The following review articles are based on presentations given at a Symposium on Neural Crest Development, held in December 1996 at St George's Hospital Medical School, London. The Symposium formed part of the Winter Meeting of the Anatomical Society of Great Britain and Ireland. There were 6 invited speakers, 5 selected short presentations, and posters. The topics ranged widely, from autocrine regulation of Schwann cell development through clinical and animal disorders involving the melanocyte lineage, the heart and other organs, to the regulation of neural crest cell differentiation by cytokines.

Recent progress in human genetics and neurobiology has led to the identification of various mutations in particular myelin genes as the cause for many of the known inherited demyelinating peripheral neuropathies. Mutations in 3 distinct myelin genes, PMP22, P0, and connexin 32 cause the 3 major demyelinating subtypes of Charcot-Marie-Tooth (CMT) disease, CMT1A, CMT1B and CMTX, respectively. In addition, a reduction in the gene dosage of PMP22 causes hereditary neuropathy with liability to pressure palsies (HNPP), while particular point mutations in PMP22 and P0 cause the severe Dejerine-Sottas (DS) neuropathy. A series of spontaneous and genetically engineered rodent mutants for genes for the above-mentioned myelin constituents are now available and their suitability to serve as models for these still untreatable diseases is an issue of particular interest. The spontaneous mutants Trembler-J and Trembler, with point mutations in PMP22, reflect some of the pathological alterations seen in CMT1A and DS patients, respectively. Furthermore, engineered mutants that either over or underexpress particular myelin genes are suitable models for patients who are similarly compromised in the gene dosage of the corresponding genes. In addition, engineered mutants heterozygously or homozygously deficient in the myelin component P0 show the pathology of distinct CMT1B and DS patients, respectively, while Cx32 deficient mice develop pathological abnormalities similar to those of CMTX patients. Mutants that mimic human peripheral neuropathies might allow the development of strategies to alleviate the symptoms of the diseases, and help to define environmental risk factors for aggravation of the disease. In addition, such mutants might be instrumental in the development of strategies to cure the diseases by gene therapy.

The chondrocyte and its pericellular microenvironment together represent the chondron, historically considered the primary structural, functional and metabolic unit of articular and other hyaline cartilages. This review summarises research over the last 10 years to establish the molecular anatomy, functional properties and metabolic contribution of the chondron in articular cartilage homeostasis, and its failure during the initiation and progression of degenerative osteoarthritis.

Nitric oxide synthase immunoreactivity (bNOS-ir) was examined in the sacral spinal cord of the cat, macaque monkey and human using an antibody to the c-terminal region of neuronal NOS. In S2 of all 3 species NOS-ir was identified in both dorsal and ventral horns. In cat, monkey and human, bNOS-r occurred in sensory neurons of superficial laminae and the base of the dorsal horn, in autonomic neurons around the central canal and in the intermediolateral sacral spinal nucleus. In all 3 species, a large proportion of somatic motor nuclei in the ventromedial (VM), ventrolateral (VL) nuclei, and Onuf's nucleus (ON) showed high bNOS-ir, while others exhibited markedly lower immunoreactivity. Validatory experiments showed separate cellular localisation for bNOS, inducible NOS (iNOS), and endothelial NOS (eNOS) with only bNOS being localised to neuronal perikarya. Comparative morphometric analyses of the relative proportions and diameters of motor neurons in the VL, VM and ON exhibiting high and low levels of bNOS-ir revealed statistically significant differences in proportions in individual nuclei, and differences in size were generally not statistically significant. Finally, a comparison between cat sacral and thoracic spinal cord showed bNOS-ir in motor neurons of S2 was subject to less animal and rostrocaudal segment variation than in T10.

The accessibility of the primary sensory neurons of the trigeminal system at stages throughout their development in avian and mammalian embryos and the ease with which these neurons can be studied in vivo has facilitated investigation of several fundamental aspects of neurotrophin biology. Studies of the timing and sequence of action of neurotrophins and the expression of neurotrophins and their receptors in this well characterised neuronal system have led to a detailed understanding of the functions of neurotrophins in neuronal development. The concepts of neurotrophin independent survival, neurotrophin switching and neurotrophin cooperativity have largely arisen from work on the trigeminal system. Moreover, in vitro studies of trigeminal neurons provided some of the first evidence that the neurotrophin requirements of sensory neurons are related to sensory modality. The developing trigeminal system has been studied most extensively in mice and chickens, each of which has particular advantages for understanding different aspects of neurotrophin biology. In this review, I will outline these advantages and describe some of the main findings that have arisen from this work.

Intralobular lymphatics in the guinea pig pancreas were demonstrated enzyme-histochemically showing the extent, distribution and fine structure by combined light and transmission electron microscopy. 5′-nucleotidase(5′-Nase)-positive lymphatic vessels were present throughout the pancreas. Intralobular lymphatics among the acini were comparatively rare and generally independent of the blood capillaries, pancreatic ducts and acini. These lymphatics revealed the usual structural features, such as typical intercellular junctions and very tenuous vascular walls without continuous basal laminae. Fine precipitates of the cerium-based reaction product for 5′-Nase activity were found to be associated with cell membranes of the lymphatic endothelium and pinocytotic vesicles. Lymphatics were not closely related to the endocrine islets, although alkaline phosphatase(ALPase)-positive blood capillaries were well developed. Collecting lymphatic vessels with valves with weaker 5′-Nase activity were also detected in the interlobular connective tissue. ALPase activity, absent in the lymphatics, was positive in the blood capillaries, suggesting that it is also a useful way of demonstrating, histochemically, the blood capillaries in the guinea pig pancreas.

During development of the central nervous system (CNS) both deafferentation and axotomy induce increased neuronal death and result in a smaller brain with diminished function at maturity. Unilateral cerebellar pedunculotomy has been used as a model to study the relative importance of these 2 types of lesion on the survival of developing CNS neurons. Within the cerebellum, unilateral pedunculotomy causes deafferentation of the hemicerebellum and axotomy in the efferent pathway from the ipsilateral deep cerebellar nuclei. This results in a smaller hemicerebellum with normal cortical laminae but no extracerebellar outflow. In order to identify the sequence of events which leads to this altered structure and therefore to understand the relative importance of afferent versus target-derived trophic support, unilateral cerebellar pedunculotomy was performed on neonatal rat pups, aged between 1 and 3 days. The cerebella were analysed for histological and vascular changes after survival times of 0, 3, 6, 9, 12, 18, 21, 24 and 48 h. The results show that the effects of axotomy on the deep cerebellar nuclear neurons begin within 3 h of the lesion and apoptotic neuronal degeneration occurs within 48 h. However, the cerebellar cortical neurons continue to undergo normal histological development for at least 48 h after deafferentation. In addition, since ischaemia induces similar effects, a study of the vascular tree was made. The results indicate that the pedunculotomy does not alter the blood supply to the cerebellum, nor induce ischaemia of the cerebellar neurons. From this it may be hypothesised that target-derived trophic support is more crucial for the survival of immature neurons than is the trophic effect of afferent input.

There is evidence for lymphatic drainage of interstitial fluid from the brain along perivascular spaces in a number of mammalian species. Ultrastructural studies suggest that there are similar drainage pathways in the human cerebral cortex. Perivascular spaces in the basal ganglia, however, differ from those in the cortex in that they dilate to form lacunes and rarely accumulate beta-amyloid (amyloid angiopathy) in Alzheimer's disease; in the cortex, lacunes are rare but amyloid angiopathy is common. The aim of the present study is to compare the structure of perivascular spaces in the basal ganglia and at the anterior perforated substance with perivascular spaces in the cerebral cortex. Eight postmortem brains from patients aged 23–80 years (mean 68 y) were examined by light microscopy, by scanning and transmission electron microscopy and by direct visualisation of etched paraffin blocks. The results show that arteries in the basal ganglia are surrounded by 2 distinct coats of leptomeninges separated by a perivascular space which is continuous with the perivascular space around arteries in the subarachnoid space. The inner layer of leptomeninges closely invests the adventitia of the vessel wall and the outer layer is continuous with the pia mater on the surface of the brain at the anterior perforated substance. Veins in the basal ganglia have no outer layer of leptomeninges and thus the perivascular space is continuous with the subpial space. The anatomy of the periarterial spaces in the basal ganglia differs significantly from that in the cerebral cortex where there is only a single periarterial layer of leptomeninges. Differences in structure of perivascular spaces around arteries may reflect relative efficiencies in the drainage of interstitial fluid from different sites in the brain. Futhermore, the structure of the perivascular spaces may contribute to the relatively high frequency of lacunes in the basal ganglia, and the low frequency of amyloid angiopathy at this site in Alzheimer's disease.

At the onset of their migration into the embryo, many neural crest cells are pluripotent in the sense that they have the capacity to generate progeny that consist of more than one cell type. More recently, we have found that there are pluripotent neural crest cell-derived cells even at sites of terminal differentiation. These findings support the notion that cues originating from the microenvironment, at least in part, direct neural crest cell type specification. Based on the rationale that growth factors that are known to support survival of neural crest cell derivatives may have additional functions in progenitor cell development, we have examined the action of pertinent growth factors. Trophic, mitogenic, antiproliferative and differentiation promoting activities were found. Stem cell factor (SCF) is trophic for pluripotent neural crest cells. Contrary to expectation, SCF plus a neurotrophin, rather than SCF alone, is trophic for committed melanogenic cells. Basic fibroblast growth factor (bFGF) is mitogenic both for pluripotent cells and committed melanogenic cells. However, the cells become dependent on another factor for survival. Whereas any neurotrophin tested can rescue bFGF-activated pluripotent neural crest cells, the factor that rescues melanogenic cells remains to be determined. Transforming growth factor β1 (TGF-β1) is a powerful antimitotic signal for all neural crest cells that overrides the bFGF/neurotrophin proliferative signal. Furthermore, SCF promotes differentiation of neural crest cells into cells of the sensory neuron lineage. Neurotrophin-3 (NT-3) specifically promotes high affinity uptake of norepinephrine by neural crest cells and is thus thought to play a critical role in the differentiation of sympathetic neuroblasts. In summary, our data indicate that neurotrophins and other pertinent growth factors affect survival, proliferation and differentiation of neural crest cells at multiple levels and in different lineages. Moreover, our findings emphasise the importance of the concerted action of combinations of growth factors, rather than of individual factors.

Investiture of new microvessels within an injured peripheral nerve trunk may determine the success that the local environment has in promoting axonal sprouting and regeneration. We therefore examined microvessel investment of 24 h–14 d proximal nerve stump preparations in rat sciatic nerves. The stumps, later destined to form neuromas, were created by sciatic nerve transection with resection of distal branches to prevent distal reinnervation. Microvessels were studied in the proximal stump in semithin whole mount sections of nerve and by analysis of India ink perfused microvessel profiles. Quantitative image analysis was made of the luminal profiles of vessels perfused with India ink from unfixed sections of the stumps, contralateral uninjured nerves and sham-exposed but uninjured nerves. Evidence of angiogenesis was observed in stumps 7 d after transection, indicated by a rise in the total numbers of perfused microvessels and in the numbers of 2–6 μm diameter perfused microvessels. There was a shift in the histogram of the percentage of perfused microvessels towards the 2–4 μm range and a reduction in the mean microvessel luminal area in the stumps. By 14 d, new microvessels were larger, indicated by an increase in total luminal area. New microvessels were prominent in the epineurial connective tissue or between layers of perineurial cells of former fascicles. Microvessels probably share a battery of trophic signals with other proliferating cellular elements in the milieu of the injured peripheral nerve trunk.

Computerised morphometry and a double stain technique were utilised to examine the corneal nerves in whole mounts. This novel stain combines the nonspecific acetylcholinesterase (NsAchE) and gold chloride (AuCl) procedures to enhance staining contrast and facilitate computerised detection of corneal nerves. Fresh rat corneas were dissected, and the Descemet's membrane-endothelium complex was removed. Then the corneas were fixed in 4% paraformaldehyde with 50 mm Na-K phosphate buffer (pH 7.2) and 8% sucrose for 30 min. They were rinsed and stained singly with NsAchE or AuCl, or were double stained using NsAchE followed by AuCl. Between NsAchE and AuCl staining the corneas were stored frozen in OCT compound at −70°C. Flat mounts of whole corneas were photographed before and after the second staining. Measurable stromal innervation density (mean±S.D.) in age-matched corneas stained with AuCl (3.90±0.36 mm/mm2) was not significantly different from that of NsAchE stained corneas. However, double staining compared with NsAchE staining of the same corneas revealed a 48±27% increase in demonstrable innervation density of the subepithelial nerve plexus (7.95±0.86 mm/mm2 vs 5.52±1.31 mm/mm2, respectively). Improved visualisation of epithelial nerves and their fine ramifications (leashes) was also obtained by double staining. This novel combination of 2 procedures enhances the detection of corneal nerves for analysis by computerised morphometry and provides a more representative estimate of total corneal innervation density.

The encephalomyocarditis virus of the diabetogenic M-strain (EMC-M) is known to cause diabetes in mice. The EMC-M virus has also been shown to cause paresis in some of the infected animals. The clinical features include an acute ascending predominantly motor paralysis, developing within days. This resembles acute idiopathic polyneuritis. The alpha motor neurons would be a possible target for the virus, so two parameters, the total number and the size distribution of motor neurons, were therefore selected for further investigation in 6 mice with neurological involvement and compared with 6 control mice. The optical fractionator method was applied for estimating the total number of motor neurons and the 3D size distribution was estimated using the rotator method in a vertical design. No difference was found in the total number of motor neurons and the size distributions were similar in the 2 groups. This design can be used as a model for the estimation of the total number of motor neurons and their size distribution in other experimental animal models.

In the absence of a systematic morphometric study of Onuf's nucleus in man, this investigation defines the limits of variation of segmental position and the range of length and volume of Onuf's nucleus in 6 normal humans displaying no neurological disease (2 males, 4 females). Serial section reconstruction methods in conjunction with the disector method provided information on the numbers, sizes and shapes of the constituent motor neurons of Onuf's nucleus. In contrast to previous descriptions, the cranial origin of Onuf's nucleus occurred in rostral S1 in 50% of subjects, and midcaudal S1 in the remaining subjects. Onuf's nucleus varied in length between 4 and 7 mm, and was 0.2–0.37 mm3 in volume. Differences in length or volume between males or females, or between the left and right side of the cord were not statistically significant. Neurons in Onuf's nucleus varied in diameter between 10 μm and 60 μm (mean 26 μm) and their mean number was 625±137. A higher density of neurons occurred at the cranial and caudal ends of the nucleus relative to the middle. While 37% of neurons were approximately spherical (shape index ∼1), 44% were ellipsoid and 19% fusiform (shape indices varying between 0.26 and 0.8). These findings are compared with previous studies of Onuf's nucleus in man and animals. The results form a basis for further studies on Onuf's nucleus in normality and neurodegenerative diseases.

A 14 d ovarian primordium was transplanted with a fetal testis (13–18 d and 21 d of gestation) or a neonatal testis (15, 20, 30 and 45 d after birth) into the renal subcapsular position of an adult male rat. Two weeks after transplantation, transplants were examined as to the degree of ovarian and testicular differentiation. In the combination of a 14 d ovary and a 13 d testis, there were 3 types of result: either the ovary or the testis alone developed or both gonads developed well. Ovaries transplanted in union with 15–18 d testes did not develop, although the testes developed normally. Some ovaries in union with 21 d testes developed normally. In combination with infantile testes, the incidence of developed ovaries increased as the age of testes advanced. These results suggest that the 13 d fetal testes begin to suppress the development of cotransplanted 14 d ovaries, that 14–18 d fetal testes maintain such suppressive effects and that this effect gradually diminishes in infantile testes as they progress toward 45 d after birth.

The cellular events leading to the generation of Schwann cells from the neural crest have recently been clarified and it is now possible to outline a relatively simple model of the Schwann cell lineage in the rat and mouse. Neural crest cells have to undergo 3 main developmental transitions to become mature Schwann cells. These are the formation of Schwann cell precursors from crest cells, the formation of immature Schwann cells from precursors and, lastly, the postnatal and reversible generation of non-myelin- and myelin-forming Schwann cells. Axonal signals involving neuregulins are important regulators of these events, in particular of the survival, proliferation and differentiation of Schwann cell precursors.

Cultures of rat thymic epithelium were used to measure the effect of thymulin secretagogues on dye-coupling and proliferation. Dye-coupling was assessed after the injection of lucifer yellow dextran which cannot permeate the connexin pore of gap junctions and the smaller, permeant cascade blue. In addition to gap junctional communication, larger intercellular bridges were demonstrated by the transfer of lucifer yellow dextran between cells. The extent of intercellular communication was found to be influenced by both cell density and the number of passages. In control cultures, intercellular communication was reduced in cell groups of low (<20 cells/group) or high cell densities (>100 cells/group) compared with groups of 20–60 cells. The highest coupling indices were found in subcultures 20–30. Taking these factors into account, significant decreases in coupling index were observed after pretreatment of test cultures with factors known to influence the secretion of thymulin (5 U/ml interleukin 1 (α and β), 1 μM progesterone, 1 μM oestrogen, 1 μM testosterone, 1 ng/ml adrenocorticotropic hormone, 100 nm rat growth hormone) but 7.5 ng/ml thymulin had no effect on dye-coupling. The nonspecific gap junction uncoupler, octanol, abolished dye-coupling. Cellular proliferation, as measured by the uptake of tritiated thymidine, showed that the same factors that reduced coupling also increased proliferation. None of these factors affected the number of multinucleate cells present, except interleukin-1β which caused a significant reduction in the average number of nuclei per cell. Thus rat thymic epithelium in vitro provides a model for the study of the direct action of factors on cells of the thymic microenvironment.

Development of the neural crest involves a remarkable feat of coordinated cell migration in which cells detach from the neural tube, take varying routes of migration through the embryonic tissues and then differentiate at the end of their journey to participate in the formation of a number of organ systems. In general, neural crest cells appear to migrate without the guidance of long-range physical or chemical cues, but rather they respond to heterogeneity in the extracellular matrix that forms their migration substrate. Molecules such as fibronectin and laminin act as permissive substrate components, encouraging neural crest cell attachment and spreading, whereas chondroitin sulphate proteoglycans are nonpermissive for migration. A balance between permissive and nonpermissive substrate components seems to be necessary to ensure successful migration, as indicated by a number of studies in mouse mutant systems where nonpermissive molecules are over-expressed, leading to inhibition of neural crest migration. The neural crest expresses cell surface receptors that permit interaction with the extracellular matrix and may also modify the matrix by secretion of proteases. Thus the principles that govern the complex migration of neural crest cells are beginning to emerge.

Connective tissues synthesise and secrete a family of matrix metalloproteinases (MMPs) which are capable of degrading most components of the extracellular matrix. Animal studies suggest that the MMPs play a role in bone turnover. Using specific polyclonal antisera, immunohistochemistry was used to determine the patterns of synthesis and distribution of collagenase (MMP-1), stromelysin (MMP-3), gelatinase A (MMP-2) and gelatinase B (MMP-9) and of the tissue inhibitor of metalloproteinases-1 (TIMP-1) within developing human osteophytic bone. The different MMPs and TIMP showed distinct patterns of localisation. Collagenase expression was seen at sites of vascular invasion, in osteoblasts synthesising new matrix and in some osteoclasts at sites of resorption. Chondrocytes demonstrated variable levels of collagenase and stromelysin expression throughout the proliferative and hypertrophic regions, stromelysin showing both cell-associated and strong matrix staining. Intense gelatinase B expression was observed at sites of bone resorption in osteoclasts and mononuclear cells. Gelatinase A was only weakly expressed in the fibrocartilage adjacent to areas of endochondral ossification. There was widespread but variable expression of TIMP-1 throughout the fibrous tissue, cartilage and bone. These results indicate that MMPs play a role in the development of human bone from cartilage and fibrous tissue and are likely to have multiple functions.

Changes in the 3-dimensional arrangement of the articular cartilage matrix during growth of the rabbit tibial plateau were studied. Knees from newborn, and 1, 2 and 6 wk-old rabbits were compared with those of adults by light and electron microscopy. The specimens were fixed, embedded en bloc in epoxy resin and sectioned vertically/coronally through the point where the articular cartilage was thickest in the adult medial tibial plateau. At birth, the proximal tibial epiphysis was cartilaginous, but nascent articular cartilage was recognisable as a densely cellular layer covering the tibial condyle. Within 30 μm of the articular surface, the chondrocytes were flattened and collagen fibres ran among these cells in a direction parallel to the surface. Deeper in the articular cartilage, rounded cells were evenly distributed within a random collagen fibril network. At the centre of the plateau, the tangential layer changed little during growth, whereas the subjacent cellular layer grew in thickness and steadily achieved a more vertical character in the organisation of its constituent collagen and cellular elements. At 1 wk, cells were separated into clusters by acellular regions filled with collagen fibrils. At 2 wk, cells within the forming radial zone were aligned in columns bracketed by vertical collagen fibres. Continuity of these vertical fibres with those in the tangential surface layer was evident at this age. The chondrocytes were surrounded by fibrous capsules typical of chondrons. By 6 wk, the bases of the radial collagen fibres in the very centre of the condyle had calcified, as had the adjacent hypertrophic hyaline cartilage. A solid subchondral plate and tidemark did not appear until skeletal maturity. From birth to age 6 wk, maximum thickness of the layer identified as primordial articular cartilage increased from 0.13 mm to 0.70 mm, and was 1.5 mm in the adult. Throughout growth, however, the thickness of the tangential layer in the centre of the plateau never exceeded 0.05 μm. In the patella, femoral head and peripheral tibial plateau, cartilage development followed the same general sequence. In contrast to the central tibial plateau, the tangential layer also grew in thickness, but at a slower rate than that of the radial zone. At all ages, the developing articular cartilage was structurally distinct from the deeper hyaline cartilage which contributed to growth of the ossification centre through enchondral ossification. The collagen matrix of articular cartilage acquires a characteristic, orderly 3-dimensional structure soon after birth. Growth in cartilage thickness occurs primarily through enlargement of the radial zone.

Methods are described for the semiquantitative analysis of the connective tissue components of human peripheral nerve using light microscopy. General histological preservation was assessed using haematoxylin and eosin staining and the distribution of collagen type IV was investigated using immunohistochemistry. Several techniques were investigated to establish the one giving optimum structural preservation, immunobinding and greatest contrast for image analysis. Frozen sections were unsuitable for this tissue and paraffin wax sections were therefore used. Alcohol fixation was rejected due to poor preservation of the endoneurium, although immunobinding was excellent. Ice-cold formalin fixation for 24 h was found to be adequate for structural preservation and antibody binding, provided that a protease step was introduced. Trypsin was found to be superior to pepsin for exposing collagen type IV epitopes. Of the detection systems investigated indirect immunofluorescence was not suitable due to considerable autofluorescence of the nerve. The avidin-biotin method provided the greatest contrast, and was therefore the detection method of choice for image analysis. The optimum techniques for image analysis were then used on control human sural nerve to ascertain the best comparative method for collagen type IV in the perineurium. A method of semiquantitative analysis is described which takes into account the fact that there is a close linear relationship between collagen content per unit of perineurium and perineurial perimeter as fascicle size increases in peripheral nerve. This means that data from 2 different sample groups can easily be compared, provided that a range of fascicle sizes is analysed in each case.