No nutrient is as important as oxygen. Oxygen might be considered in the same category as "food". We have separated it for the simple reason that while diffusion of nutrients of various size is largely a function of the chemical structure of the coating, diffusion of oxygen is mainly a function of coating thickness. Islets buried deep beneath a pack of neighboring, actively metabolizing cells will die of oxygen suffocation.
Islet Sheets have been fabricated with perfectly uniform overall thickness and can be made at any thickness desired. The vast majority of competing methods produce coatings which either are too thick or have unpredictable thickness. Islets inevitably die within these coatings, from either starvation, suffocation, or immune attack
Good oxygen kinetics
The islet must receive many nutrients to survive and thrive. The limiting nutrient is usually oxygen. The figure to the left shows the effect on islets of being more and more distant from the oxygen source. In the figure, oxygen diffuses in from the bottom; the darkening of the green color indicates depletion of oxygen. The islet on the bottom is normal and healthy. As one looks further up, further from oxygen, the islets become distressed, apoptotic, and finally are dead. Studies have shown that oxygen-starved islets have a poor insulin secretion response. Thus, the living islet in the middle is not an effective contributor to insulin responsiveness.
The basic physics of oxygen diffusion through oxygen-consuming tissue places severe limits on the thickness of the bio-artificial pancreas. The Islet Sheet must be less than a third of a millimeter thick to preserve 75% of the normal function of the islets it contains.
Oxygen diffusion through the Islet Sheet
Death from low oxygen is the most common cause of bio-artificial pancreas failure. This topic has been analyzed in some detail by researchers at MIT (Colton, C. K. (1995) Cell Transplant 4, 415-436.). In their model the islets are in a 100 micron thick slab at the center, and the sheet has membranes on both sides with thickness of 25, 50, 100 and 150 microns. The graph (Figure 13 in Colton) shows the calculated oxygen tension as a function of distance from the from the core/membrane interface. The model assumes 40 mm Hg Oxygen in the surrounding tissue, and that the implanted tissue consumes oxygen at a rate typical for islets.
Note that, because oxygen is being consumed in the core, it must diffuse to the core through the membrane, and at equilibrium a linear drop in partial pressure is established. It is startling to most people to learn that oxygen available to the islets is only 10mm Hg when the tissue level is 40mm Hg!
The situation is even worse when we consider that islets in a low oxygen environment survive but have reduced output of insulin. This concept is illustrated in the figure to the left (Dionne et al., Diabetes 42,12-21). The half-maximal secretion rate is at 27 mm Hg. Recall that the most privileged islets in the first figure have only 10 mm Hg in their immediate environment!
The take-home lessons are:
- The membranes should be as thin as possible, certainly less than 25 microns
- the overall sheet thickness should not exceed 250 microns.
Stathis Avgoustiniatos, the M.I.T. scientist who did much of the modelling work published by the Colton group, has visited us here in San Francisco. At this time he is taking our data and developing a specific model of oxygen diffusion through the Islet Sheet.