Palladium complexes built from flu and Alzheimer’s drugs curb SARS-CoV-2 in cells

A set of palladium(II) coordination compounds assembled from the aminoadamantane drugs amantadine, memantine and rimantadine strongly suppressed SARS-CoV-2 infection in human lung-derived cell cultures, according to a laboratory study published in Frontiers in Drug Discovery. The three metal complexes, named Pd-atd, Pd-mtn and Pd-rtd, showed selectivity indices of 14.2, 36.2 and 23.5 respectively, and each blocked more than 75 percent of viral entry in specific assays. The work also reports the first crystal structure of the Pd-rtd complex and traces how each compound interferes with different stages of the viral life cycle.

The research team, based at institutions including the Federal University of Uberlândia, the University of Leeds and the University of Campinas, screened the free drugs alongside the palladium(II) salt Li2PdCl4 and the three metal complexes. Amantadine, memantine and rimantadine are approved medicines used clinically against Alzheimer disease and Influenza A virus infection, which made them attractive starting points for coordination chemistry. At physiological pH these aminoadamantanes are essentially fully protonated, with pKa values between 10.1 and 10.5, so the free drugs carry a positive charge that limits passive passage through cell membranes. Binding the amine nitrogen to palladium(II) removes that charge and assembles a neutral, lipophilic, square-planar trans-[PdCl2(L)2] unit in which two ligand molecules and one metal center travel together.

Palladium(II) is a borderline-soft Lewis acid whose chlorido ligands are substitutionally labile, meaning the complexes can act as masked electrophiles toward accessible histidine, methionine and cysteine residues in viral or host proteins. The three ligands differ in ways expected to modulate this reactivity. Amantadine carries a primary amine bonded directly to the tertiary C1 carbon of the diamond-like adamantane cage, memantine shares that connectivity but adds two methyl groups at positions C3 and C5, and rimantadine places a stereogenic CH(CH3) spacer between the nitrogen and the cage, increasing conformational freedom around the Pd–N bond. These structural differences alter the steric environment of the metal, the hydrogen-bonding capacity of the coordinated N–H groups and the lipophilicity of each complex.

Chemical characterization confirmed the proposed compositions. Elemental analysis verified the metal-to-ligand molar ratios, mass spectrometry confirmed the presence of palladium(II), and infrared spectroscopy and nuclear magnetic resonance data confirmed coordination of the ligands. For Pd-rtd, single-crystal X-ray diffraction at 120 K revealed a monoclinic P21/n structure in which the palladium atom sits at a center of inversion, giving a square-planar complex with the chlorido ligands and rimantadine ligands in trans positions. The Pd–Cl bond measured 2.3018 angstroms and the Pd–N bond 2.0471 angstroms, closely matching the previously reported Pd-mtn complex, whose corresponding bond lengths were 2.3080 and 2.0548 angstroms. DMSO solvent molecules in the crystal form hydrogen bonds with the rimantadine amine groups.

Antiviral testing used human adenocarcinoma-derived alveolar basal epithelial cells engineered to express the ACE2 and TMPRSS2 receptors, infected with a SARS-CoV-2 reporter virus carrying the mNeonGreen fluorescence gene. In initial screening at the highest non-cytotoxic concentrations, the precursors Li2PdCl4, amantadine and memantine failed to inhibit viral replication, while free rimantadine inhibited 60.1 percent. The coordinated compounds performed better: Pd-atd suppressed replication by 50.8 percent at 2 micromolar, Pd-mtn by 97.2 percent at 10 micromolar and Pd-rtd by 71.1 percent at 10 micromolar, all with p-values below 0.0001. Dose-response experiments then established EC50 values of 8.4 micromolar for Pd-atd, 15.2 micromolar for Pd-mtn and 32.6 micromolar for Pd-rtd, with CC50 cytotoxicity values of 119.7, 550.9 and 767.5 micromolar respectively.

Time-of-drug-addition experiments mapped where each complex acts in the viral life cycle. When cells were pretreated with compound for one hour before infection, only Pd-atd reduced infection, by 30.2 percent. In entry assays where virus and compound were added together for one hour, Pd-atd blocked 76.4 percent of entry while Pd-mtn and Pd-rtd each blocked roughly 51 percent. Attachment assays at 4 degrees Celsius, which allow viral particles to bind cellular receptors without entering, showed strong inhibition of 69.2 percent for Pd-atd and 66.5 percent for Pd-mtn, while Pd-rtd inhibited attachment by 23 percent, a figure that rose to 37.4 percent when a 30-minute incubation at 37 degrees Celsius was added to probe post-attachment steps. Pre-incubating virus with compound before cell exposure revealed virucidal activity for Pd-mtn, which cut fluorescence levels by 62.6 percent, and for Pd-atd, which achieved a 32.5 percent reduction, while Pd-rtd showed no significant direct virucidal effect.

Only Pd-atd impaired post-entry stages of replication. Added one hour after infection, it inhibited 87.5 percent of replication, and even when added six hours post-infection it still reduced activity by 49 percent. Pd-mtn and Pd-rtd showed no significant activity in either post-entry protocol. Against variant viruses, the picture became more nuanced. Using infectious-clone chimeras carrying the Delta or Omicron BA.2 spike gene in a Wuhan-strain backbone, Pd-atd and Pd-rtd reduced Wuhan-B.1.617.2-Spike infection by 17.5 and 16.2 percent respectively, with Pd-mtn showing no effect, while all three complexes reduced Wuhan-BA.2-Spike infection by 47.7, 15.1 and 29.1 percent. With wild-type viruses, Pd-atd lowered Delta viral titers by 68.4 percent, but Pd-mtn and Pd-rtd had no effect against that variant. Against wild-type Omicron BA.2, all three complexes performed strongly, impairing replication by 99 percent for Pd-atd, 90 percent for Pd-mtn and 99 percent for Pd-rtd.

Mechanistic assays pointed toward proteins rather than nucleic acids as the likely targets. Pd-atd decreased the band intensity of SARS-CoV-2 double-stranded RNA by only 10.3 percent in a gel-retardation assay, indicating minimal interaction with the viral replication intermediate. In separate 1H NMR experiments using N-acetyl-L-cysteine as a small-molecule model for peptide and protein interactions, the Pd-rtd complex exchanged its rimantadine ligand for the NAC molecule over 24 hours, releasing free protonated rimantadine and forming a Pd-NAC product, mirroring earlier observations with Pd-atd and Pd-mtn. A dissemination assay in methylcellulose-overlayed cultures further showed that all three complexes reduced the median size of viral foci, from 0.115 square millimeters in the control to 0.0525, 0.077 and 0.080 square millimeters under Pd-atd, Pd-rtd and Pd-mtn treatment respectively, indicating reduced cell-to-cell spread.

Serial passaging of SARS-CoV-2 in the presence of each complex at 50 micromolar demonstrated that the virus can develop resistance in vitro. Pd-atd suppressed titers until the third passage, after which titers dropped sharply and then rebounded above control levels by the sixth passage. Pd-mtn maintained lower titers through the sixth passage before titers rose in the seventh, and Pd-rtd induced resistance as early as the second passage. Sequencing of the resistant populations identified two spike-protein mutations in Pd-atd-treated virus, G22049A and A22245G, corresponding to the amino acid substitutions E180K and H245R, both of which occur in circulating variants including Delta and BA.2. For Pd-rtd and Pd-mtn, no non-synonymous consensus mutations appeared, but intrahost diversity analysis revealed the H83Y mutation in nsp1 at 25 percent frequency in Pd-rtd-treated samples and the P1213L mutation in nsp3 at frequencies of 20 and 17 percent for Pd-rtd and Pd-mtn respectively. These non-structural proteins are involved in viral translation, host gene expression regulation, immune evasion and the formation of replication organelles.

The authors conclude that all three complexes likely exert antiviral activity through interactions with proteins rather than direct binding to viral nucleic acids, yet each imposes a distinct selective pressure: Pd-atd selected spike mutations and predominantly affected early stages of the cycle, while resistance to Pd-rtd and Pd-mtn involved non-structural proteins tied to RNA replication and host modulation. They note that the adamantane substitution pattern appears to modulate how accessible the palladium(II) center is to substitution by biological nucleophiles, predicting a reactivity order of Pd-atd over Pd-mtn over Pd-rtd that mirrors the EC50 values. The study is limited by its in vitro scope and the absence of comprehensive ADMET analysis, which the authors note is difficult for metal-based complexes because most existing prediction tools were developed for organic compounds. The findings warrant further investigation in vivo before any therapeutic application can be considered.

Subject of Research: Antiviral activity of palladium(II) aminoadamantane coordination complexes against SARS-CoV-2 in cell culture

Article Title: Palladium(II) aminoadamantane complexes: a promising approach to impair SARS-CoV-2 infection

Article References: Santos, I. A., Barros-Silva, L., Karla dos Santos Pereira, A., Astbury, S., McClure, C. P., de Alencar Simoni, D., Sucena, S. F., Nakahata, D. H., Ferraz de Paiva, R. E., Corbi, P. P., Harris, M., & Jardim, A. C. G. (2026). Palladium(II) aminoadamantane complexes: a promising approach to impair SARS-CoV-2 infection. Frontiers in Drug Discovery, 6, Article 1946835. https://doi.org/10.3389/fddsv.2026.1946835

Image Credits: AI Generated

DOI: 10.3389/fddsv.2026.1946835

Keywords: SARS-CoV-2, palladium(II) complexes, aminoadamantanes, amantadine, memantine, rimantadine, antiviral, metallodrugs, viral entry, drug resistance, Omicron variant, coordination chemistry