SEARCH

SEARCH BY CITATION

Keywords:

  • Alzheimer disease;
  • microglia;
  • miR146;
  • neuroinflammation;
  • presenilin

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
Thumbnail image of graphical abstract

Microglia, the resident innate immune cells of the CNS, are the primary defenders against microbes and critical to CNS remodeling. Dysregulation of microglial behavior can lead to unchecked pro-inflammatory activity and subsequent neurodegeneration. The molecular mechanisms leading to chronic inflammation and microglial dysfunction in neurodegenerative diseases are not well-understood. It is known that patients with Presenilin 2 (PS2) mutations develop autosomal dominant Alzheimer disease. We have shown that a lack of normal PS2 function is associated with exaggerated microglia pro-inflammatory responses in vitro. To identify pathways by which PS2 regulates microglia and determine how PS2 dysfunction may lead to altered inflammatory pathways, we pursued an unbiased array approach to assess differential expression of microRNAs between murine PS2 knockout (KO) and wild-type microglia. We identified miR146, a negative regulator of monocyte pro-inflammatory response, as constitutively down-regulated in PS2 KO microglia. Consistent with a state of miR146 suppression, we found that PS2 KO microglia express higher levels of the miR146 target protein interleukin-1 receptor-associated kinase-1, and have increased NFκB transcriptional activity. We hypothesize that PS2 impacts microglial responses through modulation of miR146a. PS2 dysfunction, through aging or mutation, may contribute to neurodegeneration by influencing the pro-inflammatory behavior of microglia.

Presenilin 2 (PS2), a membrane associated protease, has been implicated in the pathogenesis of Alzheimer disease. We have previously shown that PS2 plays an important role in curbing the proinflammatory response in microglia. Here, we report the novel finding that PS2 participates in maintaining the basal and cytokine induced expression of the innate immunity regulating microRNA, miR146. These data suggest one mechanism by which PS2 works to reign in proinflammatory microglial behavior and that PS2 dysfunction or deficiency could thus result in unchecked proinflammatory activation contributing to neurodegeneration.

Abbreviations used
AD

Alzheimer disease

APP

amyloid precursor protein

IRAK-1

interleukin-1 receptor-associated kinase-1

KO

knockout

LPS

lipopolysaccharide

miRNAs

MicroRNAs

PDL

poly-d-lysine

PS

presenilin protein

PSEN1

presenilin 1

PSEN2

presenilin 2

TLR

toll-like receptor

Identification of Alzheimer Disease (AD) causing mutations in the amyloid precursor protein (APP), presenilin 1 (PSEN1), and presenilin 2 (PSEN2) genes has provided valuable insights into the mechanisms involved in AD neurodegeneration. However, the pathophysiology of AD remains enigmatic despite decades of rigorous study. The clinical and neuropathological similarities between early onset familial and late onset AD suggest that APP, PSEN1, and PSEN2 may also play pivotal roles in sporadic AD. Newly identified variants in these genes are associated with late onset familial AD cases (Cruchaga et al. 2012), thus supporting the possibility that altered function of APP, PSEN1, or PSEN2 contributes to AD pathogenesis in both familial and sporadic forms. We have recently described an early onset AD patient with a novel PSEN2 mutation predicted to lead to a premature termination codon causing either haploinsufficiency or a dramatically truncated protein (Jayadev et al. 2010b). Our previous work has demonstrated that the deficiency of presenilin 2 (PS2) protein function is associated with an exaggerated pro-inflammatory state in microglia (Jayadev et al. 2010a). Therefore, we propose that loss of PS2 function through mutation or cumulative effects of aging, may contribute to the neurotoxic inflammatory milieu of AD.

Neuroinflammation is a common pathological feature of neurodegenerative disease and a core characteristic of AD. Numerous epidemiological, mechanistic and discovery driven studies strongly suggest a functional role for neuroinflammation in promoting or exacerbating neurodegeneration (McGeer et al. 1996; Hensley 2010). During neuroinflammation, microglia execute functions with both neurotoxic and neuroprotective consequences in the CNS (Ransohoff and Cardona 2010; Aguzzi et al. 2013). For instance, unchecked anti-microbial cytokine release may lead to a CNS environment as inhospitable to neurons as it is to invading pathogens, potentially contributing to neurodegeneration in the setting of AD associated chronic inflammation. By understanding the molecular mechanisms behind the regulation of microglial inflammatory pathways, we may identify more specific targets for neuroimmunomodulatory interventions to ameliorate the resultant neurodegeneration.

In vivo murine models first suggested a position for presenilin proteins at the functional intersection between CNS inflammation and neurodegeneration. Presenilins are the catalytic subunit of the multi-protein γ-secretase complex, which cleaves type 1 membrane proteins involved in a panoply of regulatory pathways including apoptosis, cell differentiation, mitochondrial integrity, calcium regulation and inflammation (Haapasalo and Kovacs 2011; Ho and Shen 2011). PS2 knockout mice in which PS1 is deleted in adult forebrain neurons show progressive neurodegeneration, cognitive deficits and marked neuroinflammation. Similar findings were not observed in wild-type mice with a similar neuronal PSEN1 conditional deletion (Beglopoulos et al. 2004; Shen and Kelleher 2007). It seems possible therefore, that PS2 dysfunction has impacts on the developed CNS and can promote neuroinflammation. However, the mechanism by which PS2 influences microglia inflammatory behavior has not been determined.

MicroRNAs (miRNAs) are a class of small non-coding 22 nucleotide RNAs that regulate gene expression through post-transcriptional regulation. MiRNAs bind the 3′untranslated region of target mRNAs to promote mRNA degradation or interfere with translation (Bartel 2004). Recent reports demonstrate that miRNAs are key regulators of the intensity of the innate immune response (O'Connell et al. 2012). Experimental data have demonstrated a role for several specific miRNAs, for example, miR155, miR146a/b, and miR132, in regulating the expression of key innate immunity signaling proteins (O'Neill et al. 2011).

MiR-146a is a potent negative regulator of innate immunity and responsive to inflammatory cytokines and viral infection (Taganov et al. 2006; Hou et al. 2009; Zhao et al. 2011). It acts as a pivotal molecule in the negative feedback regulation of the powerful pro-inflammatory pathway mediated by nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB), a transcription factor regulating inflammation, immunity and cell survival. NFκB activation induces transcription of pro-inflammatory cytokines and is thus a critical factor in downstream innate immunity signaling (Newton and Dixit 2012). As a fast-acting inflammatory signal NFκB is subject to complex regulation and miRNAs are a significant component to the ‘fine-tuning’ of NFκB activity (Kondo et al. 2012). By suppressing expression of proteins that promote NFκB activity such as interleukin-1 receptor-associated kinase-1 (IRAK-1), IRAK-2 and tumor necrosis factor receptor-associated factor-6, miR146a attenuates proinflammatory responses, functioning as a brake in the potentially harmful proinflammatory response.

Our previous work had suggested that PS2 may play a role in the control of the microglial response to classical stimulation, and therefore we set out to identify important molecular factors in that putative regulation. We found that PS2 appears to influence levels of the negative regulator of innate immunity, miR146 with concomitant effects on protein levels of the miR146 target, IRAK-1. In addition to increased IRAK-1, PS2KO microglia demonstrate elevated NFκB activity. Taken together our findings support our hypothesis that PS2 is an important modulator of neuroinflammation and that dysfunction of PS2 may contribute to the aggravated inflammatory reaction in neurodegeneration.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Primary culture of microglia

All animals were housed and all experiments conducted according to the University of Washington IUCAC guidelines and approved by University of Washington IUCAC, Protocol #2856-01. C57Bl/6 and PS2 knockout mice, B6.129P-Psen2tm1- Bdes/J, were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Mice employed in these experiments were all post-natal day 3, and were not selected by sex prior to use. Mice husbandry and experiments were performed in compliance with ARRIVE guidelines.

Primary microglia and astrocytes were isolated from brain cortices of perinatal mice age P3 and P4. Cells were grown in T75 and T175 vented cap flasks coated with poly-d-lysine (PDL) in D10C growth media [Dulbecco's modified Eagle's medium (D5671; Sigma-Aldrich, St Louis, MO, USA) + 10% Heat Inactivated Horse serum (Life Technologies, Grand Island, NY, USA; 26050-088) + 10% Ham's Nutrient Mixture F12 (N4888; Sigma) + 2 mM l-Glutamine (G8540; Sigma) + 10 mM Hepes (Invitrogen; 11344-041) + 1x Pen/Strep (Gemini Bio-Products, West Sacramento, CA, USA; 400-109)]. The day after initial isolation media were aspirated and replaced with fresh D10C supplemented with 20% L929 conditioned D10C media. Cultures were incubated at 37°C with 5% CO2 for 9–14 days at which point microglia were harvest and plated for experiments. Media in the flasks were replaced with fresh D10C +20% L929 conditioned media for subsequent harvests at 7 day intervals for up to three harvests total per flask.

Primary microglia were plated at 3 × 105 cells dish on PDL-coated 35 mm dishes in D10C media. Cells were treated on Day 2 with 10 μ/mL mouse IFN-γ (485-MI; R & D Systems, Minneapolis, MN, USA) in D10C. Cells were collected 24 h after treatment and protein lysates prepared in buffer containing 50 mM Tris-HCl; 150 mM NaCl; 5 mM EDTA, pH 7.4; 1% Na deoxycolate; 0.1% sodium dodecyl sulfate; 1% Triton X-100; 1 mM phenylmethylsulfonyl fluoride; 7 mg/mL pepstatin; 5 μg/mL aprotinin; 5 μg/mL leupeptin; and 160 mM sodium orthovanadate. Primary microglia for miR146a over-expression were plated at 3e5 cells/dish on PDL-coated 34 mm dishes in D10C media and were infected at the time of plating with either lentivirus containing an miR146a over-expression vector (MMIR-146A-PA-1; System Biosciences) or control lentivirus containing a scrambled vector (MMIR-000-PA-1; System Biosciences, Mountain View, CA, USA). Both viruses were used at 5 TU/cell. Cells were incubated overnight and media were changed the next day. Cells were incubated an additional 48 h, after which point cells were collected and protein lysates prepared using Cell Lysis Buffer (9803S; Cell Signaling Technology, Danvers, MA, USA) with 1 mM phenylmethanesulfonyl fluoride.

NFκB activity

Primary microglia were plated at 5 × 104 cells per well on PDL-coated 96-well plates in standard growth medium D10C Cells were transduced at the time of plating with lentivirus containing an expression vector with firefly luciferase expression driven by a minimal cytomegalovirus (CMV) promoter and tandem repeats of the NFκB transcriptional response element (CLS-013L; Qiagen, Valencia, CA, USA). Cells were also transduced with viral particles expressing Renilla luciferase driven by a CMV promoter as an internal control (CLS-RCL-8; Qiagen). Both sets of particles were used at 5 viral particles/cell (MOI 5).

Twenty-four hours after plating and infection (Day 2), growth media were aspirated and replaced with macrophage serum free media. On Day 4 cells were treated with 100 ng/mL lipopolysaccharide (LPS) in macrophage serum free media. After 4 h of incubation with treatment, luciferase activity assay was performed using Promega Dual-Glo reagents and protocol (E2940; Promega, Madison, WI, USA).

Western blot analysis

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis western blots were performed with Invitrogen equipment and reagents using β-mercaptoethanol (Bio-Rad Laboratories, Hercules, CA, USA), 4–12% bis-Tris gradient gels, 3-(N-morpholino)-propanesuflonic acid - sodium dodecyl sulfate (MOPS-SDS) buffer and polyvinylidene difluoride membrane (162-0177; BioRad). Membranes were incubated in 5% milk/PBS with 0.1% Tween-20 with primary anti-PS2 Loop antibody (#529594; 11000 EMD Biosciences, Billerica, MA, USA), primary anti-IRAK-1 antibody (sc-5288; Santa Cruz Biotechnology, Santa Cruz, CA, USA), or primary anti- βactin antibody (A5441; Sigma-Aldrich) followed by secondary antibody (either anti-rabbit IgG horseradish peroxidase (NA934; GE Healthcare Life Sciences, Pittsburg, PA, USA) or anti-mouse IgG horseradish peroxidase (NA931V; Amersham). Quantification of protein was performed employing Image J software (National Institutes of Health, Bethesda, MD, USA).

RNA isolation and analysis

Primary microglia were plated at 1 × 106 cells on PDL-coated 6-cm dishes in D10C media. Cells were treated on Day 2 with 10 μ/mL IFNγ in D10C. Cells were collected 24 h after treatment and RNA was isolated using the High Pure miRNA Isolation Kit (05 080 575 001; Roche Applied Science,, Indianapolis, IN, USA). Pooled RNAs from four independent experiments were used for hybridization with Affymetrix whole mouse genome miRNA microarrays. The hybridizations, intensity measurements, and bioinformatics analyses were performed at the University of Washington Functional Genomics and Bioinformatics Core Laboratory. Microarray results were normalized using the GC Robust Multi-array Average method.

Real-Time PCR was performed to quantify miR146a and miR146b transcripts using Taqman miRNA Assays (Life Technologies) with the Taqman miRNA Reverse Transcription Kit (4366596; Life Technologies) and Taqman Universal MasterMix II (4440040; Life Technologies). Assays were performed on an ABI Step One Plus machine and results analyzed using the accompanying software.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

PS2 deficiency is associated with altered levels of miRNAs in microglia

PS2 knockout or knockdown results in increased pro-inflammatory cytokine release by microglia in response to several pro-inflammatory signals (Jayadev et al. 2010a). This suggests that in the absence of PS2, inflammatory signaling pathways are intact but abnormally regulated. To determine the mechanism of increased proinflammatory cytokine release we investigated molecular pathways described to have a role in innate immune regulation. miRNAs regulate cellular pathways by fine-tuning protein expression thereby exerting modest but precise modulatory control over signaling in response to extracellular or intracellular events. Because we observed that the impact of PS2 deficiency in microglia was a modulation of response intensity, we reasoned that PS2 may influence inflammatory signaling through miRNAs. We analyzed wild-type and PS2 knockout (KO) microglia for differential miRNA expression using a PCR-array platform (Affymetrix). Pooled RNA samples from four sets of wild-type and PS2 KO microglia were screened using this approach. We observed a subset of miRNAs was differentially regulated showing a greater than 1.5 fold difference between the two gentoypes (Table 1). The deficiency of PSEN2 in microglia was associated with altered expression of several miRNAs known to be involved in the regulation of innate immune signaling including miR-146a/b. MiR146a, is up-regulated by an array of inflammatory cytokines and viral proteins (Taganov et al. 2006; Motsch et al. 2007; Perry et al. 2008; Hou et al. 2009) which results in dampening of the inflammatory response through suppression of protein expression involved in NFκB activation. We observed that miR146a and miR146b are down-regulated in PS2 deficient microglia. The finding that basal miR146 expression is influenced by the presence of PS2 suggests that PS2 may modulate cytokine responses by influencing the level of negative innate immune regulatory miRNAs.

Table 1. PS2 impacts the microglia microRNA profile
microRNALog2 AvgExprFold decreasemicroRNALog2 AvgExprFold increase
  1. Primary microglia isolated from PS2 knockout mice show both decreased and increased (left and right panel respectively) transcript of microRNAs as compared to wild-type microglia. Both baseline level of transcript and fold change are displayed and are the result of pooled RNAs from four independent experiments.

miR-4947.992.28miR-1275.833.85
miR-501-5p4.862.15miR-130a5.273.34
miR-8056.512.06miR-34c5.322.92
miR-146b 5.83 1.95 miR-3794.542.41
miR-315.641.95miR-1955.302.34
miR-10a5.021.68miR-193b4.972.32
miR-7124.521.62miR-94.682.28
let-7g6.691.58miR-125a4.672.27
miR-146a 9.94 1.58 miR-181c5.362.19
miR-4834.501.52miR-7b11.132.13
miR-181a40311.50miR-3464.962.12
miR-29b4.901.49miR-1245.732.04
miR-29c4.071.47miR-34b5.111.90
miR-2127.101.44miR-2074.661.88
miR-133b4.251.43let-7e9.961.68

Microglia isolated from PS2 KO mice have decreased miR146

To validate the miRNA microarray findings, we measured miR146 levels by quantitative PCR in four additional independent primary microglial cultures. Consistent with the microarray results, PS2 KO microglia show a relative decrease in miR146a levels to less than 60% of wild type (Fig. 1). As a component of a negative feedback loop, miR146 is induced in response to inflammatory cytokines and microbial molecules such as LPS (Taganov et al. 2006). We found that LPS treatment of wild-type and PS2KO microglia leads to an induction of miR146, and the extent of induction did not differ between genotypes (data not shown). We also assessed the induction of miR146 by the classical pro-inflammatory cytokine IFNγ. In contrast to wild-type cells, PS2KO microglia did not show a significant increase in miR146a in response to 24 h stimulation with 10 μ/mL IFNγ (2.3 fold induction after IFNγ exposure vs. 1.4, respectively, Fig. 1b).

image

Figure 1. Presenilin 2 (PS2) knockout (KO) primary microglia have suppressed expression and induction of miR146a. (a) RNA collected from primary PS2 KO or wild-type microglia was assayed for miR146a transcript by RT-PCR. Levels of miR146a are significantly decreased in PS2 KO microglia. (n = 4, *p < 0.05). RT-PCR for miR146 transcript in RNA collected from wild-type and PS2KO microglia after 24 h incubation with (b) 10 μ/mL IFNγ or (c) 100 ng/mL lipopolysaccharide (LPS) reveals a significant difference between WT and PS2KO microglia in the induction of miR146 by IFNγ but not LPS (n = 4 for each set of data, *p < 0.05).

Download figure to PowerPoint

PS2 KO microglia have increased IRAK-1, a miR-146 target in the NFκB pathway

To determine if the magnitude of change in miR146a observed in PS2 KO microglia is sufficient to impact expression of functionally relevant targets, we assessed protein levels of IRAK-1 (O'Neill et al. 2011), an intracellular mediator of MyD88 dependent NFκB activation (Newton and Dixit 2012). IRAK-1 (O'Neill et al. 2011) is a target of miR-146a and promotes nuclear localization of NFκB, thereby activating proinflammatory transcriptional activity. We hypothesized that decreased levels of miR146a in PS2KO microglia would result in reduced gene silencing and increased protein expression of IRAK-1. PS2KO microglia did express significantly increased levels of IRAK-1 in the basal state (Fig. 2) consistent with the observed decrease miR146a in basal state microglia (Fig. 1). Our findings demonstrate that altered miR146 in PS2 KO microglia correlates with protein levels of its target, IRAK-1, suggesting PS2 may modulate the proinflammatory response by supporting miR-146a dependent down-regulation of IRAK-1.

image

Figure 2. MiR146a target, interleukin-1 receptor-associated kinase-1 (IRAK-1) is increased in presenilin 2 (PS2) knockout (KO) microglia. (a) Protein lysates prepared from wild-type and PS2 KO microglia were analyzed by Western blot for expression of IRAK-1 protein, (representative immunoblot shown). Densitometry from four independent experiments is graphed demonstrating that IRAK-1 is significantly increased in PS2KO microglia. (b) IRAK-1 levels measured by western blot after miR146 over-expression in PS2KO microglia are significantly decreased compared to PS2KO microglia over-expressing scrambled control sequence (*p < 0.05).

Download figure to PowerPoint

Microglia deficient in PS2 demonstrate increased NFκB transcriptional activity

IRAK-1 was initially identified as a mediator of IL-1 signaling (Cao et al. 1996) and is now also recognized as a critical regulator of Toll-like receptor (TLR) signal transduction (Swantek et al. 2000). IRAK-1 is activated by binding the TLR associated MyD88 protein and activated IRAK-1 binds to NFκB thereby promoting nuclear localization and transcriptional activity (Flannery and Bowie 2010). Since PS2 deficient microglia have an exaggerated proinflammatory response involving multiple pro-inflammatory cytokines and chemokines, we reasoned that PS2 may influence a common upstream component of the inflammatory response such as NFκB. We investigated whether the presence of PS2 impacts the activation of NFκB using a luciferase reporter assay.

Both wild-type and PS2KO microglia manifest low-level and statistically similar measures of NFκB transcriptional activity in basal culture conditions. However, upon stimulation with the TLR4 canonical ligand LPS for 4 h, both wild-type and PS2KO microglia demonstrated induction of NFκB mediated transcription (Fig. 3) and PS2KO microglia show significantly increased NFκB activity in response to a LPS compared with wild-type cells.

image

Figure 3. NFκB transcriptional activity is enhanced in presenilin 2 (PS2) knockout (KO) microglia in response to a Toll-like receptor 4 (TLR4) agonist. NFκB activity in PS2 KO or wild-type microglia was measured by a luciferase reporter assay. PS2 KO microglia show increased NFκB activity compared to wild type (n = 5, *p < 0.05).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Mutations in presenilin genes are the causative factor in a majority of early-onset familial AD cases and presenilin function is likely to be relevant to the development sporadic AD. We propose that PS2 influences innate immunity and subsequently, neurodegeneration, in part through modulating miRNAs. Here, we demonstrate that microglia isolated from PSEN2 knockout mice have decreased levels of miR146 compared with wild-type microglia. Furthermore, we have found that suppressed miR146a corresponded with increased basal IRAK-1 and exaggerated NFκB activity.

Presenilins are intra-membrane proteases translated as 50 kD holoproteins that undergo endoproteolytic cleavage to form the catalytic component of the γ-secretase complex (De Strooper 2003). More than 50 putative γ-secretase substrates regulate a variety of cellular mechanisms including inflammation, development, and synaptic plasticity (Kopan and Ilagan 2004; Hass et al. 2009). Computational analysis of multiple genomic datasets suggests that PS2 is ‘overrepresented’ in the immune system and co-regulated with innate immune signaling molecules in the TLR pathway (Yagi et al. 2008). Animal studies have shown that PS2 is regulated with NFκB during infection further suggesting a functional relationship between PS2 and the inflammatory response (Saban et al. 2002). Our findings demonstrating an association between PS2 and miR146 suggest one molecular mechanism responsible for exaggerated inflammation and exaggerated NFκB activity observed in several models of PS2 deficiency. Taken together these observations suggest an important role for PS2 in modulating innate immune responses.

Cytokine signaling is one arm of the innate immune response that is tightly regulated at the transcriptional, translational, and post-translational levels. The altered transcriptional regulation of inflammatory cytokines in PS2 deficient microglia suggested alteration of inflammatory signaling upstream of cytokine transcription. Indeed, we found that the absence of PS2 is associated with increased NFκB transcriptional activity, consistent with an overall pattern of increased pro-inflammatory signals in PS2 KO microglia. In further support of these observations, it has been shown that PS2 over-expression in cultured neuronal cells is associated with decreased NFκB transcription activity (Nguyen et al. 2005). Therefore, it is possible that PS2 down-regulates NFκB signaling and participates in the negative feedback response after pro-inflammatory stimulation. Proinflammatory stimuli lead to a rapid cytokine response by microglia. Excessive or prolonged responses, however, can be deleterious, thus complex cellular negative feedback pathways have developed to quell the process of inflammation. When NFκB signaling induces transcription of proinflammatory mediators it also initiates negative feedback pathways (Newton and Dixit 2012), including miRNAs that contribute to the tight temporal and amplitude control of the innate immune response.

MiRNAs mediate rapid post-transcriptional regulation of intracellular processes. Through gene silencing, miRNAs influence both molecular and cellular components of CNS innate immune responses (O'Connell et al. 2012; Ponomarev et al. 2013). One miRNA with known function in innate immunity is miR146a which suppresses expression of proinflammatory proteins to curb potentially harmful consequences of unchecked inflammation (Taganov et al. 2006). MiR146 null animals have measurably increased serum levels of the proinflammatory cytokines Tumor Necrosis Factor alpha (TNFα) and IL-6 in response to a sub-lethal injection of LPS and die faster from lethal LPS injections compared to control mice (Boldin et al. 2011). Our in vitro studies of murine PS2 KO microglia demonstrated a similarly exaggerated release of proinflammatory cytokines. In vivo studies also reveal that in aging mice, miR146 deficiency leads to a chronic inflammatory phenotype (Zhao et al. 2011) and miR146, though expressed, is dysfunctional in aging macrophages (Jiang et al. 2012). We observed that LPS induced miR146 as expected (Taganov et al. 2006) in both genotypes. IFNγ, however, induces significantly less miR146 in PS2 KO microglia compared to wild type. The greater capacity of wild-type microglia to induce miR146 compared to that of PS2KO cells is consistent with the hypothesis that PS2 KO microglia have impaired negative feedback of the pro-inflammatory response. The findings reported here also suggest that miR146 dysfunction in aging mice may relate to abnormal PS2 function in aging mice, since genetic PS2 deficiency in neonatal microglia replicates miR146 dysfunction observed in aging macrophages.

The relative increase of IRAK-1 protein in PS2 KO microglia demonstrates that the level of reduced miR146 expression associated with PS2 deficiency has a functional consequence. The magnitude of increased protein expression is consistent with the ‘fine-tuning’ role of miRNA in protein translation (O'Connell et al. 2012). The IRAK family members are intracellular signaling molecules that modulate innate immune pathways, including NFκB mediated transcription. Since, PS2KO microglia express higher levels of IRAK-1 at baseline and produce exaggerated pro-inflammatory cytokine release in response to a pro-inflammatory stimulus, the finding that miR146a is also decreased in PS2KO microglia suggested that PS2 can regulate inflammatory responses through modulation of miR146a expression.

The downstream influence of PS2 on IRAK-1 expression through miR146 may be a mechanism by which PS2 tempers the magnitude of microglia responses in the context of endotoxin tolerance. Furthermore, miR146 itself has been shown to play an important role in LPS tolerance (Nahid et al. 2009). It is possible that the consequences of PS2 functional deficits in microglia result in looser control of the NFκB/IRAK-1 inflammatory arm of endotoxin response, in part through regulation of miR146. PS2KO microglia showed an increase in NFκB in response to LPS compared with wild type, consistent with PS2KO microglia releasing increased amounts of pro-inflammatory cytokines after stimulation. However, LPS stimulation of wild-type and PS2KO microglia yielded similar levels of miR146 transcript. It is known that LPS can initiate signaling beyond its canonical TLR4 ligand role, and thus miR146 regulation may be impacted by pathways in addition to those initiated by TLR4. The maximal increase of miR146 induced after 24 h LPS stimulation was not influenced by PS2 in cultured microglia (Jayadev et al. 2010a). However, PS2KO microglia exposed to IFNγ, failed to induce miR146 to levels comparable to wild-type microglia. It is possible that the regulation of miR-146 expression by IFNγ has greater dynamic range. We did note that the IFNγ stimulation resulted in a higher miR146 induction compared not only to PS2KO but also compared to wild-type microglia stimulated by LPS. This suggests that PS2 regulates both basal and IFNγ induced miR146 expression. Our data do not demonstrate that PS2 influences miR146 expression following LPS treatment even though PS2 does impact baseline miR146 expression and thus IRAK-1. Therefore, taken in conjunction with our prior report that PS2KO microglia demonstrated exaggerated cytokine release following LPS treatment, it remains likely that PS2 modifies the proinflammatory behavior of microglia in the face of LPS stimulation.

An important remaining question is the mechanism(s) by which PS2 influences the level of miR146 in microglia. Direct transcriptional control of miR146 could be influenced by PS2 mediated cleavage of a transcriptional factor. For example γ-secretase cleavage of Notch directs Notch signaling and down-stream transcriptional targets (Jorissen and De Strooper 2010). Notch1 regulates immune cell phenotype in both the adaptive and innate immune systems. However, while the Notch pathway is a candidate mediator of presenilin impact on CNS function, conditional deletion of Notch1 and Notch2 in excitatory forebrain neurons does not result in the same neurodegenerative phenotype as does the deletion of presenilins 1 and 2 (Zheng et al. 2012). Therefore, the function of presenilins in CNS neurodegeneration and neuroinflammation likely extends beyond Notch or APP proteolysis. Of note, Notch-1 itself is also a predicted target of miR146 (Mei et al. 2011), and it is possible that multiple overlapping pathways involving multiple miRNAs are involved. The γ-secretase complex may have yet to be identified roles of miRNAs regulation, which is an ongoing subject of investigation in our laboratory. In addition, PS2 has non-enzymatic functions including modulation of calcium signaling and protein trafficking (Hass et al. 2009). Thus, there are both γ-secretase dependent and independent mechanisms through which PS2 may modulate levels of miRNAs in microglia.

Here, we show that PS2 expression is correlated with the expression level and activity of miR146, the miR146 target IRAK-1 and subsequent NFκB activity in cultured microglia. Given what is known about the negative regulatory role of miR146 in innate immunity, the data suggest a mechanistic explanation for why PS2 influences microglia pro-inflammatory cytokines release following innate immune activation by LPS and Aβ (Jayadev et al. 2010a). Presenilin function is currently considered a therapeutic target by γ-secretase inhibitors to treat AD. However, our data and those of others strongly suggest that γ-secretase activity impacts a wide array of biological processes (Hass et al. 2009; Wolfe 2012). Inhibiting PS function, through mutation, age related change or pharmacological inhibition may worsen chronic neuroinflammation, contribute to AD pathogenesis and provide further explanations for the lack of success of γ-secretase inhibitors in clinical trials. Our demonstration of the association between loss of PS2 function and innate immune miRNAs offer possible alternate targets for therapeutic intervention.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This study was supported by R01NS072395 (SJ), RAG08083 (SJ), R01NS073848 (GAG), with core facilities support from P30HD02274 and DK56465-P30 (Core Center of Excellence in Hematology). The authors thank Dr Richard Morrison for his advice and helpful discussions. The authors have no conflicts to declare.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References