The filum terminale of the frog spinal cord, a non transformed preparation: I. Morphology and uptake of gamma-aminobutyric acid

Gross and microscopic study of the filum terminale: does the filum contain functional neural elements?

OBJECT

The filum terminale (FT) is considered a fibrous structure that extends from conus medullaris of the spinal cord to coccyx. Based on previous studies and from their own experience with intraoperative electrophysiological monitoring of the sacral nervous system, the authors postulate that the FT contains functional neural elements in some individuals.

METHODS

The FT was dissected from 13 fresh stillborn cadavers (7 male, 6 female; mean gestational age 36 weeks and 1 day). The gross anatomical features were recorded, and connections between the FT and the nerve roots of the cauda equina were noted. These connections, when present, were sectioned for histological studies. The fila (both interna and externa) were also sectioned for histological and immunohistochemical studies. In addition, FT specimens were obtained from 5 patients undergoing sectioning of the FT in an untethering surgical procedure.

RESULTS

There were 5 gross connections between the FT and nerve roots demonstrating nerve fibers that were positive for S100. The FT showed islands of cells that were positive for GFAP in 10 cases, synaptophysin in 3 cases, S100 in 11 cases, and nestin in 2 cases. The nerve fibers in the FT were myelinated in 2 cases. The conus ended at the L-1 or L-2 vertebral level in all 13 specimens. The dural sac terminated at the S-2 vertebral level in most of the specimens. The 5 FT specimens that were obtained from patients revealed nerve bundles that were positive for S100 in 4 cases and cells that were positive for GFAP in 3 cases.

CONCLUSIONS

There are gross anatomical connections between the FT and nerve roots that contain nerve fibers. Apart from fibrous stroma, the FT may contain nerve bundles and cells that stain positive for GFAP, synaptophysin, S100, and nestin. These microscopic findings and previous intraoperative electrophysiological studies suggest a probable functional role for the FT in some individuals. At birth, the conus ends at a higher vertebral level (lower L-1 or upper L-2) than L-3.

Membrane currents and morphological properties of neurons and glial cells in the spinal cord and filum terminale of the frog

Using the patch-clamp technique in the whole-cell configuration combined with intracellular dialysis of the fluorescent dye Lucifer yellow (LY), the membrane properties of cells in slices of the lumbar portion of the frog spinal cord (n=64) and the filum terminale (FT, n=48) have been characterized and correlated with their morphology. Four types of cells were found in lumbar spinal cord and FT with membrane and morphological properties similar to those of cells that were previously identified in the rat spinal cord (Chvátal, A., Pastor, A., Mauch, M., Syková, E., Kettenmann, H., 1995. Distinct populations of identified glial cells in the developing rat spinal cord: Ion channel properties and cell morphology. Eur. J. Neurosci. 7, 129-142). Neurons, in response to a series of symmetrical voltage steps, displayed large repetitive voltage-dependent Na(+) inward currents and K(+) delayed rectifying outward currents. Three distinct types of non-neuronal cells were found. First, cells that exhibited passive symmetrical non-decaying currents were identified as astrocytes. These cells immunostained for GFAP and typically had at least one thick process and a number of fine processes. Second, cells with the characteristic properties of rat spinal cord oligodendrocytes, with passive symmetrical decaying currents and large tail currents after the end of the voltage step. These cells exhibited either long parallel or short hairy processes. Third, cells that expressed small brief inward currents in response to depolarizing steps, delayed rectifier outward currents and small sustained inward currents identical to rat glial precursor cells. Morphologically, they were characterized by round cell bodies with a number of finely branched processes. LY dye-coupling in the frog spinal cord gray matter and FT was observed in neurons and in all glial populations. All four cell types were found in both the spinal cord gray matter and FT. The glia/neuron ratio in the spinal cord was 0.78, while in FT it was 2.0. Moreover, the overall cell density was less in the FT than in the spinal cord. The present study shows that the membrane and morphological properties of glial cells in the frog and rat spinal cords are similar. Such striking phylogenetic similarity suggests a significant contribution from distinct glial cell populations to various spinal cord functions, particularly ionic and volume homeostasis in both mammals and amphibians.

Extracellular diffusion parameters in spinal cord and filum terminale of the frog

Extracellular space (ECS) diffusion parameters were studied in isolated frog spinal cord grey matter and filum terminale (FT), that is predominantly composed of glial cells and axons. We compared the cell swelling induced by K(+) application, hypotonic stress and tetanic stimulation of afferent input. The ECS diffusion parameters, volume fraction alpha (alpha = ECS volume/total tissue volume), tortuosity lambda (lambda(2) = free/apparent diffusion coefficient in the tissue) and non-specific cellular uptake k', were determined by the real-time iontophoretic method using TMA(+)-selective microelectrodes. Stimulation-evoked changes in extracellular K(+) concentration ([K(+)](e)) were measured by K(+)-selective microelectrodes. Histological analysis revealed that in the central region of the FT, the cell density was lower than in SC, neurons and oligodendrocytes were scarce, GFAP-positive astrocytes were abundant, and they showed thicker and more densely stained processes than in spinal cord. In the FT, alpha was 58% higher and lambda significantly lower than in the spinal cord. In 50 mM K(+), alpha in spinal cord decreased from about 0.19 to 0.09, i.e., by 53%, whereas in FT from about 0.32 to 0.20, i.e., by only 38%; lambda increased significantly more in FT than in spinal cord. Hypotonic solution (175 mmol/kg(-1)) resulted in similar decreases in alpha, and there were no changes in lambda in either spinal cord or FT. Stimulation of VIII or IX dorsal root (DR) by 30 Hz evoked an increase in [K(+)](e) from 3 to 11-12 mM in spinal cord, but to only 4-5 mM in FT. In the spinal cord this stimulation led to a 30% decrease in alpha and a small increase in lambda whereas in the FT the decrease in alpha was only about 10% and no increase in lambda was found. We conclude that in spinal cord, a complex tissue with a higher density of cellular elements than the FT, 50 mM K(+), hypotonic stress as well as DR stimulation evoked a greater decrease in ECS volume than in FT. Nevertheless, the K(+)-induced increase in tortuosity was higher in FT, suggesting that a substantial part of the K(+)-evoked increase in lambda was due to astrocytic swelling.

The ventriculus terminalis and filum terminale of the human spinal cord

Serial sections of the conus medullaris and the filum terminale of 23 randomly selected human spinal cords were studied by light and electron microscopy, and following immunoperoxidase staining for glial fibrillary acidic protein (GFAP), vimentin, neuron-specific enolase (NSE), amyloid beta protein, and S-100 protein. The intradural portion of the filum contains bundles of GFAP-positive glial fibers, scattered silver- and NSE-positive neurons, segments of peripheral nerve, blood vessels, fibrous connective tissue, and fat. Glial cell clusters varying from five to 100 cell layers thick at times constitute the bulk of the filum. The periependymal glial cells possess moderate amounts of eosinophilic cytoplasm and relatively uniform round to ovoid nuclei containing evenly distributed chromatin. They are distributed diffusely with no specific pattern of organization, although some of them showed a tendency to form acinar structures. A minority of the glial cells showed GFAP immunoreactivity, and some were immunoreactive for vimentin. Electron microscopy demonstrated the presence of periependymal cells showing cilia, microvilli, and the formation of intercellular junctional complexes, as well as cells containing bundles of glial filaments within the cytoplasm. Degenerated NSE-positive neurons and degenerated neurites resembling neuritic plaques were also demonstrated. However, immunoperoxidase staining for amyloid beta protein was negative in these structures. Thus, the filum terminale is endowed with an abundance of glial cells and neurons and is not simply a fibrovascular tag. Periependymal glial cells in the filum terminale should not be mistaken for neoplasm. The presence of neuropil with profuse astroglial and neuronal components within the filum terminale suggests a possible functional role for these structures.

Sodium dependency of GABA uptake into glial cells in bullfrog sympathetic ganglia

The kinetics of sodium dependency of GABA uptake by satellite glial cells was studied in bullfrog sympathetic ganglia. GABA uptake followed simple Michaelis-Menten kinetics at all sodium concentrations tested. Increasing external sodium concentration increased both Km and Vmax for GABA uptake, with an increase in the Vmax/Km ratio. The initial rate of uptake as a function of the sodium concentration exhibited sigmoid shape at 100 microM GABA. Hill number was estimated to be 2.0. Removal of external potassium ion or 10 microM ouabain reduced GABA uptake time-dependently. The effect of ouabain was potentiated by 100 microM veratrine. These results suggest that at least two sodium ions are involved with the transport of one GABA molecule and that sodium concentration gradient across the plasma membrane is the main driving force for the transport of GABA. The essential sodium gradient may be maintained by Na+, K(+)-ATPase acting as an ion pump.

Organization of the filum terminale in the frog

The histological organization of the filum terminale of the spinal cord in Rana catesbeiana and Rana pipiens was characterized to determine if this region possessed an organized neuropil or whether it was merely a glial remnant that persisted after absorption of the larval tail. The excised filum was maintained in vitro. Intracellular electrophysiological recording was performed with horseradish peroxidase injection. Tyrosine hydroxylase and serotonin distribution were revealed by immunocytochemical methods. Astroglia were the dominant cell type and displayed an elaborate variety of forms. The mean membrane potential was logarithmically related to the extracellular potassium concentration but displayed a sub-Nernstian slope. Oligodendroglia were also seen, as well as ependyma that extended from the central canal to the pial surface. Neuronal activity was revealed by occasional intracellular penetration of elements that displayed spontaneous excitatory postsynaptic or action potentials. The major evidence for the presence of neurons was the demonstration of tyrosine hydroxylase (TH) immunoreactivity in a large population of cerebrospinal fluid-contacting neurons that abutted the ventral half of the central canal. The axons of these cells entered a ventral bundle and ascended the cord; some fibers left this tract and apparently terminated on large arcuate neurons within the filum. Serotoninergic fibers were primarily confined to a subpial location at the dorsal midline. We conclude that the filum terminale of the frog has a sparse but functional neuropil that is organized around the central canal and supported by a profusion of elaborate glial forms.

Regional cell count and respiration in the brain of Rana catesbeiana

Cell counts were made in three divisions of the bullfrog brain: (1) olfactory lobestelencephalon, (2) diencephalon-mesencephalon, and (3) the hindbrain; average cellular respiration in potassium-enriched Ringer solution and the rate of respiration in sodium-free medium were tested for their possible relationship to the numbers of glia and numbers of neurons in each region. After restoring some of the sodium ions to sodium-free minces, the percent increase in respiration was plotted for each brain segment and compared to the percent of glia per brain division to determine the glial respiratory contribution. Where glial numbers were greater, the potassium-stimulated respiration was greater; where neuronal numbers were greater, the sodium-free respiration was most prominent. The respiration increased most above the sodium-free rate, when sodium ions were added or when the full amount of sodium was present, in the hindbrain, where the glia appear to metabolize more vigorously and the neurons least actively.

Ubiquitous presence of the tailed, asymmetric forms of acetylcholinesterase in the peripheral and central nervous systems of the frog (Rana temporaria)

Five molecular forms of acetylcholinesterase can be solubilized from the peripheral and central nervous systems of the frog: they will be referred to as the 3.6, 6, 10.5, 14 and 18 S forms. They seem to be analogous to the forms present in endplate-rich and endplate-free regions of frog skeletal muscle. In particular the 18 and 14 S forms represent the collagen-tailed forms of frog acetylcholinesterase. These heavy forms are found in all peripheral and central tissues examined, including whole brain or regions of brain: cerebellum, telencephalon, optic tectum, spinal cord, spinal ventral and dorsal roots and sciatic nerve, as well as in glial or Schwann cellrich tissues devoid of neuronal elements, such as the filum terminale or the severed stump of the nerve, several weeks after section. The 18 S form may represent up to 30% of total acetylcholinesterase activity. It thus seems that the 14 S and 18 S forms are very widely distributed throughout most neuronal and non-neuronal tissues in amphibians.

The filum terminale of the frog spinal cord, a nontransformed glial preparation: II. Uptake of serotonin

The uptake of 5-HT was measured in the frog filum terminale (FT), a preparation composed almost exclusively of normal glia. [3H]5-HT was taken up by the FT via a high-affinity, sodium-dependent, temperature-sensitive transport system having a Km of 0.7 microM. In addition, a variety of drugs and aromatic amines known to selectively inhibit 5-HT uptake by synaptosomes and brain slices affected the uptake of 5-HT by the FT in qualitatively similar manner. The FT was shown to accumulate [3H]5-HT at rates significantly greater than the lumbar enlargement which contains both neurons and glia. The glial accumulation of 5-HT by the FT was verified by autoradiography. These findings strongly support the suggestion that glia may modulate aminergic transmission by competing with neurons for the reuptake of neuronally released amines.

Primary afferent depolarization. Distribution of the gamma-aminobutyric acid system in frog spinal cord

In the frog spinal cord primary afferent depolarization (PAD) constitutes a powerful inhibitory control mechanism. It has been suggested that gamma-aminobutyric acid (GABA) is the transmitter substance involved in the genesis of PAD. In these studies we show that maximal glutamic acid decarboxylase activity is localized roughly 400-600 micrometers from the dorsal surface, and that correlates well with the intraspinal distribution of field potentials associated with PAD. Measurements of GABA in serial spinal cord sections cut in a dorsal--ventral direction shows that high levels of GABA are seen at 400--600 micrometers, with a peak at 800 micrometers from the dorsal surface. Stimulation at frequencies shown to produce PAD augments the release of endogenous GABA from a superfused frog hemicord preparation.

GABA uptake in astrocytes in primary cultures: coupling with two sodium ions

The influence of sodium ions on GABA uptake into astrocytes in primary cultures has been investigated performing kinetic analysis of GABA uptake at different sodium concentrations in the range 16 to 151 mM. These investigations reveal that sodium affects both the Km and the Vmax of the saturable component of the astroglial GABA uptake. Uptake rates as a function of the sodium concentration at high GABA concentrations (greater than or equal to 50 microM) were clearly sigmoid whereas at lower GABA concentrations this sigmoid shape was not obvious. Accordingly, Hill plots of the sodium dependency at high GABA concentrations exhibited straight lines with slopes of 2.0 to 2.5, suggesting that the coupling ratio between sodium and GABA is at least two. Corresponding Hill plots at lower GABA concentrations exhibited slopes of 1.6 to 1.8. Moreover, plots of 1/v versus 1/Na2 gave better fits to straight lines than plots of 1/v versus 1/Na which were curvilinear upward. Again, this curvilinearity was more pronounced at high GABA concentrations that at low GABA concentrations. From these results it is concluded that GABA uptake into astrocytes in primary cultures requires the binding of at least two sodium ions per GABA molecule transported.