|Home | About | Journals | Submit | Contact Us | Français|
Traumatic brain injury (TBI) is one of the most robust environmental risk factors for Alzheimer’s disease (AD). Compelling evidence is accumulating that a single event of TBI is associated with increased levels of Aβ. However, the underlying molecular mechanisms remain unknown. We report here that the BACE1 interacting protein, GGA3, is depleted while BACE1 levels increase in the acute phase post-injury (48hrs) in a mouse model of TBI. We further demonstrated the role of GGA3 in the regulation of BACE1 in vivo by showing that BACE1 levels are increased in the brain of GGA3 null mice. We next found that head trauma potentiates BACE1 elevation in GGA3 null mice in the acute phase post-TBI and discovered that GGA1, a GGA3 homologue, is a novel caspase-3 substrate depleted at 48 hrs post-TBI. Moreover, GGA1 silencing potentiates BACE1 elevation induced by GGA3 deletion in neurons in vitro indicating that GGA1 and GGA3 synergistically regulate BACE1. Accordingly, we found that levels of both GGA1 and GGA3 are depleted while BACE1 levels are increased in a series of post-mortem AD brains. Finally, we show that GGA3 haploinsufficiency results in sustained elevation of BACE1 and Aβ levels while GGA1 levels are restored in the subacute phase (7 days) post-injury. In conclusion, our data indicate that depletion of GGA1 and GGA3 engender a rapid and robust elevation of BACE1 in the acute phase post-injury. However, the efficient disposal of the acutely accumulated BACE1 solely depends on GGA3 levels in the sub-acute phase of injury.
Alzheimer's disease (AD) is a complex disease influenced by the actions of multiple genes, their interactions with each other and with the environment (Reitz et al., 2011). Traumatic brain injury (TBI) is one of the most robust environmental risk factors for AD. TBI has been suggested to accelerate the onset of AD and the severity of the injury positively correlates with increased risk (Jellinger, 2004). Compelling evidence is mounting that a single TBI event is associated with increased levels of Aβ and amyloid deposition both in humans and animal models (Johnson et al., 2010). Experimental TBI in rodents has been reported to increase levels of BACE1 (Blasko et al., 2004; Loane et al., 2009), suggesting that BACE1 elevation may be responsible for increased Aβ production following TBI. However the molecular mechanisms responsible for this post-injury elevation of BACE1 remain unknown.
BACE1 is a stress related protease that is also upregulated in AD brains (Cole and Vassar, 2008). We have shown that BACE1 increases following cerebral ischemia in rodents and proposed that caspase-mediated depletion of the BACE1 interacting molecule GGA3 is the underlying mechanism of BACE1 elevation. GGA3 depletion stabilizes BACE1 by impairing its sorting to lysosomes where it is normally degraded (Koh et al., 2005; Tesco et al., 2007; Kang et al., 2010). We also reported that levels of GGA3 are decreased and inversely correlated with BACE1 levels in post-mortem AD brains (Tesco et al., 2007). Levels of the GGA3 homologue, GGA1, are also decreased in AD brains (Wahle et al., 2006). GGA1 overexpression has been shown to decrease Aβ levels (von Arnim et al., 2006; Wahle et al., 2006), most likely due to the increased retrograde transport of BACE1 from the endosomes to the trans-Golgi network (Wahle et al., 2005). Accordingly, GGA1 RNAi-mediated downregulation results in increased Aβ (Wahle et al., 2006) and BACE1 accumulation in the endosomes (He et al., 2005).
Here we report that GGA3 is depleted while BACE1 levels increase in the acute phase post-injury in a mouse model of TBI. We confirmed the role of GGA3 in the regulation of BACE1 in vivo by showing that BACE1 levels are increased in the brain of GGA3 null mice. We then found that head trauma potentiates BACE1 elevation in GGA3 null mice concurrently with caspase-mediated depletion of GGA1. Furthermore, GGA1 silencing potentiates BACE1 elevation induced by GGA3 deletion in neurons in vitro. Collectively, these data indicate that GGA3 and GGA1 cooperatively regulate BACE1 degradation. Accordingly, decreased levels of GGA1 but not GGA2 are associated with depletion of GGA3 and elevation of BACE1 in a series of post-mortem AD brains. Finally, we show that GGA3 haploinsufficiency results in sustained elevation of BACE1 and Aβ production (while GGA1 levels are restored) in the sub-acute phase of injury. These findings indicate that depletion of GGA1 and GGA3 leads to a rapid and robust elevation of BACE1 in the acute phase post-injury. However, the efficient disposal of the acutely accumulated BACE1 depends solely on GGA3 levels in the sub-acute phase post-injury.
The monoclonal antibody m3.2 (rodent APP, sAPPα, β-CTF, Aβ) and Ab14 (against PS1) were a generous gift from Dr. P. Matthews and Dr. S. Gandy, respectively. Polyclonal anti-GGA1 (H-215) was purchased from Santa Cruz Biotechnology, polyclonal anti-GGA1 (a generous gift from Dr. M. Robinson), monoclonal GGA3 (612310) from BD Transduction Laboratories, polyclonal anti-GGA3 (4167) from Cell Signaling Technology, monoclonal anti-GGA2 antibody from BD Transduction Laboratories, monoclonal GAPDH (MAB374) from Millipore, polyclonal anti-BACE1 (PA1-757) from Thermo Scientific, polyclonal anti-BACE1 (D10E5) from Cell Signaling Technology, monoclonal anti-myc (9B11) from Cell Signaling Technology, polyclonal anti-caspase-3 (9665) from Cell Signaling Technology, polyclonal anti-APP CTF (A8717) from Sigma, monoclonal anti-NeuN (MAB377) from Millipore, polyclonal anti- β-galactosidase (559761) from MP Biomedicals and monoclonal anti-GFAP (GA5) from Millipore. Secondary anti-mouse IgG HRP (Thermo Scientific), anti-rabbit Ig G HRP (GE Healthcare), mouse IgG trueblot secondary (eBioscience).
5–7 month old and 18–24 month old Gga3−/−, Gga3+/−, Gga3+/+ mice of both sexes were utilized in these experiments. Mice were housed under standard conditions and food and water were available ad libitum. All animal experiments were carried out with the approval of Tufts University and Massachusetts General Hospital Institutional Animal Care and Use Committees.
The strain was created by microinjection of E14Tg2a.4 from 129P2/OlaHsd embryonic stem (ES) cells generated by BayGenomics (see http://baygenomics.ucsf.edu). The gene-trap vectors used within BayGenomics contain a splice-acceptor sequence upstream of a reporter gene, β-geo (a fusion of β-galactosidase and neomycin phosphotransferase II). These vectors insert randomly into introns. Chimeric males were mated to C57BL/6J females (Jackson laboratories) and the resulting heterozygous male was purchased. We have developed a PCR-based protocol to genotype the mice using 3 primers.
To confirm that only one copy the gene trap vector (β-galactosidase and neomycin resistance insert) inserted into the Gga3 gene the first 100 mice bred were also subjected to PCR analysis for the neomycin resistance gene using the following primers:
30 µm paraformaldehyde fixed frozen sections were incubated in 0.1% cresyl violet acetate (Sigma) at 37°C for 30 minutes. Stained sections were briefly rinsed in water and differentiated in 95% ethanol. Sections were dehydrated in ethanol and cleared in xylene (Sigma) prior to mounting in histomount (Invitrogen).
30 µM (coronal) and 60 µM (longitudinal) paraformaldehyde fixed frozen sections were stained O/N at 37°C using β-galactosidase reporter gene staining kit (Sigma) as per manufacturers instructions. Sections were mounted in GelMount aqueous mounting medium (EMS).
30 µM paraformaldehyde fixed frozen sections from Gga3+/+ and Gga3−/− mice were blocked for 1hr at RT in 5% Goat serum. Blocked sections were incubated in anti-β glactosidase antibody (1:10,000) in combination with either anti-NeuN (1:100) or anti-GFAP (1:500) in blocking solution at 4°C overnight. Sections were washed three times in PBS followed by incubation in AlexaFluor488 or 568 secondary antibodies for 2hrs at RT. Sections were washed and nuclei stained with DAPI. Sections were mounted on gelatin coated slides with fluorescent mounting medium (Dako). Fluorescent tissue sections were imaged on a Nikon A1R confocal microscope with Plan Apo VC 20× (air) and 60× (oil immersion) objectives. Background fluorescent staining of the negative control tissue from Gga3+/+ mice was used to set baseline for the laser strength and gain for image capture of the β-galactosidase staining in Gga3−/− tissue sections. Z-stacks were captured in 2µm increments and analyzed in NIS elements software (Nikon).
Briefly, 6 month old Gga3−/−, Gga3+/−, Gga3+/+ mice were anesthetized with 4% isoflurane (Anaquest, Memphis, TN, USA) in 70% N2O and 30% O2 using a Fluotec 3 vaporizer (Colonial Medical, Amherst, NH, USA) and positioned in a stereotaxic frame. Anesthesia was maintained using 2% to 3% isoflurane N2O/O2. Following a mid-line incision a 5 mm craniotomy was made using a portable drill over the left parietotemporal cortex, and the bone flap was removed. Mice were then subjected to CCI using a pneumatic cylinder with a 3-mm flat-tip impounder, velocity 6 m/sec, depth 0.6 mm, and impact duration 100 msec. Cotton swabs were used to absorb and control any bleeding post-impact. The bone flap was discarded and the scalp was sutured closed. Mice were allowed to recover in their cage.
2 weeks post-TBI mice were unrecoverably sedated with isoflurane followed by perfusion by 4% paraformaldehyde. Brains were carefully removed and fixed overnight at 4°C in 4% paraformaldehyde, followed by cryopreservation in 30% sucrose for 3 days at 4°C. Lesion volume analysis was performed as previously described (Wang et al., 2000). Briefly, cryopreserved brains were sectioned on a sliding microtome. 25µM sections were cut and every 20th section was collected (500µM intervals). Lesion volume (mm3) was determined using MCID Analysis software by carefully tracing the area of the cavitary lesion in each collected section. Each lesion area was measured three times and the average taken of the measurements.
H4-751 cells cultured in DMEM supplemented with 10% FBS and 200µg/ml G418 were incubated in the presence of 1µM STS (Calbiochem) with or without 50µM of the general caspase inhibitor zVAD (Calbiochem) for 8hrs at 37°C. Cell lysates were subjected to electrophoresis on a 4–12% Bis-Tris acrylamide gel and western blotting was performed using an anti-GGA1 antibody at a 1:1000 dilution (H-215, Santa Cruz).
The HA-GGA3 pcDNA4 plasmid was a generous gift from Dr. Waguri. The myc-GGA1 pCR3.1 plasmid was a generous gift from Dr. Juan Bonifacino (NIH). The D306AGGA1 plasmid was generated by site-directed mutagenesis of the my-CCA1 pCR3.1 plasmid. GGA3, GGA1 and D306AGGA1 were in vitro translated (IVT) in the presence of cold methionine using TNT Quick Coupled Transcription/Translation Systems as recommended by the manufacturer (Promega, Madison WI). 3.5µL of the GGA3 and GGA1 IVT reactions were incubated with or without increasing amounts (200 through 600ng) of recombinant caspase-3 (Pharmingen, San Diego, CA) in caspase reaction buffer at 37°C 16hrs. 3.5µl of the GGA1 and D306AGGA1 IVT reactions were incubated with or without 12µM recombinant caspase-3 for 16hrs at 37°C.
Site-directed mutagenesis was performed using the Quickchange site-directed mutagenesis kit (Stratagene) according to manufacturers instructions. Briefly, The D306A GGA1 plasmid was generated by PCR based site-directed mutagenesis of the myc-GGA1 pCR3.1 plasmid using the following primers:
Frozen tissue (Hemibrains or Hippocampi) from 5–7 month old and 18–24 month old Gga3−/−, Gga3+/− and Gga3+/+ mice were homogenized in modified RIPA buffer for immunoblotting. Briefly, snap frozen tissue from each mouse was homogenized in 10 volumes of modified RIPA buffer supplemented with protease and phosphatase inhibitors. The homogenate was then centrifuged at 20,000 × g for 15 mins at 4°C. The supernatant was collected and further clarified by a second centrifugation. Protein concentration was determined using the BCA method (Thermo Scientific). Homogenates were divided into aliquots and stored at −80°C until analysis.
Beta-secretase activity was measured in tissue homogenates using a highly sensitive FRET based cleavage assay as described by Irizarry et al., (2002) (Fukumoto et al., 2002) with modifications. Briefly, 96 well microplates were coated with anti-BACE1 antibody (D10E5, Cell Signaling, 1:1000) at RT for 8 hrs. Excess unbound antibody was removed by washing with PBS. The coated plate was blocked using 1%BSA/PBS overnight at 4°C. 50–100ug of RIPA lysed hemibrain homogenates from Gga3+/+, Gga3+/− and Gga3−/− mice were incubated for 1hr at 37°C followed by extensive washing with PBS. Bound BACE1 activity was measured by using 10µM fluorogenic β-secretase substrate IV (Millipore) in assay buffer (50mM sodium acetate pH4.5, 10mM NaCl, 0.002% Triton-X, 1mM DTT) in the dark at 37°C. The fluorescent signal resulting from cleavage of the substrate was measured at intervals over a 24hr time period using a Synergy 2 plate reader (excitation 340nm, emission 485nm) (Biotek). To control for non-specific cleavage of the fluorogenic substrate an equal amount of hemibrain lysate from a Bace1 null mouse was included in each assay.
Frozen tissue (Hemibrains or Hippocampi) from 5–7 month old and 18–24 month old Gga3−/−, Gga3+/− and Gga3+/+ mice were homogenized in a diethylamine extraction buffer (DEA) for analysis of endogenous secreted Aβ. Briefly, snap frozen tissue (hemibrain or hippocampus) from each mouse was homogenized in 10 volumes of chilled DEA extraction buffer (0.2% DEA, 50mM NaCl, 2mM PNT, 1mM AEBSF, protease and phosphatase inhibitor cocktail) and centrifuged at 100,000 × g for 1hr at 4°C in a Beckman Ultima Ultracentrifuge. The supernatant was collected and neutralized with 1/10th volume of 0.5M Tris-HCl (pH 6.8). Protein concentrations were determined via the BCA method. Homogenates were frozen at −80°C until ELISA analysis.
6 month old Gga3−/−, Gga3+/− and Gga3+/+ mice subjected to experimental TBI were euthanized by isoflurane sedation followed by decapitation 48hrs and 7days post-CCI. Ipsilateral contusions (cortex and hippocampus) were dissected and snap frozen in liquid nitrogen. The identical area in the contralateral (un-injured) hemisphere was also dissected and snap frozen in liquid nitrogen to serve as an internal control. The snap frozen tissues were homogenized in 10 volumes of modified RIPA buffer supplemented with protease inhibitors. The homogenate was then centrifuged at 20,000 × g for 15 mins at 4°C. The supernatant was collected and further clarified by a second centrifugation. Protein concentration was determined using the BCA method (Thermo Scientific). Homogenates were divided into aliquots and stored at −80°C until analysis.
15–50µg of RIPA extracted protein lysates were electrophoresed on 4–12% Bis-Tris NUPAGE gels (Invitrogen). Proteins were electroblotted onto PVDF (BioRad) membrane and blocked in 5% skim milk/TBST. Membranes were incubated in primary antibody O/N at 4°C, washed 3× in TBST and incubated in secondary antibody either anti-mouse-HRP or anti-rabbit-HRP (1:10,000 dilution) for 1hr at RT. Membranes were detected chemiluminescently using either ECL (Thermo Scientific), ECL-Plus (GE Healthcare) or Femto (Thermo Scientific) chemiluminescent reagents. Chemiluminescent signal was captured on an LAS4000 Fuji Imager.
Endogenous Aβ x-40 was detected in naÔve hemibrains and hippocampi (DEA soluble) and RIPA soluble CCI contusion extracts using the WAKOII rodent/human ELISA according manufacturers instructions. The WAKOII rodent/human ELISA employs the well-characterized BNT77/BA27 antibody system to detect Aβx-40 (Wako Chemicals USA, Inc. Richmond, VA). For detection of endogenous Aβ40 in the naive hemibrains and hippocampi, 100µl of DEA extract was analyzed. RIPA extracts (1–2.5µg/µl) prepared from the CCI contusions were used to detect RIPA soluble Aβ40 in the injured and contralateral hemispheres of mice subjected to CCI.
Cortical neurons were extracted from postnatal day 1 (P1) mouse pups as described in (Ninan and Arancio, 2004). Briefly, neocortex was dissected and digested with 0.25% trypsin at 37°C for 15 mins. Cells were cultured in Neurobasal A supplemented with 2% B27, 1% FBS, 0.4mM L-Glut, 6.6ng/ml 5 fluorodeoxyuridine and 16.4ng/ml uridine. A 50% media change was performed every four days. Mission shRNA plasmids expressing shRNA’s against murine GGA1 (TRCN0000115330) and a negative control (Sigma) were packaged into lentiviruses as described by Sena-Esteves et al., (2004) (Sena-Esteves et al., 2004). Lentiviruses were titred using the Quicktiter Lentivirus ELISA kit (Cell Biolabs). Cortical neurons were infected on DIV3 with lentiviruses expressing shRNA against murine GGA1 or a negative control at an MOI 5 for 6 hrs at 37°C. After 6 hrs the virus was replaced with conditioned Neurobasal A media from cortical neurons not subject to viral infection.
20AD and 19ND temporal cortex were obtained from the Brain Donation Program, Sun Health Research Institute, Sun City, Arizona, USA. Human tissue was collected with informed consent of subjects or next of kin and with ethical approval from the Sun Health IRB. Protein lysates were prepared by homogenization of frozen temporal cortex in modified RIPA buffer supplemented with protease inhibitors (Thermo Scientific).
Digital Images were collected using either a Versadoc (BioRad) or LAS-4000 (Fuji) imager. Densitometry analysis was performed on a Macintosh computer using QuantityOne software (BioRad). Statistical analysis was performed using Instat3 software (GraphPad Software Inc.). Unpaired or paired t-test was employed for data sets that passed normality test. Unpaired t-test with Welch correction was employed for data sets that passed a normality test but had different standard deviations. Mann-Whitney test was employed for data sets that did not pass a normality test.
Head trauma was induced by controlled cortical impact model (CCI) as previously described (Bermpohl et al., 2006) in C57Bl/6J mice. At various time intervals post-CCI, mice were euthanized and brains collected. Brain homogenates were prepared from the ipsilateral/injured (I) and contralateral/control (C) hemispheres and Western blot analysis was performed as previously described (Tesco et al., 2007). Following CCI, GGA3 was depleted while BACE1 increased (Fig.1). Since APP is also a substrate for caspase cleavage (LeBlanc, 2005), we tested whether APP undergoes caspase-mediated cleavage in the mouse brain following TBI and found that full-length APP protein levels were slightly decreased owing to the generation of a 90kDa fragment previously reported to be the N-terminal APP caspase fragment in cells undergoing apoptosis and in ischemic rat brain (Tesco et al., 2003) (Tesco et al., 2007). Thus, GGA3 depletion and elevated BACE1 levels occur concomitantly with caspase activation (Fig.1).
Next, RIPA-soluble Aβ40 levels were measured in contralateral (C) and injured (I) hemisphere of 6 C57Bl/6J mice 48 hrs post-injury using a commercial ELISA kit employing the well-characterized BNT77/BA27 antibody system (Wako Chemicals USA, Inc. Richmond, VA). Aβ levels were increased by ~50% (10.2 ± 0.82 vs 6.87 ± 0.24 pmol/g protein, p=0.0038) in the injured hemisphere compared to the contralateral hemisphere. As a negative control, Aβ levels were also measured in the injured and contralateral hemisphere of APP−/− mice (purchased from Jackson Laboratory) (data not shown). Our aim was to assess β-secretase activity, and since increases in BACE1 activity have previously been shown to increase both Aβ40 and Aβ42 (Vassar et al., 1999), we did not measure Aβ42 in these experiments. Collectively, these findings indicate that GGA3 depletion, mediated by caspase cleavage, and the consequent BACE1 elevation may be a common underlying mechanism of increased Aβ production following cerebral ischemia and TBI.
To investigate GGA3 regulation of BACE1 in vivo we analyzed mice with heterozygous or homozygous deletions of the Gga3 gene (Gga3+/− and Gga3−/−, respectively). Gga3 +/− founder mice were generated by the Mutant Mouse Regional Resource Centers at University of California, Davis (MMRRC) using a gene-trapping method (see http://www.genetrap.org). The gene-trap vectors contain a splice-acceptor sequence upstream of a reporter gene, β-geo (a fusion of β-galactosidase and neomycin phosphotransferase II) (Stryke et al., 2003). We developed a three primer PCR-based protocol to genotype the mice (Fig.2A,B) and determined that the gene-trap vector inserted at nucleotide 1173 of intron 1 of the Gga3 mouse gene (ENSMUST00000019135.916). Of the first 100 mice bred, every neomycin-positive mouse was also positive for the 1800 kb Gga3 null PCR product, confirming that this line has only one insertion of the gene trap construct. Intercrosses of Gga3+/− mice produced Gga3−/− mice in a normal Mendelian fashion that are healthy, viable and fertile. Analysis of neural tissue from 6 month old littermates revealed no gross anatomical defects in GGA3 null mice (Fig.2C). Beta-galactosidase staining of Gga3−/− mouse brain was used to determine the expression pattern of GGA3 in the adult mouse brain. GGA3 is ubiquitously expressed throughout the brain with the highest levels of expression in the hippocampus, cortex and cerebellum (Fig.2D). Confocal analysis of Gga3−/− tissue sections stained with an antibody against β-galactosidase revealed an expression pattern of GGA3 throughout the mouse brain identical to that seen with the enzymatic β-galactosidase staining (data not shown). Next, we determined that GGA3 is mainly expressed in neuronal cells by performing confocal microscopy analysis of brain sections co-stained with anti-β-galactosidase antibody and neuronal (NeuN) or glial (GFAP) markers. Co-localization studies were performed in the cortex, hippocampus (CA1, CA3, dentate gyrus) and midbrain. Co-staining in the CA1 region of the hippocampus is shown as an example (Fig. 2E–J).
Western blot analysis using two different anti-GGA3 antibodies revealed that the GGA3 protein is absent in brain extracts from Gga3−/− mice while the levels are reduced by ~ 50% in Gga3+/− mice (Fig.2K). Given that previous reports have shown that GGA1 levels are significantly decreased (~40%) in AD brains (Wahle et al., 2006) and that GGA1 overexpression decreases Aβ levels (von Arnim et al., 2006; Wahle et al., 2006), we assessed the levels of GGA1 in GGA3 null mice and found that the genetic ablation of GGA3 does not produce a compensatory increase in GGA1 (Fig.2K–L).
BACE1 and Aβ levels were assessed in primary cortical neuronal cultures (DIV 8) from Gga3+/+ and Gga3−/− P1 pups. Levels of BACE1 were found to be increased 2 fold (p=0.0003) in DIV 8 Gga3 −/− cortical neurons compared to Gga3+/+ neurons (Fig.3A,B). Levels of GGA1, APP and PS1 were unchanged (Fig.3A). Consistent with the BACE1 elevation, Aβx-40 levels were increased by ~ 3-fold (p=0.0001) in Gga3−/− compared to Gga3+/+ cultures (Fig.3C).
We measured BACE1 protein levels in 5–7 month old Gga3+/+, Gga3+/− and Gga3−/− mice of both sexes and found that BACE1 was increased ~30% in the brains of Gga3−/− mice compared to Gga3+/+ littermate controls. BACE1 levels were comparable between Gga3+/− and Gga3+/+ mice (Fig.3D). Further analysis revealed that the effect of the Gga3 genetic deletion is specific for BACE1 as their was no detectable change in SorLA, another GGA-binding protein (Rogaeva et al., 2007), PS1, the catalytic component of the γ-secretase complex, or APP levels between genotypes (Fig.3E).
β-secretase activity was assessed by two different methods: measuring β-CTF (C99) levels in brain extracts using the m3.2 antibody (kind gift from Paul Matthews) and a highly sensitive FRET based β-secretase cleavage assay. We found no statistical difference in β-CTF (C99) levels between Gga3−/−, Gga3+/− and Gga3+/+ littermate controls (Fig.3F). Additionally no statistical difference was observed in β-secretase activity between Gga3−/−, Gga3+/− and Gga3+/+ littermate controls as measured by cleavage of the fluorogenic β-secretase IV substrate (Fig.3G). Accordingly, no difference was observed in Aβx-40 levels between Gga3−/− and Gga3+/+ mice (Fig.3H). The differing results obtained in vitro and in vivo could be attributed to the presence of non-neuronal cells in brain extracts when compared to neuronal-enriched cultures. BACE1 expression is predominately neuronal (Vassar et al., 1999). As a consequence, BACE1 and Aβ levels are robustly increased in Gga3−/− primary neuronal cultures containing negligible amount of non-neuronal cells. While the presence of non-neuronal cells in brain extracts may mask the effect of Gga3 deletion on BACE1 and Aβ, resulting in a smaller increase in BACE1 and a failure to increase Aβ levels in the Gga3−/− brain. Another alternative explanation is suggested by previous reports showing that the effect of BACE1 overexpression or haploinsufficiency on Aβ levels (endogenous or human transgenic) is minimal in young adult mice (Luo et al., 2001; McConlogue et al., 2007; Hirata-Fukae et al., 2008). BACE1 haploinsufficiency results in a minimal reduction of Aβ in BACE1(+/−)/APP transgenic mice at 3 months of age and a ~90% and 50% reduction of Aβ levels at 13 and 18 months of age, respectively (McConlogue et al., 2007). More importantly, β-secretase activity has been shown to increase with age in human, monkey and mouse brain (Fukumoto et al., 2004). Given that the impact of BACE1 levels on Aβ production seems to be age-dependent, the deletion of Gga3 may not result in increased Aβ production in young adult mice explaining the discrepancies we observed in vitro and in vivo.
To assess the effect of Gga3 deletion on BACE1 elevation, β-secretase activity and Aβ production during aging, we analyzed the hippocampi from 6 month old and 18–24 month old Gga3+/+ and Gga3−/− mice of both sexes. We chose to analyze the hippocampi of mice rather than their hemibrains as both GGA3 and BACE1 are highly expressed in this brain region. We found that, in agreement with our previous data in hemibrain extracts (Fig.3D), BACE1 levels were increased by 25% (p=0.0001) in the hippocampi of Gga3−/− mice at 6 months of age compared to their WT littermate controls. This increase in BACE1 was replicated in the hippocampi of aged Gga3−/− mice when compared to their WT littermate controls (27% p <0.0001) (Fig.4A,B). Interestingly, overall levels of BACE1 did not increase in the hippocampi of Gga3+/+ and Gga3−/− mice with aging but rather decreased in both Gga3+/+ and Gga3−/− aged mice when compared to their 6 month old genetic counterparts (~−17 p <0.0001 and −18% p=0.0011, respectively). However the percentage elevation between genotypes remained the same (Fig.4B). In contrast, we did not detect any difference in PS1 levels (as a measure of gamma-secretase) between Gga3−/− and their WT littermate controls at either 6 months or 18–24 months of age (data not shown). Analysis of APP levels revealed that, in agreement with our previous data in hemibrain extracts (Fig.3E), there was no difference in APP levels in the hippocampi of 6 month old Gga3+/+ and Gga3−/− mice. Instead, there was a significant decrease (13%, p=0.0061) in APP levels in aged Gga3−/− mice compared to their WT littermate controls (Fig.4A,C). As was observed with BACE1 levels over aging, APP levels declined overall with aging in both Gga3+/+ and Gga3−/− mice, −21% (p<0.0001) and −31% (p<0.0001), respectively (Fig.4C). When we analyzed levels of β-CTF (C99) normalized to full length (fl) APP, we did not detect a difference in C99 levels between Gga3−/− mice and their WT littermate controls at 6 months of age in agreement with our previous data (Fig. 3F). However, C99 levels were increased by 21% (p =0.0043) in aged Gga3−/− mice when compared to their WT littermate controls (Fig.4A,D). The increase in β-secretase activity in aged Gga3−/− mice compared to their WT littermate controls as observed by an increase in C99 levels was confirmed by using a sensitive FRET based β-secretase cleavage assay, which showed that β-secretase activity was increased by 70% (p=0.0128) in the hippocampi of aged Gga3−/− mice compared to WT littermate controls (Fig 4E). However, despite increased β-secretase levels and activity in the hippocampi of aged Gga3−/− mice we did not detect increased Aβ levels in the hippocampus of the aged Gga3−/− mice compared to their WT controls (Fig 4F). Similar results were also obtained in cortical extracts from both 6 month old and 18–24 month old mice (data not shown). The significant decrease in APP along with an increase in C99 levels observed in aged Gga3−/− mice is most likely the result of increased β-secretase activity. Accordingly, previous reports have shown that levels of murine full-length APP (APP-fl) are decreased while levels of βAPP-CTFs are increased in transgenic (tg) mice expressing human (h) BACE1 (Bodendof et al. 2002; Rockenstein et al. 2005; Lee et al. 2005). Our data demonstrating an increase in β-secretase activity without a further increase in BACE1 protein levels in the aged Gga3 null mice are in agreement with a previous study reporting that β-secretase activity increases without change in BACE1 protein levels in the mouse brain with aging (Fukumoto et al., 2004). Although, Gga3 deletion leads to ~ 30% increase in BACE1 levels and increased β-secretase activity with aging, such increase does not seem to be sufficient to exert a measurable effect on Aβ levels.
To investigate the role of GGA3 on BACE1 elevation following experimental TBI, 6 month old Gga3−/−, Gga3+/− and Gga3+/+ mice of both sexes were subjected to CCI. Western blot analysis of tissues collected 48 hrs post-TBI revealed that BACE1 levels increase in the injured hemisphere of all mice (Fig.5A). BACE1 levels were significantly increased (13%, p=0.0359) in the injured hemisphere of Gga3−/− mice compared to Gga3+/+ mice. The BACE1 levels were 30% (p =0.0011) higher in the contralateral hemisphere of Gga3−/− mice compared to the Gga3+/+ mice. This is comparable to the difference in BACE1 levels observed between Gga3−/− and Gga3+/+ naïve mice (i.e. mice not subjected to TBI, see Fig.3D). As a consequence of the elevated BACE1 levels in the Gga3−/− contralateral hemisphere, the percentage increase in BACE1 levels between the injured and contralateral hemispheres in Gga3−/− mice was only 37% (p=0.0031) compared to 60% (p<0.0001) in Gga3+/+ mice (Fig.5A). Therefore, post-injury elevation of BACE1 is reduced by ~ 50% in the Gga3 null mice. The percentage increase in BACE1 levels in the Gga3+/− mice (33%, p=0.0020) was similar to that observed in the Gga3−/− mice, however GGA3 haploinsufficiency did not potentiate BACE1 elevation in the injured hemisphere but rather slightly increased BACE1 levels in the contralateral hemisphere (Fig.5A). CCI produces a focal injury, however modest alterations have been reported in the contralateral hemisphere depending on the severity of the injury (Hall et al., 2005), which may account for the increase observed in the contralateral hemisphere of Gga3+/− mice.
Next we assessed β-secretase activity in brain extracts of mice subjected to CCI, by measuring β-CTF (C99) levels and found that C99 levels were significantly increased in the injured hemispheres of Gga3+/+ and Gga3−/− but not Gga3+/− mice. As was observed with BACE1 levels, the greatest percentage increase in β-secretase activity occurred in the Gga3+/+ mice (92%, p=0.0017) compared Gga3−/− mice (70%, p<0.0001) (Fig.5B). APP C99 levels were significantly increased (25%, p=0.0172) in the contralateral hemisphere of Gga3−/− mice compared to Gga3+/+ mice accounting for the smaller percentage increase observed (Fig.5B). The increase in β-secretase activity observed in the contralateral hemisphere of Gga3−/− mice compared to Gga3+/+ mice post-TBI is interesting as we did not observe an increase in naïve mice (i.e. not subjected to TBI), and it is most likely due to the modest global effects that have been noted by other researchers following CCI (Hall et al., 2005). Aβ40 levels were significantly increased in the injured compared to contralateral hemispheres of Gga3+/+ and Gga3−/− but not Gga3+/− mice similarly to APP C99 levels. (Fig.5C). Levels of PS1 and APP were unchanged between the injured and contralateral hemisphere across genotypes (data not shown).
While TBI potentiates BACE1 elevation in Gga3 null mice, Aβ levels were similar in the injured hemisphere of Gga3+/+ and Gga3−/− mice. One possible explanation is that following TBI, levels of Aβ degrading enzymes (e.g. neprilysin) are upregulated (Chen et al., 2009), which may account for the observed dissociation between BACE1 and Aβ levels. Alternatively, it has been reported that high expression levels of BACE1 suppress Aβ production both in vitro and in vivo (Creemers et al., 2001; Lee et al., 2005). Thus, it is possible that the robust accumulation of BACE1 in Gga3 null mice following TBI produces a similar suppression of Aβ production.
Analysis of cavitary lesion volume in a subset of the Gga3+/+, Gga3+/− and Gga3−/− mice demonstrated that while Gga3 depletion potentiates BACE1 elevation in the acute phase following TBI it is insufficient to cause a measurable increase in lesion volume two weeks post-injury (mean ± SEM of 7 mice per genotype: Gga3+/+: 7.25±0.32mm3; Gga3+/−: 7.22±0.19mm3; Gga3−/−: 7.57±0.29mm3).
In summary, we determined that TBI potentiates BACE1 elevation in Gga3−/− mice at 48 hrs post-injury. Consequently, these findings indicate that in addition to the GGA3 mediated post-translational stabilization of BACE1 other mechanisms also contribute to BACE1 accumulation in the acute phase post-injury.
In an effort to find additional mechanisms responsible for the BACE1 elevation observed 48 hrs post-TBI in the Gga3−/− mice and given that previous reports have shown that GGA1 levels are significantly decreased (~40%) in AD brains (Wahle et al., 2006), we set out to determine whether GGA1 is depleted concurrently with GGA3 following TBI. We found that GGA3 levels were decreased in the injured vs. contralateral hemisphere of Gga3+/+ and Gga3+/− mice (55% p=0.0002 and 48% p=0.0184, respectively) (Fig.6A,B). Levels of GGA1 were also decreased in the injured hemispheres of Gga3−/− (61%, p<0.0001), Gga3+/− (53%, p=0.0019) and Gga3+/+(67%, p<0.0001) mice 48 hrs post-injury (Fig.6A,C). However, unlike GGA3 there was no difference in residual GGA1 levels observed in the injured hemispheres across genotypes.
Caspase activation is a well-known mechanism of programmed cell death following TBI in both humans and experimental models (Clark et al., 1999; Clark et al., 2000; Knoblach et al., 2002; Chen et al., 2004). To determine whether GGA1 is depleted during caspase activation H4-APP751 cells were treated with staurosporine (STS) alone or in association with a general caspase inhibitor (zVAD) for 16 hrs. Western Blot analysis with an anti-GGA1 antibody, targeted to the hinge and GAE domains of GGA1 (amino acids 286–500), revealed that full length GGA1 is cleaved into several fragments during apoptosis and that caspase inhibition (zVAD treatment) prevents GGA1 depletion (Fig.7A).
To determine if GGA1 is a caspase-3 substrate we subjected in vitro translated (IVT) GGA3 and GGA1 to incubation with increasing concentrations of recombinant caspase-3 overnight at 37°C. Western blot analysis of the IVT extracts using antibodies specific for GGA3 and GGA1 demonstrate that GGA1 is capable of being cleaved by caspase-3 and its cleavage generates a fragment pattern similar to that of GGA3 (Fig.7B). We have previously shown that caspase-3 cleaves GGA3 at D313 and generates a dominant negative molecule (Tesco et al., 2007), therefore we assessed whether GGA1 is cleaved at the corresponding aspartic acid residue (a.a. 306) (Fig. 7C). We mutagenized the D306 residue in the wild–type myc-tagged GGA1 plasmid to an alanine and subjected both wild type Myc-tagged GGA1 (w.t.) and mutated GGA1 (D306A) to in vitro translation followed by caspase-3 cleavage. Western Blot analysis with an anti-Myc antibody revealed that the D306A mutation prevented the generation of a specific caspase-3-derived fragment (Fig.7D, Fragment 1 D.N.). The GGA1 truncated molecule ending at D306 contains the VHS and GAT domain, which has been shown to function as a dominant negative by attenuating the retrograde transport of BACE1 from endosomes to the TGN (Wahle et al., 2005; Wahle et al., 2006). Thus, during apoptosis, caspase-mediated cleavage of GGA1 results not only in the degradation of GGA1, but also in the production of a GGA1 dominant negative molecule.
In order to investigate whether GGA1 depletion potentiates the BACE1 elevation induced by GGA3 deletion, GGA1 was silenced in primary cortical neurons (PCN) collected from P1 Gga3−/− pups using a lentivirus encoding a GGA1 shRNA. PCN's were infected with either a lentivirus encoding a GGA1shRNA or negative control shRNA at a multiplicity of infection (MOI) of 5 for 6 hrs on DIV3. Cells were collected at DIV15 and analyzed by western blotting (Fig.8A). BACE1 levels were increased 60% (p =0.0029) in Gga3−/− cortical neurons depleted of GGA1 compared to those expressing a negative control shRNA (Fig.8A). While levels of PS1 were unchanged, APP levels (in particular the immature isoform of APP) were significantly increased in cortical neurons depleted of GGA1 (26% increase p =0.0179) (Fig.8A). This increase in APP is specifically due to the depletion of GGA1 as we have previously shown that GGA3 depletion does not affect APP levels either in vitro or in vivo (see Tesco et al., 2007 and Fig.3A). The observed increases in BACE1 and APP levels appear to be specific for the deletion of GGA3 and GGA1 as caspase activation was ruled out by the detection of unchanged levels of full-length caspase-3 (Fig. 8A). These data indicate that GGA3 and GGA1 synergistically regulate BACE1 degradation and that caspase-mediated depletion of GGA1 is a leading candidate mechanism to explain BACE1 elevation in Gga3−/− mice at 48 hrs post-injury.
Several studies have demonstrated that BACE1 levels and activity are elevated in the brains of AD patients (Fukumoto et al., 2002; Holsinger et al., 2002; Tyler et al., 2002; Yang et al., 2003; Li et al., 2004). We have previously shown that GGA3 levels are significantly decreased (55%) and inversely correlated with BACE1 levels in the temporal cortex of patients with AD (Tesco et al., 2007). We re-analyzed this cohort of patients for GGA1 and GGA2 levels and demonstrated that like GGA3, GGA1 levels are also decreased in the temporal cortex of AD sufferers (30% decrease, p= 0.02, Fig.8 B,C), while levels of GGA2 levels are unchanged. Taken together these data suggest that depletion of both GGA3 and GGA1 contributes to the BACE1 elevation observed in AD brains.
We have previously demonstrated that GGA3 regulates BACE1 degradation by trafficking BACE1 to the lysosomes (Tesco et al., 2007; Kang et al., 2010). Thus, GGA3 is expected to play a key role in the disposal of the BACE1 accumulated during the acute phase post-injury. Given that GGA3 levels are decreased by ~50% in the temporal cortex of AD patients, Gga3+/− mice best represent the GGA3 reduction observed in the AD brains. In an attempt to address the important question of how acute brain injuries (e.g. stroke and head trauma) result in chronic accumulation of Aβ, we investigated the effect of GGA3 haploinsufficiency on BACE1 levels in the sub-acute phase of injury (7 days post-TBI). Six month old Gga3+/− and Gga3+/+ mice of both sexes were subjected to CCI. Western blot analysis of tissues collected 7 days post-TBI revealed that BACE1 levels were similar in the injured and contralateral hemisphere of Gga3+/+ mice. In contrast, BACE1 levels were still increased by ~20% (p=0.0025) in the injured hemisphere of Gga3+/− compared to the contralateral hemisphere (Fig.9A). Accordingly, C99 and Aβ40 levels were increased by ~40% (p=0.0313) and ~25% (p=0.0075), respectively, in the injured vs. contralateral hemisphere in Gga3+/− but not Gga3+/+ mice (Fig.9B,C). The observed increase in Aβ40 levels appears to be due to enhanced β-secretase activity, as APP and PS1 levels (a measure of gamma secretase) remain unchanged (data not shown). GGA1 levels were restored to normal while GGA3 was still slightly depleted in the injured hemisphere of both Gga3+/+ and Gga3+/− mice at 7 days post-injury (Fig.9D–F). This demonstrates that while multiple mechanisms including depletion of GGA3 and GGA1 are responsible for the elevation of BACE1 in the acute phase post-injury, in the sub-acute phase of injury haploinsufficiency of GGA3 is solely responsible for a sustained increase in BACE1 level and activity, and Aβ production.
We report here a novel GGA1/3-mediated mechanism underlying BACE1 elevation following TBI. We have found that GGA3 is depleted while BACE1 levels increase in the acute phase post-TBI. We have demonstrated the role of GGA3 in the regulation of BACE1 in vivo by showing that BACE1 levels are increased in the brain of GGA3 null mice and that Gga3 deletion leads to increased β-secretase activity with aging. We next asked to what extent the deletion of GGA3 affects BACE1 elevation following TBI; we found that head trauma potentiates BACE1 elevation in GGA3 null mice at 48 hrs post-TBI. Consequently, these findings indicate that in addition to the GGA3 mediated post-translational stabilization of BACE1 other mechanisms also contribute to BACE1 accumulation in the acute phase post-injury.
In an effort to find other mechanisms responsible for the BACE1 elevation observed at 48 hrs post-TBI, we discovered that GGA1 is depleted by caspase cleavage both in vitro following apoptosis and in vivo at 48 hrs post-TBI. Furthermore, GGA1 silencing potentiates BACE1 elevation induced by GGA3 deletion in neurons in vitro. Thus, we conclude that depletion of GGA1 by RNAi-mediated silencing or by caspase activation following TBI potentiates the BACE1 elevation produced by GGA3 deletion in vitro and in vivo, respectively. Importantly we have shown that decreased levels of GGA1 are associated with the depletion of GGA3 and BACE1 elevation observed in a series of post-mortem AD brains. These findings confirm and extend a previous report showing that GGA1 levels were significantly decreased (~40%) in AD brains (Wahle et al., 2006). Collectively, our data indicate that depletion of GGA1 and GGA3 synergistically elevate BACE1 levels and suggest that the BACE1 elevation observed in AD brains is mediated by the concurrent depletion of GGA1 and GGA3. Our data support a molecular mechanism by which BACE1 levels are regulated by caspase-mediated depletion of GGA1 and GGA3 in the acute phase of brain injury.
While caspase activation is a well-known mechanism of programmed cell death following TBI in both humans and experimental models (Clark et al., 1999; Clark et al., 2000; Knoblach et al., 2002; Chen et al., 2004), the role of caspase activation in neurodegenerative diseases has been matter of debate for a very long time. In support of a role for caspase activation in AD, we demonstrated that caspase-3 is activated in the same series of AD brains analyzed here (Tesco et al., 2007). Moreover, recent reports have provided compelling evidence that caspase activation is an early event, which plays a key role in neurodegeneration. Using in vivo multiphoton microscopy in association with fluorescent dyes, de Calignon et al. demonstrated that caspase activation precedes tangle formation in Tau transgenic mice (de Calignon et al., 2010). Surprisingly, this same study showed that neurons, in which caspase activation occurs, do not die acutely, but develop tangles. Similarly, caspase-3 activation is increased in hippocampal dendritic spines and is an early event associated with synaptic dysfunction in Tg2576 mice (D'Amelio et al., 2011). Caspase-3 activation also mediates the inhibition of LTP induced by Aβ42 toxicity (Frohlich et al., 2011). Thus, caspase-mediated depletion of GGA1 and GGA3 is a candidate mechanism contributing to BACE1 elevation in AD brains.
To date, several mechanisms have been proposed to explain the increased accumulation of BACE1 in AD brains: depletion of GGA3 (Tesco et al., 2007); increased phosphorylation of translation factor eIF2α (O'Connor et al., 2008); increased expression of a non coding anti-sense BACE1 transcript (Faghihi et al., 2008)) and decreased expression of the BACE1 regulating microRNA’s, miR-29 and miR-107 (Hebert et al., 2008) (Wang et al., 2008). Additionally, increasing evidence suggests that BACE1 is a stress induced protease. BACE1 levels have been shown to increase in cells exposed to oxidative stress (Tamagno et al., 2002; Tamagno et al., 2003; Tamagno et al., 2005; Tong et al., 2005), apoptosis (Tesco et al., 2007), in in vivo animal models following traumatic brain injury (TBI) (Blasko et al., 2004), cerebral ischemia (Wen et al., 2004) and impaired energy metabolism (Velliquette et al., 2005).
In addition to the post-translational regulation of BACE1 via caspase mediated depletion of GGA3 and GGA1, BACE1 levels can also be regulated at the transcriptional and translational level. The BACE1 promoter contains transcription factor binding sites for NF-kB, Sp1, YY1, PPARγ, HIF-1α, STAT1 and STAT3 (Christensen et al., 2004; Nowak et al., 2006; Sastre et al., 2006; Sun et al., 2006; Bourne et al., 2007; Wen et al., 2008; Cho et al., 2009). While at the translational level the phosphorylation of eIF2α under conditions of energy deprivation (Velliquette et al., 2005) and the loss of microRNA’s miR-107 (Wang et al., 2008), miR-29a/b (Hebert et al., 2008), miR-298 and miR-328 (Boissonneault et al., 2009) have been shown to increase BACE1 levels. Additionally, the BACE1 mRNA 50-untranslated region (50UTR) has been shown to act as a translational repressor (De Pietri Tonelli et al., 2004; Lammich et al., 2004; Rogers et al., 2004; Zhou and Song, 2006; Mihailovich et al., 2007).
A number of known and hypothesized BACE1 regulating factors are acutely altered (usually 3–24 hrs post-injury) following experimental TBI in rodents. These include activation/upregulation of the well known transcriptional molecules: STAT1 (Zhao et al., 2011), STAT3 ((Zhao et al., 2011; Oliva et al., 2012), HIF-1α (Anderson et al., 2009), NF-κβ (Sanz et al., 2002) and TNFα (a potent activator of NF-κβ pathways) (Lotocki et al., 2004). Moreover eiF2α phosphorylation is increased in the hippocampus of rodents 24 hrs post-injury following fluid percussion injury (Singleton et al., 2002). While TBI induced by controlled cortical impact in rodents has been shown to acutely decrease the levels of miR-107 and miR-328 (Redell et al., 2009; Wang et al., 2010). Thus, in addition to GGA3 and GGA1 depletion, any, or all of these additional BACE1 regulatory mechanisms may also contribute to the elevation of BACE1 observed in the acute phase post-injury. However, in order to confirm the contribution of these other mechanisms, post-injury levels of BACE1 would need to be analyzed in animals in which these molecules or pathways have been pharmacologically or genetically inhibited. At this stage these studies are currently missing. To date this is the first study providing evidence for a molecular mechanism of BACE1 elevation following TBI taking advantage of a novel mouse model null for Gga3.
In this study, we also attempted to address the important question of how acute brain injuries (e.g. stroke and head trauma) result in chronic neurodegeneration. We have previously demonstrated that GGA3 regulates BACE1 degradation by trafficking BACE1 to the lysosomes (Tesco et al., 2007; Kang et al., 2010). Thus, GGA3 is expected to play a key role in the disposal of the BACE1 that accumulates during the acute phase post-injury. Consequently, we set out to investigate the effect of GGA3 haploinsufficiency (which best resembles the depletion of GGA3 observed in AD brains) on BACE1 levels in the sub-acute phase of injury (7 days post-TBI). We found that GGA3 haploinsufficiency results in sustained elevation of BACE1 and Aβ levels in the sub-acute phase of injury when GGA1 levels are restored.
In conclusion, our data indicate that depletion of GGA1 and GGA3, and most likely additional transcriptional and post-transcriptional mechanisms (Rossner et al., 2006; Vassar et al., 2009), engender a rapid and robust elevation of BACE1 in the acute phase post-injury. However, the efficient disposal of the acutely accumulated BACE1 solely depends on GGA3 levels in the sub-acute phase of injury. As a consequence, impaired degradation of BACE1, e.g. owing to GGA3 haploinsufficiency, represents an attractive molecular mechanism linking acute brain injury to chronic Aβ production and neurodegeneration. Persistent Aβ elevation would be predicted to result in further caspase activation and ensuing GGA1 and GGA3 depletion. According to both our previous and current findings, this would serve to further elevate BACE1 and Aβ levels leading to a vicious cycle in individuals affected by TBI. As such, our data strongly support the hypothesis that subjects with lower levels of GGA3 may be at increased risk to develop AD following acute brain injury, whether it be stroke, TBI, or some other form of major brain insult. Regulation of BACE1 levels seems to be mediated by molecular mechanisms influenced by both genetic and environmental factors. Thus, BACE1 elevation may be the first step in increasing Aβ and triggering AD pathology, at least in the sporadic cases. The identification of the molecular mechanisms that regulate BACE1 is expected to lead the discovery of novel therapeutic targets for the treatment and or prevention of AD.
This work was supported by Award Number R01AG033016 (to GT) and 1R01AG025952 (to GT) from the National Institute On Aging, and a grant from Cure Alzheimer’s Fund (to GT). We thank Dr. Rudy Tanzi for his useful comments in the preparation and editing of this manuscript.
Conflict of interests: none