Calcium-Sensing Receptor Antagonist NPS 2143 Restores Amyloid Precursor Protein Physiological Non-Amyloidogenic Processing in Aβ-Exposed Adult Human Astrocytes

Physiological non-amyloidogenic processing (NAP) of amyloid precursor holoprotein (hAPP) by α-secretases (e.g., ADAM10) extracellularly sheds neurotrophic/neuroprotective soluble (s)APPα and precludes amyloid-β peptides (Aβs) production via β-secretase amyloidogenic processing (AP). Evidence exists that Aβs interact with calcium-sensing receptors (CaSRs) in human astrocytes and neurons, driving the overrelease of toxic Aβ42/Aβ42-os (oligomers), which is completely blocked by CaSR antagonist (calcilytic) NPS 2143. Here, we investigated the mechanisms underlying NPS 2143 beneficial effects in human astrocytes. Moreover, because Alzheimer’s disease (AD) involves neuroinflammation, we examined whether NPS 2143 remained beneficial when both fibrillary (f)Aβ25–35 and a microglial cytokine mixture (CMT) were present. Thus, hAPP NAP prevailed over AP in untreated astrocytes, which extracellularly shed all synthesized sAPPα while secreting basal Aβ40/42 amounts. Conversely, fAβ25–35 alone dramatically reduced sAPPα extracellular shedding while driving Aβ42/Aβ42-os oversecretion that CMT accelerated but not increased, despite a concurring hAPP overexpression. NPS 2143 promoted hAPP and ADAM10 translocation to the plasma membrane, thereby restoring sAPPα extracellular shedding and fully suppressing any Aβ42/Aβ42-os oversecretion, but left hAPP expression unaffected. Therefore, as anti-AD therapeutics calcilytics support neuronal viability by safeguarding astrocytes neurotrophic/neuroprotective sAPPα shedding, suppressing neurons and astrocytes Aβ42/Aβ42-os build-up/secretion, and remaining effective even under AD-typical neuroinflammatory conditions.


Results
The specific involvement of CaSRs in the following studies was shown by the inhibitory effects of a highly selective antagonist, NPS 2143, on the NO release and cAMP levels elicited by various treatments (see Supplementary Information and Fig. S1). The reversemer peptide Aβ 35-25 was always ineffective (not shown).

Intracellular and secreted
NPS 2143 rescues the fAβ 25-35 ± CMT-induced block in the physiological shedding of sAPPα. Several studies have highlighted the key physiological roles hAPP NAP plays through the shedding of neurotrophic and neuroprotective sAPPα via α-secretase activity 18,25,65 . Therefore, we investigated how sAPPα shedding is altered in human astrocytes exposed to fAβ 25-35 ± CMT ± NPS 2143.

NPS 2143 drives plasma membrane translocation of hAPP in fAβ
Recent evidence indicates that promoting the delivery of hAPP to the plasma membrane or inhibiting the internalization of hAPP favours hAPP NAP 66 . Hence, we investigated the distribution of hAPP in cortical adult human astrocytes following fAβ [25][26][27][28][29][30][31][32][33][34][35] ± CMT ± NPS 2143-treatment. We biotinylated proteins on the astrocyte cell surface and then assessed the amount of biotinylated hAPP at the plasma membrane, with the remaining non-biotinylated hAPP regarded as "intracellular" hAPP (see Methods for details).
Thus, antagonizing Aβ•CaSR signalling intensified hAPP trafficking to the plasma membrane, an effect that was remarkably intensified by CMT.

Discussion
Currently, scant information is available about the mechanisms modulating hAPP proteolysis in cortical untransformed human neural cells. Our previous findings revealed that Aβ•CaSR signalling promotes the AP of hAPP, eliciting substantial increases in endogenous Aβ 42 accumulation and secretion from adult human astrocytes and postnatal neurons [22][23][24] . Our present results show for the first time that a highly selective CaSR antagonist (calcilytic) rescues the physiological NAP of hAPP, maintaining neurotrophic and neuroprotective sAPPα shedding while fully suppressing pathological AP in Aβ-exposed human astrocytes and neurons, even in the presence of microglial CMT (Fig. 6). Although our results were obtained only in astrocytes, we previously demonstrated that the effects of Aβ•CaSR signalling and its antagonism by a calcilytic on Aβ 42 metabolism are very similar in human neurons and astrocytes [22][23][24] . Hence, an extension of the present findings to neurons is feasible until experimentally proven. The concurrent suppression of NO overproduction and adenylate cyclase inhibition (Fig. S1) strengthens the view that NPS 2143 specifically antagonized Aβ•CaSR signalling in astrocytes.
Therefore, the present results prove for the first time that a calcilytic agent can effectively correct the balance of hAPP processing altered by Aβ•CaSR-signalling-mediated mechanisms in cortical human astrocytes (and likely neurons) (Fig. 6). The calcilytic rescues the α-secretase-mediated extracellular shedding of neurotrophic and neuroprotective sAPPα at the expense of the BACE1/β-secretase-mediated neurotoxic Aβ 42 overrelease. The former would safeguard neuronal trophism, viability, and synaptic connections. The latter is inherently dangerous because overproduced and oversecreted soluble Aβ 42 /Aβ 42 -os and their insoluble fibrillar derivatives are endowed with a pernicious self-propagating potential. These peptides can react with CaSRs in adjacent and farther neurons and astrocytes, triggering self-spreading and self-perpetuating vicious waves of Aβ•CaSR signalling, and their consequent accumulation releases further Aβ 42 /Aβ 42 -os surpluses, which likely sustain LOAD progression [22][23][24] . Most remarkably, calcilytics can break such vicious cycles and hence stop Aβ 42 oversecretion and intra-brain diffusion, thus safeguarding neuronal viability and function [22][23][24] . In addition, the calcilytic NPS 2143 elicits a robust downregulation of total CaSR levels in astrocytes, thereby inducing a lasting cell desensitization to exogenous Aβs•CaSR-driven noxious effects 22 .
Regarding the mechanisms of calcilytics, NPS 2143 notably (i) drives a substantial translocation of hAPP, as well as both the precursor (85 kDa; not shown) and active (55 kDa) forms of ADAM10 α-secretase, to the astrocyte plasma membrane; (ii) greatly increases the total ADAM10 α-secretase specific activity, particularly in the presence of CMT; (iii) maintains the intracellular levels of active ADAM10 (55 kDa) at or above basal values; and (iv) restores neurotrophic and neuroprotective sAPPα extracellular shedding close to untreated control levels while hindering the intracellular accumulation of sAPPα. Consequently, by concomitantly also suppressing any surplus Aβ 42 release, NPS 2143 maintains the secreted Aβ 42 /sAPPα and Aβ 42 /Aβ 40 ratios near controls levels.
As with other type-I transmembrane proteins, hAPP is synthesized in the endoplasmic reticulum (ER). Then, hAPP undergoes maturation (glycosylation) while migrating to the Golgi/TGN compartment, where it is mainly found in neurons 66 . Finally, hAPP reaches the plasma membrane via the constitutive secretory pathway, where it is inserted to be cleaved by ADAM10 (mainly) α-secretase, shedding the sAPPα ectodomain, which also occurs in a post-Golgi compartment 67 . In addition, through the recognition of its YENPTY motif and clathrin-coated pits, hAPP can be quickly endocytosed from the plasma membrane and trafficked back to the membrane, delivered through endosomes to the lysosomal system for proteolysis 66 or alternatively cleaved by BACE1/β-secretase, shedding Aβs (the amyloidogenic pathway), particularly if retained in acidic late endosomes, the TGN or ER 67 . Thus, favouring the plasma membrane trafficking or retention of hAPP blocks Aβ production while enhancing controls. In all cases blots have been cropped to size for clarity. Right panel. Densitometric evaluation of holo-APP specific bands for each treatment and time point. Points on the curves are means ± SEM of 3 independent experiments, with 0-h values normalized as 1.0. One-way ANOVA analysis of (i) fAβ [25][26][27][28][29][30][31][32][33][34][35] ± NPS 2143 data set: F = 14.520, P < 0.001; (ii) fAβ 25 sAPPα extracellular shedding via ADAM10 cleavage. hAPP trafficking is regulated by factors that promote Aβ generation, such as the SNX family (SNX17 and SNX33), dynamin I, and the RAB GTPase family (RAB1B, RAB6, RAB8, and RAB11) 68 . In addition, factors that regulating α−, β−, and γ-secretase trafficking are able to alter hAPP processing and, hence, impact the production of sAPPα or Aβs 68 . Further investigations will clarify the roles played by such factors in human astrocytes.
The activation of a number of cell surface receptors, e.g., muscarinic acetylcholine receptors, platelet-derived growth factor (PDGF) receptors, serotonin/5-hydroxytryptamine (5-HT 4 ) receptors, and metabotropic glutamate receptors, reportedly exerts differential effects on hAPP AP or NAP 69 . Our findings add CaSRs to this group of receptors. These receptors activate various signalling pathways that regulate extracellular Aβ secretion and sAPPα shedding via changes in cytosolic [Ca 2+ ] i , cAMP, inositol 1,4,5-triphosphate, small Rac GTPases, and in the activity of a number of protein kinases, including PKA, PKC, mitogen activated protein kinase kinase (MAPKK), extracellular signal-regulated kinase (ERK), phosphatidylinositol-3-kinase (PI3K), and Src tyrosine kinase 69 . Reduced cholesterol levels also heighten ADAM10 activity and hinder hAPP endocytosis, thus enhancing sAPPα shedding from cultured cells 69 . Similar effects can be obtained via ADAM10 overexpression 70 , pharmacological muscarinic activation 32 or phorbol myristate acetate treatment in hAPP-transfected CHO cells 27 . Conversely, the AP of hAPP was favoured at the expense of sAPPα extracellular shedding following overexpression of BACE1/β-secretase 40 or the Swedish mutant form of hAPP (SweAPP), which is linked to a familial EOFAD and is more effectively cleaved by BACE1/β-secretase within the TGN 69 . In addition, knocking down ADAM10 25 and expressing a dominant-negative ADAM10 mutant in mice 70 both increased hAPP AP.
ADAM family members belong to the metzincin superfamily and are typically synthesized as inactive precursors (zymogens) 71 . The proteolytic removal of a conserved cysteine switch in the prodomain is necessary to activate these zymogens 71 . Our findings indicate that cleavage by proprotein convertases (e.g., furin and PC7 in HEK293 cells 72 ) into the 55-kDa ADAM10 active form occurs at the cell surface of human astrocytes rather than in late compartments of the secretory pathway. However, the complex mechanisms modulating α-secretase cleavage activity are not fully elucidated. ADAM10 is not the sole constitutive α-secretase in neurons 25,73 . The present findings indicate that antagonizing Aβ·CaSR signalling with a calcilytic agent, in the absence but more effectively in the presence of CMT, increases the regulated ADAM10 α-secretase specific activity in adult human astrocytes. Treatment with NPS 2143 drives the plasma membrane translocation of both ADAM10 and hAPP in fAβ 25-35 + CMT-exposed astrocytes. This finding reveals that Aβ·CaSR signalling alone restrains the vesicular transport of hAPP and ADAM10 to the plasma membrane, while raising hAPP intracellular levels and AP.
Although it increased ADAM10 α-secretase specific activity, CMT addition had little to no impact on daily and cumulative (i.e., over 72 h) extracellular sAPPα secretion. Only NPS 2143 addition restored sAPPα secretion to the levels of untreated astrocytes, showing the importance of blocking Aβ•CaSR signalling is in restoring hAPP NAP. Regarding the intracellular storage of sAPPα, which did not occur in the untreated astrocytes, CMT addition altered the kinetics but not the total amount stored in fAβ 25-35 -exposed astrocytes. As expected, NPS 2143 reduced most but not all) of the sAPPα storage caused by fAβ [25][26][27][28][29][30][31][32][33][34][35] treatment, even in the presence of CMT. The reasons why these minor sAPPα fractions were retained regardless of NPS 2143 and CMT treatment are not currently understood. Intracellular sAPPα accumulation has also been observed in other cellular models, including cultured human thyroid cells 50 .
Increased cleavage of hAPP by α-secretase was previously suggested as a therapeutic approach to AD 32 . Our present results strengthen the role of calcilytics as prospective drugs for AD therapy (Fig. 6). In this regard, calcilytics benefits largely overcome the mild hyperparathyroidism they induce in humans, given that AD "inexorably kills the patient cognitively several years before his/her actual physical demise" [22][23][24] . Therefore, the negative consequences of calcilytics should prove negligible if clinical trials prove that they can halt AD development.

Methods
Cell cultures. Untransformed human adult astrocytes were isolated from anonymized surgical fragments of normal adult human temporal cortex (brain trauma leftovers) provided by several Neurosurgery Units after obtaining written informed consent from all the patients and/or their next-of-kin. Experimental use of isolated astrocytes was approved by the Ethical Committee of Verona University-Hospital Integrated Company. All human cells experiments were performed in accordance with the relevant guidelines and regulations of Verona University-Hospital Integrated Company. Cultures of astrocytes were set up, as previously described 22  Experimental protocol. Since astrocytes do not actively divide in the adult human brain, we employed them once they had reached mitotic quiescence. At experimental "0 h", culture flasks served partly as untreated controls receiving a change of fresh medium and partly received fresh medium with 20 µM of either fibrillar (f) Aβ [25][26][27][28][29][30][31][32][33][34][35] or reversemer Aβ  added. This dose of the fAβs had been found to be ideal in previous studies 22-24 . Part of the treated cultures received 20 µM of fAβ [25][26][27][28][29][30][31][32][33][34][35] once (at 0 h) plus a cytokine mixture trio (CMT), that is IL-1β (20 ng mL −1 ), TNF-α (20 ng mL −1 ), and IFN-γ (70 ng mL −1 ) (all from PeproTech, London, England). A second and a third CMT bolus was added at 24-h and 48-h. The CaSR allosteric antagonist (calcilytic) NPS 2143 HCl (2-chloro-6-[(2 R)-3-1,1-dimethyl-2-(2-naphtyl) -ethylamino-2-hydroxy-propoxy]-benzonitrile HCl; Tocris Bioscience, UK) 54 was dissolved in DMSO and next diluted in the growth medium at a final concentration of 100 nM. At experimental "0-h", "24-h", and "48-h" part of the astrocyte cultures were exposed for 30 min to NPS 2143 dissolved in fresh medium. Next, the NPS 2143-containing medium was removed and fresh (at 0.5-h) medium or the previously astrocyte-conditioned (at 24.5 and 48.5-h) media were added again to the cultures. Cultures and cell-conditioned media were sampled at 24 hourly intervals. Phosphoramidon (10 μM; Sigma), an inhibitor of thermolysin and other proteases, was added to the media at "0-h" experimental time.
Western immunoblotting (WB). At selected time points, control and treated astrocytes were scraped into cold PBS, sedimented at 200 × g for 10 min, and homogenized in T-PER ™ tissue protein extraction reagent (Thermo Scientific, Rockford, USA) containing complete EDTA-free protease inhibitor cocktail (Roche, Milan). Equal amounts (10-30 µg) of protein from the samples were loaded on NuPAGE Novex 4-12% Bis-Tris polyacrylamide gel (Life Technologies Italia) and next blotted onto nitrocellulose membranes (0.

Biotinylation and isolation of astrocytes' plasmalemmal proteins. The Pierce TM Cell Surface
Protein Isolation Kit (Thermo Scientific) served to biotinylate and isolate cell surface proteins. According to the supplier's procedure the cell culture media were removed and astrocytes were washed twice with ice-cold PBS followed by incubation with 0.25 mg mL −1 Sulfo-NHS-SS-Biotin in ice-cold PBS on a rocking platform for 30 minutes at 4 °C. The biotinylation reaction was quenched by adding 500 μl of the provided Quenching Solution (Pierce). Astrocytes were harvested by gentle scraping and pelleted by centrifugation at 500 × g for 5 minutes at 4 °C. After washing with TBS astrocyte pellets were lysed using the provided Lysis Buffer (Pierce) containing a protease inhibitor cocktail (Roche) for 30 minutes on ice with intermittent vortexing. To get rid of cell remnants, the lysates were centrifuged at 10,000 × g for 2 minutes at 4 °C. To purify biotinylated proteins on Immobilized NeutrAvidin Gel, the clarified supernatant was incubated for 1-h at room temperature (RT) to allow the biotinylated proteins to bind to the NeutrAvidin Gel. The unbound proteins, representing the intracellular fraction, were collected by centrifugation of the column at 1,000 × g for 2 minutes. Any remaining unbound proteins were removed by washing thrice with Wash Buffer (Pierce). Finally, the biotinylated surface proteins were eluted from the biotin-NeutrAvidin Gel by incubation with 400 µL of the SDS-PAGE Sample Buffer containing 50 nM DTT for 1-h at RT in the end-over-end tumbler, and were collected by column centrifugation at 1,000 × g for 2 minutes.
Assays of α-secretases specific activities. The ADAM10 and ADAM17 enzymatic activities were assayed by means of fluorescent methods using EnSens TM ADAM10 and EnSens TM ADAM17 activity detection kits (Enzium, Inc., Philadelphia, USA) in the cell lysates. Despite the highly-overlapping substrate specificities of ADAM10 and ADAM17, EnSens TM substrates are able to differentiate between the two enzymes. Astrocytes' lysates (20 μg) were incubated with the fluorogenic EnSens TM ADAM10 and EnSens TM ADAM17 substrates, respectively for 1-h at RT, protected from light according to the supplier's protocol. The fluorescence was recorded at excitation and emission wavelengths of 625-635 nm and 655-665 nm, respectively. The results were expressed as specific activity (means ± SEMs of ΔF µg −1 protein pertaining to each experimental group). 40 , and sAPPα released into in cell-conditioned growth media. Quantifications of Aβ 42 , and Aβ 40 and sAPPα were carried out by means of specific Aβ 42 , and Aβ 40 Human/Rat High-Sensitive ELISA Kits (both from Wako, Japan) as previously described 22 and by means of specific Human sAPPα High-Sensitive ELISA Kit (from IBL International). Briefly, the astrocytes conditioned media samples were added with a protease inhibitor cocktail (Roche) and centrifuged for 10 minutes at 13,000 rpm to remove any cellular debris. Supernatants were tested in triplicate according to the manufacturer protocol.