Using double fluorescence with primary antibodies against TH and DBH, we demonstrated that the vast majority of catecholaminergic fibers in the rat Vmot were immunoreactive for both enzymes. We therefore presume that the identified fibers and related structures are mainly NAergic and not dopaminergic. The presumed NA fibers and their varicosities appeared as early as the day of birth. These structures showed an age-dependent increase in both fiber density and the size of their varicosities, which, according to morphological criteria, reached mature levels at P15 and P30, respectively. Although a lower density of fibers and fewer varicosities were observed at early postnatal ages (before P15), ultrastructural examination revealed that functional synaptic connections of the presumed NA fibers on somata, dendrites, and other axonal terminals in the Vmot were seen at all postnatal ages examined. These results suggest that the NA system functions immediately after birth and exerts its effect on synaptic transmission, both pre- and postsynaptically, and on global neuronal excitability during the entire postnatal development of neuronal circuits in the Vmot.
Noradrenergic Innervation in Vmot
Although we did not provide any direct evidence for the possible sources of the NA innervation in the Vmot, it is reasonable to presume that it arises from the lateral tegmental catecholaminergic cell groups, referred to as the A5 and A7 catecholamine cell groups by many investigators (Vornov and Sutin, 1983; Grzanna et al., 1987; Lyons and Grzanna, 1988). Consistent with this suggestion, we found that all neurons of the A5 and A7 groups were both TH-IR and DBH-IR. Direct morphological evidence for this idea comes from the report of Vornov and Sutin (1983), who injected horseradish peroxide into the Vmot of normal and noradrenergically hyperinnervated rats and found that the A7 cell group was the sole source of NA afferents to the Vmot. Grzanna et al. (1987) and Lyons and Grzanna (1988) also reported that multiple cell groups in the pons provide noradrenergic innervation to the Vmot. Moreover, these studies suggest that the Vmot receives NA afferents mainly from the A7 cell group and, to a lesser extent, from the A5 cell group. It should be borne in mind that the A7 group is located rostral, and the A5 group ventral, to the Vmot (see also Fig. 1D–F). If the A7 neurons send NA fibers to the Vmot caudally with their fibers running in a horizontal direction, and the A5 neurons send NA fibers to the Vmot dorsally with their fibers running in a vertical direction, this could explain our observation that projection in the horizontal direction was more abundant than that in the vertical direction at all developmental ages in the present study.
Development of Noradrenergic Axon Profiles in Vmot
In this study, we found that both the density and size of varicosities of the NA fibers increased during postnatal development and reached adult levels at approximately P15 and P30, respectively. In rats, maturation of both the distribution pattern and density of the NA innervation by P15 has been reported in many other areas of the central nervous system, including the visual and motor cortex (Latsari et al., 2002) and the septal area (Antonopoulos et al., 2004). Nevertheless, our EM morphometric measurements suggested that the bouton areas showed no age-dependent change. This result seems to conflict with the results of the LM study, which showed an age-dependent increase in varicosity size. One possible explanation could be that the shape of the bouton might resemble a cylinder, with a small diameter and a long axis. It is possible that, during development, the diameter remains relatively constant, while the axis increases in length. Since an ultrathin section only covers a short distance of the axis and reveals mainly the diameter, whereas the LM section covers the entire bouton structure, this would explain the age independence of the EM observations on bouton area and the age dependence of the LM observations. This argument is supported by the fact that fewer serial sections were needed to reveal synaptic contacts at an early postnatal age (P5) than in the adult.
Although there was a lower density of adrenergic fibers and fewer varicosities at early postnatal ages, functional synaptic contacts could be seen as early as P5. This conclusion was based on the observations that several important ultrastructural parameters related to the function of synapses (Peters et al., 1991), including the maximal apposition length, maximal active zone length, and morphology of synaptic vesicles, in early postnatal stages were the same as in the adult.
Ultrastructure of NA Terminals
The ultrastructural characteristics of NA terminals in various regions of the brain have been studied by several investigators (Beaudet and Descarries, 1978; Itakura et al., 1981; Olschowka et al., 1981; Papadopuolos et al., 1989; Hagihira et al., 1990). One important issue that is still controversial is the proportion of NA terminals forming specific synaptic contacts. Studies using autoradiography after topical application of [3H] noradrenaline (Beaudet and Descarries, 1978) or potassium permanganate fixation (Itakura et al., 1981) have suggested that monoaminergic axon terminals do not form synaptic contacts with distinct membrane specializations and that monoamines seem to be released nonsynaptically in the central nervous system, influencing not only adjacent, but also more distant, neurons, the so-called volume transmission (Vizi et al., 2004). However, apart from criticisms of the effectiveness of these two experimental approaches for the study of the fine structure of monoaminergic terminals, the number of consecutive sections examined (only three in each study) was probably inadequate. Since a single section contains only a small portion of a typical varicosity (the diameter of a varicosity is roughly 15 times greater than the thickness of an ultrathin section) and since the site of membrane specialization is a relatively small portion of the total surface area of a varicosity, it is likely that many randomly oriented sections through synaptic varicosities do not contain the specialized zone (Olschowka et al., 1981).
In the present study, the ultrastructure of a TH-IR axon terminal was investigated throughout an entire varicosity in order to judge whether the labeled terminals formed synaptic contacts with any neuronal structures. Our study revealed that more than half (74% at P15 and 62% in the adult) of TH-IR terminals made synaptic contacts with other neuronal elements with clear membrane specialization in the Vmot. Thus, it is reasonable to assume that the difference between the results presented here and those of other workers could be at least in part due to the different number of consecutive sections examined. Consistent with the present results, Arce et al. (1994), Hagihira et al. (1990), and Papadopoulos et al. (1989), who also examined the morphological characteristics of NA-containing axon terminals in serial ultrathin sections, demonstrated that the majority of NA terminals in the superior colliculus, spinal dorsal horn, and visual and frontoparietal cortex form conventional synapses, suggesting that the action of noradrenaline is mainly mediated via a synaptic mechanism. Taken together, the results of the present and other studies based on multiple serial ultrathin sections demonstrate that the action of central noradrenaline is mediated by conventional synaptic transmission and is therefore characterized by a high degree of functional specificity determined by the spatial distribution of specialized junctions between presynaptic varicosities and adrenergic receptors.
However, since we also found that a substantial percentage of NA varicosities did not make contacts with postsynaptic targets, but contained clusters of synaptic vesicle aggregates near the plasma membrane, the possibility that certain central NA varicosities might exert their influence through a nonsynaptic mechanism has to be taken into account. In particular, recent studies have suggested the existence of extrasynaptic receptors and that, in some experimental or pathological conditions, e.g., blockage of transmitter uptake, transmitters could spill over and quickly fill up the extracellular space, allowing the activation of extrasynaptic receptors in the central nervous system (Sykova, 2004; Vizi et al., 2004). As for the peripheral nervous system, it has also been reported that noradrenaline released from nonsynaptic varicosities might diffuse beyond adjacent postsynaptic elements and influence large numbers of neurons (Descarries et al., 1977; Vizi et al., 2004). The effect of noradrenaline released in this diffuse fashion would be relatively unselective and determined primarily by the distribution of adrenergic receptors on nonadjacent neuronal elements.
Regardless of whether or not a TH-IR bouton targeted a postsynaptic element, we found that two types of synaptic vesicles could be identified in most TH-IR terminals, namely, small, pleomorphic, agranular vesicles and prominent dense-core vesicles. It has been suggested that the small pleomorphic vesicles probably contain catecholamines (Doyle and Maxwell, 1991), while the dense-core vesicles contain transmitters other than catecholamines (Hökfelt and Ljungdahl, 1972). Recent evidence indicates that some of these dense-core vesicles might contain neuropeptides (Meright et al., 1989). In fact, the coexistence and corelease of catecholamines and neuropeptides, such as neuropeptide Y (Charnay et al., 1982; Everitt et al., 1984; Yamazoe et al., 1985; Melander et al., 1986) and galanin (Xu et al., 1998; Simpson et al., 2006), have been suggested in the central nervous system.
The majority of active zones formed by NA terminals in the Vmot were classified as being symmetrical, with a slight accumulation of electron-dense material on the presynaptic and postsynaptic sides of the junction. In agreement with the present results, similar observations have been reported for synapses on the cat lumbosacral spinal dorsal horn using TH as marker (Doyle and Maxwell, 1991) and in the rat visual and motor cortex (Latsari et al., 2002) and septal area (Antonopoulos et al., 2004) using DBH as marker. In our study, we found that the majority of postsynaptic targets were dendrites (59% at P5, 74% at P15, and 89% in the adult) and the remainder cell bodies or unlabeled axon terminals. These results suggest that NA terminals might be directly involved in governing neuronal excitability and in modulating synaptic input onto trigeminal motoneurons, both pre- and postsynaptically. Consistent with this idea are results from previous electrophysiological studies, which showed that noradrenaline released from NA terminals facilitates both the excitatory postsynaptic potentials recorded in trigeminal motoneurons and the monosynaptic reflex evoked by stimulation of group Ia primary afferent axons from the masseter muscle (the masseteric jaw closing reflex) (Morilak and Jacobs, 1985; Vornov and Sutin, 1986; Stafford and Jacobs, 1990). Since the masticatory rhythm is due to alternation of excitatory and inhibitory postsynaptic potentials in jaw-closer motoneurons (Goldberg, 1972; Chandler et al., 1985; Enamoto et al., 1987), the NA innervation could play an important role in the control of rhythmical jaw movements. This view is supported by the demonstration that low-current iontophoretic application of noradrenaline potently facilitates both glutamate-induced motoneuronal discharges and ongoing rhythmical motoneuronal activity during rhythmical jaw movements evoked by repetitive cortical stimulation (Katakura and Chandler, 1990).
In conclusion, our data provide direct morphological evidence that adrenergic innervation modulates rhythmical jaw movement control, as suggested by previous physiological studies. Our results also suggest that the adrenergic innervation might exert its effect via both synaptic and nonsynaptic mechanisms, with the former being more important, and that these mechanisms might operate as early as the day of birth.