One of the hot technologies that’s been driving the multiyear biopharma boom is cell therapy, which comes in many forms. But it generally involves engineering human cells for our own purposes, and while that promises some things that are basically impossible by other means, it also comes with a lot of unknowns. This new paper is a good illustration. People have been trying for quite a while to produce pancreatic islet cells for transplantation as a treatment for Type I diabetes (as well as find a way to keep direct tissue transplants functioning), and it’s been a tall order indeed. But there’s been progress in making insulin-secreting beta cells from stem cells, and of course stem cells themselves were the subject of their own funding boom some years back.
I thought the public enthusiasm over stem cell technologies was overdone, and the paper under discussion is a good illustration of why. There have been reports since 2014 of producing beta cells through differentiation of human stem cells, and that was an exciting advance. But on closer inspection, the differences between these and good ol’ fashioned beta cells become apparent. The new paper has yet another differentiation protocol to produce such cells (and indeed, functional islets as a mixture of alpha- and beta cells), and the authors do a pretty thorough comparison to see where things diverge.
The good news is that these neo-islets are glucose-responsive, and they can be seen maturing over weeks in cell culture and during months of engraftment in mice, ending up with what seems to be a reasonable facsimile of a native islet. The weird news is that they manage to do this despite having profound metabolic differences as compared to “real” beta cells – and those goes all the way down to the way glucose in metabolized in their mitochondia. In normal beta-cells, that’s a key signaling pathway for glucose sensing and insulin secretion, but not in these new cells:
In contrast to primary islets, SC-islets present significantly lower mitochondrial TCA metabolite enrichment, minimal ATP/ADP and NAD+/NADH ratio shifts, and absent respiration spikes during glucose stimulation, all of which are key aspects of the canonical triggering pathway in functional beta cells61. Modulation of the KATP-channel in SC-islets does show that glucose-induced Ca2+influx and insulin release largely depend on KATP-channel closure. Therefore, while many elements of the stimulus-secretion coupling process are functional in at least a subset of SC-beta cells, the SC-islets as a whole are not robustly coupled metabolically to the canonical triggering pathway.
So all of the stuff that you learn about glucose sensing and its coupling to insulin secretion sort of goes out the window, which I’ll bet no one saw coming. The quoted paragraph says “subset” because it appears that only part of the new beta-cell population is insulin-responsive at all, and why this occurs is still under investigation. From the metabolic profiling work, it appears that the phosphoenolpyruvate cycle and various intracellular redox signals might be the mechanisms by which these stem-cell-derived-beta-cells respond to glucose, but that’s still being worked out. There must be some compensatory mechanisms operating, since it seems clear that the canonical ones aren’t. It’s remarkable that a glucose-responsive phenotype emerges, but not the one that anyone was expecting.
So that brings up the question of how such cells would perform in real diabetes therapy – you’d especially want to know their strengths and weaknesses before you start putting them into human patients. How do they respond to exercise, for example, or to a prolonged fast? Will the islets continue to mature, or is this as far as they get? What seems sure is that more advanced differentiation protocols would be useful – maybe we can invent ourselves past some of these questions with beta-cells that act more like what we’re used to (or what evolution has provided us with).
But all this illustrates the trickiness of making new human tissues from stem cells. Recapitulting human tissue development is no easy task; the amount of signaling that goes into these processes is mind-boggling. And as these beta-cell efforts show, we don’t quite understand the details, both what to leave in and what to leave out. In this case, we ended up with something that still seems to work fairly well, somehow, but many times you won’t. So all the talk about growing and transplanting new stem-cell derived nerve tissue, new liver and pancreas tissue, new cardiac muscle, etc. still (after all these years) comes under the “Should be possible but not really yet” heading. It is a long hard road, and we’re only partway along it – we can make islet-ish tissue, neural-like tissue, muscle-oid cells, that sort of thing, but are these useful for human therapy or not? We might be within range of useful effects with these new islets, but as mentioned, that remains to be proven. Overall, it’s a good thing that the hype has died down over the years so the real work can go on. . .
