Background
Inflammatory conditions are marked by the production of mediators such as cytokines, chemokines, reactive oxygen species, and acute phase proteins that are key elements of the accompanying physiological and metabolic changes. C-reactive protein, complement proteins, and serum amyloid A protein (SAA) are some of the principal acute phase proteins, mainly generated in the liver and released into the systemic circulation in response to inflammation [
1,
2]. SAA is the generic name of a family of proteins that share high levels of sequence homology but are encoded by different genes [
3]. Humans possess four SAA genes (SAA1, SAA2, SAA3, and SAA4) mapped in a 150-kb region of chromosome 11p15.115. Mice also have four SAA genes whose protein products are highly homologous to their human counterparts [
3]. Inducible expression is characteristic of all acute-phase SAAs including SAA1 and SAA2. Extra-hepatic expression of SAA has been reported as well [
4].
Central nervous system (CNS) disorders are characterized by central activation of innate immunity and activation of a potent peripheral acute phase response that influences central inflammation and contributes to poor outcome [
5]. Syrian hamsters injected systemically with lipopolysaccharide (LPS) had elevated levels of mRNA for Apo-SAA in all tissues examined, including brain [
6]. While not detectable in normal brain, SAA protein has been found in Alzheimer disease (AD) brain, along with SAA gene expression in multiple sclerosis (MS) brain tissue [
7]. Elevated SAA concentration was described in cerebrospinal fluid of AD subjects [
8], as well as SAA immunoreactivity that co-localized with amyloid β-peptide deposits in AD brain [
9]. Induction of a systemic acute phase response in SAA transgenic mice enhanced amyloid β-peptide deposition [
10]. Further, Chung et al. [
11] reported a much stronger immunostaining of SAA in brain of patients with neurologically confirmed AD and MS in comparison to unaffected regions and non-AD/MS brain. Barbierato et al. [
12] recently demonstrated that cortical glia responds to pro-inflammatory agents (LPS, tumor necrosis factor alpha (TNF-α), Apo-SAA) by upregulating their expression of
Saa1.
Interleukin-1β (IL-1β), a master regulator of neuroinflammation [
13] mainly produced by activated inflammatory cells of the myeloid lineage [
14], contributes importantly to cellular activation and cytokine production. IL-1β plays a key role in the pathogenesis of acute and chronic diseases of both the peripheral nervous system and CNS [
15‐
17]. LPS, a potent stimulus for IL-1β synthesis by microglia is rather inefficient, given that most of the secreted cytokine remains in the immature (inactive) pro-form [
18]. One of the molecules mainly involved in IL-1β maturation is the purinoceptor P2X
7 (P2X
7R), an ATP-gated ion channel that chiefly acts through the recruitment of the NLRP3 inflammasome-caspase-1 complex [
14,
19]. This activation process involves first recognition by toll-like receptors (TLRs, a sub-family of pattern recognition receptors) of exogenous (e.g., bacterial- and virus-derived pathogens) or endogenous (e.g., components of cell damage) stimuli to induce transcription and translation of IL-1β (‘priming’). This is followed by a secondary signal such as ATP to trigger formation of the inflammasome complex that leads to caspase 1 activation and cleavage/release of IL-1β [
20‐
22]. P2X
7R-triggered IL-1β maturation and export may thus represent a major contributor to this cytokine pool [
20,
23].
SAA appears to be an endogenous ligand for both TLR4 [
24‐
27] and TLR2 [
28‐
31], despite having little structural resemblance to the bacteria-derived ligands of either receptor. Although SAA can upregulate the NLRP3 inflammasome in peripheral immune cells [
25] and provoke mediator production in a variety of non-neural cells, nothing is known about its ability to stimulate IL-1β release from CNS glia in the presence of ATP, a multi-target danger signal in the brain [
32] in a P2X
7R-dependent manner. This is especially important, given the growing body of data indicating that genetic or pharmacological manipulation of P2X
7Rs alters responsiveness in animal models of CNS neurological disorders [
33]. Recent studies suggest also that P2X
7Rs regulate the pathophysiology of psychiatric disorders, including mood disorders [
33]. The present study was undertaken to examine the ability of ATP to promote the intracellular production, and release, of IL-1β from cortical microglia stimulated with Apo-SAA, and the involvement of P2X
7R, TLR4, and TLR2.
Methods
Tissue culture media, antibiotics, fetal calf serum (FCS), and NP40 cell lysis buffer (10×) were purchased from Life Technologies (San Giuliano Milanese, Italy); lipopolysaccaride (LPS) (Ultra-Pure LPS-EB from E. coli 0111:B4 strain; only activates TLR4), Pam3CSK4, and ethyl-(6R)-6-(N-(2-chloro-4-fluorophenyl)sulfamoyl)cyclohex-1-ene-1-carboxylate (CLI-095 or TAK 242) were from InvivoGen (Cayla-Invivogen Europe, Toulouse, France); A740003 from Tocris Bioscience, Pittsburgh, PA, USA); poly-L-lysine hydrobromide (mol wt 70,000–150,000), papain, DNase I (bovine pancreas), trypsin inhibitor, L-leucyl-L-leucine methyl ester (L-LME), protease inhibitor cocktail, Pefabloc® SC (100 mM), CU-CPT22, and all other biochemicals were purchased from Sigma-Aldrich (Milan, Italy) unless noted otherwise; recombinant human Apo-SAA (consensus SAA molecule corresponding to human Apo-SAA1α, except for the presence of an N-terminal methionine, the substitution of asparagine for aspartic acid at position 60, and arginine for histidine at position 71) from Peprotech (endotoxin level < 0.1 ng/μg protein; London, UK); QIAzol from Qiagen (Milan, Italy. Falcon tissue culture plastic-ware were purchased from BD Biosciences. Sterilin Petri plastic dishes (10 cm Ø) were obtained from Sarstedt (Verona, Italy).
Cell culture
Microglia were isolated from mixed glial cell cultures as previously described [
34]. All experiments were conducted in compliance with Italian Ministry of Health (art. 31, D.L. 26/2014) guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee of the University of Padua (958/2016-PR). In brief, cells dissociated from postnatal day 1 rat pups (Charles River, Calco, Italy; strain: CD) cerebral cortices were plated in 75 cm
2 poly-L-lysine-coated tissue culture flasks (1.5 brains per flask) and grown in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) with 2 mM glutamine, 50 units/ml penicillin/50 μg/ml streptomycin, 50 μg/ml gentamycin, and 10% FCS (‘growth medium’). Culture medium was changed after 24 h. The cultures reached confluence by 7 days at which time microglia were recovered by shaking the flasks on an orbital shaker at 200 rpm for 1 h (37 °C). The remaining cell monolayers were highly enriched in astrocytes (< 5% microglia, as determined by flow cytometry using cell type-specific antibodies) [
35]. The culture supernatant containing microglia was transferred to Sterilin plastic Petri dishes and incubated for 45 min at 37 °C (5% CO
2, 95% air) to allow adhesion of microglia. The adherent microglial cells (> 99% pure, as determined by flow cytometry using cell type-specific antibodies) [
35] were detached by mechanically scraping into growth medium and re-plated in this same medium, on poly-L-lysine-coated 24-well or 96-well culture plates (250,000 and 50,000 cells per well for mRNA and cytokine analysis, respectively). For some experiments, the astrocyte monolayers were depleted of residual microglia using a 60-min exposure (50 mM) to the lysosomotropic agent L-LME [
36], as described previously [
37,
38]. Astrocyte plating densities were the same as used for microglia.
Quantitative real-time polymerase chain reaction (q-PCR)
Total RNA was extracted from cells by QIAzol, according to the manufacturer’s instructions. RNA integrity and quantity were determined by RNA 6000 Nano assay in an Agilent BioAnalyser (A
260/280 ratio > 1.8). Reverse transcription was performed with Superscript III reverse transcriptase (Invitrogen). The q-PCR reaction was performed as described previously [
37]. Primer sequences are listed in Table
1. Amounts of each gene product were calculated using linear regression analysis from standard curves, demonstrating amplification efficiencies ranging from 90 to 100%. Dissociation curves were generated for each primer pair, showing single product amplification. Data are normalized to β-actin mRNA level.
Table 1
PCR primers used in this study
β-ACT | F | 5′-CCCCATTGAACACGGCATTGTCA-3′ |
R | 5′-ACCCTCATAGATGGGCACAGTGT-3′ |
IL-1β | F | 5′-TGTGGCAGCTACCTATGTCT-3′ |
R | 5′-GGGAACATCACACACTAGCA-3′ |
NLRP3 | F | 5′-TGATGCATGCACGTCTAATCTC-3′ |
R | 5′-CAAATCGAGATGCGGGAGAG-3′ |
SAA1 | F | 5′-ACACGGAGCAGAGGACTCAAG-3′ |
R | 5′-GGTCGAAAGTGGTTGGGGTC-3′ |
TNF-α | F | 5′-CATCTTCTCAAAACTCGAGTGACAA-3′ |
R | 5′-TGGGAGTAGATAAGGTACAGCCC-3′ |
TLR2 | F | 5′-TCCATGTCCTGGTTGACTGG-3′ |
R | 5′-AGGAGAAGGGCACAGCAGAC-3′ |
TLR4 | F | 5′-GATTGCTCAGACATGGCAGTTTC-3′ |
R | 5′-CACTCGAGGTAGGTGTTTCTGCTAA-3′ |
IL-1β production and release
Purified microglia and enriched astrocytes were plated in poly-L-lysine coated 96-well plates (50,000 cells per well) in growth medium and allowed to adhere overnight. These plating densities do not affect glial cell vitality/function [
39,
40]. Cells were primed by pre-treating with 0.1 μg/ml LPS (optimal concentration chosen from preliminary experiments) [
18,
37,
41] or different concentrations of recombinant human Apo-SAA (as indicated in the experiment) for 2 h in serum-free culture medium prior to stimulation with 5 mM ATP [
41] for 1 h. None of the treatments, at the concentrations tested, affected cell viability (data not shown; see also [
12,
42‐
44]). Cell supernatants were collected and stored at − 20 °C until the day of assay (avoiding repeated freeze-thaw cycles). Cell lysates were prepared by adding to each 96-well culture 100 μl lysis solution containing: 89 μl NP40 lysis buffer, 10 μl of 10× protease inhibitor cocktail, and 1 μl of 100 mM Pefabloc SC. IL-1β content of culture medium and cell lysates was analyzed using commercially available enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturer’s instructions (Antigenix America, Huntington Station, NY, USA). Standards with known amounts of IL-1β and TNF-α were used to convert values into absolute concentrations of the cytokine in pg/ml.
Statistics
Data are given as mean ± sem, unless stated otherwise. Statistical analyses to determine group differences were performed either by two-sample equal variance Student’s t test, or by one-way analysis of variance followed by Dunnett’s or Bonferroni’s post hoc tests for comparisons involving more than two data groups. Significance was taken at p < 0.05.
Discussion
Growing evidence indicates that CNS disorders are characterized by central activation of innate immunity, as well as activation of a potent peripheral acute phase response that influences central inflammation and leads to poor disease outcome [
5]. The acute phase response plays a critical role in the innate immune response to tissue injury [
63]. Among acute phase proteins such as C-reactive protein, complement proteins, and SAA, the last one can be considered a “danger signal” that influences the inflammation process [
64]. Its low basal level and high inducibility are in keeping with danger signal molecules [
65], being produced in response to potentially harmful environmental cues, including trauma, infection, surgery, and severe stress. A number of studies imply a role for SAA in inflammation-associated neuropathologies [
7‐
11], although the underlying molecular processes remain to be fully explored. Here, we show that in neonatal cortical microglia, Apo-SAA time-dependently upregulates NLRP3 inflammasome and IL-1β mRNA expression and intracellular production of IL-1β and stimulates release of IL-1β in the presence of ATP, a multi-target danger signal in the brain [
32] in a P2X
7R-dependent manner. The rise in extracellular release of IL-1β in the presence of ATP was accompanied by a fall in the intracellular content, consistent with NLRP3/caspase 1-complex 1 activation and cleavage of the pro-form of IL-1β to the mature, active secreted species [
20‐
22]. The action of Apo-SAA was not limited to cortical microglia, as similar ATP-dependent release of IL-1β was also seen for cerebellar microglia. IL-1β is viewed as a master regulator of neuroinflammation [
13] that contributes importantly to cellular activation and cytokine production. This cytokine plays a key role in the pathogenesis of acute and chronic diseases of both the peripheral nervous system and CNS [
15‐
17]. It merits mention that the effects of Apo-SAA on gene expression and cytokine production were of a far greater magnitude than those obtained using the optimal concentration of LPS as benchmark, thus highlighting the pro-inflammatory potency of this acute phase protein. Whether adult microglia would respond differently was not tested.
A comparison of the SAA concentrations used in the present investigation with levels of SAA previously detected in human cerebrospinal fluid and plasma suggests the potential for physiological relevance to the in vivo setting. In one report, SAA levels in cerebrospinal fluid of AD subjects were found to be much higher than in normal controls [
8], and generally within the range of the highest concentration used here. Serum concentrations of SAA in relapsing-remitting MS patients have been reported elevated with a mean level of 12.1 ± 8.7 μg/ml [
66], and significantly increased (mean value 10 μg/ml,
p = 0.030 vs. control) in neuromyelitis optica patients [
67]. These values contrast with active concentrations of 0.15–1.5 μg/ml in the present in vitro study.
An expanding body of data demonstrates that pharmacological or genetic manipulation of P2X
7Rs alters their responsiveness in animal models of CNS neurological disorders [
33,
68]. The P2X
7R has been suggested to also regulate the pathophysiology of psychiatric disorders, including mood disorders [
33]. P2X
7R-triggered IL-1β maturation and export is thus likely to represent an important contributor to this cytokine pool [
20,
23]. SAA is not detectable in normal brain but has been reported in AD brain, together with its gene in MS brain [
7]. Miida et al. [
8] described a raised SAA concentration in cerebrospinal fluid of AD. SAA immunoreactivity was reported to co-localize with amyloid β-peptide deposits in AD brain [
9]. P2X
7R-positive microglia surrounded amyloid plaques in a mouse transgenic AD model [
69], and microglia around amyloid plaques in AD brain are immunopositive for IL-1β [
70]. Collectively, these findings propose a link between P2X
7R, Apo-SAA, and IL-1β in AD pathophysiology. In addition, the ability of Apo-SAA to regulate its own gene expression suggests the potential for autocrine/paracrine effects of SAA. Since microglia in the AD brain adopt distinct functional and molecular phenotypes, it is conceivable that the response of “AD microglia” to SAA would differ from that of wild-type microglia.
A number of reports indicate the capability of SAA to act as an agonist for both TLR4 [
24‐
27] and TLR2 [
28‐
31,
71]. Ligand engagement of TLR4 by LPS and TLR2 by Pam
3CSK
4 leads to the upregulation of
Tlr2 and downregulation of
Tlr4 in cortical microglia [
35]. Consistent with its putative action as a ligand for both TLR2 and TLR4, Apo-SAA produced a time-dependent robust and significant increase in
Tlr2 expression in cortical microglia, with a concomitant reduction in the relative level of
Tlr4. Conceivably, this action of SAA could result in a ‘feed-forward’ mechanism, whereby SAA increases expression of its receptor and amplification of a priming response. Attempts at using pharmacological tools to dissect participation of TLR4 and TLR2 in the actions of Apo-SAA were equivocal. The selective TLR4 antagonist CLI-095 completed blocked the ability of LPS to synthesize/release IL-1β and partially, but significantly, that of Apo-SAA. While the TLR2 antagonist CU-CPT22 failed to alter the stimulatory effect of Apo-SAA or LPS, it also proved ineffective on microglia treated with the TLR1/2 agonist Pam
3CSK
4. Our inability to confirm the earlier report for CU-CPT22 action against Pam
3CSK
4 [
61] could be due to differences in cell type used (primary microglia vs RAW264.7 macrophages) or treatment times (not specified in [
61]), even though we used two incubation times and the same concentrations of ligand and antagonist as in [
61]. Another consideration is that CU-CPT22 was designed to compete with Pam
3CSK
4 binding to TLR1/2, thus disrupting formation of the TLR1/TLR2 heterodimer [
61]. Although outside the scope of the present study, the use of microglia from TLR2
−/−
animals could provide a tool to address this question.
A failure of remyelination is responsible, in large part, for the long-term neurologic consequences of MS. An intriguing study by Sloane et al. [
72] described upregulated TLR2 expression by oligodendrocytes in MS lesions, with pathogen-derived TLR2 agonists, but not agonists for other TLRs, inhibiting oligodendrocyte precursor cell (OPC) maturation in vitro. Ablated expression of TLR2 also enhanced remyelination in a lysolecithin animal model of MS [
72]. Intense immunohistochemical staining of SAA has been detected in the brains of patients with neurologically confirmed MS in comparison to an unaffected region and non-MS brains, with the major site of staining being the myelin sheaths of axons in affected cortex [
11]. Pro-inflammatory cytokines such as IL-1β and TNF-α [
73] may play important roles in expression of SAA1 and SAA2. The pathophysiology of a variety of neurological disorders, including MS, is associated with TNF-α [
74,
75], a master pro-inflammatory product of activated microglia and peripheral macrophages implicated in the pathogenesis of CNS demyelination [
76,
77]. Apo-SAA treatment of rat cortical microglia increased production of TNF-α and IL-1β, while TNF-α time-dependently raised
Saa1 expression in cultured OPCs [
12]. Our findings in the context of the above considerations, together with evidence for P2X
7R in the development of experimental autoimmune encephalomyelitis [
78] and microglia-oligodendrocyte crosstalk [
79], propose a vicious cycle of Apo-SAA, IL-1β, and TLR2 leading to the demise of OPCs.