|Home | About | Journals | Submit | Contact Us | Français|
The contribution of mutations in amyloid precursor protein (APP) and presenilin (PSEN) to familial Alzheimer's disease (AD) is well established. However, little is known about the molecular mechanisms leading to amyloid beta (Aβ) generation in sporadic AD. Increased brain ceramide levels have been associated with sporadic AD, and are a suggested risk factor. Serine palmitoyltransferase (SPT) is the first rate limiting enzyme in the de novo ceramide synthesis. However, the regulation of SPT is not yet understood. Evidence suggests that it may be post-transcriptionally regulated. Therefore, we investigated the role of miRNAs in the regulation of SPT and amyloid beta (Aβ) generation. We show that serine palmitoyltransferase (SPT) is upregulated in a subgroup of sporadic AD patient brains. This is further confirmed in mouse model studies of risk factors associated with AD. We identified that the loss of miR-137, -181c, -9 and 29a/b-1 increases SPT and in turn Aβ levels, and provides a mechanism for the elevated risk of AD associated with age, high saturated fat diet and gender. Finally, these results suggest SPT and the respective miRNAs may be potential therapeutic targets for sporadic AD.
It is well established that amyloid precursor protein (Aβ) accumulation in familial Alzheimer's disease (AD) is due to mutations in amyloid precursor protein (APP) and presenilin (PSEN) genes (reviewed in ref. (Cruts and Van Broeckhoven, 1998)). However, the mechanisms contributing to Aβ accumulation in sporadic AD is less well understood. Research thus far, consistently demonstrates that ceramide, a sphingolipid, is increased in AD patients (He et al., 2010; Cutler et al., 2004) and may contribute to the disease pathogenesis. Membrane ceramides are not only the major component of lipid rafts but they also contribute to AD pathology by facilitating the mislocation of BACE1 and γ-secretase to lipid rafts, and thereby promoting amyloid beta (Aβ) formation (Lee et al., 1998; Vetrivel et al., 2005). Inhibiting de novo ceramide synthesis has been shown to decrease the production of Aβ while exogenous addition of ceramide increased Aβ production (Puglielli et al., 2003; Patil et al., 2007). Numerous studies suggest a connection between ceramides and Aβ, and indicate increased ceramide levels may be an important risk factor for sporadic AD (Puglielli et al., 2003; Mattson et al., 2005).
SPT is the first rate limiting enzyme in the de novo ceramide synthesis pathway (Hannun and Obeid, 2008). Activation of SPT elevates ceramide levels (Perry et al., 2000) and inhibition of SPT decreases ceramide levels (Hojjati et al., 2005; Patil et al., 2007) and neuronal cell death by Aβ (Cutler et al., 2004), supporting SPT as an important regulator of ceramide. SPT is a heterodimer composed of serine palmitoyltransferase long chain 1 (SPTLC1) and either serine palmitoyltransferase long chain 2 (SPTLC2) or serine palmitoyltransferase long chain 3 (SPTLC3) (Rotthier et al., 2010). In the brain, SPTLC3 is lowly expressed while SPTLC1 and SPTLC2 are the major subunits (Hornemann et al., 2006). However, the regulation of these subunits and in turn SPT is not well understood. Cell culture studies demonstrate that SPT activity increases in response to various stimuli (i.e. etoposide or resveratrol), but without concomitant changes in SPTLC1 and SPTLC2 mRNA levels (Perry et al., 2000; Scarlatti et al., 2003), which have led researchers to hypothesize that SPT may be post-transcriptionally regulated.
Gene expression may be post-transcriptionally regulated through miRNAs, endogenous small RNAs of 21–25-nucleotides, that bind to 3′UTR of the target mRNA to cause translational repression or degradation of the mRNAs (He and Hannon, 2004). MiRNAs have been associated with neuronal differentiation, synaptic plasticity and memory formation (Sempere et al., 2004; Mehler and Mattick, 2006; Schratt et al., 2006). From miRNA expression profiles, several miRNAs are differentially expressed in AD patients (Lukiw, 2007; Cogswell et al., 2008; Hebert et al., 2008; Wang et al., 2008) and several have been reported to be specific or enriched in the brain (Sempere et al., 2004). Indeed a recent study reported altered expressions of several miRNAs in response to Aβ (Schonrock et al., 2010), suggesting the involvement of miRNA in sporadic AD. Therefore, we investigated whether miRNAs mediated the post-transcriptional regulation of SPT with respect to sporadic AD.
The AD (n=7) and control (n=7) neocortical brain samples were from the University of Kentucky (UK) Alzheimer's disease center tissue bank (ADC) as frozen tissues. The samples have been clinically diagnosed by neurologists, neuropathologists, neuropsychologists, and other staff members in the ADC clinic. Most samples have been obtained in <4hrs post-mortem interval (PMI). All individuals were between the ages of 88-99 years. The reference number (ref. #), gender, Braak stage, MMSE scores, frontal neuritic plaque numbers (NP), neurofibrillary tangle numbers (NFT) and ApoE genotype of the individuals are listed in Table 1. The cause of death of these individuals is multifactorial or unclear with pneumonia being the classical cause of death. The above information was provided by the University of Kentucky (UK) Alzheimer's disease center tissue bank (ADC).
Wild type C57BL/6 mice purchased from Jackson laboratories were used in the developmental study. Wild type mice on a hybrid background, C3H/He (Charles River) × C57BL/6 were used in the diet and gender specific studies. All procedures conducted were approved by the Institutional Animal Care and Use Committee at Michigan State University.
Primary astrocytes were isolated and cultured from <24hr old wild type Sprague-Dawley rat pups and 3 week old TgCRND8 (Centre for Research in Neurodegenerative Diseases) transgenic mouse pups, containing the APP 695-cDNA with both the Indiana and the Swedish mutations, in a hybrid C3H/He × C57BL/6 background (Chishti et al., 2001) as described (Patil et al., 2007). The TgCRND8 mice express the APP transgene at 5-folds higher than the endogenous APP under the control of the Syrian hamster prion promoter (Chishti et al., 2001).
Cells, mouse brain cortices and human brain neocortices (homogenized) were lysed in buffer: 1% (v/v) Triton, 0.1% (w/v) sodium dodecyl sulfate, 0.5% (w/v) deoxycholate, 20 mm Tris, pH 7.4, 150 mm, NaCl, 100 mm NaF, 1 mm Na3VO4, 1 mm EDTA, 1 mm EGTA ,1 mm phenylmethylsulfonyl fluoride and protease inhibitor cocktail (all chemicals from Sigma). The lysis was spun at 10,000 rpm for 10 min then the total protein concentration of the supernatant was measured by Bradford assays and was mixed with reducing loading buffer and heated at 94°C for 5 min. Immunoblot analysis was performed as described (Patil et al., 2007). Protein quantifications were conducted by normalizing to GAPDH or β-Actin. Western blots were quantified using Quantity One (BioRad) version 4.5.
Primary wild type and transgenic astrocytes were plated in 12 well plates and transfected for 24 – 72 hrs with 100-150nM of Syn-miRNA miScript miRNA mimic or anti-miR-RNA miScript miRNA inhibitor (Qiagen) or 500 ng – 2 μg of human SPTLC1 cDNA or 1.5 – 2 μg of luciferase vector construct using Lipofectamine RNAi/MAX or Lipofectamine 2000 following manufacturer's instructions. The SPTLC1 cDNA plasmid and SPTLC1 and SPTLC2 luciferase 3’UTR expression clones, containing the luciferase reporter gene and Renilla tracking gene and driven by the SV40 promoter, were purchased from Genecopoeia. The luciferase assay was conducted with dual luciferase assay kit (Luc-Pair™ miR Luciferase Assay Kit) (Genecopoeia). The Luciferase expression levels were normalized to Renilla expression levels.
Total mRNA was extracted using RNeasy Plus Mini Kit (Qiagen) and total RNA was quantified using ND-1000 nanodrop spectrophotometer. RNA quality control was performed by assessing OD 260/280 ratio. RNA quality of the control and human brain samples were performed using the Agilent Bioanalyzer 2100. In addition the PCR products were run on agarose gels. qRT-PCR was conducted using iQSYBR Green Supermix (BioRad) and MyiQ real-time PCR detection system following reverse transcription using iScript™ cDNA Synthesis Kit according to manufacturer's instructions. Primers include human SPTLC1: 5’ TGGAAGAGAGCACTGGGTCT 3’ and 5’ GCTACCTCCTTGATGGTGGA 3’; human SPTLC2: 5’ GAGACGCCTGAAAGAGATGG 3’ and 5’ TGGTATGAGCTGCTGACAGG 3’; human GAPDH: 5’GAGTCAACGGATTTGGTCGT 3’ and 5’ TTGATTTTGGAGGGATCTCG 3’; mouse Sptlc1:5’ AGTGGTGGGAGAGTCCCTTT 3’ and 5’ CAGTGACCACAACCCTGATG 3’ ; mouse Sptlc2: 5’ CCTGTCAGCAGCTCATACCA 3’ and 5’ CACACTGTCCTGGGAGGAAT 3’; mouse Gapdh: 5’ AACTTTGGCATTGTGGAAGG 3’ and 5’ ACACATTGGGGGTAGGAACA 3’; rat Sptlc1: 5’ ACCTGGAGCGACTGCTAAAA 3’ and 5’ ATCCCATAGTGCTCGGTGAC 3’; rat Sptlc2: 5’ TTGAGACTCACTGGCCCTCT 3’ and 5’ GGCCAGGAGGAGTCACATAA 3’; rat Gapdh: 5’ AGACAGCCGCATCTTCTTGT 3’ and 5’ CTTGCCGTGGGTAGAGTCAT 3’. Relative human, mouse and rat SPTLC1 and SPTLC2 expressions were calculated using the comparative CT method normalizing to their corresponding GAPDH expressions.
Total miRNAs were extracted using miRNeasy Mini Kit (Qiagen) and RNeasy MinElute Cleanup Kit (Qiagen) total RNA was quantified using ND-1000 nanodrop spectrophotometer. RNA quality control was performed by assessing OD 260/280 ratio. In addition the PCR products were run on agarose gels. qRT-PCR was conducted using miScript SYBR Green PCR Kit (Qiagen) and MyiQ real time PCR detection system following reverse transcription using miScript Reverse Transcription Kit (Qiagen) according to manufacturer's instructions. All miRNA primers were purchased from Qiagen and the relative expressions were calculated using the comparative CT method using RNU6B as the normalizing control.
Lipids were extracted from homogenized human brain neocortices and mouse brain cortices according to Bligh and Dyer (Bligh and Dyer, 1959). Tandem mass spectrometry (MS/MS) was performed using Quattro Premier XE (Waters), Acquity ultra performance liquid chromatography (Waters) (LC-MS/MS) and Mass Lynx 4.1 software. External ceramide standards were purchased from Matreya. C12:0 (Avanti, Polar Lipid Inc.) was used as the internal standard.
LCB1 (BD Transduction Laboratories™), SPTLC1 (proteintech), SPTLC2 (abcam), GAPDH (cell signaling), β-Actin (Sigma), β-Amyloid (cell signaling), β-Amyloid -4G8 clone (Covance).
Statistical significances were determined by using 2 tailed t tests and Spearman correlation (2 tailed-T distribution test).
The levels of ceramide and SPT protein expression were measured in the frontal brain cortices of 7 sporadic AD patients and 7 controls (see Table 1 for information on the patients). Of the vast number of distinct ceramide species (over 50 species), d18:1;18:0 and d18:1;16:0 are reported to be the major sphingolipid species in rat neurons (Valsecchi et al., 2007) and human brain (Ladisch et al., 1994). Consistent with previous reports (He et al., 2010; Cutler et al., 2004), ceramide levels, d18:1; 16:0 (P=0.037, student's t test) and d18:1; 18:0 (P=0.033), were significantly increased in this subgroup of AD patients (Figure 1A). Several reports have shown that the sphingomyelin levels either increased (Pettegrew et al., 2001; Bandaru et al., 2009) or remained unchanged (Han et al., 2002) in AD brains. In contrast, other researchers have shown that the sphingomyelin levels decreased (He et al. 2010; Cutler et al., 2004) in AD brain. We found that the sphingomyelin 18:1, 16:0 levels increased (P=0.045) while the 18:1, 18:0 levels remained unchanged in the subgroup of AD patient brain cortices studied (Figure 1B). This suggests that the increased ceramide levels in these patients are from the de novo synthesis pathway. Accordingly, SPTLC1 (P=0.004) and SPTLC2 (P=0.007) protein expression were significantly elevated in the autopsy AD brain samples (Figure 1C and D). However, SPTLC1 and SPTLC2 mRNA levels remained predominantly unchanged in the AD samples as measured by quantitative RT-PCR (qRT-PCR) (Figure 1E).
Previously, our group found that palmitate (a saturated fatty acid) increased de novo ceramide synthesis in astrocytes through SPT (Patil et al., 2007). Thus we treated wild-type primary rat astrocytes with palmitate for 24 hrs and found that the SPTLC1 (P=0.032) and SPTLC2 (P=0.015) protein levels (Figure 1F) increased without a concomitant change in their mRNA levels (Figure 1G), which is consistent with previous reports (Perry et al., 2000; Scarlatti et al., 2003). Overall, these results support that increased SPTLC1 and SPTLC2 expressions may be post-transcriptionally regulated in a subgroup of sporadic AD patients and in primary astrocytes cultured with palmitate. Thus, we proceeded to further elucidate the potential regulation of SPT by miRNAs.
Prediction algorithms miRbase (Griffiths-Jones et al., 2008), Targetscan (Lewis et al., 2005), Pictar (Krek et al., 2005) and miRanda (Betel et al., 2008) were used to select potential miRNAs that bind the human 3’UTR of SPTLC1 or SPTLC2 with strongly conserved (in mammals) target sites. Likely miRNA candidates were filtered according to the following criteria, it must be 1) predicted by at least 2 algorithms, and 2) down-regulated in AD patients or enriched in the brain. Of the miRNAs predicted by 2 or more algorithms to bind the 3’UTR of SPTLC1, miR-15a and miR-181c (Hebert et al., 2008) are reported to be down regulated in sporadic AD patients, while miR-137 and miR-124 (Sempere et al., 2004) are reported to be enriched in the brain. Of the miRNAs predicted by 2 or more algorithms to bind the 3’UTR of SPTLC2, miR-29a, miR-29b-1 and miR-9 are reported to be down-regulated in sporadic AD patients while miR-9 is also reported to be enriched in the brain.
Two luciferase reporter constructs were generated containing the 3’UTR of human SPTLC1 or SPTLC2. The miRNAs (sense) were co-transfected with the constructs and the luciferase expression was detected in wild-type primary rat astrocytes. While miR-137 (P=0.000016, student's t test) and miR-181c (P=0.0003) significantly decreased the luciferase expression of the construct containing the 3’UTR of SPTLC1, miR-15a and miR-124 did not (Figure 2A). The luciferase expression of the construct containing the 3’UTR of SPTLC2 decreased significantly upon co-transfection with miR-9 (P=1.023E-08), miR-29a (P=1.2E-07) or miR-29b-1(P=0.007) (Figure 2B). These results were confirmed by transfecting primary rat astrocytes with either the sense-miRs or anti-miRs (anti-sense) of their respective miRNAs following analysis of the endogenous miRNA expression levels in primary rat astrocytes (Figure 2I). MiR-137 and miR-181c significantly suppressed the endogenous SPTLC1 expression and cellular ceramide levels while anti-miR-137 and anti-miR-181c significantly enhanced the endogenous SPTLC1 (Figure 2C and 2E) and cellular ceramide (Figure 2G) levels upon transient tranfection. Similarly, miR-9, miR-29a and miR-29b-1 significantly suppressed the endogenous SPTLC2 and cellular ceramide levels while anti-miR-9 and anti-miRs-29a/b-1 significantly enhanced the endogenous SPTLC2 (Figure 2D and 2F) and cellular ceramide (Figure 2H) levels upon transient transfection.
The expression levels of miR-137 (P=0.006, student's t test), miR-181c (P=0.006), miR-9 (P=0.045), miR-29a (P=0.03) and miR-29b-1 (P=0.03) (Figure 3A-C), miR-15 (P=0.048) and miR-124 (P=0.002) (data not shown) were significantly down-regulated in the frontal cortices of the subgroup of sporadic AD patient.
Statistically significant negative correlations were observed between SPTLC1 and miR-137 (r=–0.807, P=0.0005, Spearman's correlation) (Figure 3D), miR-181c (r=-0.569, P=0.034) (Figure 3E), miR-15a (r=–0.59, P=0.026) and miR-124 (r=–0.67, P=0.009) (data not shown) in the control and the subgroup of AD patients. Significant negative correlations were also observed between SPTLC2 and miR-9 (r=–0.675, P=0.008) (Figure 3F), miR-29a (r=–0.603, P=0.023) (Figure 3G) and miR-29b-1 (r=–0.714, P=0.004) (Figure 3H) in the subgroup of AD patients. This negative correlation between the subunits of SPT and their corresponding miRNA expressions, coupled with the transient transfection results, suggest the possibility that changes in miR-137 or miR-181c, and miR-9, miR-29a or miR-29b-1 contribute, at least in part, to the overall protein expressions of SPTLC1 and SPTLC2, respectively, in AD.
Given that AD is an age related disorder(Bachman et al., 1992) we assessed the expressions of SPTLC1 (Figure 4A and B), SPTLC2 (Figure 4A and C) and their corresponding miRNAs (Figure 4D-F) with development. The protein, mRNA and miRNA expressions were evaluated in wild-type mice brain cortices from post-natal day 0 (P0) up to 18 months. This provided an independent confirmation of the correlation between SPTLC1, SPTLC2 and their respective miRNAs under non-pathological settings. During development, the expression levels of miR-137, miR-181c (Figure 4D) and miR-124 (data not shown) increased while SPTLC1 expression levels decreased with age (Figure 4A and B). Consistent with previous reports (Hebert et al., 2008), expression levels of miR-29a and miR-29b-1 were found to increase (Figure 4E) with development, while the expression levels of SPTLC2 decreased with age (Figure 4A and 4C). The Sptlc1 and Sptlc2 mRNA expression levels remained unchanged (stable) over the period analyzed (Figure 4F), a signature of miRNA regulation. These results suggest that miR-137, miR-181c, miR-29a and miR-29b-1 are developmentally regulated, with the highest expressions in adult mice. Concomitantly, protein analyses indicate that SPTLC1 and SPTLC2 have lower expression levels in adult mice, thereby further supporting a negative relationship between SPTLC1/2 and their corresponding miRNAs.
Increasing evidence in animal models suggest that a high fat diet aggravates the Aβ burden and thereby the AD pathology (Julien et al.). Indeed, high fat/ high cholesterol diets have been found to increase plasma ceramide levels in rodents (Shah et al., 2008). Moreover, prior research in our lab demonstrated that palmitate, a saturated fatty acid, increases ceramide levels and induces AD-like pathology in primary neuronal cell culture mediated by astrocytes (Patil et al., 2007). Therefore, the expression levels of ceramide, SPTLC1 and SPTLC2 and their corresponding miRNAs were measured in brain cortices of wild-type male mice fed a 60% kcal high fat diet for a period of 5 months (starting at 4 months of age). While ceramide (Figure 5A), SPTLC1 and SPTLC2 (Figure 5B) expression levels increased in mice fed a high fat diet, Sptlc1 and Sptlc2 mRNA levels remained unchanged (Figure 5C), supporting our hypothesis that SPTLC1/2 may be post-transcriptiponally regulated by miRNAs. Indeed, miR-137 (P=0.005, student's t test) (Figure 5D), miR-181c (P=0.026) (Figure 5D), miR-15a (P=0.01) (data not shown) and miR-9 (P=0.0027) (Figure 5E) expression levels were down-regulated in mice fed a high diet. In agreement with our in vivo results miR-137 (P=5.6E-05) (Figure 5F), miR-181c (P=2.2E-06) (Figure 5F), and miR-9 (P=1.9E-05) (Figure 5G) expression levels were down-regulated in wild-type primary rat astrocytes treated with palmitate, whereas SPTLC1/2 protein expression levels were upregulated (Figure 1F). However, miR-29a and miR-29b-1 expressions did not change with either a high fat diet (in vivo) or palmitate treatment (in vitro).
Evidence suggests that AD pathology may be more prevalent in females than in males (Bachman et al., 1992). Therefore, we evaluated the SPTLC1, SPTLC2 and miRNA expression levels in the brain cortices of female and male wild-type mice (9 months of age). Ceramide species, d18:1; 18:0 (P=0.0042, student's t test), d18:1; 16:0 (P=0.0045, student's t test) (Figure 6A), SPTLC1 (P=0.018) and SPTLC2 (P=0.014) (Figure 6B and 6C) protein expression levels were higher in females as compared to males, while the Sptlc1 and Sptlc2 mRNA levels remained unchanged (Figure 6D), further indicating that SPTLC1/2 may be post-transcriptionally regulated by miRNAs. Concomittantly, miR-137 (P=0.011), miR-181c (P=0.038) (Figure 6E), miR-124 (data not shown) miR-29a (P=0.031) and miR-29b-1 (P=0.004) (Figure 6F) expression levels are downregulated, but not miR-9, in female mice, while SPTLC1/2 protein expression levels are increased, further supporting a negative relationship between SPTLC1/2 and their target miRNAs.
A casual relationship between miR-29a/b-1, BACE1 activity and Aβ has been established by (Hebert et al., 2008). Therefore, we assessed whether a relationship exists between miR-137/181c and Aβ, mediated by SPTLC1. Statistically significant positive correlations were observed between SPTLC1 (western blot) and Aβ(42) protein levels (from ELISA) (r=0.76, P=0.002, Spearman's correlation) (Figure 7A), and SPTLC2 (western blot) and Aβ(42) protein levels (r=0.67, P=0.007) (Figure 7B) in the control and the subgroup of AD patients. Additionally, statistically significant negative correlations were observed between Aβ(42) and miR-137 (r=-0.75, P=0.003), miR-181c (r=–0.57, P=0.037), miR-9 (r=–0.7, P=0.007), miR-29a (r=–0.64, P=0.01) and miR-29b-1 (r=–0.569, P=0.037) in the control and the subgroup of AD patients. Furthermore, we performed gain- and loss-of-function experiments in primary astrocytes derived from transgenic mice expressing the human APP Swedish mutation. In these cells, over-expressing miR-137 or miR-181c down-regulated the endogenous expression levels of SPTLC1 (P=0.001) and Aβ (P=0.01) (Figure 7D and F). The functional affects were reversed upon transfection with the complementary anti-miRs-137 and -181c (Figure 7C and E). Thus, the loss of the suppressing activity of miR-137 and miR-181c led to increased Aβ production in cell culture. Additionally, transient overexpression of SPTLC1 (P=0.033) restored/increased Aβ expression levels in cells co-transfected with miR-137/-181c (P=0.005) (Figure 7D and F). In order to assess the direct role of miR-137, miR-181c, and thus SPTLC1 on Aβ expression, “target protectors” were designed against the targeted site on SPTLC1 for miR-137 and miR-181c. Primary astrocytes expressing the human APP Swedish mutation were transiently transfected with miR-137 or miR-181c along with their respective “target protectors” (Figure 7G, H, I and J). Both SPTLC1 and Aβ expression levels decreased significantly upon transfection with miR-137 (Figure 7G and I) or miR-181c (Figure 7H and J) along with a negative target protector. SPTLC1 and Aβ expression levels remained unchanged upon transfection with miR-137 (Figure 7G and I) or miR-181c (Figure 7H and J) along with their respective target protectors. Additionally, the transfection of anti-miR-137 (Figure 7G and I) or anti-miR-181c (Figure 7H and J) significantly increased Aβ and SPTLC1 expression levels.
We found a subgroup of sporadic AD patients exhibited increased levels of ceramides (this study and refs. (He et al., 2010; Cutler et al., 2004)) suggesting that ceramide may be a potential target for the treatment of AD. Increased ceramide levels have been associated with increased neutral SMase (N-SMase) levels in AD where Aβ induced N-SMase production (Jana and Pahan, 2010). In this study we observed that the Aβ levels increased with overexpression of SPTLC1. Therefore, ceramide rise through the de novo synthesis pathway upregulates Aβ levels, and the Aβ in turn may induce N-SMase activity to reinforce the production of ceramide, and thereby propagate a continual cycle of ceramide-Aβ generation.
We identified that a subgroup of sporadic AD patients exhibit increased levels of ceramides with concomitant increase in SPTLC1 and SPTLC2 protein expression levels in their brain cortices. This coupled with our animal and cell culture studies suggests that SPT may be a novel target for the treatment of AD. Further, SPTLC1/2 mRNA levels in these AD patient samples did not differ significantly from the levels in the control samples. This in combination with the luciferase assays and primary cell culture data suggests SPTLC1/2 may be post-transcriptionally regulated through miRNAs. Along these lines, we found negative correlations/relationships between the expression levels of miR-137/-181c and SPTLC1, and between miR-9/-29a/b-1 and SPTLC2 protein expressions, in sporadic AD brains, and developing, diet and gender specific mouse brains.
Apart from changes in miRNA expressions, binding sites for the transcription factor NFκB have been identified in the promoter region of Sptlc2 (Chang et al., 2011). However, this regulation may be tissue- and stimuli-specific as the tested experimental conditions did not impact brain SPT activity (Memon et al., 2001). Further, a significant increase in SPTLC2 protein levels was observed in human glioma tissue with only a slight increase in SPTLC2 mRNA (An et al., 2009). Interestingly, miR-29b is downregulated in glioblastomas (Cortez et al., 2010) suggesting miR-29b could be involved in elevating the SPTLC2 protein levels. In this present study we observed that SPTLC1 and SPTLC2 mRNA expression levels remained unchanged in the brain cortices of AD patients, as well as in primary astrocytes treated with palmitate and in mice fed a high fat diet. This was also observed with development and in both genders of mice. Further, correlation analyses coupled with the transfection studies in cells suggest changes in the miRNA levels, miR-137, -181c, -9 and 29a/b-1, could contribute to altered SPTLC1 and SPTLC2 expression levels in this subgroup of sporadic AD patient samples. Of the miRNAs identified to regulate SPT expression, increased expression levels of miR-137 has been shown to induce neurogenesis in hippocampus (Szulwach et al., 2010) while miR-9 is involved in neurogenesis and differentiation (Gao 2010; Coolen and Bally-Cuif, 2009). In addition, (Hebert et al., 2008) and (Cogswell et al., 2008) observed down-regulated miR-9 levels in AD patient brains, whereas (Lukiw, 2007) detected an up-regulation in AD. In contrast, miR-29a/b-1 was observed to be consistently down-regulated by (Hebert et al., 2008), (Wang et al., 2008) and (Shioya et al., 2010) in AD brains. Similarly, (Hebert et al., 2008) also detected down-regulated miR-181c expression levels in these patients. Consistent with these reports we found miR-181c, -9 and 29a/b-1 levels are down-regulated in the frontal cortex of the sub-group of AD patients in this study. In addition, we observed that miR-137 was also down-regulated in the frontal cortex of these 7 AD patients. In support of this, chromosomes 1p13.3-q31.1 region which includes the map location of miR-137, chromosome 1p21, has been linked to late-onset-AD (Butler et al., 2009). The map location of miR-181c, chromosome 19p13.13, has also been linked to late-onset-AD (Butler et al., 2009). We observed the suppression of SPTLC1 by miR-137 and miR-181c reduced Aβ expression levels in a target specific manner while over-expression of SPTLC1 and inhibition of miR-137 and miR-181c increased Aβ expression levels. This coupled with the fact that over-expression of BACE1 did not increase Aβ levels even though it increased β-CTF levels (Hebert et al., 2008), leaves open a possible role of ceramide, mediated by SPT, in transporting BACE1 and γ-secretase to the lipid rafts for amyloidogenic processing of APP. Inactive BACE1 and γ-secretase resides outside of the lipid rafts under non-pathological settings allowing non-amyloidogenic processing of APP, while under disease state the ceramides facilitate the trafficking of these pathogenic secretases to lipid rafts where they become active to produce Aβ (Cordy et al., 2003; Vetrivel et al., 2005; Ebina et al., 2009). Ceramide also increases Aβ production by stabilizing BACE1 (Puglielli et al., 2003; Costantini et al., 2007; Patil et al., 2007) through increased acetylation (Ko and Puglielli, 2009). MiR-9 and miR-29a/b previously have been identified as potential suppressors of BACE1 and thus associated with sporadic AD (Hebert et al., 2008). Given that SPT is also regulated by miR-9 and miR-29a/b-1, it further strengthens the contribution of SPT to the etiology of sporadic AD.
Of the miRNAs identified to regulate SPTLC1 expression, miR-137 was shown to be negatively regulated in adult neural stem cells, epigenetically and transcriptionally by MeCP2 and Sox2 through direct binding to the 5’ regulatory region (Szulwach et al., 2010). Research conducted to treat Rett syndrome, a disease caused by mutations in MeCP2, indicates that a high fat diet may increase MeCP2 levels (Haas et al., 1986; Liebhaber et al., 2003), providing a potential explanation for the reduced miR-137 levels in mice fed a high fat diet. The other miRNA identified to regulate SPTLC1, miR-181c, is positively regulated by Akt1 at the transcriptional level (Androulidaki et al., 2009). Akt activity is reduced in response to high fat diet (Tremblay et al., 2001), providing a possible mechanism for our observations that miR-181c expression levels are reduced in mice fed a high fat diet. Of the miRNAs identified to regulate SPTLC2 expression, miR-9 is negatively regulated by RE1-silensing transcription factor (REST) but positively regulated by cAMP-response element binding protein (CREB) (Laneve et al., 2010). High fat has been shown to suppresses CREB protein expression in the liver (Inoue et al., 2005) providing a possible explanation for the reduced miR-9 expression levels observed in high fat diet fed mice.
Many studies suggest that dysregulation of miRNA expression is aging-associated, and contributes to AD (Niwa et al., 2008). Additionally, increased DNA methylation have led to down-regulated expressions of miR-137 (Langevin et al., 2010) and miR-29 family (Koturbash et al., 2011) in females, indicative of differential expression of miRNA in a gender-specific manner and further supporting our observations in the mice study. Furthermore, maternal high fat diet has been shown to influence differential expression of mouse hepatic miRNAs in offsprings including down-regulation of miR-29a (Zhang et al., 2009). We observed a reduction of miR-137, -181 and -29a/b-1 expressions in females compared to males and a down-regulation of miR-137, -181c and -9 expression levels with high dietary fat intake. This raises an intriguing possibility that women consuming high fat diets may be at higher risk for SPT dysregulation and thus AD. Therefore, our results lend support to epidemiological factors such as age, gender and diet epigenetically regulating miRNAs and contributing to loss-of-function of miR-137, -181c, -9 and 29a/b-1, resulting in reduced suppression of SPT expression, and thereby increasing the ceramide levels and Aβ generation seen in sporadic AD.
We thank Dr. Peter Nelson and the UK ADC NIA P30-AG0-28383 for providing the human autopsy brain samples. We thank the MSU mass spectrometry facility for their support. We thank Aditi Upadhye for her assistance. This work was supported in part by the National Institute of Health (R01GM079688, R01GM089866 and R21RR024439), the National Science Foundation (CBET 0941055 and CBET-1049127) and the MSU Foundation.