ApoA-I mimetic peptides promote pre-β HDL formation in vivo causing remodeling of HDL and triglyceride accumulation at higher dose

Article abstract of DOI:10.1016/j.bmc.2010.09.074

Reverse cholesterol transport promoted by HDL-apoA-I is an important mechanism of protection against atherosclerosis. We have previously identified apoA-I mimetic peptides by synthesizing analogs of the 22 amino acid apoA-I consensus sequence (apoA-I_cons) containing non-natural aliphatic amino acids. Here we examined the effect of different aliphatic non-natural amino acids on the structure-activity relationship (SAR) of apoA-I mimetic peptides. These novel apoA-I mimetics, with long hydrocarbon chain (C _5-8) amino acids incorporated in the amphipathic α helix of the apoA-I_cons, have the following properties: (i) they stimulate in vitro cholesterol efflux from macrophages via ABCA1; (ii) they associate with HDL and cause formation of pre-β HDL particles when incubated with human and mouse plasma; (iii) they associate with HDL and induce pre-β HDL formation in vivo, with a corresponding increase in ABCA1-dependent cholesterol efflux capacity ex vivo; (iv) at high dose they associate with VLDL and induce hypertriglyceridemia in mice. These results suggest our peptide design confers activities that are potentially anti-atherogenic. However a dosing regimen which maximizes their therapeutic properties while minimizing adverse effects needs to be established.

Full text of DOI:10.1016/j.bmc.2010.09.074

Bioorganic & Medicinal Chemistry 18 (2010) 8669–8678
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
ApoA-I mimetic peptides promote pre-b HDL formation in vivo
causing remodeling of HDL and triglyceride accumulation at higher dose
Ester Carballo-Jane a, , Zhu Chen a, Edward O’Neill a, Jun Wang a, Charlotte Burton a, Ching H. Chang a,
⇑
Xun Chen a, Suzanne Eveland a, Betsy Frantz-Wattley a, Karen Gagen a, Brian Hubbard a, Marina Ichetovkin a,
Silvi Luell a, Roger Meurer a, Xuelei Song a, Alison Strack a, Annunziata Langella b, Simona Cianetti b,
Francesca Rech b, Elena Capitò b, Simone Bufali b, Maria Veneziano b, Maria Verdirame b, Fabio Bonelli b,
Edith Monteagudo b, Antonello Pessi b, Raffaele Ingenito b, Elisabetta Bianchi b,
⇑
a Merck Research Laboratories, 126 East Lincoln Avenue, RY80Y-140, Rahway, NJ 07065, USA
b Merck Research Laboratories, Istituto di Ricerche di Biologia Molecolare P. Angeletti, 00040 Pomezia, Rome, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 28 July 2010
Revised 20 September 2010
Accepted 21 September 2010
Available online 4 November 2010
Reverse cholesterol transport promoted by HDL-apoA-I is an important mechanism of protection against
atherosclerosis. We have previously identified apoA-I mimetic peptides by synthesizing analogs of the 22
amino acid apoA-I consensus sequence (apoA-Icons) containing non-natural aliphatic amino acids. Here
we examined the effect of different aliphatic non-natural amino acids on the structure–activity relation-
ship (SAR) of apoA-I mimetic peptides. These novel apoA-I mimetics, with long hydrocarbon chain (C5–8
)
amino acids incorporated in the amphipathic helix of the apoA-Icons, have the following properties: (i)
a
Keywords:
Peptide mimetics
ApoA-I
HDL (high density lipoproteins)
Pre-b HDL
Atherosclerosis
they stimulate in vitro cholesterol efflux from macrophages via ABCA1; (ii) they associate with HDL and
cause formation of pre-b HDL particles when incubated with human and mouse plasma; (iii) they asso-
ciate with HDL and induce pre-b HDL formation in vivo, with a corresponding increase in ABCA1-depen-
dent cholesterol efflux capacity ex vivo; (iv) at high dose they associate with VLDL and induce
hypertriglyceridemia in mice. These results suggest our peptide design confers activities that are poten-
tially anti-atherogenic. However a dosing regimen which maximizes their therapeutic properties while
minimizing adverse effects needs to be established.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
including infusion of apoA-I itself, and pharmacological agents that
upregulate apoA-I expression.7,8 Here we focus on another ap-
Many factors may contribute to the ability of HDL, or its main
protein constituent, apoA-I, to protect against atherosclerosis.
One of the primary mechanisms of atheroprotection is believed
to be reverse cholesterol transport (RCT), the process of cholesterol
efflux from arterial macrophages to HDL resulting in delivery of ex-
cess cholesterol to the liver for excretion in bile and ultimate elim-
ination in feces.1,2
proach, the development of apoA-I mimetic peptides that recapit-
ulate certain pharmacological features of apoA-I while being easier
and less expensive to manufacture.
Despite the importance of ABCA1 in nascent HDL biogenesis and
atheroprotection, relatively little is known about the structural ele-
ments of apoA-I that mediate ABCA1-dependent cholesterol efflux.
ApoA-I is a 243 amino acid protein containing multiple
a-helical
RCT is a regulated process mediated by specific transporters.3 In
particular, the ATP-binding cassette transporter A1 (ABCA1) is be-
lieved to initiate the process, by mediating cholesterol efflux from
macrophages to lipid-free or lipid-poor apoA-I,4 generating nascent
HDL particles. Many studies have documented the anti-athero-
genic effects of apoA-I in animal models.5,6 These observations
highlight the therapeutic potential of apoA-I-based treatments,
repeats separated by proline residues. Most of the repeats are 22
amino acids long with a unique secondary structure defined as
class A amphipathic a-helix. In this helix, negatively charged ami-
no acid residues are clustered at the center of the polar face and
positively charged residues are at the interface between the hydro-
philic and hydrophobic faces. The apoA-I segments that have been
implicated in mediating cellular lipid efflux are those that possess
the highest lipid binding affinity.9,10
Although individual 22-mer helices of apoA-I are incapable of
stimulating ABCA1-dependent cholesterol efflux, longer peptides
consisting of more than one repeat are able to promote cholesterol
efflux.9 In addition, a variety of peptides unrelated to apoA-1 but
⇑
Corresponding authors. Tel.: +1 732 594 8489 (E.C.-J.); tel.: +39 06 91093434
(E.B.).
E-mail addresses: ester_carballojane@merck.com (E. Carballo-Jane), e.bian-
chi@irbm.it (E. Bianchi).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.09.074

8670
E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
Table 1
with amphipathic helical characteristics have been shown to pro-
mote cholesterol efflux from cells and to mimic additional activi-
ties of apoA-I, such as anti-oxidation and anti-inflammation.11–14
In particular, the apoA-I mimetic peptide D4F was reported to have
anti-atherogenic effects in animal models.15,16
Single-point mutant sequence of apoA-Icons
ApoA -Icons
analogsa
Sequence
ApoA -Icons
X-2
X-3
X-6
X-9
X-10
X-13
X-14
X-17
X-20
X-21
PVLDEFREKLNEELEALKQKLK
PXLDEFREKLNEELEALKQKLK
PVXDEFREKLNEELEALKQKLK
PVLDEXREKLNEELEALKQKLK
PVLDEFREXLNEELEALKQKLK
PVLDEFREKXNEELEALKQKLK
PVLDEFREKLNEXLEALKQKLK
PVLDEFREKLNEEXEALKQKLK
PVLDEFREKLNEELEAXKQKLK
PVLDEFREKLNEELEALKQXLK
PVLDEFREKLNEELEALKQKXK
We have previously discovered that starting from an inactive
22-mer consensus sequence of the tandem repeats of apoA-I
17
(apoA-Icons
)
it was possible to design apoA-I mimetic peptides
that could: i) mediate cholesterol efflux via ABCA1 from macro-
phages in vitro, and ii) cause an increase in pre-b HDL formation
in vitro when incubated with both human and mouse plasma.18
Key to the activity of these peptides was increased lipophilicity
of the peptides achieved by introducing alkenyl long hydrocarbon
chain amino acids into the sequence, resulting in the formation of
an all-hydrocarbon cycle by ring closing metathesis. The observa-
tion that the introduction of a few alkenyl amino acids per se influ-
enced the activity of apoA-I mimetic peptides prompted us to
further explore the potential of introducing a new set of aliphatic
a
X represents the amino acid substitution at num-
bered position of apoA-Icons
.
non-natural amino acids on the hydrophobic face of apoA-Icons
Here we describe the detailed structure–activity relationship
(SAR) of novel peptides containing long hydrocarbon chain (C5–8
amino acids incorporated at all positions of the hydrophobic por-
tion of the class A amphipathic -helix of apoA-Icons. We started
.
)
a
by synthesizing a series of novel non-natural amino acids that
had both alkenyl and alkanyl side chains of different configuration.
We then carried out SAR studies on peptide analogs containing
these non-natural amino acids. This led to the identification of spe-
cific positions at which introduction of a long hydrocarbon chain
confers on the 22-mer peptide pharmacological features character-
istic of apoA-I (Fig. 1). We found that novel apoA-I mimetic pep-
tides which can stimulate in vitro cholesterol efflux from
macrophages via ABCA1 also induce pre-b HDL formation in hu-
man and mouse plasma in vitro, trigger HDL remodeling and pre-
b HDL formation upon administration into animals (with a con-
comitant increase in serum ABCA1-dependent cholesterol efflux
Figure 2. Panel of non-natural amino acids with long hydrocarbon side chains
introduced for the SAR of apoA-Icons peptide.
capacity ex vivo), and, at high concentration, incorporate into VLDL
and induce hypertriglyceridemia in mice.
2. Results and discussion
2.1. Design and synthesis of apoA-Icons peptide analogs
Figure 1. Helical wheel representation and peptide sequence of the 22-mer
consensus peptide of the tandem repeats of apoA-I, apoA-Icons. The black bar
highlights the interface between the hydrophilic and hydrophobic faces. High-
lighted with a green or red ball are positions were aliphatic non-natural amino
acids were introduced. In red are the positions identified by the SAR as most
suitable for gaining activity in the cholesterol efflux assay.
Figure 1 shows the sequence and helical wheel representation
of the 22 amino acid consensus sequence, apoA-Icons, with its un-
ique secondary structural element defined as a class A amphipathic
a-helix. In the helix, negatively charged amino acid residues are
clustered at the center of the polar face and positively charged

E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
8671
residues are at the interface between the hydrophilic and hydro-
phobic faces. On the helical wheel of apoA-Icons, the hydrophobic
face comprises residues numbered on the sequence as 2, 3, 6, 9,
10, 13, 14, 17, 20, and 21. We synthesized analogs of apoA-Icons
with hydrophobic non-natural amino acids at each of these posi-
tions following the scheme shown in Table 1.
The non-natural amino acids were coupled manually using HBTU
as activator until a negative ninhydrin test was observed. All the
peptides were purified by reverse-phase chromatography at purity
above 95%, as determined by analytical HPLC chromatography.
After assessing the cholesterol efflux activities of each of the
single substitution analogs, the substitutions in the most active
analogs were combined in analogs of apoA-Icons with multiple sub-
stitutions (Table 3).
The non-natural amino acid substitutions have long carbon
(C5–8) alkenyl or alkanyl side chains, in both R or S configuration,
as shown in Figure 2. Our goal was to stabilize the class A a-helical
motif and modulate lipophilicity of the amphipathic helix. Amino
acids Fmoc-R5H, Fmoc-R6H and Fmoc-S7H were synthesized as pre-
viously described,18 while the synthesis of alkenyl amino acids
Fmoc-S6H, Fmoc-R7H, Fmoc-S8H, and alkanyl amino acids Fmoc-
S7H3 and Fmoc-S8H3 were described in Section 4.
2.2. Structure–activity relationship in an in vitro cholesterol
efflux assay
ApoA-Icons analog peptides with single substitutions were
tested for their ability to promote cholesterol efflux from macro-
phages in an in vitro assay (Table 2). As a negative control, we used
the apoA-Icons peptide (compound 1) which had been shown to be
inactive in the cholesterol efflux assay.18 As positive controls, we
used the HDL peptide mimetics, D4F (2) and 5A (3) previously re-
ported by others to induce cholesterol efflux.11,12 In our assay D4F
The 22-mer peptide apoA-Icons analogs were synthesized by so-
lid phase using Fmoc/tBu chemistry on a peptide synthesizer using
2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium
hexa-
fluorophosphate (HBTU) as coupling reagent and a Fmoc-Rink
AM-PS based resin. All the experimental details are available in
Section 4.3.
(2) and 5A (3) exhibited an EC50 of
3.4 1.06 M, respectively. As a reference we used purified hu-
man apoA-I protein (4), which had an EC50 of 0.1 M. We first
2
l
M
0.25 lM and
l
M
l
l
tested the apoA-Icons analogs (5–14) in which a single C6 carbon
chain amino acid, S6H, of S configuration (Fig. 2), was introduced
at position 2, 3, 6, 9, 10, 13, 14, 17, 20, or 21, following the scheme
in Table 1. These positions are all located on the hydrophobic face
of the peptide helical sequence (Fig. 1). Of these analogs, only three
exhibited any cholesterol efflux activity: analog S6H-13 (10), with
Table 2
In vitro cholesterol efflux from macrophages
Compd No.
Name
Chol. efflux (EC50
>100
, l
M)a
1
2
3
4
5
6
7
8
9
ApoA-Icons
D4F
5A
h-apoA-I protein
S6H-2
S6H-3
2
0.25
3.4 1.06
0.1 0.01
>100*
>100
>100
>100
>100
1.4
>100
>100
the substitution at position 13, had an EC50 of 1.4
l
M. Analog
S6H-20 (13), with the substitution at position 20, had an EC50 of
213
lM, while analog S6H-2 (5), with the substitution at position
S6H-6
S6H-9
2, exhibited weak activity only at 100
lM, the highest concentra-
tion used. We then tested the apoA-Icons analogs (15–24) in which
a single C6 carbon chain amino acid, R6H, (Fig. 2) having an oppo-
site configuration with respect to S6H, was introduced at hydropho-
bic positions. Analog R6H-13 (20), with the substitution at position
S6H-10
S6H-13
S6H-14
S6H-17
S6H-20
S6H-21
R6H-2
R6H-3
R6H-6
R6H-9
R6H-10
R6H-13
R6H-14
R6H-17
R6H-20
R6H-21
R5H-6
R5H-13
S7H-6
S7H-9
S7H-10
S7H-13
S7H-14
S7H-17
S7H-20
R7H-2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
213
13, had an EC50 of about 31
l
M, while analogs R6H-2 (15) and R6H
-
>100
20 (23) exhibited weak activity at 100
lM. Thus, introduction of
>100*
>100
both S6H and R6H at position 13 confers relatively potent choles-
terol efflux activity, although the S configuration seems to be pre-
ferred. If a shorter amino acid side chain (C5) was used, R5H-2 (25)
and R5H-13 (26), no activity was observed in the cholesterol efflux
assay.
Substitution by amino acids with a 7 carbon chain, S7H, (analogs
27–33) confirmed the observation that introduction of an aliphatic
amino acid at position 13 confers potent cholesterol efflux capacity
on the inactive peptide. Analog S7H-13, (30), exhibited activity in
the cholesterol efflux assay, with potency in a range similar to that
>100
>100
>100
31 18
>100
>100
>100*
>100
>100
>100*
>100*
>100
>100
2.4 0.6
>100
Table 3
In vitro cholesterol efflux from macrophages of apoA-Icons mutants obtained by
introduction of two or more aliphatic non-natural amino acids at specific positions of
the sequence
>100
>100*
>100*
>100
Compd No. (L-no)
Name
Chol. efflux (EC50, lM)
R7H-6
R7H-9
>100
>100
9.5 3.3
>100
>100
>100
>100
0.7 0.1
0.7 0.3
0.7 0.2
46
47
48
49
50
51
52
53
54
55
56
57
58
R6H-6/S7H-13
S7H-2/S7H-13
S7H-6/S7H-13
S7H-13/S7H-20
S7H3-6/S7H3-13
S8H-6/S8H-13
S8H3-6/S8H3-13
S7H-6/S7H-13, S7H-20
S7H-2/S7H-13/S7H-20
S7H-2/S7H-6/S7H-13
R7H-6/S7H-13/S7H-20
S7H-2/R7H-6/S7H-13
S7H-2/S7H-6/S7H-13/S7H-20
0.8 0.2
0.4 0.2
0.4 0.1
0.9 0.3
0.3 0.2
0.4 0.3
0.6 0.2
1.8 0.3
R7H-10
R7H-13
R7H-14
R7H-17
R7H-20
R7H-21
S7H3-13
S8H-13
S8H3-13
5
2
1.9 0.9
1.0 0.5
0.9 0.2
6.7 4.6
The activity is gained by introduction of one aliphatic non-natural amino acid at
positions of apoA-Icons
.
a
>100*, Weakly active at the highest concentration tested.

8672
E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
observed with analog S6H-13 (10). Weak activity was observed
with analogs S7H-6 (27) and S7H-20 (33) when tested at a concen-
tration of 100 lM. Further, when analogs with single R7H substitu-
tions (34–42) were tested, only the variant with a substitution at
position 13, R7H-13 (38), showed significant activity. Substitutions
with a non-natural amino acid having an even longer (C8) carbon
alkenyl side chain resulted in analog S8H-13 (44), which displayed
20 in many possible combinations, either with triple substitutions
(53–57), or with quadruple substitutions (58), all exhibited re-
duced potency when compared with their respective double sub-
stitution analogs.
2.3. In vitro evaluation of pre-b HDL formation
significantly increased activity (EC50 0.7 0.35
Introduction of fully saturated non-natural amino acids S7H3 (C7
chain) and S8H3 (C8 chain) (Fig. 2) at position 13 resulted in analogs
l
M).
Selected apoA-Icons analogs which had shown activity in the
in vitro cholesterol efflux assay were further characterized for their
ability to induce pre-b HDL formation upon incubation with plas-
ma. When using human plasma, all peptides tested (43, 44, 45,
46, 48, and 56) were capable of generating a dose-dependent in-
crease in pre-b HDL, with peptide 48 being the most potent among
peptides examined (Fig. 3a). In an independent experiment pep-
tides 44, 46 and 48 were compared with control peptides 1
(apoA-Icons) and 2 (D4F) (Supplementary Fig. 1). As expected pep-
tide 1 was completely inactive while peptide 48 was more potent
than D4F (2), 44 and 46. Note that the stack-ranking of peptides’
potency in pre-b HDL formation is slightly different from their
stack-ranking in the cholesterol efflux assay. One possible explana-
tion is that the two assays reflect distinct features of the peptides.
Reports in the literature suggest that apoA-I mimetic peptides in-
duce formation of pre-b HDL by displacing apoA-I from mature
HDL.19 Thus, pre-b HDL formation reveals how readily a peptide
interacts with mature HDL and subsequently displaces apoA-I,
while the cholesterol efflux assay primarily assesses a peptide’s
ability to interact with the cell membrane and elicit efflux of free
cholesterol. Given that peptide 48 appeared to be the most potent
peptide in both assays, the two features may be related and collec-
tively determine a peptide’s therapeutic potential.
S7H3-13 (43) and S8H3-13 (45), which had potencies 0.7 0.08
lM
and 0.7 0.23 M, respectively, slightly better than alkenyl mutant
l
S7H-13 (30) but in a similar range with alkenyl mutant S8H-13 (44).
In summary, analysis of the single substitution analogs in the
in vitro cholesterol efflux assay indicate that activity was gained
mainly by the introduction of an aliphatic chain longer than C6
at position 13. Weak activity was also generated by the introduc-
tion of an aliphatic chain longer than C6 at positions 2, 6, and 20.
Interestingly, if positions 2, 6, 13, 20 are highlighted on the helical
wheel representation of the apoA-Icons they are all localized on a
narrow corridor of the hydrophobic portion along the amphipathic
helix (Fig. 1). To determine if cholesterol efflux activity could be
further improved with multiple substitutions we synthesized ana-
logs with substitutions at position 13 and at positions 2, 6, or 20.
Activity of the double substitution analogs in the cholesterol efflux
assay is shown in Table 3. In general, combining a substitution at
position 13 with one at any of positions 2, 6, or 20 improved the
cholesterol efflux activity of the singly substituted peptide. Intro-
duction of the aliphatic amino acid S7H at positions 2 and 13 (ana-
log 47) or positions 6 and 13 (analog 48) brought an improvement
in potency, in the range of 0.4
l
M, with respect to the potency
We next examined whether peptide 48 was able to elicit pre-b
HDL formation upon incubation with mouse plasma. In normal
shown by the single mutant analog having S7H at position 13
(30). Also analogs 47 and 48 were more potent than the singly
substituted analog 30 but showed a potency in the same range of
single mutants at position 13 with a saturated side chain (43) or
mice the majority of circulating HDL exists in the mature,
a-
migrating form (larger than 8.2 nm), with very low levels of pre-
b HDL present in the 7.1–8.2 nm range.20–22 Incubation of mouse
plasma with increasing concentrations of peptide 48 resulted in a
dose-dependent increase in the intensity of the pre-b band
(Fig. 3b), suggesting that the peptide displaced mouse apoA-I from
mature HDL to form pre-b particles. Similar observations were re-
ported with peptides 4618 and D4F (2).12
a longer (C8) carbon chain (44, 45). The double mutant with S7H
,
at positions 13 and 20 (analog 49) seemed to be less potent, with
an EC50 near the micromolar range. Similarly, the double mutant
previously described18 with R6H at position 6 and S7H at position
13 (46) showed a slightly lower potency suggesting that the S con-
figuration was overall preferred.
Wool et al. have reported that the apoA-I mimetic L4F (the L-
Introduction of aliphatic amino acids at more than two posi-
tions was detrimental for activity in the cholesterol efflux assay,
as shown by peptides 53–58 in Table 3. When an analog with a
substitution at position 13 was combined with those at 2, 6, or
amino acid version of D4F), upon incubation with plasma, localizes
mainly within the HDL fraction, although in the presence of iso-
lated LDL, it associates with this lipoprotein.23 Peptide 48 was
incubated with either human or mouse plasma for 2 h. Plasma
Figure 3. (a) Increasing concentrations of selected apoA-I mimetic peptides were incubated with human plasma at 37 °C for 2 h. Pre-b HDL formation was assessed by
quantitative ELISA. (b) Increasing concentrations of peptide 48 were incubated with C57Bl/6 mouse plasma at 37 °C for 2 h. Pre-b HDL formation was assessed by native
Western blot. Mature,
a HDL and pre-b HDL were identified based on electrophoretic migration.

E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
8673
Figure 4. (a) Peptide 48 (0.25 mg/ml) was incubated with C57Bl/6 mouse plasma at 37 °C for 2 h, and then subjected to lipoprotein analysis by FPLC. (b) Peptide 48 (0.5 mg/
ml) was incubated with human plasma at 37 °C for 2 h, and then subjected to lipoprotein analysis by FPLC. (c) Peptide 48-treated human plasma was alternatively analyzed
by tandem size exclusion chromatography, to allow separation of pre-b fractions from mature
a-HDL. Cholesterol and peptide concentrations were measured in resulting
fractions. (d) ApoA-I concentrations in resulting fractions were determined using an in-house human apoA-I ELISA.
lipoproteins were then fractionated by size exclusion chromatog-
raphy, and each fraction was analyzed for cholesterol and peptide
content. As shown in Figure 4a and b, peptide 48 preferentially
associated with HDL, regardless of the presence or abundance of
other lipoproteins. These results suggested that, in vitro, peptide
48 has a higher affinity for HDL than LDL or VLDL. Peptide incorpo-
ration did not change cholesterol levels in VLDL+LDL, or in HDL
(Fig. 4c). Analysis of apoA-I content in the fractions indicated that
in control samples (human plasma incubated with PBS) apoA-I ap-
peared as a large peak in fractions containing mature HDL and only
as a small peak in fractions containing pre-b HDL (Fig. 4d). In com-
parison, peptide incorporation into mature HDL resulted in an in-
crease in apoA-I in pre-b HDL containing fractions (Fig. 4d).
These observations confirmed that under the conditions tested,
peptide 48 was able to incorporate into mature HDL, while displac-
ing apoA-I to form pre-b HDL.
peptide is active, serum samples from treated animals were used
to measure ex vivo cholesterol efflux. Cholesterol efflux capacity
was increased in the serum harvested from peptide-treated mice.
This increase was time-dependent and correlated with an increase
in serum pre-b HDL (Fig. 5b). Peptide concentration in serum also
correlated with both pre-b HDL level and cholesterol efflux capac-
ity. Taken together, these results suggest that peptide administra-
tion to apoeꢀ/ꢀ mice had triggered HDL remodeling and pre-b HDL
formation in a fashion similar to that observed in our in vitro stud-
ies, and that the increase in pre-b HDL was likely the main contrib-
utor to the increased ABCA1-dependent efflux capacity of the
serum samples.
These results indicate that peptide 48 possesses several of the
noted beneficial attributes of other apoA-I mimetic peptides,
namely the ability to generate pre-b HDL formation in vivo and
to increase serum cholesterol efflux capacity, which is an impor-
tant pharmacodynamic readout for apoA-I mimetic peptides.
However, we observed no increase in HDL cholesterol, at any
peptide dose, either in vitro or in vivo. As epidemiological observa-
tions point to an inverse relationship between plasma HDL-C con-
centration and cardiovascular risk,25 the inability of these peptides
to increase plasma HDL-C suggests the possibility that they may
not have a therapeutic effect.
2.4. In vivo evaluation of pre-b HDL formation and serum
cholesterol efflux
As in vitro incubation of peptide 48 with human or mouse plas-
ma resulted in the formation of pre-b HDL particles, we proceeded
to assess whether a similar phenomenon would occur in vivo. Pilot
studies in C57Bl/6 mice indicated that administration of peptide 48
to mice, at doses resulting in plasma levels comparable to the con-
centration of peptide that promoted pre-b HDL formation in vitro,
induced a time- and dose-dependent accumulation of pre-b HDL in
plasma (data not shown). In order to determine if HDL mimetic
peptides would elicit a similar response in a relevant disease mod-
el, we tested the ability of peptide 48 to induce pre-b HDL forma-
tion in apoeꢀ/ꢀ mice. Administration of peptide 48 (60 mg/kg,
subcutaneously) to apoeꢀ/ꢀ mice resulted in a time-dependent in-
crease in circulating pre-b HDL (Fig. 5a). Pre-b HDL is the main
acceptor of cellular cholesterol efflux mediated by the ABCA1
transporter.24 To determine if the pre-b HDL induced by the
2.5. Effects of apoA-I mimetic peptides on serum triglyceride
levels
Studies in which full length apoA-I was administered to hu-
mans, either as lipid-free protein or formulated with diverse phos-
pholipids, showed elevations in circulating triglycerides which
paralleled the exposure to exogenous apoA-I.26–28 A similar obser-
vation has been made in mice, rats, and rabbits.29–31 It has been
suggested that apoA-I can induce hypertriglyceridemia by inhibi-
tion of the degradation of triglyceride-rich lipoproteins.26 These
observations prompted us to examine the effects of our peptides

8674
E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
Figure 5. (a) Peptide 48 was administered to apoeꢀ/ꢀ mice subcutaneously at 60 mg/kg. Serum samples were collected at various time points post-dosing and analyzed for
pre-b HDL formation. (b) ABCA1-dependent cholesterol efflux capacity of the corresponding serum samples was measured in a J774 cholesterol efflux assay. (c) Peptide
concentrations in the serum samples.
Figure 6. (a) Selected peptides were administered to C57Bl/6 mice at 20 or 60 mg/kg subcutaneously. Circulating triglyceride levels in serum samples at 1 or 3 h post-dosing
were determined. (b) The same peptides shown in panel 6a were tested for their ability to inhibit LPL activity in vitro. Peptide levels in plasma in panel 6a (60 mg/kg, 3 h)
were at or above the IC50 for LPL inhibition.
on circulating triglyceride levels in mice. As shown in Figure 6a,
administration of 20 mg/kg of selected apoA-I mimetic peptides
to C57Bl/6 mice had no effect on triglyceride levels, at 1 or 3 h
post-administration (the time points where we observed maximal
production of pre-b HDL). In contrast, with the exception of pep-
tide 1 (apoA-Icons), administration of 60 mg/kg of the same pep-
tides resulted in time-dependent accumulation of triglycerides.
Further investigation of this observation indicated that apoA-I mi-
metic peptides can inhibit lipoprotein lipase (LPL) in an in vitro LPL
activity assay (Fig. 6b). All of the peptides tested were capable of
inhibiting LPL, but the least potent of them was peptide 1 (apoA-
sulted in a time-dependent accumulation of triglycerides (Fig. 2s in
Supplementary data). Both D4F and 5A were also capable of inhib-
iting LPL with an IC50 of 24
lM
3 and 57
lM
32, respectively, in
the same concentration range of the other peptides tested, suggest-
ing that this might be a class effect and not a specific feature of our
apoA-I mimetics.
To further investigate the phenomenon of hypertriglyceridemia,
we studied the effects of peptide 48 on lipoprotein composition
after administration to C57Bl/6 mice. As shown in Figure 7, admin-
istration of peptide 48 at 30 mg/kg (a dose that did not induce
hypertriglyceridemia) resulted in a distribution profile in which
the peptide associated mostly with HDL, with peak values ob-
served at 3 h (corresponding approximately to the Cmax of the pep-
tide). The accumulation of peptide 48 in the HDL fraction did not
induce any significant changes in either cholesterol or triglyceride
content of HDL, LDL or VLDL. In comparison, when peptide 48 was
dosed at 100 mg/kg, significant hypertriglyceridemia was observed
Icons). Except for control peptide 1 (apoA-Icons), plasma levels for
the peptides in the study described in Figure 6a were at or above
the IC50 values of in vitro LPL inhibition, suggesting that triglycer-
ide accumulation in vivo may have resulted from LPL inhibition.
Similarly, in an independent experiment, we found that adminis-
tration of D4F (2) and 5A (3) at 60 mg/kg, but not at 20 mg/kg, re-

E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
8675
Figure 7. Peptide 48 was administered to C57Bl/6 mice at 30 or 100 mg/kg subcutaneously. Serum samples at various time points post-dosing were subjected to lipoprotein
fractionation by FPLC. Peptide concentration, cholesterol, and triglyceride levels were determined in resulting fractions.
and the distribution profile of the peptide differed significantly
from that observed after a 30 mg/kg dose. We found that although
at 1 h the majority of the peptide was associated with HDL, by 3 h
there was no observable increase of peptide in the HDL fraction,
with all new peptide accumulating in the VLDL fraction. Peptide
continued to accumulate in the VLDL fraction up to 6 h. We also
observed increases in both cholesterol and triglycerides in VLDL,
concurrent with peptide accumulation in VLDL. Triglyceride con-
tent of HDL did not change over the 24 h of the study, but HDL cho-
lesterol was surprisingly reduced in parallel with the increase in
VLDL cholesterol. At 24 h, we detected an increase in cholesterol
in the fractions corresponding to LDL, which could be explained
by catabolism of VLDL after clearance of peptide 48. A similar
observation of increased cholesterol content in the VLDL and LDL
fractions has been reported by Shah et al. after administration of
recombinant human apoA-IMilano to apoeꢀ/ꢀ mice.32 Taken to-
gether, these observations suggest that the association of apoA-I
mimetic peptide with HDL is likely a saturable process, and that
once maximal association is reached, peptide becomes available
for association with other lipoproteins, in this case VLDL. Given
the ability of apoA-I mimetic peptides to inhibit LPL activity, this
initiates a cycle in which VLDL triglycerides accumulate, and this
lipoprotein becomes available for the incorporation of any choles-
terol the peptide may mobilize.
tide sequence into which non-natural amino acids were introduced
at specific positions of the hydrophobic portion. In vitro our
mimetics can stimulate cholesterol efflux from macrophages and
promote pre-b HDL formation in both mouse and human plasma.
Their potencies in these activities are superior to other peptide
mimetics in development and are in a similar range as apoA-I.
When incubated at moderate concentrations with human plasma
our peptides are distributed mostly in HDL fractions. When admin-
istered to mice, our peptides trigger HDL remodeling and pre-b
HDL formation. Moreover we observed a concomitant increase in
ABCA1-dependent efflux capacity of serum from peptide-dosed
animals.
However, regarding therapeutic potential of apoA-I mimetics,
there are some concerns. The administration at higher dose of
apoA-I mimetics in our study, including previously reported D4F
and 5A, resulted in triglyceride accumulation due to LPL inhibition.
In fact once our peptides saturated HDL, they incorporated into
VLDL and increased plasma triglyceride associated with VLDL. Pre-
vious reports about triglycerides elevations with apoA-I protein
and our data with D4F and 5A suggest that this might be a class is-
sue. Therefore a narrow therapeutic window in terms of safe dos-
ing might exist for apoA-I peptide mimetics.
4. Experimental section
4.1. Materials
Thus, special attention has to be paid to the dosing regimen of
apoA-I mimetic peptides, and an appropriate therapeutic window
needs to be identified to ensure HDL association is not saturated,
and there is no surplus of peptide available to interact with VLDL
thereby inducing hypertriglyceridemia.
All amino acid derivatives, the Rink linker p-[(R,S)-a-[9H-Fluo-
ren-9-yl-methoxyformamido]-2,4-dimethoxybenzyl]-phenoxyace-
tic acid, and coupling reagents were purchased from EMD
Biosciences. The Fmoc-Linker AM-PS based resin 1% cross-linked
was purchased from Biosearch Technologies Inc.
3. Conclusions
Our structure–activity relationship studies led us to the identi-
fication of potent apoA-I peptide mimetics based on a 22-mer pep-
ACS grade solvents were purchased from Sigma–Aldrich (Ger-
many) and used as obtained.

8676
E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
Determination of total cholesterol, triglycerides and fatty acids
(Millipore, MillexÒ-LH, 0.45
lm) and the solvent evaporated under
was performed by colorimetric assays purchased from WAKO
Diagnostics (WAKO Chemicals, USA, Inc., Richmond, VA).
[3H] cholesterol was purchased from Perkin–Elmer (Waltham,
MA). 8-(4-chlorophenylthio)-adenosine 30:50-cyclic monophos-
phate (cpt)-cAMP, Bovine serum albumin (BSA), Tris, HEPES (and
other buffer reagents), bovine lipoprotein lipase, protease inhibitor
cocktail, lipase inhibitors (paraoxan, di-ethyl-p-nitrophenyl phos-
phate) were obtained from Sigma (Sigma–Aldrich, St. Louis, MO).
Culture medium (DMEM and others) and fetal calf serum (FCS)
were purchased from Invitrogen (Invitrogen Corporation, Carlsbad,
CA).
reduced pressure. The crude oil is taken in ethyl acetate, washed
three times with H2O and dried over Na2SO4. The white solid is
washed with petrol ether and dried under vacuum to afford the de-
sired amino acids as determined by 1H NMR analysis (74% yield):
4.3.1. Fmoc-S7H3-OH—(2S)-2-{[(9H-fluoren-9-ylmethoxy)-
carbonyl]amino}nonanoic acid
Obtained as white solid; 1H NMR (400 MHz, DMSO-d6) d: 7.88
(d, 2H), 7.70 (d, 2H), 7.54 (d, 1H), 7.38 (t, 2H), 7.28 (t, 2H), 4.24
(m, 2H), 3.92 (m, 1H), 1.65 (m, 2H), 1.42–1.18 (m, 10H), 0.82 (m,
3H). MS (ES) C24H29NO4 required 395.49, found: 396.5.
Human pre-b HDL Elisa was obtained from Daiichi (Daiichi Pure
Chemicals, Tokyo, Japan). Purified human apoA-I protein, Rabbit
anti-mouse apoA-I antibody and monoclonal mouse anti-human
apoA-I antibody were purchased from Biodesign International
(Saco, ME); HRP-conjugated goat anti-rabbit secondary antibody
was purchased from Pierce (Rockford, IL); anti-human apoA-I poly-
clonal antibody was purchased from Academy Biomedical (Hous-
ton, TX). Purified human VLDL was obtained from Intracel
(Frederick, MD). Immulon high-binding plates were purchased
from ThermoLabsystems.
4.3.2. Fmoc-R7H3-OH—(2R)-2-{[(9H-fluoren-9-ylmethoxy)-
carbonyl]amino}nonanoic acid
Obtained as white solid; 1H NMR (400 MHz, DMSO-d6) d: 7.90
(d, 2H), 7.72 (d, 2H), 7.56 (d, 1H), 7.40 (t, 2H), 7.28 (t, 2H), 4.26
(m, 2H), 3.88 (m, 1H), 1.60 (m, 2H), 1.42–1.14 (m, 10H), 0.82 (m,
3H). MS (ES) C24H29NO4 required 395.49, found: 396.5.
4.3.3. Fmoc-S8H3-OH—(2S)-2-{[(9H-fluoren-9-ylmethoxy)-
carbonyl]amino}decanoic acid
Obtained as white solid; 1H NMR (400 MHz, DMSO-d6) d: 7.88
(d, 2H), 7.72 (d, 2H), 7.58 (d, 1H), 7.42 (t, 2H), 7.32 (t, 2H), 4.28
(m, 2H), 3.90 (m, 1H), 1.62 (m, 2H), 1.40–1.16 (m, 10H), 0.88 (m,
3H). MS (ES) C25H31NO4 required 409.5, found: 410.7.
4.2. Synthesis of non coded hydrocarbon side chain amino acids
The Fmoc hydrocarbon side chain amino acids S6H, R7H, and S8H
were synthesized following the same synthetic procedure as previ-
ously reported.18
4.4. Peptide synthesis
4.2.1. Fmoc-S6H-OH—(2R)-2-{[(9H-fluoren-9-ylmethoxy)-
carbonyl]amino}oct-7-enoic acid
The apoA-I mimetic peptides were synthesized by solid phase
using Fmoc/tBu chemistry on a peptide synthesizer SYMPHONY
(Protein Technologies, Inc). All the acylation reactions were per-
formed for 60 min with a five-fold excess of activated amino acid
(HBTU/DIEA) over the resin free amino groups following the end
of peptide assembly on the synthesizer. The side chain protecting
groups were: tert-butyl for Asp and Glu; tert-butyloxy carbonyl
(BOC) for Lys; trityl for Asn and Gln; 2,2,4,6,7-pentame-
thyldihydrobenzofuran-5-sulfonyl for Arg. The Proline at the N-ter-
minal was introduced as Ac-Proline-OH by using HBTU/DIEA as
Obtained after lyophilization as white powder (54% yield); 1H
NMR (400 MHz, DMSO-d6) d: 7.92 (d, 2H), 7.76 (d, 2H), 7.55 (d,
1H), 7.45 (t, 2H), 7.32 (m, 2H), 5.82 (m, 1H), 5.01 (m, 2H), 4.30
(m, 3H), 3.94 (m, 1H), 2.05 (m, 2 H); 1.8–1.6 (m, 2H), 1.38 (m,
4H). MS (ES) C23H25NO4 required 379.44, found: 380.4.
4.2.2. Fmoc-R7H-OH—(2S)-2-{[(9H-fluoren-9-ylmethoxy)-
carbonyl]amino}non-8-enoic acid
Obtained after lyophilization as white powder (40% yield); 1H
NMR (400 MHz, DMSO-d6) d: 7.96 (d, 2H), 7.75 (d, 2H), 7.65 (d,
1H), 7.45 (t, 2H), 7.32 (t, 2H), 5.84 (m, 1H), 5.01 (m, 2H), 4.42 (m,
3H), 3.94 (m, 1H), 2.04 (m, 2H); 1.65 (m, 2H), 1.36 (m, 6H). MS
(ES) C24H27NO4 required 393.48, found: 394.4. (>95% ee as revealed
by HPLC analysis).
coupling reagents in DMF. All the non-natural amino acids (R7H,
S7H, R8H, S8H, S7H3, R7H3, S8H3, R8H3,) were coupled manually by
using HBTU/DIEA as activators and the coupling repeated if neces-
sary. After the coupling the remainder of the synthesis was per-
formed automatically as described above.
At the end of the synthesis, the dry peptide-resins were individ-
ually treated with the cleavage mixture, 88% TFA, 5% phenol, 2% tri-
isopropylsilane, and 5% water for 1.5 h at room temperature. In
case of peptides containing olefinic-side chain amino acids, the
cleavage mixture was composed by 95% TFA and 5% water. Each re-
sin was filtered and the solution was added to cold methyl-t-butyl
ether in order to precipitate the peptide. After centrifugation, the
peptide pellets were washed three times with fresh cold methyl-
t-butyl ether to remove the organic scavengers. Final pellets were
dried, resuspended in H2O, 20% acetonitrile, and lyophilized.
Analytical HPLC was performed on a ReproSil-Pur 300 C4 or C18
4.2.3. Fmoc-S8H-OH—(2S)-2-{[(9H-fluoren-9-ylmethoxy)-
carbonyl]amino}dec-9-enoic acid
Obtained after lyophilization as white powder (42% yield); 1H
NMR (400 MHz, DMSO-d6) d: 7.88 (d, 2H), 7.70 (d, 2H), 7.54 (d,
1H), 7.38 (t, 2H), 7.28 (t, 2H), 4.24 (m, 2H), 3.92 (m, 1H); 1.62
(m, 2H), 1.28 (m, 8 H). MS (ES) C25H29NO4 required 407.50, found:
408.5.
4.3. Synthesis of Fmoc-S7H3-OH, Fmoc-R7H3-OH, and Fmoc-S8H3
-
OH
column (150 ꢁ 4.6 mm, 5
or C18 column (150 ꢁ 4.6 mm, 3
rate 1 ml/min at 45 °C or Acquity UPLC BEHÒ C18 column
m, 130 Å) (Waters) flow rate 0.4 ml/min at
45 °C. Crude peptides were purified by reverse-phase HPLC using
preparative Waters Prep LC 4000 System or GX-281 Gilson Trilu-
l
m, 300 Å) (Dr. Maisch GmbH) or ACE C4
lm, 300 Å) (CPS Analitica) flow
Direct Pd/C-catalyzed hydrogenation of corresponding unsatu-
rated Fmoc-amino acids (Fmoc-S7H-OH, Fmoc-R7H-OH, and Fmoc-
S8H-OH)18 in MeOH gave the all hydrocarbon non-natural amino
acids. We used the standard procedure where the mixture of the
substrate (500 mg) and 10% Pd/C (70 mg, 0.05 mol %) in MeOH
(20 ml) in a round-bottom flask after two vacuum/H2 cycles to re-
place air inside with hydrogen, was vigorously stirred at room tem-
perature (ca. 20 °C) under ordinary hydrogen pressure (balloon) for
30 min. The reaction mixture was filtered using a membrane filter
(100 ꢁ 2.1 mm, 1.7
l
tion LC equipped with
(250 ꢁ 20 mm, 10 m) (Dr. Maisch GmbH) or RCM Delta-Pak C4
cartridges (40 ꢁ 200 mm, 15 m) or RCM Delta-Pak C18 cartridges
(40 ꢁ 200 mm, 15 m) or ACE C18 (250 ꢁ 21 mm, 10 m, 300 Å)
(CPS analitica) and using as eluents (A) 0.1% TFA in water and (B)
a
ReproSil-Pur 300 C4 column
l
l
l
l

E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
8677
0.1% TFA in acetonitrile, flow rate 30 or 80 ml/min, respectively.
The purified peptides were characterized by electrospray mass
spectrometry on a Micromass LCZ platform and/or Acquity SQD
Waters.
difference between cholesterol measured in the presence and the
absence of cAMP.33
4.9. Ex vivo serum cholesterol efflux assay
4.5. Peptide concentration measurement
ABCA1-dependent serum cholesterol efflux assays were per-
formed similarly to previously described.21 Briefly, J774 macro-
phages were seeded in 24 well tissue culture plates, labeled with
[3H] cholesterol for 24 h, then incubated in media containing
0.2% BSA for 18 h to allow cellular cholesterol radiolabel to equili-
brate. Incubations with or without 8-(4-chlorophenylthio)-adeno-
sine 30:50-cyclic monophosphate (cpt)-cAMP were both included.
Cells were then washed once, and incubated with acceptor media
containing mouse serum diluted to 2% at 37 °C for 4 h. [3H] re-
leased into the media was compared with total [3H] at time zero
to determine the percent release of [3H]cholesterol (fractional ef-
flux). ABCA1-dependent efflux was derived by subtracting frac-
tional efflux of non-cAMP-treated cells from fractional efflux of
cAMP-treated cells.
Fractions collected from chromatography were mixed at 1:3 ra-
tios with an organic solvent (acetonitrile with 0.1% trifluoroacetic
acid). A structurally distinct peptide was also added as internal
standard. Samples were spined at 2000 rpm for 10 min. Superna-
tant was taken and subjected to LC–MS analysis of peptide concen-
tration using the RapidFire System (Biotrove, Woburn, MA). Serum
levels of peptide were measured by LC–MS/MS after protein pre-
cipitation with acetonitrile.
4.6. Lipid analysis
Total cholesterol or triglyceride in serum or column fractions
were determined by enzymatic colorimetry by commercially avail-
able assays according to manufacturer’s instructions (WAKO
Diagnostics).
4.10. Native gel-Western quantitation of mouse pre-b HDL
Two microliters of peptide 48 solutions of increasing concentra-
4.7. In vivo administration of peptides
tion were added to 20 ll C57Bl/6 mouse plasma, incubated at 37 °C
for 2 h, and then subjected to one-dimensional, non-reducing, anti-
mouse apoA-I Western blotting. Briefly, samples were diluted 1:2
with sample buffer (60% sucrose, 0.1% bromophenol blue) and re-
solved by nondenaturing, non-reducing one-dimensional 4–20%
polyacrylamide gel electrophoresis at 50 V for 30 min and 120 V
for 3 h. Gels were transferred to nitrocellulose overnight (28 V at
4 °C). After transfer, the membrane was blocked with 0.5% BSA
and washed three times with TBST (10 mM Tris, pH 7.4, 150 mM
NaCl, and 0.1% Tween 20). Rabbit anti-mouse apoA-I antibody (Bio-
design International) and HRP-conjugated goat anti-rabbit second-
ary antibody were used as the primary and secondary antibodies,
respectively. Blots were developed with ECL reagent (GE Health-
care, Piscataway, NJ), and scanned using a Typhoon 9400 Imager
(GE Healthcare, Piscataway, NJ). Pre-b HDL was identified based
on electrophoretic mobility (7.1–8.2 nm), and the band was quan-
titated using ImageQuant software, with the results expressed as
arbitrary units (AU) representing signal intensity of the band after
background subtraction of an equal-sized area.
All studies described here were approved by the Merck Re-
search Laboratories (Rahway, NJ) Institutional Animal Care and
Use Committee. Male mice C57Bl/6 (Taconic, Germantown, NY)
or apoeꢀ/ꢀ (Jackson Laboratory, Bar Harbor, ME), at least 12 weeks
of age, were used in these studies. Animals were maintained in
environmentally controlled temperature and humidity facilities,
with a 12 h light/dark cycle (7:00 am–7:00 pm). All animals were
fed standard rodent chow for the duration of the studies. Animals
had ad libitum access to food and water.
On the day of the study, mice were injected subcutaneously
with either control vehicle (50 mM mannitol in phosphate buffered
saline) or test peptides, at 20–100 mg/kg. At predetermined time
points, animals were euthanized by CO2 asphyxiation, and blood
collected by cardiac puncture. Blood was allowed to clot for 1 h
at room temperature and serum separated by centrifugation at
10,000g for 10 min. Serum samples were immediately frozen at
ꢀ80 °C for further analysis (serum-mediated cholesterol efflux,
pre-b HDL measurement, triglyceride measurement) or filtered
through a 0.65
lm filter (Millipore, Billerica, MA), mixed with
4.11. ELISA quantitation of human pre-b HDL
lipase and protease inhibitors, and used for lipoprotein isolation
as described below. 3–5 animals were used per dose and time
point.
Increasing concentrations of various peptides were incubated
with normal human plasma at 37 °C for 2 h, as described above.
The amount of pre-b HDL in the resulting samples was determined
by a commercially available human pre-b HDL ELISA (Daiichi Pure
Chemicals, Tokyo, Japan).
4.8. In vitro Cholesterol efflux assay
Mouse macrophage cells (RAW 264.7, ATCC, Manassas, VA)
were maintained in DMEM medium with10% FCS and 1% PSG in
T225 flasks to a confluency of ꢂ80%. For efflux assays, cells were
seeded into 48 well plates at a concentration of 1 ꢁ 105 cells/
0.2 ml/well in the same medium supplemented with [3H]
cholesterol.
4.12. ELISA quantitation of human apoA-I
Human apoA-I concentrations in samples were determined
using a sandwich ELISA. Immulon high-binding plates were coated
overnight with 0.5 lg/ml monoclonal mouse anti-human apoA-I
After 24 h, cells were washed once and starved overnight with
serum-free medium supplemented with 0.1% lipid-free BSA. Final-
ly, peptides with or without cAMP in serum-free medium were
added and incubated for another 24 h.
antibody at 4 °C. Plates were then blocked with TBST (150 mM
NaCl, 50 mM Tris, pH 7.5, 0.05% Tween20) supplemented with 1%
BSA for 1 h at room temperature, washed with TBST, and then
incubated with diluted samples for 2 h at room temperature. Plates
Subsequently, medium and cell lysate (0.5% Triton X-100 in
1 mM HEPES, pH 7.5) were collected and aliquots were mixed
separately with scintillation cocktail and counted. Cholesterol
efflux was calculated as the percentage of radioactivity associated
with medium over the sum of radioactivity of both medium and
lysate. cAMP-dependent cholesterol efflux was determined by the
were then washed, incubated with 0.5 lg/ml biotinylated goat
anti-human apoA-I polyclonal antibody for 1 h at room tempera-
ture. After washes, plates were subjected to the Streptavidin/Euro-
pium DELFIA detection system (Envision plate reader, model
#2103, Perkin–Elmer, Shelton, CT). Purified human apoA-I was
used as a standard.

8678
E. Carballo-Jane et al. / Bioorg. Med. Chem. 18 (2010) 8669–8678
4.13. Peptide-lipoprotein association in vitro
References and notes
1. Linsel-Nitschke, A.; Tall, A. R. Nat. Rev. Drug Disc. 2005, 4, 193.
2. Rader, D. J. J. Clin. Invest. 2006, 116, 3090.
3. Wang, X.; Collins, H. L.; Ranalletta, M.; Fuki, I. V.; Billheimer, J. T.; Rothblat, G.
H.; Tall, A. R.; Rader, D. J. J. Clin. Invest. 2007, 117, 2216.
4. Wang, N.; Silver, D. L.; Costet, P.; Tall, A. R. J. Biol. Chem. 2000, 275, 33053.
5. Rubin, E. M.; Krauss, R. M.; Spangler, E. A.; Verstuyft, J. G.; Clift, S. M. Nature
1991, 353, 265.
6. Plump, A. S.; Scott, C. J.; Breslow, J. L. Proc. Natl. Acad. Sci. USA 1994, 91, 9607.
7. Nissen, S. E. et al JAMA 2003, 290, 2292.
Peptide 48 was mixed with C57Bl/6 mouse plasma at a final
concentration of 0.25 mg/ml or with human plasma at a final con-
centration of 0.5 mg/ml. Plasma was then incubated at 37 °C for
1 h, and lipoprotein particles fractionated by size exclusion chro-
matography (see below). To measure peptide association with
pre-b HDL particles, human plasma was incubated with peptide
48 at 0.1 mg/ml, incubated at 37 °C for 1 h, and then pre-b HDL
was separated from HDL by tandem size exclusion (tandem HP-
SEC) chromatography as described below. A PBS-spiked control
sample was also included in these studies.
8. Tardif, J.-C. JAMA 2007, 297, 1675.
9. Natarajan, P.; Forte, T. M.; Chu, B.; Phillips, M. C.; Oram, J. F.; Bielicki, J. K. J. Biol.
Chem. 2004, 279, 24044.
10. Gillotte, K. L.; Zaiou, M.; Lund-Katz, S.; Anantharamaiah, G. M.; Holvoet, P.;
Dhoest, A.; Palgunachari, M. N.; Segrest, J. P.; Weisgraber, K. H.; Rothblath, G.
H.; Phillips, M. C. J. Biol. Chem. 1999, 274, 2021.
11. Remaley, A. T.; Thomas, F.; Stonik, J. A.; Demosky, S. J.; Bark, S. E.; Neufeld, E. B.;
Bocharov, A. V.; Vishnyakova, T. G.; Patterson, A. P.; Eggerman, T. L.;
Santamarina-Fojo, S.; Brewer, H. B. J. Lipid Res. 2003, 44, 828.
12. Navab, M.; Anantharamaiah, G. M.; Srinivasa, T. R.; Hama, S.; Hough, G.;
Grijalva, V. R.; Wagner, A. C.; Frank, J. S.; Datta, G.; Garber, D.; Fogelman, A.-M.
Circulation 2004, 3215.
4.14. Lipoprotein analysis by size exclusion chromatography
Plasma or serum lipoproteins were separated by a fast protein
liquid chromatography (FPLC) system using a Superose 6 HR col-
umn (Amersham Pharmacia Biotech, Piscataway, NJ). The column
was equilibrated with PBS (pH 7.4) plus 1 mM EDTA and was run
at a flow rate of 0.2 ml/min. The serum samples were pooled (three
animals per pool), supplemented with protease inhibitor cocktail
(Sigma) and lipase inhibitor (paraoxan, di-ethyl-p-nitrophenyl
13. Sethi, A. A.; Stonik, J. A.; Thomas, F.; Demosky, S. J.; Amar, M.; Neufeld, E.;
Brewer, H. B.; Davidson, W. S.; D’Souza, W.; Sviridov, D.; Remaley, A. T. J. Biol.
Chem. 2008, 283, 32273.
14. Tabet, F.; Remaley, A. T.; Segaliny, A. I.; Millet, J.; Yan, L.; Nakhla, S.; Barter, P. J.;
Rye, K.-A.; Lambert, G. Arterioscler. Thromb. Vasc. Biol. 2010, 3, 246.
15. Navab, M.; Anantharamaiah, G. M.; Hama, S.; Garber, D. W.; Chaddha, M.;
Hough, G.; Lallone, R.; Fogelman, A. M. Circulation 2002, 105, 290.
16. Li, X.; Chyu, K.-Y.; Faria Neto, J. R.; Yano, J.; Nathwani, N.; Ferreira, C.;
Dimayuga, P. C.; Cercek, B.; Kaul, S.; Shah, P. K. Circulation 2004, 110, 1701.
17. Anantharamaiah, G. M.; Venkatachalapathi, Y. V.; Brouillette, C. G.; Segrest, J. P.
Arteriosclerosis 1990, 10, 95.
phosphate, Sigma), filtered with 0.65
lm microcentrifuge filters,
and then 100 l samples were loaded onto the column. Fractions
l
(0.27 ml) were collected and subjected to analyses as described
below.
Lipoprotein analysis by tandem HP-SEC. Tandem HP-SEC was
performed at ambient temperature using a Superdex 200 HR 10/
30 column connected in series to a Superdex 75 HR 10/30 column
in the AKTA FPLC system. The column was equilibrated with PBS
(pH 7.4) plus 2 mM EDTA and was run at a flow rate of 0.35 ml/
min. The serum samples were pooled (three animals per pool),
supplemented with protease inhibitor cocktail and lipase inhibitor,
filtered with 0.65 lm microcentrifuge filters, and then 100 ll sam-
ples were loaded onto the column. Fractions (0.25 ml) were col-
lected and subjected to lipid/protein analyses as described above.
18. Ingenito, R.; Burton, C.; Langella, A.; Chen, X.; Zytko, K.; Pessi, A.; Wang, J.;
Bianchi, E. Bioorg. Med. Chem Lett. 2010, 20, 236.
19. Wool, G. D.; Reardon, C. A.; Getz, G. S. J. Lipid Res. 2008, 49, 1268.
20. Ishida, B. Y.; Frolich, J.; Fielding, C. J. J. Lipid Res. 1987, 28, 778.
21. Duong, P. T.; Weibel, G. L.; Lund-Katz, S.; Rothblat, G. H.; Phillips, M. C. J. Lipid
Res. 2008, 49, 1006.
22. Troutt, J. S.; Alborn, W. E.; Mosior, M. K.; Dai, J.; Murphy, A. T.; Beyer, T. P.;
Zhang, Y.; Cao, G.; Konrad, R. J. J. Lipid Res. 2008, 49, 581.
23. Wool, G. D.; Vaisar, T.; Reardon, C. A.; Getz, G. S. J. Lipid Res. 2009, 50, 1889.
24. Rye, K. A.; Barter, P. J. Arterioscler. Thromb. Vasc. Biol. 2004, 24, 421.
25. Castelli, W. P.; Garriosn, R. J.; Wilson, P. W.; Abbott, R. D.; Kalousdian, S.;
Kannel, W. B. JAMA 1986, 256, 2835.
26. Nanjee, M. N.; Crouse, J. R.; King, J. M.; Hovorka, R.; Rees, S. E.; Carson, E. R.;
Morgenthaler, J. J.; Lerch, P.; Miller, N. E. Arterioscler. Thromb. Vasc. Biol. 1996,
16, 1203.
27. Nanjee, M. N.; Doran, J. E.; Lerch, P. G.; Miller, N. E. Arterioscler. Thromb. Vasc.
Biol. 1999, 19, 979.
28. Shaw, J. A.; Bobik, A.; Murphy, A.; Kanellakis, P.; Blombery, P.; Mukhamedova,
N.; Woollard, K.; Lyon, S.; Sviridov, D.; Dart, A. M. Circ. Res. 2008, 103, 1084.
29. Pedersen, T. X.; Bro, S.; Andersen, M. H.; Etzerodt, M.; Jauhiainen, M.;
Moestrup, S.; Nielsen, L. B. Atherosclerosis 2009, 202, 372.
30. Ha, Y. C.; Trifunovic, Z.; Howlett, G. J. Biochim. Biophys. Acta 1992, 1125, 223.
31. Van Lenten, B. J.; Wagner, A. C.; Navab, M.; Anantharamaiah, G. M.; Hama, S.;
Reddy, S. T.; Fogelman, A. M. J. Lipid Res. 2007, 48, 2344.
4.15. LPL inhibition assay
LPL activity was measured as free fatty acids released from
VLDL. Assays were carried out in 96-well plates with a 50
tion system that includes PBS, 0.1% BSA, 1 U bovine lipoprotein li-
pase (Sigma), 12.5 g purified human VLDL (Intracel, Frederick,
ll reac-
l
MD), and various amounts of peptides. Reaction mixture was first
incubated on ice for 1 h, and then incubated at room temperature
for another hour. Free fatty acids generated in the reaction were
quantitated in the same plate using a commercially available kit
(Wako).
32. Shah, P. K.; Yano, J.; Reyes, O.; Chyu, K. Y.; Kaul, S.; Bisgaier, C. L.; Drake, S.;
Cercek, B. Circulation 2001, 103, 3047.
33. Chen, X.; Burton, C.; Song, X.; McNamara, L.; Langella, A.; Cianetti, S.; Chang, C.
H.; Wang, J. Int. J. Biol. Sci. 2009, 5, 489.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.09.074.