|Home | About | Journals | Submit | Contact Us | Français|
The overproduction and extracellular buildup of amyloid-β peptide (Aβ) is a critical step in the etiology of Alzheimer’s disease. Recent data suggest that intracellular trafficking is of central importance in the production of Aβ. Here we use a neuronal cell line to examine two structurally similar clathrin assembly proteins, AP180 and CALM. We show that RNA interference-mediated knockdown of AP180 reduces the generation of Aβ 1-40 and Aβ 1-42, whereas CALM knockdown has no effect on Aβ generation. Thus AP180 is among the traffic controllers that oversee and regulate amyloid precursor protein processing pathways. Our results also suggest that AP180 and CALM, while similar in their domain structures and biochemical properties, are in fact dedicated to separate trafficking pathways in neurons.
Extracellular accumulation and aggregation of amyloid-β peptide (Aβ) plays a pivotal in the pathogenesis of Alzheimer’s disease (AD) . Within the brain, Aβ is mostly produced by neurons and is derived from the amyloid precursor protein (APP). APP can undergo two different cleavage processes, non-amyloidogenic processing and amyloidogenic processing; the latter route leads to overproduction and extracellular accumulation of Aβ [3,4]. Recent advances in understanding APP processing have led to the realization that many of the molecules that play fundamental roles in intracellular trafficking also function to direct APP away or towards the amyloidogenic processing pathway [5–10]. We therefore examined the possible roles of AP180 and CALM (clathrin assembly lymphoid myeloid protein) – two regulators of clathrin-mediated trafficking – in Aβ production.
AP180 and CALM belong to clathrin assembly proteins because of their widely recognized role in promoting the assembly of clathrin-coated vesicles [11–16]. AP180 and CALM share a remarkably similar domain structure  and both are present in neurons [13, 17, 18]. However, we previously found that AP180 and CALM carry out different functions in neurons , raising the possibility that AP180 and CALM might be involved in different populations of clathrin-coated vesicles and thus regulate different clathrin-mediated pathways. Consistent with this idea, we have found in this study that AP180 but not CALM influences the level of extracellular Aβ.
Human neuroblastoma SH-SY5Y cells expressing human Swedish mutant APP695 cDNA were obtained from Dr. Wataru Araki (National Institute for Longevity Sciences) . The pSuper-AP180shRNA has been described in detail in a previous study . The pSuper-CALMshRNA has also been previously characterized and described [19, 21]. Cells were transfected using Lipofectamine following the manufacturer’s (Invitrogen) protocol and analyzed 3–4 days after transfection.
The Aβ 1-40 and 1-42 in the media of transfected cells were measured using enzyme-linked immunosorbent assay (ELISA) kits (Biosource International Inc.). For each experiment, samples were assayed in triplicate and all experiments were repeated at least three times. To confirm changes in Aβ 1-40, we also used a non-commercial ELISA protocol . The data from both ELISA protocols were similar.
The following antibodies were used: mouse monoclonal anti-AP180 (clone AP180-I; Sigma), goat polyclonal anti-CALM (sc5395 and sc6433; Santa Cruz Biotechnology), mouse monoclonal anti-APP N-terminus (clone 22C11; Chemicon/Millipore), and rabbit polyclonal anti-APP C-terminus (IBL Co., LTD, Japan). For immunoblotting, cell lysates were separated by SDS-PAGE and transferred to nitrocellulose membranes. Blots were incubated with primary antibodies followed by appropriate HRP-conjugated secondary antibodies, and visualized using ECL chemiluminescence. For immunolabeling, cells were fixed with 4% paraformaldehyde and 4% sucrose, and permeabilized with 0.1% Triton X-100. Cells were then incubated with primary antibodies followed by appropriate fluorescently tagged secondary antibodies.
The goal of this study was to determine whether AP180 and CALM have effects on Aβ production. We chose to use the neuroblastoma SH-SY5Y cells expressing the AD-associated Swedish mutant APP  for several reasons. First, these cells have been used as a simple system for studying APP processing and Aβ production [8, 20]. Second, because SH-SY5Y cells are neural, we were able to compare the neuron-specific AP180 and the ubiquitously expressed CALM. Third, higher transfection efficiency in cell lines compared with primary cultured neurons provides a practical means for biochemical assays.
To suppress the expression of AP180 or CALM, we transfected the SH-SY5Y cells with AP180 shRNA or CALM shRNA. The specificity and efficacy of these shRNAs in reducing the level of AP180 or CALM in the SH-SY5Y cells were analyzed by immunoblotting and immunolabeling. AP180 shRNA was originally designed to silence the rat AP180 gene and has proven to be highly effective in the knockdown of AP180 in rat neurons . However, the shRNA targeting region of rat and human AP180 differs in two nucleotides (Figure 1A). The difference could potentially render AP180 shRNA ineffective in human SH-SY5Y cells, as gene silencing by RNAi is known to be specific [23, 24]. To address this question, we examined the cells after they had been transfected with AP180 shRNA for 3–4 days. Immunoblotting of the cell lysates showed that the level of AP180 in the AP180 shRNA-transfected cells was significantly lower (<50%) compared to those transfected with the control vector (Figure 1B, upper panel). The level of AP180 in the CALM shRNA-transfected cells, however, was not reduced (Figure 1B, also upper panel), suggesting the specificity of the AP180 shRNA. To confirm the immunoblotting observation and to determine if the residual AP180 was derived from non-transfected cells, we co-transfected the cells with EGFP to mark transfected cells and carried out immunofluoresence labeling. While not all cells were transfected, those EGFP-expressing transfected cells were devoid of AP180 labeling (Figure 1C).
CALM shRNA was also designed to silence rat CALM . Unlike AP180 shRNA, the sequence within the CALM shRNA-targeting region is identical between rats and human, as well as between the long- and short-splice variants (see Materials and Methods). This CALM shRNA has been characterized in rat neurons  and in human embryonic kidney (HEK) 293 cells . As expected, the shRNA suppressed the expression of CALM, as determined by immunoblotting (Figure 1B, middle panel) and immunolabeling (Figure 1D). Thus, AP180 shRNA and CALM shRNA can specifically and effectively suppress the expression of AP180 and CALM in the SH-ST5Y cells.
To determine if suppressing AP180 or CALM would influence Aβ production, we measured Aβ1-40 and Aβ1-42 levels in the media of the shRNA transfected cells using ELISA. We observed that the AP180-knockdown cells produced significantly less Aβ1-40 (63% ± 10.8) compared to the control cells (Figure 2A). The level of Aβ1-40 from the CALM-knockdown cells was slightly reduced (88.8% ± 6.2), but the reduction was not statistically significant (Figure 2A). The AP180-knockdown cells also produced less Aβ1-42 (78.2% ± 1.3) (Figure 2A). In contrast, the level of Aβ1-42 in the CALM-knockdown cells remained unchanged (100% ± 2.2) (Figure 2B). Therefore, suppressing AP180 reduces the production of both Aβ1-40 and Aβ1-42. Suppressing CALM, on the other hand, does not alter Aβ levels.
To assess whether the reduction of Aβ resulted from a decrease of total APP, we next measured APP level in the lysates of the AP180 knockdown cells using immunoblotting. We used two different APP antibodies: one reacts against APP’s N-terminus  and the other against APP’s C-terminus. We found that both antibodies revealed a major protein band at ~120 kDa and that the intensity of this band was not different between the AP180 knockdown cells and the control cells (Figure 2B). This result suggests that AP180 knockdown has no effect on the level of APP.
Our results add AP180 to the growing list of endocytic proteins that are actively involved in Aβ production by neurons. Endocytosis and subsequent intracellular sorting direct APP to subcellular compartments where the amylidogenic cleavages of APP that generate Aβ take place [2, 3]. Therefore, the production of Aβ is not only controlled by the enzymes that carry out the cleavage, but also by cell’s ability to control the traffic and the destination of APP.
Our finding also raises several questions about AP180. First, does AP180 exert the same effects on mutant APP and on the wild-type APP? In the initial course of this study, we planned to analyze two different SH-SY5Y cells, expressing either the AD-associated Swedish mutant APP or the wild-type APP. However, we found that the level of AP180 in the wild-type APP expressing cells was nearly undetectable (data not shown), precluding us from carrying out analysis on AP180 in those cells. The disappearance of AP180 could directly result from the overexpression of wild-type APP. It is, then, tempting to ask if a two-way regulatory relationship exists between APP and AP180 and if the mutation in APP alters this relationship.
Second, how can we consolidate the data described here and the previous observations from tissues of AD patients? We previously found that AP180 was significantly and consistently reduced in the postmortem brain tissues of patients suffering late-stage AD [26, 27]. The in vitro finding that AP180 suppression leads to Aβ reduction seems to conflict with the observation of low AP180 but high Aβ in vivo [26, 27]. The seeming contradiction, however, can be ascribed to the complexity of the disease. Important issues to address in future studies include at what disease stage the loss of AP180 occurs, whether its loss simply reflects the disease or contributes to the disease pathology, and how AP180 cooperates with other endocytic proteins in regulating APP trafficking and Aβ production.
Third, how are AP180 and CALM related to each other? AP180 and CALM have been regarded as homologues because of their overlapping amino acid sequences, similar domain structures, and ability to promote the assembly of clathrin-coated vesicles [11–16]. In neurons, however, AP180 and CALM do not co-localize [18, 28], and they play different roles in the growth of neurons . Our finding that AP180, but not CALM, affects Aβ production supports the notion put forward previously: that AP180 and CALM are two functionally separate clathrin assembly proteins.
This research was supported by the Intramural Research Program of the Natinal Institute on Aging of the NIH, and extramural grants (R01AG026478 and K01AG022455 to Y.M.). We also thank Ms. Chiho Hirata-Fukae for her technical assistance.