The following review articles are based on presentations given at a Symposium on Computer Modelling for Anatomists and Clinicians held in July 1997 at the Department of Anatomy, University of Edinburgh. The Symposium formed part of the Summer Meeting of the Anatomical Society of Great Britain and Ireland. There were 6 invited speakers, and 4 of these have produced review articles which are published in this issue of the Journal. In 3 of the presentations, recent advances in the use of 3D methodology to analyse mouse and human developmental anatomy were described. The reviews by Kaufman et al. and Baldock et al. (the latter to be published in a forthcoming issue of the Journal) complement each other in that they describe different aspects of a collaborative project (the Mouse Atlas Project, or MAP) involving the preparation of a text database of mouse developmental anatomy and a digital Atlas of normal mouse development. In the first of these presentations the principles underlying the methodology employed to reconstruct a range of early post-implantation mouse embryos spanning the stages between the primitive streak and the early limb-bud that had previously been serially sectioned was briefly described. Because of the efficiency of the computer technology involved (termed the warping program), the grey level images of the embryo histology could be arbitrarily resectioned to view the embryo as if sectioned in any orientation. Furthermore, all anatomically-defined tissues could be individually delineated (or painted) and viewed either in isolation or in combination with other appropriately labelled tissues and organs.

Where tendons and ligaments are subject to compression, they are frequently fibrocartilaginous. This occurs at 2 principal sites: where tendons (and sometimes ligaments) wrap around bony or fibrous pulleys, and in the region where they attach to bone, i.e. at their entheses. Wrap-around tendons are most characteristic of the limbs and are commonly wider at their point of bony contact so that the pressure is reduced. The most fibrocartilaginous tendons are heavily loaded and permanently bent around their pulleys. There is often pronounced interweaving of collagen fibres that prevents the tendons from splaying apart under compression. The fibrocartilage can be located within fascicles, or in endo- or epitenon (where it may protect blood vessels from compression or allow fascicles to slide). Fibrocartilage cells are commonly packed with intermediate filaments which could be involved in transducing mechanical load. The ECM often contains aggrecan which allows the tendon to imbibe water and withstand compression. Type II collagen may also be present, particularly in tendons that are heavily loaded. Fibrocartilage is a dynamic tissue that disappears when the tendons are rerouted surgically and can be maintained in vitro when discs of tendon are compressed. Finite element analyses provide a good correlation between its distribution and levels of compressive stress, but at some locations fibrocartilage is a sign of pathology. Enthesis fibrocartilage is most typical of tendons or ligaments that attach to the epiphyses of long bones where it may also be accompanied by sesamoid and periosteal fibrocartilages. It is characteristic of sites where the angle of attachment changes throughout the range of joint movement and it reduces wear and tear by dissipating stress concentration at the bony interface. There is a good correlation between the distribution of fibrocartilage within an enthesis and the levels of compressive stress. The complex interlocking between calcified fibrocartilage and bone contributes to the mechanical strength of the enthesis and cartilage-like molecules (e.g. aggrecan and type II collagen) in the ECM contribute to its ability to withstand compression. Pathological changes are common and are known as enthesopathies.

The menisci and their insertions into bone (entheses) represent a functional unit. Thanks to their firm entheses, the menisci are able to distribute loads and therefore reduce the stresses on the tibia, a function which is regarded essential for cartilage protection and prevention of osteoarthrosis. The tissue of the hypocellular meniscal body consists mainly of water and a dense elaborate type I collagen network with a predominantly circumferential alignment. The content of different collagens, proteoglycans and nonproteoglycan proteins shows significant regional variations probably reflecting functional adaptation. The meniscal horns are attached via meniscal insertional ligaments mainly to tibial bone. At the enthesis, the fibres of the insertional ligaments attach to bone via uncalcified and calcified fibrocartilages. This anatomical configuration of gradual transition from soft to hard tissue, which is identical to other ligament entheses, is certainly essential for normal mechanical function and probably protects this vulnerable transition between 2 biomechanically different tissues from failure. Clinical treatment of meniscal tears needs to be based on these special anatomical and functional characteristics. Partial meniscectomy will preserve some of the load distribution function of the meniscus only when the meniscal body enthesis entity is preserved. Repair of peripheral longitudinal tears will heal and probably preserve the load distribution function of the meniscus, whereas radial tears through the whole meniscal periphery or more central and complex tears may be induced to heal, but probably do not preserve the load distribution function. There is no proof that replacement of the meniscus with an allograft can reestablish some of the important meniscal functions, and thereby prevent or reduce the development of osteoarthrosis which is common after meniscectomy. After implantation, major problems are the remodelling of the graft to inferior structural, biochemical and mechanical properties and its insufficient fixation to bone which fails to duplicate a normal anatomical configuration and therefore a functional meniscal enthesis.

Laminins are a family of multifunctional macromolecules, ubiquitous in basement membranes, and represent the most abundant structural noncollagenous glycoproteins of these highly specialised extracellular matrices. Their discovery started with the difficult task of isolating molecules produced by cultivated cells or extracted from tissues. The development of molecular biology techniques has facilitated and accelerated the identification and the characterisation of new laminin variants making it feasible to identify full-length polypeptides which have not been purified. Further, genetically engineered laminin fragments can be generated for studies of their structure-function relationship, permitting the demonstration that laminins are involved in multiple interactions with themselves, with other components of the basal lamina, and with cells. It endows laminins with a central role in the formation, the architecture, and the stability of basement membranes. In addition, laminins may both separate and connect different tissues, i.e. the parenchymal and the interstitial connective tissues. Laminins also provide adjacent cells with a mechanical scaffold and biological information either directly by interacting with cell surface components, or indirectly by trapping growth factors. In doing so they trigger and control cellular functions. Recently, the structural and biological diversity of the laminins has started to be elucidated by gene targeting and by the identification of laminin defects in acquired or inherited human diseases. The consequent phenotypes highlight the pivotal role of laminins in determining heterogeneity in basement membrane functions.

The skin structure of 2 Bathyergid rodents, the naked mole-rat (Heterocephalus glaber) and the common mole-rat (Cryptomys hottentotus) is compared, to investigate whether thermoregulatory differences may be attributed to different skin features. Histological and ultrastructural studies of the dorsal skin of these closely related species show morphological and structural similarities but differences in the degree of skin folding, thickness of the integument and dermal infrastructure were evident. The skin of the common mole-rat conforms with expected morphological/histological arrangements that are commonly found in mammalian skin. Many features of the skin of the naked mole-rat, such as the lack of an insulating layer and the loosely folded morphological arrangement contribute to poikilothermic responses to changing temperatures of this mammal. Further evidence for poikilothermy in the naked mole-rat is indicated by the presence of pigment containing cells in the dermis, rather than the epidermis, as commonly occurs in homeotherms. Lack of fur is compensated by a thicker epidermal layer and a marked reduction in sweat glands. Differences in skin morphology thus contribute substantially to the different thermoregulatory abilities of the 2 Bathyergids. The skin morphology is related to the poor thermoinsulatory ability of the animals while simultaneously facilitating heat transfer from the environment to the animal by thigmothermy and/or other behavioural means.

We have been involved with a group of computer scientists and anatomists in the development of computer-based methodologies that not only combine the advantages of scanning electron microscopy and conventional histology, but provide the additional dimension of tissue recognition. The latter is achieved by the appropriate labelling of tissues and structures by delineation or ‘painting’. Individually segmented anatomically defined tissues can be highlighted in a particular colour and viewed either in isolation or in combination with other appropriately labelled tissues and organs. Tissues can be shown in any orientation either as a transparent overlay on computer-generated histological sections or as 3-D images without the histological background. An additional feature of the system is that computer graphics technology combined with 3-D glasses now also allows the viewer to see the object under analysis in stereo. This facility has been found to be particularly helpful in drawing attention to topological relationships that had not previously been readily noted. As the mouse is now the mammalian model of choice in many areas of developmental research, it is of critical importance that a basic level of skill is available in the research community in the interpretation of serially sectioned material, for example, for the rapidly expanding field in which gene expression studies play a significant role. It is equally important that there is an understanding of the dynamic changes that occur in relation to the differentiation of the various organ systems seen in these early stages of development. What we emphasise here is the additional information that it is possible to gain from the use of this tool which, in our view, could not readily have been gained from the analysis of scanning electron micrographs or by studying conventional serial histological sections of similar stages of mouse embryonic development. The methodology has been developed as part of a large project to prepare a database of mouse developmental anatomy covering all stages from fertilisation to birth in order to allow the accurate spatial mapping of gene expression and cell lineage data onto the digital Atlas of normal mouse development. In this paper we show how this digital anatomical Atlas also represents a valuable teaching aid and research tool in anatomy.

Experimental animal models are frequently used to study articular cartilage, but the relevance to man remains problematic. In this study animal models were compared by examination of the collagen fibre arrangement in the medial tibial plateau of human, cow, pig, dog, sheep, rabbit and rat specimens. 24 cartilage samples from each species were prepared and maximum cartilage thickness in the central tibial plateau measured. Samples were fixed, dehydrated, freeze-fractured and imaged by scanning electron microscopy (SEM). At low magnification, 2 different arrangements of collagen fibres were observed: leaf-like (human, pig, dog) and columnar (cow, sheep, rabbit, rat). The porcine collagen structure was the most similar to that of man. This arrangement was consistent from the radial to the upper zones. Under higher magnification at the surface of the leaves, the collagen was more randomly oriented, whereas the columns consisted of parallel collagen fibrils. The maximum thickness of cartilage did not correlate with the type of collagen arrangement but was correlated with the body weight of the species (r=0.785). When using animal models for investigating human articular cartilage function or pathology, the differences in arrangement of collagen fibres in tibial plateau cartilage between laboratory animals should be considered especially if morphological evaluation is planned.

En bloc staining with basic fuchsin is an established method for demonstrating microdamage in bone. Using transmitted light microscopy, variations in light intensity, depth of focus and magnification are necessary to distinguish fully-stained microcracks generated in vivo, from partially-stained or unstained artefactual cracks due to cutting and machining. This process is both difficult and time-consuming. In this study, 2 methods were used to examine fuchsin-stained microcracks in human rib sections, transmitted light and epifluorescence microscopy. No differences were found in crack number, density or length between the 2 methods indicating comparable accuracy. Using green epifluorescence, only microcracks containing fuchsin fluoresced orange against the darkfield background, enabling unstained, artefactual cracks to be screened out. Under UV epifluorescence, microcracks stained through the full 100 μm depth of the section fluoresced purple. Partially-stained artefactual cracks failed to fluoresce and were screened out. Epifluorescence is a simple, rapid and accurate screening method for differentiating fully-stained from artefactual microcracks in bone.

The aim of this study was to determine the degree to which the epidermis and oral epithelium of species other than man express cytokeratin (CK) intermediate filaments, which are markers of epithelial differentiation. Fixed, wax-embedded samples of skin, buccal mucosa and gingiva from rhesus monkey, marmoset, cow, sheep, pig, ferret, hamster, axolotl and trout were tested for CK expression using a panel of antihuman CK antibodies and an immunoperoxidase procedure. Human skin and oral mucosa were also stained to act as positive control. The results showed that antihuman CK antibodies stained animal tissues, but the patterns of staining were not always identical to the established human CK profile. Of particular interest was the expression of CK18, typically only detected in ‘simple’ epithelium in man, in bovine, ferret and hamster stratified epithelium from different sites. However, there was evidence of variable anti-CK antibody cross-reactivity, both as a result of intrinsic variations in CK polypeptide structure and as artifacts of fixation. We conclude that some CK are conserved between species, but that biological variables, for example local functional requirements, and technical factors affect the results. These considerations need to be borne in mind in animal studies of epithelial differentiation employing CK immunohistochemistry. Biochemical characterisation is ultimately necessary to determine specific differences between human and animal CK.

Careful histological observation of the development of the anlage of the inner ear in chicken embryos led us to question the traditional view of otic placode (OP) formation. First, morphological studies in the cephalic region carried out on stages preceding the appearance of the placodal epithelium revealed that the medial placodal cells are continuous temporally and spatially with cells belonging to the neural fold (NF). Second, both the formation of the basal lamina between the dorsal region of the neural tube (NT) and ectoderm and the pattern of formation of the neural crest present distinctive characteristics between otic levels and regions located anteriorly and posteriorly. Third, numerical comparisons of parameters for the NT and the OP between different levels of the rhombencephalon allowed us to assign a differential behaviour in the growth pattern of the otic region. These results indicated that the medial part of the OP is not derived from already independent ectoderm that increases in thickness under the influence of the NT (as previously accepted) but that it develops directly from the NFs. Although we do not exclude other possibilities, we propose that at least a proportion of the OP cells originate directly from cells committed to be neural crest. After this incorporation, basal laminal formation would delimit the NT from the OP without transition of the otic cells to ectoderm. This hypothesis would imply that part of the otic cells originate directly from neuroepithelial cells having a neuroectodermal (rather than the previously established ectodermal) origin.

Time-lapse photography was used to examine the effects of γ-aminobutyric acid (GABA) on the outgrowth and motility of neurites in cultures from mouse spinal cord. GABA at concentrations of 100, 10 and 1 μm caused significant inhibition of neurite outgrowth and the motility of growth cones was significantly reduced by treatment with 100 and 10 μm GABA. This effect was mimicked by the GABAB receptor agonist baclofen, whereas the GABAA receptor agonist muscimol had no effect. The effect of GABA on outgrowth and motility seems to be dependent on the type of serum employed. The results reported here were obtained only when heat-inactivated serum was used and not when non heat-inactivated serum was added to the culture medium. They suggest that GABA has a role in the regulation of process outgrowth within the embryonic mouse spinal cord.

Despite the fact that development of the human embryo heart is of considerable clinical importance, there is still disagreement over the process and the timing of events. It is likely that some of the conflicting accounts may have arisen from difficulties in describing and visualising 3-dimensional structures from 2-dimensional sections. To help overcome this problem and to improve our understanding of the development of the heart, we have devised techniques for the production of interactive 3D models reconstructed from serial histological sections of human embryos. Our method uses commercial software designed for the creation of 3D models and virtual reality environments. The ability to construct interactive visual images which both illustrate and communicate complex 3D information contributes to our understanding of the complex developmental changes occurring in embryogenesis.

The reactive formation of lymphoid follicles and germinal centres in lymph nodes, induced by subcutaneous transfer of in vitro activated splenic adherent cells into syngeneic mice, were studied. Adherent cells were obtained by incubating spleen cell suspensions for 24 h and activated by incubating for 1 h in the medium containing keyhole limpet haemocyanin (KLH) absorbed onto alumina. Some of the treated adherent cells were irradiated with 10 Gy x-rays, while others were either not stimulated or were stimulated with alumina-KLH but killed by repeated freezing and thawing. Examination of adherent cell smears immunostained with antibodies against, F4/80, Mac-1, Mac-2 and NLDC-145 indicated that many adherent cells displayed macrophage markers but few displayed the interdigitating cell marker. Animals transfused with KLH-treated adherent cells with or without irradiation showed a marked increase in the number of lymphoid follicles and germinal centres in draining lymph nodes, whereas those transfused with adherent cells which had not been KLH-treated or which had been killed after KLH treatment displayed no significant change in the number of follicles. These results were interpreted as indicating that following transfusion, antigen-activated adherent macrophages migrated into the draining lymph nodes and induced the reactive formation of lymphoid follicles and germinal centres outside preexisting follicles.

The axon determines whether or not it is myelinated by the Schwann cell. At maturity there is a positive correlation between sheath thickness and axon calibre. This correlation is initially very low or absent, but gradually strengthens during development. This increase could come about because the axon continuously controls Schwann cell myelinating activity, so that a given axon calibre is associated with a particular myelin sheath thickness, an interaction which would entail the Schwann cell continuously monitoring and responding to axon size. This seems unnecessarily complex. This theoretical study shows that the strong correlation between the 2 parameters within a given myelinated fibre population may come about in a much simpler way than outlined above. This is demonstrated by modelling the growth and myelination of a hypothetical population, utilising data from earlier studies on cervical ventral motoneuron axon development. The hypothesis tested shows that the only instructive interactions by the axon on the Schwann cell necessary for the strong correlation between the 2 parameters to emerge are for the initiation of myelination, its continuation and its termination. These could result from a single stimulus being switched on, persisting for a time and being switched off. Under this influence, the Schwann cell is assumed to proceed to form the myelin sheath at a constant rate which it itself inherently determines, in the absence of any quantitative influence exerted by the axon. This continues until the stimulus for myelination ceases to emanate from the axon. The validity of the hypothesis is demonstrated, because the resulting myelin-axon relationships correspond closely to those observed during development.

This paper reviews recent work in radiological image registration and provides a classification of image registration by type of transformation and by methods employed to compute the transformation. The former includes transformation of 2D images to 2D images of the same individual, transformation of 3D images to 3D images of the same individual, transformation of images to an atlas or model, transformation of images acquired from a number of individuals, transformations for image guided interventions including 2D to 3D registration and finally tissue deformation in image guided interventions. Recent work on computing transformations for registration using corresponding landmark based registration, surface based registration and voxel similarity measures, including entropy based measures, are reviewed and compared. Recently fully automated algorithms based on voxel similarity measures and, in particular, mutual information have been shown to be accurate and robust at registering images of the head when the rigid body assumption is valid. Two approaches to modelling soft tissue deformation for applications in image guided interventions are described. Validation of complex processing tasks such as image registration is vital if these algorithms are to be used in clinical practice. Three alternative validation strategies are presented. These methods are finding application outside the original domain of radiological imaging.

Sympathetic nerves to the enlarged fetus-containing region of the uterus undergo degenerative changes during late pregnancy and show slow regrowth after parturition. It is not known whether this unusual response of sympathetic nerves to smooth muscle hypertrophy is due to the sensitivity of short adrenergic neurons to hormonal changes, or whether the nerves respond to changes in the neurotrophic capacity of the target. We have investigated this question using in oculo transplantation. Small pieces of myometrium from the uterine horn of virgin guinea pigs, or from the region previously occupied by the placenta and fetus in postpartum guinea pigs, were transplanted into the anterior eye chamber. After 3 wk in oculo, the pattern of reinnervation of the transplants was assessed on whole mount stretch preparations stained for tyrosine hydroxylase. The histology of the transplants was examined in toluidine blue-stained semithin sections. Myometrial transplants from virgin donors and uterine artery transplants from both virgin and postpartum donors became organotypically reinnervated by sympathetic fibres from the host iris. In contrast, sympathetic nerves did not reinnervate myometrial transplants from postpartum donors, although they approached the transplants and became distributed in the surrounding connective tissue. All transplanted tissues showed a normal histological appearance. Both the myometrium and uterine artery from postpartum donors retained a hypertrophic appearance after 3 wk in oculo. We interpret these results to indicate that the degeneration of sympathetic nerves in late pregnancy, as well as their slow regrowth to the uterus after delivery, may be due to changes in uterine smooth muscle rather than a particular sensitivity of short adrenergic neurons to hormonal changes.

The maturity of current 3D rendering software in combination with recent developments in computer vision techniques enable an exciting range of applications for the visualisation, measurement and interactive manipulation of volumetric data, relevant both for diagnostic imaging and for anatomy. This paper reviews recent work in this area from the Image Sciences Institute at Utrecht University. The processes that yield a useful visual presentation are sequential. After acquisition and before any visualisation, an essential step is to prepare the data properly: this field is known as ‘image processing’ or ‘computer vision’ in analogy with the processing in human vision. Examples will be discussed of modern image enhancement and denoising techniques, and the complex process of automatically finding the objects or regions of interest, i.e. segmentation. One of the newer and promising methodologies for image analysis is based on a mathematical analysis of the human (cortical) visual processing: multiscale image analysis. After preprocessing the 3D rendering can be acquired by simulating the ‘ray casting’ in the computer. New possibilities are presented, such as the integrated visualisation in one image of (accurately registered) datasets of the same patient acquired in different modality scanners. Other examples include colour coding of functional data such as SPECT brain perfusion or functional magnetic resonance (MR) data and even metric data such as skull thickness on the rendered 3D anatomy from MR or computed tomography (CT). Optimal use and perception of 3D visualisation in radiology requires fast display and truly interactive manipulation facilities. Modern and increasingly cheaper workstations (<$10000) allow this to be a reality. It is now possible to manipulate 3D images of 2563 at 15 frames per second interactively, placing virtual reality within reach. The possibilities of modern workstations become increasingly more sophisticated and versatile. Examples presented include the automatic detection of the optimal viewing angle of the neck of aneurysms and the simulation of the design and placement procedure of intra-abdominal aortic stents. Such developments, together with the availability of high-resolution datasets of modern scanners and data such as from the NIH Visible Human project, have a dramatic impact on interactive 3D anatomical atlases.

In the golden hamster (Mesocricetus auratus), pubertal establishment of spermatogenesis includes a defined period (d 26–30 of life) during which elongation of spermatids is selectively arrested. The resulting appearance of germ cell associations not conforming to stage and the phenomenon of desynchronisation-related germ cell degeneration are analysed both quantitatively and qualitatively by means of light and ‘retrospective’ electron microscopy. From d 26 onwards, the portion of tubules containing non-stage conforming germ cell associations gradually increases up to 37.5% of sectioned tubules on d 32. Concomitantly, the degree of desynchronisation rises to a maturational gap between spermatids and associated younger germ cells of 7 stages of the seminiferous epithelium cycle, i.e. of fully half a cycle. Beyond d 32, the frequency of desynchronised tubule segments decreases again. Some of the arrested round spermatids and, eventually, all belatedly elongating spermatids degenerate and are lost from the epithelium. Thus a regular maturation of advanced spermatids does not succeed under non-stage conforming conditions. Possibly it is not the desynchronisation between the associated germ cell generations and the spermatids by itself that impedes normal further development of the latter cells. Instead this may be due to the maturational delay of the stage-aberrant cells by several stages compared to the seminiferous epithelium as a whole and, especially, in relation to the stage-conditioned functional state of the neighbouring Sertoli cells.

Associated skeletons, which are specimens preserving more than one body part from the same individual, are especially important for taxonomic and functional analyses. This study concentrates on the subset of associated skeletons which preserve the reciprocal surfaces of a joint. It uses laser scanning to explore whether the shapes of the reciprocal surfaces of a joint of an individual are significantly more congruent than the surfaces of randomly-matched pairings taken from the same species. Laser scanning was used to capture the distal articular surface of the left tibia of OH35 and the trochlear articular surface of the talus of OH8, both from Bed I, Olduvai Gorge, Tanzania. The degree of congruency between those articular surfaces was tested against the congruency of the talocrural joint of AL 288-1 (Australopithecus afarensis), and the congruency of both associated and randomly-matched talocrural joints of modern humans, chimpanzees and gorillas. The results suggest that OH35 and OH8 do not come from the same individual and may not come from the same species. Although this analysis leaves open the taxonomic affinity of OH35, it demonstrates the potential of laser scanning for capturing 3D data in palaeoanthropology. It also demonstrates the potential for using the relative congruency of reciprocal joint surfaces as a test of the likelihood that isolated limb bones are components of a single individual.

The objective of this study was to investigate the normal distribution of cartilage thickness in the major joints of the lower limb in elderly individuals. A 12.5 MHz ultrasound transducer was used to measure the cartilage thickness in the right and left hip, knee and ankle joint of 10 individuals aged between 62 and 99 y. Distribution patterns of cartilage thickness were derived by b-spline interpolation and the average distribution computed in each surface. The maximum cartilage thickness in the hip joint was 2.6 (±0.36) mm and the mean thickness 1.3 (±0.17) mm. The CV% (a measure of thickness inhomogeneity within the joint surface) was 32%. In the knee, the maximal and mean values were 3.8 (±0.46) mm and 1.9 mm (±0.24) mm, respectively (CV%=34%), and in the ankle 1.7 (±0.25) mm and 1.0 (±0.16) mm (CV%=32%). Systematic differences existed between both sides in the knee, the distal femur showing a significantly greater thickness on the right. While the mean and maximal thicknesses were systematically higher in the knee than in the hip, and in the hip higher than in the ankle (P<0.05), there were no systematic differences in the thickness inhomogeneity of the 3 joints. Only the malleolus showed a somewhat more uniform thickness than the other joint surfaces. The variablity between individuals was similar for all joints for mean thickness, but the interindividual variability of the maximal thickness values was highest in the knee and lowest in the ankle. Whereas the cartilage thickness distributions in the joints of the lower limb have been suggested to reflect the pressure distribution within the articular surface, the absolute thickness is proposed to be a function of dynamic loading (range of motion) during gait, rather than being a reflection of the static articular pressure.