Protein misfolding is a common feature in several highly diffused (Alzheimer’s (AD) and Parkinson’s (PD)) or rare (Creutzfeldt–Jacob’s (CJD) or Niemann–Pick’s (NP)) human severe amyloid-related neurodegenerations and results in the formation and deposition of amyloid aggregates [3-6]. The amyloid superfamily is a crowded house, which includes well-known proteins, like Amyloid-b1-42 (Aβ1-42), a-Synuclein (aS), and Prion-Protein (PrP) implicated in AD, PD, and CJD, respectively as well as hormones such as Insulin or Calcitonin (CT) [7]. The exact molecular mechanisms at the base of the amyloid neurotoxicity are still unknown even though the existence of “common mechanisms” focused on low molecular weight Prefibrillar Oligomers (PFOs) and not on Mature Fibers (MFs) or monomers, have been proposed [8,9].

Major efforts have been made to clarify the link between oligomer structures and neurotoxicity, based on the analysis of their effects in vitro [10-12] (neuronal cultures and acute brain slices [13] and in vivo (direct administration in the brain [14,15]. However, the culprit has not yet been identified. As Benilova et al. [16] point out, this can be due to the intrinsic metastability of the aggregates leading to difficulties in the comparison of results from different research groups. However, there is a general agreement that small and soluble PFOs are toxic species [7].

Salmon Calcitonin (sCT) - a 32 amino acid polypeptide hormone used in the treatment of osteoporosis - is neurotoxic in vitro and forms amyloid aggregates at a very low rate [17]. Thanks to this characteristic, we successfully investigated the effects of native sCT-PFOs from the early stages, avoiding any fixing procedure [10,18]. We recently demonstrated that sCT-PFOs, but not MFs, induced abnormal permeability to Ca²⁺, reduced cellular viability, and impaired Long-Term Potentiation (LTP) and expression of Post-Synaptic Density (PSD-95) protein and pre-synaptic protein Synaptophysin. In our paper, we proposed that the two pre-existing neurotoxicity paradigms used to explain amyloid neurotoxicity, “membrane-permeabilization” and “receptor-mediation”, could not separately explain our data but must coexist, constituting a novel “unified neurotoxicity mechanism” [1].

We recently noticed the paper published by Yasumoto et al. in “The FASEB Journal” concerning High Molecular Weight-Aβ1-42 (HMW-Aβ1-42) and Low Molecular Weight-Aβ1-42 (LMW-Aβ1-42) oligomers inducing neurotoxicity via plasma membrane damage [2]. This paper follows other important papers published by the Ono’s and Teplow’s group in this field, as an example see [19,20].

Interestingly, using a different protein (Aβ1-42), they obtained experimental results very similar to our findings relative to sCT.

There are some issues we would like to address here:

1. Yasumoto’s and our samples can be matched: HMW-Aβ1-42 corresponds to sCT-PFOs.

Yasumoto et al. [2] clearly stated that the HMW-Aβ1-42 sample spans from trimers to Proto Fibrils (PFs), including tetramers, pentamers, and hexamers. We showed exactly the same in our paper for sCT-PFOs. Moreover, Yasumoto’s results concerning the Ca²⁺-influx (figure 5) and LTP (figure 7) alterations due to HMW-Aβ1-42 sample are very similar to our results obtained by sCT-PFOs (figure 2 and figure 5);

2. The two Yasumoto’s HMW-Aβ1-42 and LMW-Aβ1-42 samples are of the same kind.

This can be deduced from the small difference in the oligomer mean diameters (figure 1F) and by the similar biological effects they induced;

3. Yasumoto didn’t show results concerning Aβ1-42 monomers. This was likely due to the well-known high aggregation rate of Aβ1-42. We presume that Yasumoto et al. didn’t have an actual monomeric sample while the very slow aggregation rate of sCT allowed us to overcome this problem;

4. In both papers, the neuronal membrane disruption was indicated as the trigger event for neurotoxicity. In the title and in figure 8 Yasumoto et al. indicated “neuronal membrane disruption” as the main cause “..of Ab oligomer-induced neuronal injury and death”. This is the same conclusion of our paper dealing with sCT.

5. In both papers, blocking NMDA receptors abolished the neurotoxicity.

In this paper, concerning sCT, we proposed that the triggering event activating a “unified neurotoxicity mechanism” was the membrane permeabilization exerted by PFOs. Amyloid-pores, unable “per se” to heavily raise the intracellular Ca concentration, were capable of triggering “receptor-mediated” excitotoxicity. We based this conclusion on the protective effect obtained by blocking the NMDA receptors by MK-801 [1]. Notably, Yasumoto et al. reported the same result for Aβ1-42 by blocking NMDA receptors by Memantine (figure 5B) [2].

Here we speculate about the surprising similarity and complementarity of our and Yasumoto’s results.

Concerning oligomers, we noticed that our findings relative to sCT-PFOs were very similar (see item 1) to Yasumoto’s results relative to Aβ1-42, LMW, and HMW oligomers (see item 2).

Concerning monomers, we showed in our paper that sCT monomers were unable to activate the proposed “unified neurotoxicity mechanism” [1]. Despite Yasumoto et al. didn’t show results relative to monomers (see item 3), it is known that purified Aβ1-42 monomers are protective and non-toxic [21] and unable to produce “amyloid-pores” in an excised patch of membranes [22].

Based on these considerations, we think that the “unified neurotoxicity mechanism” we proposed for sCT [1] can also be applied to interpret Yasumoto‘s results relative to Aβ1-42.

More in particular, in both cases, the triggering event could be the formation of “amyloid-pores” described in the literature for both proteins [22,23]. In a recent paper, we described in detail this mechanism in the case of sCT-PFOs [18]. As summarized in figure 1, after the interaction with particular membrane structures rich in charged lipids as GM1-monosialogangliosides, named “lipid-rafts”, sCT-PFOs were incorporated in the lipid bilayer and formed channels able to permeabilize the neuron membrane. Conversely, monomers interacted with the same structures but, remaining outside the membranes, didn’t affect the ionic permeability (Figure1).

However, this phenomenon is only the trigger for the Ca²⁺ homeostasis unbalance and is not able “per se” to damage neurons without the involvement of the MNDA-receptors. This is demonstrated by the inhibition of the Ca²⁺-influx achieved by MK-801 for sCT [1] and Memantine for Aβ1-42 [2].

Thus, we wonder how it was possible to obtain such comparable results with such different proteins. We speculate that some common features in the hydrophobic profiles of the two proteins could be the structural basis for the assembly of PFOs with a similar hydrophobic “outfit”. We can see in Figure 2 that the two peptides are similar in dimension (32 and 42 residues) and that the reversed sequence of sCT and the normal sequence of Aβ1-42, show a core (highlighted by arrows) of about five highly hydrophobic residues (red) enclosed within mixed hydrophobic and polar segments (blue). Notably, the reversed sequence of sCT looks like to the specific segment of Aβ1-42 exposed to the medium. Consider that the last set of ten hydrophobic C-terminus residues of Aβ1-42, tend to be excluded by the medium while the remaining part tends to be exposed.

To better investigate the similarity in the sequences highlighted in figure 2, we simulated the hydrophobic profiles by means of three hydrophobicity scales. In particular, the hydrophobic value of the single amino acid residue is linearly averaged with respect to the ±2 nearest residues (window size of 5 with a weight of 50% at the edge). The results are reported in Figure 3.

In the Kyte & Doolittle’s (K&D) scale [24] the hydrophobic character is derived from the physicochemical properties of the amino acid side chains and results mainly from the inspection of the amino acid structures. In the Rose’s [25] and Janin’s [26] scales, starting from the protein 3-D structures, the hydrophobic character is defined as the tendency for a residue to be found inside of a protein rather than on its surface.

Notably, the hydrophobic profiles of the two proteins, obtained using all scales after a shift of 6 residues, showed a “common profile” in the central part.

In conclusion, our idea is that this “common hydrophobic core” leads, under similar chemical-physical conditions, to a “common PFOs structure” characterized by a common hydrophobic/polar “outfit” able to activate the “unified neurotoxicity mechanism”, described for sCT [1]. Moreover, in our opinion, the fast aggregation of Aβ1-42, concerning the slow sCT could be due to the long hydrophobic tail of Aβ1-42 missing in the sCT (see figure 2). The hydrophobic force occurring between the tails of the Aβ1-42 constituting the aggregate would at the basis of the faster formation of the β-sheet configuration in the core of PFOs and eventually in the MFs.

The research into the existence of this “common outfit” in the amyloid family, performed by molecular dynamics simulations combined with biophysical techniques, would be very useful to elucidate this outstanding similarity and to design innovative strategies to counteract the adverse effects exerted on cellular membranes at the basis of such important diseases.

This work was supported by the Italian “Ministero della Salute” with the “Progetto Ordinario di Ricerca Finalizzata (RF-2013-02355682)” entitled: “Calcitonin oligomers as a model to study the amyloid neurotoxicity: the focal role played by lipid rafts in the prevention and cure”.

The authors declare no conflict of interest.