The Amyloid-β Peptide Suppresses Transforming Growth Factor-β1-induced Matrix Metalloproteinase-2 Production via Smad7 Expression in Human Monocytic THP-1 Cells*

Accumulation of the amyloid-β (Aβ) peptide in the brain is a crucial factor in the development of Alzheimer disease. Expression of transforming growth factor-β1 (TGF-β1), an immunosuppressive cytokine, has been associated in vivo with Aβ accumulation in transgenic mice and recently with Aβ clearance by activated microglia, suggesting its deleterious and beneficial effects in neuronal cells. In this study, we demonstrated that TGF-β1 stimulated the production of matrix metalloproteinase-2 (MMP-2) in a time- and dose-dependent manner in a human monocytic THP-1 cell line. Notably, we found that Aβ1–42 consistently inhibited the TGF-β1-induced production of MMP-2, the endogenous gene containing Smad response elements, whereas the reverse peptide, Aβ42–1, evidenced little effect. Additionally, Aβ1–42 reduced TGF-β1-induced increase in plasminogen activator inhibitor-1 (PAI-1). This inhibitory effect of Aβ1–42 was also seen in human astroglial T98G cell line. Furthermore, Aβ1–42 significantly induced the expression of Smad7, which appears in turn to mediate the Aβ suppression of the TGF-β1-induced MMP-2 production. Indeed, Smad7 overexpression mimicked the inhibitory effect of Aβ1–42 on TGF-β1-induced MMP-2 production. Importantly, Aβ1–42 markedly suppressed the transactivation of the transfected reporter construct, p3TP-Lux, which contains TGF-β1-inducible Smad response elements. This was concomitant with a decreased MMP-2 production in TGF-β1-treated cells. Inhibition of cellular Smad7 levels via the small interference RNA method significantly ameliorated the Aβ1–42-mediated suppression of TGF-β1-inducible transcription reporter activity, thereby restoring MMP-2 induction, whereas Smad7 transfection down-regulated TGF-β1-inducible transcription reporter activity. Collectively, these data suggest that Aβ1–42 may play an important role in the negative regulation of TGF-β1-induced MMP-2 production via Smad7 expression.

Alzheimer disease (AD) 1 is a progressive neurodegenerative disorder, which is characterized by the loss of higher cognitive functions. The increased production of amyloid-␤ (A␤) peptide, a 39 -42 residue of the proteolytic product of amyloid precursor protein, and fibrillar A␤ deposition in dense senile plaques have been correlated with the progression of cognitive dysfunction in AD (1). Immunoreactivity to numerous inflammatory mediators has been detected in sections of AD brains including pro-inflammatory cytokines, acute phase proteins, and several proteins involved in the classical complement pathway (2,3). In AD brains, the co-localization of a broad variety of these inflammation-related proteins and clusters of reactive microglia and astrocytes with A␤ deposits is consistent with the notion that chronic inflammation plays a significant role in the progression to enhanced neurodegeneration (2,4). However, the precise role of these molecules in the neuropathology of AD has yet to be clarified.
Transforming growth factor-␤1 (TGF-␤1), a potent immunosuppressive cytokine found in neurite plaques, has been implicated recently as a cofactor for AD progression, largely due to its ability to promote perivascular inflammation and/or amyloid deposition (5,6). Postmortem brain tissue analyses of AD patients show increased TGF-␤1 expression, which can be closely correlated with the degree of cerebral amyloid angiopathy (CAA) (7), a major pathological feature of AD and related disorders (8). A genetic polymorphism of the TGF-␤1 gene may be associated with a higher risk of developing AD (9). In contrast to its amyloidogenic effect, TGF-␤1 may also exert a more complex role since TGF-␤1 facilitates increased A␤ clearance from the brain parenchyma to the cerebral blood vasculature following the activation of parenchymal microglial cells and also induces plaque burden reduction in TGF-␤1/human amyloid precursor protein bigenic mouse brains compared with human amyloid precursor protein mice (10). The TGF-␤s have also been shown to protect neuronal cell cultures from A␤-and glutamate-induced neurotoxicity by up-regulating either antiapoptotic (Bcl-2, Bcl-XL) or calcium-stabilizing factors (calbindin) (11,12). Thus, in the A␤ function of neuronal cells, TGF-␤s apparently exert both deleterious and beneficial effects via regional and/or cell type-specific TGF-␤ isoforms and receptor expression (7, 10 -13).
The TGF-␤ receptors are transmembrane serine/threonine kinases that propagate signals downstream (14 -16). TGF-␤ binding induces the phosphorylation of the TGF-␤ type I receptor by the TGF-␤ type II receptor, the essential process in TGF-␤ signaling. Smads associate with these activated TGF-␤ receptors and play a crucial role in TGF-␤ signal transduction. Smad2 and Smad3 are direct substrates of the TGF-␤ receptor kinase, and they interact with a common partner, Smad4. Smad4-containing heterometric Smad complexes then translocate from the cytoplasm into the nucleus where they function as transcriptional regulators. Thus, the activation of gene transcription via Smad protein binding to Smad-binding elements (SBEs) represents terminal event in TGF-␤ signaling (17,18). In contrast to the receptor-activated Smads, inhibitory Smad7 binds stably to TGF-␤ receptors and interferes with the ligandinduced phosphorylation of both Smad2 and Smad3 (14 -16). Importantly, Smad7 is considered to function as a negative regulator of the TGF-␤/Smad signaling cascade and abnormal Smad7 expression is implicated in human disease (19,20).
Increased levels of matrix metalloproteinases (MMPs) have been observed in the plasma and the proximity of extracellular amyloid plaques in the brain tissues from the AD patients (21,22). Of important interest, pathogenic A␤ (Dutch variant A␤), possibly associated with CAA, was shown to stimulate the expression and activation of MMP-2. Thus, these activities may contribute to the loss of vessel wall integrity in CAA, resulting in hemorrhagic stroke (23). In contrast, earlier studies showed that MMPs are involved in A␤ degradation (24,25). These findings together suggest a critical role of MMPs in the pathophysiology of AD.
Despite extensive studies on the dual effects of TGF-␤s in amyloid plaque metabolism, the role of A␤ in the modulation of TGF-␤ function has remained enigmatic. Because dysregulation of TGF-␤ function is also implicated in chronic inflammation, leading to enhanced neurodegeneration, it is important to uncover any synergistic or antagonistic modulations of TGF-␤mediated effects elicited by A␤ to generate effective therapeutic strategies involving TGF-␤ and A␤ in AD pathology. Furthermore, few studies try to understand how TGF-␤1 affects MMP production and how A␤ may modulate this process at the molecular level.
In this regard, we sought to determine the regulatory effects of A␤ on the TGF-␤-induced stimulation of MMP-2 and plasminogen activator inhibitor-1 (PAI-1), endogenous genes containing SBEs (26,27), in human monocytic THP-1 cells as a model for microglia and in astroglial T98G cells. We also assessed the effects of A␤1-42 on the transcriptional activation of a transfected reporter gene containing TGF-␤-inducible SBEs and also on Smad7 expression. We then attempted to ascertain in what way Smad7 expression was related to TGF-␤-induced effect on MMP-2. To this end, we analyzed the effects of Smad7 overexpression on TGF-␤1-mediated MMP-2 production and TGF-␤1-induced transcription reporter activity and looked at decreases in cellular Smad7 levels via the siRNA method on the A␤ suppression of TGF-␤1-inducible effects. Our data constitute evidence that TGF-␤-mediated MMP-2 production primarily occurs through the activation of the Smad pathway. This study posits a novel role for A␤ as a negative regulator of TGF-␤-mediated MMP-2 production via Smad7 expression.
A␤ Peptide Preparation-A␤1-42 and A␤42-1 were purchased from U. S. Peptide (Fullerton, CA), and stock solutions were prepared as described previously with some modifications (28). A␤ peptides were dissolved at 5 mM in dimethyl sulfoxide (Me 2 SO) to be diluted next at 250 M in double-distilled water before experiments. This preparation contains predominantly monomer and dimers with larger oligomers up to 6-mers (data not shown). These small A␤ oligomers, similar to small SDS-stable A␤ oligomers (29) isolated from AD brains (30), are the neurotoxic species, which are better correlated with synaptic loss in AD than plaques (31,32). A␤ peptides, if used in the indicated concentrations in the presence of TGF-␤1, did not significantly affect the viability of THP-1 cells when LDH release was measured. The effects of Me 2 SO alone in the used range of concentrations were not detectable.
Cell Culture and Treatment for Preparation of Conditioned Media-THP-1 (a mononuclear cell line of human origin) and T98G (the human astroglial cell line) were obtained from ATCC (Manassas, VA) and cultured in RPMI 1640 medium (Sigma). THP-1 has been widely used as a model of human monocytes/macrophages or microglia, not only because of its functional and morphological similarities including its capacity to perform signal transduction pathways but also because of functional differences in the metabolisms of rodent and human microglial cells, as was described previously (33). Confluent cells suspended in serum-free RPMI 1640 media supplemented with glucose (0.5%) were seeded into 96-well culture plates (ϳ1 ϫ 10 5 /well) and incubated for 2 h at 37°C before experimental manipulation. The cells were then treated with TGF-␤1 and/or A␤ peptides as indicated. To determine the effects of specific inhibition of TGF-␤1-induced responses, cells were pretreated for 30 min with various chemical inhibitors at the concentrations specified followed by incubation with TGF-␤1 and/or A␤ peptides. After incubation for the indicated periods, conditioned media were collected for subsequent analysis by gelatin zymography and cells were lysed for Western blot. To confirm TGF-␤1-mediated MMP-2 production, TGF-␤1 was preincubated with anti-TGF-␤1-neutralizing antibody, anti-TNF-␣-neutralizing antibody, or preimmune IgG for 30 min at room temperature at the indicated concentrations before the addition to the cell cultures. None of the drugs at the concentrations used affected cell viability.
Analyses of Gelatinolytic Zymography-The gelatinolytic activities in the cell-free supernatants, normalized for equal amounts of protein, were determined by zymography with gelatin according to previously published methods (33). Clear bands on the zymograms were quantitated by densitometric scanning to determine the intensity of MMP-2 activity. The arbitrary densitometric units were expressed or converted to a fold measurement of the response of the vehicle-treated controls for each individual experiment.
Smad7 Overexpression-THP-1 cells grown in 96-well plates were transiently transfected with Lipofectamine 2000 (Invitrogen) and with 0.2 g/well of either Smad7 expression plasmid or control pcDNA3 (19). After 24 h of transfection, cells were starved for 2 h in serum-free RPMI 1640 medium and then treated with TGF-␤1 as indicated. Western blot analysis and zymography were performed to assess Smad7 expression and MMP-2 gelatinolytic activities as described below.
Transient Transfection and Luciferase Reporter Assay-THP-1 cells grown in 96-well plates were transiently transfected with Lipofectamine 2000 and with 0.2 g/well TGF-␤-responsive luciferase reporter construct and plasmid p3TP-Lux containing TGF-␤-inducible SBEs (34,35). Also, 0.01 g/well pRL-TK-Renilla was co-transfected to normalize for variations in transfection efficiency between assays (36,37). After 4 h of transfection, the cells were treated with TGF-␤1 and/or A␤ peptides in serum-free RPMI 1640 medium for 4 h at the indicated concentrations. In some experiments, to measure the effect of Smad7 depletion, cells were transfected with either synthetic siRNA for Smad7 or nonspecific control pool for 24 h, as described below, prior to transfection with p3TP-Lux. The effect of Smad7 overexpression was also analyzed by transfecting either Smad7 expression plasmid or control pcDNA3 for 24 h, as described above, prior to transfection with p3TP-Lux. Firefly and Renilla luciferase activities were measured in either duplicate or triplicate and normalized using a dual luciferase reporter assay system (Promega, Madison, WI) as described previously (37). Luciferase activity values were normalized with respect to protein concentrations, and the results are presented as fold increases.
siRNA Studies-Synthetic siRNA for Smad7 and nonspecific control pool were purchased from Cellogenetics (Gaithersburg, MD), and transfection of the RNA oligonucleotide was performed using Lipofectamine 2000. THP-1 cells were treated with Lipofectamine 2000 (mock transfection), siRNA, or nonspecific RNA pool at the concentrations indicated. After 24 h of transfection, cells were starved for 2 h and then treated with TGF-␤1 and/or A␤ peptides as indicated. Western blot analysis and zymography were performed to measure cellular Smad7 depletion and MMP-2 gelatinolytic activities as described above.
Western Blot Analysis-Cytoplasmic fractions were prepared as described previously (38), and whole cell lysates were prepared by protein extraction, using a buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, and 0.1% SDS supplemented with protease inhibitors and phosphatase inhibitors as described above. For Western blot analysis, equal amounts of protein (ϳ30 g) were subjected to reducing SDS-PAGE. After electroblotting and blocking, blots were then probed at room temperature for 3 h with the primary antibody and incubated for 1 h with the specific secondary antibody conjugated with horseradish peroxidase. Proteins were visualized using an ECL Western blotting detection system (Amersham Biosciences). After the antibodies were stripped by incubating the membranes with stripping buffer (62.5 mM Tris-HCl, pH 6.7, 2% SDS, 100 mM 2-mercaptoethanol) at 50°C for 30 min, they were processed for re-labeling with ␤-actin monoclonal antibody (1:2000, Sigma).
Statistical Analysis-All of the values were expressed as means Ϯ S.E. Student's t tests were used for unpaired results to evaluate differences between groups. Differences in p values of Ͻ0.05 were considered to be significant.

Induction of MMP-2 by TGF-␤1-TGF-␤1
is a pleiotropic inflammatory mediator with diverse immunomodulatory properties (39). TGF-␤1 may affect monocyte chemotaxis and migration into tissues by enhancing the monocytic expression of MMPs. MMP-2 is released from the cells as inactive precursor (72 kDa of pro-MMP-2), which is activated by MT1-MMP (MMP-14) (40). To assess the potential involvement of MMP-2 in the A␤ modulation of TGF-␤1 function, we examined TGF-␤1-induced MMP-2 production in monocytic THP-1 cells because MMP-2 contains TGF-␤-inducible SBE domains in its promoter regions (17,26). MMP-2 activity was determined by gelatin zymography, and the specificity of MMP-2 protein was verified by Western blot analysis as a 68-kDa intermediateactive MMP-2 ( Fig. 1, E), both of which were recently described (41). However, 62 kDa of full-active MMP-2 and inactive MMP-2 (72 kDa of pro-MMP-2) were not seen on the zymograms. The treatment of THP-1 cells with TGF-␤1 at a concentration of 10 ng/ml induced MMP-2 expression in a time-dependent manner. MMP-2 levels began to increase at 12 h and reached peak 40 h after treatment (Fig. 1, A). This induction of MMP-2 by TGF-␤1 in THP-1 also occurred in a dose-dependent manner (Fig. 1, B). Maximal induction of MMP-2 was found to occur at a concentration of 20 ng/ml TGF-␤1. Increases in TGF-␤1 concentration beyond this level had no further effect on MMP-2 levels (data not shown). In contrast, TGF-␤1 treatment had only a minimal effect on MMP-9 activity in THP-1 cells. Similar results were obtained using TGF-␤2 and TGF-␤3. These results confirm the presence of the three functional TGF-␤ receptors (Fig. 1, C). This TGF-␤-mediated MMP-2 production was clearly specific, because only anti-TGF-␤1-neutralizing antibody was able to reverse the process (Fig. 1, D). Moreover, MMP-2 induction was consistently inhibited when de novo mRNA expression and protein synthesis were inhibited by actinomycin and cycloheximide, respectively, which indicates that TGF-␤-mediated MMP-2 induction is dependent on both transcriptional and translational activities.
MMP-2 Induction by TGF-␤1 Is Dependent on Intracellular Smad Signaling-To unravel the mechanism by which TGF-␤1 induces MMP-2 expression in THP-1 cells, we investigated the results of blocking various signaling pathways of TGF-␤1 on MMP-2 expression since TGF-␤1 has been shown to activate several distinctive signaling pathways, such as Smad2, p38 MAPK, and Akt (also known as protein kinase B) (42,43). However, LY294002 and worthmanin, inhibitors of phosphatidylinositol 3-kinase, an activator of Akt, evidenced only minimal effects on TGF-␤1-mediated MMP-2 induction (Fig. 2, A). Likewise, SB202190, a specific inhibitor of p38 MAPK, and PD98059, known to selectively block the activity of MAPK kinase kinase (MEK1), an activator of ERKs, as well as an inhibitor of JNK MAPK (JNKI), also failed to abolish TGF-␤1induced MMP-2 production. On the evidence presented in pre- vious papers (33,37,38,44) where these inhibitors were used, the concentrations employed in the experiment should have been enough to inhibit their respective kinases without significantly affecting cell viability. To further delineate the potential involvement of Smad signaling in MMP-2 induction, we inhibited Smad signaling by overexpressing inhibitory Smad7. Because Smad7 competes with Smad-2 and Smad-3 to bind to the activated TGF-␤ type I receptor, an overexpression of Smad7 should result in the attenuation of Smad2/3 signaling (14 -16). Interestingly, Smad7 overexpression in THP-1 cells significantly inhibited TGF-␤1-induced MMP-2 expression, in contrast to its effects in vector-transfected cells (Fig. 2, B and  C). These results are consistent with the notion that, in THP-1 cells, MMP-2 induction by TGF-␤1 is primarily mediated through Smad signaling.
A␤1-42 Inhibits TGF-␤1-induced MMP-2 Production in THP-1 Cells-To date, the study on TGF-␤1 has centered on its involvement in amyloid plaque metabolism. Because dysregulation of TGF-␤1 function is also implicated in chronic inflammation, leading to neurodegeneration, it is important to uncover any synergistic or antagonistic modulation of TGF-␤1mediated effects by A␤1-42 to generate effective therapeutic strategies involving TGF-␤1 and A␤. To this end, we examined the effect of A␤1-42 on TGF-␤1-induced MMP-2 production. Interestingly, A␤1-42 treatment of THP-1 cells resulted in a decrease of TGF-␤1-induced MMP-2 activity over the level seen in TGF-␤1-treated cells, whereas the reverse peptide, A␤42-1, elicited only minimal effects under the same experimental conditions (Fig. 3). In contrast, A␤1-42 at a concentration of 10 M slightly induced MMP-9 expression in THP-1 cells, as was previously observed in our study (33). However, this A␤1-42mediated MMP-9 production was not significantly modulated by TGF-␤1. These observations suggest that A␤1-42 could act as a negative modulator of TGF-␤1-induced production of MMP-2.
A␤1-42 Induced Smad7 in THP-1 Cells-Given the critical role of Smad7 in negatively modulating TGF-␤1 effects on MMP-2 (as described above), we reasoned that A␤1-42 might be inducing Smad7 expression, thereby exerting its effects. Therefore, A␤1-42 would have a detectable effect on the cellular levels of Smad7, acting as a negative regulator of the TGF- ␤/Smad signaling cascade. As shown in Fig. 6, A␤1-42 consistently induced the cytoplasmic expression of Smad7 in both TGF-␤1-treated and TGF-␤1-untreated cells, whereas the reverse peptide had no significant effect (Fig. 6, A). Moreover, this A␤1-42-mediated Smad7 induction was dose-dependent and inhibited when de novo mRNA expression and protein synthesis were inhibited, indicating that A␤1-42-mediated Smad7 production is dependent on both transcriptional and translational activities (Fig. 6, B). Together, these results indicate that the molecular mechanism underlying A␤1-42 suppression of TGF-␤1-mediated effect on MMP-2 production in THP-1 cells appears to involve Smad7 expression.

A␤1-42 Inhibits Transcriptional Activation of TGF-␤1-inducible SBE-containing Promoters in THP-1 Cells-Given that
SBEs are present in the promoter regions of MMP-2 (26), we further examined the effects of A␤1-42 on the well characterized TGF-␤1-inducible SBE-containing promoter-reporter construct, p3TP-Lux (34,35), which was utilized in THP-1 cells to examine the induction of transcriptional activity by TGF-␤1. As shown in Fig. 7, TGF-␤1 effectively stimulated the transcriptional activation of luciferase expression. TGF-␤1 treatment of transfected THP-1 cells elicited statistically significant inductions of SBE-directed luciferase activity, which were ϳ4-fold greater than the activity detected in the control cell cultures. Furthermore, A␤1-42 inhibited the transcriptional activation of TGF-␤1-inducible luciferase reporter activity by 2.2-fold, whereas the reverse peptide, A␤42-1, exhibited little, if any, influence. Importantly, Smad7 transfection mimicked the inhibitory effect of A␤1-42 on TGF-␤1-inducible luciferase reporter activity. These results confirm a direct correlation between A␤1-42-mediated Smad7 induction and the reduction of TGF-␤1-inducible transcriptional activation.
Depletion of Smad7 with siRNA Reversed the Inhibitory Effects of A␤1-42 on TGF-␤1-induced MMP-2 Production-To further verify the inhibitory effect of A␤1-42 on TGF-␤1-induced MMP-2 activity via Smad7 induction, Smad7 depletion was performed using the siRNA method. Inhibition of Smad7 expression with Smad7-siRNA (Fig. 8, A) significantly reversed A␤1-42-mediated suppression of the TGF-␤1-induced MMP-2 production in a dose-dependent manner. The maximal reversal effect occurred at a final RNA concentration of 100 nM (Fig. 8, B). On the other hand, little effect was observed in the nonspecific control siRNA-transfected THP-1 cells. Furthermore, consistent with the rescue of the TGF-␤1-mediated MMP-2 activity, the A␤1-42 suppression of TGF-␤1-inducible transcriptional reporter activity was restored in the Smad7-siRNA-treated cells (Fig. 8, C). These results clearly confirm that the inhibitory effect of A␤1-42 on TGF-␤1-induced MMP-2 production is mediated primarily via Smad7 expression. DISCUSSION The results presented in this paper verify a critical role of A␤1-42 in suppressing TGF-␤1-mediated MMP-2 induction via Smad7 expression in human monocytic THP-1 cells. Several lines of observation support this conclusion. First, overexpression of Smad7 significantly decreased TGF-␤1-induced monocytic MMP-2 production in a time-and dose-dependent manner, whereas pharmacological inhibition of different signal transduction pathways resulted in only minimal effects. Second, A␤1-42 treatment inhibited TGF-␤1-induced MMP-2 production, concomitant with increased Smad7 expression. Third, A␤1-42 markedly suppressed the TGF-␤-inducible transactivation of transfected reporter constructs containing SBEs in TGF-␤1-treated cells. Fourth, Smad7 overexpression mimicked the inhibitory effect of A␤1-42 on TGF-␤1-inducible luciferase reporter activity. Finally, the reduction of cellular levels of Smad7 via siRNA treatment reversed the A␤1-42-mediated suppression of TGF-␤-inducible transactivation reporter activity and TGF-␤1-mediated MMP-2 production. Overall, these data point to the presence of negative modulation of functional TGF-␤1-induced MMP-2 production by A␤1-42 via Smad7 expression in human monocytic THP-1 cells. This antagonistic regulation of TGF-␤1 by A␤1-42 is further supported by our additional findings that A␤1-42 could also reduce TGF-␤1mediated increase of PAI-1, another endogenous gene containing SBEs, and that these inhibitory effects of A␤1-42 were also observed in human astroglial T98G cells.
Earlier examination of the promoter sequences of MMP-2 (gelatinase A) and MMP-9 (gelatinase B) (26,45) revealed the presence of a putative SBE, which corresponds to the CAGAC consensus site reported by Dennler's group (17). Furthermore, a previous study (46) also reported that both MMP-2 and MMP-9 were transcriptionally regulated by TGF-␤ through direct promoter activation. Results from the present study clearly demonstrate that transcription of the MMP-2 gene containing the three putative SBEs can be functionally activated by all three TGF-␤ isoforms, thus confirming both the functionality of the Smad-mediated TGF-␤ signaling pathway and the presence of the three TGF-␤ receptors in human monocytic THP-1 cells. This Smad-mediated MMP-2 production, in response to TGF-␤1, is in good agreement with recent studies (47). However, the minimal effect on monocytic MMP-9 activity exhibited by TGF-␤s in this study is not in good agreement with published reports (47)(48)(49)(50) in which contrasting results are reported. Treatment with TGF-␤s has been demonstrated to result in either down-regulation or up-regulation of MMP-9 production, possibly due to differences in cell types and/or experimental conditions.
On the other hand, PAI-1 is a serpin that suppresses fibrinolysis by inhibiting the activity of plasminogen activator (PA). Together with PA, PAI-1 is expressed in the central nervous system and may play a role in the regulation of PA activity. Given that PAI-1 has physiological functions for the survival of neurons other than its role as PA inhibitor (51-53), decreased PAI-1 production from astrocytes due to the negative regula- tion of TGF-␤1-mediated effects by A␤1-42, as seen in our study, might be detrimental to neurons and accelerate A␤induced neurodegeneration. It is also interesting to note that PA was shown to inhibit A␤-induced neurotoxicity through rapid clearance of A␤ (54).
Interestingly, our results showed that overexpression of inhibitory Smad7 significantly decreased TGF-␤1-induced MMP-2 production, suggesting that TGF-␤-mediated Smad signaling is primarily responsible for MMP-2 production in THP-1 cells. In fact, the Smad pathway is not the sole one that transduces the TGF-␤ signal and, hence, differential activation of the varied TGF-␤ signaling pathways has been proposed as a mechanism for selectivity of TGF-␤ action in different cell types (42,43). However, pharmacological inhibition of different signal transduction pathways, as shown in our study, did not significantly affect TGF-␤1-induced MMP-2 production. This finding is in good contrast to MMP-9 production, which occurs via multiple integrated signaling pathways within the cell, as described previously (33). Furthermore, depletion of Smad7 with siRNA enhanced TGF-␤1-inducible luciferase reporter activity; concomitant with increased MMP-2 activity in TGF-␤1treated cells. Thus, our observations strongly support the notion that TGF-␤-mediated MMP-2 production primarily occurs through the activation of the Smad pathway, which is in turn inhibited by the overexpression of Smad7, a negative regulator of the TGF-␤/Smad signaling cascade.
To our knowledge, this study constitutes the first report that A␤1-42 and TGF-␤1 have opposite effects on MMP-2 production. Smad7-siRNA data from the present study support the  notion that the negative modulation by A␤1-42 of a functional Smad-mediated TGF-␤ signaling system primarily involves Smad7 induction. This conclusion is further substantiated by our observation that Smad7 overexpression mimicked the inhibitory effect of A␤1-42 on TGF-␤1-induced MMP-2 production. Furthermore, A␤1-42 consistently suppressed the TGF-␤1-mediated activation of p3TP-Lux, a well characterized TGF-␤-responsive construct, which contains SBEs capable of conferring TGF-␤ responsiveness on a luciferase reporter. This construct has been used extensively to characterize transcriptional responsiveness to both TGF-␤1 and Smads in a variety of primary cells in culture and established cell lines (17,34,47). We observed that Smad7 depletion by the siRNA method significantly ameliorated or reversed the A␤1-42-mediated suppression of both TGF-␤1-inducible luciferase reporter activity and TGF-␤1-mediated MMP-2 induction. Conversely, Smad7 overexpression mimicked the inhibitory effect of A␤1-42 on TGF-␤1-inducible luciferase reporter activity. These observations further confirmed the inhibitory effect of A␤1-42 on TGF-␤1-mediated MMP-2 production via Smad7 expression, which results in the inhibition of Smad2/3 phosphorylation and Smad complex nuclear translocation (14 -16). Further studies are necessary regarding which receptor(s) and/or signal transduction pathways are involved in A␤-mediated Smad7 expression. This is because the Smad signaling system through crosstalk with other signal transduction pathways appears to be an important integrator of multiple signaling pathways within the cell (42,43).
Although the precise role of TGF-␤ in the neuropathology of AD remains unknown, earlier studies (7,12) have detected TGF-␤1 immunoreactivity in senile plaques and neurofibrillary tangle-bearing neurons in the brains of AD patients. Elevated TGF-␤1 levels were found in the cerebrospinal fluid and serum of patients with AD than in control patients without dementia (55), and the expression of TGF-␤ receptors I and II was increased in the reactive glia of AD brains (56). However, a more recent study has demonstrated that decreased TGF-␤1 plasma levels are a potential biomarker for AD (57). Moreover, TGF-␤ has recently been shown to play a protective role, creating an environment favorable for the survival of the cell in the face of death-inducing insults including A␤, hypoxia/ischemia, glutamate excitotoxicity, oxidative damage, and the human immunodeficiency virus (58). Consistent with this notion, chronic neuroinflammation could represent the result of a failure in the anti-inflammatory mechanism, in which a significant reduction in the biological function of immunosuppressive cytokines, such as TGF-␤1, results in the defective suppression of the inflammatory process and the unrestrained production of proinflammatory cytokines such as TNF-␣, thus leading to enhanced neurodegeneration. In this context, the aberrant antagonistic activities of A␤1-42 against the physiological actions of TGF-␤ as reported in this study would be detrimental and may even play an important role in the pathogenesis of AD. Nevertheless, recent studies suggest the possibility of an indirect neurotoxic role for TGF-␤ in terms of its promoting effects with regard to perivascular inflammation and/or amyloid deposition (5,6,59). These deleterious effects when bolstered by a genetic polymorphism of the TGF-␤1 gene may be associated with higher risks of AD (9). Other possible synergistic risk factors include a direct receptor-independent interaction between TGF-␤ and A␤, which enhances A␤ oligomerization, leading to potentiation of the neurotoxic effects of A␤ (60), although an earlier study (61) proposes the TGF-␤ receptor as one possible cellular A␤ interaction site. Maximal inhibitory effect resulting from A␤1-42 pretreatment, as shown in our study, suggested that the functional antagonism between A␤1-42 and TGF-␤1 appears to involve a direct activation of signaling cascade downstream of A␤ receptors. However, the possibility of simple sequestration of TGF-␤1 due to a direct interaction between TGF-␤1 and A␤1-42 cannot be completely ruled out. Further studies with a dominant active TGF-␤ receptor will clarify this point.
Contrasting results regarding the involvement of MMPs in AD are also present in the literature. Previous studies indicated that MMPs are involved in A␤ degradation (24, 25) and FIG. 8. Smad7-siRNA reversed the inhibitory effect of A␤1-42 on TGF-␤1-enhanced MMP-2 production and TGF-␤1-mediated reporter activity. THP-1 cells were transfected with increasing concentrations of Smad7-small interference RNA or the control-small interference RNA (c-siRNA) pool for 24 h and then were treated with TGF-␤1 (10 ng/ml) in the presence or absence of A␤1-42 (5 M) for 4 h. Cytoplasmic fractions were then immunoblotted as described in Fig. 2. In A, inhibition of Smad7 expression was observed with Smad7-siRNA transfection (Fig. 6, A). Relative levels of Smad7 expression, normalized to ␤-actin, are shown on top of the panel. In parallel, transfected cells as described in A were incubated for 30 h for purposes of zymographic analysis in B. Relative levels of MMP-2 release were shown on top of the panel. Only Smad7-siRNA transfection restored TGF-␤1-mediated MMP-2 induction in a dose-dependent manner. Results are representative of five independent experiments. In C, Smad7-siRNA or c-siRNAtransfected cells were further transfected with the TGF-␤-responsive p3TP-Lux and with pRL-TK-Renilla for 4 h and then exposed to TGF-␤1 (10 ng/ml) in the presence or absence of A␤1-42 (5 M) or A␤42-1 (5 M) for 4 h. Firefly and Renilla luciferase activities were measured as described in Fig. 5. Only Smad7-siRNA transfection restored TGF-␤1mediated reporter activity of THP-1 cells in the presence of A␤1-42. The data are expressed as the means Ϯ S.E. *, p Ͻ 0.05, Smad7-siRNAtransfected cells in response to A␤1-42 plus TGF-␤1 versus c-siRNAtransfected cells in response to A␤1-42 plus TGF-␤1 or mock-transfected cells in response to A␤1-42 plus TGF-␤1 (n ϭ 5). that increased MMP expression levels, in particular MMP-9, have been observed in plasma and in proximity to extracellular amyloid plaques in brain tissues from AD patients, whereas plasma levels of MMP-2 were found to be unchanged (21,22). Recent study has shown a functional polymorphism in the gene for MMP-9 (62), which supports a critical role for MMP-9 in the pathophysiology of AD. Likewise, a recent study has also demonstrated that Dutch variant A␤, possibly associated with CAA, a major pathological feature of AD and related disorders (8), stimulates the expression and activation of MMP-2. This study also asserts that this process may contribute to loss of vessel wall integrity in CAA, resulting in hemorrhagic stroke (23). Taken together, these findings suggest that MMPs may influence the stability of extracellular matrix or other MMP substrates and thus may play a role in the neurotrophic/neurotoxic events associated with AD.
In conclusion, as both glial cells and monocytes have been observed to express TGF-␤ and to respond to TGF-␤ stimulation, our findings support the notion that specific concentrations and combinations of A␤1-42 and TGF-␤1, encountered during the immune response at inflammatory foci, provide a balanced system for regulating MMP-2-mediated activities such as extracellular matrix degradation and PAI-1 activities closely associated with neuroprotection. Therefore, dysfunctions in either TGF-␤1 or A␤1-42 might result in improper MMP-2 and PAI-1 activities, which could result in chronic inflammation followed by secondary neuronal cell death leading to AD. Overall, our results prompted us to propose a molecular mechanism by which A␤1-42 could negatively modulate TGF-␤-mediated MMP-2 production via Smad7 expression as depicted in Fig. 9. However, the in vivo relevance of this TGF-␤1 inhibitory role of A␤1-42 in AD pathology remains unknown.