Rational design of η⁶-arene Ru(II) complexes for amyloid-β targeting: influence of coordination lability and aromaticity

Abstract

The interaction of transition-metal complexes with amyloidogenic peptides represents a promising strategy for controlling pathological protein aggregation through coordination chemistry and supramolecular effects. Herein, we investigate a series of η⁶-arene Ru(II) complexes bearing a glucosylated N-heterocyclic carbene (NHC) ligand and distinct ancillary ligand environments as modulators of amyloid-β (Aβ_1–42 and Aβ_21–40) aggregation. By systematically varying the nature of the arene, overall charge, and ligands lability, we elucidate how the design of the complexes governs peptide binding, aggregation pathways, fibril morphology, and cytotoxicity. Thioflavin T fluorescence assay revealed pronounced, peptide-dependent inhibition of amyloid aggregation. RuPhenCym strongly suppresses Aβ_1–42 fibrillization, whereas the neutral, dichloro complex RuCl₂Tol exhibited enhanced efficacy toward the shorter Aβ_21–40 fragment. Circular dichroism spectroscopy demonstrated that Ru complexes modulate peptide secondary structure, promoting early β-structured species formation and altering aggregation kinetics. Scanning electron microscopy showed substantial remodeling of fibril morphology, including reduced fiber length and increased heterogeneity, indicative of off-pathway aggregation. Electrospray ionization mass spectrometry provided direct evidence of adducts formation in the presence of RuCl₂Tol, highlighting the crucial role of labile chloride ligands in generating coordination vacancies that enable peptide binding. Importantly, RuCl₂Tol and RuPhenCym significantly attenuate Aβ_1–42-induced cytotoxicity in SH-SY5Y neuroblastoma cells without exhibiting intrinsic cellular toxicity. Overall, this study establishes clear structure–activity relationships linking ligand environment, coordination chemistry, and biological outcome in η⁶-arene Ru(II)–amyloid systems. These findings identify glucosylated η⁶-arene Ru(II) complexes as tunable bioinorganic platforms for the selective modulation of amyloid aggregation, providing a rational framework for the development of metal-based agents targeting neurodegenerative disorders.

---

Introduction

Amyloid-β (Aβ) peptides are generated from amyloid precursor protein (APP) by β- and γ-secretase cleavage. Among them, Aβ_1–42 is strongly implicated in Alzheimer's disease because it aggregates more readily and forms neurotoxic oligomers and fibrils compared with Aβ_1–40, a property largely attributed to its hydrophobic C-terminal region.¹ To elucidate sequence determinants of aggregation, shorter Aβ fragments have been widely studied. The Aβ_21–40 fragment retains most of the hydrophobic C-terminal domain while excluding the N-terminal region involved in metal binding and electrostatic interactions. Experimental studies show that Aβ_21–40 can self-assemble into amyloid-like fibrils and reproduce key aggregation features of full-length Aβ, making it a useful minimal model for C-terminal-driven amyloid formation and for testing aggregation modulators² Complementary studies on the neighboring Aβ_21–30 segment indicate that this region adopts relatively stable folded conformations that may act as early structural nuclei during peptide folding and aggregation, further supporting the central role of residues 21–40 in amyloid assembly.^2–4 Ruthenium complexes have emerged as promising modulators of amyloid aggregation through a variety of molecular structures and mechanisms. Ru(II) polypyridyl complexes, which often employ bipyridine or phenanthroline ligands, can engage in strong noncovalent interactions with Aβ via π–π stacking and hydrophobic contacts,⁵ and in some cases chelate redox-active metal ions as Cu(II) which are associated with Aβ plaques.⁶ A series of Ru(II) polypyridyl compounds bearing a hydrophobic ancillary ligand (tetra-xylene bipyridine glycoluril, txbg) showed nearly complete inhibition of Aβ aggregation at micromolar concentrations and low toxicity in neuroblastoma cells.⁷ “Piano-stool” Ru(II) arene complexes also demonstrated potent anti-amyloid activity: two Ru(II) complexes of the form [Ru(p-cymene)Cl(L) (L = bidentate nitrogen ligand)] inhibited aggregation of Aβ_1–42 and the most active compound protected neuronal cells (Neuro-2a) and computational docking suggested both hydrophobic interactions and coordination to peptide binding sites.⁸ More recently, a series of Ru(II)–arene azole complexes was shown to be exceptionally stable in aqueous environments, to coordinate to Aβ, and to strongly inhibit aggregation, with one compound (“RuBO”) also reducing cytotoxicity.⁹ Another important approach involves photoactivatable Ru(II) polypyridyl complexes: in one study, distorted Ru(II) complexes bearing sterically encumbered bipyridine ligands were irradiated, triggering the loss of one ligand and generating coordinatively unsaturated species capable of binding histidine residues in Aβ. This ligand loss caused a rapid change in aggregation behavior, redirecting Aβ into large, amorphous, insoluble aggregates rather than typical fibrils and these off-pathway species were shown to be more susceptible to proteolytic degradation.¹⁰ In a more recent study, two Ru(II) polypyridyl complexes with extended planar phenanthroline ligands were engineered to pre-associate with Aβ via hydrophobic interactions, and, upon photoactivation, release ligands that promote peptide binding and scavenge Cu–Aβ species, thereby reducing redox cycling and reactive oxygen species (ROS) production.¹¹ In addition to ligand-exchange systems, Ru(II) complexes that release biologically active small molecules under light have also been explored. For instance, CO-releasing Ru(II) compounds with benzimidazole and terpyridine moieties modulate the aggregation of model amyloidogenic peptides via the formation of adducts after loss of labile ligands on irradiation.¹²

Mechanistically, the effects of Ru complexes derive from a combination of noncovalent binding (hydrophobic and π–π), coordination to peptide residues (especially His), displacement or sequestration of metal ions like Cu(II), and, in photoactivatable systems, ligand release or ROS/singlet oxygen generation.^6,13,14 Indeed, at the molecular level, these complexes coordinate to the imidazole side chains of His within N-terminal portion of Aβ, serving as critical metal-binding sites influencing peptide conformational dynamics. This coordination interferes with the ability of Aβ to adopt the β-sheet-rich structures necessary for nucleation and fibril elongation. Beyond histidine coordination, ruthenium complexes engage in hydrophobic/aromatic interactions facilitated by their aromatic ligands, thus destabilizing the hydrophobic core that promotes amyloid assembly.^10,15,16 They also form hydrogen bonds through amine ligands, further stabilizing non-fibrillar peptide conformations.

In several previous studies, we examined the effects of two η⁶-arene glucosylated Ru(II) complexes on the self-aggregation of several amyloidogenic peptides with different sequences and aggregation kinetics. These complexes contain two chloride ligands, a glucosylated N-heterocyclic carbene (NHC), and different arene ligands, Their activity was tested on small amyloid models, including the NPM1_264–277 fragment of nucleophosmin-1,^17–19 the BSP_27–32 fragment of the human Bloom syndrome protein,²⁰ and the β2-microglobulin fragment β2m_83–88 derived from proteolytic cleavage.^21,22 In the present study, for the first time, we investigated three additional Ru(II) complexes—RuCl₂Tol, RuBipyCym, and RuPhenCym (Fig. 1), for their ability to modulate the self-aggregation of Aβ_1–42 and Aβ_21–40 peptides.

Fig. 1 Schematic structure of (A) RuCl₂Tol, (B) RuBipyCym and (C) RuPhenCym Ru compounds. (D) Sequences of amyloid peptides investigated in this study.

All three compounds exhibit a piano-stool geometry and contain an NHC ligand bearing a peracetylated β-D-glucose moiety bound through the anomeric nitrogen. RuCl₂Tol is a neutral complex featuring toluene as the arene ligand quite similar to those already investigated toward small, and His-rich amyloid peptides,^17,22 while RuBipyCym and RuPhenCym are cationic p-cymene complexes incorporating either 2,2′-bipyridine or 1,10-phenanthroline as chelating ligands, with trifluoromethanesulfonate as the counterion. In them bipyridine and phenanthroline ligands in place of chlorides (Fig. 1) are predicted to have a reduced ability as leaving ligands in ligand exchange reactions since in both cases it requires chelate opening.

The investigations were conducted using a broad set of spectroscopic and biophysical methods. Initially, the effects of Ru complexes on ThT fluorescence emission were examined for both amyloids. CD spectroscopy was used to evaluate changes in peptide conformational preferences, while SEM analysis provided insight into the structural impact of Ru complexes on amyloid fibers. ESI-MS was then employed to investigate adduct formation between the amyloid peptides and Ru complexes and viability cellular assays in SH-SY5Y cells allowed to evaluate different effects on typical amyloid cytotoxicity.

Results and discussion

Effects of η⁶-arene Ru(II) complexes on amyloid aggregation

Thioflavin T (ThT) assay was performed to evaluate the effects of Ru complexes on the aggregation kinetics of Aβ_1–42 and Aβ_21–40 peptides, at different ratios, (Fig. 2): maximum fluorescence intensities and half-times of aggregation (t_1/2) are reported in Table S1, however, ThT intensity should be interpreted cautiously, as it reflects dye binding to β-structured aggregates rather than a direct quantitative measure of fibrils concentration. As expected, Aβ_1–42 rapidly aggregated (Fig. 2A) as shown by a short lag phase and a steep increase in ThT signal (maximum fluorescence of 200.6 a.u. with a t_1/2 of 95 min). Among the Ru complexes, RuPhenCym exhibited the strongest inhibitory effect, reducing the ThT signal (18.2 a.u.), and, at 1 : 1 ratio, completely hampered the evaluation of t_1/2, while RuBipyCym showed minimal effect at 1 : 0.5 and only partial inhibition at 1 : 1 ratios (94.6 a.u., t_1/2 = 169 min). RuCl₂Tol displayed an intermediate efficacy between RuBipyCym and RuPhenCym activity (135.2 a.u. at 1 : 0.5 and 40.5 a.u. at 1 : 1 ratios).

Fig. 2 Aggregation kinetics monitored by ThT-fluorescence of (A) Aβ_1–42 (50 μM), (B) Aβ_21–40 (100 μM) alone and in the presence of Ru-complexes 1 : 1 and 1 : 0.5 peptide : complex molar ratio.

Aβ_21–40 aggregates more slowly than Aβ_1–42 (t_1/2 = 286 min) (Fig. 2B). Its aggregation was affected by the presence of both RuBipyCym and RuPhenCym : RuBipyCym showed minimal inhibitory activity, while RuPhenCym was effective mainly at the 1 : 1 and less at 1 : 0.5 ratios, with the overall inhibition weaker than for Aβ_1–42. Interestingly, RuCl₂Tol exhibited the strongest effect, with substantial reduction of maximum fluorescence already at 1 : 0.5 (51 a.u.) and an almost complete suppression at 1 : 1 ratio (40 a.u.).

These observations suggest a peptide-dependent mechanism of inhibition: indeed, the greater effect of RuPhenCym toward Aβ_1–42 may arise from the presence of extended aromatic interactions between the η⁶-coordinated arene and the planar N^N aromatic chelating ligands, with aromatic residues present in Aβ_1–42 (e.g., His^6,13,14 Tyr,¹⁰ and Phe^4,19,20) absent in Aβ_21–40. In contrast, RuCl₂Tol, which bears labile chlorides, resulted more effective against the shorter Aβ_21–40 peptide, likely due to its increased accessibility toward the smaller and more flexible sequence. None of Ru complexes alone exhibited significant signals in this assay (Fig. S1).

Conformational effects of η⁶-arene Ru(II) complexes on amyloid peptides

To investigate the conformational effects of the Ru-complexes on Aβ peptides, CD experiments were performed at a 1 : 1 peptide-to-complex ratio (Fig. 3). At t = 0 the observed β-content (Table S2) likely reflects rapid re-aggregation Aβ_1–42 alone in PBS,²³ and, during time, we observed a complete transition toward β-sheet structure (Fig. 3A).⁴ The presence of all three Ru-complexes promoted early β-structured species formation already at t = 0, with RuPhenCym showing the most pronounced effect over time (Fig. 3D). This is supported by CD deconvolution, which indicates a β-content of 40% with RuPhenCym (Table S2) at t = 0, reaching 50% after 48 h. A similar but slower effect was observed in the presence of RuBipyCym (Fig. 3C). For Aβ_21–40 alone spectra suggested a predominant random coil content (Fig. 3E) and RuCl₂Tol and RuBipyCym exhibited minimal effects, with a β-content (Table S3) similar to that of the peptide alone, and a progressive reduction of the Cotton effect over time, but no clear β-sheet transition. Interestingly, since t = 0, in the presence of RuBipyCym and RuPhenCym complexes, an additional positive band at λ 250 nm (Fig. 3G and H) is observed, suggesting a different solvent exposure of the aromatic ligands of the complexes, indicating the formation of adducts with the peptide. In addition, the presence of RuPhenCym increased the β-content already at t = 0 (38.5%) (Table S3) and, after 48 h, a more defined β-structure was observed. However, CD spectroscopy cannot distinguish between non-toxic and toxic β-sheet-containing species, including oligomers and these species may correspond to either on-pathway intermediates or off-pathway aggregates, and their precise nature remains to be established.

Fig. 3 Overlays of CD spectra of (A–D) Aβ_1–42, (E–H) Aβ_21–40 peptides in the absence and presence of Ru complexes, subtracted by spectra of complexes alone. All samples were prepared at 1 : 1 Aβ : Ru molar ratio.

Effects of η⁶-arene Ru(II) complexes on fibril morphologies

To assess morphological changes in aggregates derived from Aβ_1–42 and Aβ_21–40 peptides in the absence and presence of Ru-complexes, scanning electron microscopy (SEM) was employed. Samples were prepared by mixing peptides and complexes at a 1 : 1 molar ratio and analyzed after 6 h of stirring. Representative micrographs are shown in Fig. 4. Aβ_1–42 alone formed long, well-defined fibers (Fig. 4A, A′ and Table 1), whereas Aβ_21–40 fibers were shorter and thicker, (Fig. 4E, E′ and Table 1), in line with already reported data.^24,25 The presence of Ru-complexes altered the morphology of the fibers with distinct fiber populations and heterogeneous lengths and diameters. In the presence of RuCl₂Tol, Aβ_1–42 fibers were relatively uniform, slightly thicker than those of the peptide alone (Fig. 4B, B′ and Table 1). In the presence of RuBipyCym, fibers were shorter (Fig. 4C, C′ and Table 1), while diameters remained similar to those observed with RuCl₂Tol. The presence of RuPhenCym induced a more heterogeneous morphology, giving rise to three aggregates that differ in both length (in the range of 660–1050 μm) and diameter (in the range of 8–17 μm) (Fig. 4D, D′ and Table 1).

Fig. 4 SEM images of, upper panel, Aβ_1–42 alone (A and A′) and in the presence of RuCl₂Tol (B and B′), RuBipyCym (C and C′) and RuPhenCym (D and D′) at a magnification of 220× (A–D) and 2500× (A′–D′); lower panel, of Aβ_21–40 alone (E and E′) and in the presence of RuCl₂Tol (F and F′), RuBipyCym (G and G′) and RuPhenCym (H and H′) at a magnification of 220× (E–H) and 2500× (E′–H′).

Table 1 SEM analyses: average diameter and length of fibers obtained for Aβ_1–42 and Aβ_21–40 peptides in the absence and presence of Ru complexes

	Average length (μm)		Average diameter (μm)
Aβ1–42		1721 ± 3		12 ± 2
Aβ1–42 : RuCl2Tol 1 : 1	Fiber 1	675 ± 4	Fiber 1	15 ± 2
	Fiber 2	801 ± 2	Fiber 2	13 ± 2
Aβ1–42 : RuBipyCym 1 : 1	Fiber 1	1076 ± 4	Fiber 1	15 ± 2
	Fiber 2	610 ± 3	Fiber 2	13 ± 2
Aβ1–42 : RuPhenCym 1 : 1	Fiber 1	1052 ± 2	Fiber 1	8 ± 2
	Fiber 2	927 ± 2	Fiber 2	8 ± 2
	Fiber 3	662 ± 4	Fiber 3	17 ± 2
Aβ21–40		665 ± 4		42 ± 3
Aβ1–42 : RuCl2Tol 1 : 1		83 ± 2		5 ± 2
Aβ1–42 : RuBipyCym 1 : 1		252 ± 3		12 ± 2
Aβ1–42 : RuPhenCym 1 : 1		227 ± 4		14 ± 2

---

For Aβ_21–40, the addition of Ru-complexes strongly reduced fiber size, resulting in a single fiber population, in contrast to the multiple populations observed for Aβ_1–42. In the presence of RuCl₂Tol, fibers became very short and thin (Fig. 4C, C′ and Table 1), whereas RuBipyCym (Fig. 4G and G′) and RuPhenCym (Fig. 4H and H′) led to slightly longer fibers with diameters of 12–14 μm (Table 1). The Ru complexes alone did not show any fiber formation (data not shown). Granular features observed in some micrographs are attributed to buffer salt residues (Fig. S5).

ESI-MS analysis: formation of η⁶-arene Ru(II) complex–peptide adducts

To evaluate the formation of direct adducts between Aβ peptides and Ru complexes, they were separately incubated at a 1 : 1 molar ratio (Aβ : Ru) and analyzed after two hours by electrospray ionization-mass spectrometry (ESI-MS). Fig. 5 and 6 show the analysis of each peptide in the presence of RuCl₂Tol.

Fig. 5 ESI-MS spectra of the Aβ_1–42 peptide in the absence (A) and presence of RuCl₂Tol (B). The peaks marked with bn derive from spontaneous source fragmentation of Aβ_1–42 (b series elements). Asterisks (*) highlight the species present in the control (RuCl₂Tol alone).

Fig. 6 ESI-MS spectra of the Aβ_21–40 peptide in the absence (A) and presence of RuCl₂Tol (B). The peaks marked with bn derive from spontaneous source fragmentation of Aβ_21–40 (b series elements). Asterisks (*) highlight the species present in the control RuCl₂Tol alone.

The incubation of Aβ_1–42 with RuCl₂Tol produced a peak at m/z 1680.21 a.m.u., indicating a stable adduct formed by the peptide and RuCl₂Tol with the loss of two chlorides ligands and of two acetyl ligands from the sugar portion (Fig. 5 and Table 2). The tendency to lose acetyl fragment in ESI-MS experimental conditions was already reported for this class of compounds.²⁶ This evidence suggested that the longer peptide, Aβ_1–42, may provide multiple donor sites for RuCl₂Tol. Similar experiments with RuBipyCym or RuPhenCym, exhibited novel peaks due to the formation of nonspecific associations with trifluoromethanesulfonate (CF₃SO₃⁻) ions (Fig. S2, S3 and Tables S4, S5). The lack of stable peptide adducts with RuBipyCym and RuPhenCym can be explained by the presence of the bidentate ligands in their coordination spheres, that do not undergo ligand exchange the ruthenium center, hampering peptide coordination at the metal center and detection of the relative adducts.

Table 2 Experimental m/z values detected in the spectra of Aβ_21–40 and Aβ_1–42 alone and with the addition of RuCl₂Tol. The ion species corresponding to each experimental m/z, the expected m/z value (theoretical) and their charge states are also reported

	Description	m/z (charge)		Theoretical m/z
		Control	+ Complex
Aβ1–42 : RuCl2Tol	Aβ1–42	1505.35 (+3)	1505.50 (+3)	1505.71
		1129.15 (+4)	1128.81 (+4)	1129.54
		903.61 (+5)	903.57 (+5)	903.83
		753.28 (+6)	753.30 (+6)	753.36
		645.85 (+7)	—	645.88
	b38	1371.78 (+3)	—	1371.34
	b34	1256.68 (+3)	1256.94 (+3)	1256.63
	Aβ1–42 + RuCl2Tol (−2Cl^−, −2Ac)	—	1680.21 (+3)	1679.91
			1008.04 (+5)	1007.8
			840.17 (+6)	839.96
Aβ21–40 : RuCl2Tol	Aβ21–40	1886.57 (+1)	1886.58 (+1)	1886.19
		943.70 (+2)	943.71 (+2)	943.6
		629.62 (+3)	—	629.4
	b18	1670.53 (+1)	1670.52 (+1)	1669.86
	b16	1555.45 (+1)	1555.43 (+1)	1555.81
	b15	1456.46 (+1)	1456.46 (+1)	1456.75
	b14	1325.42 (+1)	1325.44 (+1)	1325.71
	b12	1155.39 (+1)	1155.42 (+1)	1155.6
	b11	1042.37 (+1)	1042.22 (+1)	1042.52
	Aβ21–40 + RuCl2Tol (−2Cl^–, −1Ac)	—	1225.05 (+2)	1224.8

---

In the case of Aβ_21–40 (Fig. 6), signals of the b-series were generated from spontaneous in-source fragmentation events, as previously reported.²⁷ The presence of RuCl₂Tol led to the appearance of a distinct peak at m/z 1225.15 a.m.u. (Table 2), corresponding to the peptide bound to RuCl₂Tol complex upon the loss of two chloride ligands and one acetyl group. In the absence of N-terminal histidine residues, coordination may involve side chains such as Glu²² and Asp.²³ These interactions, combined with the increased flexibility of the shorter peptide, may facilitate coordination to RuCl₂Tol following chloride dissociation. In the presence of RuBipyCym or RuPhenCym no similar adducts were detected (Fig. S2, S3 and Tables S4, S5), while novel signals are consistent with nonspecific adducts involving trifluormethanesulfonate ions.

Effects of η⁶-arene Ru(II) on the Aβ_1–42 cytotoxicity

MTT assay was performed to assess the potential protective effect of Ru complexes against Aβ_1–42 amyloid-induced cytotoxicity²⁸ at three different time points (0, 2 and 24 h) (Fig. 7). As a preliminary step, the effects of Ru-complexes alone toward SH-SY5Y neuroblastoma cells were evaluated. Complexes did not show significant toxicity up to 24 h of incubation (Fig. S4). As expected, treatment with Aβ_1–42 resulted in a significant, time-dependent reduction of cell viability (almost 20%), confirming its intrinsic neurotoxic effects.²⁹ The presence of the Ru-complexes determined different effects: both RuCl₂Tol and RuPhenCym exhibited a rescue of cell viability confirming that both complexes act as inhibitors of toxic aggregation. Conversely, RuBipyCym, consistently with its limited anti-aggregative activity, was unable to reduce the cytotoxicity induced by Aβ_1–42.

Fig. 7 Effects of Ru-complexes on the cytotoxicity of Aβ_1–42 peptides in SH-SY5Y neuroblastoma cells. Cell viability was evaluated by MTT assay after incubation of Aβ_1–42 peptides in the absence or presence of Ru-complexes under stirring conditions at t = 0, 2, and 24 h. Control cells (untreated) were set as 100% viability. Data are presented as mean ± standard deviation (SD) from three independent experiments. Statistical analysis was performed using one-way ANOVA followed by Dunnett's multiple comparisons test, comparing all treatment groups to the control condition. No statistically significant differences were observed.

Experimental

Peptides and metal complexes syntheses

Aβ_1–42 and Aβ_21–40 were purchased from NovoPro Bioscience Inc. (Shanghai, China). All peptides were treated with 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) to guarantee a monomeric state, lyophilized and stored at −20 °C until use.

Ru complexes were synthetized as reported in.²⁶

Fluorescence assays

ThT emission assays were carried out in black plates (96 well) under stirring on fluorescence reader Envision 2105 (PerkinElmer). Measurements were collected every 7 min (λ_ex 440 nm and λ_em 483 nm). Assays were performed in duplicate at 25 °C employing a peptide concentration of 100 µM for Aβ_21–40, and 50 µM for Aβ_1–42 in 50 mM NaCl and 20 mM phosphate buffer (pH 7.4), using a ThT final concentration of 50 µM for Aβ_21–40, and 5 µM for Aβ_1–42. Different ThT concentrations were employed to ensure optimal signal-to-noise ratios for each peptide system. Different ratios with metal complexes (stock solutions 2 mM in water) were tested.

Circular dichroism (CD) spectroscopy

CD spectra of Aβ_1–42 at 50 µM and Aβ_21–40 at 100 µM, in 10 mM phosphate buffer pH 7.4, in the absence and presence of the metal compounds in 1 : 1 peptide to complex molar ratio, were registered on a Jasco J-815 spectropolarimeter (JASCO, Tokyo, Japan), at 25 °C using a 0.1 cm path-length quartz cuvette. CD spectra were recorded overtime using solutions prepared by the dilution of freshly prepared stock solutions of Aβ peptides (∼1 mM).³⁰ Deconvolutions of CD spectra were obtained by BESTSEL software (https://bestsel.elte.hu/).³¹

ESI-MS

The solution of Aβ_21–40 or Aβ_1–42 at a concentration of 50 μM in 15 mM ammonium acetate (AMAC) buffer at pH = 7.0 was incubated with metal compounds in a peptide to metal compound molar ratio of 1 : 1. The solutions were diluted 10 times with 15 mM AMAC and then analyzed using a LTQ XL ion trap mass spectrometer equipped with an electrospray ionization (ESI) source operating at a needle voltage of 3.5 kV and 320 °C with a complete Ultimate 3000 HPLC system, including a pump MS, an autosampler, and a photodiode array (all from Thermo Fisher Scientific). Spectra of the isolated peptides and the three complexes alone were recorded as controls.

SEM analysis

Aβ_1–42 (50 μM) and Aβ_21–40 (100 μM) alone and in the presence of Ru-complexes (1 : 1 peptide : complex molar ratio) in 10 mM phosphate buffer, pH 7.4, were morphologically analyzed after 6 h of aggregation using field-emission SEM (Phenom_XL, Alfatest, Milan, Italy). After this time, ∼30 μL of solution was drop-cast on an aluminum stub and vacuum dried to prepare each sample. For 75 s, a thin layer of gold was sputtered at a current of 25 mA. The sputter-coated samples were then introduced into the microscope chamber, and micrographs were acquired with a secondary electron detector (SED) operating at an accelerating voltage of 10 kV. The Ru-complexes and buffer alone (Fig. S5) were analyzed as controls.

Cell culture

SH-SY5Y human neuroblastoma cells (CRL-2266; ATCC, Manassas, VA, USA) were grown in 100 × 20 mm culture dishes at a density of 1 × 10⁶ cells per dish. Cells were maintained in DMEM (Corning; Cat. No. 10-013-CV) supplemented with 10% fetal bovine serum (Sial; Cat. No. YourSIAL-FBS-SA), 100 U mL⁻¹ penicillin, 100 μg mL⁻¹ streptomycin (Corning; Cat. No. 30-002-CI), and 25 mM HEPES (Sigma-Aldrich; Cat. No. H0887). Cultures were incubated at 37 °C in a humidified environment containing 5% CO₂ and were subcultured when approximately 80% confluence was reached.

MTT assay

Aβ_1–42 (stock solution 100 μM in 50 mM phosphate buffer at pH 7.4) in the absence and in the presence of the Ru-complexes at a 1 : 1 peptide to metal complex molar ratio (after 0, 2 and 24 h of stirring) were diluted in a cell culture medium at a final concentration of 25 μM and added to the cells for 24 h. Prior to cell treatment, samples were visually inspected to exclude precipitation. No visible aggregation or precipitation was observed under the experimental conditions. Control cells were incubated with phosphate buffer diluted in the cell culture medium at the same final concentration used for the peptide. Following treatment, cell viability was assessed using the MTT assay according to the manufacturer's protocol. Briefly, 100 μL of MTT solution (final concentration 0.5 mg mL⁻¹; Sigma-Aldrich) was added to each well and incubated for 2 h. The medium was then removed, and DMSO was added to dissolve the formazan crystals generated by metabolically active cells. Absorbance was measured at 570 nm using a microplate reader (VICTOR Nivo™, Revvity). A background absorbance value of 0.2, determined from cell-free wells treated with MTT, was subtracted from all readings. Cell viability was calculated by normalizing the mean absorbance values of cells treated with Aβ_1–42, in the absence or presence of Ru-complexes, to those of buffer-treated control cells and expressed as a percentage of control. In addition, the intrinsic cytotoxic effects of the Ru complexes were independently evaluated.

Conclusion

The binding of Ru(II) complexes to Aβ peptides is highly tuned by the nature of ligands in the coordination sphere: charge, aromaticity, chelate state.^4,29 This study demonstrates that η⁶-arene Ru(II) complexes can effectively modulate amyloid-β aggregation and proposed mechanisms are highly dependent on peptide length, composition and ligand fields. Among the investigated compounds, RuPhenCym and RuCl₂Tol emerged as the most active species, displaying complementary modes of action. RuPhenCym predominantly interferes with Aβ_1–42 aggregation via aromatic and supramolecular interactions, leading to the stabilization of soluble β-sheet-containing species and a pronounced reduction of fibril growth. The presence of soluble β-structured species is inferred from spectroscopic data but cannot be conclusively established without complementary size-distribution techniques such as DLS or SEC. In contrast, RuCl₂Tol, characterized by a more labile coordination sphere, forms stable peptide adducts via coordination mechanism and efficiently disrupts aggregation of both Aβ_1–42 and Aβ_21–40. While these observations are consistent with distinct coordination- versus supramolecular-driven mechanisms, definitive identification of binding sites and aggregate nature will require further structural investigations. Non-covalent interactions are proposed based on indirect evidence, and further studies (e.g., NMR or computational modeling) will be required to identify specific binding sites. The ability of RuCl₂Tol and RuPhenCym to significantly rescue Aβ_1–42-induced cytotoxicity in neuronal cells establishes a direct link between the modulation of aggregation pathways and biological outcome, underscoring the relevance of metal-mediated control of amyloid assembly for mitigating neurotoxicity. Importantly, the lack of intrinsic cellular toxicity of the Ru complexes under the investigated conditions further supports their suitability as potential neurodrugs. Collectively, these results highlight η⁶-arene Ru(II) complexes as versatile platforms for the development of metal-based agents targeting pathological amyloid aggregation. The tunability of ligand frameworks and coordination lability enables selective engagement with toxic amyloid species, offering a rational foundation for the design of next-generation bioinorganic modulators with potential theranostic relevance in neurodegenerative disorders associated with amyloid misfolding.

Author contributions

Funding acquisition, conceptualization, original draft: D. M., supervision, review & editing D. M. and S. L. M.; investigation and formal analysis: S. L. M., C. D. S., D. F., F. R., A. A.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: Fig. S1 ThT-fluorescence assay of Ru complexes alone at 100 µM, Fig. S2. ESI-MS spectra of the Aβ_21–40 peptide in the absence (A) and presence of RuBipyCym (B) and RuPhenCym (C). The peaks marked with bn derive from spontaneous source fragmentation of Aβ_21–40. Fig. S3. ESI-MS spectra of the Aβ_1–42 peptide in the absence (A) and presence of RuBipyCym (B) and RuPhenCym (C). The peaks marked with bn derive from spontaneous source fragmentation of Aβ_1–42. Fig. S4. MTT assay of Ru complexes at t = 0, 2 and 24 h, Table S1: t_1/2 and maximum intensity values related to ThT experiments. Fig. S5. SEM micrographs of phosphate buffer at 10 mM. Tables S2 and S3: Deconvolution of CD spectra reported in Fig. 3, Table S4: Experimental m/z values detected in the spectra of Aβ_21–40 and Aβ_1–42 alone and with the addition of RuBipyCym. Table S5: Experimental m/z values detected in the spectra of Aβ_21–40 and Aβ_1–42 alone and with the addition of RuPhenCym. See DOI: https://doi.org/10.1039/d6dt00454g.

Acknowledgements

This work was supported by #NEXTGENERATIONEU (NGEU), Ministry of University and Research (MUR), National Recovery and Resilience Plan (NRRP), project MNESYS (PE0000006) – A Multiscale integrated approach to the study of the nervous system in health and disease (DN. 1553 11.10.2022). D. F. acknowledge the Italian Ministry of Health – Ricerca Corrente Project.

References

1. M. Oprica, E. Hjorth, S. Spulber, B. Popescu, M. Ankarcrona, B. Winblad and M. Schultzberg, J. Cell. Mol. Med., 2007, 11, 810–825.
2. S. La Manna, C. Di Natale, V. Panzetta, P. A. Netti, A. Merlino, K. Kowalski and D. Marasco, Inorg. Chem. Front., 2024, 11, 6577–6587.
3. A. Baumketner, S. L. Bernstein, T. Wyttenbach, N. D. Lazo, D. B. Teplow, M. T. Bowers and J. E. Shea, Protein Sci., 2006, 15, 1239–1247.
4. S. La Manna, C. Di Natale, V. Panzetta, M. Leone, F. A. Mercurio, I. Cipollone, M. Monti, P. A. Netti, G. Ferraro, A. Terán, A. E. Sanchez-Pelaez, S. Herrero, A. Merlino and D. Marasco, Inorg. Chem., 2023, 63, 564–575.
5. G. Leech, A. L. Acosta, S. Mahato, P. C. Barrett, R. O. Hodges, S. A. McFarland and T. Storr, Chem. Sci., 2025, 16, 20914–20923.
6. Y.-B. Peng, C. Tao, C.-P. Tan and P. Zhao, J. Inorg. Biochem., 2021, 224, 111591.
7. N. A. Vyas, S. N. Ramteke, A. S. Kumbhar, P. P. Kulkarni, V. Jani, U. B. Sonawane, R. R. Joshi, B. Joshi and A. Erxleben, Eur. J. Med. Chem., 2016, 121, 793–802.
8. S. Singh, G. R. Navale, S. Agrawal, H. K. Singh, L. Singla, D. Sarkar, M. Sarma, A. R. Choudhury and K. Ghosh, Int. J. Biol. Macromol., 2023, 239, 124197.
9. R. M. Hacker, D. M. Grimard, K. A. Morgan, E. Saleh, M. M. Wrublik, C. J. Meiss, C. C. Kant, M. A. Jones, W. W. Brennessel and M. I. Webb, Dalton Trans., 2024, 53, 18845–18855.
10. J. C. Bataglioli, L. M. Gomes, C. Maunoir, J. R. Smith, H. D. Cole, J. McCain, T. Sainuddin, C. G. Cameron, S. A. McFarland and T. Storr, Chem. Sci., 2021, 12, 7510–7520.
11. G. Leech, A. L. Acosta, S. Mahato, P. C. Barrett, R. O. Hodges, S. A. McFarland and T. Storr, Chem. Sci., 2025, 20914–20923.
12. D. Florio, M. Cuomo, I. Iacobucci, G. Ferraro, A. M. Mansour, M. Monti, A. Merlino and D. Marasco, Pharmaceuticals, 2020, 13, 171.
13. G. K. Yawson, S. E. Huffman, S. S. Fisher, P. J. Bothwell, D. C. Platt, M. A. Jones, G. M. Ferrence, C. G. Hamaker and M. I. Webb, J. Inorg. Biochem., 2021, 214, 111303.
14. G. Son, B. I. Lee, Y. J. Chung and C. B. Park, Acta Biomater., 2018, 67, 147–155.
15. G. K. Yawson, M. F. Will, S. E. Huffman, E. T. Strandquist, P. J. Bothwell, E. B. Oliver, C. F. Apuzzo, D. C. Platt, C. S. Weitzel and M. A. Jones, Inorg. Chem., 2022, 61, 2733–2744.
16. J.-M. Suh, G. Kim, J. Kang and M. H. Lim, Inorg. Chem., 2018, 58, 8–17.
17. D. Florio, A. Annunziata, V. Panzetta, P. A. Netti, F. Ruffo and D. Marasco, Inorg. Chem., 2024, 63, 16001–16010.
18. A. De Santis, S. La Manna, I. R. Krauss, A. M. Malfitano, E. Novellino, L. Federici, A. De Cola, A. Di Matteo, G. D'Errico and D. Marasco, Biochim. Biophys. Acta, Gen. Subj., 2018, 1862, 967–978.
19. S. La Manna, P. L. Scognamiglio, V. Roviello, F. Borbone, D. Florio, C. Di Natale, A. Bigi, C. Cecchi, R. Cascella, C. Giannini, T. Sibillano, E. Novellino and D. Marasco, FEBS J., 2019, 286, 2311–2328.
20. K. L. Morris, A. Rodger, M. R. Hicks, M. Debulpaep, J. Schymkowitz, F. Rousseau and L. C. Serpell, Biochem. J., 2013, 450, 275–283.
21. Z. S. Al-Garawi, K. L. Morris, K. E. Marshall, J. Eichler and L. C. Serpell, Interface Focus, 2017, 7, 20170027.
22. D. Florio, S. La Manna, A. Annunziata, I. Iacobucci, V. Monaco, C. Di Natale, V. Mollo, F. Ruffo, M. Monti and D. Marasco, Dalton Trans., 2023, 52, 8549–8557.
23. F. Bossis and L. L. Palese, Biochim. Biophys. Acta, Proteins Proteomics, 2013, 1834, 2486–2493.
24. S. La Manna, D. Florio, J. A. Platts, E. Gabano, M. Ravera and D. Marasco, Dalton Trans., 2025, 54, 10416–10425.
25. S. La Manna, M. Leone, I. Iacobucci, A. Annuziata, C. Di Natale, E. Lagreca, A. M. Malfitano, F. Ruffo, A. Merlino and M. Monti, Inorg. Chem., 2022, 61, 3540–3552.
26. A. Annunziata, M. E. Cucciolito, M. Di Ronza, G. Ferraro, M. Hadiji, A. Merlino, D. Ortiz, R. Scopelliti, F. Fadaei Tirani and P. J. Dyson, Organometallics, 2023, 42, 952–964.
27. S. La Manna, D. Florio, I. Iacobucci, F. Napolitano, I. D. Benedictis, A. M. Malfitano, M. Monti, M. Ravera, E. Gabano and D. Marasco, Int. J. Mol. Sci., 2021, 22, 3015.
28. D. A. Butterfield and D. Boyd-Kimball, Brain Pathol., 2004, 14, 426–432.
29. S. La Manna, V. Panzetta, C. Di Natale, I. Cipollone, M. Monti, P. A. Netti, A. Terán, A. E. Sánchez-Peláez, S. Herrero and A. Merlino, Inorg. Chem., 2024, 63, 10001–10010.
30. B. Tizzano, P. Palladino, A. De Capua, D. Marasco, F. Rossi, E. Benedetti, C. Pedone, R. Ragone and M. Ruvo, Proteins: Struct., Funct., Bioinf., 2005, 59, 72–79.
31. A. Micsonai, F. Wien, L. Kernya, Y.-H. Lee, Y. Goto, M. Réfrégiers and J. Kardos, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, E3095–E3103.

---

Footnotes

† Co-first authors.

‡ Present address: IMDEA Nanociencia, C. Faraday, 9, Fuencarral-El Pardo, 28049 Madrid.

---