Ketorolac beats ketoprofen: Lower photodecarboxylation, photohemolysis and phototoxicity

Article abstract of DOI:10.1039/c3md00258f

Ketorolac shows reduced photohemolytic activity and low phototoxicity against human skin fibroblasts when compared to ketoprofen. The low decarboxylation quantum yield together with the efficient non-radiative deactivation of the triplet and singlet excit

Full text of DOI:10.1039/c3md00258f

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Ketorolac beats ketoprofen: lower
photodecarboxylation, photohemolysis and
phototoxicity†
Cite this: Med. Chem. Commun., 2013,
4, 1619
a
a
b
´
´
Christopher D. McTiernan, Chiara Fasciani, Maria Gonzalez-Bejar, Daniel Roca-
b
a
ac
*
*
´
Sanjuan, Emilio I. Alarcon and Jose Carlos Netto-Ferreira
Received 10th September 2013
Accepted 16th October 2013
Ketorolac shows reduced photohemolytic activity and low phototoxicity against human skin fibroblasts
when compared to ketoprofen. The low decarboxylation quantum yield together with the efficient non-
radiative deactivation of the triplet and singlet excited states of ketorolac are believed to be responsible
for this behaviour.
DOI: 10.1039/c3md00258f
www.rsc.org/medchemcomm
The presence of a benzophenone-like moiety in many phar-
Interestingly, when red blood cell suspensions (z3.3 ꢀ 106
maceutical compounds, including non-steroidal anti-inama- cells per ml) containing 10, 30, or 60 mM ketoprofen, ketorolac,
tory drugs (NSAID), has been related to their phototoxicity due or their corresponding decarboxylated photoproducts are
to the generation of reactive intermediates upon drug photo- exposed to UVA light for 15 min varying degrees of photo-
decomposition.1–3 For example, the main pathway for ketopro- hemolytic activity are observed (see Fig. 1). For example, while
fen photodegradation upon UVA light exposure, at physiological ketoprofen shows photohemolytic activity in the whole range of
pH, is photodecarboxylation.1 Such decarboxylation has been concentrations, ketorolac displays only a 20% hemolysis at its
related to the generation of free radicals and the formation of highest concentration. Noticeably, the ketoprofen photo-
hemolytic holes.4 Further, formation of methemoglobin free product showed a dose dependent behaviour in its hemolytic
radicals has been also reported upon UVA exposure of erythro- activity. In contrast, the ketorolac photoproduct did not show
cyte suspensions containing ketoprofen.5 Thus, both processes any toxicity even at the highest concentration tested.
are believed to play key roles in the photohemolysis of human
In addition, phototoxicity experiments carried out on skin
erythrocytes mediated by ketoprofen. Ketorolac, on the other broblasts cultures revealed non-toxicity of ketorolac and rela-
hand, is a powerful analgesic commonly administrated aer tive phototoxicities for the other compounds as follows: keto-
dental procedures, including root canals and tooth extractions, profen photoproduct > ketoprofen z ketorolac photoproduct,
as well as in eye-drops for glaucoma pain control.6 Although this see Fig. 2. Although a direct comparison between the hemolysis
NSAID is commonly prescribed and shares a benzoyl group with and broblasts toxicity data is not possible due to differences in
ketoprofen, see Scheme 1, there have been documented studies uptake and metabolism between the two cell lines, our cumu-
on its toxicity in humans, see for example ref. 7 & 8, and little is lative data suggests that ketorolac is not toxic upon UVA expo-
known of its photophysics and photochemistry. The following sure. Differences in the decarboxylation quantum yield between
presents our ndings on the photohemolysis, cell phototoxicity, ketorolac and ketoprofen are likely responsible for the differ-
photophysics, and photochemistry for ketorolac.
ences in toxicity data.
In fact a decarboxylation quantum yield as low as 8.0 ꢀ 10ꢁ5
was measured for ketorolac (Fig. 3). This value is almost ve
orders of magnitude smaller than that reported for ketoprofen.1
However, the reason for the difference in the photoreactivity of
these two compounds is not yet clear. Therefore, further pho-
tophysical and photochemical characterization of ketorolac was
carried out.
aDepartment of Chemistry and Centre for Catalysis Research and Innovation,
University of Ottawa, 10 Marie Curie, Ottawa K1N 6N5, Canada. E-mail: emilio@
photo.chem.uottawa.ca; josecarlos@photo.chem.uottawa.ca
b
´
´
´
Instituto de Ciencia Molecular, Universidad de Valencia, C/Catedratico Jose Beltran,
2, Paterna, 46980, Valencia, Spain
c
´
´
Departamento de Quımica, Universidade Federal Rural do Rio de Janeiro, Seropedica,
23851-970, Rio de Janeiro, Brazil
The absorption spectrum of ketorolac presents a high
intensity band around 300 nm (Fig. S1†). According to theo-
retical CASPT2 computations (see ESI; Fig. S2–S4 and Tables S1–
S4†), the lowest-energy band can be ascribed to an excited state
featuring a transfer of electron density from the pyrrole moiety
to the keto group (see Fig. 4).
† Electronic supplementary information (ESI) available: Materials, syntheses and
characterization, theoretical methods and analyses, phototoxicity experiments
carried out on skin broblasts cultures, ketorolac UV-Visible spectra in
acetonitrile or in phosphate buffer, 1/kobs at 620 nm as function of 1-methyl
naphthalene, singlet oxygen phosphorescence, phosphorescence. See DOI:
10.1039/c3md00258f
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1619–1622 | 1619

View Article Online
MedChemComm
Concise Article
Scheme 1 Chemical structures of Ketorolac, Ketoprofen, and their corresponding decarboxylated photoproducts identified as “photo”.
Fig. 3 Time evolution profile for the photodecomposition of 10 mM ketorolac to
afford its corresponding decarboxylated photoproduct, see experimental. All
measurements were carried out in nitrogen-saturated solutions at RT.
Fig. 1 Photohemolytic activity for ketorolac, ketoprofen and their respective
decarboxylation products (Scheme 1).
Fig. 4 CASSCF electron-density difference between the ground and the lowest-
lying excited singlet states at the CASSCF optimized gas-phase structure of the
later. Regions with an excess and deficiency of electron density are in yellow and
blue, respectively.
carried out using 1-methyl naphthalene rendered a rate
constant of 1.5 ꢀ 109 Mꢁ1 sꢁ1, fully compatible with an energy
transfer mechanism9,10 (Fig. S5).† Thus, the 620 nm transient
species was assigned to the ketorolac triplet with a quantum
yield of only 0.18. Also a singlet oxygen quantum yield of 0.06
was obtained when ketorolac was employed as sensitizer, cor-
responding to a singlet oxygen efficiency of z0.33 (Fig. S6†).
This value is in full agreement with those observed for singlet
oxygen efficiency of other 3n,p* systems such as acetophe-
none.11 Further, low temperature phosphorescence measure-
ments indicate that the ketorolac triplet state shows a 3n,p*
Fig.
2
Phototoxicity experiments carried out on skin fibroblasts. Results
obtained using an MTS viability assay. The parental drugs and their corresponding
decarboxylated photoproduct were added as 1 mM solutions in FCS supple-
mented DMEM containing 10% DMSO. See SI for complete details (plate P6).
A red-shi (14 nm) is observed in aqueous solutions proving character with 2.73 eV energy; Fig. S7.† A similar conclusion is
the polarizability of the singlet state (Fig. S1†). This agrees with drawn from the bimolecular rate constants measured using
its charge transfer nature and the theoretical analysis of solvent several triplet quencher molecules (Table 1) and CASPT2 results
3
effects (see ESI†). Moreover, no uorescence emission was (even though the presence of a triplet p,p* state is also esti-
detected for this compound in acetonitrile or buffer solution.
mated at close energies and might contribute to the band at
Upon laser excitation (266 nm) of deareated ketorolac in higher wavelengths, Tables S5–S6†).
acetonitrile, the presence of two intermediate species was
Interestingly, activation energies for the ketorolac triplet
observed (Fig. 5). The rst one has a lifetime of 2.5 ms with state as low as 1.0 and 1.5 kcal molꢁ1 for acetonitrile and PBS
absorption maxima at 350 and 620 nm, and is readily quenched buffer, respectively, were obtained from Arrhenius plots
by oxygen (2.0 ꢀ 109 Mꢁ1 sꢁ1) (Table 1). Quenching experiments (Fig. S8†), with pre-exponential factors in the order of z108 sꢁ1
.
1620 | Med. Chem. Commun., 2013, 4, 1619–1622
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
Theoretical analyses point out to the population of the anti-
bonding p* orbital of the ketone group of ketorolac in the lowest-
lying singlet and triplet excited states. As a consequence, the
3
relaxation of the n,p* state to the corresponding equilibrium
structure results in an enlargement of the CO bond distance from
˚
1.22 to 1.36 A (Table S6†), thus increasing the basicity of the keto
group as explained elsewhere.12,13
The formation of an enol-like intermediate is in good
agreement with the reported oxidation products in litera-
ture.14,15 The reason why the enol intermediate for ketorolac
appears in time scales where the triplet is still decaying (Fig. 5B)
is probably due to the presence of mixing states for the ketor-
olac triplet as revealed from computational calculations (Tables
S5 and S6†). In this case only the short-lived 3n,p* would lead to
the enol formation.
Finally, we have addressed how the ketorolac triplet-
excited state is competitively deactivated but have not dis-
cussed its lack of uorescence as well as its low triplet
quantum yield. From our modest computational approach,
we expect that the corresponding radiationless decay mech-
anism will be initially driven by the intra-molecular charge
transfer, ICT, nature of the singlet-excited state (see Fig. 4).
An enlargement of the CO bond distance and a rearrange-
ment of the relative orientation between the benzoyl and
pyrrole moieties might bring the system to an internal
conversion process, as it happens in other molecules such as
the cytosine DNA base.16,17 More advanced computational
strategies shall, however, be required in future studies to
accurately determine the deactivation channel of the bright
state of ketorolac.
Fig. 5 (A) Ketorolac (50 mM) transient absorption spectra at different times after
266 nm laser excitation. (B) 370 nm and 620 nm decay profiles at short time scale.
(C) Inset: 370 nm transient time profile.
Table 1 Bimolecular quenching rate constants for Ketorolac triplet state with
different quenchers
Quencher
kq (Mꢁ1 sꢁ1) ꢀ 106
Oxygen
2000 ꢂ 100
4.80 ꢂ 0.10
0.47 ꢂ 0.01
0.65 ꢂ 0.05
1500 ꢂ 100
3000 ꢂ 100
270 ꢂ 10
1,4-Cyclohexadiene
Conclusions
2-Propanola
2-Propanolb
In summary, our ndings show a lack of uorescence and low
triplet quantum yield and decarboxylation efficiency for
ketorolac. The formation of an enol intermediate is proposed
to be responsible for the formation of a long-lived interme-
diate from the triplet-excited state and an ICT excited state
may be expected to initiate the radiationless decay of the
molecule on the singlet manifold. Such a complex photo-
physical pathway helps explain the low photohemolytic
activity and reduced phototoxicity of ketorolac. These results
open a new paradigm on the, until now, well-known photo-
physical behavior of NSAIDs.
L-Tryptophan methyl ester
L-Tyrosine methyl ester
20-Deoxyguanosine
a
Measured in acetonitrile. b Measured in dichloromethane.
Thus, the presence of an adiabatic deactivation route where
spin multiplicity is conserved seems to be the pathway
responsible for triplet deactivation.
The second intermediate observed upon laser excitation
(Fig. 5B and C) showed a rise time of z260 ns with a strong
absorbance at 370 nm. This species has a lifetime of z86 ms and
is readily quenched by oxygen at a diffusion-controlled rate
constant of 1 ꢀ 1010 Mꢁ1 sꢁ1. A similar intermediate was detec-
Acknowledgements
ted for ketorolac ester and 2-benzoylpyrrole (Fig. S9 and S10†), We gratefully acknowledge NSERC and The Spanish Ministry of
indicating that the origin of this intermediate is directly related Economy and Competitiveness (CTQ2011-27758, CTQ2010-
to the benzoylpyrrole moiety. Further, although time-resolved 14892; M.G.B. Juan de la Cierva contract) for support of this
experiments for ketorolac in buffer solutions also led to similar research. J.C.N.-F. thanks the University of Ottawa for a Visiting
intermediates, the triplet lifetime was only 0.2 ms. Note that Professor fellowship. C.F. gratefully acknowledges NSERC and
under these conditions HCl addition produced a 60% quenching the Vanier CGS program. Lastly, the authors would like to thank
of the long-lived transient (data not shown). This result suggests Michel Grenier for his help in the time resolved measurements
the formation of an enolate-like intermediate for ketorolac as the and Professor Juan C. Scaiano for helpful discussions and
most likely responsible for this long-lived intermediate. comments on the article.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 1619–1622 | 1621

View Article Online
MedChemComm
Concise Article
9 N. J. Turro, V. Ramamurthy and J. C. Scaiano, Principles of
molecular photochemistry, University Science Books,
Sausalito, California, 2009.
10 I. Carmichael and G. L. Hug, J. Phys. Chem. Ref. Data, 1986,
15, 1–250.
11 A. A. Gorman, I. Hamblett, C. Lambert, A. L. Prescott,
M. A. J. Rodgers and H. M. Spence, J. Am. Chem. Soc., 1987,
109, 3091–3097.
Notes and references
1 G. Cosa, M. Lukeman and J. C. Scaiano, Acc. Chem. Res., 2009,
42, 599–607.
2 G. Cosa, L. Llauger, J. C. Scaiano and M. A. Miranda, Org.
Lett., 2002, 4, 3083–3085.
3 L. Llauger, M. A. Miranda, G. Cosa and J. C. Scaiano, J. Org.
Chem., 2004, 69, 7066–7071.
4 L. L. Costanzo, G. D. Guidi, G. Condorelli, A. Cambria and
M. Fama, Photochem. Photobiol., 1989, 50, 359–
365.
5 Y.-Y. He, D. C. Ramirez, C. D. Detweiler, R. P. Mason and
C. F. Chignell, Photochem. Photobiol., 2003, 77, 585–
591.
12 J. F. Ireland and P. A. H. Wyatt, J. Chem. Soc., Faraday Trans.
1, 1973, 69, 161–168.
13 A. Romani, F. Elisei, F. Masetti and G. Favaro, J. Chem. Soc.,
Faraday Trans., 1992, 88, 2147–2154.
14 L. Gu, H.-S. Chiang and A. Becker, Int. J. Pharm., 1988, 41,
95–104.
6 D. A. Handley, P. Cervoni, J. E. McCray and J. R. McCullough,
J. Clin. Pharmacol., 1998, 38, 25–35.
15 L. Gu, H.-S. Chiang and D. Johnson, Int. J. Pharm., 1988, 41,
105–113.
7 L. A. Boras-Uber and N. C. Brackett Jr, Am. J. Med., 1992, 92,
450–452.
8 M. L. Randi, T. Tison, G. Luzzatto and A. Girolami, BMJ.,
1993, 306, 186.
16 N. Ismail, L. Blancafort, M. Olivucci, B. Kohler and
M. A. Robb, J. Am. Chem. Soc., 2002, 124, 6818–6819.
17 M. Merchan and L. Serrano-Andres, J. Am. Chem. Soc., 2003,
125, 8108–8109.
1622 | Med. Chem. Commun., 2013, 4, 1619–1622
This journal is ª The Royal Society of Chemistry 2013