Angewandte
Communications
Chemie
of lactam/lactones of various ring sizes. CO release rate for
whether these compounds would function in a biological
BW-CO-101 is the fastest in this series (t1/2, 1.9 min) due to
formation of a five-membered lactam ring after CO release.
Replacement of the alkyne terminal hydrogen with a bulky
TBDPS group (BW-CO-102) did not make any difference in
terms of CO release rate. Such results further suggest the
important role of entropic factors in the rate acceleration.
As to balancing stability and ready reactivity under
physiological conditions, Breslow and others[21] have long
recognized that water and glycoprotein[22] can significantly
accelerate Diels–Alder reactions. Much of this is attributed to
the hydrophobic interaction between the reactants in aqueous
solution. Thus, as expected, the CO release rate of BW-CO-
101 is much faster in aqueous solutions (e.g. DMSO/PBS)
than in organic solvents (e.g. CH2Cl2) (Figure S33), which
afforded the necessary stability during synthesis and storage.
Compared to five-membered ring formation, six-membered
lactam ring formation is less favorable;[23] As a result, CO
release rate for BW-CO-103 is much slower (t1/2, 1.2 h)
compared to that of BW-CO-101/102 (Table 1). Substituting
the terminal alkyne hydrogen for a methyl group in BW-CO-
103 results in an internal alkyne (BW-CO-104) and reduces
the CO release rate by about 4-fold (t1/2, 6.2 h). Comparing the
reaction rate between lactone formation and lactam forma-
tion, one can see that the CO release rate from BW-CO-105,
which leads to lactone formation, is extremely slow with
a half-life of more than one week (as compared to 1.9 min for
BW-CO-101), which are consistent with literature prece-
dents.[24] Interestingly, BW-CO-106, which is analogous to
BW-CO-101 without the N-methyl group, showed extremely
slow CO release rate (t1/2 > one week). We hypothesized that
this slow reaction could be due to intramolecular hydrogen
bond formation between the amide hydrogen and the cyclo-
pentadienone carbonyl group (Figure S41), which could lead
to a conformation unfavorable for the intended cycloaddition.
NMR experiments indicate that the NH proton is at 7.75 ppm
in DMSO and 8.02 ppm in CDCl3. Such results are consistent
with weakened intramolecular hydrogen bond in DMSO
(Figures S41–S43).
system by delivering CO in sufficient quantity to achieve the
known biological effects reported by others. Therefore, BW-
CO-102- 104, and 107 with different release rates were chosen
to test their inhibitory effect against lipopolysaccharide (LPS)
induced TNF-a secretion in RAW 264.7 cells. Initially, we
tested the cytotoxicity of CO prodrugs along with their
inactive products in RAW 264.7 cells. The results showed that
all the tested compounds did not affect cell viability at up to
100 mm after 24 h incubation (Figure S44), except for BW-CO/
CP-107, which did not show any effect on cell viability at up to
50 mm. Clearly, as a class of compounds, toxicity is not an issue,
but idiosyncratic toxicity may occur with individual com-
pounds. For the TNF-a inhibition assay, the RAW264.7 cells
were pretreated with CO prodrugs or their inactive products
for 5 h. This was followed by the treatment of LPS for another
1 h. Then the cell culture supernatants were subjected to
ELISA assay for TNF-a levels.
Shown in Figure 2 are the ELISA assay results. BW-CO-
103, 104 and 107 dose-dependently inhibited LPS-induced
TNF-a secretion. BW-CO-102 also suppressed TNF-a secre-
tion, but it did not show any dose-dependent effect, which
might be ascribed to the fast CO release kinetics of BW-CO-
102 (t1/2 = 1.8 min). The CO release from BW-CO-102 was
probably finished before it entered into the cells. CO is known
to be volatile and hardly soluble in cell culture medium; thus
the balance between volatility and rapid diffusion into cell is
one factor that has to be considered, especially in an open cell
culture system. Consequently, the intracellular CO concen-
tration might be similar among different concentration
groups. Meanwhile, the inactive products BW-CP-102–104
did not show any TNF-a suppression effects. Therefore, it is
clear that the TNF-a inhibition effect observed for BW-CO-
102–104 is associated with the CO released from these
prodrugs. Interestingly, the inactive compound BW-CP-107
also dose-dependently inhibited TNF-a secretion, and hence
the observed TNF-a inhibition effect for BW-CO-107 is an
additive effect between CO release and the inactive product
BW-CP-107.
With the hydrophobic nature of these prodrugs, we were
interested in finding ways to improve water solubility. Thus,
BW-CO-107 with glucose conjugation was synthesized. The
introduction of the glucose moiety slightly decreased (by
around 1 fold, t1/2, 2.1 h) the CO release rate compared to
BW-CO-103 (t1/2, 1.2 h). Meanwhile, the improved water
solubility of BW-CO-107 allowed for the determination of
CO release rate in a more hydrophilic environment (1% of
DMSO in PBS). As expected, the CO release rate increased
by around 11 fold (t1/2, 0.18 h) in 99% aqueous solution as
compared to the release rate in DMSO/PBS (5:1, t1/2, 2.1 h).
These results further highlight the cycloaddition reaction rate
acceleration by water.
Although we observed the anticipated anti-inflammatory
effect, it is also important to confirm CO release. The strong
blue fluorescence of the cycloaddition products greatly
facilitates the real time monitoring of CO release. For this
purpose, BW-CO-103 was chosen for imaging studies in Raw
264.7 cells. As shown in Figures S45 and S46, the intensity of
the fluorescence increased in dose- and time-dependent
fashions. As an additional confirmation of CO release under
the experimental conditions used, we also used a literature
fluorescent CO probe (COP-1)[25] to visulize intracelluar CO
release from BW-CO-103, which has “turn-on” fluorescence
upon CO detection (Figure S47).
Having confirmed the CO prodrugsꢀ definitive CO-
associated anti-inflammatory effects in Raw 264.7 cells, we
next probed whether they would have the general properties
needed for in vivo application. Cigarette smoking has been
shown to be protective against the development of ulcerative
colitis,[26] and CO, as one major component of cigarette
smoke, has been reported to ameliorate intestinal inflamma-
tion in various colitis models.[27] Therefore, BW-CO-103 was
As shown in Table 1, there are numerous possibilities for
structural modifications (e.g. n, R1, R2, and X groups) to serve
different purposes including tuning CO release rates, and
improving pharmacokinetics profiles and water solubility,
among others.
With the success in preparing the prodrugs and seeing CO
release with tunable rates, next we were interested in seeing
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!