Fucosyl-GM1 expression and amyloid-β protein accumulation in PC12 cells

Gangliosides are sialic acid-containing glycolipids, which are ubiquitously expressed in all eukaryotic cells and are localized mainly in the outer leaflet of the plasma membrane. As the biological functions of gangliosides, mediation and modulation of cell–cell recognition, adhesion, and signal transduction are indicated from numerous studies (Hakomori, 1990; Yu, 1994; Yu et al., 2004; Yu and Yanagisawa, 2006). Recently, glycosphingolipids, including gangliosides with sphingomyelin and cholesterol, have been shown to form a cell surface microdomain structure, caveolae or lipid raft (also known as glycosphingolipid-enriched microdomain, detergent-resistant membrane, or detergent-insoluble membrane), which are important for signal transduction and cell adhesion (Anderson, 1998; Hakomori et al., 1998; Simons and Toomre, 2000). In addition, it has been reported that a monosialoganglioside, GM1 (see Fig. 1), plays a role in initiating aggregate formation of amyloid-β protein (Aβ) by acting as a seed in Alzheimer's disease brain (Yanagisawa et al., 1995; Hayashi et al., 2004).

The PC12 cell is a rat pheochromocytoma cell line frequently used to investigate the function of gangliosides in events leading to neuronal differentiation (Greene and Tischler, 1976). To detect gangliosides expressed in PC12 cells, cell-staining experiments with cholera toxin B subunit (Ctxb) have been frequently performed. Ctxb is a component (103 amino acid residues) of heat-labile enterotoxin produced by the bacterium Vibrio cholerae and has been widely used as a probe to detect GM1 in a number of biochemical, molecular biological, and cell biological studies because of the remarkably high affinity with GM1 (Cuatrecasas, 1973; Holmgren et al., 1973; King and van Heyningen, 1973; de Haan and Hirst, 2004). As a result of cell staining experiments, several researchers have reported the expression of GM1 in PC12 cells. This is in sharp contrast to a number of reports showing that the expression pattern of gangliosides in PC12 cells is quite different from that of normal brain gangliosides (e.g., GM1), and all of the monosialogangliosides expressed in PC12 cells are fucosylated (e.g., fucosyl-GM1; see Fig. 1; Margolis et al., 1983, 1984; Ariga et al., 1987).

In this study, we have reevaluated ganglioside expression and Ctxb reactivity in PC12 cells and the subcloned line, PC12D cells (Katoh-Semba et al., 1987), by cell staining and biochemical techniques and clarified that Ctxb reactivity in PC12 cells does not represent GM1 expression at all. In fact, much of the reactivity could be traced to interaction of Ctxb with fucosyl-GM1 and not with GM1. Furthermore, by using the method of detecting Aβ accumulation in PC12 cells with fluorescence-labeled Aβ (FL-Aβ) as established by Wakabayashi et al. (2005), we found the distinct possibility that induction of Aβ aggregate formation by acting as a seed in Alzheimer's disease brain is not limited to GM1, but occurs with rather fucosyl-GM1, at least in PC12 cells. The nomenclature for gangliosides follows the system of Svennerholm (1964).

MATERIALS AND METHODS

RESULTS

Reactivity of Ctxb in PC12 and PC12D Cells

The rat pheochromocytoma cell line PC12 (Greene and Tischler, 1976) has frequently been used as a model for investigating events leading to NGF-induced neuronal differentiation. Although it is generally thought that PC12 cells express GM1 ganglioside, a series of biochemical studies (Margolis et al., 1983, 1984; Ariga et al., 1987), indicated that the expression pattern of gangliosides in PC12 cells is quite different from the pattern in normal brain: all of the monosialogangliosides expressed in the PC12 cells are fucosylated (e.g., fucosyl-GM1). Because there has been considerable interest in the expression of GM1 in PC12 cells, we undertook a detailed study to determine whether PC12 cells indeed express GM1. We compared Ctxb reactivity and the expression pattern of gangliosides in PC12 cells and the subcloned PC12D cells. First, we confirmed the integrity of our PC12 cells by examining their neurite extension activity in response to NGF treatment. As reported in the original study (Greene and Tischler, 1976), neurite extension was found in the PC12 cells treated with NGF (Fig. 2A,B). Next, we analyzed the Ctxb reactivity in these cells by flow cytometry. Although the intensity was not high, approximately 30–50% of the PC12 cells could be recognized by Ctxb (Fig. 2C). However, as a result of TLC with the resorcinol reagent, GM1 was not detectable in either PC12 or PC12D cells, under the conditions used (left panel in Fig. 2D). On the other hand, on the TLC plate, Ctxb was found to be reactive with other gangliosides with the Rf value similar with fucosyl-GM1 (right panel in Fig. 2D). To evaluate a possibility of cross-reactivity of Ctxb with fucosyl-GM1, we performed TLC-overlay assay with antifucosyl-GM1 antibody. As shown in Figure 2E, fucosyl-GM1 was expressed in PC12 and PC12D cells. Therefore, the Ctxb reactivity in PC12 and PC12D cells obviously was due to cross-reactivity of Ctxb with fucosyl-GM1.

Accumulation of Amyloid-β Protein in PC12 Cells

Alzheimer's disease, the most common form of dementia, is a progressive degenerative disease of the brain characterized by a loss of memory and cognitive function. Biochemically, this disease is characterized by the presence of amyloid plaques and neurofibrillary tangles in the brain. The amyloid plaque, the hallmark in Alzheimer's disease, is led by the aggregate formation of a misfolded protein, Aβ. It has recently been reported that GM1 plays a role in initiating Aβ aggregate formation by acting as a seed (Yanagisawa et al., 1995; Hayashi et al., 2004). Recently, a method to detect Aβ accumulation in living PC12 cells with florescein-labeled Aβ (FL-Aβ) was established by Wakabayashi et al. (2005). As shown in Figure 2D, however, PC12 cells do not express GM1. To examine the possibility that Aβ is accumulated in PC12 cells utilizing fucosyl-GM1 as the seed, we cultured PC12 cells in the presence of FL-Aβ (1–40 or 1–42) following the method established by Wakabayashi et al. (2005). As shown in Figure 3, FL-Aβ accumulation in the PC12 cells was clearly observed, as Wakabayashi et al. (2005) had reported. Because PC12 cells do not express GM1 (Fig. 2), fucosyl-GM1, by interacting with Aβ, might be the component that served as a seed for amyloid accumulation, as does GM1 in normal brain cells.

Reactivity of Ctxb in a Subcloned Cell Line of PC12D Cells

After we developed a long-term culture of PC12D cells with multiple passages, we established a new subclone of the PC12D cells (PC12DT). PC12DT cells have a flat cell body and retain some of the phenotypes of the original PC12 cells (Fig. 4A). Interestingly, flow cytometry of the PC12DT cells showed that approximately 70% of the cells were reactive to Ctxb (Fig. 4B). The percentage of the Ctxb reactivity and the high intensity were almost equivalent to those of HNK-1, a glycoconjugate antigen expressed in the parent PC12 cells (Margolis et al., 1987). Analysis by TLC (Fig. 4C) revealed that the expression pattern of gangliosides was quite different in the PC12DT cells compared with the expression patterns of the parent PC12 and PC12D cells (Fig. 2D). In the PC12DT cells, fucosyl-GM1 disappeared, but the expression of GM1 and GD1a was up-regulated. TLC-overlay assay showed that in PC12DT cells GM1, but not fucosyl-GM1, was the component reacting with Ctxb (Fig. 4C). Consistently with this observation, in PC12DT cells, the expression of fucosyltransferases, FUT1 and FUT2, regulating fucosyl-GM1 synthesis (Iwamori and Domino, 2004) was found to be down-regulated (Fig. 4D). These results suggest that PC12 cells previously reported by other investigators to express GM1 might have already transformed from the parent PC12 cells.

DISCUSSION

In this study, we reevaluated the controversial GM1 expression in PC12 cells and found that Ctxb reactivity in PC12 cells arose from cross-reactivity with another ganglioside, fucosyl-GM1. It has been reported that Ctxb recognizes GM1 and other glycolipids, such as GD1b, fucosyl-GM1, B-GM1, GM2, GT1b, GD1a, GM3, and asialo-GM1 (Masserini et al., 1992; Angström et al., 1994; Kuziemko et al., 1996; MacKenzie et al., 1997; Fukuta et al., 1998; Lauer et al., 2002). Thus, it is not surprising that the target of Ctxb binding in PC12 cells is actually fucosyl-GM1. On the other hand, we found a GM1-expressing transformant, PC12DT cells, despite the fact that no GM1 was expressed in the original PC12 cells (Margolis et al., 1983, 1984; Ariga et al., 1987). In considering the results of our study, it appears that Ctxb reactivity in PC12 cells most likely is due to cross-reactivity to fucosyl-GM1 or perhaps to transformation of the cells to acquire the ability to express GM1 ectopically, although the latter is highly unlikely. In another study (Yanagisawa et al., 2006), we evaluated the reactivity of Ctxb in mouse primary neural precursor cells. Although most of the neural precursor cells were reactive with Ctxb in flow cytometry, we showed that the actual GM1 expression in the cells was smaller than could be chemically detected by TLC. Compared with the binding affinity of an antibody to a glycolipid antigen [e.g., the K_D value of antiasialo-GM1 antibody to asialo-GM1, 3.0 × 10^–8 (Harrison et al., 1998); the K_D value of anti-GD3 antibody to GD3, 1.77 × 10^–8 (Kaminski et al., 1999)], the binding affinity of the Ctxb to GM1 is remarkably high [the K_D value, 4.6 × 10^–12 (Kuziemko et al., 1996); 7.3 × 10^–10 (MacKenzie et al., 1997)]. Therefore, there is a distinct possibility that reactivity of Ctxb in neural precursor cells might be overestimated because of the high binding affinity and nonspecificity of Ctxb. Although Ctxb has been a useful tool in detecting GM1 on the cell surface, as shown in this and other studies, its use can lead to an overestimation of GM1 expression level as a result of its high affinity and broad cross-reactivity. Thus, to prevent inaccurate estimates of GM1 expression based solely on cell staining with Ctxb, the use of a confirming biochemical experiment, such as TLC or TLC-overlay assay, of extracted gangliosides is thought to be essential.

An unexpected observation from our present study concerns the presumed role of GM1 in initiating Aβ aggregate formation by acting as a seed in Alzheimer's disease brain (Yanagisawa et al., 1995; Hayashi et al., 2004). It has been demonstrated by surface plasmon resonance that not only GM1 but also other gangliosides, such as GQ1bα, GT1α, GQ1b, GT1b, GD3, GD1a, GD1b, LM1, GM2, GM3, and GM4, have a varying degree of affinity to Aβ (Ariga et al., 2001). Neutral glycolipids have a much lower affinity for Aβ than acidic glycolipids. The authors also emphasized the importance of a NeuAcα2–3 residue on the carbohydrate chain for binding with Aβ. The same structural component is found in fucosyl-GM1 (Fig. 1). In our present study, we detected Aβ accumulation in PC12 cells expressing fucosyl-GM1. Thus, our finding suggests the possibility that gangliosides other than GM1 (i.e., fucosyl-GM1) might also play a role in promoting amyloid formation. This study is the first to indicate a possible role of fucosyl-GM1 for providing seeding for Aβ accumulation.