The PC12 cell as model for neurosecretion

Initial experiments at undifferentiated PC12 cells

Initial studies of PC12 cells focused on the detection of zeptomole quantities of catecholamines released in a regulated manner from the soma of undifferentiated cells (Chen et al. 1994). Current transients for the oxidation of catecholamines were on the timescale expected for exocytotic release (9.3 ms half-width) and were thus similar to events observed at adrenal chromaffin cells. However, catecholamines are present in much lower levels in vesicles of PC12 cells. This is at least partially because of the smaller size of the PC12 vesicle (Schubert et al. 1980) compared with the adrenal cell vesicle (Coupland 1968). The quantity of dopamine oxidized from a single vesicle after stimulated release was first reported to be 190 zmol (114 300 molecules) for a specific passage of PC12 cells (Chen et al. 1994), although it appears to vary in different PC12 cell subcultures. For the cell line obtained from the American Type Culture Collection; Rockville, MD (original experiments), it was approximately 18 times smaller than the amount of release observed from a single adrenal cell vesicle (Wightman et al. 1991).

An example of an amperometric recording of depolarization-induced exocytosis from a single PC12 cell is shown in Figure 1a. Quantification is accomplished by integrating the charge (Q) under individual current transients, representing single vesicle release, obtained using amperometry. Using Faraday’s law: Q = nFN, where n is the number of electrons in the electrooxidation (n = 2 for catecholamines), and F is Faraday’s constant (96 485 C/mole), the number of moles (N) of oxidized catecholamines can be determined. The reported quantity of catecholamines from individual vesicles in a cell type is an average value. Histograms generated by plotting the percentage of total release events versus vesicle content in each bin demonstrated the distribution of event sizes. Figure 1b shows a typical histogram generated from exocytotic events measured from PC12 cells. The average vesicle content for this set of PC12 cells was 199 ± 14 zmol (Zerby & Ewing 1996a); however, the distribution ranges from 20 to 800 zmol (Fig. 1b). Histograms plotted in this manner give a skewed (non-Gaussian) distribution and have been modelled mathematically by Wightman et al. (1991) for adrenal cells. The most feasible explanation for this distribution is that vesicles have a range of radii with relatively uniform concentration of catecholamines (Wightman et al. 1991). Thus, if the cube root of the vesicular content (Q^1/3 or mole^1/3) is utilized for generation of the histograms, then a normal (Gaussian) distribution is obtained (Fig. 1c).

The initial experiments to examine distributions of the cube root of vesicle contents for release from PC12 cells used pooled data from many cells and have been fitted using single Gaussian functions (Finnegan et al. 1996, Zerby & Ewing 1996b, Taylor & Peers 1999). However, it is not always accurate to pool data from individual event at many cells as cell-to-cell variation in the number of events can bias pooled means to only a few cells (Colliver et al. 2000a, Westerink et al. 2000). Thus, it is usually more accurate to examine pooled means than to pool data from many events covering many cells. In addition, it has been shown that in cells that release large numbers of events, the majority of these cells contain two distinct classes of catecholamine-containing vesicles (Fig. 2). The mean vesicle contents between these two classes of vesicles varies by approximately twofold, at 141 and 293 zmol, and each makes up roughly half of the total release events observed (Westerink et al. 2000). It is therefore likely that PC12 cells contain distinct classes of vesicles, although at present it cannot be excluded that the observed results are because of the release of aggregate vesicles or compound fusion.

Latency of release following chemical stimulation

Amperometry has been used to investigate the effect of different mechanisms of stimulation on the latency of exocytosis from PC12 cells following stimulation (Zerby & Ewing 1996b). PC12 cells possess both nicotinic and muscarinic receptors that can trigger exocytosis through two different mechanisms upon activation. Application of nicotine causes opening of sodium channels, which results in sufficient depolarization of the cell membrane to open voltage-sensitive calcium channels, thereby allowing rapid influx of calcium and subsequent exocytosis (Stallcup 1979). On the other hand, the application of muscarine activates muscarinic receptors that act through intracellular second messengers to release calcium from intracellular stores triggering exocytosis (Berridge & Irvine 1984). To obtain rapid membrane depolarization of the cell, KCl can be applied to promote exocytosis by direct membrane depolarization. The average catecholamine content of the vesicles is unaltered by these different stimuli, but the latencies (time between application of the stimulant and secretion events) vary significantly (Fig. 3; Zerby & Ewing 1996b). The mean latencies for each type of stimulation have been observed to be 6 ± 1 s (105 mM K⁺); 37 ± 5 s (1 mM nicotine); and 103 ± 11 s (1 mM muscarine). The 6-s delay until release following potassium stimulation apparently represents the diffusion time from the stimulation pipette as the delay is generally <1 s when a continuous, fast superfusion system is used to apply KCl (Westerink et al. 2000). The relatively long latencies until the onset of exocytosis after nicotine and muscarine are surprising as one usually expects latencies comparable with that observed following KCl-induced membrane depolarization. The longer times might reflect a relatively low number of sodium channels on these cells or a slow step in the G-protein coupling by these receptors in PC12 cells, as well as other potential rate-limiting mechanisms.

Experiments at differentiated (NGF-treated) PC12 cell varicosities

PC12 cells can be differentiated with NGF to better resemble a neuronal phenotype. Upon treatment with NGF, PC12 cells extend processes and along these processes or neurites varicosities form (bulbous regions, 1–2 μm in diameter). Varicosities have been previously shown to contain aggregates of small vesicles (20–70 nm in diameter) (Greene & Tischler 1976). Experiments carried out on days 10–14 of NGF-treated PC12 cell cultures showed no release from the cell body, only very infrequent responses from the smooth regions of the neurites, and frequent release when the electrode was located at a varicosity (Zerby & Ewing 1996a). This response was most frequently observed at varicosities located at the intersections of several neurites. The average vesicular catecholamine content observed for exocytosis at varicosities was 178 ± 9 zmol (107 000 molecules) and was not significantly different from that observed at undifferentiated cells; however, a more narrow distribution was observed at the former (Fig. 4). This study demonstrated that functional changes occur during differentiation, specifically the relocation of the sites of exocytosis without significant alteration in the overall mean vesicle catecholamine content, although the narrow distribution of vesicle content might indicate that a tighter distribution of vesicle radii is necessary to pack them into the relatively small varicosities.

Dexamethasone increases calcium channel function and exocytosis

Dexamethasone, a potent synthetic glucocorticoid, is used to treat a wide variety of conditions as an anti-inflammatory agent or immunosuppressant. Dexamethasone has also been used to differentiate PC12 cells into neuroendocrine chromaffin-like cells. These dexamethasone-differentiated cells have been monitored with patch clamp and amperometry and it was found that treatment for 5–7 days dramatically increases quantal size, excitability and coupling between calcium channels and vesicle release sites, leading to rapid exocytosis and endocytosis (Elhamdani et al. 2000).

These findings can be extended by comparison of amperometric measurements of exocytosis in undifferentiated and dexamethasone-differentiated PC12 cells from the same research group (Westerink et al. 2000, Westerink & Vijverberg 2002a). Approximately 90% of dexamethasone-differentiated PC12 cells responded with vesicular neurotransmitter release when depolarized with KCl, compared with ~50% in undifferentiated cells. During stimulation the release frequency in dexamethasone-differentiated PC12 cells amounts to ~5 Hz, whereas it averages only ~2 Hz in undifferentiated cells. The amount of releasable vesicles in dexamethasone-differentiated PC12 is also much larger (~500 vs. ~50). Finally, vesicle content increased from ~200 to ~650 zmol following differentiation with dexamethasone. Surprisingly, the cube root of vesicle content is distributed normally in single dexamethasone-differentiated PC12 indicating a more homogenous vesicle population. These characteristics allowed dexamethasone-differentiated PC12 to become a preferred model for studying modulation of neurotransmitter release at the single-cell level in neuropharmacological and neurotoxicological studies.

SNARE proteins

Secretory vesicles become docked at the cell membrane through formation of the SNARE-complex consisting of vesicle-associated synaptobrevin and plasma membrane-associated synaptosomal-associated protein of 25 kDa (SNAP-25) and syntaxin. Obviously, disruption of the SNARE complex or cleavage of SNARE proteins prevents vesicle docking, priming and release. This was elegantly demonstrated in PC12 cells transfected with Botulinum neurotoxin C1 light chain (BoNT/C1), which cleaves syntaxin and SNAP-25. In PC12 cells expressing BoNT/C1, ATP-evoked exocytosis was inhibited almost completely as revealed using amperometry (Fisher & Burgoyne 1999). More recently, syntaxin was shown to be involved in fusion pore formation as well. The fusion pore is a channel-like structure connecting the vesicle and plasma membranes, which is formed during the initial phase of exocytosis. Leakage of catecholamines through the fusion pore results in the detection of foot signals that precede the actual exocytotic event (Chow et al. 1992, Alvarez de Toledo et al. 1993). An amperometric study showed the influence of specific proteins on the process of release via the pore. Point mutations in syntaxin, a protein thought to form transmembrane segments in the pore, reduced the flux of catecholamines through the fusion pore as well as its conductance (Han et al. 2004), underlining the importance of syntaxin in regulated exocytosis.

Synaptotagmin

The proposed Ca²⁺-sensor synaptotagmin (Syt), which exists in several isoforms, prevents the SNARE complex from further interactions until an increase in the intracellular Ca²⁺ concentration displaces Syt and catalyses membrane fusion. Using amperometry, it was shown that silencing Syt-I in PC12 cells significantly reduced evoked exocytosis. The reduction in exocytosis was reversed following rescue with human Syt-I (Moore et al. 2006), indicating that Syt-I plays a major role in regulated secretion. It is not clear if Syt-I is regulatory by enhancing or inhibiting full expansion of the pore. As it appears that membrane mechanics provide enough energy for expansion of the pore (Cans et al. 2003), it is highly possible that protein regulation is inhibitory. However, as exocytosis is not completely abolished, these findings also indicate that one or more proteins can replace Syt-I. PC12 cells over-expressing Syt-I display a prolonged open time of the fusion pore, whereas Syt-IV decreases the fusion pore open time (Wang et al. 2001). Transfection with Syt-IV, or increasing endogenous Syt-IV by forskolin treatment, increases the frequency and duration of kiss-and-run events. Full fusion is inhibited by mutation of a Ca²⁺ ligand in the C2A domain of Syt-I, whereas kiss-and-run exocytosis is inhibited by mutation of a homologous Ca²⁺ ligand in the C2B domain of Syt-IV (Wang et al. 2003).

In permeabilized PC12 cells, elevating the intracellular Ca²⁺ concentration reduced the fusion pore lifetime, indicating that Ca²⁺ acts during the actual fusion process. Both, opening of the fusion pore and dilatation of the pore were accelerated by Ca²⁺, suggesting separate Ca²⁺ control over each of these steps. This was further confirmed by the fact that Ca²⁺ ligand mutations in either the C2A or C2B domains of Syt-I reduced fusion pore opening, but had opposite actions on the rate of fusion pore closure (Wang et al. 2006). Exactly how the Syt proteins regulate pore is still a matter of debate and ongoing interest.

SCAMP2

Another protein critically involved in fusion pore formation is secretory carrier membrane protein 2 (SCAMP2). Expression of a point mutant of SCAMP2 in PC12 cells resulted in the inhibition of exocytosis. Interestingly, dilatation of the opened fusion pores was also hampered, resulting in a relative abundance of stand-alone foot signals, thus linking SCAMP2 to the process of fusion pore formation (Liu et al. 2005). As such, amperometry at (transfected) PC12 cells is a powerful tool to further complete the detailed picture of the molecular regulatory mechanisms underlying exocytosis.

L-DOPA increases quantal size

Several early studies used PC12 cells to investigate the effects of different pharmacological agents on the average catecholamine content of PC12 cell vesicles. For example, the quantal size of release events can be increased by treatment of the cells with the dopamine precursor L-3, 4-dihydroxyphenylalanine (L-DOPA) (Kozminski et al. 1998). In one set of experiments, treatment of cells with 50 μM L-DOPA for 40–90 min increased the average quantal size for release by 251% (Pothos et al. 1996). When release events are examined at different times following exposure to L-DOPA, this treatment steadily increases vesicle content, already within 15 min of exposure, and after 60 min of exposure apparently saturates at ~175% of the contents before incubation with L-DOPA (Fig. 5A). Interestingly, L-DOPA initially increases the mean number of exocytosis events in undifferentiated PC12 cells (Fig. 5b). This increase, however, is transient and the release frequency rapidly declines following repeated stimulations indicating depletion of the pool of releasable vesicles (Westerink et al. 2000).

Autoreceptors on PC12 cells

Besides upregulation of vesicle contents by L-DOPA, vesicle content can also be reduced by exposure to amphetamine or reserpine (Sulzer et al. 1995, Kozminski et al. 1998). Importantly, Pothos et al. (1998) used amperometry to demonstrate that quantal size can also be modulated physiologically, e.g. by activation of D₂ autoreceptors. In these experiments, PC12 cells were treated with the D₂ agonist quinpirole resulting in a ~50% decrease in the quantal size of release events. In contrast to the results for amphetamine and reserpine, which appear to deplete vesicular stores by operating on the vesicle monoamine transporter (VMAT), the experiments with quinpirole provide evidence for a receptor-mediated mechanism that can deplete or modulate quantal release, probably through feedback on the TH-mediated dopamine synthesis pathway.

Variability of vesicle size and the fusion pore

PC12 cells can be loaded with L-DOPA to increase quantal size (Pothos et al. 1996, Kozminski et al. 1998, Colliver et al. 2000b, Westerink et al. 2000). However, it was unclear by what means this enhancement was taking place. Colliver et al. (2000b) used amperometry and transmission electron microscopy to demonstrate that the vesicle size in PC12 increased in the presence of excess L-DOPA and decreased in the presence of reserpine. This was later corroborated for adrenal chromaffin cells by Lindau and coworkers using patch-amperometry (Gong et al. 2003).

More recently, Sombers et al. (2004) have used the ability to vary vesicle size to examine the effect of vesicle size on release via the fusion pore. Although L-DOPA and reserpine increase and decrease vesicle size, respectively, the dense core of the vesicle is mostly unchanged (Colliver et al. 2000b) and it appears that the increased amount of dopamine in the vesicle following L-DOPA exposure is mainly in the clear halo of the vesicle (Sombers et al. 2005). The dense core consists primarily of the catecholamine-storage protein chromogranin A and is thought to expand during exocytosis, facilitating fusion pore expansion (Amatore et al. 2005). Although this is a likely outcome with adrenal chromaffin cell vesicles, it appears that the fusion pore in PC12 cell vesicles can pass through a transition state involving a lipid nanotube. Under conditions where the dense core expands, placing tension on the vesicular membrane, the nanotube is stabilized and constricts. This mechanism has been proposed to explain a higher occurrence of events with a foot for the smaller vesicles with larger ratio of dense core to vesicle volume, following treatment with reserpine (Sombers et al. 2004) and might provide a mechanism for forced release of dopamine from the halo of the vesicle through the fusion pore.

In some cases the fusion pore opens and closes without a full exocytosis event occurring. This phenomenon, termed ‘kiss and run’ has been observed in dopaminergic midbrain neurons using amperometry (Staal et al. 2004). If these events represent a significant communication pathway, then the cell’s ability to modify vesicle size could be a mechanism of neuronal plasticity (learning and memory).

Heavy metals

A number of recent studies indicate that environmental and chemical factors can strongly modulate exocytosis in PC12 cells. Of these environmental factors, heavy metals are probably best studied.

Heavy metals have been known for many years to disturb neurotransmission and ion channel function (reviewed in, e.g. Cooper & Manalis 1983, Vijverberg et al. 1994). In PC12 cells, the heavy metal Cd²⁺ has been shown to block voltage-gated Ca²⁺ channels (VGCCs; Shafer 1998) and, consequently, inhibit neurotransmitter release (Taylor & Peers 1998, Westerink & Vijverberg 2002b). Additionally, using a combination of amperometry and membrane capacitance measurements, it has been shown that Cd²⁺, as well as Sr²⁺ and Ba²⁺, can induce vesicular catecholamine release in PC12 cells, probably by acting as a partial agonist of the proposed Ca²⁺-sensor synaptotagmin (Kishimoto et al. 2001).

The neurotoxic heavy metal Pb²⁺ was shown to not only block VGCCs (Shafer 1998), but also to enhance spontaneous neurotransmitter release in populations of PC12 cells (Bressler et al. 1996). This effect was delayed and, as it occurred irrespective of the presence of extracellular Ca²⁺, was attributed to intracellular effects of Pb²⁺ (reviewed in Suszkiw 2004, Toscano & Guilarte 2005). Recent amperometric recordings, in combination with the imaging of the intracellular Ca²⁺ and Pb²⁺ concentrations, demonstrated the vesicular origin of the enhancement of spontaneous neurotransmitter release in dexamethasone-differentiated PC12 cells (Fig. 6). Pb²⁺ concentration-dependently induced exocytosis from intact as well as from ionomycin-permeabilized PC12 cells through direct effects on intracellular mechanisms following a concentration-dependent delay (Fig. 6a). Pb²⁺-induced exocytosis occurred only after partial saturation of intracellular high-affinity buffer components (Fig. 6b; Westerink & Vijverberg 2002a). Pb²⁺-induced exocytosis could be blocked almost completely by inhibition of Ca²⁺/calmodulin-dependent protein kinase II (CaM kinase II), indicating that CaM kinase II plays a major role in Pb²⁺-induced catecholamine exocytosis from PC12 cells (Fig. 6c) (Westerink et al. 2002), although synaptotagmin I is also likely to be involved (Bouton et al. 2001). Thus, PC12 cells have successfully been used to demonstrate that heavy metals can exert inhibitory (block of VGCCs) as well as stimulatory (mimicking Ca²⁺ as a trigger for exocytosis) effects on neurotransmission.

Organic solvents

Organic solvents, particularly toluene, are occasionally used as a drug of abuse and are known to cause a variety of neurotoxic and neuropathological effects. Exposure to toluene causes post-synaptic effects on a variety of receptors (reviewed in Bowen et al. 2006). Using amperometry at PC12 cells, it was shown that toluene induced catecholamine exocytosis in dexamethasone-differentiated PC12 cells, which depended on the influx of extracellular Ca²⁺ through VGCCs (Westerink & Vijverberg 2002b). However, in another study, it was shown that toluene dose-dependently inhibits the depolarization-induced rise in intracellular Ca²⁺ in NGF-differentiated PC12 cells (Tillar et al. 2002). These rather contradictory results might be explained by the different effects of NGF and dexamethasone on the density and expression patterns of VGCCs. Nonetheless, the increase in catecholamine release as observed in the amperometric study is in agreement with elevated levels of extracellular dopamine in rat brain (Riegel & French 1999).

Persistent organic pollutants

Several persistent environmental compounds, like polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs), are known to accumulate in biological tissues and cause behavioural symptoms as well as alteration in neurotransmission. Potential mechanisms underlying the effects of PCBs on the catecholaminergic system include changes in dopamine homeostasis, inhibition of vesicular catecholamine uptake and alteration of the intracellular Ca²⁺ concentration (reviewed in Fonnum et al. 2006, Mariussen & Fonnum 2006). Surprisingly, to date there is only one study in which amperometry was used to demonstrate that acute exposure of PC12 cells to the non-planar congener PCB 4 and to the coplanar congener PCB 126, but not exposure to the non-planar congener PCB 128, increases spontaneous catecholamine exocytosis. Unexpectedly, acute and subchronic exposure to these PCBs failed to cause changes in the contents of catecholamine-containing vesicles in this study (Westerink & Vijverberg 2002c), suggesting that the toxicological focus may need to be redirected from the turnover of catecholamines to other effects including the mechanisms underlying the subtle changes in basal neurotransmitter release and its functional consequences.

Brominated flame retardants, among them PBDEs, have been shown to inhibit vesicular neurotransmitter uptake and induce neuronal cell death (reviewed in Fonnum et al. 2006, Mariussen & Fonnum 2006). Only recently, it was shown that a single exposure of young mice to BDE-47 reduced long-term potentiation recorded in hippocampal slices. This reduction in long-term potentiation was probably caused by a post-synaptic mechanism and coincided with a decrease in post-synaptic glutamate receptor subunits NR2B and GluR1 and autophosphorylated-active αCaM kinase II. The involvement of a post-synaptic mechanism was further underlined by in vitro research in dexamethasone-differentiated PC12 cells revealing that acute exposure to high concentrations of BDE-47 (20 μM) induced only a modest increase in intracellular Ca²⁺ and basal catecholamine exocytosis (Dingemans et al. 2007).

These combined findings illustrate the usefulness of neuroendocrine PC12 cells as models for past and future neurotoxicological studies.