By David N. Leff

The Holiest Grail that diabetologists pray for is a way to transplant islets of Langerhans into the bloodstreams of diabetic patients ¿ 16 million of whom are in the U.S. alone.

A German anatomist and microscopist, Paul Langerhans (1847-1888) was the first person ¿ in 1869 ¿ to see and describe the strange blobs scattered throughout the pancreatic gland. But it took another half century for a new generation of seekers to recognize those islets as the source of insulin. This is the hormone that keeps diabetes sufferers alive by controlling the day-to-day conversion of the glucose sugar in their diet into energy and fat.

It¿s not for lack of trying that researchers have yet ¿ on a practical clinical basis ¿ to replace failed islets with those harvested from cadavers, or other sources. But the potential demand far outstrips the supply. Experimental implanted islet cells trigger immune rejection, for which recipients would in effect have to subsist on a lifetime of immunosuppressants as the price of restoring their insulin output.

Today¿s issue of Science, dated April 27, 2001, reports a molecular-genetic way out of this dead end. Its title: ¿Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets.¿ Its senior author is neuroscientist Ron McKay, who heads the Laboratory of Molecular Biology at the National Institute of Neurological Disorders and Stroke. The paper¿s lead author is cell biologist Nadya Lumelsky, a Special Expert in McKay¿s lab.

¿Our overall finding,¿ Lumelsky told BioWorld Today, ¿is that we can use embryonic stem cells to generate pancreatic progenitor cells, and differentiate them into islet-like cells producing insulin. They also differentiate into cells producing other pancreatic endocrine hormones ¿ specifically, glucagons and somatostatin, plus pancreatic polypeptides.

¿All these different cell types,¿ she continued, ¿assemble into structures in vitro that are reminiscent topologically of pancreatic islets. I¿m not saying these clusters are authentic islets of Langerhans, but they are organized in a similar fashion.¿

Embryonic Stem Cells Kick Off Program

With mouse embryonic stem cells for starters, the co-authors approximated in culture dishes the steps that primitive embryos go through in creating islets of Langerhans.

¿The overall idea that led us to this work,¿ Lumelsky said, ¿was that if you take the repertoire of transcription factors, which are involved in regulation of neural and pancreatic development, they are almost identical. And many other features bring them close together. So, it has been suggested that they are exposed to similar kinds of signaling molecules during that early stage.

¿Using this information,¿ she went on, ¿we had hypothesized that maybe we could use strategies we knew work for generation of neural cells and apply them for generation of pancreatic cells. Which is what we did. And we could actually see generation of both cell types ¿ neurons and pancreatic cells ¿ in our culture dishes.¿

Two key factors ¿ a mystery protein called nestin, and basic fibroblast growth factor (bFGF) ¿ guided the five-stage, three-week formulation of the insulin-secreting pseudo-islets:

¿ Expanding the embryonic stem cells over two to three days in the presence of a factor that prevents differentiation of these stem cells inhibiting other cells.

¿ Removing that factor after four days allows formation of ES cell aggregates where they initiate their differentiation program.

¿ Growing those clumps in serum-free medium for a week killed many cells, but those expressing restin remained.

¿ Exposing those nestin-positive cells to bFGF for six days transformed them into pancreatic precursor cells.

¿ Removing bFGF after another six days caused some of those cells to become insulin-secreting clusters resembling their role model ¿ pancreatic islets.

¿We don¿t know the function of nestin,¿ Lumelsky said. ¿It¿s a structural protein ¿ a part of the so-called intermediate filaments, which form the cell¿s cytoskeleton. It didn¿t do anything functional for us. Because we knew that nestin is expressed in early neural cells, we used its expression merely as a criterion to know that we were targeting the right cell population.¿

Lumelsky described their lab-made clusters as ¿somewhat roundish to perfectly spherical, about 50 microns in diameter. And the 6-centimeter culture dish we grew them on probably contained several hundred.¿

Results In Vitro, In Vivo ¿ Upside, Downside

As proofs of principle, she and her co-authors ran two sets of tests, one in culture, the other in diabetic mice. ¿We achieved response of insulin to glucose in vitro,¿ she observed. ¿It¿s an important finding in itself, because we know that at least in culture the cells sense glucose and release insulin in response to it. The cells were also blocked by inhibitors of native islets, and activated by their activators. So at least in vitro, their physiology is somewhat similar to native islet physiology.

¿Going in vivo,¿ Lumelsky recounted, ¿we haven¿t done much. What we have done is graft cells under the skin of the mice. This might not be the optimal place to graft, but we wanted to do an experiment that would give us fast results. We didn¿t ask whether we can cure an animal. We wanted to know if cells were surviving, and if they were still expressing insulin.

¿Also, are they going to vascularize, because that would be one of the requirements for cells to function. Unless there is blood supply in the graft, their insulin will never sense glucose in the blood.

¿Answering these questions,¿ she went on, ¿we confirmed that the cells survive at least six weeks and still produce insulin. We don¿t know what happens after that. And we saw angiogenesis after about 10 days.¿

McKay observed: ¿It really looks as if you¿re getting bits of the animal ¿ groups of cells that are assembled together. And this first demonstration of islet-like structures produced from ES cells may provide a source of functional pancreatic islets for treatment of diabetes.¿

The major downside is that insulin output amounted to only 2 percent of that secreted by normal islets. Moreover, it did not correct hyperglycemia ¿ the hallmark of diabetes ¿ in the mice. These are the hurdles the group is now tackling.

¿Our immediate goal,¿ Lumelsky concluded, ¿is to increase insulin production by the cells. We have several ways that we think should work.¿