Collagen VI is a major component of the islet-exocrine interface of the adult pancreas, the content being more than double that of collagen I or IV. However, the proportional collagen VI content was not dependent on the age of the donor. These data may facilitate the design of new collagenases, targeting major substrates such as collagen VI in order to improve clinical islet isolation.

The intraportal site is the most common site for islet transplantation. Many other sites have been tried experimentally, including the spleen, which has successfully lead to insulin independence in a number of animal models. Nevertheless, there are no detailed reports of total pancreatectomy and splenic islet autotransplantation in humans. Five patients underwent total pancreatectomy and splenic islet autotransplantation for chronic pancreatitis. Four patients had a pylorus-preserving total pancreatectomy and one patient a duodenal-preserving pancreatectomy. In three cases islets were embolized into both the portal vein and spleen. Two patients received splenic islet transplants alone. Islets were transplanted by retrograde venous infusion via the short gastric veins (n = 3), splenic vein stump (n = 1), and the left gastroepiploic vein (n = 1). The total volumes of transplanted pancreatic digest in those receiving combined intraportal and splenic autografts (n = 3) were 15.8, 13.0, and 13.5 ml. The volumes in those receiving a splenic-alone autograft (n = 2) were 12.0 and 5 ml. The mean rise in portal pressure was 18 cm of water. Complications related to the splenic autograft included a wedge splenic infarct, an emergency splenectomy, and a portal vein thrombosis in one patient having a combined intraportal and splenic autograft. Two patients developed insulin independence. two patients were still insulin independent at 1-year follow-up, and all had normal HbA1c levels (mean 5.6, range 5.2-6.3). Splenic islet autotransplantation, after total pancreatectomy, does lead to insulin independence. However, in our experience the combined procedure has a high morbidity because of splenic infarction and venous thrombosis.

The use of the COBE 2991 cell processor (COBE Laboratories, Colorado) for large-scale islet purification using discontinuous density gradients has been widely adopted. It minimizes many of the problems such as wall effects, normally encountered during centrifugation, and avoids the vortexing at interfaces that occurs during acceleration and deceleration by allowing the gradient to be formed and the islet-containing interface to be collected while continuing to spin. We have produced cross-sectional profiles of the 2991 bag during spinning which allow the area of interfaces in such step gradients to be calculated. This allows the volumes of the gradient media layers loaded on the machine to be adjusted in order to maximize the area of the gradient interfaces. However, even using the maximal areas possible (144.5 cm2), clogging of tissue at such interfaces limits the volume of digest which can be separated on one gradient to 15 ml. We have shown that a linear continuous density gradient can be produced within the 2991 bag, that allows as much as 40 ml of digest to be successfully purified. Such a system combines the intrinsic advantages of the 2991 with those of continuous density gradients and provides the optimal method for density-dependent islet purification.

No abstract