washed with a saturated solution of

NaHCO₃(adding solution until the production of gaseous CO₂

stopped). The aqueous layer was then extracted with EtOAc

(2 ꢃ 50 mL) and the combined organic layers were dried over

MgSO₄. The solvent was removed under reduced pressure and the

oily residue was puriﬁed by ﬂash chromatography (DCM/acetone

7:3, R_f¼ 0.45) to afford 0.87 g, 3.1 mmol of 7 (73% yield) as a col-

Bis(4-nitrophenyl)carbonate (0.28 g, 0.92 mmol, 1.5 equiv.),

2,4,6-trimethylpyridine (0.15 g, 1.2 mmol, 2.0 equiv.) and Lindlar

Catalyst (0.13 g, 25% by wt.) were added successively to a stirred

solution of 9 (0.50 g, 0.61 mmol, 1.0 equiv.) in i-PrOH (15 mL). The

reaction ﬂask was evacuated and ﬂushed with hydrogen gas. The

resultant mixture was stirred under hydrogen at 0 ^ꢂC for 3 h. After

completion of the reaction, the catalyst was ﬁltered through a pad

of celite and the ﬁlter cake was washed with DCM (125 mL). The

ﬁltrate was washed with HCl 0.5 M (100 mL), the aqueous layer was

extracted with DCM (50 mL) and the combined organic layers were

dried over MgSO₄. The solvent was removed under reduced pres-

sure and the oily residue was puriﬁed by ﬂash chromatography

(petroleum ether/acetone 85:15, R_f¼ 0.33) to afford 0.44 g,

0.46 mmol of 10 (75% yield) as a slightly yellow oil. ¹H NMR

ourless oil. ¹H NMR (300 MHz, CDCl₃)

d 4.23e4.09 (m, 2H, -COO-

CH₂-CH-), 3.96e3.87 (m, 1H, -CH₂-CH-CH₂-), 3.73e3.55 (m, 2H,

-CH-CH₂-OH), 3.52 (t, J ¼ 6.7 Hz, 2H, -CH₂-Cl), 2.44 (br, 2H, 2 ꢃ -OH),

2.34 (t, J ¼ 7.5 Hz, 2H, -CH₂-COO-), 1.82e1.69 (m, 2H, -CH₂-CH₂-Cl),

1.69e1.55 (m, 2H, -CH₂-CH₂-COO-), 1.48e1.21 (m, 10H, 5 ꢃ -CH₂-).

¹³C NMR (75 MHz, CDCl₃)

174.4, 70.4, 65.3, 63.5, 45.3, 34.2, 32.7,

(300 MHz, CDCl₃)

8.22 (d, J ¼ 9.1 Hz, 2H, Ar-H), 7.30 (d, J ¼ 9.1 Hz,

29.3, 29.2, 29.2, 28.9, 26.9, 25.0. ESI-MS (ESIþ): 303 m/z [MþNa]^þ.

2H, Ar-H), 5.40e5.16 (m, 5H, 2 ꢃ -CH¼CH- and -CH₂-CH-CH₂-), 4.29

(dd, J ¼ 11.9, 4.2 Hz, 2H, -CH₂-CH-CH₂-), 4.13 (dd, J ¼ 11.9, 5.9 Hz, 2H,

-CH₂-CH-CH₂-), 3.32e3.20 (m, 2H, -CH₂-NH-), 2.30 (t, J ¼ 7.3 Hz, 6H,

3 ꢃ -CH₂-COO-), 2.07e1.90 (m, 8H, 2 ꢃ -CH₂-CH¼CH-CH₂-),

3.2.3. (Z)-3-((10-chlorodecanoyl)oxy)propane-1,2-diyl dioleate (8)

Oleoyl chloride (purity 90%) (2.49 g, 8.27 mmol, 2.8 equiv.) in

DCM (5 mL) was added dropwise at 0 ^ꢂC to a solution of 7 (0.84 g,

3.0 mmol, 1.0 equiv.) and pyridine (0.70 g, 8.8 mmol, 3.0 equiv.) in

DCM (20 mL) and the resulting solution was vigorously stirred at

room temperature overnight. The reaction mixture was then

diluted with DCM (25 mL) and washed with HCl 0.5 M (50 mL). The

aqueous layer was extracted with DCM (2 ꢃ 25 mL) and the com-

bined organic layers were dried over MgSO₄. The solvent was

removed under reduced pressure and the oily residue was puriﬁed

by ﬂash chromatography (petroleum ether/acetone 95:5, R_f¼ 0.23)

to afford 2.40 g, 2.97 mmol of 8 (99% yield) as a slightly yellow oil.

1.67e1.51 (m, 8H, -CH₂-CH₂-NH- and

ꢃ

-CH₂-CH₂-COO-),

1.42e1.13 (m, 50H, 25 ꢃ -CH₂-), 0.91e0.80 (m, 6H, 2 ꢃ -CH₃). ¹³C

NMR (75 MHz, CDCl₃) d 173.4, 173.3, 172.9, 156.2, 153.2, 144.8, 130.1,

130.1, 129.8, 129.8, 125.2, 122.0, 69.0, 62.2, 41.5, 34.3, 34.1, 34.1, 32.0,

29.9, 29.8, 29.8, 29.6, 29.4, 29.4, 29.3, 29.3, 29.2, 29.2, 29.1, 29.1, 27.3,

27.3, 26.8, 25.0, 24.9, 24.9, 22.8, 14.2. ESI-MS (ESIþ): 973 m/z

[M þ NH₄]^þ, 978 m/z [MþNa]^þ.

3.2.6. (Z) -3-((10-(((4-((E)-3,5-dimethoxystyryl)phenoxy)carbonyl)

amino)decanoyl)oxy) propane-1,2-diyl dioleate (Type-1)

¹H NMR (300 MHz, CDCl₃)

5.42e5.20 (m, 5H, 2 ꢃ -CH¼CH- and

Pterostilbene (0.28 g, 1.1 mmol, 3.0 equiv.) and DMAP (0.13 g,

1.1 mmol, 3.0 equiv.) were added to a solution of 10 (0.35 g,

0.36 mmol, 1.0 equiv.) in HPLC grade THF (8.0 mL) and the resulting

solution was stirred at 50 ^ꢂC for 18 h. The reaction mixture was

diluted with DCM (25 mL) and washed with HCl 0.5 M (50 mL). The

aqueous layer was extracted with DCM (2 ꢃ 25 mL) and the com-

bined organic layers were dried over MgSO₄. The solvent was

removed under reduced pressure and the oily residue was puriﬁed

twice by ﬂash chromatography (1st: petroleum ether/acetone 8:2,

R_f¼ 0.40; 2nd: DCM/acetone 99:1) to afford 0.17 g, 0.16 mmol of

Type-1 (43% yield) as a colourless oil. ¹H NMR (300 MHz, CDCl₃)

-CH₂-CH-CH₂-), 4.29 (dd, J ¼ 11.9, 4.2 Hz, 2H, -CH₂-CH-CH₂-), 4.14

(dd, J ¼ 11.9, 6.0 Hz, 2H, -CH₂-CH-CH₂-), 3.52 (t, J ¼ 6.7 Hz, 2H, -CH₂-

Cl), 2.31 (t, J ¼ 7.5 Hz, 6H, 3 ꢃ -CH₂-COO-), 2.09e1.93 (m, 8H, 2 ꢃ -

CH₂-CH¼CH-CH₂-), 1.82e1.69 (m, 2H, -CH₂-CH₂-Cl), 1.68e1.53 (m,

6H, 3 ꢃ -CH₂-CH₂-COO-), 1.49e1.17 (m, 50H, 25 ꢃ -CH₂-), 0.92e0.83

(m, 6H, 2 ꢃ -CH₃). ¹³C NMR (75 MHz, CDCl₃)

d 173.4, 173.4, 173.0,

130.2, 130.1, 129.8, 129.8, 69.0, 62.2, 45.3, 34.3, 34.2, 34.1, 32.8, 32.0,

29.9, 29.8, 29.7, 29.5, 29.4, 29.3, 29.3, 29.3, 29.3, 29.2, 29.2, 29.0,

27.4, 27.3, 27.0, 25.0, 25.0, 24.9, 22.8, 14.3. ESI-MS (ESIþ): 827 m/z

[M þ NH₄]^þ, 832 m/z [MþNa]^þ, 848 m/z [MþK]^þ.

7.47 (d, J ¼ 8.4 Hz, 2H, H-2⁰and H-6⁰), 7.16e7.01 (m, 3H, Ar-

3.2.4. (Z)-3-((10-azidodecanoyl)oxy)propane-1,2-diyl dioleate (9)

Sodium azide (1.90 g, 29.3 mmol, 10.0 equiv.) was added to a

solution of 8 (2.37 g, 2.93 mmol, 1.0 equiv.) in anhydrous DMF

(20 mL) and the resulting solution was stirred at 60 ^ꢂC overnight.

CH¼CH-Ar, H-3⁰and H-5⁰), 6.96 (d, J ¼ 16.2 Hz, 1H, Ar-CH¼CH-Ar),

6.66 (m, 2H, H-2 and H-6), 6.41e6.37 (m, 1H, H-4), 5.41e5.22 (m,

5H, 2 ꢃ -CH¼CH- and -CH₂-CH-CH₂-), 5.12e5.04 (m, 1H, -NH-), 4.30

(dd, J ¼ 11.9, 4.2 Hz, 2H, -CH₂-CH-CH₂-), 4.14 (dd, J ¼ 11.9, 5.9 Hz, 2H,

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

-CH₂-CH-CH₂-), 3.82 (s, 6H, 2 ꢃ -OCH₃), 3.30e3.20 (m, 2H, -CH₂-

NH-), 2.36e2.26 (m, 6H, 3 ꢃ -CH₂-COO-), 2.07e1.94 (m, 8H, 2 ꢃ -

CH₂-CH¼CH-CH₂-), 1.68e1.51 (m, 8H, -CH₂-CH₂-NH- and 3 ꢃ -CH₂-

CH₂-COO-), 1.40e1.18 (m, 50H, 25 ꢃ -CH₂-), 0.92e0.83 (m, 6H, 2 ꢃ -

diluted with DCM (25 mL) and washed with HCl 0.5 M (40 mL). The

aqueous layer was extracted with DCM (3 ꢃ 25 mL) and the com-

bined organic layers were dried over MgSO₄. The solvent was

removed under reduced pressure and the oily residue was puriﬁed

by ﬂash chromatography (DCM/acetone 92:8, R_f¼ 0.72) to afford

1.47 g, 4.57 mmol of 11 (93% yield) as a slightly yellow oil. ¹H NMR

CH₃). ¹³C NMR (75 MHz, CDCl₃)

d 173.4, 173.3, 172.9, 161.1, 154.5,

150.8, 139.4, 134.4, 130.1, 130.1, 129.8, 129.8, 128.6, 128.5, 127.5,

121.9, 104.7, 100.1, 69.0, 62.2, 55.5, 41.4, 34.3, 34.1, 34.1, 32.0, 29.9,

29.8, 29.8, 29.6, 29.4, 29.3, 29.3, 29.2, 29.2, 29.2, 29.2, 27.3, 27.3,

26.8, 25.0, 25.0, 24.9, 22.8, 14.2. ESI-MS (ESIþ): 1073 m/z [M þ H]^þ,

1090 m/z [M þ NH₄]^þ, 1095 m/z [MþNa]^þ, 1111 m/z [M þ K]^þ. Purity

ꢁ95% (HPLC).

(300 MHz, CDCl₃)

d 4.73e4.64 (m, 1H, -CH₂-CH-CH₂-), 4.07 (dd,

J ¼ 12.9, 3.3 Hz, 2H, -CH₂-CH-CH₂-), 3.78 (dd, J ¼ 13.0, 3.5 Hz, 2H,

-CH₂-CH-CH₂-), 3.50 (t, J ¼ 6.7 Hz, 2H, -CH₂-Cl), 2.35 (t, J ¼ 7.5 Hz,

2H, -CH₂-COO-), 1.80e1.68 (m, 2H, -CH₂-CH₂-Cl), 1.68e1.53 (m, 2H,

-CH₂-CH₂-COO-), 1.47e1.20 (m, 16H, 2 ꢃ -CH₃and 5 ꢃ -CH₂-). ¹³

NMR (75 MHz, CDCl₃) d 173.6, 98.5, 66.0, 62.6, 62.1, 45.2, 34.4, 32.7,

29.3, 29.2, 29.1, 28.9, 26.9, 26.2, 25.0, 21.1. ESI-MS (ESIþ): 343 m/z

3.3. Synthesis of (Z)-2-((10-(((4-((E)-3,5-dimethoxystyryl)

phenoxy)carbonyl) amino)decanoyl)oxy) propane-1,3-diyl dioleate

(Type-2 prodrug)

[M þ Na]^þ.

3.3.2. 1,3-dihydroxypropan-2-yl 10-chlorodecanoate (12)

The synthetic strategy adopted to synthesize the Type-2 lipid

mimetic derivative of pterostilbene is similar to that developed for

the Type-1 prodrug, but includes additional initial passages to

synthesize the non-commercially available 1,3-isopropylidene

protected glycerol (SI-2, See Supporting Information S4), which

was obtained as described by Forbes et al. [52]. The subsequent

passages in the synthesis are analogous to those described above

for Type-1 prodrug (see Scheme 2).

11 (1.42 g, 4.42 mmol) was dissolved in AcOH/H₂O 4:1 v/v

(13 mL) and stirred at 50 ^ꢂC for 35 min. The reaction mixture was

then diluted with EtOAc (80 mL) and washed with a saturated so-

lution of NaHCO₃(adding the solution until the production of

gaseous CO₂stopped). The aqueous layer was then extracted with

EtOAc (2 ꢃ 50 mL) and the combined organic layers were dried over

MgSO₄. The solvent was removed under reduced pressure and the

oily residue was puriﬁed by ﬂash chromatography (DCM/acetone

7:3, R_f¼ 0.43) to afford 1.08 g, 3.84 mmol of 12 (87% yield) as a

3.3.1. 2,2-dimethyl-1,3-dioxan-5-yl 10-chlorodecanoate (11)

colourless oil. ¹H NMR (300 MHz, CDCl₃)

d 4.94e4.84 (m, 1H, -CH₂-

10-chlorodecanoyl chloride (SI-1, 1.33 g, 5.89 mmol, 1.2 equiv.)

(see Supporting Information for its synthetic procedure) in DCM

(5 mL) was added dropwise at 0 ^ꢂC to a solution of SI-2 (0.65 g,

4.9 mmol, 1.0 equiv.) and pyridine (0.59 g, 7.4 mmol, 1.5 equiv.) in

DCM (10 mL), and the resulting solution was vigorously stirred at

room temperature overnight. The reaction mixture was then

CH-CH₂-), 3.85e3.73 (m, 4H, -CH₂-CH-CH₂-), 3.51 (t, J ¼ 6.7 Hz, 2H,

-CH₂-Cl), 2.67 (br, 2H, 2 ꢃ -OH), 2.35 (t, J ¼ 7.3 Hz, 2H, -CH₂-COO-),

1.80e1.68 (m, 2H, -CH₂-CH₂-Cl),1.68e1.54 (m, 2H, -CH₂-CH₂-COO-),

1.47e1.22 (m, 10H, 5 ꢃ -CH₂-). ¹³C NMR (75 MHz, CDCl₃)

d 174.2,

75.0, 62.3, 45.2, 34.4, 32.7, 29.3, 29.2, 29.1, 28.9, 26.9, 25.0. ESI-MS

(ESIþ): 303 m/z [M þ Na]^þ.

Scheme 2. Synthesis of Type-2 lipid mimetic prodrugs of pterostilbene. Reagents and conditions: i) SI-1 (1.2 equiv.), SI-2 (1.0 equiv.), pyridine (1.5 equiv.), DCM, r.t. overnight, 93%

isolated yield; ii) 11, AcOH/H₂O 4:1 v/v, 50 ^ꢂC for 35 min, 87% isolated yield; iii) oleoyl chloride (2.5 equiv.), 12 (1.0 equiv.), pyridine (3.1 equiv.), DCM, r.t. overnight, 98% isolated

yield; iv) sodium azide (10.0 equiv.), 13 (1.0 equiv.), DMF, 60 ^ꢂC overnight, 97% isolated yield; v) bis(4-nitrophenyl)carbonate (1.5 equiv.), 2,4,6-trimethylpyridine (2.0 equiv.), Lindlar

catalyst (25% by wt.), 14 (1.0 equiv.), H₂, i-PrOH, 0 ^ꢂC for 3 h, 77% isolated yield; vi) pterostilbene (3.0 equiv.), DMAP (3.0 equiv.), 15 (1.0 equiv.), THF, 50 ^ꢂC for 18 h, 46% isolated yield.

A. Mattarei et al. / European Journal of Medicinal Chemistry 135 (2017) 77e88

3.3.3. (Z)-2-((10-chlorodecanoyl)oxy)propane-1,3-diyl dioleate

(13)

-CH₂-CH-CH₂-), 3.32e3.21 (m, 2H, -CH₂-NH-), 2.30 (t, J ¼ 7.5 Hz, 6H,

3 ꢃ -CH₂-COO-), 2.08e1.91 (m, 8H, 2 ꢃ -CH₂-CH¼CH-CH₂-),

Oleoyl chloride (purity 90%) (2.99 g, 8.94 mmol, 2.5 equiv.) in

DCM (6 mL) was added dropwise at 0 ^ꢂC to a solution of 12 (1.00 g,

3.57 mmol, 1.0 equiv.) and pyridine (0.85 g, 11 mmol, 3.1 equiv.) in

DCM (25 mL) and the resulting solution was vigorously stirred at

room temperature overnight. The reaction mixture was then

diluted with DCM (30 mL) and washed with HCl 0.5 M (60 mL). The

aqueous layer was extracted with DCM (2 ꢃ 30 mL) and the com-

bined organic layers were dried over MgSO₄. The solvent was

removed under reduced pressure and the oily residue was puriﬁed

by ﬂash chromatography (petroleum ether/acetone 95:5, R_f¼ 0.21)

to afford 2.85 g, 3.52 mmol of 13 (98% yield) as a slightly yellow oil.

1.68e1.50 (m, 8H, -CH₂-CH₂-NH- and

ꢃ

-CH₂-CH₂-COO-),

1.42e1.17 (m, 50H, 25 ꢃ -CH₂-), 0.93e0.80 (m, 6H, 2 ꢃ -CH₃). ¹³C

NMR (75 MHz, CDCl₃) d 173.3, 172.9, 156.1, 153.2, 144.8, 130.1, 129.8,

125.2, 122.0, 69.0, 62.2, 41.5, 34.2, 34.1, 32.0, 29.9, 29.8, 29.8, 29.6,

29.4, 29.4, 29.3, 29.3, 29.2, 29.2, 29.1, 27.3, 27.3, 26.8, 24.9, 24.9, 22.8,

14.2. ESI-MS (ESIþ): 973 m/z [M þ NH₄]^þ, 978 m/z [M þ Na]^þ.

3.3.6. (Z) -2-((10-(((4-((E)-3,5-dimethoxystyryl)phenoxy)carbonyl)

amino)decano yl)oxy) propane-1,3-diyl dioleate (Type-2)

Pterostilbene (0.30 g, 1.2 mmol, 3.0 equiv.) and DMAP (0.14 g,

1.2 mmol, 3.0 equiv.) were added to a solution of 15 (0.37 g,

0.39 mmol, 1.0 equiv.) in HPLC-grade THF (8.0 mL) and the resulting

solution was stirred at 50 ^ꢂC for 18 h. The reaction mixture was

diluted with DCM (25 mL) and washed with HCl 0.5 M (50 mL). The

aqueous layer was extracted with DCM (2 ꢃ 25 mL) and the com-

bined organic layers were dried over MgSO₄. The solvent was

removed under reduced pressure and the oily residue was puriﬁed

two times by ﬂash chromatography (1st: petroleum ether/acetone

8:2, R_f¼ 0.39; 2nd: DCM/acetone 99:1) to afford 0.19 g, 0.18 mmol

of Type-2 (46% yield) as a white solid. ¹H NMR (300 MHz, CDCl₃)

¹H NMR (300 MHz, CDCl₃)

5.42e5.20 (m, 5H, 2 ꢃ -CH¼CH- and

-CH₂-CH-CH₂-), 4.29 (dd, J ¼ 11.9, 4.2 Hz, 2H, -CH₂-CH-CH₂-), 4.14

(dd, J ¼ 11.9, 5.9 Hz, 2H, -CH₂-CH-CH₂-), 3.52 (t, J ¼ 6.7 Hz, 2H, -CH₂-

Cl), 2.31 (t, J ¼ 7.5 Hz, 6H, 3 ꢃ -CH₂-COO-), 2.09e1.91 (m, 8H, 2 ꢃ -

CH₂-CH¼CH-CH₂-), 1.83e1.69 (m, 2H, -CH₂-CH₂-Cl), 1.68e1.53 (m,

6H, 3 ꢃ -CH₂-CH₂-COO-), 1.49e1.16 (m, 50H, 25 ꢃ -CH₂-), 0.88 (t,

J ¼ 6.3 Hz, 6H, 2 ꢃ -CH₃). ¹³C NMR (75 MHz, CDCl₃)

d 173.4, 172.9,

130.1, 129.8, 69.0, 62.2, 45.2, 34.3, 34.2, 32.7, 32.0, 29.9, 29.8, 29.7,

29.5, 29.4, 29.3, 29.3, 29.2, 29.2, 29.2, 29.1, 29.0, 27.4, 27.3, 27.0, 25.0,

22.8, 14.3. ESI-MS (ESIþ): 827 m/z [M þ NH₄]^þ, 832 m/z [M þ Na]^þ,

848 m/z [M þ K]^þ.

7.47 (d, J ¼ 8.5 Hz, 2H, H-2⁰and H-6⁰), 7.17e7.01 (m, 3H, Ar-

CH¼CH-Ar, H-3⁰and H-5⁰), 6.96 (d, J ¼ 16.2 Hz, 1H, Ar-CH¼CH-Ar),

6.66 (d, J ¼ 2.1 Hz, 2H, H-2 and H-6), 6.39 (t, J ¼ 2.0 Hz, 1H, H-4),

5.42e5.21 (m, 5H, 2 ꢃ -CH¼CH- and -CH₂-CH-CH₂-), 5.11e5.01 (m,

1H, -NH-), 4.30 (dd, J ¼ 11.9, 4.3 Hz, 2H, -CH₂-CH-CH₂-), 4.15 (dd,

J ¼ 11.9, 5.9 Hz, 2H, -CH₂-CH-CH₂-), 3.82 (s, 6H, 2 ꢃ -OCH₃),

3.33e3.19 (m, 2H, -CH₂-NH-), 2.37e2.24 (m, 6H, 3 ꢃ -CH₂-COO-),

2.09e1.91 (m, 8H, 2 ꢃ -CH₂-CH¼CH-CH₂-), 1.68e1.50 (m, 8H, -CH₂-

CH₂-NH- and 3 ꢃ -CH₂-CH₂-COO-), 1.38e1.21 (m, 50H, 25 ꢃ -CH₂-),

3.3.4. (Z)-2-((10-azidodecanoyl)oxy)propane-1,3-diyl dioleate (14)

Sodium azide (2.24 g, 34.4 mmol, 10.0 equiv.) was added to a

solution of 13 (2.79 g, 3.44 mmol,1.0 equiv.) in dry DMF (25 mL) and

the resulting solution was stirred at 60 ^ꢂC overnight. The reaction

mixture was then diluted with EtOAc (70 mL) and washed with

brine/H₂O 1:1 (6 ꢃ 35 mL). The organic layer was dried over MgSO₄,

the solvent was removed under reduced pressure and the oily

residue was puriﬁed by ﬂash chromatography (petroleum ether/

acetone 95:5, R_f¼ 0.25) to afford 2.73 g, 3.34 mmol of 14 (97% yield)

0.91e0.83 (m, 6H, 2 ꢃ -CH₃). ¹³C NMR (75 MHz, CDCl₃)

d 173.4,

172.9,161.1,154.6,150.8,139.4,134.4,130.1, 129.8,128.7,128.5,127.5,

121.9, 104.7, 100.1, 69.0, 62.2, 55.5, 41.4, 34.3, 34.1, 32.0, 29.9, 29.8,

29.6, 29.4, 29.4, 29.3, 29.3, 29.2, 29.2, 29.1, 27.3, 27.3, 26.8, 25.0,

22.8, 14.2. ESI-MS (ESIþ): 1073 m/z [M þ H]^þ, 1090 m/z [M þ NH₄]^þ,

1095 m/z [M þ Na]^þ, 1111 m/z [M þ K]^þ. Purity ꢁ95% (HPLC).

as a slightly yellow oil. ¹H NMR (300 MHz, CDCl₃)

d 5.45e5.20 (m,

5H, 2 ꢃ -CH¼CH- and -CH₂-CH-CH₂-), 4.29 (dd, J ¼ 11.9, 4.3 Hz, 2H,

-CH₂-CH-CH₂-), 4.14 (dd, J ¼ 11.9, 5.9 Hz, 2H, -CH₂-CH-CH₂-), 3.25 (t,

J ¼ 6.9 Hz, 2H, -CH₂-N₃), 2.36e2.26 (m, 6H, 3 ꢃ -CH₂-COO-),

2.07e1.95 (m, 8H, 2 ꢃ -CH₂-CH¼CH-CH₂-), 1.67e1.52 (m, 8H, -CH₂-

CH₂-N₃and 3 ꢃ -CH₂-CH₂-COO-), 1.38e1.20 (m, 50H, 25 ꢃ -CH₂-),

3.4. HPLC-UV analysis

0.87 (t, J ¼ 6.6 Hz, 6H, 2 ꢃ -CH₃). ¹³C NMR (75 MHz, CDCl₃)

173.3,

Samples (2 mL) were analyzed by HPLC-UV (1290 Inﬁnity LC

172.9, 130.1, 129.8, 69.0, 62.2, 51.6, 34.3, 34.2, 32.0, 29.9, 29.8, 29.7,

29.5, 29.4, 29.3, 29.3, 29.2, 29.2, 29.2, 29.1, 29.0, 27.4, 27.3, 26.8,

25.0, 22.8, 14.2. ESI-MS (ESIþ): 834 m/z [M þ NH₄]^þ, 839 m/z

[M þ Na]^þ.

System, Agilent Technologies) using a reverse phase column and a

UV diode array detector (190e500 nm). Type-1, Type-2 prodrugs

and their hydrolysis products were analyzed with a Zorbax Extend-

C18 column (Rapid Resolution HT, 1.8

m, 50 ꢃ 3.0 mm i.d.; Agilent

Technologies, cat. n. 727975-302); solvents A and B were water

containing 0.1% triﬂuoroacetic acid (TFA) and THF/MeCN 7:3,

respectively. The gradient for B was as follows: 30% for 0.5 min,

then from 30% to 100% in 6.5 min, 100% for 3 min; the ﬂow rate was

0.5 mL/min and column compartment was maintained at 50 ^ꢂC.

Pterostilbene and its phase II metabolites were analyzed with a

3.3.5. (Z)-2-((10-(((4-nitrophenoxy)carbonyl)amino)decanoyl)oxy)

propane-1,3-diyl dioleate (15)

Bis(4-nitrophenyl)carbonate (0.28 g, 0.92 mmol, 1.5 equiv.),

2,4,6-trimethylpyridine (0.15 g, 1.2 mmol, 2.0 equiv.) and Lindlar

catalyst (0.13 g, 25% by wt.) were added successively to a stirred

solution of 14 (0.51 g, 0.62 mmol, 1.0 equiv.) in i-PrOH (15 mL). The