The peptide encompassing the amino acid sequences SSSHPIFHRGESFV-NH₂, NGF(1- 14), and its acetylated form, Ac-SSSHPIFHRGESFV-NH₂, Ac-NGF(1–14), were synthesized with the C-termini amidated as previously reported (Travaglia et al., 2011). The fluorescent peptides, NGF(1–14)FAM and Ac-NGF(1–14)FAM, were labeled with 5,6-carboxyfluorescein through the side chain of an amidated additional lysine residue (FAM) (purchased from CASLO, Lyngby, Denmark).

A FV1000 laser-scanning microscope (Olympus), equipped with diode (LD405), Argon multiline (458, 488, 515) and HeNe (543 and 633) lasers, fitted with a low chromatic aberration objective PlanApo60x (NA: 1.4, oil immersion, W.D.: 0.12 mm) was used. The images were scanned with Kalman filtering on at the resolution of 512 × 512 pixels. For multichannel imaging, fluorescent dyes were imaged sequentially to eliminate cross talk between the channels, namely: (i) the blue (ex405/em 425–475), for the emission of the DAPI-stained nuclei, (ii) the green (ex488/em 500–530), for the emission of the carboxyfluorescein (FAM) group in the NGF(1–14)FAM and Ac-NGF(1–14)FAM peptides, and (iii) the red (ex543/em 560–700), for the BODIPY moiety of CS1 copper probe (Miller et al., 2006).

PC12 cells were seeded on glass bottom dishes (WillCo Wells B.V., Amsterdam, NL), pre-coated with 0.01% polylysine, at a density of 5 × 10⁴, and maintained in D-MEM complete medium for 24–36 h. The day of the experiment cells were rinsed with serum-free medium.

Live cell imaging experiments of the peptide cellular uptake were performed directly on the microscope stage operating in xy-time scan mode, with a total scan time of 20 min and a scan interval set to 30 s. More in detail, after the first 1–2 scans, peptides were quickly added from a 100X concentrated stock solution in water to the Petri dish containing the cells in 1 mL DMEM. The solution was vigorously mixed to ensure a homogeneous peptide dilution to the final concentration of 10 μM in the whole liquid volume, in order to avoid artifacts in the real time live cell imaging owing to different diffusion rates of the peptide molecules related to concentration gradients. The acquisition parameters were kept constant for all the experiments. Confocal imaging of metal trafficking was performed by treatments of PC12 cell for 5 min with the peptide solution (10 μM), either NGF(1–14) or Ac-NGF(1–14) in DMEM. After that, cells were rinsed (2 × 1mL) with phosphate buffer saline solution (PBS, 10 mM, pH = 7.4 at 25°C) and stained by 15 min incubation with CS1 intracellular monovalent copper probe (final concentration of 2 μM from 1 mM stock solution in DMSO) and the cell-permeant nuclear counterstain Hoechst 33342 (NucBlue® Live ReadyProbes® Reagent, Life Technologies), followed by buffer rinsing (2 × 1 mL). In two parallel experiments, before the treatments with the peptide NGF(1–14) and Ac-NGF(1–14), cells were pre-incubated in the medium supplemented with 50 μM of copper chelator BCS (bathocuproine disulfonic acid disodium salt, purchased from Sigma Aldrich, from a 10 mM stock solution in DMSO) for 3 h or 100 μM of CuSO₄ for 1 h. A 0.5% (v/v) DMSO was added in all conditions as control. At the end of treatments and staining, the cells were fixed in fresh 2% paraformaldehyde and deeply rinsed (3 × 2 mL) with PBS.

The N-terminal domain of NGF binds specifically to domain 5 of TrkA (TrkA-D5), affording distinct weak non-covalent interactions (Travaglia et al., 2015). Here we aimed to assess, through molecular simulations, whether the protection of the N-terminal amino group by acetylation alters the structural features of NGF(1–14) upon the binding toTrKA-D5.

Analogously with the simulations carried out for the wild-type NGF(1–14), parallel tempering simulations were run in parallel, in order to sample the conformations of Ac-NGF(1–14) at physiological pH. Three main clusters were observed accounting for 68%, 22%, and 10% of population percentage.

The simulated conformations showed a tendency to preserve a loop state that can be further stabilized via backbone hydrogen bonds in the central domain, to form a short alpha helix (Figure 1A). Such an alpha helix conformation, also observed in NGF(1–14) (Travaglia et al., 2015), was thus preserved upon the acetylation of the N-terminal amino group. In general, the various conformational states of Ac-NGF(1–14) were not affected from the N-terminal acetylation. However, upon docking the Ac-NGF(1–14) with TrkA-D5 relevant differences were observed in the three main binding poses (Figure 1B).

In the first binding pose, the acetylation of the amine group caused S1 moving far away from TrkA-D5, resulting in a different rearrangement of the residues of Ac-NGF(1–14); H4 faces H343 of TrkA-D5, at variance with the contact with H291 observed in NGF(1–14).

The second binding pose indicated that V14 of Ac-NGF(1–14) approaches H343 of TrkA-D5, a contact also observed in NGF(1–14). However, the contacts involving S1 of NGF(1–14) with TrkA-D5 were completely absent in Ac-NGF(1–14), owing to the acetylation of the N-terminal amine group.

The third binding pose highlighted a close contact of H8 of Ac-NGF(1–14) with E295 of TrkA-D5, E11 of Ac-NGF(1–14) with H291 of TrKA-D5 and a C-terminal contact involving V14 of AcNGF(1–14) with H343 of TrkA-D5. These contacts, observed also in NGF(1–14), were consistent with the other binding poses and, at variance with NGF(1–14), the S1 residue of Ac-NGF(1–14) was not facing any residues of TrkA-D5 (Figure 1C).

It is worth noting that the protection of the free-amino N-terminal group leads to the adoption of hydrogen bonds and salt-bridge interactions between the Ac-NGF(1–14) peptide and TrkA-D5. The number of non-covalent interactions was slightly higher for the Ac-NGF(1–14), if compared with those formed in NGF(1–14) peptide (Figure 2). Specifically, the following interactions were involved: (i) the OH groups of S1 and S2 side chains with the carbonyl groups of A293, E295 of TrkA-D5, (ii) the H4 imidazole hydrogen with the OH group of S304 of TrkA-D5, (iii) the E11 carboxyl group both with R289 amine side chain residue, forming salt-bridge. The E11 carboxyl group can also be engaged in a hydrogen bond with the imidazole hydrogen of H291 of TrkA-D5.

Interestingly, both peptides show the tendency to form a higher number of non-covalent interactions than in the co-crystallized protein structure (Wiesmann et al., 1999).

The acetylation of the N-terminal amino group of NGF(1–14) thus induces distinct structural features toward the binding with TrkA-D5, with a tendency to form weak non-covalent interactions.

In PC12 cells neurite outgrowth was used as a marker of differentiation. We recently showed that NGF(1–14) was unable to induce a visible differentiated phenotype (Travaglia et al., 2015). Since the neurite outgrowth is a macro marker of neuronal differentiation, the profile expression of specific genes involved in neuronal develop and differentiation, Map2, Tubb3, Elavl4, GAP-43 were evaluated in PC12 cells, untreated or treated for 3 days with the protein (NGF, 50 ng/ml) or the peptides (NGF(1–14) or Ac-NGF(1–14), 10 μM) in the presence or absence of CuSO₄ or ZnSO₄ (1 μM) (Figure 3). CuSO₄ or ZnSO₄ induced a two-fold increase of the Map2 and Elavl4 mRNA levels, although only for Map2 were significant (Figures 3A,B), while NGF elicited a 2.5-12-fold increase of Map2, Tubb3, Elavl4, and GAP-43 mRNA levels compared to the untreated control (p < 0.001). This effect significantly enhanced, up to two-fold, by CuSO₄ or ZnSO₄ addition. Conversely, Map2, Tubb3, Elavl4, and GAP-43 mRNA levels were not affected by CuSO₄ and ZnSO₄ addition in the cells stimulated with NGF(1–14) or Ac-NGF(1–14).

Endocytic trafficking of neurotrophins and their receptors is a fundamental feature of neurotrophin signaling. It is unclear whether the different receptor localization causes different intracellular pathway and thus different biological activities. Literature data show that neuronal differentiation requires the internalization of the ligand-receptor complex to clathrin-coated vesicles, while cell survival is induced by receptors located on the cell membrane.

Figure 4 shows that NGF induced TrkA receptor internalization in absence of CuSO₄ and ZnSO₄; the presence of metal ions slightly inhibited this effect (Figures 4A,B).

Conversely, CuSO₄ and ZnSO₄ elicited the NGF(1–14)-induced internalization of TrkA. The acetylated peptide showed a weak receptor internalization not affected by the presence of CuSO₄ and ZnSO₄.

The P75^NTR internalization paralleled the TrkA pattern (Figures 4A,C). Similar to TrkA receptor internalization, NGF activity was more inhibited by CuSO₄ and almost totally by ZnSO₄.

The TrkA/ P75^NTR internalization ratio decreased in the presence of metal ions for NGF and NGF(1–14) but not for Ac-NGF(1–14) (Figure 4D).

To evaluate whether the metals in the medium are essential in the TrkA signaling triggered by NGF and NGF(1–14) western blot analyses were performed in PC12 cells un-pre-treated or pre-treated 24 h before with BCS (a membrane-impermeable extracellular strong metal chelating agent), in the presence or absence of CuSO₄ and ZnSO₄ (Figure 5).

In cells incubated with NGF (50 ng/ml), Tyr-490 phosphorylation of TrkA was partially inhibited by pre-treatment with BCS (50 μM) (P < 0.01); this effect was partially restored by the presence of CuSO₄ and ZnSO₄ (P < 0.05) (Figure 5A).

In cells incubated with NGF(1–14) in the presence or in absence of CuSO₄, Tyr-490 phosphorylation of TrkA, was significantly more potent (P < 0.05). The presence of BCS totally blocked the Tyr-490 phosphorylation of TrkA stimulated by NGF(1–14) (50 μM) (P < 0.01); this effect was partially inhibited by the presence of metal ions (p < 0.05) (Figure 5B).

The Tyr-490 phosphorylation of TrkA stimulated by Ac-NGF(1–14) (50 μM) was inhibited by pre-treatment with BCS, but this effect was not restored by the presence of CuSO₄ and ZnSO₄ (Figure 5C). The AKT phosphorylation on Ser-473, in response to the stimulation with the ligands used, followed a pattern similar to that observed for TrkA phosphorylation, although ZnSO₄ was less effective to restore the inhibitory effect of BCS (Figure 6). AKT phosphorylation on Ser-473 was significantly more activated in the presence of CuSO₄ only when stimulated with NGF(1–14) (P < 0.05) (Figures 6A,B).

The ERK 1/2 phosphorylation was also inhibited by pre-treatment with BCS (Figure 7), but not in the presence of metal ions when stimulated with NGF(1–14) (P < 0.05) (Figure 7B). On the other hands, ERK 1/2 phosphorylation was activated in the presence of CuSO₄ when stimulated with either NGF or NGF(1–14) (P < 0.05) (Figures 7A,B) and in the presence of ZnSO₄, only for cells stimulated with NGF(1–14) (P < 0.05) (Figure 7B).

It must be noted that in a previous work NGF(1–14) 100 μM found unable to induce ERK1/2 phosphorylation in PC12 cells (Travaglia et al., 2015). The different effect observed in the present work is explained in terms of the aging of the cell model used. Indeed, although PC12 cells were used in both cases, here new early passages (<20) PC12 cells were used, with an expected higher responsiveness to NGF (Kinarivala et al., 2016).

The inhibitory effect of BCS on CREB phosphorylation at Ser 133 was evident in response to both NGF and NGF(1–14) (P < 0.01) (Figure 8). When stimulated with NGF or NGF(1–14), the Ser-133 phosphorylation of CREB increased markedly, in the presence of CuSO₄ (Figures 8A,B). This effect less evident in cells stimulated with Ac-NGF(1–14) (Figure 8C).

Brain-derived neurotrophic factor (BDNF) is a neurotrophin modulating survival, neurogenesis and differentiation of neuronal cells, the branching and survival of differentiated neurons, and the formation and maturation of the dendritic spine and synapses, influencing learning and memory.

To investigate the effect of NGF(1–14) on BDNF mRNA levels, in the presence or not of metal ions, PC12 cells were untreated or treated for 24 h with NGF (50 ng/ml), NGF(1–14) (10 μM), or Ac-NGF(1–14) (10 μM) in the presence or absence of CuSO₄ or ZnSO₄ (1 μM).

Figure 9 shows that CuSO₄ and ZnSO₄ a 3- and 5-fold increase of the BDNF mRNA levels respectively (P < 0.01 and P < 0.001). The NGF a 5-fold increase of BDNF mRNA levels compared to the untreated control (P < 0.01); this effect significantly in the presence of CuSO₄ or ZnSO₄ (P < 0.05).

Remarkably, NGF(1–14) and Ac-NGF(1–14) respectively a 5- and 8-fold increase of the BDNF mRNA levels (P < 0.01 and P < 0.001). At variance with NGF, for NGF(1–14) a significant increase of the BDNF mRNA levels in comparison to NGF(1–14) alone (P < 0.05) was found in the presence of CuSO₄ or ZnSO₄. The presence of metals not affect the increase of BDNF mRNA induced by Ac-NGF(1–14).

Live cell confocal imaging in real-time of the cells upon the addition of N-terminal free amino peptide and its acetylated form was performed to visualize, at cellular and nuclear level, the peptides uptake by the PC12 cells. To this purpose, fluorescent NGF(1–14)FAM and Ac-NGF(1–14)FAM, with the carboxyfluorescein (FAM) moiety covalently bound to the peptide C-terminus, were used.

The representative time-course images showed in Figure 10 pointed out that: (i) the cellular uptake of the peptides was a quick process, occurring on the timescale of minutes at the used experimental conditions (i.e., 10 μM peptide at 37°C); (ii) the peptide molecules completely permeated the PC12 cells, including the nuclei; (iii) the process of cellular uptake was reversible, as the cells, turned-on by the green emission of the FAM-labeled peptide molecules, rapidly recovered to their original autofluorescence. The representative images for the whole set of samples are reported in Figure S1.

Figure 11 reports the quantitative analysis of the FAM fluorescence (in terms of integrated density, ID, values) for NGF(1–14)FAM (Figure 11A), Ac-NGF(1–14)FAM (Figure 11B), NGF(1–14)FAM:Cu⁺² (Figure 11C), and Ac-NGF(1–14)FAM:Cu²⁺ (Figure 11D).

The evident trend displayed that both NGF(1–14)FAM and Ac-NGF(1–14)FAM were internalized by the cells. The uptake of Ac-NGF(1–14)FAM (Figure 11C) was slightly lower than the cell entry measured for NGF(1–14)FAM (Figure 11A). However, whereas the latter exhibited an exponential ID decay, which can be interpreted as an efflux process of the peptide after the initial cellular uptake (in the time scale investigated and at the used experimental conditions), the acetylated peptide showed a sinusoidal curve trend, likely related to dynamic efflux/re-entry processes.

For both peptides, the two different trends in the measured IDs, i.e., exponential decay for NGF(1–14)FAM and sinusoidal for Ac-NGF(1–14)FAM, were maintained also when supplemented to cells together with copper ions.

In addition, the ratio of the fluorescence intensity measured in the nuclei with respect to that measured in the whole cell (ID_nuclei/ID_total), revealed another significant difference between the free-amino peptide and the acetylated one. Indeed, at the maximum of cell internalization (i.e., about 1.5 min on the x-axis in the plots of Figure 11), such a ratio was about 25% for NGF(1–14)FAM, both in the absence and in the presence of copper, about 18% for Ac-NGF(1–14)FAM and 15% for Ac-NGF(1–14)FAM:CuSO₄. The gap observed for the acetylated peptide in the absence or presence of copper ions was still maintained and even increased in the minutes following the initial uptake, thus Ac-NGF(1–14)FAM exhibited a generally lower presence in the nuclei compared to the total peptide uptake.

To evaluate the effect of NGF(1–14) and Ac-NGF(1–14) on the intracellular copper trafficking, confocal microscopy imaging was used to scrutinize the cells stained with the intracellular probe of monovalent copper CS1 (Rampazzo et al., 2011).

The micrographs in Figure 12 show the response to the Cu⁺ reporter in a comparison of the PC12 cells for the different treatment conditions, i.e., free-amino vs. acetylated peptide, basal vs. copper-supplemented and copper-deprived medium.

The red emission of the fluorophore component (the BODIPY) of CS1 was enhanced when the Cu⁺ chelator moiety (a thioether-rich receptor) binds to the intracellular monovalent copper (Miller et al., 2006). Figure 12 clearly demonstrates that CS1 emission strongly increased for the cells treated with NGF(1–14) (Figure 12B) with respect to the control (Figure 12A), with a red fluorescence visible also in the nuclei. This observation is explained in terms of NGF(1–14) that acts as ionophore, i.e., able to bind extracellularly the Cu²⁺ ions in the basal medium and transport them at the cell membrane, where the redox conversion to the intracellular Cu⁺ occurs (Nevitt et al., 2012).

Since copper content in the basal media can reach concentrations up to micromolar (Huang et al., 2004), the experiments were repeated by supplementing the cell medium, before the peptide addition, with 100 μM CuSO₄ for 1 h (Figures 12D–F) or with BCS pre-tretment for 3 h (Figures 12G–I).

As to cells pre-incubated with copper sulfate before of the peptide addition, only for NGF(1–14) (Figure 12E) a visible increase of fluorescence was detected, confirming its active role in the extracellular Cu²⁺ uptake/intracellular Cu⁺ mobilization machinery (Giuffrida et al., 2014).

For both CuSO₄ (Figure 12D) and CuSO₄+Ac-NGF(1–14) (Figure 12F) –treated cells, no significant increase of intracellular copper was observed in comparison to control. Such a result is not surprising owing to the used experimental conditions (i.e., the culture medium supplemented for 1 h with 100 μM CuSO₄). Indeed, it is known that elevating the copper concentrations in the growth medium is able to enhance the cytoplasmic copper concentrations only after several hours of incubation time, depending on the cell line (Trusso Sfrazzetto et al., 2016).

The pre-incubation of the cells in the medium supplemented with BCS led to a relatively small but still noticeable decrease of the fluorescence intensity (Figure 12G) compared to the basal conditions. The addition of NGF(1–14) (Figure 12H) or Ac-NGF(1–14) (Figure 12I) did not result in any visible change of emission.

The quantitative analysis results (Figure 12J) evidenced statistically significant differences only for cells treated with NGF(1–14), which therefore is confirmed to play a role as ionophore, both in the basal and in the copper-supplemented medium. Noteworthy, such an effect of NGF(1–14)-induced increase of intracellular copper content is also significant in the nuclei.

Peptide mimetics is raising a considerable interest owing to the versatility in the synthesis and the tunability of properties of small size peptides compared to the full-length protein (La Mendola et al., 2010; Forte et al., 2014). For example, a peptide can contact a few selected key regions of the target receptor rather than extensive protein surface (Skaper, 2011) allowing an easier design and evaluation of the functional binding activity.

In these regards, the N-terminal domain of NGF has been disclosed as essential for the stability of the NGF/TrkA complex (Berrera et al., 2006), where the formation of H-bonds at the protein−protein interface stabilized the functional NGF/TrkA-D5 binding (Settanni et al., 2003).

We recently reported that NGF(1–14), a linear peptide encompassing the 1–14 sequence of the human NGF protein, is able to activate the TrkA pathway, to induce the phosphorylation of CREB Ser-133, and to affect the PC12 proliferation rate (Travaglia et al., 2015). The NGF(1–14) peptide is the first example of linear peptide with NGF mimetic activity.

Another relevant finding on the NGF(1–14) peptide is that induced only minor morphological changes of PC12 cells, thus suggesting the selective activation of specific signaling pathways of the whole NGF protein (Travaglia et al., 2015).

It is known that the p75/TrKA receptors cooperation is necessary to transduce the entire pattern of NGF signaling. Therefore, to better understand whether the “selective” activation of TrkA receptor might drive/modulate specific NGF signaling, in the present study we performed a new series of experiments to elucidate on the NGF N-terminal peptide activity, in comparison with its acetylated form, and in the presence or not of metal ions. In particular, we carried out combined parallel tempering/docking simulations in order to assess the effects of the N-terminal acetylation of the NGF(1–14) peptide on its molecular recognition of domain 5 of TrkA (TrkA-D5).

These simulations were achieved through the structural data provided by the NGF/TrkA co-crystallization structure (Wiesmann et al., 1999). Noteworthy, we found that while the conformational states of Ac-NGF(1–14) were not affected by the acetylation, an abrupt alteration in the binding modes of Ac-NGF(1–14) to TrkA-D5 was instead observed. The most influenced residue was predicted as the N-terminal serine (S1), which is obviously highly exposed to the effects induced from the insertion of the acetyl group. In particular, S1 completely lost the contacts with histidine H343 or glutamic E295, both featuring the binding between NGF(1–14) and TrkA-D5. The forth histidine residue of Ac-NGF(1–14) H4, lost the contact with H291, which instead was present in NGF(1–14). However, the acetylated peptide, Ac-NGF(1–14), spans a larger number of non-covalent interactions with TrkA-D5 than the analog NGF(1–14), by the formation of hydrogen bonds or salt-bridge interactions.

It is to note that the NGF binding to TrkA, and the following induction of biochemical pathways, is mainly based on weak non-covalent interactions (Wiesmann et al., 1999). Therefore, the enhancing of such interactions can lead to an optimal peptidomimetic activity (Massa et al., 2003). Weak interactions have also been observed in other NGF-mimetic peptides, where docking simulations predicted the binding region of the peptide as a shallow hydrophobic cavity of TrkA (Scarpi et al., 2012). Furthermore, the presence of β-turn and loop structures have been related to an enhanced propensity for NGF mimicking activity, by considering the neurotrophin β-turn domains as the selective portions for the specific receptor binding (Brahimi et al., 2014). Intriguingly, the typical TrkA binding sites involve the NGF loop 2 (residues 40–49: GEVNINSVF), loop 4 (residues 91–97: TMDGKQA), the N terminus (residues 1–8: SSSHPIFH), and the C terminus (residues 111–115: VLSRK). Loops 1 and 4 based peptides (Xie and Longo, 2000; Massa et al., 2003) or a combination of them (Colangelo et al., 2008) have been rigorously studied and identified as NGF peptide mimetics in a number of contributions; the N-terminus domain, however, has been less investigated in the years.

Metal ions, especially copper and zinc, are known to affect the NGF activity (Hwang et al., 2007; Travaglia et al., 2012b). Copper may be required for various aspects of NGF-stimulated neuronal differentiation. For example, copper has been found to promote neurotransmitter release from synaptosomes (Wang, 1999) and to facilitate neurite outgrowth.

Early clinical trials of Cu/Zn ionophores including clioquinol (CQ) and PBT2 have demonstrated promising outcomes in Alzheimer's disease (AD) patients (Lannfelt et al., 2008; Crouch et al., 2009; Faux et al., 2010). It has been generally observed that also the level of zinc transporters (ZnT3, ZIP 1, and 6 as well as ZnT1, 4, and 6) undergo various age-related changes in the brain, and in AD patients (Prakash et al., 2015). Indeed, treatment of ZnT3 KO mice with clioquinol has been found to restore the hippocampal zinc levels and up-regulated key proteins particularly relevant for cognition functions (e.g., presynaptic SNAP-25 and synaptophysin, post-synaptic spinophilin and PSD-95, cell supporting pro-BDNF and DCX, glutamate receptors AMPAR, and NMDAR2a/2b). CQ, being a chaperone able to trespass the BBB, rather than acting as metal chelator to remove metal ions, can redistribute zinc to brain post-synaptic targets leading to events such as TrkB activation and pro-BDNF conversion with beneficial effects on cognitive function in amyloid protein pre-cursor transgenic mice, AD model, and ZnT3 KO mice (White et al., 2006).

These effects may provide neuroprotective signaling changes in neurons, as also recently demonstrated by the observation of morphological changes, i.e., robust induction of neuritogenesis and neurite elongation, in PC12 neuronal-related cells after treatment by bis(thiosemicarbazonato)-copper complexes (Bica et al., 2014). Several new therapeutic strategies are currently aimed at regulating and restoring metal homeostasis (Opazo et al., 2014; Bharti et al., 2016). Interestingly, also four neurotrophic factor, namely BDNF, CNTF, PEDF, GDNF, have been reported to increase the intracellular Zn level in RPE (retinal pigmental epithelium) cells, modulating the expression of zinc transporters which increase zinc uptake (Leung et al., 2008).

In light of the above observations, NGF-mimicking peptides could both to re-establish a normal metal homeostasis and to potentiate the neurotrophic signaling. The experimental data presented here demonstrate significant differences between NGF and the NGF-mimetic peptides on PC12 proliferation/differentiation, in particular in the presence of copper and zinc ions (Figure 3). These findings open to speculations and to further investigations on the early events that occur upon receptor binding (Chao et al., 2006; Matusica and Coulson, 2014).

Here we confirm the NGF ability to induce TrkA and p75NTR receptors internalization, and demonstrate that NGF(1–14) and Ac-NGF(1–14) imitate such an effect (Figure 4). Moreover, different activity was observed in the presence of metals ions: the NGF-induced receptors internalization was inhibited, whereas that of NGF(1–14) ameliorated, and no significant differences in the presence or absence of metals were observed for Ac-NGF(1–14).

The absence of metals effects on the Ac-NGF(1–14)-induced receptors internalization supports the hypothesis of a direct effect of metals on peptides conformation although a metal-related change in the endosome machinery cannot be excluded.

The typical TrkA engagement by NGF (Biarc et al., 2013) was also reproduced using the NGF-mimetic peptides NGF(1–14) and AcNGF(1–14), which were able to phosphorylate Tyr-490, thus triggering an intracellular signaling with strength similar to that elicited after NGF stimulation (Figure 5). This, in turn, was followed by Akt phosphorylation (Ser-473) (Figure 6), and the downstream ERK1/2 phosphorylation that paralleled the pattern observed for Akt phosphorylation (Figure 7).

The Ras-MAPK/Erk signaling cascade is essential for neurotrophin-promoted differentiation of neuronal cells, and transient vs. prolonged MAPK-activation might also be associated with mitogenic- or proliferation-promoting responses (Katz et al., 2007). The ERK1/2 are MAPKs activated by TrkA and are known to phosphorylate and activate the downstream transcription factor CREB (as well as Elk-1, and MEF2), to regulate target gene expression and to contribute to neuronal differentiation and survival (Riccio et al., 1999; Pearson et al., 2001).

Indeed, by the investigation of CREB phosphorylation on Ser-133, we observed that the stimulation of TrkA by NGF and NGF-mimetic peptides was able to activate the canonical end of the cascade (Figure 8). Such a finding demonstrates that our NGF-mimetic peptides can reproduce the signal transduction of the whole protein, being therefore good candidates for further pre-clinical studies.

Many evidences suggest that metal ions could play a role in the NGF and NGF(1–14) signal transduction triggering (Travaglia et al., 2011). Here we demonstrate that copper addition stimulates the NGF(1–14)-mediated TrkA phosphorylation (Figure 5), whereas both the NGF and NGF(1–14) induce Akt, ERK and CREB phosphorylation (Figures 6–8).

In addition, the crucial role of the metal is evidenced also by a general inhibitory effect on both NGF and related peptides in the signaling cascade in the presence of BCS (Figures 5–8). Indeed, at the same experimental condition of metal deprivation, the recovery of NTs-related activities is achieved by copper supplementation to NGF and NGF(1–14) but, as expected, is ineffective for Ac-NGF(1–14) (Figures 5–8). The functional recovery found upon zinc addition, again except than for Ac-NGF(1–14)-mediated signaling, clearly demonstrates the critical role also of this metal (Figures 5–8).

It is to note that NGF(1–14) binds to Zn²⁺ with a higher stability constant than Ac-NGF(1–14), similarly to the trend observed for the binding affinities toward Cu²⁺ of the two peptides (Travaglia et al., 2011). This might explain the higher ionophore activity showed by NGF(1–14) in comparison to its acetylated analogous (Figures 10–12).

The metals homeostatic machinery could be affected or involved in the NGF route to signaling, and the strong disturbance of BCS pre-treatment can affect cell physiology and specific signaling knots. The issue has never been explored and therefore represents a new challenge in the study of neurotrophins activities, as well as memory formation. The restorative effect of NGF treatment is known to be related to its cell internalization and to CREB phosphorylation, followed by CREB-induced gene expression. These peptides are the first linear N-terminal NGF sequences with a direct role in the activation of the Akt-CREB signaling cascade and therefore with high potentials for further pre-clinical approaches. Noteworthy, nanomedicine approaches, e.g., by the immobilization of these peptides at the surface of proper nanoparticle platform, can help in the actual translation of these systems from the bench to the bedside (Di Pietro et al., 2016).

The results in this new study show the ability of NGF(1–14) and its acetylated derivative to up-regulate BDNF mRNA, in PC12 cells and further demonstrate the involvement of copper and zinc ions in these processes. Indeed, the treatment with CuSO₄ or ZnSO₄, induces an increase respectively of 3-, or 5-fold) of the BDNF mRNA levels in comparison to the untreated control, that is comparable to that elicited by NGF treatment (5-fold increase) (Figure 9). Interestingly, while the co-treatment with CuSO₄ or ZnSO₄ significantly down-regulate BDNF mRNA for NGF, a 8-fold increase of the BDNF mRNA levels is found for Ac-NGF(1–14) (Figure 9).

These results appear to be irrespective of the effects on the upstream kinases, but it is possible to argue that different routes, here not explored, could be responsible of such important outcome. In any case, the BDNF mRNA expression parallels the CREB phosphorylation. The relevance of BDNF in the process of synaptic plasticity and memory formation, as well as in neuroprotective responses, makes our peptides suitable for a large plethora of hypothetical approaches. At the same time, the metal ion involvement here exposed puts in evidence the importance of a fundamental bioinorganic perspective.

In correlation to the observed effects on the CREB activation and the increase of mRNA levels of BDNF, the confocal microscopy results of a reversible cellular uptake, especially in the nuclei, for both NGF(1–14) and Ac-NGF(1–14) peptides. This is relevant to confirm the role of the selected sequence (1–14) of NGF to mimic the whole protein. In particular, the LSM studies demonstrate that both NGF(1–14) and its acetylated derivative Ac-NGF(1–14) can be internalized by PC12 cells in a fast process (time scale of seconds-minutes) of cellular influx and efflux. The main difference observed between the two peptides is a more dynamic uptake process for Ac-NGF(1–14), which results in a re-entry effect of the peptide, not observed for NGF(1–14).

A second relevant LSM result is the role of copper ions in the modulation of peptide sub-cellular localization. In particular, Ac-NGF(1–14) that binds copper with a lower stability constant and a different coordination environment than NGF(1–14) (Travaglia et al., 2011), exhibits a nuclear localization with respect to the total cellular uptake (quantified by the ID_nuclei/ID_total ratio) significantly lower in the presence of copper ions. On the contrary, the ID_nuclei/ID_total ratio in the cellular uptake of NGF(1–14) is comparable both without and with the addition of copper ions.

Thirdly, LSM measurements demonstrate the crucial role in the copper trafficking played by the free amino group at the N-terminal residue. Indeed, only NGF(1–14)-treated cells increase their intracellular content of monovalent copper, also in the nuclei. This results hold both in the basal (typically containing Cu²⁺ ions at concentrations up to the micromolar range) and in the copper-supplemented medium. According to the reversible cellular uptake/efflux process described above, an actual role of NGF(1–4) as ionophore is therefore demonstrated.

As known, metallostasis (Milardi and Rizzarelli, 2011) is a complex and dynamic machinery in the spatio-temporal scale (La Mendola et al., 2016; Trusso Sfrazzetto et al., 2016). The used approach to track the intracellular copper needs a careful analysis of results for the correct data interpretation. Indeed, the fluorescent synthetic probe CS1, selective and specific toward Cu⁺ ions, binds to the intracellular monovalent copper in competition with the other copper ligands. The comparison between the two peptides NGF(1–14) and Ac-NGF(1–14), with the unique difference of the presence of the free amino group, strongly supports the more relevant ionophore activity of NGF(1–14), thus providing a smart design route to achieve a modulation of copper homeostasis.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.