Spontaneous cell sorting of fibroblasts and keratinocytes creates an organotypic human skin equivalent.

We show that an inherent ability of two distinct cell types, keratinocytes and fibroblasts, can be relied upon to accurately reconstitute full-thickness human skin including the dermal-epidermal junction by a cell-sorting mechanism. A cell slurry containing both cell types added to silicone chambers implanted on the backs of severe combined immunodeficient mice sorts out to reconstitute a clearly defined dermis and stratified epidermis within 2 wk, forming a cell-sorted skin equivalent. Immunostaining of the cell-sorted skin equivalent with human cell markers showed patterns similar to those of normal full-thickness skin. We compared the cell-sorted skin equivalent model with a composite skin model also made on severe combined immunodeficient mice. The composite grafts were constructed from partially differentiated keratinocyte sheets placed on top of a dermal equivalent constructed of devitalized dermis. Electron microscopy revealed that both models formed ample numbers of normal appearing hemidesmosomes. The cell-sorted skin equivalent model, however, had greater numbers of keratin intermediate filaments within the basal keratinocytes that connected to hemidesmosomes, and on the dermal side both collagen filaments and anchoring fibril connections to the lamina densa were more numerous compared with the composite model. Our results may provide some insight into why, in clinical applications for treating burns and other wounds, composite grafts may exhibit surface instability and blistering for up to a year following grafting, and suggest the possible usefulness of the cell-sorted skin equivalent in future grafting applications.

Variability of human coronary artery geometry: an angiographic study of the left anterior descending arteries of 30 autopsy hearts.

A computer-based system is described to measure objectively the geometric parameters of arteries from pairs of projection angiograms. This technique, which employs back-projection to define the vessel axes in 3-D space, was used to obtain selected parameters of coronary artery geometry from radiographic images of autopsy hearts. Results of the first 30 cases are presented, focussing on the distribution of the geometric parameters of the left anterior descending coronary artery (LAD) and its first two major branches. The derived parameters include the angle between the left circumflex artery and the LAD; the angles between the LAD and its early diagonal and septal perforator branches; distances between branch points; and tortuosity. The geometric parameters vary considerably, presumably contributing to a corresponding variability in local hemodynamic and mechanical stresses. Most parameters are uncorrelated. One exception is the angle at the origin of the second diagonal branch, which is positively correlated (p < 0.01) with the distance between the ostia of the first two diagonal vessels; this correlation could reflect the existence of "target" perfusion regions. No relation between geometric parameters and age or gender was seen. In this sample, blacks had a larger angle at the left main bifurcation than whites (p < 0.05).

Relation between coronary artery geometry and the distribution of early sudanophilic lesions.

The relationship between the angle formed by the left circumflex (LCX) and left anterior descending (LAD) coronary arteries at the bifurcation of the left main coronary artery (LM) and the location of sudanophilia in the proximal portions of the LCX and LAD, was investigated, using a normalized index of disease severity (relative proximal involvement, RPI) to isolate the local effect of geometry on the predisposition to disease from that of other risk factors. Multiplane angiograms of the left coronary artery system of 15 hearts were digitized and processed to form a three-dimensional representation of the centerline of the lumen (termed the 'axis') of each coronary segment; the LM branch angle was objectively computed from the reconstruction. The coronary vessels were fixed in situ, removed from the hearts, opened longitudinally and stained with Sudan IV. The RPIs of the LAD and LCX were obtained by dividing the percentage sudanophilia seen en face in the first 1 cm of each vessel by the percentage sudanophilia in the first 5 cm. RPI was correlated negatively with LM branch angle, suggesting that a small LM branch angle may be a 'geometric risk factor' for proximal atherosclerotic disease in the daughter vessels. The correlation was stronger for the LCX than the LAD. These results would appear to be consistent with the notion that atherosclerosis is favored where a larger extent of the vessel experiences fluid dynamic wall shears near zero for an appreciable part of the pulsatile cycle.