Computational study reveals how ketoprofen interacts with biomolecules

Ketoprofen, one of the most widely prescribed nonsteroidal anti-inflammatory drugs in the world, has long been regarded as a reliable workhorse against pain, fever and inflammation, particularly in conditions such as osteoarthritis, rheumatoid arthritis, ankylosing spondylitis and menstrual pain. A new computational study, however, suggests that the drug’s behaviour inside the body may be considerably more complicated than its familiar mechanism of blocking cyclooxygenase enzymes and suppressing prostaglandin synthesis would imply. Using density functional theory, a team of researchers has mapped, at the level of individual atoms and electrons, how ketoprofen interacts with ten biologically important molecules, and the results point to a web of potential off-target interactions that could subtly reshape amino acid chemistry, antioxidant defences and even glucose metabolism, especially during chronic or high-dose use.

The study, conducted with the DMol3 module of BIOVIA Materials Studio 2023, employed the generalized gradient approximation with the Perdew–Burke–Ernzerhof exchange–correlation functional and a double numerical basis set with polarization functions. The researchers, Eman H. Salem, Abbas M. Abbas and Adel S. Orabi, optimized the geometries of ketoprofen and its ten molecular partners—tryptophan, uric acid, glutathione, asparagine, glycine, proline, glucose, valine, cysteine and glutamic acid—selected to span amino acids, antioxidants, aromatic metabolites and carbohydrates. Convergence criteria were strict, with energy changes limited to 1.0 × 10−5 Hartree, maximum forces to 2.0 × 10−3 Hartree per Ångström and maximum displacements to 5.0 × 10−3 Ångströms. All calculations were performed in the gas phase, a deliberate simplification the authors acknowledge as a limitation, since the aqueous environment of living cells would modulate but not necessarily eliminate the electronic effects they observed.

The heart of the analysis lies in frontier molecular orbital theory, the quantum chemical framework that treats the highest occupied molecular orbital, or HOMO, as the electron-donating frontier of a molecule and the lowest unoccupied molecular orbital, or LUMO, as its electron-accepting frontier. The gap between these two orbital energies serves as a proxy for chemical reactivity: small gaps imply soft, polarizable, highly reactive species, while large gaps indicate hard, kinetically stable ones. Among the biomolecules examined, glutamic acid emerged as the most reactive, with the smallest energy gap of just 0.009 Hartree, while tryptophan proved the most stable, with a gap of 0.227 Hartree. Uric acid displayed the highest HOMO energy at −0.079 Hartree, marking it as the best electron donor in the set, whereas DL-proline possessed the lowest LUMO energy at −0.113 Hartree, making it the most eager electron acceptor. Ketoprofen itself carried the lowest HOMO energy of all compounds studied, a signature of its inclination to act as an electrophile in encounters with biological partners.

From these orbital energies the team derived a battery of global reactivity descriptors grounded in Koopmans’ theorem: ionization potential, electron affinity, chemical hardness, softness, chemical potential, electronegativity and the electrophilicity index. Glutathione, the cell’s principal low-molecular-weight antioxidant, showed the lowest chemical hardness at 0.005 Hartree and a softness value of 96.34, confirming its exceptional reactivity. Glutamic acid and glutathione also registered the highest electrophilicity indices, 1.296 and 1.091 respectively, indicating strong appetites for electrons, whereas glycine, uric acid and tryptophan sat at the opposite end of the scale. When the researchers asked which molecule would be most likely to interact with ketoprofen—an electrophile with an electronegativity of 0.105 Hartree and moderate hardness—uric acid stood out clearly. With the lowest electronegativity, 0.031 Hartree, the lowest electrophilicity index, 0.010, and high softness, uric acid behaves as the strongest nucleophile in the panel, making it the theoretically best-matched partner for the drug.

To move from global tendencies to atom-by-atom predictions, the team computed Fukui functions, local reactivity descriptors derived from Mulliken atomic charges in neutral, anionic and cationic states. These functions pinpoint where a molecule’s electron density is most inclined to shift upon accepting or donating electrons. For ketoprofen, the carboxyl oxygen designated O8 proved to be the star of the show. It exhibited the highest values for all three Fukui functions—0.150 for electrophilic attack, 0.307 for nucleophilic attack and 0.229 for radical attack—revealing a remarkable dual character in which the atom can both donate and accept electrons, a property the authors attribute to its lone pairs and self-electronegativity. The C7 atom of the benzoyl group followed closely as a second electrophilic hotspot. In tryptophan, the C9 site showed the highest nucleophilic Fukui value of the entire dataset at 0.660, while cysteine’s sulfur atom S5 registered a strong 0.420, flagging the thiol sulfur as a potent electron donor poised to engage the drug’s electrophilic centres.

One of the most striking findings concerns cysteine, whose electronic personality changes dramatically upon binding to ketoprofen. Before interaction, cysteine’s HOMO sat at −0.207 Hartree with an energy gap of 0.157 Hartree; after complexation, the gap swelled to 0.318 Hartree, with the LUMO localized largely on the ketoprofen moiety. This widening indicates that ketoprofen accepts electron density from cysteine while simultaneously dampening the amino acid’s chemical reactivity and electron-transfer capability. Because cysteine’s thiol group underpins redox balance, disulfide-bridge formation in proteins and detoxification pathways, the authors suggest that ketoprofen’s electronic grip on the molecule could, in principle, interfere with oxidative stress defences and enzymatic functions that depend on free thiol chemistry. The effect is predicted to be weak and reversible, but its direction is chemically unambiguous.

Glutamic acid tells a different and equally consequential story. The neurotransmitter and metabolic intermediate is extraordinarily reactive in its free state, with that minuscule 0.009 Hartree gap and an electrophilicity index of 1.296. When complexed with ketoprofen, however, its electrophilicity plummets to 0.231 Hartree, essentially matching that of the drug itself, while the HOMO–LUMO gap of the complex remains close to ketoprofen’s original 0.112 Hartree. Fukui mapping identified the O8 carboxyl oxygen of ketoprofen, with a nucleophilic index of 0.308 and a Mulliken charge of −0.493 elementary charges, as the dominant site of interaction. The binding energy for this pair was computed at −6.49 Hartree, by far the most negative value in the study and a strong indication of thermodynamically favourable complex formation. Critically, the authors conclude that the interaction is moderately weak and reversible, that the carboxylate groups responsible for glutamic acid’s biological activity remain intact, and that excitatory neurotransmission through NMDA and AMPA receptors should be largely preserved, with any effect limited to a temporary dip in free glutamic acid concentration.

Binding energies across the full panel revealed a heterogeneous landscape. Only four complexes proved thermodynamically favourable: ketoprofen–DL-proline at −0.0299 Hartree, ketoprofen–glucose at −0.00486 Hartree, ketoprofen–valine at −0.0028 Hartree and ketoprofen–glutamic acid. The remaining pairs, including tryptophan, uric acid, glutathione, asparagine, glycine and cysteine, yielded positive binding energies, signalling weak or unfavourable associations under the gas-phase conditions of the calculation. Notably, the researchers found no consistent linear correlation between the global electronic descriptors and the binding energies, a result they interpret as evidence that geometry, steric effects and hydrogen-bonding patterns matter as much as orbital energies in determining complex stability. Electronic descriptors, they conclude, are best treated as complementary rather than predictive tools.

The glucose interaction carries particular practical weight. Upon complexation, glucose’s energy gap widened from 0.034 to 0.111 Hartree, indicating reduced reactivity, and a calculated charge transfer of roughly 0.42 electrons flowed from glucose to ketoprofen, confirming the drug’s electron-acceptor role. The authors propose that ketoprofen may form stable hydrogen bonds or weak ester-like linkages with glucose, potentially reducing its bioavailability and perturbing normal glucose-related biochemical pathways. They flag this as a theoretical concern especially relevant to diabetic patients or individuals with impaired glucose regulation, while stressing that the finding is purely computational and awaits experimental confirmation. Prior work has, in fact, exploited glucose moieties to shuttle ketoprofen across the blood–brain barrier via glucose transporters, lending plausibility to a genuine chemical affinity between the two molecules.

Tryptophan, meanwhile, emerged as the most electronically transformed partner in the panel. Complexation dropped its energy gap from 0.227 to 0.074 Hartree, softened the molecule and raised its electrophilicity, producing a highly reactive adduct. Because tryptophan is the essential precursor of serotonin and other neuroactive compounds, the authors warn that strong electronic engagement with ketoprofen could, under chronic exposure, subtly interfere with neurotransmitter biosynthesis. A parallel concern applies to glutathione: even though its formal binding energy was unfavourable, its pronounced softness and electrophilic character suggest meaningful interactions with electron-rich regions of the drug, raising the possibility that long-term ketoprofen use could tax antioxidant capacity at a time when NSAIDs are already associated with oxidative stress. Cysteine, paradoxically, ranked as the least likely significant partner despite its reactive thiol, owing to its low softness.

The authors are careful to frame their conclusions within the limits of the method. No solvent model was applied, counterpoise correction for basis set superposition error was not performed, and the work is explicitly theoretical, intended as a molecular-level hypothesis generator rather than a clinical verdict. Their overarching message is that quantum chemical modelling of the kind demonstrated here—combining frontier orbital analysis, global reactivity descriptors, Fukui mapping and binding energy calculations—offers drug developers a fast, inexpensive screen for off-target chemistry that might otherwise surface only after years of clinical use. Ketoprofen remains, in their assessment, neither toxic nor dangerous on its own terms, but the study suggests it is an electronically promiscuous molecule capable of transiently engaging sensitive players in cellular biochemistry. Translating these predictions into certainty will require the in vitro and in vivo experiments the researchers explicitly call for, but the work stands as a demonstration that the hidden life of a familiar drug, down to its last reactive oxygen atom, can now be read directly from the mathematics of its electrons.