Carbon monoxide-releasing micelles for immunotherapy

Article abstract of DOI:10.1021/ja1075025

With the discovery of important biological roles of carbon monoxide (CO), the use of this gas as a therapeutic agent has attracted attention. However, the medical application of this gas has been hampered by the complexity of the administration method. To overcome this problem, several transition-metal carbonyl complexes, such as Ru(CO)₃Cl(glycinate), [Ru(CO) ₃Cl₂]₂, and Fe(η⁴-2-pyrone)(CO) ₃, have been used as CO-releasing molecules both in vitro and in vivo. We sought to develop micellar forms of metal carbonyl complexes that would display slowed diffusion in tissues and thus better ability to target distal tissue drainage sites. Specifically, we aimed to develop a new CO-delivery system using a polymeric micelle having a Ru(CO)₃Cl(amino acidate) structure as a CO-releasing segment. The CO-releasing micelles were prepared from triblock copolymers composed of a hydrophilic poly(ethylene glycol) block, a poly(ornithine acrylamide) block bearing Ru(CO)₃Cl(ornithinate) moieties, and a hydrophobic poly(n-butylacrylamide) block. The polymers formed spherical micelles in the range of 30-40 nm in hydrodynamic diameter. Further characterization revealed the high CO-loading capacity of the micelles. CO-release studies showed that the micelles were stable in physiological buffer and serum and released CO in response to thiol-containing compounds such as cysteine. The CO release of the micelles was slower than that of Ru(CO) ₃Cl(glycinate). In addition, the CO-releasing micelles efficiently attenuated the lipopolysaccharide-induced NF-κB activation of human monocytes, while Ru(CO)₃Cl(glycinate) did not show any beneficial effects. Moreover, cell viability assays revealed that the micelles significantly reduced the cytotoxicity of the Ru(CO)₃Cl(amino acidate) moiety. This novel CO-delivery system based on CO-releasing micelles may be useful for therapeutic applications of CO.

Full text of DOI:10.1021/ja1075025

Published on Web 12/03/2010
Carbon Monoxide-Releasing Micelles for Immunotherapy
Urara Hasegawa, Andr e´ J. van der Vlies, Eleonora Simeoni, Christine Wandrey, and
Jeffrey A. Hubbell*
Institute of Bioengineering (IBI) and Institute of Chemical Sciences and Engineering (ISIC),
Ecole Polytechnique F e´ d e´ rale de Lausanne (EPFL), Lausanne CH-1015, Switzerland
Received August 19, 2010; E-mail: jeffrey.hubbell@epfl.ch
Abstract: With the discovery of important biological roles of carbon monoxide (CO), the use of this gas as
a therapeutic agent has attracted attention. However, the medical application of this gas has been hampered
by the complexity of the administration method. To overcome this problem, several transition-metal carbonyl
4
complexes, such as Ru(CO)
3
Cl(glycinate), [Ru(CO)
3
Cl
2
]
2
, and Fe(η -2-pyrone)(CO)
3
, have been used as
CO-releasing molecules both in vitro and in vivo. We sought to develop micellar forms of metal carbonyl
complexes that would display slowed diffusion in tissues and thus better ability to target distal tissue drainage
sites. Specifically, we aimed to develop a new CO-delivery system using a polymeric micelle having a
Ru(CO)
from triblock copolymers composed of a hydrophilic poly(ethylene glycol) block, a poly(ornithine acrylamide)
block bearing Ru(CO) Cl(ornithinate) moieties, and a hydrophobic poly(n-butylacrylamide) block. The
3
Cl(amino acidate) structure as a CO-releasing segment. The CO-releasing micelles were prepared
3
polymers formed spherical micelles in the range of 30-40 nm in hydrodynamic diameter. Further
characterization revealed the high CO-loading capacity of the micelles. CO-release studies showed that
the micelles were stable in physiological buffer and serum and released CO in response to thiol-containing
compounds such as cysteine. The CO release of the micelles was slower than that of Ru(CO)
In addition, the CO-releasing micelles efficiently attenuated the lipopolysaccharide-induced NF-κB activation
of human monocytes, while Ru(CO) Cl(glycinate) did not show any beneficial effects. Moreover, cell viability
assays revealed that the micelles significantly reduced the cytotoxicity of the Ru(CO) Cl(amino acidate)
3
Cl(glycinate).
3
3
moiety. This novel CO-delivery system based on CO-releasing micelles may be useful for therapeutic
applications of CO.
Introduction
ride (LPS)-challenged mouse macrophages and human dendritic
cells, CO has been shown to inhibit the production of pro-
Carbon monoxide (CO), one of the byproducts of heme
catabolism via heme oxygenase-1 (HO-1), has recently been
recognized as an important signaling mediator in mammals
despite its bad reputation as an air pollutant. CO plays versatile
roles in tissue protection via anti-inflammatory, antiproliferative,
and antiapoptotic effects. Anti-inflammatory effects have
attracted particular growing attention due to possible applications
to many inflammation-based diseases, and an increasing number
of reports concerning the mechanism through which CO exerts
inflammatory cytokines and increase production of anti-inflam-
4
-7
1
,2
matory cytokines.
In addition, CO suppresses T cell prolif-
8
eration. Though the mechanism is still not fully understood, it
seems that CO elicits these anti-inflammatory effects through
p38 MAPK and/or JNK pathways. Moreover, the effect of
CO has also been confirmed in various disease models such as
4
,6
1
,2
3,4,9
sepsis, chronic graft rejection, and ischemia/reperfusion injury.
Thus, CO appears to have a high potential utility for future
therapeutic applications.
3-9
anti-inflammatory effects have appeared.
In lipopolysaccha-
Despite the interesting biological activities, the development
of methods for delivering CO is in its infancy. Since CO is
toxic at high concentrations, the precise control of the location,
timing, and dosages of CO is the critical factor for sufficient
therapeutic responses. In addition, the fact that CO is a gaseous
molecule makes its controlled delivery more challenging. The
simplest administration method is the inhalation of CO at a
known concentration. However, this method is a very nonspe-
cific approach that can cause adverse effects. These difficulties
led chemists to explore molecules capable of releasing CO in
biological systems (CO-releasing molecules) for safer and more
(
(
(
1) Ryter, S. W.; Otterbein, L. E. Bioessays 2004, 26, 270–80.
2) Ryter, S. W.; Alam, J.; Choi, A. M. Physiol. ReV. 2006, 86, 583–650.
3) Nakao, A.; Kimizuka, K.; Stolz, D. B.; Neto, J. S.; Kaizu, T.; Choi,
A. M.; Uchiyama, T.; Zuckerbraun, B. S.; Nalesnik, M. A.; Otterbein,
L. E.; Murase, N. Am. J. Pathol. 2003, 163, 1587–98.
(4) Otterbein, L. E.; Bach, F. H.; Alam, J.; Soares, M.; Tao Lu, H.; Wysk,
M.; Davis, R. J.; Flavell, R. A.; Choi, A. M. Nat. Med. 2000, 6, 422–
8.
(
(
5) Sarady, J. K.; Otterbein, S. L.; Liu, F.; Otterbein, L. E.; Choi, A. M.
Am. J. Respir. Cell Mol. Biol. 2002, 27, 739–45.
6) Morse, D.; Pischke, S. E.; Zhou, Z.; Davis, R. J.; Flavell, R. A.; Loop,
T.; Otterbein, S. L.; Otterbein, L. E.; Choi, A. M. J. Biol. Chem. 2003,
2
78, 36993–8.
(
7) Remy, S.; Blancou, P.; Tesson, L.; Tardif, V.; Brion, R.; Royer, P. J.;
Motterlini, R.; Foresti, R.; Painchaut, M.; Pogu, S.; Gregoire, M.; Bach,
J. M.; Anegon, I.; Chauveau, C. J. Immunol. 2009, 182, 1877–84.
8) Pae, H. O.; Oh, G. S.; Choi, B. M.; Chae, S. C.; Kim, Y. M.; Chung,
K. R.; Chung, H. T. J. Immunol. 2004, 172, 4744–51.
(9) Otterbein, L. E.; Zuckerbraun, B. S.; Haga, M.; Liu, F.; Song, R.;
Usheva, A.; Stachulak, C.; Bodyak, N.; Smith, R. N.; Csizmadia, E.;
Tyagi, S.; Akamatsu, Y.; Flavell, R. J.; Billiar, T. R.; Tzeng, E.; Bach,
F. H.; Choi, A. M.; Soares, M. P. Nat. Med. 2003, 9, 183–90.
(
10.1021/ja1075025  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 18273–18280 9 18273

A R T I C L E S
Hasegawa et al.
Figure 1. Schematic illustration of micelle formation of PEG-b-OrnRu-b-nBu triblock copolymer and CO release after cellular uptake.
convenient gas-delivery systems. The pioneering work of
Motterlini and co-workers showed that transition-metal carbonyl
carriers is a very promising method due to their unique
characteristics such as high drug-loading capacity, easy formula-
tion, and low toxicity. Taking these advantages into consider-
ation, we aimed to develop a new gas-delivery system based
on a polymeric micelle.
complexes such as Ru(CO)
3
Cl(glycinate), [Ru(CO)
had high potential as CO-releasing
Among them, Ru(CO) Cl(glycinate) and its
derivatives (Ru(CO) Cl(amino acidate)) are very promising
water-soluble CO-releasing molecules. In particular, the
pharmacological actions of Ru(CO) Cl(glycinate) have been
3 2 2
Cl ] , and
4
Fe(η -2-pyrone)(CO)
molecules.
3
1
0-14
3
3
Recently, our laboratory reported a micelle-based gas-delivery
system for NO, the first gaseous species to be identified as a
1
5
2
6
3
cell signaling agent. The NO-releasing micelles have a core
composed of N-diazenium diolate (NONOate) moieties, which
decompose to liberate NO in the presence of protons. The
micelles showed remarkably prolonged NO release due to the
slow diffusion of protons into the NONOate core. This promis-
ing result prompted us to develop a second gas-delivery micelle
system for CO.
extensively studied. This compound has been shown to alleviate
damage in ischemia/reperfusion injury, inhibit allograft rejection,
suppress nitric oxide (NO) production from macrophages, and
attenuate cardiovascular inflammation and thrombin-induced
1
1,16-18
neuroinflammation.
Though the discovery of these CO-releasing molecules opens
up new possibilities, there are still several issues to overcome
for medical applications, particularly those in which downstream
tissue sites draining the injection site are targeted. These small
molecular drugs diffuse rapidly within the body after admin-
istration and may liberate CO prior to reaching these target
tissues. Thus, there is a considerable need for developing a safe
and efficient CO-delivery system.
We report here a novel CO-delivery system using a polymeric
micelle as a CO carrier. The CO-releasing micelles were
prepared from triblock copolymers composed of a hydrophilic
poly(ethylene glycol) block, a poly(Ru(CO) Cl(ornithinate acry-
3
lamide) block capable of releasing CO, and a hydrophobic
poly(n-butylacrylamide) block. The micelles were characterized
by dynamic light scattering (DLS), transmission electron
microscopy (TEM), and analytical ultracentrifugation (AUC).
The CO-release property was examined, and possible biological
molecules capable of inducing CO-release were identified. In
addition, the anti-inflammatory effect and cytotoxicity of the
CO-releasing micelles were assessed with a human monocyte-
derived cell line.
In the field of drug delivery, polymeric nanocarriers are
recognized as a powerful system for delivering a wide variety
of therapeutic agents such as hydrophobic drugs, nucleotides,
19-22
and proteins.
It has been shown that the precise size control
in the subhundred nanometer range enables targeted delivery
2
3-25
to specific tissues such as tumor tissues and lymph nodes.
In particular, the use of polymeric micelles, spherical supramo-
lecular assemblies from amphiphilic block copolymers, as drug
Results and Discussion
Synthesis of PEG-b-OrnRu-b-nBu Triblock Copolymer. Sev-
eral transition-metal carbonyl complexes have been reported as
sources of CO and have been shown to attenuate inflammation,
promote wound healing, and reduce transplant rejection. Among
(
(
(
10) Motterlini, R.; Clark, J. E.; Foresti, R.; Sarathchandra, P.; Mann, B. E.;
Green, C. J. Circ. Res. 2002, 90, E17-24.
11) Clark, J. E.; Naughton, P.; Shurey, S.; Green, C. J.; Johnson, T. R.;
Mann, B. E.; Foresti, R.; Motterlini, R. Circ. Res. 2003, 93, e2-8.
12) Sawle, P.; Hammad, J.; Fairlamb, I. J.; Moulton, B.; O’Brien, C. T.;
Lynam, J. M.; Duhme-Klair, A. K.; Foresti, R.; Motterlini, R.
J. Pharmacol. Exp. Ther. 2006, 318, 403–10.
them, Ru(CO)
substitution between [Ru(CO)
most promising CO-releasing molecules. To incorporate this
Ru(CO) Cl(amino acidate) structure into a micelle, we
designed a triblock copolymer of poly(ethylene glycol)-b-
poly[Ru(CO) Cl(ornithinate acrylamide)]-b-poly(n-butylacry-
3
Cl(glycinate), which is synthesized by ligand
3
Cl and glycine, is one of the
2 2
]
(
13) Fairlamb, I. J.; Duhme-Klair, A. K.; Lynam, J. M.; Moulton, B. E.;
O’Brien, C. T.; Sawle, P.; Hammad, J.; Motterlini, R. Bioorg. Med.
Chem. Lett. 2006, 16, 995–8.
3
(
(
(
(
(
14) Fairlamb, I. J.; Lynam, J. M.; Moulton, B. E.; Taylor, I. E.; Duhme-
Klair, A. K.; Sawle, P.; Motterlini, R. Dalton Trans. 2007, 3603–5.
15) Johnson, T. R.; Mann, B. E.; Clark, J. E.; Foresti, R.; Green, C.;
Motterlini, R. J. Inorg. Biochem. 2003, 96, 160–160.
3
lamide) (PEG-b-OrnRu-b-nBu), where PEG is the hydro-
philic block used to stabilize the micelle, OrnRu is a
Ru(CO) Cl(ornithinate) block that releases CO, and nBu is the
3
hydrophobic block, which drives micellization (Figure 1).
16) Sawle, P.; Foresti, R.; Mann, B. E.; Johnson, T. R.; Green, C. J.;
Motterlini, R. Br. J. Pharmacol. 2005, 145, 800–10.
17) Bani-Hani, M. G.; Greenstein, D.; Mann, B. E.; Green, C. J.; Motterlini,
R. J. Pharmacol. Exp. Ther. 2006, 318, 1315–22.
18) Vannacci, A.; Marzocca, C.; Giannini, L.; Mazzetti, L.; Franchi-
Micheli, S.; Failli, P.; Masini, E.; Motterlini, R.; Mannaioni, P. F.
Inflammation Res. 2006, 55, S05–6.
(23) Yokoyama, M.; Okano, T.; Sakurai, Y.; Fukushima, S.; Okamoto, K.;
Kataoka, K. J. Drug Targeting 1999, 7, 171–86.
(
(
19) Allen, T. M.; Cullis, P. R. Science 2004, 303, 1818–22.
20) Vinogradov, S. V.; Bronich, T. K.; Kabanov, A. V. AdV. Drug DeliVery
ReV. 2002, 54, 135–47.
(24) Reddy, S. T.; Rehor, A.; Schmoekel, H. G.; Hubbell, J. A.; Swartz,
M. A. J. Controlled Release 2006, 112, 26–34.
(25) Reddy, S. T.; van der Vlies, A. J.; Simeoni, E.; Angeli, V.; Randolph,
G. J.; O’Neil, C. P.; Lee, L. K.; Swartz, M. A.; Hubbell, J. A. Nat.
Biotechnol. 2007, 25, 1159–64.
(
21) Takae, S.; Miyata, K.; Oba, M.; Ishii, T.; Nishiyama, N.; Itaka, K.;
Yamasaki, Y.; Koyama, H.; Kataoka, K. J. Am. Chem. Soc. 2008,
1
30, 6001–9.
(26) Jo, Y. S.; van der Vlies, A. J.; Gantz, J.; Thacher, T. N.; Antonijevic,
S.; Cavadini, S.; Demurtas, D.; Stergiopulos, N.; Hubbell, J. A. J. Am.
Chem. Soc. 2009, 131, 14413–8.
(
22) van der Vlies, A. J.; O’Neil, C. P.; Hasegawa, U.; Hammond, N.;
Hubbell, J. A. Bioconjugate Chem. 2010, 21, 653–62.
1
8274 J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010

CO-Releasing Micelles
A R T I C L E S
Scheme 1. Synthesis of (a) Boc-Orn-OtBu AAm, (b) PEG-CTA, and (c) PEG-b-OrnRu-b-nBu Triblock Copolymers
The PEG-b-OrnRu-b-nBu triblock copolymers were prepared
as shown in Scheme 1. First, PEG-b-(Boc-Orn-OtBu)-b-nBu-
CTA triblock copolymers 4a-g were synthesized by reversible
addition-fragmentation transfer (RAFT) polymerization. The
degree of polymerization of each segment was successfully
controlled by changing the monomer/chain-transfer agent (CTA)
ratio (Tables S1 and S2, Supporting Information). Gel perme-
ation chromatography (GPC) showed unimodal size distributions
of PEG-b-(Boc-Orn-OtBu)-CTA diblock copolymers 3a-g
the radical-induced reduction reaction apparently is capable
of cleaving this linkage as shown by the disappearance in
the GPC trace. (For further discussion, see the Supporting
Information).
Next, the Boc and OtBu protecting groups were removed in
trifluoroacetic acid (TFA) and treated with sodium methoxide
(
NaOMe) to convert the Orn segment to the sodium salt form
followed by Sephadex LH-20 column chromatography purifica-
tion. The absence of the peaks corresponding to the protecting
(
Figures S6 and S7a, Supporting Information). In the case of
1
groups was confirmed by H NMR (Figure S5b, Supporting
PEG-b-(Boc-Orn-OtBu)-b-nBu-CTA triblock copolymers 4a-g,
a prominent peak with a small shoulder in the higher molecular
weight region was observed (Figure S7b). Thereafter, the CTA
end group was removed by radical-induced reduction with
Information).
Finally, the obtained PEG-b-OrnNa-b-nBu triblock copoly-
mers 6a-g were complexed with [Ru(CO) Cl ] and purified
3 2 2
by Sephadex LH-20 column chromatography. IR spectra of the
polymers showed the three characteristic bands in the carbonyl
27
1
-ethylpiperidine hypophosphite (EPHP). Successful removal
of the end group was confirmed by the absence of the peaks
1
corresponding to the pyrrole protons in the H NMR spectrum
3
vibration region similar to those of Ru(CO) Cl(glycinate) (Table
(
Figure S4b, Supporting Information). After CTA removal, the
1
and Figure S8, Supporting Information). The broadening of
shoulder peak at higher molecular weight was no longer
detectable in the GPC trace (Figure S7c). The facts that (1) the
high molecular weight peak in the GPC trace has an absorbance
around 300 nm, which is similar to the π-π* transition of the
dithioester group and (2) a monomodal molecular weight
distribution was measured by GPC after the radical-induced
reduction suggest that the high molecular weight peak consists
of two polymers linked by a dithioester-containing group and
1
the signals corresponding to the Orn block in H NMR spectra
indicates the complexation between the OrnNa block and
3 2 2
[Ru(CO) Cl ] (Figure S5b, Supporting Information). To obtain
the degree of conversion of OrnNa to OrnRu, the ruthenium
content of PEG-b-OrnRu-b-nBu polymers 7a-g was measured
by inductively coupled plasma optical emission spectroscopy
(
ICP-OES). As shown in Table 1, the conversion of the OrnNa
unit to an OrnRu unit was 58-75% possibly due to the bulkiness
3
of the Ru(CO) Cl(ornithinate) group that interferes with the full
complexation.
(
27) Chong, Y. K.; Moad, G.; Rizzardo, E.; Thang, S. H. Macromolecules
007, 40, 4446–4455.
2
J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010 18275

A R T I C L E S
Hasegawa et al.
Table 1. Characterization of PEG-b-OrnRu-b-nBu Triblock
However, the transition from spherical micelles to other
structures was not observed for the PEG-b-OrnRu-b-nBu
triblock polymers under the ratios of block sizes explored.
The micelles (7c, 7d, and 7g) were further characterized by
sedimentation velocity measurements. A typical sedimentation
coefficient distribution, c(s), and a molar mass distribution, c(M),
are shown in Figure 3 for the 7d micelle. Figure 3a shows a
prominent unimodal peak with a maximum near 20 S (90% of
Copolymers
no. of
Ru concn OrnRu units complexation
degree of Ru
c
b
a
-1
sample
ν(CO) (cm
)
(wt %)
per polymer
(%)
7
7
7
7
7
7
7
a
b
c
d
e
f
2135, 2049, 1968
2134, 2047, 1968
2134, 2047, 1968
2134, 2050, 1968
2134, 2049, 1970
2134, 2048, 1970
2134, 1046, 1968
9.2
10.0
12.9
6.9
4.7
8.9
20
24
37
18
15
27
37
75
65
70
70
58
63
72
102
-13
the total absorbance) (S ) Svedberg ) 10 s), a small amount
with lower sedimentation coefficients (sw,20) (about 6% of the
total absorbance), and about 4% at higher sw,20 values. Con-
centration independence of c(s) was confirmed by sedimentation
velocity for all three micelles studied (Table S3, Supporting
Information). After conversion of c(s) into c(M), the aggregation
number and the number of OrnRu units of the micelles were
estimated from the average molar mass value. As shown in
Table 2, the number of OrnRu units of the 7c, 7d, and 7g
micelles were 2300, 1200, and 3300 per micelle, respectively.
Assuming that each OrnRu unit releases 1 equiv of CO, the 7c,
g
10.3
34.6
Ru(CO)
3
Cl(glycinate) 2135, 2045, 1981
a
b
c
Determined by ATR-IR. Determined by ICP-OES. Degree of Ru
complexation ) [OrnRu]/([OrnNa] + [OrnRu]).
Table 2. Characterization of PEG-b-OrnRu-b-nBu Micelles
hydrodynamic
diameter
a
molar mass
partial specific of the micelle number (no. of OrnRu units
aggregation
no. of
c
z average
(nm)
b
PDI volume (mL/g)
sample
(g/mol)
chains/micelle) per micelle
7
d, and 7g micelles can deliver 2300, 1200, and 3300 equiv of
7
7
7
7
7
7
7
a
b
c
d
e
f
29
33
32
37
41
44
39
0.19
0.21
CO per micelle.
6
6
3
0.19
0.15
0.11
0.14
0.12
0.7893
0.8536
1.82 × 10
1.79 × 10
63
66
2.3 × 10
1.2 × 10
The spherical nature of the micelles was confirmed by
UltraScan two-dimensional spectrum analysis (2DSA) and
subsequent Monte Carlo analysis (2DSA-MC), yielding a
frictional ratio (f/f
of the frictional ratios (f/f
concentrations of the 7d micelles (Figure S10, Supporting
Information). Moreover, the sedimentation coefficient range of
the main fraction of the 7d micelles is almost identical with
the range obtained by SEDFIT and shown in Figure 3.
3
6
3
g
0.8568
3.14 × 10
88
3.3 × 10
0
) of about 1.1 for the main fraction. The plots
) versus sw,20 are presented for three
a
2
0
z average and polydispersity index (PDI ) µ /Γ ) calculated by the
2
b
c
cumulant method (DLS). Determined by densitometry. Weight
average of c(M) calculated by SEDFIT.
Micelle Formation from PEG-b-OrnRu-b-nBu Triblock
Polymers. To prepare CO-releasing micelles, we first designed
PEG-b-OrnRu diblock copolymers assuming that the OrnRu
block would be hydrophobic enough to drive micelle formation.
However, in contrast to our expectation, the diblock copolymers
formed polydisperse aggregates larger than 200 nm (data not
shown). This might be due to the bulkiness of the ruthenium
carbonyl group, which may disfavor formation of a compact
structure. Moreover, its complicated solution chemistry as well
as the remaining OrnNa residues may make the OrnRu segment
less hydrophobic than expected. This initial observation,
therefore, led us to design a new block copolymer having a
hydrophobic third segment, an nBu block, for driving micelle
formation.
CO-Release Property of the Micelles. It is believed that
Ru(CO) Cl(glycinate) releases CO in the body via substitution
3
of chloride or glycinate ligands with electron-withdrawing
30
ligands such as thiol compounds and imidazole. Although the
precise mechanism is not known, the successful results observed
in vitro and in vivo also imply that Ru(CO)
releases a sufficient amount of CO in the physiological milieu.
3
Cl(glycinate)
31
To quantify the amount of CO release in aqueous medium, the
spectrophotometric change of deoxymyoglobin (Mb) to carboxy-
28
1
0
Mb (MbCO) has often been used. In this assay, the sample
solution is added to the Mb solution containing sodium dithionite
(
2 2 4
Na S O ) as a reducing agent, and then the formation of MbCO
is quantified. It has been confirmed that Ru(CO) Cl(glycinate)
3
1
1
liberates 1 equiv of CO with a half-life of ∼1 min. Thus, we
first used the Mb assay to examine the CO-releasing capacity
of the 7g micelles. The micelles released 0.76 equiv of CO per
DLS showed that the triblock copolymers formed monodis-
perse micelles about 30 nm in diameter for the polymers with
5
kDa PEG (7a, 7b, 7c) and about 40 nm in diameter for the
3
OrnRu unit, whereas Ru(CO) Cl(glycinate) released 0.90 equiv
polymers with 11 kDa PEG (7d, 7e, 7f, and 7g) (Table 2). The
micelles had a high colloidal stability in distilled water, and
the aggregation was not observed for 1 week at both room
temperature and 4 °C (Figure S9, Supporting Information). TEM
images showed that all triblock copolymers formed spherical
micelles (Figure 2). Generally, block polymers with a large
hydrophilic segment form spherical micelles, whereas those with
a smaller hydrophilic segment tend to form wormlike micelles
of CO (Figure S11b, Supporting Information). This slightly
lower CO release from the micelles can be attributed to the
micelle structure, which keeps some Ru(CO) Cl(ornithinate)
3
moieties intact inside the micelles. Considering that the 7g
micelles have 3300 OrnRu units according to the AUC analysis,
the CO-loading capacity of the micelles is calculated to be 2500
equiv of CO per micelle. This high loading will be advantageous
to enhance the therapeutic efficiency of CO.
Though we confirmed that the micelles are able to release
CO using the Mb assay, the CO-release mechanism under the
physiological conditions is still not clear. In particular, further
investigation is needed to indentify molecules capable of
29
or vesicles. We expected to observe morphological changes
for the polymers composed of short hydrophilic blocks (PEG)
and long hydrophobic blocks (OrnRu and nBu), such as 7c.
(
(
28) Johnson, T. R.; Mann, B. E.; Teasdale, I. P.; Adams, H.; Foresti, R.;
Green, C. J.; Motterlini, R. Dalton Trans. 2007, 1500–8.
(30) Alberto, R.; Motterlini, R. Dalton Trans. 2007, 1651–60.
(31) Johnson, T. R.; Mann, B. E.; Clark, J. E.; Foresti, R.; Green, C. J.;
Motterlini, R. Angew. Chem., Int. Ed. 2003, 42, 3722–9.
29) Forster, S.; Plantenberg, T. Angew. Chem., Int. Ed. 2002, 41, 689–
7
14.
1
8276 J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010

CO-Releasing Micelles
A R T I C L E S
Figure 2. TEM images of PEG-b-OrnRu-b-nBu micelle formulations: (a) 7a, (b) 7b, (c) 7c, (d) 7d (e) 7e, (e) 7g. Samples were negatively stained with 2
wt % uranyl acetate(aq).
Figure 3. Typical (a) sedimentation coefficient distribution and (b) molar mass distribution of PEG-b-OrnRu-b-nBu 7d micelles (sedimentation velocity at
3
0 000 rpm, 20 °C, λ ) 270 nm, micelle concentration 1 mg/mL).
inducing CO release in the body. Since Na
in the body and Mb mainly localizes in muscle tissues, it is
not appropriate to consider the data obtained by this method as
2
S
2
O
4
does not exist
gas constant, T is the temperature, c is the CO concentration in
the liquid phase, and k is Henry’s law constant of CO in water
(1052.63 (L ·atm)/mol at 25 °C).
3
2
a real CO-release profile in the body. Furthermore, Na
2
S
2
O
4
Since thiol compounds such as cysteine, glutathione, and
3
3
induces CO liberation of Ru(CO) Cl(glycinate) (see Figure S13c
3
proteins with free cysteine residues are abundant inside cells,
in the Supporting Information). Therefore, we sought to find a
method to measure CO release without any additives in
physiologically relevant conditions.
The experiments were carried out in a desiccator equipped
with a CO detector and a test tube for loading a sample solution
these molecules may be potential candidates to induce CO
release in the body. Therefore, we focused on the effects of
cysteine on CO release from Ru(CO)
releasing micelles. As shown in Figure 4a, CO release by
Ru(CO) Cl(glycinate) was dependent on the cysteine concentra-
3
Cl(glycinate) and the CO-
3
(
Figure S12, Supporting Information). Assuming that the gas
tions. In the presence of 10 mM cysteine, the amount of released
CO increased to 0.12 equiv of CO per molecule within 1.5 h,
whereas almost no CO release was observed in the presence of
0.1 mM cysteine for 2 h. Moreover, after incubation of
and liquid phases reach equilibrium and the pressure in the
desiccator stays at 1 atm, one can calculate the amount of
released CO (NCO) using the following equation:
3
Ru(CO) Cl(glycinate) with 1 mM cysteine, the addition of 10
mM cysteine induced additional CO release (Figure S13a,
Supporting Information). These data clearly show that cysteine
is able to interact with Ru(CO) Cl(glycinate) and subsequently
3
induces CO release. The micelles showed a similar CO-
release profile but with slower release rates than that of
pVg
RT
Vg
Vl
k
NCO
)
+ cV ) p
+
(1)
l
(
)
RT
where p is the partial pressure of CO, V
of the gas phase (250 mL) and liquid phase (5 mL), R is the
g
and V
l
are the volume
(
32) Springer, B. A.; Sligar, S. G.; Olson, J. S.; Phillips, G. N. Chem. ReV.
994, 94, 699–714.
(33) Saito, G.; Swanson, J. A.; Lee, K. D. AdV. Drug DeliVery ReV. 2003,
55, 199–215.
1
J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010 18277

A R T I C L E S
Hasegawa et al.
Figure 4. CO release of Ru(CO)
3
Cl(glycinate) and the 7f CO-releasing micelles as a function of time. CO release of (a) 1 mM Ru(CO)
3
Cl(glycinate) and
(
b) 1 mM Ru(CO)
cysteine. (c) Stability of 1 mM Ru(CO)
0% FBS (blue). (d) Stability of 1 mM Ru(CO)
cysteine after 60 min of incubation with 80% FBS (blue).
3
Cl(ornithinate) moieties of the micelles in 50 mM PBS (pH7.4) in the presence of 10 (blue), 1 (red), 0.1 (green), and 0 (yellow) mM
Cl(glycinate) in 80% FBS (yellow) and CO release upon addition of 10 mM cysteine after 60 min of incubation with
Cl(ornithinate) moieties of the micelles in 80% FBS (yellow) and CO release upon addition of 10 mM
3
8
3
Ru(CO)
hindrance of the PEG corona, which hampers the diffusion of
cysteine to the Ru(CO) Cl(ornithinate) moieties inside the
3
Cl(glycinate). This slow release may be due to the steric
blood plasma. In addition, the cytoplasm contains glutathione
3
8
levels in the millimolar range (0.5-10 mM). Therefore, it is
3
likely that Ru(CO) Cl(glycinate) and the micelles release CO
3
micelles. We also assessed the effect of glutathione, another
major thiol compound in the body, on the CO release of
primarily inside cells and to a lesser extent within the extra-
cellular space.
3
Ru(CO) Cl(glycinate) (Figure S13b). As was observed for
Inhibition of LPS-Induced Inflammation. The anti-inflam-
matory effect of Ru(CO) Cl(glycinate) and the CO-releasing
3
cysteine, CO release was detected upon the addition of 10 mM
glutathione. These data suggest that the compounds bearing thiol
moieties are molecular cues to induce CO release from
micelles was assessed using THP-1 Blue cells, which are derived
from human monocyte THP-1 cells. This cell line is stably
transfected with a reporter plasmid expressing a secreted
embryonic alkaline phosphatase (SEAP) gene under the control
of a promoter which is inducible by the pro-inflammatory
transcription factor NF-κB. Upon the addition of LPS or other
Toll-like receptor ligands, THP-1 Blue cells secrete SEAP via
NF-κB activation. Thus, NF-κB activation can be assessed by
measuring the SEAP levels in the medium. The cells were
3
Ru(CO) Cl(glycinate) and micelles.
Despite the low concentration of cysteine (∼8 µM) and
3
4,35
glutathione (∼2 µM) in human serum,
the presence of other serum components may induce CO release
or inactivate Ru(CO) Cl(glycinate) and the micelles. For this
reason, we tested their stability in 80% fetal bovine serum (FBS).
As can be seen in Figure 4c,d, both Ru(CO) Cl(glycinate)
it is anticipated that
3
3
and the micelles did not liberate CO for 2 h and the amount
of released CO was negligible even after 3 h. However,
upon addition of 10 mM cysteine, CO release of both
cultured in the medium containing Ru(CO)
CO-releasing micelles for 1 h followed by the addition of LPS
0.1 µg/mL). After 18 h of culture, the SEAP levels in the
3
Cl(glycinate) or the
(
Ru(CO)
3
Cl(glycinate) and the micelles was observed after
Cl(glycinate)
medium were analyzed. As shown in Figure 5a, the micelles
suppressed the LPS-induced NF-κB activation in a dose-
dependent manner. In particular, the addition of a 200 µM
incubation in serum for 1 h. Therefore, Ru(CO)
3
and the micelles are stable in serum and release CO when
exposed to high concentrations of cysteine. This thiol-dependent
CO release is interesting in terms of intracellular delivery of
CO. It has been reported that a high concentration of cysteine
exists in the endosome/lysosome compartment of mammalian
3
concentration (concentration of Ru(CO) Cl(ornithinate) moi-
eties) of the micelles substantially suppressed the increase of
the SEAP levels. Importantly, the addition of the micelles alone
did not alter the SEAP levels. Since serum does not trigger CO
release from the micelles (Figure 4d), this anti-inflammatory
effect may only be attributed to CO release after cellular uptake
of the micelles (see the Supporting Information, section S9).
This release mechanism is presumably induced by cysteine in
3
6,37
cell lines,
whereas much lower cysteine levels are found in
(
34) Jones, D. P.; Carlson, J. L.; Samiec, P. S.; Sternberg, P., Jr.; Mody,
V. C., Jr.; Reed, R. L.; Brown, L. A. Clin. Chim. Acta 1998, 275,
1
75–84.
(
35) Jones, D. P.; Carlson, J. L.; Mody, V. C.; Cai, J.; Lynn, M. J.;
Sternberg, P. Free Radical Biol. Med. 2000, 28, 625–35.
36) Lloyd, J. B. Biochem. J. 1986, 237, 271–2.
the endo/lysosome. In contrast, Ru(CO)
increased the SEAP level after LPS stimulation. Surprisingly,
the addition of Ru(CO) Cl(glycinate) alone also activated the
NF-κB pathway. It is likely that the anti-inflammatory effects
3
Cl(glycinate) strongly
(
(
37) Pisoni, R. L.; Acker, T. L.; Lisowski, K. M.; Lemons, R. M.; Thoene,
J. G. J. Cell Biol. 1990, 110, 327–35.
3
1
8278 J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010

CO-Releasing Micelles
A R T I C L E S
Figure 5. Biological activity of Ru(CO)
releasing micelles 7g. THP-1 Blue cells were preincubated with Ru(CO)
3
Cl(glycinate) and CO-releasing micelles in vitro. (a) Anti-inflammatory effect of Ru(CO)
Cl(glycinate) or the micelles (Ru(CO)
and 200 µM) for 1 h and cultured for 18 h in the presence (0) or absence (9) of 0.1 µg/mL LPS. The activation of the transcription factors was assessed
by SEAP levels in the media. (b) Cytotoxicity of Ru(CO) Cl(glycinate) (2) and the 7g CO-releasing micelles (O). THP-1 Blue cells were incubated with
Ru(CO) Cl(glycinate) or the micelles for 3 days. The cell viability was assessed by an MTT assay.
3
Cl(glycinate) and CO-
3
Cl(amino acidate) concentration 10, 50, 100,
3
3
3
of Ru(CO)
3
Cl(glycinate) were ablated by other side effects
amounts of CO in response to biological stimuli. Nevertheless,
as shown in this study, these compounds and/or their byproducts
can have adverse effects. In contrast, CO-releasing micelles can
deliver CO with much less toxicity. Therefore, the CO-releasing
micelles could provide a safer CO-delivery system for immu-
notherapy as well as a useful platform for basic research to
elucidate the physiological functions of CO.
caused by the Ru compound, but the mechanims of these effects
are unknown.
It has been reported that inhalation of exogenous CO gas (250
ppm in the air supply) attenuates the LPS-induced NF-κB
activation and the pro-inflammatory cytokine production of
mouse macrophages and human monocytes.
ment of human dendritic cells with 30 µM [Ru(CO)
another CO-releasing molecule, did not show any effect on the
3
9
5,40
However, treat-
Cl
3
2
]
2
,
In addition, it is of importance for immunotherapeutic
applications to delivery of immunomodulatory drugs to lymph
nodes, where a substantial fraction of immature dendritic cells
NF-κB activation despite the significant suppression of cytokine
7
42
expression. This confusing result of the CO-releasing molecule
and macrophages exists. It has been well-established that
may be attributable to side effects of the Ru compound, although
it is still possible that CO shows different actions depending
on the cell type.
particle size and surface chemistry are the most crucial factors
43
for lymphatic uptake from the interstitial space. We previously
reported that PEGylated poly(propylene sulfide) nanoparticles
in the range of 20-45 nm in diameter were efficiently delivered
Since the NF-κB pathway is known to respond to diverse
4
1
24
danger stimuli including cellular injury and stress signals, we
to the draining lymph nodes following intradermal injection.
hypothesized that the strong NF-κB activation observed in the
presence of Ru(CO) Cl(glycinate) was induced by its cytotox-
3
icity. To demonstrate this, we performed an MTT assay for
metabolic activity to assess cell viability in the presence of
Thus, the use of drug-delivery vehicles in this size range is a
promising way to deliver drugs to lymph nodes without using
any targeting ligands. The CO-releasing micelles are 30-40
nm in diameter, which is sufficiently small to allow lymph node
targeting. Thus, CO-releasing species of colloidal dimensions,
rather than low molecular weight CO-releasing molecules as
Ru(CO)
pected, the viability of the cells decreased upon the addition
of Ru(CO)
Cl(glycinate) (IC50 ≈ 600 µM); however, the
micelles did not show any toxic effects up to 1 mM
Ru(CO) Cl(ornithinate) moieties (Figure 5b). These data clearly
show that Ru(CO) Cl(glycinate) or its byproducts after CO
3
Cl(glycinate) or the CO-releasing micelles. As ex-
1
0-14
amply described in the literature,
may be promising for
3
CO release in lymph nodes draining an injection site for
immunotherapy.
3
3
liberation are toxic to the THP-1 Blue cells and subsequently
Conclusion
induced the NF-κB activation. In the case of the micelles, the
toxic Ru(CO) Cl(ornithinate) moieties are hidden by a PEG
3
corona. Thus, the stealth feature of PEG appears to inhibit
unfavorable cellular responses.
Since the use of CO gas has potential risks in clinical
applications, including the increase of carboxyhemoglobin
levels, there is a need to develop a more efficient and convenient
delivery system. The transition-metal carbonyl complexes have
emerged as promising CO-releasing molecules to release known
The use of CO for medical applications is challenging due
to the difficulties in controlled delivery of this gas. To develop
an administration method suitable for therapeutic applications,
numerous interesting prodrugs of CO have been reported.
Despite the successful control of CO-releasing properties by a
chemical approach, rapid diffusion of these small molecules after
administration remains a fundamental problem in controlled
delivery for some applications, including ours, namely, targeting
immune cells in lymph nodes that drain the injection site. We
describe herein a novel approach to deliver CO using micelles
(
(
(
(
38) Meister, A.; Anderson, M. E. Annu. ReV. Biochem. 1983, 52, 711–
0.
39) Korashy, H. M.; El-Kadi, A. O. Free Radical Biol. Med. 2008, 44,
6
(42) Wilson, N. S.; El-Sukkari, D.; Belz, G. T.; Smith, C. M.; Steptoe,
R. J.; Heath, W. R.; Shortman, K.; Villadangos, J. A. Blood 2003,
102, 2187–94.
7
95–806.
40) Chhikara, M.; Wang, S.; Kern, S. J.; Ferreyra, G. A.; Barb, J. J.;
Munson, P. J.; Danner, R. L. PLoS One 2009, 4, e8139.
(43) Reddy, S. T.; Swartz, M. A.; Hubbell, J. A. Trends Immunol. 2006,
27, 573–9.
41) Tripathi, P.; Aggarwal, A. Curr. Sci. 2006, 90, 519–531.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010 18279

A R T I C L E S
Hasegawa et al.
as a CO donor. The amphiphilic PEG-b-OrnRu-b-nBu triblock
copolymers formed monodisperse CO-releasing micelles in the
range of 30-40 nm in diameter. CO release was induced by
thiol compounds, including cysteine. The micelles successfully
attenuated the LPS-induced inflammatory response of human
Dr. Karen Dane and Mr. Stephane Kontos (EPFL) for critical
reading of the paper. This work has benefited from research funding
from the European Community’s Seventh Framework Programme
via Project FP7-Health-2009 ENDOSTEM 241440 (activation of
vasculature associated stem cells and muscle stem cells for the repair
and maintenance of muscle tissue).
3
monocytes. In addition, the toxicity of Ru(CO) Cl(amino
acidate) moieties was significantly reduced by the stealth feature
of PEG. These results show that the CO-releasing micelles can
provide an efficient and safe administration method for CO-
based therapy.
1
Supporting Information Available: Experimental Section, H
NMR spectra, GPC charts, IR spectra, colloidal stability of the
micelles, UltraScan two-dimensional spectrum analysis (2DSA),
Monte Carlo analysis (2DSA-MC), CO-release capacity of the
micelles by the Mb assay, experimental setup for CO release
measurement, CO release profiles in the presence of cysteine,
glutathione, and Na S O , and cellular uptake of the micelles.
2 2 4
This material is available free of charge via the Internet at http://
pubs.acs.org.
Acknowledgment. We thank Dr. Michael Groessl (Laboratory
of Organometallic and Medicinal Chemistry, EPFL) for his help
and advice in the ICP-OES experiment. We thank P. Schuck (NIH,
Bethesda) for providing the SEDFIT software and B. Demeler
(
Health Science Center at San Antonio, University of Texas) for
providing the UltraScan software and free access to the supercom-
puting resources through NSF Teragrid Science Gateway. We thank
JA1075025
1
8280 J. AM. CHEM. SOC. 9 VOL. 132, NO. 51, 2010