Nerve ultrastructure of the arteries of the brain stem]

Ultrastructure of the nerve apparatus in the arteries of the brain base has been studied in cats. The structure of peri- and adventitial nerves has been investigated electron microscopically. Three types of efferent axons and four types of synaptic vesicles (small agranular and granular, large granular, large electron opaque vesicles) have been revealed. Vesicle-containing axons in the brain arteries approach the external smooth muscle cells of about 80 nm. Terminal axonal dilatations possessing direct and mediated connections with muscular cells of the middle tunica have been revealed.     

Autonomic nerve terminals in relation to contractile and non-contractile structures in the conduit coronary artery of the dog

The ramus interventricularis anterior (RIA), its first- and second-order branch were prepared for EM (perfused with glutaraldehyde under pressure, or simply fixed with KMnO4). No nerve fibres were found in the tunica media of either of the three consecutive segments. In the tunica adventitia axons with varicosities were found at a distance from the tunica media of 0.5-15 microns (about 50% 0.5-4.5 microns) in the RIA, 0.4-12 microns (about 50% 0.5-3.4 microns) in the first-order branch and 0.3-6.0 microns (about 50% 0.3-2.3 microns) in the second-order branch. Varicosities contain small, dense-cored vesicles (35-60 nm) and large, dense-cored vesicles (70-90 nm, exceptionally up to 120 nm); the other type contains small, clear vesicles (35-60 nm) and few large, dense-cored vesicles (70-90 nm). The remarkably large distance between the nerve terminals and smooth muscle cells fits well with the small range of sympathetic control of the conduit coronary artery. Close apposition of nerve terminals to fibroblasts (30-200 nm) was revealed in all three consecutive coronary portions. Moreover, terminal axons often lose the Schwann cell cover on the abluminal site and face the fibroblast.     

Innervation of mouse lymph nodes: nerve endings on muscular vessels and reticular cells

Lymph node nerve endings have been studied in 1- to 48-day-old mice. Serial sections of Epon-embedded lymph nodes were observed under the electron microscope to find the nerve endings. Most lymph node nerve fibers finally reach the smooth muscle cells of arterioles and muscular venules. Both kinds of vascular endings are similar, although endings are less numerous on venules. Nerve endings consist of one or more nerve processes surrounded by a usually incomplete Schwann cell sheath; frequently, axons show wide areas directly facing the muscle cells. The distance between such a naked axon and a myocyte ranges from 100 to 800 nm. Small granulated and clear vesicles are especially abundant in varicosities of nerve processes that are located very close to muscle cells. Nerve endings of lymph node vasculature probably correspond to vasomotor sympathetic adrenergic endings, regulating the degree of contraction of vessels which have a muscular layer. Other kinds of nerve endings also exist in lymph nodes: some axons appear free in the stroma and contact the surfaces of reticular cells; the latter also extend delicate cytoplasmic processes that surround the axons. The functional significance of nerve cell-reticular cell contacts is unknown.     

Innervation of the renal vasculature of the toad (Bufo marinus)

The innervation of the dorsal aorta and renal vasculature in the toad (Bufo marinus) has been studied with both fluorescence and ultrastructural histochemistry. The innervation consists primarily of a dense plexus of adrenergic nerves associated with all levels of the preglomerular vasculature. Non-adrenergic nerves are occasionally found in the renal artery, and even more rarely near the afferent arterioles. Many of the adrenergic nerve profiles in the dorsal aorta and renal vasculature are distinguished by high proportions of chromaffin-negative, large, filled vesicles. Close neuromuscular contacts are common in both the renal arteries and afferent arterioles. Possibly every smooth muscle cell in the afferent arterioles is multiply innervated. The glomerular capillaries and peritubular vessels are not innervated, and only 3-5% of efferent arterioles are accompanied by single adrenergic nerve fibres. Thus, nervous control of glomerular blood flow must be exerted primarily by adrenergic nerves acting on the preglomerular vasculature. The adrenergic innervation of the renal portal veins and efferent renal veins may play a role in regulating peritubular blood flow. In addition, glomerular and postglomerular control of renal blood flow could be achieved by circulating agents acting via contractile elements in the glomerular mesangial cells, and in the endothelial cells and pericytes of the efferent arterioles. Some adrenergic nerve profiles near afferent arterioles are as close as 70 nm to distal tubule cells, indicating that tubular function may be directly controlled by adrenergic nerves.     

Innervation of the large arteries and heart of the toad (Bufo marinus) by adrenergic and peptide-containing neurons

Summary The innervation of the major arteries and heart of the toad (Bufo marinus) was examined by use of glyoxylic acid-induced catecholamine fluorescence and peptide immunohistochemistry. All arteries possessed a moderate to dense plexus of adrenergic axons, which also showed neuropeptide Y-like immunoreactivity (NPY-LI). Some adrenergic axons in the intracardiac vagal trunks showed NPY-LI, but the varicose adrenergic axons innervating the cardiac muscle of the atria and ventricle, and the coronary blood vessels did not display NPY-LI. About half of the nerve cell bodies in the anterior sympathetic chain ganglia with dopamine--hydroxylase-LI (DBH-LI) also contained NPY-LI. The nerve cell bodies with DBH-LI alone were generally larger (median diameter 30 m) than those with both DBH-LI and NPY-LI (median diameter 20 m). Some cell bodies showing DBH-LI alone were surrounded by boutons with NPY-LI but not DBH-LI. Axons that displayed simultaneously both substance P-LI (SP-LI) and calcitonin gene-related peptide-LI (CGRP-LI) also formed a plexus around all arteries studied, being particularly dense around the mesenteric and pulmonary arteries. These axons are most likely sensory since SP-LI was reduced by capsaicin treatment, and nerve cell bodies with both SP-LI and CGRP-LI were found in dorsal root ganglia and the vagal ganglion. A dense plexus of axons showing somatostatin-LI was located around the pulmonary artery and its main intrapulmonary branches. A few nerves with vasoactive intestinal polypeptide-LI were found around the dorsal aorta and pulmonary artery. No perivascular nerves with enkephalin-LI were observed. Reversed-phase, high-pressure liquid chromatography of acid extracts of the large arteries showed that the major peaks of NPY-LI and SP-LI coeluted with porcine NPY (1–36) and synthetic SP (1–11), respectively. Thus, the location and structure of these peptides in perivascular nerves has been highly conserved during vertebrate evolution.     

Contractile capacity of isolated flaps from the aorta of rats with stable arterial hypertension caused by the prolonged administration of cerebrosides

A study was made of the contractility of smooth muscle cells of abdominal aorta strips of noninbred white rats with stable arterial hypertension induced by protracted intraperitoneal injection of cerebrosides isolated from cattle brain. It was demonstrated that as compared with normotensive animals, smooth muscle cells of the animals' arteries are characterized by the increased influx of extracellular calcium via slow potential-dependent calcium channels and hypersensitivity to noradrenaline and serotonin.     

Preliminary observations, at the ultrastructural level, on the reactivity of cerebral arteries from New Zealand rabbits to combined atherogenic stimuli (hypercholesterol diet and hypertension)

Both in monkeys (Rhesus and Cynomolgus) and in New Zealand rabbits fed an atherogenic diet, a marked delay in the appearance of atherosclerotic lesions of the cerebral arteries in comparison with other arterial districts has been observed. This appearance has been described in monkeys as relatively earlier if hypertension is added to the atherogenic diet. Preliminary observations on a little group of rabbits on a 3 months hypercholesterolic diet, subjected to Goldblatt aortic coarctation, have shown an increase of blood pressure and a severe gross atherosclerotic involvement of aorta, resembling the one obtainable after 6 months of atherogenic diet. Histologically, the aorta predominantly shows lesions of the fatty streaks type with necrotic areas in the deep; the carotid lesions show some lipid in smooth muscle cells disseminated in a sub-endothelial "edematous" space (rich in protein). The cerebral arteries do not show any lesion. At TEM, the aortic lesions look sometimes as advanced plaques with an initial fibrosis at the surface; the carotid lesions are characterized by a granular deposit in the sub-endothelial space in which some smooth muscle cells (with lipid in the cytoplasm) are present; in the cerebral arteries only the presence of collagen fibers among the smooth muscle cells of the media, never observed in the animals fed the atherogenic diet alone, has sometimes been noted.     

Participation of smooth muscle cells in the formation of atherosclerotic plaques]

A study of the participation of the smooth muscle cells in the formation of atherosclerotic lesions was made on the autopsy material with the use of specific antiserum to the smooth muscle actomyosin and of indirect Coons' method. Typical forms of atherosclerotic lesions in the aorta, cerebral vessels and coronary arteries were studied. Smooth muscle cells were detected in the thickened intima alongside the atherosclerotic lesions, in fatty streaks, in the fibrous tissue of the atherosclerotic plaque, but they were not found in the atheromatous masses. The proliferation and migration of the smooth muscle cells is regarded as an essential factor in the pathogenetic mechanisms of atherosclersis.     

Effect of sympathectomy on the phenotype of smooth muscle cells of middle cerebral and ear arteries of hyperlipidaemic rabbits

This study was carried out to determine whether sympathectomy influences the phenotypic modulation of smooth muscle cells in the peripheral and cerebral arteries of heritable hyperlipidaemic rabbits. Unilateral superior cervical ganglionectomy (common origin of innervation to the middle cerebral artery and the central ear artery) was performed on four Watanabe heritable hyperlipidaemic rabbits. Cross-sections of the ipsi- (sympathectomized) and the contralateral (intact) cerebral and ear arteries were prepared 2 months later and labelled with monoclonal antibodies against vimentin and desmin, two markers of the differentiation of smooth muscle cells, and α-smooth muscle actin, a marker of these cells. Sections from control and sympathectomized arteries were analysed with a confocal laser scanning microscope. Compared with contralateral intact ear arteries, the sympathectomized ear artery developed a thickened intima with dedifferentiated smooth muscle cells, expressing α-smooth muscle actin but no desmin, whereas the middle cerebral artery remained unchanged. These results suggest that sympathectomy may favour the progression of atherosclerosis in peripheral but not in cerebral arteries of Watanabe heritable hyperlipidaemic rabbits     

Effect of atriopeptin on the adrenergic mechanism regulating blood vessel tonus

The direct action of atriopeptin on the cell regulation mechanism of the smooth muscle in isolated segments of the portal vein, aorta, pancreatic and cerebral arteries have been studied. It was found that atriopeptin induce the direct relaxation of the smooth muscle in the main vessels only (aorta, portal vein). In the cerebral and pancreatic arteries atriopeptin stops norepinephrine-induced contractions. The data obtained show that the action of the atriopeptin is mediated by Na+-K+ pump activation of smooth muscle cells and restricts vasoconstriction of catecholamine effect.     

Mucosal nerves and smooth muscle relationships with gastric glands of the opossum: an ultrastructural and three-dimensional reconstruction study

The precise anatomical relation by which autonomic nerve endings contact gastric epithelial cells to enhance the rate of gastric secretions is not fully understood. The aim of the present study was to clarify this issue by using the technique of serial section reconstruction of areas of the gastric mucosa. The work also explored the possibility of a functional role for a system of smooth muscle strands in the gastric mucosa that emanate from the muscularis mucosa, run in the lamina propria, and are associated in a unique manner with the gastric glands. Electron microscopic serial sections of the gastric mucosa were performed to visualize the entire limiting membrane of gastric epithelial cells to determine any nerve associations (especially varicose endings) with these cells. Evaluation of serial sections of five separate parietal cells showed that their basal membrane did not come in close contact (nearest distance 500 nm) with any nerve axon or varicosity. Moreover, the axons passing in the area of these cells ultimately showed varicose endings associated with smooth muscle cells in the adjacent connective tissue (often separated by only 20 nm), with mast cells or with vascular elements. Additionally, the lateral membrane of these five parietal cells did not contact any endocrine cell in the epithelium, although other parietal cells in the area were adjacent to endocrine cells. Chief cells in the immediate area also did not form any close associations with nerve varicosities. Random analysis of 5,000 additional epithelial cells in these sections showed no close associations to nerve elements with significant accumulations of neurosecretory vesicles (varicosities). Because of the observed existence of innervation to the smooth muscle strands in the area of the gastric glands, serial 1-micron epoxy sections of the gastric mucosa were prepared, and profiles of smooth muscle and gastric glands were entered into a computer-assisted reconstruction system. Three-dimensional reconstruction techniques were employed to reveal the existence of a unique association between the mucosal smooth muscle strands and the gastric glands. The muscle strands arose from the muscularis mucosa at regular intervals and became branched to form an intricate wrap around a series of gastric glands that empty into one gastric pit.(ABSTRACT TRUNCATED AT 400 WORDS)     

Innervation of the ureterovesical junction musculature of man

The ultrastructure of the innervation of the human ureterovesical junction was studied. Three different nerve terminals were distinguished among the smooth muscle cells. 1. Nerve processes containing predominantly small granular vesicles (40--60 nm in diameter). 2. Other nerve fibres contained predominantly small round agranular vesicles (30--50 nm in diameter). 3. Processes with large granulated vesicles (80--120 nm in diameter). The first type may be adrenergic, the second cholinergic and the third may originate from the local nerve cells. The gap between the nerve fibres and muscle cells was 300 to 500 nm wide and no synaptic thickenings were observed. This suggests that the transmitter may influence several muscle cells, and the different nerve fibres may directly innervate the smooth muscle cells.     

Regional functional heterogeneity of aortic smooth muscle cells in rats

The aorta is a magistral artery, which has been traditionally looked upon as a vessel whose properties are invariable throughout its length. However, in the most recent decade, there have been accumulated data that provide evidence that different aorta sections arise from different embryonic origins and that the population of smooth muscle cells making up the vessel’s wall is, consequently, heterogenic. Tracing the fate of smooth muscle cells, the basic components of the vessel, with the aid of genetic marking methods revealed that the cells’ response to various factors is largely determined by the embryonic origin of a certain cell population. However, functional differences between the smooth muscle cells making up different aorta sections remain poorly understood. The aim of the current work was to compare the functional characteristics of the populations of aortic wall smooth muscle cells obtained from the aorta sections differing by their embryonic origin. Towards this end, we obtained smooth muscle cell cultures from the three aorta sections of linear rats, namely, the neural crest derived ascending thoracic aorta, the somites derived descending thoracic aorta, and splanchnic mesoderm derived abdominal aorta. Using immunocytochemistry and Western blotting, the cells from the different regions of aorta were compared on the basis of smooth muscle actin, vimentin, and SM22 content in them. Cell proliferation rate was estimated using the growth curves method. We have demonstrated that the three smooth muscle cell populations arising from different embryonic origins differ in their morphological characteristics as well as by smooth muscle actin and SM22 content. We have shown that smooth muscle cells from the ascending aorta proliferate more actively than the corresponding cells from the descending thoracic aorta. Thus, the functional properties of the populations of rat aortic smooth muscle cells are different and depend on the embryonic origin of the aorta section from which they were obtained.     

Electron microscopic observations on the juxtamedullary efferent arterioles and Arteriolae rectae in kidneys of rats

Summary Efferent arterioles leaving juxtamedullary glomeruli in the kidneys of rats have a media comprized of a layer of closely packed smooth muscle cells. This muscle coat continues along the length of the efferent arterioles and arteriolae rectae to a level deep in the outer medullary zone, where smooth muscle cells are gradually replaced by pericytes characteristic of the non-muscular arterial vasa recta.Bundles of unmyelinated nerve fibers accompany the efferent arterioles and arteriolae rectae to the level where smooth muscle is no longer found in the media of the latter vessels. Close associations between smooth muscle cells and axons are marked by axonal dilatations which lie adjacent to muscle cells. There is no modification of either the axonal or the muscle cell membrane at these sites, nor do axons penetrate the basal lamina of muscle fibers. Large granular vesicles and small granular and agranular vesicles occur in most axons at the dilations, although the granular material in the small granular vesicles is usually sparse and in dispersed form.The nerves are considered to be primarily adrenergic because of strong catecholamine fluorescence demonstrated by other workers in association with the efferent arteries and arteriolae rectae. Poor definition of the small granular vesicles, which are commonly supposed to contain catecholamines, is ascribed to extraction of catecholamines during processing, discharge of granules prior to fixation, or inability of these axons to store catecholamines in quantity under physiological conditions.Financial assistance during the progress of this work was obtained from the Medical Research Council of Canada.     

VEGF promotes vascular sympathetic innervation

The sympathetic nervous system, via postganglionic innervation of blood vessels and the heart, is an important determinant of cardiovascular function. The mechanisms underlying sympathetic innervation of targets are not fully understood. This study tests the hypothesis that target-derived vascular endothelial growth factor (VEGF) promotes sympathetic innervation of blood vessels. Western blot and immunohistochemical analyses indicate that VEGF is produced by vascular cells in arteries and that VEGF receptors are expressed on sympathetic nerve fibers innervating arteries. In vitro, exogenously added VEGF and VEGF produced by vascular smooth muscle cells (VSMCs) in sympathetic neurovascular cocultures inhibited semaphorin 3A (Sema3A)-induced collapse of sympathetic growth cones. In the absence of Sema3A, VEGF and VSMCs also increased growth cone area. These effects were mediated via VEGF receptor 1. In vivo, the neutralization of VEGF inhibited the reinnervation of denervated femoral arteries. These data demonstrate that target-derived VEGF plays a previously unrecognized role in promoting the growth of sympathetic axons.     

Evidence for exclusive adrenergic innervation of feather muscles (mm. pennati) in the chicken

Summary Feather follicles in the avian skin are interconnected by well-defined bundles of smooth muscle cells, which are responsible for the erection and depression of feathers and thus play an important role in thermoregulation. The depressing and erecting muscle bundles were found to receive a very dense supply of unmyelinated nerve fibres that displayed ultrastructural and histochemical characteristics of noradrenergic axons (formaldehyde- and glyoxylic acid-induced catecholamine fluorescence; uptake to 5-hydroxydopamine). No nerve fibres were encountered showing histochemical acetylcholinesterase activity. There was no indication of the presence of peptidergic or purinergic nerve endings.The neuromuscular space usually ranged from 40–60 nm in width and contained a basal lamina. Occasionally, this space was reduced to approximately 20 nm. At such close neuromuscular contacts a basal lamina was lacking, and focal densities beneath the pre- and postsynaptic plasma membrane were observed. Since no gap junctions between muscle cells were detected, the dense supply with noradrenergic nerve fibres indicates a high amount of directly innervated smooth muscle cells.An additional finding of the present study was the observation that high local concentrations of 5-hydroxydopamine led to degeneration of noradrenergic nerve endings.Supported by a grant from the Deutsche Forschungsgemeinschaft (Dr. 91)     

Electron microscopic observations on the innervation of the sphincter of oddi in the dog

The ultrastructure and acetylcholinesterase activity of the intrinsic innervation of the sphincter of Oddi of eight adult dogs was studied by electron microscopy. A rich distribution of unmyelinated axons embedded individually or as groups within Schwann cell cytoplasm ("innervation fasciculee"), is to be observed. A few myelinated fibres were also observed. Many of the axons are acetylcholinesterase-positive. Three main types of nerve terminals are distinguished according to their vesicle populations. Individual nerve cells or small groups of nerve cells were scattered between the smooth muscle bundles and in the lamina glandularis mucosae. The cytoplasm of some neurons contains many electron dense spherical bodies resembling "myeloid bodies", and many lysosomes. Nerve terminals synapse onto both neuronal perikarya and their dendrites. Within the nerve fascicles, close appositions between the terminals occur frequently probably representing the most peripheral inter-neuronal integrative link in the neural regulation of the function of the sphincter of Oddi. -- The gap between nerve terminals and smooth muscle cells usually measures several thousands of A. Closer appositions are seldom seen, and no synaptic complexes can be observed.     

PDGF-D contributes to neointimal hyperplasia in rat model of vessel injury

In this study, we determined the role of PDGF-D, a new member of the PDGF family, in a rat model of balloon injured artery made with a 2F catheter in Sprague-Dawley male rats. PDGF-D expression was studied in the injured and control segments of abdominal aorta. The function of PDGF-D was evaluated in rat vascular smooth muscle cells stably transfected with PDGF-D gene. We found that in normal abdominal aorta, PDGF-D was highly expressed in adventia, moderate in endothelia, and unidentified in media. Stable transfection of PDGF-D gene into vascular smooth muscle cells increased the cell migration by 2.2-fold, and the proliferation by 2.3-fold, respectively, and MMP-2 production and activity as well. These results support the fact that PDGF-D is involved in the formation of neointimal hyperplasia induced by balloon catheter injury and may serve as a target in preventing vascular restenosis after coronary angioplasty.     

A comparative study on the metabolic activation of 3,4-benzo[a]pyrene to mutagens by aorta smooth muscle cells of rat and rabbit

The mutagenic activation of benzo[a]pyrene (BaP) after exposure to aorta smooth muscle cells of different origin was examined. Three test systems with different genetic endpoints--sister-chromatid exchange (SCE), gene mutation at the hypoxanthine guanine phosphoribosyl transferase (HGPRT) locus and unscheduled DNA synthesis (UDS)--were used. Treatment of rat and rabbit aorta smooth muscle cells with BaP (1-6 micrograms/ml) resulted in a significant increase of SCEs, HGPRT mutations and UDS. So smooth muscle cells are capable of converting BaP to metabolites with a DNA-damaging action. In order to investigate the relation between the formation of mutagenic BaP metabolites and the susceptibility to atherosclerosis we compared the mutagenic potential of BaP using aorta smooth muscle cells of different species (rat, rabbit) and locations (thoracic and abdominal aorta). Rabbits and abdominal aortas are more susceptible to atherosclerosis than rats and thoracic aortas. The SCE, HGPRT and UDS assays revealed that smooth muscle cells of different origin possessed the same metabolic potential towards BaP. There was no correlation between the mutagenic potency of BaP metabolites and the susceptibility to atherosclerosis. As smooth muscle cells have a low metabolic capacity towards BaP, probably other factors in addition to the metabolic capacity of smooth muscle cells are responsible for species and tissue differences in susceptibility to atherosclerosis.     

NEUROBIOLOGY OF ECHINODERMATA

During the past ten years much information has been added to our knowledge of nerve and muscle systems of echinoderms. 1. Electron-microscopy has shown that all the main nerve trunks consist of large numbers of small, parallel-running unmyelinated axons which are packed tightly together. Glial cells are generally absent. Discrete regions of neuropile are recognizable by the interweaving of axons, and the presence of vesicles. It has not yet been found possible to locate synapses with certainty in the nervous system, but it appears that they are chiefly confined to neuropile. The obvious nerve cords are massive accumulations of neurons which do not appear to interact locally. 2. Peripheral axons are difficult to distinguish because both interstitial and muscle cells have processes which often resemble axons. Ultrastructural analysis of this problem is aggravated by difficulties in fixation. However, the electron-microscope has shown that much of the echinoderm body wall contains a thick subepithelial plexus of processes from epithelial cells. Epithelial cells may thus act as sensory cells and supply axons to the plexus. 3. With the exception of striated muscles in some pedicellariae, all echinoderm muscles so far examined are of the smooth type. These muscles characteristically contain large filaments, and in this way do not resemble vertebrate smooth muscle. Some muscles are innervated by simple axonal contact, in others the muscles themselves send processes towards the nervous tissue. 4. Physiological studies of electrical activity in nerve and muscle systems have not added significantly to our knowledge of function. Several authors have demonstrated that massed electrical activity is conducted decrementally along the radial nerve cords, but this does not explain any known aspect of coordination. The only records of electrical activity from single neurons (Takahashi, 1964) have not been repeated. 5. There is strong evidence for two types of neurons in the central nervous system of echinoderms. One of these contains acetylcholine, the other dopamine and/or noradrenaline. Electron-microscopical histochemistry has given good indication that catecholamines are bound in echinoderm nerve tissue to particles similar to those reported in other invertebrate nervous tissues, and there is good evidence that acetylcholine is bound to synaptic vesicles which are morphologically identical to those present in the mammalian brain. The available data further indicate that acetylcholine is a transmitter in sensory and motor neurons, while dopamine and/or noradrenaline are transmitters in interneurons. Such interneurons may be involved in the coordination of the movement of the tube-feet. Other substances which have been implicated in neuro-effector mechanisms in other animal groups have not been found or are present in very small quantities. 6. Studies on the reproductive physiology of starfish have shown that several substances in the radial cords play important roles in its control. Such substances cannot at present be called neurosecretions because it is not known if they are derived from neurons. 7. Pharmacological studies on isolated muscle tissues have not added significantly to our knowledge of their control. The potency of ACh in causing contraction is well documented, and anticholinesterases are similar in effect. Catecholamines, although clearly very important in the nervous system, do not produce clear-cut effects. The published reports of relaxation to noradrenaline may well be due to direct effects on the muscle. No definite information has been obtained on the role of the adrenergic parts of the nervous systems of echinoderms, other than showing that they are not involved in motor responses. Extensive studies with a wide variety of drugs have produced inconsistent and largely negative results.