J Neural Transm (2009) 116:79–88 DOI 10.1007/s00702-008-0147-z
ALZHEIMER’S DISEASE AND RELATED DISORDERS - ORIGINAL ARTICLE
Inhibition of heme synthesis alters Amyloid Precursor Protein processing Luisa Benerini Gatta Æ Massimiliano Vitali Æ Rosanna Verardi Æ Paolo Arosio Æ Dario Finazzi
Received: 4 July 2008 / Accepted: 20 October 2008 / Published online: 11 November 2008 Ó Springer-Verlag 2008
Abstract Decay of mitochondria, energy failure and increased oxidative stress are features commonly detected in brains from Alzheimer’s disease (AD) patients. Recent findings indicate that neuronal heme deficiency may contribute to the appearance of those cytopathologies and potentially alter the course of AD. We repressed heme synthesis in cells by inhibiting ferrochelatase enzyme with small interfering RNA and N-methylprotoporphyrin IX. The treatments induced a severe perturbation of mitochondria and energy production, with decrease of the subunit II of Cytochrome c Oxidase, alteration of the membrane potential and a 50% reduction of intracellular ATP. The state and processing of the Amyloid Precursor Protein (APP) was also affected, with the appearance of APP aggregates and a significant decrease (30–40%) of sAPPa secretion, associated with perturbation of ADAM10 and TACE, enzymes involved in the a-secretase cleavage. The production of sAPPb was increased, without augment of Amyloid b generation. Our findings strengthen the
Electronic supplementary material The online version of this article (doi:10.1007/s00702-008-0147-z) contains supplementary material, which is available to authorized users. L. B. Gatta M. Vitali P. Arosio D. Finazzi (&) Dipartimento Materno Infantile e Tecnologie Biomediche, University of Brescia, viale Europa 11, 25123 Brescia, Italy e-mail:
[email protected] R. Verardi Servizio di Immunoematologia e Medicina Trasfusionale, Spedali Civili di Brescia, 25123 Brescia, Italy P. Arosio D. Finazzi III Laboratorio di Analisi Chimico Cliniche, Spedali Civili di Brescia, 25123 Brescia, Italy
hypothesis that a reduced availability of heme may play a role in AD pathogenesis. Keywords Heme Ferrochelatase N-Methylprotoporphyrin Alzheimer’s disease Oxidative stress Amyloid b Mitochondria
Background According to the amyloid cascade hypothesis (Golde et al. 2006; Hardy and Selkoe 2002) the excessive accumulation of Amyloid beta (Ab), a peptide derived from the Amyloid Precursor Protein (APP), is the primary biochemical event in Alzheimer’s disease (AD) pathogenesis. This hypothesis explains the familial forms of AD, rare genetically determined cases characterized by mutations either in APP or in Presenilins 1 or 2 genes (Bertram and Tanzi 2004), which lead to perturbation of APP intracellular processing, with a net increase of Ab42 production or a modification of the Ab40/Ab42 ratio (Bentahir et al. 2006; Kumar-Singh et al. 2006). As for the late-onset, sporadic forms of AD, which are the vast majority of the cases, it is not clear yet what starts the neurodegenerative process. Even if soluble Ab oligomers, presumably the most toxic forms of the peptide (Lambert et al. 1998; Walsh et al. 2002) are often increased in AD brain (Lacor et al. 2004), no conclusive explanation for such an accumulation is put forward. It is probable that many and different biological factors concur in determining the pathological process of sporadic AD, either by inducing minor but persistent changes in APP metabolism, or by modifying the intrinsic toxicity of Ab, or the susceptibility of selected population of nerve cells. Positron emission tomography analyses have revealed dramatic reductions in glucose utilization in the brain
123
80
regions most severely affected by AD neurodegeneration (de Leon and George 1983; Reichmann et al. 1993). Decreases in mitochondrial enzyme activities, such as the Pyruvate Dehydrogenase Complex, the a-Ketoglutarate Dehydrogenase Complex and, above all, Cytochrome c Oxidase Complex have been detected in the brain from AD patients compared to control subjects (Gibson et al. 1998; Mutisya et al. 1994; Reichmann et al. 1993). Mitochondrial dysfunction, with a fall in ATP production, can be linked to an augmentation of ROS production (Richter et al. 1995) and increased levels of products derived from oxidative damage are detectable in AD in the affected brain regions (Nunomura et al. 2004; Markesbery and Lovell 1998; Smith et al. 1991; Sultana et al. 2006). Their appearance seems to precede by many years the histological features and clinical symptoms of AD (Butterfield et al. 2006; Keller et al. 2005; Pratico et al. 2002). Altogether these data suggest that the triad of mitochondrial perturbations, hypometabolism and oxidative stress may contribute to AD development. Along this line, we found of interest the observation that a deficit in brain heme metabolism may concur to the derangement of mitochondria associated to aging and neurodegeneration (Atamna and Frey 2007; Dwyer et al. 2006). Ferrochelatase (FECH) is the terminal enzyme in the heme biosynthesis pathway and its inhibition by N-methylprotoporphyrin IX (NMP) caused increased oxidative stress, iron accumulation and the appearance of APP aggregates (Atamna et al. 2002), suggesting a link between heme deficiency, mitochondria decay and APP metabolism. We were interested in further investigating this line of evidence and set up a cellular model in which FECH activity was inhibited by treatment with specific small interfering RNA (siRNA) or NMP. We obtained evidence of mitochondria dysfunction, augmented oxidative damage and perturbations of APP intracellular processing, findings which support the hypothesis of an involvement of heme homeostasis in AD development.
Materials and methods Cell lines Human Embryonic Kidney 293 cells overexpressing the human APP 695 isoform (HEK/APP) were a kind gift by Dr. Paganetti (Novartis, Basel, Switzerland). Neuroblastoma SH-SY5Y cells were from Centro Substrati Cellulari, IZSLER, Brescia, Italy. HEK/APP were maintained in Dulbecco’s modified Eagle’s medium with 1 g/l glucose; SH-SY5Y cells were cultured in 50% MEM-EBSS, 50% HAM-F12. Both media were supplemented with 10% fetal bovine serum, 40 lg/ml gentamicine, and 2 mM glutamine. Differentiation of SH-SY5Y cells was obtained by
123
L. B. Gatta et al.
treatment with retinoic acid (10-5 M) for 14 days and checked for by observation at the microscope. siRNA design and delivery and NMP treatment FECH (Genbank accession no NM000140) specific siRNA were designed using TROD software (T7 RNAi Oligo Designer v1.1.2: http://www.unige.ch/sciences/biologie/ bicel/websoft/RNAi.html) and synthesized by T7 polymerase. The most effective one (sense sequence: 50 gcagcuuaaaugccauuuauuu 30 ) was then obtained from Ambion (Ambion, Applied Biosystems, CA, USA). siRNA for Tumor Necrosis Factor a Converting Enzyme (TACE) (sense sequence: 50 gaaacacuacuaacuuuuuuu 30 ) and A Disintegrin And Metalloproteinase Domain 10 (ADAM10) (sense sequence 50 gaauuuaaaguagaaacauuu 30 ) were used as controls. Lyophilized siRNAs were reconstituted according to manufacturer instructions, mixed with 6 ll of RNAi Transfection Reagent (Qiagen, MD, USA) and delivered to the cells (50,000–70,000/well) at the final concentration of 30 nM in complete culture medium. After 48 h, cells were either collected or exposed to a second dose of the same siRNA and incubated for further 72 h. The same temporal scheme was applied to expose HEK/ APP cells to 10 lM NMP (Frontier Scientific, UT, USA) for 48 or 120 h. mRNA quantification Efficiency and specificity of each siRNA was monitored by quantifying FECH mRNA by a real-time reverse transcription-polymerase chain reaction (RT-PCR) assay using the GeneAmp 5700 Sequence Detection System (Applied Biosystems). Total RNA was extracted from transfected and untransfected cells by RNAzol B (Biotec Laboratories Inc., TX, USA). The cDNAs were synthesized by oligo-dT primers and Improm-II RT (Promega Corporation, WI, USA) from 1 lg of total RNA, in a total volume of 20 ll. Two ll of the cDNA synthesis reaction was used for the specific amplification of the target transcript and of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as normalization control. The PCR was performed in a total volume of 25 ll at 50°C for 2 m, 95°C for 10 m followed by 40 cycles at 95°C for 15 s and at 60°C for 1 m, with Taqman Master Mix and the Assay-onTM Demand Gene Expression Product (Hs00164616_m1, Applied Biosystems) for FECH mRNA, or GAPDH forward primer (50 gaaggtgaaggtcggagtc 30 ), GAPDH reverse primer (50 gaagatggtgatgggatttc 30 ) and GAPDH probe (50 caagcttcccgttctcagcc 30 ) for GAPDH mRNA. The threshold cycle (CT) was determined for each sample in duplicates and quantification was performed using the comparative CT method (DDCT).
Inhibition of heme synthesis alters Amyloid Precursor Protein processing
Flow cytometry analysis Cells exposed to siRNA were trypsinized, washed and resuspended in complete medium. The PPIX fluorescence was initiated by k excitation at 488 nm and collected after passing trough a 610–670 nm long pass filter (FL3 detector) of the FACScalibur flow cytometry (Becton Dickinson FACS, CA, USA) (Brunner et al. 2001). Debris and aggregates were excluded from analysis using forward and side scatter signals. Cells vitality at the end of each incubation was investigated by the annexin-V/propidium iodide incorporation. Harvested cells were incubated for 10 min at room temperature in the dark with 100 ll of annexin V incubation reagent, including both fluorescein isothiocyanate-conjugated annexin V (FITC-Annexin V) and propidium iodide (PI) (Serotec, Oxford Biomedical Research, UK). FITCAnnexin V emission was measured as a green signal by the FL1 detector (530 nm); PI was measured as a violet signal (623 nm peak fluorescence) by the FL3 detector. Mitochondrial transmembrane potential (Dw) was evaluated by loading cells for 30 min at 37°C with 5 lg/ml of the JC-1 fluorophore (5,50 ,6,60 -tetrachloro-1,10 ,3,30 tetraethil-benzimidazolylcarbocyanine iodide) (Molecular Probes, Invitrogen Corporation, CA, USA) and flow cytometry or microscopy analysis. The color of the dye changes from green to orange as the mitochondrial membrane becomes more polarized. This property is due to the reversible formation of JC-1 aggregates upon membrane polarization. By flow cytometry, the mitochondrial depolarization is indicated by decrease in the red/green fluorescence intensity ratio (FL2/FL1 ratio). ATP quantification ATP levels were determined by a luciferin/luciferase-based ATP assay kit (CellTiter-GloTM Luminescent Cell Viability Assay, Promega) and the luminescence analysed according to manufacturer protocol.
81
by 10% SDS-PAGE transferred to a PDVF (GE Healthcare, Sweden) membrane and processed as described in the western blotting section using the anti-DNP antibody (1:150). Western blotting The intracellular levels of COXII, APP, ADAM10, TACE, b-APP Cleaving Enzyme (BACE), Transferrin Receptor (TFR) and actin proteins were determined by western blotting. Cells were washed, resuspended in lysis buffer (20 mM Tris–HCl pH 8, 200 mM LiCl, 1 mM EDTA, 0.1% Nonidet P-40) with Protease Inhibitor Cocktail (Sigma), and kept in ice for 1 h. After centrifugation at 12,0009g for 20 min at 4°C, the supernatant was collected and protein concentration determined by Bradford assay. A total of 2–20 lg of proteins were loaded on 10–18% SDSPAGE under reducing condition and transferred onto a PDVF membrane. Membranes were incubated in the blocking solution (2% ECL blocking solution in TBS, 0.05% Tween 20), for 1 h at room temperature, then exposed to primary antibodies diluted in blocking solution, for 16 h at 4°C. The antibodies used were: anti-COX subunit II (monoclonal 12C4, Molecular Probes; 1:1,000), anti-APP NH2-terminal (monoclonal 22C11, Chemicon Inc.; 1:2,000); anti-APP COOH-terminal (Sigma; 1:1,000); anti-Ab 1-17 (monoclonal 6E10, Sigma; 1:1,000); antisAPP b (IBL Co. LTD, Japan; 1:50); anti-ADAM10 COOH-terminal and anti-TACE COOH-terminal (ProSci Inc., CA, USA; 1:2,000 and 1:1,000, respectively); antiBACE COOH-terminal (ProSci Inc.; 1:100); anti-BACE NH2-terminal (Sigma, 1:100); anti-Actin (Sigma, 1:1,000); anti-TFR (Zymed Laboratories Inc., CA, USA; 1:500); anti-ubiquitin (Chemicon Inc.; 1:2,000) After washing in 19 TBS-T, the blots were incubated with the appropriate secondary antibodies (Dako, Denmark), conjugated with horseradish peroxide for 1 h at room temperature. Immunoreactive proteins were detected by ECL advance (GE Healthcare), visualized and quantified by the KODAK Image Station 440CF (Kodak, NY, USA).
Measurement of intracellular protein carbonyls Protein carbonyl levels were determined by the OxyBlotTM Protein Oxidation Detection Kit (Chemicon Inc., MA, USA). Cells were treated or not with 300 lM H2O2 for 30 m and lysed in 20 mM Tris–HCl pH 8, 200 mM LiCl, 1 mM EDTA, 0.1% Nonidet P-40, Protease Inhibitor Cocktail (Sigma, MI, USA), 1% 2-Mercaptoethanol (2ME). Equal amount (5 lg) of each sample was incubated for 20 m at room temperature with 5 ll of 12% sodium dodecyl sulfate (SDS) and 10 ll of 19 2.4-dinitrophenylhydrazine (DNPH) stock solution. The samples were neutralized with 7.5 ll of neutralization solution, separated
Quantification of Ab by enzyme-linked immunosorbent assay (ELISA) Amyloid b (1–40) levels were determined with a commercial sandwich ELISA assay following the manufacturer instruction (IBL). Briefly, serial dilutions of the Ab peptide (from 1,000 to 15.63 pg/ml) and 100 ll of the conditioned supernatants were incubated in the pre-coated wells at 4°C overnight. Each well was washed seven times with 100 ll of wash buffer (0.05% Tween 20 in phosphate buffer), 100 ll of the detecting antibody were added and the plate incubated for 1 h at 4°C. After extensive washing, 100 ll
123
82
**
FECH mRNA
100 80 60 40 20 0 M
b
30
F
T
F
T
*
25
(arbitrary units)
of the chromogen were added for 30 m at room temperature. The plate was read at 450 nm. The absorbance value was converted to pg/ml according to the Ab standard curve and further normalized by the protein concentration in the lysates of each sample.
a (%)
PPIX levels
Fig. 1 Inhibition of heme synthesis by FECH siRNA and NMP c treatment in HEK/APP cells. a FECH mRNA levels in HEK/APP cells exposed to FECH siRNA (white bar F) or TACE siRNA (grey bar T) for 120 h were quantified by real time RT-PCR and expressed as a percentage of the amount detected in each experiment in mocktransfected cells (black bar M). The graph shows mean values and standard deviations of at least three different experiments, ** P \ 0.001. b The accumulation of protoporphyrin IX (PPIX) was monitored in mock-transfected HEK/APP cells or transfected for 120 h with FECH and TACE siRNA by flow cytometry. The graph shows means and standard deviations of the relative PPIX levels of each single condition calculated from at least three different experiments, * P \ 0.017. c HEK/APP cells exposed to siRNA or NMP treatment were labeled with FITC-Annexin V (apoptotic cells) and PI (necrotic cells) and analyzed by flow cytometry. The graph shows the percentage of live (white bars), apoptotic (black bars) and necrotic (gray bars) cells out of the total of the gated cells. The modest increase in dead cells observed in treated versus control cells is not statistically significant. Data are means and standard deviations of at least three different experiments. M Mock-transfected cells; F cells exposed to FECH siRNA; T cells exposed to TACE siRNA; N cells exposed to NMP
L. B. Gatta et al.
20 15 10 5 0 M
c 100
Comparison of values between different samples was performed by paired, two tailed Student’s t test. Differences were defined as significant for P values lower than 0.05 and the Bonferroni correction was applied when appropriate.
Cell Number (%)
Statistical analysis 75
50
25
Results Inhibition of heme synthesis by FECH siRNA or NMP treatment in HEK/APP cells A well-documented in vitro model of heme deficiency relies upon the exposure of cells to NMP, a competitive inhibitor of FECH enzyme (Atamna et al. 2001; Chernova et al. 2006; Gamble et al. 2000; Jacobs et al. 1998), which causes a 40% or greater reduction in heme synthesis (Chernova et al. 2006 Jacobs et al. 1998). The drug is specific and mildly toxic, but it may interfere with hemeregulated pathways, thus we planned to include in our experimental model the down-regulation of FECH expression by siRNA silencing. We synthesized five different siRNAs and compared their efficiency at downregulating the FECH mRNA 48 h after the transfection. The most effective siRNA was selected for the chemical synthesis. Treatment of HEK/APP cells with 50 pmol of this siRNA reduced the levels of FECH mRNA by about
123
0 M
F
T
N
80% (not shown). We obtained a stronger and persistent reduction of FECH mRNA by transfecting cells twice, with two consecutive doses of siRNA with a 48 h time interval. Figure 1a shows the intracellular levels of FECH mRNA as measured by quantitative real time RT-PCR 120 h after the first transfection. Little FECH mRNA (10.6 ± 5.3%) persisted in HEK/APP cells transiently transfected by FECH siRNA as compared to mock-transfected cells. A control siRNA, that efficiently reduces TACE expression (Fig. 5a), did not modify FECH mRNA level (Fig. 1a). As indirect measure of FECH enzyme activity we monitored the intracellular accumulation of protoporphyrin IX (PPIX), the physiological substrate of FECH, at the end of the treatment with siRNAs. The autofluorescence measured by flow cytometry (excitation at
Inhibition of heme synthesis alters Amyloid Precursor Protein processing
83
T b
M
a Actin CoxII M % 100 80 60 40 20 0
F
T
*
N
** M
40µm
CoxII
F
T
F DNP:
c
% 100
*
*
N
N
+
-
d
FL2/FL1 97 68
80 60
43
40
29
20
21
0
H2O 2: M
F
T
N
a
-
M
b
+
c
d
+
N
e
f
+
-
F
Fig. 2 Inhibition of heme synthesis results in mitochondrial failure and increased oxidative stress in HEK/APP cells. a The intracellular amount of the subunit two of the mitochondrial Cytochrome c Oxidase (COXII) was determined in cells exposed either to siRNAs or NMP by western blotting and expressed as percentage of the levels found in mock-transfected cells (M). Actin levels were also determined as an internal control for protein load. A representative experiment and the graph with means and standard deviations from at least three different experiments are shown. * P \ 0.017; ** Statistical significance could not be calculated because of undetectable levels in some experiments. b, c Cells were loaded with the JC-1 fluorophore and analysed by microscopy (b) and flow cytometry (c). JC-1 is a lipophilic fluorescent cation that incorporates into the mitochondrial membrane, and forms aggregates at physiological membrane potential. Upon aggregation JC-1 gives an orange
fluorescence. Intact living cells stained with JC-1 therefore exhibit a pronounced orange fluorescence of mitochondria. The depolarization of the membranes converts the dye to the monomeric state that gives a green fluorescence; the red to green fluorescence intensity ratio (FL2/FL1) is a qualitative indicator for the Dw value. It is expressed as a percentage of the value calculated in mock-transfected cells (M), * P \ 0.017. d Intracellular protein carbonyls were determined by the OxyBlotTM Protein Oxidation Detection Kit. Cells were treated (lanes b, d, f) or not (lanes a, c, e) with H2O2 300 lM for 30 min. Lysates were exposed (?) or not (-) to 2,4-dinitrophenylhydrazine (DNPH) and analyzed by western blotting with anti-DNP antibody. One representative experiment out of three is shown in the figure. M Mock-transfected cells; F cells exposed to FECH siRNA; T cells exposed to TACE siRNA; N cells exposed to NMP
488 nm, emission at 620 nm) revealed an increase of the mean fluorescence intensity (23.4 ± 7.4, P \ 0.017) in HEK/APP cells exposed to FECH siRNA; a non-significant increase (mean intensity 9.8 ± 6.5) was detected in cells treated by TACE siRNA. The accumulation of the enzyme substrate let infer that the siRNA-induced degradation of FECH mRNA was associated to a significant decrease of FECH activity. Both the treatment with the chemical compound and the FECH gene silencing had minor effects upon cells viability, since only about 10% of the cells showed signs of necrosis/apoptosis, as assessed by propidium iodide incorporation and annexin V binding to the cellular surface (Fig. 1c). Altogether we could conclude that both the specific reduction of FECH enzyme synthesis by siRNA and the suppression of FECH enzyme activity by the competitive inhibitor in HEK/APP cells were effective and non-toxic 120 h after the treatment.
Inhibition of heme synthesis results in mitochondrial failure and increased oxidative stress in HEK/APP cells Heme has an important role in the assembly and function of many mitochondrial enzymes. To investigate the functional state of these organelles, we monitored three critical parameters. We first analyzed the levels of Cytochrome c Oxidase subunit COXII (Fig. 2a); the western blotting revealed a significant reduction of COXII protein (residual amount 51.0 ± 28.6%) in the silenced cells and its marked decrease in cells exposed to NMP. The control TACE siRNA caused a non-significant reduction. These data document an evident decrease of Complex IV levels and let infer a severe perturbation of the ETC function. When we analyzed the intracellular ATP amounts by a luciferinbased assay, we found a significant reduction by 40 and 50% in cells treated with FECH siRNA and with NMP, respectively (data not shown).
123
84
L. B. Gatta et al.
APP
APP
a
b
130 95
M
F
T
N
M
F
T
N
UBIQUITIN
APP
c
d
250 130 95 72
The inhibition of FECH by specific siRNA or NMP determined signs of increased oxidative stress, as assessed by the immunoblotting for oxidated proteins. The basal level of carbonylated proteins was higher in treated HEK/ APP cells than in control cells (Fig. 2d, lanes a, c, e), and the difference was even increased after incubation with 300 lM H2O2 (Fig. 2d, lanes b, d, f). Inhibition of heme synthesis alters the state of APP in HEK/APP cells
55
36
APP (DNP)
DNP
e
f
250 130 95 72
°
55
36
NMP H2O2
-
+
+ -
+ +
-
+
+ + - +
-
Fig. 3 Inhibition of heme synthesis alters the state of APP in HEK/ APP cells. a, b Equal amounts of total lysates from cells treated or not with siRNAs or NMP were immunoblotted with specific antibodies for full length APP. Full length APP (a, b) appears as a doublet of bands of about 110 and 120 kD. Aggregates are visible in cells treated with FECH siRNA (F) and NMP (N), particularly after long exposure (b). The presence of aggregates is associated with a reduction in the intensity of the upper band. c, d Equal amounts of total lysates from cells treated or not with NMP and H2O2 (300 mM/30 m) were immunoblotted with specific antibodies for full length APP (c) and for ubiquitin (d). Actin levels were used as internal control for protein load (not shown). e, f Equal amounts of total lysates from cells treated or not with NMP and H2O2 were exposed to DNPH and probed with an antibody for DNP (e). Then the membrane was stripped and probed with an anti-APP antibody (f). ° indicate a sample which was not exposed to the DNPH treatment
The signs of mitochondrial failure were further strengthened by the evaluation of the mitochondrial Dw, which depends upon inner membrane integrity and active ions transport. Cells were loaded with the carbocyanine JC1 dye and analyzed by microscopy (Fig. 2b) and flow cytometry (Fig. 2c). After treatment with FECH siRNA or NMP, we observed a significant increase of the green intracellular fluorescence with a decrease of the FL2/FL1 ratio by about 40% for both experimental conditions. Similar results were obtained by incubating cells with the DiOC6 dye, which incorporates within the membrane of healthy mitochondria (data not shown).
123
In control cells the immunoblotting for APP revealed a doublet of bands of about 110 and 120 kD, corresponding to the mature, fully glycosylated form and the immature one, respectively (Scheuermann et al. 2001). This migration pattern was altered in HEK/APP cells treated with FECH siRNA or NMP since aggregates of higher molecular mass were evident (Fig. 3a), especially after long exposures of the blotting (Fig. 3b). These data were also confirmed with a different antibody recognizing the COOH-terminal fragment of APP (Fig. 3c). Interestingly, the upper band of about 120 kD was less intense in treated cells than in control ones, whereas the lower band of 110 kD did not appear to be changed (Fig. 3a, c). Nonetheless the total amount of intracellular APP did not change, as indicated by the densitometric analysis (not shown). In an attempt to explore the nature of such alterations, lysates from cells treated or not with NMP and H2O2 were probed with anti-APP and anti-ubiquitin antibodies. The H2O2 treatment neither induced the appearance of APP aggregates nor modified the effects due to NMP (Fig. 3c). The ubiquitin immunoblotting revealed a smear of bands in control cells, the intensity of which was attenuated by NMP treatment (Fig. 3d). No bands with high molecular weight and corresponding to the APP migration pattern were evident in NMP-treated samples. The oxidation of proteins was more evident in samples exposed to NMP (Figs. 2d, 3e), but carbonylation did not seem to modify the migration pattern of APP in the experimental conditions we applied (Fig. 3f). The intracellular distribution of APP was studied by immunocytochemistry in cells cultured in the presence or not of NMP (Supplementary figure 1). In control cells the APP phenotype was characterized by a diffuse punctate pattern, which was more intense in the paranuclear area, probably corresponding to the Golgi apparatus; the surface of the cells was also stained, as better appreciated in non-permeabilized cells (not shown). The pattern did not change upon exposure to NMP for 120 h, either in intensity or in the cellular distribution of the fluorescent signal.
Inhibition of heme synthesis alters Amyloid Precursor Protein processing
85
M
a 130 sAPPα
95
T
Ad
N
TACE
b
130
C83
11
c 130
sAPPβ
95
95
b
ADAM10 95
d 11
AICD
M
e
F
a
F
Actin
N
d
**
*
100
T
c 42 %
sAPPα
ADAM10
*
100 60
*
60 20
0
f
M
**
300 200 100 0
g
Aβ
**
100 50 0
M
F
F
T
Ad
N
sAPPβ
T
N
Fig. 4 Inhibition of heme synthesis alters the processing of APP in HEK/APP cells. a–d Conditioned media from cells treated or not with siRNA or NMP were immunoblotted with specific antibodies for sAPPa (a). Total lysates from the same cells were analyzed with specific antibodies for the COOH-terminal fragment of APP (C83 in b and AICD in d), for sAPPb (c), and for actin (not shown). Graphs in e and f show the densitometric analysis of three different immunoblottings for sAPPa (e) and sAPPb (f). * P B 0.025; ** P \ 0.017. g Media from cells exposed or not to siRNA or NMP were collected after 120 h of incubation and analyzed by an ELISA assay specific for Ab1-40. Values are expressed as percentage of those obtained from the analysis of mock-transfected cells (M). F Cells exposed to FECH siRNA; T cells exposed to TACE siRNA; N cells exposed to NMP. * P B 0.025
Inhibition of heme synthesis alters the processing of APP in HEK/APP cells The immunoblotting for APP protein derivatives showed that FECH inhibition by NMP treatment led to a significant decrease (about 45%) of sAPPa in the conditioned medium
Fig. 5 Inhibition of heme synthesis perturbs TACE and ADAM10 proteins in HEK/APP cells. TACE and ADAM10 are membrane proteins with a-secretase activity. Total lysates from HEK/APP cells exposed or not to siRNAs for FECH, TACE, and ADAM10 or NMP were analyzed by immunoblotting with specific antibodies for TACE (a), ADAM10 (b) and Actin (c), which was used as an internal reference for protein load. a TACE immunoblotting reveals a specific band of 120 kD, which is significantly reduced by TACE siRNA (T). TACE aggregates are visible in cells exposed to FECH siRNA (F) or NMP (N). b ADAM10 shows up as two bands of approximately 100 kD. ADAM10 silencing decreased the intensity of both bands. Reductions in the intensity of the bands are also evident upon exposure to FECH siRNA (F) or NMP (N). d The graph shows mean and standard deviations of the densitometric analysis of ADAM10 protein levels from at least three experiments. * P \ 0.017; M Mocktransfected cells; F cells exposed to FECH siRNA; T cells exposed to TACE siRNA; Ad cells exposed to ADAM10 siRNA; N cells exposed to NMP
of HEK/APP cells (Fig. 4a, e) and that the exposure to FECH siRNA produced a smaller but still significant reduction of 28%. In parallel, we detected lower amounts of C83 COOH-terminal fragment in the lysates from treated cells (Fig. 4b), with NMP treatment being more effective (60% reduction) than FECH silencing (40% reduction). We then analyzed the possible effects on the bsecretase cut by western blotting with a sAPPb-specific antibody: NMP and FECH silencing determined evident raises in the level of sAPPb both in the cell lysates (Fig. 4c, f) and in the conditioned media (Supplementary figure 2). The total amount of proteins secreted by cells was not affected by inhibition of heme synthesis (Supplementary figure 3). Surprisingly, we did not find alterations in the amount of secreted Ab peptide in cells exposed to NMP for 120 h, but a significant reduction (24%) upon treatment with FECH siRNA (Fig. 4g). Noteworthy, NMP
123
86
determined a significant 30% reduction of the APP-intracellular domain (AICD, Fig. 4d). Inhibition of heme synthesis perturbs TACE and ADAM10 proteins in HEK/APP cells To study the effects of FECH inhibition on APP proteolytic processing, we analysed TACE and ADAM10 proteins, the enzymes more heavily involved in the a-secretase processing of APP; western blotting analysis for TACE revealed a specific band of about 120 kD (Fig. 5a), which was clearly reduced by TACE siRNA. The treatment with NMP or FECH siRNA determined the appearance of a smear of aggregates above the specific band (Fig. 5a); the aggregates were absent in mock-transfected cells and in cells exposed to control siRNAs (TACE and ADAM10). The immunoblotting for ADAM10 evidenced a doublet of specific bands of about 100 kD, which were clearly reduced upon exposure to ADAM10 siRNA (Fig. 5b); a significant decrease of the protein levels was also present upon inhibition of heme synthesis, with residual amounts of 67 and 50% in FECH siRNA and NMP treated cells, respectively (Fig. 5b, d). No aggregates were visible in the immunoblotting for ADAM10, actin (Fig. 5b, c), and the integral membrane protein TFR (not shown). Inhibition of heme synthesis alters mitochondrial function and APP metabolism in differentiated neuroblastoma SH-SY5Y cells To verify if the inhibition of heme synthesis could determine the same effects in a neuronal cell line not overexpressing APP protein, we performed the same experiments in differentiated neuroblastoma SH-SY5Y cells. Cells were exposed to retinoic acid for 2 weeks and then treated with FECH siRNA or NMP. All treatments were not toxic to the cells (not shown); the silencing of FECH expression was less efficient (about 70%) than in HEK/ APP cells (about 90%) (Supplementary figure 4a) and its effects on mitochondrial function and APP metabolism were less evident that in HEK/APP cells (Supplementary figure 4, b–e). The NMP treatment reduced COXII levels in mitochondria, induced the appearance of APP aggregates and clearly reduced the secretion of sAPPa (Supplementary figure 4b–e).
Discussion In spite of considerable advances in understanding the pathogenesis of AD, the search for the initiating event has been elusive until now. Our experimental model suggests that heme metabolism may have a role in the derangement
123
L. B. Gatta et al.
of APP processing and the establishment of the typical pathophysiological features of AD. We found conditions to repress FECH expression with specific FECH siRNA. Even though we could not measure FECH protein levels or activity directly, the efficacy of the treatment was confirmed by the intracellular accumulation of heme precursor protoporphyrins. FECH silencing was performed in parallel to the chemical inhibition of the enzyme by NMP; the biological effects induced by the latter were more robust, especially in differentiated SH-SY5Y cells. Presumably this is related to the incomplete silencing of FECH mRNA we obtained in HEK/APP (about 90%) and in SH-SY5Y cells (about 70%). The inhibition of heme production by siRNA or NMP led to dysfunction of mitochondria and energy failure, as indicated by the reduction of mitochondrial COXII protein and ATP cellular amount and by the perturbation of the mitochondrial membrane potential. Under these experimental settings, we detected a significant increase in intracellular protein carbonyls, an index of abnormal production of ROS. The increased oxidative damage is presumably linked to the alteration of the ETC (Atamna et al. 2001; Sohal 1993), even though the accumulation of heme precursors (Afonso et al. 1999) and/or the possible reduction of catalase activity may contribute to the phenomenon. When we investigated the possible effects upon APP metabolism, we found that the intracellular amount of the protein was not changed, but a small fraction of it underwent aggregation, confirming previous findings (Atamna et al. 2002). The appearance of aggregates was associated with reduced levels of the mature form of APP. Whether this is due to a selective susceptibility of the fully glycosylated form of APP or to an attenuation of the transport and maturation processes is not known, even though we did not observe gross modification of APP intracellular distribution upon NMP treatment. APP dimers and aggregates have been documented in normal cells and dimerization of APP may influence Ab production (Scheuermann et al. 2001). While the physiological APP dimerization is reversible, the aggregation we documented appeared to be of a covalent nature, being resistant to SDS and reducing agents (2-ME) and did not seem to depend on ubiquitination. Even though the acute exposure to H2O2 neither induced APP aggregates nor increased their amount in NMP-treated cells, the aggregates may be due to protein cross-linking induced by excessive ROS production; the interaction between APP and redox active metals and the accumulation of iron in heme deficient cells (Atamna et al. 2002) may exacerbate the process. The inhibition of heme synthesis was also associated to the appearance of higher molecular weight forms of TACE, another transmembrane protein. The process did not involve all transmembrane proteins in the cells, since ADAM10 and Transferrin Receptor migrated in SDS-
Inhibition of heme synthesis alters Amyloid Precursor Protein processing
PAGE according to their mass and without evidence of aggregates. Further studies are needed to understand the nature and the effects of the phenomenon. We also observed effects upon APP metabolism. While the total secretion of protein was not affected by the inhibition of heme synthesis, there was a boost of the bsecretase cleavage, with a significant elevation of sAPPb levels; concomitantly the a-secretase cleavage was reduced. TACE and ADAM10 are the major enzymes with a-secretase activity (Nunan and Small 2000); NMP and FECH silencing induced partial aggregation of TACE and a drop of ADAM10 protein levels, which may explain the decrease of the a-cleavage. We cannot exclude that APP aggregation is involved in this phenomenon, either by hindering the direct interaction of the substrate with the enzymes or by altering the intracellular trafficking patterns. Since the a- and b-processing of APP are inversely proportional (Skovronsky et al. 2000), the raise of the latter may be a direct consequence of the reduction of the former. Furthermore, BACE levels and activity have been shown to augment upon oxidative stress (Tamagno et al. 2002; Tong et al. 2005) and energy failure (Velliquette et al. 2005), which we have documented in heme deficient cells. Regardless of the increase in the b-processing of APP, we did not detect any raise in Ab secretion in cells with reduced heme synthesis. It is well known that ATP deficiency hinders the gamma-secretase activity and the production of the amyloidogenic peptide (Hoyer et al. 2005; Netzer et al. 2003). Therefore, it is possible that the energy failure observed in our experimental settings resulted in a reduced generation of AICD and of Ab, even when the b-cleavage was massively increased, such as upon NMP treatment. Our experimental settings determine an acute and drastic inhibition of heme production, probably not to be seen in vivo unless deficiencies of the heme biosynthetic pathway are present, such as in porphyrias. We are not aware of studies that analyzed the possible association between such metabolic disorders and AD, but growing evidence in the literature suggests that heme metabolism may have a role in the development of late-onset AD. Heme biosynthesis declines with age (Bitar and Weiner 1983) and in AD brain heme-a, the limiting factor for the assembly of Cytochrome c Oxidase Complex (Wielburski and Nelson 1994), seems to be particularly affected (Atamna and Frey 2004). The damage to mitochondria and the raise of ROS levels may contribute to switch the APP processing towards the amyloidogenic pathway (Tamagno et al. 2002; Tong et al. 2005). On the other hand, Ab itself has been shown to sequester heme forming a complex that has peroxidase activity (Atamna and Boyle 2006), but also protects mitochondria from Ab-dependent damage (Lloret et al. 2008). Such an interaction could alter Ab toxicity by
87
inhibiting its aggregation (Howlett et al. 1997) and lead to a vicious cycle, in which heme deficiency increases Ab generation and Ab peptide exacerbates heme paucity. Our data suggest that the inhibition of heme production recapitulates some pathophysiological features of AD and alters APP metabolism, thus strengthening the hypothesis that heme deficiency may contribute to sporadic AD pathogenesis. We believe that a better comprehension of the biochemical mechanisms linking heme metabolism to APP proteolysis could shed further light in the understanding of the neurodegenerative processes leading to AD. Acknowledgments We would like to thank Dr. Isabella Zanella (University of Brescia) and Dr. Daniela Galimberti (University of Milan) for critical reading of the manuscript. LBG and MV were supported by fellowships from the University of Brescia, and the work was supported by funds from the University of Brescia to DF. Conflicts of interest statement disclosed by the authors.
No conflicts of interest need to be
References Afonso S, Vanore G, Batlle A (1999) Protoporphyrin IX and oxidative stress. Free Radic Res 31(3):161–170 Atamna H, Frey WH 2nd (2004) A role for heme in Alzheimer’s disease: heme binds amyloid beta and has altered metabolism. Proc Natl Acad Sci USA 101(30):11153–11158 Atamna H, Boyle K (2006) Amyloid-beta peptide binds with heme to form a peroxidase: relationship to the cytopathologies of Alzheimer’s disease. Proc Natl Acad Sci USA 103(9):3381– 3386 Atamna H, Frey WH 2nd (2007) Mechanisms of mitochondrial dysfunction and energy deficiency in Alzheimer’s disease. Mitochondrion 7(5):297–310 Atamna H, Liu J, Ames BN (2001) Heme deficiency selectively interrupts assembly of mitochondrial complex IV in human fibroblasts: relevance to aging. J Biol Chem 276(51):48410– 48416 Atamna H, Killilea DW, Killilea AN, Ames BN (2002) Heme deficiency may be a factor in the mitochondrial and neuronal decay of aging. Proc Natl Acad Sci USA 99(23):14807–14812 Bentahir M, Nyabi O, Verhamme J, Tolia A, Horre K, Wiltfang J, Esselmann H, De Strooper B (2006) Presenilin clinical mutations can affect gamma-secretase activity by different mechanisms. J Neurochem 96(3):732–742 Bertram L, Tanzi RE (2004) The current status of Alzheimer’s disease genetics: what do we tell the patients? Pharmacol Res 50(4):385–396 Bitar M, Weiner M (1983) Modification of age-induced changes in heme and hemoproteins by testosterone in male rats. Mech Ageing Dev 23(3–4):285–296 Brunner H, Hausmann F, Krieg RC, Endlicher E, Scholmerich J, Knuechel R, Messmann H (2001) The effects of 5-aminolevulinic acid esters on protoporphyrin IX production in human adenocarcinoma cell lines. Photochem Photobiol 74(5):721–725 Butterfield DA, Poon HF, St Clair D, Keller JN, Pierce WM, Klein JB, Markesbery WR (2006) Redox proteomics identification of oxidatively modified hippocampal proteins in mild cognitive
123
88 impairment: insights into the development of Alzheimer’s disease. Neurobiol Dis 22(2):223–232 Chernova T, Nicotera P, Smith AG (2006) Heme deficiency is associated with senescence and causes suppression of N-methylD-aspartate receptor subunits expression in primary cortical neurons. Mol Pharmacol 69(3):697–705 de Leon MJ, George AE (1983) Computed tomography in aging and senile dementia of the Alzheimer type. Adv Neurol 38:103–122 Dwyer BE, Stone ML, Zhu X, Perry G, Smith MA (2006) Heme deficiency in Alzheimer’s disease: a possible connection to porphyria. J Biomed Biotechnol 3:24038–24042 Gamble JT, Dailey HA, Marks GS (2000) N-Methylprotoporphyrin is a more potent inhibitor of recombinant human than of recombinant chicken ferrochelatase. Drug Metab Dispos 28(4):373–375 Gibson GE, Sheu KF, Blass JP (1998) Abnormalities of mitochondrial enzymes in Alzheimer disease. J Neural Transm 105(8–9):855– 870 Golde TE, Dickson D, Hutton M (2006) Filling the gaps in the abeta cascade hypothesis of Alzheimer’s disease. Curr Alzheimer Res 3(5):421–430 Hardy J, Selkoe D (2002) The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics. Science 297(5580):353–356 Howlett D, Cutler P, Heales S, Camilleri P (1997) Hemin and related porphyrins inhibit beta-amyloid aggregation. FEBS Lett 417(2):249–251 Hoyer A, Bardenheuer HJ, Martin E, Plaschke K (2005) Amyloid precursor protein (APP) and its derivatives change after cellular energy depletion. An in vitro-study. J Neural Transm 112(2):239–253 Jacobs JM, Sinclair PR, Sinclair JF, Gorman N, Walton HS, Wood SG, Nichols C (1998) Formation of zinc protoporphyrin in cultured hepatocytes: effects of ferrochelatase inhibition, iron chelation or lead. Toxicology 125(2–3):95–105 Keller JN, Schmitt FA, Scheff SW, Ding Q, Chen Q, Butterfield DA, Markesbery WR (2005) Evidence of increased oxidative damage in subjects with mild cognitive impairment. Neurology 64(7):1152–1156 Kumar-Singh S, Theuns J, Van Broeck B, Pirici D, Vennekens K, Corsmit E, Cruts M, Dermaut B, Wang R, Van Broeckhoven C (2006) Mean age-of-onset of familial Alzheimer disease caused by presenilin mutations correlates with both increased Abeta42 and decreased Abeta40. Hum Mutat 27(7):686–695 Lacor PN, Buniel MC, Chang L, Fernandez SJ, Gong Y, Viola KL, Lambert MP, Velasco PT, Bigio EH, Finch CE, Krafft GA, Klein WL (2004) Synaptic targeting by Alzheimer’s-related amyloid beta oligomers. J Neurosci 24(45):10191–10200 Lambert MP, Barlow AK, Chromy BA, Edwards C, Freed R, Liosatos M, Morgan TE, Rozovsky I, Trommer B, Viola KL, Wals P, Zhang C, Finch CE, Krafft GA, Klein WL (1998) Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proc Natl Acad Sci USA 95(11):6448–6453 Lloret A, Badı´a MC, Mora NJ, Ortega A, Pallardo´ FV, Alonso MD, Atamna H, Vin˜a J (2008) Gender and age-dependent differences in the mitochondrial apoptogenic pathway in Alzheimer’s disease. Free Radic Biol Med 44(12):2019–2025 Markesbery WR, Lovell MA (1998) Four-hydroxynonenal, a product of lipid peroxidation, is increased in the brain in Alzheimer’s disease. Neurobiol Aging 19(1):33–36 Mutisya EM, Bowling AC, Beal MF (1994) Cortical cytochrome oxidase activity is reduced in Alzheimer’s disease. J Neurochem 63(6):2179–2184
123
L. B. Gatta et al. Netzer WJ, Dou F, Cai D, Veach D, Jean S, Li Y, Bornmann WG, Clarkson B, Xu H, Greengard P (2003) Gleevec inhibits betaamyloid production but not Notch cleavage. Proc Natl Acad Sci USA 100(21):12444–12449 Nunan J, Small DH (2000) Regulation of APP cleavage by alpha-, beta- and gamma-secretases. FEBS Lett 483(1):6–10 Nunomura A, Chiba S, Lippa CF, Cras P, Kalaria RN, Takeda A, Honda K, Smith MA, Perry G (2004) Neuronal RNA oxidation is a prominent feature of familial Alzheimer’s disease. Neurobiol Dis 17(1):108–113 Pratico D, Clark CM, Liun F, Rokach J, Lee VY, Trojanowski JQ (2002) Increase of brain oxidative stress in mild cognitive impairment: a possible predictor of Alzheimer disease. Arch Neurol 59(6):972–976 Reichmann H, Florke S, Hebenstreit G, Schrubar H, Riederer P (1993) Analyses of energy metabolism and mitochondrial genome in post-mortem brain from patients with Alzheimer’s disease. J Neurol 240(6):377–380 Richter C, Gogvadze V, Laffranchi R, Schlapbach R, Schweizer M, Suter M, Walter P, Yaffee M (1995) Oxidants in mitochondria: from physiology to diseases. Biochim Biophys Acta 1271(1):67– 74 Scheuermann S, Hambsch B, Hesse L, Stumm J, Schmidt C, Beher D, Bayer TA, Beyreuther K, Multhaup G (2001) Homodimerization of amyloid precursor protein and its implication in the amyloidogenic pathway of Alzheimer’s disease. J Biol Chem 276(36):33923–33929 Skovronsky DM, Moore DB, Milla ME, Doms RW, Lee VM (2000) Protein kinase C-dependent alpha-secretase competes with betasecretase for cleavage of amyloid-beta precursor protein in the trans-golgi network. J Biol Chem 275(4):2568–2575 Smith CD, Carney JM, Starke-Reed PE, Oliver CN, Stadtman ER, Floyd RA, Markesbery WR (1991) Excess brain protein oxidation and enzyme dysfunction in normal aging and in Alzheimer disease. Proc Natl Acad Sci USA 88(23):10540– 10543 Sohal RS (1993) Aging, cytochrome oxidase activity, and hydrogen peroxide release by mitochondria. Free Radic Biol Med 14(6):583–588 Sultana R, Poon HF, Cai J, Pierce WM, Merchant M, Klein JB, Markesbery WR, Butterfield DA (2006) Identification of nitrated proteins in Alzheimer’s disease brain using a redox proteomics approach. Neurobiol Dis 22(1):76–87 Tamagno E, Bardini P, Obbili A, Vitali A, Borghi R, Zaccheo D, Pronzato MA, Danni O, Smith MA, Perry G, Tabaton M (2002) Oxidative stress increases expression and activity of BACE in NT2 neurons. Neurobiol Dis 10(3):279–288 Tong Y, Zhou W, Fung V, Christensen MA, Qing H, Sun X, Song W (2005) Oxidative stress potentiates BACE1 gene expression and Abeta generation. J Neural Transm 112(3):455–469 Velliquette RA, O’Connor T, Vassar R (2005) Energy inhibition elevates beta-secretase levels and activity and is potentially amyloidogenic in APP transgenic mice: possible early events in Alzheimer’s disease pathogenesis. J Neurosci 25(47):10874– 10883 Walsh DM, Klyubin I, Fadeeva JV, Rowan MJ, Selkoe DJ (2002) Amyloid-beta oligomers: their production, toxicity and therapeutic inhibition. Biochem Soc Trans 30(4):552–557 Wielburski A, Nelson BD (1994) Heme a induces assembly of rat liver cytochrome c oxidase subunits I-III in isolated mitochondria. FEBS Lett 177(2):291–294