CaSR Antagonist (Calcilytic) NPS 2143 Hinders the Release of Neuroinflammatory IL-6, Soluble ICAM-1, RANTES, and MCP-2 from Aβ-Exposed Human Cortical Astrocytes

Available evidence shows that human cortical neurons’ and astrocytes’ calcium-sensing receptors (CaSRs) bind Amyloid-beta (Aβ) oligomers triggering the overproduction/oversecretion of several Alzheimer’s disease (AD) neurotoxins—effects calcilytics suppress. We asked whether Aβ•CaSR signaling might also play a direct pro-neuroinflammatory role in AD. Cortical nontumorigenic adult human astrocytes (NAHAs) in vitro were untreated (controls) or treated with Aβ25–35 ± NPS 2143 (a calcilytic) and any proinflammatory agent in their protein lysates and growth media assayed via antibody arrays, enzyme-linked immunosorbent assays (ELISAs), and immunoblots. Results show Aβ•CaSR signaling upregulated the synthesis and release/shedding of proinflammatory interleukin (IL)-6, intercellular adhesion molecule-1 (ICAM-1) (holoprotein and soluble [s] fragment), Regulated upon Activation, normal T cell Expressed and presumably Secreted (RANTES), and monocyte chemotactic protein (MCP)-2. Adding NPS 2143 (i) totally suppressed IL-6′s oversecretion while remarkably reducing the other agents’ over-release; and (ii) more effectively than Aβ alone increased over controls the four agents’ distinctive intracellular accumulation. Conversely, NPS 2143 did not alter Aβ-induced surges in IL-1β, IL-3, IL-8, and IL-16 secretion, consequently revealing their Aβ•CaSR signaling-independence. Finally, Aβ25–35 ± NPS 2143 treatments left unchanged MCP-1′s and TIMP-2′s basal expression. Thus, NAHAs Aβ•CaSR signaling drove four proinflammatory agents’ over-release that NPS 2143 curtailed. Therefore, calcilytics would also abate NAHAs’ Aβ•CaSR signaling direct impact on AD’s neuroinflammation.


Introduction
Alzheimer's disease (AD) is the world's most prevalent form of dementia [1]. Global population aging has increased its incidence, making AD a serious familial, healthcare, and societal burden. The main AD's neuropathology hallmarks are amyloid-β (Aβ) senile plaques, hyperphosphorylated Tau (hp-Tau) protein neurofibrillary tangles (NFTs), and a chronic diffuse neuroinflammation due to activated innate immune pathways in glial cells [2,3]. The inexorably spreading neuropathology causes a worsening neural circuitry breakdown due to the slowly escalating death of neurons motorcycle accidents who underwent hasty neurosurgery. Briefly, the leftovers were dipped into MCDB 153 medium (Sigma-Aldrich, Milan, Italy), put into a Dewar flask at 4 • C, and carried to the laboratory. There they were soon cut into tiny pieces. The cells were released via a mild treatment with 0.25% (w/v) trypsin (Gibco, Thermo Fisher Scientific, Monza, MB, Italy) in Hank's Basal Salt Solution (BSS; Gibco, Thermo Fisher Scientific) at 18 • C and the residual pieces were triturated with a series of Pasteur pipettes with decreasing (from 5 to 1 mm) bore diameters. The isolated cells were planted in 25 cm 2 culture flasks (BD Biosciences, Le Pont de Claix, France) containing 2 mL of a medium consisting of 89% (v/v) of a 1:1 mixture of Ham's F-12 and MCDB 153 media (Sigma-Aldrich,), 10% (v/v) heat-inactivated (at 56 • C for 30 min) fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific), and 1% (v/v) of a penicillin-streptomycin solution (Gibco, Thermo Fisher). Basic fibroblast growth factor (bFGF or FGF-2; 20 ng mL −1 ; PeproTech EC Ltd., London, UK), insulin-like growth factor-I (IGF-I; 20 ng mL −1 ; PeproTech), platelet-derived growth factor (PDGF; 20 ng mL −1 ; PeproTech), and epidermal growth factor (EGF; 10 nM; Sigma-Aldrich) were added to the medium to enhance the initial proliferation and selection of the astrocytes in the mixed cell population. This complete medium was replaced every 2-3 days. When the primary mixed cultures became 70% confluent (after 1-4 weeks), the cells were detached from the flask surfaces with 0.25% (w/v) trypsin and 0.02% (w/v) EDTA (Gibco, Thermo Fisher) in Hank's BSS, split 1:4 and planted in new flasks. After the third subculture, a pure (100%) population of astrocytes was obtained that no longer needed growth factors. Immunocytochemistry and western immunoblot analysis of the cells of these pure cultures revealed the expression of only astrocyte-specific markers, such as glial fibrillary acid protein (GFAP) and glutamine synthase (GS) (Figure 1). No cells expressed neurons' (enolase), microglia's (CD-68), oligodendrocytes' (galactocerebroside), or endothelial cells' (factor VIII) markers. The astrocytes kept proliferating slowly (doubling time, 2-3 weeks) and expressing their characteristic markers in 90% (v/v) Ham's F12/MCDB 153 medium (Gibco, Thermo Fisher) and 10% (v/v) heat-inactivated fetal bovine serum (Gibco, Thermo Fisher) with no growth factors added. They stopped growing but kept expressing their distinctive markers upon reaching confluence or after they were incubated in high-Ca 2+ (1.8 mM) Dulbecco's Modified Eagle Medium (DMEM, Gibco, Thermo Fisher). Thus, they were phenotypically "locked-in". The proliferatively quiescent cells in confluent astrocyte cultures started cycling again when subcultured. The astrocytes kept expressing the CaSR both when they proliferated and 1.6-fold more intensely (p < 0.002) when they became mitotically quiescent after the exposure to the 1.8 mM Ca 2+ -containing DMEM. On the other hand, astrocytes' CaSR expression levels were independent of the actual levels of extracellular Ca 2+ [27]. Moreover, astrocytes' CaSRs specifically bound exogenous Aβs, and the Aβ•CaSR complexes thus formed were quickly internalized [15,35,55]. At least 15-18 subcultures could be obtained over 2.5 years from a tiny piece (3-4 mm 3 ) of normal cortex. Only astrocytes from the fourth to the eighth subculture were used for the experimental work.

Experimental Protocol
Because astrocytes do not normally proliferate in the adult human brain, as in earlier works [15,20,21,35,36,52], we used confluent, proliferatively quiescent, NAHAs pure cultures in 1.8 mM Ca 2+ DMEM (Gibco, Thermo Fisher). At experimental "0 h", culture flasks (10 6 NAHAs each) served partly as untreated controls receiving a change of fresh medium and partly received fresh medium with 20 µM of fibrillar (f)Aβ25−35. Exposure of NAHAs to fAβ25−35 lasted for the entire duration of experiments. This dose of fAβ25−35 had been found to be ideal in earlier studies [19,20]. The CaSR allosteric antagonist (calcilytic) NPS 2143 HCl (2-chloro-6-[(2 R)-3-1,1-dimethyl-2-(2-naphtyl)-ethylamino-2hydroxy-propoxy]-benzonitrile HCl; Tocris Bioscience, Bristol, UK) was dissolved in DMSO and next diluted in the growth medium at a final concentration of 100 nM. At experimental "0 h", "24 h", "48 h", and "72 h", part of the fAβ25−35 astrocyte cultures was 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 Pictures of pure (100%) in vitro cultures of cortical nontumorigenic adult human astrocytes (NAHAs) that express their cell type-specific markers glial fibrillary acid protein (GFAP) (top panels) and glutamine synthase (GS) (bottom panels) as detected in immunoblots (left panels) and via immunocytochemistry (right panels). NAHAs cultures were set up as detailed in the Materials and Methods). After two weeks of staying in vitro, they were sampled and processed for western immunoblotting and immunocytochemistry as described in the Materials and Methods. Original magnification of the microscope pictures: GFAP, X 100; GS, X 200.

Immunocytochemistry
Immunostaining of astrocytes, which had been seeded into 24-well plates for primary tissue cultures (Becton-Dickinson, Franklin Lakes, NY, USA), was carried out at 4 • C. Astrocytes (2.0 × 10 4 /chamber) were washed twice with PBS (phosphate-buffered saline) containing BSA (1.0% w/v) and NaN 3 (0.1% w/v), and incubated for 60 min at room temperature with mouse monoclonal antibodies (at 1.0 µg mL −1 ) against GFAP and GS (both from Santa Cruz Biotechnology Inc., Heidelberg, Germany). The cells were washed three times with PBS-BSA solution, next incubated for 60 min at room temperature with specific secondary antibodies conjugated to horseradish peroxidase (all from Santa Cruz Biotechnology). Specific immunostainings were developed with 3, 3 -diaminobenzidine (Sigma-Aldrich). After a final wash with PBS-BSA solution, specimens were examined under an inverted Zeiss IM35 microscope (Carl Zeiss Vision Italia, Milan, Italy) and photographed with an Olympus 3300 TM (Olympus Life Sciences, Milan, Italy) digital camera. Appropriate parallel controls were run with no primary or secondary antibody.

Antibody Array
We found and quantified the proinflammatory cytokines and chemokines detectable in NAHAs-conditioned media (Table 1) by using the Cytokine Kit RayBio TM Array 3 (RayBiotech, Inc., Peachtree Corners, GA, USA), according to the manufacturer's protocols. Briefly, NAHAs were treated for 48 h and 96 h with fAβ 25-35 20 µM ± NPS 2143 100 nM. NPS 2143 is a well-established highly selective and specific NAM of the CaSR [22,24]. A total of 2 mL of NAHAs-conditioned media sampled 48 h or 96 h after the onset of the treatments was incubated with the antibody array membranes previously treated for 30 min with Odissey TM blocking buffer (LI-COR Biosciences, Lincoln, NE, USA). Next, after a 2 h incubation at room temperature, the array membranes were thoroughly washed and incubated for 2 h with 1.0 mL of primary biotin-conjugated antibody, diluted 1:250 in Odissey TM blocking buffer. Finally, the membranes were incubated at room temperature for 1 h with 2 mL of DyLight800-conjugated streptavidin (KPL; SeraCare Life Sciences, Milford, MA, USA) diluted 1:7500 in Odissey TM blocking buffer. The positive signals of the detected cytokines and chemokines were acquired with an Odissey TM (LI-COR Biosciences) scanner and later quantified using the Image Studio TM (version 5.2; LI-COR Biosciences) software. The intensities of the positive signals from each array were normalized via comparisons to corresponding positive controls.   20 µM ± NPS2143 and directly stored them at −80 • C to be later assayed for their IL-6 or MCP-2 or RANTES or soluble (s)-ICAM-1 contents. The same cell-conditioned medium samples were tested for assessing IL-6, RANTES, MCP-2 and s-ICAM-1 by means of different ELISA kit. Five independent experiments were repeated using NAHAs from as many individuals. To do this we used the following commercial kits: Human IL-6 PicoKine TM ELISA (Boster Biological Technology Co., Ltd., Pleasanton, CA USA); RayBio TM Human MCP-2 ELISA (RayBiotech); RayBio TM Human RANTES ELISA (RayBiotech); and ICAM-1 (CD54) Human Simple Step ELISA (Abcam, Cambridge, UK). We carried out the tests according to the instructions of the respective manufacturers. The sensitivity of the assays was < 0.3 pg mL −1 for IL-6, 1.5 pg mL −1 for MCP-2, 3 pg mL −1 for RANTES, and 1.6 pg mL −1 for s-ICAM-1.

Western Immunoblotting
First, 1 × 10 6 NAHAs were seeded in 25 cm 2 flasks and cultured with 4 mL of medium, (two flasks for each experimental time point). At selected timepoints, we scraped untreated and treated NAHAs into cold PBS, sedimented them at 200g for 10 min, and homogenized the pellets in T-PER TM tissue protein extraction reagent (Pierce, Rockford, IL) that included a complete EDTA-free protease inhibitor cocktail (Roche, Milan, Italy). We determined the protein contents of the samples by using the Bio-Rad Protein Assay (Bio-Rad). Briefly, equal amounts (20-30 µg) of protein from the lysates were heat-denatured for 10 min at 70 • C in a proper volume of 1× NuPAGE LDS sample buffer supplemented with 1× NuPAGE reducing agent (Invitrogen). Next, the lysates were loaded on a NuPAGE Novex 4-12% Bis-Tris polyacrylamide gel (Invitrogen, Life Technologies, Monza, MB, Italy). After electrophoresis in NuPAGE MES SDS running buffer using the Xcell SureLock TM Mini-Cell (Invitrogen) (50 min run-time at 200 V constant), proteins were blotted onto nitrocellulose membranes using the iBlot TM Dry Blotting System (Invitrogen). Membranes were probed with rabbit antihuman IL-6, or rabbit antihuman ICAM-1, or rabbit antihuman RANTES IgG polyclonal antibodies (all at a final dilution of 0.5 µg mL −1 ; Boster Biological Technology), or mouse antihuman MCP-2 antibody (at a final dilution 1:500; GeneTex Inc., Irvine, CA, USA), or mouse antihuman GFAP antibody and mouse antihuman GS antibody, or goat anti-lamin B1 antibody (at a final dilution of 1.0 µg mL −1 ; all from Santa Cruz Biotechnology Inc.). Subsequent processing steps were as previously detailed [20]. We used lamin B1 as the loading control. We carried out the densitometric analysis of the immunoblots' specific protein using Image Studio TM (version 5.2, LI-COR) software.

Statistical Analysis
Statistical analysis of the data was performed using SigmaStat ® 3.5 Advisory Statistics for Scientists (Systat Software, Richmond, CA) and Analyse-it (Analyse-it Software Ltd., UK). Densitometric data were normalized to lamin B1 (loading control) and next analyzed by one-way ANOVA. When ANOVA's upshots were significant (p < 0.05) we used post hoc Tukey's test for all pairwise comparisons and multiple comparisons vs. untreated control values. Null hypotheses were rejected when p < 0.05.
These findings revealed for the first time that the AD-typical chronic neuroinflammation can be advanced by the NAHAs' Aβ•CaSR signaling-elicited secretion of proinflammatory cytokine IL-6, chemokines RANTES and MCP-2, and of the s-ICAM-1 fragment.

Calcilytic NPS 2143 Effectively Hinders the Secretion of IL-6, MCP-2, RANTES, and s-ICAM-1 from NAHAs
To expand and validate the above antibody array findings, we used more sensitive ELISA tests to assay the amounts of IL-6, MCP-2, RANTES, and s-ICAM-1 fragment released into the growth media from NAHAs treated for 24 h, 48 h, 72 h, and 96 h with fAβ 25-35 ± NPS 2143.

IL-6 Secretion into NAHA-Conditioned Growth Media
The results of time-course ELISA assay revealed that Aβ•CaSR signaling directly and linearly increased the secreted amounts of IL-6 for up to 72 h, when it peaked at a 2.6-fold (p < 0.05) value over untreated controls and remained so high even at 96 h ( Figure 3A). However, adding CaSR NAM NPS 2143 significantly (p < 0.05) suppressed at all time points the fAβ 25-35 -elicited IL-6 surplus secretion over parallel control values ( Figure 3A). Accordingly, the cumulative (as estimated from the areas under the corresponding curves) 0 h to 96 h surplus release of IL-6 over control levels (+80.4%; p < 0.05) elicited by fAβ  itself decreased to insignificance (+14.9%, p > 0.05 vs. controls; −81.5%, p < 0.05 vs. fAβ alone) when NAHAs were cotreated with fAβ 25-53 +NPS 2143 ( Figure 3A; Table 2), thereby confirming the corresponding array data (cf. Figure 2). Therefore, pathological Aβ•CaSR signaling elicited most of the IL-6 oversecretion from NAHAs as calcilytic NPS 2143 effectively suppressed it.     Table 2). Therefore, pathological Aβ•CaSR signaling was specifically responsible for the quota (62.3%) of the IL-6 surplus shed over control levels that calcilytic NPS 2143 suppressed.
In conclusion, the Aβ•CaSR signaling involvement was shown by the ability of CaSR NAM NPS 2143 to significantly reduce the amounts of the just mentioned neuroinflammatory agents that NAHAs secreted/shed.  Figure 4A). Thus, by hindering extracellular release, NPS 2143 treatment favored an early intracellular accumulation of IL-6 which concurred with a decrease in IL-6 secretion (cf. Figure 3A).

Discussion
The present findings offer the first evidence ever that Aβ•CaSR signaling drives an increased synthesis and extracellular secretion/shedding of four proinflammatory agents, i.e., IL-6, ICAM-1/s-ICAM-1, RANTES, and MCP-2, from proliferatively quiescent cortical NAHAs ( Figure 5). These results add a further noxious dimension to the previously reported multiple damaging effects that the pathological signaling from Aβ•CaSR complexes elicits in cultured NAHAs [18][19][20][21]35,36,39,54]. Most important, our results also show that a paradigmatic CaSR NAM, i.e., NPS 2143, can significantly suppress or abate NAHAs' secretion/shedding of the same four proinflammatory mediators. Abundant data in the literature prove that these same agents concur to evoke significant neuroinflammatory effects in vivo (cf. Table 3). Activated astrocytes and microglia, the brain's innate immune system cells, partake in neuroinflammation by producing and releasing copious amounts of cytokines, chemokines, and other agents [50][51][52][53]66,67]. Chemokine-attracted circulating immune cells cross a dysfunctional BBB and release proinflammatory agents intensifying the CNS tissue damage (see for reviews [2,3]). As the abundant literature on the topic makes clear (see for reviews [3,68]), the main mechanisms of inflammatory signaling develop as an integrated pattern forming a biological signaling network. In fact, once cytokines are secreted in response to initial signals, they can bind to their own receptors and trigger both cytosolic and nuclear signal amplification pathways (i.e., NF-κB, JNK [c-Jun N-terminal kinase), p38 MAPK (mitogen activated protein kinase), STAT (signal transducers and activators of transcription), and PI3K (phosphatidylinositol-3-kinase) that crosstalk with one another resulting in complex intracellular signaling networks. Reciprocal astrocytes-microglia interactions [9] lead to self-strengthening positive feedback loops perpetuating and spreading the neuroinflammation. Reportedly, IL-1β, IL-6, TNF-α, and IFN-γ are the main pro-inflammatory cytokines involved in AD brains [51,66,67,69]. Several chemokines, e.g., MCP-1 (or CCL2), RANTES (or CCL5), CCL23, IL-8 (or CXCL8), and IP-10 (or CXCL10) are involved too as they recruit peripheral immune cells into the CNS [51,67,69]. Neuroinflammation exacerbates the course of both acute (e.g., stroke) and chronic (e.g., AD, Parkinson, etc.) diseases by self-propagation and by causing neuronal excitotoxicity and loss of synapses [9,[50][51][52][53]66]. Regarding AD, some reports suggest that microglial inflammatory mediators like IFN-γ and TNF-α cause Aβs overproduction and deposition by hindering Aβs clearance in mutant APP transgenic mice and in cocultures of astrocytes and microglia from the same mutant and wild-type mice [70,71]. At variance with these results in rodents, microglial cytokines only transiently accelerated endogenous Aβs release from Aβ-exposed NAHAs [21]. The use of antibody arrays has allowed us to find the four proinflammatory agents that are driven by Aβ•CaSR signaling in NAHAs and are hindered by the CaSR NAM presently employed. Cells 2020, 9,1386 16 of 26 recruit peripheral immune cells into the CNS [51,67,69]. Neuroinflammation exacerbates the course of both acute (e.g., stroke) and chronic (e.g., AD, Parkinson, etc.) diseases by self-propagation and by causing neuronal excitotoxicity and loss of synapses [9,[50][51][52][53]66]. Regarding AD, some reports suggest that microglial inflammatory mediators like IFN-γ and TNF-α cause Aβs overproduction and deposition by hindering Aβs clearance in mutant APP transgenic mice and in cocultures of astrocytes and microglia from the same mutant and wild-type mice [70,71]. At variance with these results in rodents, microglial cytokines only transiently accelerated endogenous Aβs release from Aβexposed NAHAs [21]. The use of antibody arrays has allowed us to find the four proinflammatory agents that are driven by Aβ•CaSR signaling in NAHAs and are hindered by the CaSR NAM presently employed.

IL-6
The present findings show that Aβ•CaSR signaling strongly increases both the synthesis and

IL-6
The present findings show that Aβ•CaSR signaling strongly increases both the synthesis and release of IL-6 from NAHAs. Conversely, calcilytic NPS 2143 suppresses most of this IL-6 surplus release while increasing its amounts in NAHAs lysates. The latter finding might be ascribed to the blocking effect of the IL-6 secretory pathway at the Golgi apparatus level [72] that CaSR NAM NPS 2143 exerts, which affects the secretion of other compounds including endogenous Aβ 42 [20] and likely also RANTES and MCP-2 (see below). Further studies will clarify whether the slow decline in intracellular levels of IL-6 occurring after early peaking is due to decreased synthesis or increased lysis or both. It is well established that IL-6 and its downstream JAK/STAT3 signaling pathway exert pleiotropic effects closely related to our previous and present findings, that is the upregulation of the transcription of several genes including ICAM-1, RANTES, MCP-2, VEGF-A, and CASR [73][74][75]. On this basis, we posit that IL-6 overexpression driven by Aβ•CaSR signaling partakes in the control of CASR, ICAM-1/s-ICAM-1, RANTES, and MCP-2 genes upregulation in NAHAs. Moreover, overproduced IL-6 accumulates around and inside senile plaques in the cerebral cortex and hippocampi of AD patients [76,77]. IL-6 also increases Tau protein hyperphosphorylation in neurons of AD brains through the cdk5/p35 and the MAPK-p38 signaling pathways [78]. Conversely, the results of studies about IL-6 in AD-model animals have so far been contradictory. In the astrocytes residing in the hippocampus and cerebellum of transgenic mice, increased IL-6 levels upregulated the expression of GFAP, glutamine synthase, STAT-3, phosphorylated STAT-3, and phosphorylated pp42/44 MAPK, but downregulated that of Synuclein 1, GAD65/67, GluA1, and GluN1 [79]. Recent reports revealed that the JAK/STAT3 pathway acts as pivotal driver of astrocyte reactivity [80] and that inhibiting the Stat3-mediated astrogliosis ameliorates the neuropathology in mouse models of AD [81]. Brugg et al. [82] reported that peripheral stimulation with LPS induced transient elevations in both IL-6 and IL-1β mRNAs followed by changes in the expression pattern of APP isoforms (i.e., decreases in APP695 and increases in APP KPI levels) in the cerebellum but not in the cerebral cortex of mouse brain. This concurrent upregulation of both IL-6 and APP during acute neurological stress or chronic neurodegeneration suggested an interlinked expression of these two proteins. However, as a cautionary note, we must mention here that NAHAs are insensitive to LPS exposure (our unpublished results) and, as the present findings show, IL-6 upregulation is completely or nearly completely controlled by Aβ•CaSR signaling in NAHAs.
Based on the just mentioned data in the literature, our findings suggest that by increasing both IL-6 and RANTES expression Aβ•CaSR signaling boosted a sizeable chunk of s-ICAM-1 fragment's shedding from NAHAs. The calcilytic-non-suppressible portion of s-ICAM-1 s increased shedding might have been brought about by the concurrently Aβ-elicited IL-1β upregulation [86,87,95], which as our results show, is independent of Aβ•CaSR signaling. Further studies will assess this hypothesis. Conversely, the observed ICAM-1 holoprotein increases in NAHA lysates are likely to be due to different operative mechanisms. On its own part, Aβ•CaSR signaling might increase ICAM-1 holoprotein synthesis and its insertion and cleavage at the NAHAs plasma membrane. On the other hand, the calcilytic might hinder plasma the ICAM-1 holoprotein membrane insertion and cleavage. Although the details about the molecular mechanisms involved are not understood, our results clearly show that the addition of NPS 2143 to Aβ-exposed NAHAs did significantly cut down the potential proinflammatory actions of both ICAM-1 holoprotein and s-ICAM-1 fragment. The following evidences will allow us to better appreciate the prospective relevance of these findings. Remarkable increases in extravascular s-ICAM-1 around GFAP + astrocytes connote the orbitofrontal cortex of normally aging people and might mark an increasing with age risk of neuroinflammatory diseases [88]. In postmortem AD brains, s-ICAM-1 aggregates localize to peri-plaque astrocytes, early and late stage amyloid senile plaques, and cerebral vessels [96][97][98]. Serum levels of s-ICAM-1 are low in healthy subjects [89]. In AD cases, s-ICAM-1 levels raise in both blood and CSF, mirror the upsurges in transmembrane ICAM-1 holoprotein, positively correlate with illness severity [90,91,95], and partake in BBB's dysfunction thus advancing immune cells infiltration into the CNS and hence neuroinflammation.

RANTES
Astrocytes are the main source of RANTES, a powerful chemokine that attracts and activates eosinophil and basophil leukocytes. The present results show that, after a 24 h delay, Aβ-exposed NAHAs released significantly higher RANTES amounts while reducing its levels in the cells' lysates. The 24 h time lag preceding the onset of RANTES over-release suggests that its driving mechanism(s) is(are) complex requiring the synthesis of some intermediate agent(s). Microglial cytokines like TNF-α and IFN-γ upregulate RANTES expression in human and simian astrocytes [66]. However, these cytokines were undetectable in our experimental system. Of greater interest is the notion that IL-1β mediates RANTES expression in human fetal astrocytes via the activation of IFN regulatory factor 3 (IRF3]. In its turn, IFR3 induces a group of IFN-stimulated antiviral response genes (ISG) including, besides RANTES, IFN-β, IRF7, and CXCL10/IFN-γ-inducible protein-10 [62,99]. Whether this IL-1β-triggered mechanism also works in the Aβ-treated NAHAs that, as our results show, overexpress IL-1β ( Figure S1), seems likely yet remains to be proven. Lin et al. [100] also reported that the activation of PI3K and MAPK signaling pathways upregulated RANTES expression in curcumin-treated primary rat astrocytes. Because Aβ•CaSR signaling activates MEK/ERK signaling in NAHAs [101], this mechanism could also help up-regulate RANTES. This view is supported by our observation that CaSR NAM NPS 2143 significantly curbed a substantial fraction (i.e., 42%; Table 2) of RANTES release from NAHAs likely via a block of its secretory pathway through the Golgi apparatus-a mechanism also shared by IL-6, MCP-2, and endogenous Aβ 42 [21,102,103]. On the other hand, the fall of immunodetectable RANTES in lysates from Aβ-exposed NAHAs suggests that over synthesized RANTES is secreted into the medium with no delay just as happens for VEGF-A [35]. Moreover, our observation that after peaking at 24-h the RANTES overaccumulation progressively vanished in Aβ+NPS 2143-treated NAHAs might result from (i) an initial block of RANTES secretory pathway through the Golgi apparatus; and (ii) a later decline of RANTES de novo synthesis coupled to a rescue of proteasomal activity [21]. Clearly, a detailed definition of the mechanisms involved requires further investigations. Our results suggest as likely that RANTES over-release from Aβ-exposed NAHAs partakes in the multiple proinflammatory effects evoked by Aβ•CaSR signaling and by an Aβ exposure in general, which is in keeping with the views of authors positing that RANTES is a relevant player in the inflammatory cascade that advances AD neurodegeneration [62,104,105]. As it does in murine astrocytes, RANTES by itself also strongly stimulates the production and release of proinflammatory mediators, including s-ICAM-1 [94,106]. It is worth mentioning here the interactions of secreted RANTES with its three receptors, i.e., CCR1, CCR3, and CCR5, which drive autocrine mechanisms affecting astrocytes secretory functions [106]. Of note, the activation of RANTES receptors entails their coupling with inhibitory G i/o protein, which cuts down adenylyl cyclase activity and thus lowers the synthesis of neurotrophic cyclic 3 ,5 -adenosine monophosphate (cAMP) in mouse astrocytes [93]. Previously, we showed that CaSR NAM NPS 2143 rescued the Aβ-curtailed cAMP production and release from NAHAs [21]. Whether the signaling pathways of CASR and RANTES receptors crosstalk between each other is to be determined in NAHAs. Finally, on a discordant note, Grammas et al. [105,107] reported that RANTES upregulation assists neurons' survival in vitro by protecting them against the noxious effects of AD neurotoxins, thrombin, and sodium nitroprusside. Therefore, further studies will clarify whether RANTES plays a double-face role according to the actual stage of AD neuroinflammation.

MCP-2
Human MCP-2 is a small chemokine encoded by the CCL8 gene. MCP-2 pertains to a subfamily of the C-C or β-chemokines also including MCP-1 and MCP-3 sharing a 60-62% sequence identity [61]. The present findings prove for the first time that Aβ-exposed NAHAs quite intensely oversecrete MCP-2 with respect to basal values. They also show that adding calcilytic NPS 2143 cuts down a significant fraction (62.3%; Table 2) of the Aβ•CaSR signaling-driven MCP-2 oversecretion. MCP-2 only moderately accumulated in lysates from Aβ-treated NAHAs apparently because its oversecretion balanced most of its overproduction. However, the linearly progressive and much more pronounced MCP-2 accumulation in lysates from Aβ+NPS 2143-treated NAHAs suggests that NPS 2143 blocked MCP-2 secretion through the Golgi pathway, as did with IL-6, RANTES, and endogenous Aβ 42 [21,102,103], without interfering with its accelerated synthesis. Further work will assess this postulation. As a chemokine, MCP-2 is both less potent and less effective than MCP-1, MCP-3, and RANTES [61,108]. But, as our findings show, basal MCP-1 expression did not change in Aβ-treated NAHAs ( Figure S1). Reportedly, MCP-2 activates the chemotaxis of human lymphocytes T, NK cells, and monocytes, which take part in inflammatory responses, and of eosinophils, basophils, and mast cells, which partake in allergic reactions. Concerning AD, most of the existing literature focuses on the proinflammatory role of MCP-1. Conversely, the available data about MCP-2 role(s) in AD are scanty. Just like MCP-1, CCL7, CCL12, and CCL13, MCP-2 enhances the chemotaxis of proinflammatory cells towards inflamed areas of the CNS [109]. As happens with MCP-1, CSF MCP-2 levels might increase in AD patients, suggesting its association with neurodegeneration [108]. In addition, CSF levels of MCP-2 might be a good risk predictive marker for early stage AD and other psychoses [110]. Aβ•CaSR signaling and other hitherto unidentified factors may drive MCP-2 oversecretion from NAHAs. Next, MCP-2 would recruit monocytes and/or other leukocyte populations into the brain, thereby enhancing neuroinflammation and neuronal injury, thus contributing to AD's progression via mechanisms not detectable by the presently used experimental system. Based on our findings, it seems workable that by antagonizing Aβ•CaSR signaling, calcilytics could significantly diminish MCP-2 role in AD's neuroinflammation. Clearly, further studies should specifically address and clarify MCP-2 s role(s) in human AD. Table 3. IL-6, s-ICAM-1, RANTES, and MCP-2 drive complex neuroinflammatory responses.

Agent
Proinflammatory Roles

Other Proinflammatory Agents Not Affected by Antagonizing Aβ•CaSR Signaling
See Supplementary Data.

Conclusions
The present results are the first evidence that Aβ•CaSR signaling is directly involved in AD's neuroinflammation via the over-release/shedding of four proinflammatory agents from NAHAs. This is a further addition to the previously reported panoply of detrimental actions driven by the Aβ•CaSR signaling in NAHAs and human neurons [14,[17][18][19][20]33,34,52]. Notably, the Aβ•CaSR signaling unique ability to simultaneously set off and release an amazing multiplicity of noxious effectors from human cortical neurons and astrocytes, testifies for its relevance to AD. Our findings further stress the view that pathological Aβ•CaSR signaling is a potential therapeutic target in AD [18,20,21,[36][37][38]. On the other hand, they also show that extracellular Aβs surpluses induce NAHAs to over-release several proinflammatory agents through Aβ•CaSR-independent mechanisms. But these effects are placed downstream from Aβ•CaSR signaling, the upshots of which include the release of Aβs surpluses from human neurons and astrocytes [21]. As a final notation, we are aware that the other outstanding neuroinflammation player, i.e., human microglia, is missing from the experimental system presently used. However, we are confident that future work will overcome this limitation and throw further light onto the intricate mechanisms that hold sway in human AD-related neuroinflammation Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4409/9/6/1386/s1, Figure S1: Other Cytokines and Chemokines Arrays, Figure S2: PDGF-BB Array.