Amyloid aggregation is a degenerative process characterized by deposition at tissue levels of organized insoluble super-molecular protein assemblies with a typical cross-β secondary structure. Such degeneration gives rise to amyloidosis, a composite range of diseases classically divided into neurodegenerative (e.g., Alzheimer’s disease, Parkinson’s disease, etc.) and systemic (e.g., cystic fibrosis, light chain amyloidosis) amyloidosis. More than 20 different human proteins, intact or fragmented, proved their amyloidogenicity in vivo, among which we may count amyloid β (Aβ) peptide (in Alzheimer’s disease), α-synuclein (in Parkinson’s disease), islet amyloid polypeptide (in type II-diabetes), light chains of immunoglobulins, variants of human lysozyme [1
], and transthyretin (TTR) [3
]. It is nowadays widely accepted that amyloid aggregation is a general tendency of polypeptide chains [4
] that, in fact, may be induced to form amyloid aggregation in appropriate conditions [7
Lysozyme, a 130-residue-long bacteriolytic enzyme largely distributed in different tissues, organs, and external secretions, has been highlighted as an interesting model for the study of amyloid aggregation. Although wild-type lysozyme is not directly involved in amyloid diseases, several naturally occurring single point mutations (e.g., Ile56Thr, Phe57Ile, Trp64Arg, and Asp67His) are connected with familial non-neuropathic systemic amyloidosis [8
]. In addition, the wild-type lysozyme either from humans, horses, or hens, under appropriate conditions, is able to form amyloid fibrils in vitro [9
In this work, we used the hen egg-white lysozyme (HEWL—14.3 kDa, 129 amino acids, 40% identity with the human enzyme) inducing its aggregation through a heat treatment in acidic conditions [12
]. Despite the fact that HEWL is not associated with in vivo diseases [13
], it has been demonstrated that high temperatures and low pH induce the breakage of X-Asp peptide bonds, leading to the formation of peptide fragments (among which one contains the residues corresponding to those mutated in human familiar diseases Ile56Thr and Asp67His). Such fragments have a high tendency to form amyloid aggregates [11
] and amyloid-like fibrils in a few days. In addition, a direct toxic effect of HEWL aggregates added to cell cultures or injected in rat brains, mimicking the toxic effect of Aβ peptide, has been demonstrated [14
A substantial body of literature over the years documents that extracts from natural herbs and plants or common dietary elements such as wine [15
] or green-tea [16
] are of great benefit to general human health, mainly due to their antioxidant power [17
]. Among the molecules proposed to be essential for achieving such effects are polyphenols [19
], a wide and heterogeneous group of substances well characterized in terms of structure [20
]. Apart from protecting cells from oxidative stress, several phenolic compounds have been shown to be effective in inhibiting amyloid aggregation in various protein models such as transthyretin, microglobulin, α-syuclein, Aβ peptide and lysozyme, with supposed specific action mechanisms independent of their antioxidant properties [21
]. Of particular relevance for this work, catecholic- and hydroquinone-containing phenols were reported to act as inhibitors of amyloid aggregation for their ability to induce quinoprotein formation [26
Dihydrobenzo[1,4]oxathines, and in particular compounds possessing a polyhydroxylated 4-thiaflavane skeleton, have received great attention in chemical, medicinal, and pharmaceutical research. During the last decades, their syntheses as well as their abilities as antioxidants, hypertensive agents, estrogen receptor modulators, adrenoreceptor antagonists, and artificial sweeteners have been reported in papers and patents [27
]. As shown in Figure 1
, such 4-thiaflavane derivatives are structural mimickers of several natural polyphenols. In particular, compounds 4
(see Figure 1
) are thia-substituted biomimetic examples of catechin derivatives showing a free catechol residue on the B ring. In this work, we investigated whether selected hydroxylated 4-thiaflavane derivatives may act as an inhibitor of amyloid aggregation for the widely used and accepted HEWL amyloid model.
The possibility of blocking or reverting amyloid aggregation with small molecules may have a great impact on worldwide health. Amyloidosis, a group of over 20 different and heterogeneous diseases, are directly linked to the accumulation of amyloid matter into organs and are severely affecting human population. Amyloidosis, including neurodegenerative diseases such as Alzheimer’s disease or Parkinson’s disease as well as systemic diseases such as reactive systemic amyloidosis, TTR, and light chain amyloidosis are increasingly recognized as important death factors for public health systems.
Many studies have so far demonstrated that molecules with peculiar structural features are able to impair the formation or the elongation of amyloid fibrils, among which are a number of synthetic or natural compounds with polyphenolic rings [23
]. Several activity–structure studies have shown that, though the effect of the addition of such exogenous compounds could be beneficial in some cases in terms of the prevention of amyloid formation or the reduction of amyloid load, the activity spectrum is bound to a defined experimental set and confined to certain amyloid related proteins and peptides, such as Aβ peptides, α-synuclein, and TTR.
In this study, we used HEWL as a model of amyloid aggregation. The reason for this choice was based on the need to decouple the effect of metal ions on aggregation to the documented chelating properties of 4-thiaflavanes [35
]. In fact, many amyloid models have been found to be deeply affected by the presence of copper or iron ions, and in traces, leading to alterations of the aggregation kinetics, of the lag phase, or of the morphology of the resulting fibrils. Such effects have never been reported for HEWL, so we considered it a good model for our class of compounds, allowing us to establish a direct link between the structure of the molecules and the effect on aggregation inhibition. In fact, we found that the poor solubility of compounds such as 1
is not of benefit in reducing amyloid signals, which is contrary to what has been found for other compounds, such as curcumin, which has been declared to be extremely active in aggregation in vitro despite its absolute insolubility in aqueous buffers. In our conditions, the most active 4-thiaflavanes were molecules with a higher degree of hydroxylation, leading to increased solubility and a reasonably higher capability of interacting with nascent aggregates, blocking their elongation into higher order assemblies such as mature fibrils. Moreover, despite the lack of a direct relationship between antioxidant activity and the inhibition of amyloid aggregation, it is worth mentioning that compounds 4
, containing a catechol moiety, showed a much higher antioxidant activity with respect to compounds 1
, and 3
A relevant body of literature exists trying to shed light on the true nature of the cytotoxicity induced by amyloid material. Initially, it was strongly believed that toxicity was due to early aggregates only (a stage at which large portions of hydrophobic protein regions are exposed, waiting to gain stabilization, with polymerization and burial of these regions inside highly ordered structures). Recently, it has been shown that this cannot be considered a general rule because of the existence of proteins for which mature fibrils largely surpass the toxicity of early aggregates [37
]. HEWL has been previously shown to behave in the “classic” way, losing toxicity as fibrils grow and as early aggregates are progressively sequestered from the medium to the fibril. We verified this behavior in our experimental conditions, selecting day 3 as the one displaying the maximal toxicity that was progressively lost at longer times, until day 10, when fibrils were completely mature, as shown by AFM. At this time point, we measured extremely reduced ThT signals for compounds 3
, and 5
, but we found a toxicity comparable to that of untreated HEWL. When the incubation time was increased to allow HEWL controls to develop into non-toxic mature fibrils, the HEWL samples treated with active compounds were found to maintain a toxicity comparable to samples tested at early aggregation stages. Nevertheless, the ThT signal proved to be, in most cases, highly reduced. Our results seem to suggest that the effect of 4-thiaflavane derivatives was to impair (in part or almost completely, depending on the compound) the elongation of HEWL fibrils, leading to the formation of round particles with diameters similar to that of amyloid species and that are highly toxic to cells, stable over time, and poorly responsive to ThT binding.
According to a recent finding, molecules containing catecholic and (less efficiently) hydroquinone moieties may be able to drive the formation of quinoproteins, i.e., to covalently modify proteins in a hot acidic environment [26
] and hamper HEWL fibril formation. Compounds 4
bear in the B ring a catechol-like structure that, given aggregation conditions (acidic pH and high temperature), may induce protein derivatization and form ortho
-quinonic adducts. Similarly, compounds 1
, and 3
present a phenol ring that, in the same conditions, can undergo acid hydrolysis and further oxidation, leading to the formation of p-quinoic adducts. Our results on hydroxylated 4-thiaflavanes seem to suggest that the abovementioned reaction mechanisms may drive the observed reduction in amyloid formation without a loss of cell toxicity. The ability of compound 3
to drastically reduce the load in fibrils (despite its structural similarity with the less effective compounds 1
) partly contrasts this vision and suggests that additional research is required to definitively validate the above hypothesis.
Although our findings strongly discourage the usage of the investigated 4-thiaflavanes as potential drugs for amyloidosis, the opposite route seems interesting. Since these compounds have been shown to be safe for cells at the tested doses, their contribution to cell toxicity following incubation with toxic aggregates is minimal, allowing a coherent assay of aggregate toxicity. In fact, one of the most challenging aspects of amyloid studies is that intermediate, toxic structures are transient in nature and their effect is confined in a restricted, highly variable temporal frame.
Further biophysical studies will be needed to elucidate the structures and the features of the toxic aggregates stabilized by the 4-thiaflavane derivatives tested in this work and their mechanism of action. Of particular interest is the observation that such particles give reduced or impaired ThT signals, a fact that is counterintuitive given their toxicity, which is similar to 3-day HEWL early aggregates, which proved to give rise to aggregates that efficiently bind ThT. Furthermore, it cannot be excluded that the structural features of such aggregates may be different in response to different compounds, although the results collected so far seem to suggest as the major player a population of small, globular particles with a quite uniform size.
If confirmed on other amyloid systems, these results suggest that hydroxylated 4-thiaflavanes are promising stabilizing agents for toxic aggregates that are useful for studying, for example, the effect of single or multiple mutations on a uniform and homogeneous population of toxic species.