PSMA PET tracers bind the same target in distinct ways, simulations and cell assays show

Clinically used PSMA-targeting radiopharmaceuticals are routinely treated as a single diagnostic class in prostate cancer PET imaging guidelines, yet a new integrated computational and laboratory study reports that the fluorinated tracers [18F]PSMA-1007, [18F]DCFPyL and [18F]JK-PSMA-7 differ measurably in how they engage their shared molecular target. The work, published in EJNMMI Radiopharmacy and Chemistry on 29 August 2026, combined molecular docking, molecular dynamics simulations and real-time cellular binding assays to define the mechanistic determinants of ligand-specific behaviour among these clinically used 18F-fluorinated radiotracers, with [68Ga]Ga-PSMA-11 included as a widely used clinical reference.

Prostate-specific membrane antigen, or PSMA, is the imaging target for these radiopharmaceuticals, and the study set out to characterise the binding of the three fluorinated compounds within the PSMA active site, paying particular attention to zinc coordination and water-mediated interactions. In the simulations, all three fluorinated radiopharmaceuticals bound within the PSMA active site through the conserved Glu–urea–Lys motif and maintained stable interactions with the binuclear zinc centre. Molecular simulations further revealed ligand-specific differences in the occupation of the active-site funnel and showed substantial contributions from water-mediated hydrogen-bond networks, particularly for [18F]PSMA-1007. Relative binding energies were estimated using the Molecular Mechanics Poisson–Boltzmann Surface Area method.

The cellular experiments used the LigandTracer system to estimate apparent association and dissociation rate constants in real time. Here the tracers separated clearly. [68Ga]Ga-PSMA-11 exhibited the highest mean apparent association rate and the most pronounced decrease in cell-associated signal during the dissociation phase. In contrast, [18F]PSMA-1007 showed the lowest mean apparent association rate and no measurable signal decrease during the experimental dissociation period, while [18F]DCFPyL and [18F]JK-PSMA-7 displayed intermediate kinetic profiles between these extremes.

Competition assays with 2-(phosphonomethyl)pentanedioic acid, a PSMA inhibitor, reinforced this pattern of graded behaviour. The assays yielded progressively higher apparent IC50 estimates moving from [68Ga]Ga-PSMA-11 to [18F]PSMA-1007. Notably, no measurable decrease in [18F]PSMA-1007-associated signal was observed during the 30-minute real-time competition experiment, even at the highest inhibitor concentration tested, underscoring how differently this fluorinated tracer behaved under the assay conditions compared with the gallium-68 reference compound.

The authors emphasise that these findings were obtained under the experimental conditions examined, taking into account the simplifications adopted for the design of the study and for the analysis of the results. The conclusion drawn is deliberately scoped: despite sharing a conserved PSMA-binding motif, the clinically used PSMA radiopharmaceuticals exhibited distinct molecular interaction patterns and apparent cellular binding kinetics. The study frames these results as providing a mechanistic basis for tracer-specific behaviour and as supporting further investigation of whether such differences influence clinical imaging characteristics, rather than demonstrating any clinical effect directly.

Methodologically, the study is an integrated in silico and in vitro investigation. Molecular docking and molecular dynamics simulations characterised binding within the PSMA active site, including zinc coordination and water-mediated interactions, and the Molecular Mechanics Poisson–Boltzmann Surface Area method supplied relative binding energy estimates. On the laboratory side, the LigandTracer platform delivered real-time estimates of apparent association and dissociation rate constants for the three fluorinated tracers and [68Ga]Ga-PSMA-11, and the competition assays with 2-(phosphonomethyl)pentanedioic acid added an orthogonal readout of apparent binding potency.

The reported kinetic ordering places [68Ga]Ga-PSMA-11 at the fast-associating, readily dissociating end of the spectrum, [18F]PSMA-1007 at the slow-associating, apparently persistent end, and the two remaining fluorinated tracers in between. Because the dissociation-phase signal decrease for [18F]PSMA-1007 was not measurable during the experimental observation period, the study reports this as an absence of measurable decrease rather than as a determined dissociation rate. Similarly, the IC50 values are described as apparent estimates, reflecting the conditions of the competition assays rather than absolute thermodynamic affinities.

Water-mediated hydrogen-bond networks emerged as a substantial contributor to binding in the simulations, with the contribution being particularly notable for [18F]PSMA-1007. Differences in how each ligand occupied the active-site funnel, the region surrounding the zinc-containing catalytic centre, provide a structural correlate for the ligand-specific kinetic profiles observed in the cellular assays. The conserved Glu–urea–Lys motif, shared by all three fluorinated compounds, anchored each ligand in the active site and maintained contact with the binuclear zinc centre throughout the simulated trajectories.

The study’s framing positions these mechanistic differences as a potential foundation for understanding why radiotracers that target the same receptor can display distinct in vivo behaviour, a point the background section raises by noting that each radiopharmaceutical exhibits distinct in vivo behaviour even though clinical guidelines often group them as a single diagnostic class. The molecular interactions and binding kinetics underlying these differences had remained incompletely understood, and this integrated approach offers a first mechanistic account grounded in both simulation and direct cellular measurement.

For readers following PSMA PET imaging, the practical significance lies in the evidence that tracer choice may rest on more than target selectivity alone. The study documents ligand-specific differences in active-site occupancy, water-network contributions, apparent association rates, dissociation behaviour and apparent IC50 estimates across four clinically used compounds. Whether these experimentally observed kinetic and interaction differences translate into differences in clinical imaging performance remains an open question that the authors identify for further investigation, consistent with the study’s stated aim of providing a mechanistic basis for tracer-specific behaviour rather than clinical guidance.

Subject of Research: Binding kinetics and molecular interactions of clinically used PSMA diagnostic radiopharmaceuticals studied through molecular simulations and real-time cellular binding assays

Article Title: Ligand-specific binding kinetics and molecular interactions of clinically used PSMA diagnostic radiopharmaceuticals: an integrated in silico and in vitro study

Article References: Casini, B., Dighero, E., Riondato, M., Manconi, A., Trombetti, G. A., Ghelardoni, M., Losacco, S., Sambuceti, G., Marini, C., Bauckneht, M., D’Ursi, P., & Fossa, P. (2026). Ligand-specific binding kinetics and molecular interactions of clinically used PSMA diagnostic radiopharmaceuticals: an integrated in silico and in vitro study. EJNMMI Radiopharmacy and Chemistry. https://doi.org/10.1186/s41181-026-00485-5

Image Credits: AI Generated

DOI: 10.1186/s41181-026-00485-5

Keywords: PSMA, radiopharmaceuticals, PET imaging, prostate cancer, molecular docking, molecular dynamics, binding kinetics, LigandTracer, 18F-PSMA-1007, Ga-PSMA-11, zinc coordination, in silico