The Betatrophin Blues

All the recent high-profile papers that are having troubles are bumming me out.

The latest example is the “Betatrophin” Diabetes paper from Harvard last year.

Yesterday it was called into major doubt by a new Cell paper from a group led by Jesper Gromada at Regeneron.

The authors of the original 2013 Betatrophin paper–Doug Melton’s team at Harvard– indicated in their own accompanying, somber perspectives piece also in yesterday’s issue of Cell that they agree that their 2013 report was largely incorrect. This unfortunate turn of events is based on both the new Regeneron paper and the Melton lab’s own follow up work.

Rewind back to 2013 when the Betatrophin finding first came out and it was hard not to get into the story at that point given how it was being portrayed. A Harvard press release made little doubt that this was a seminal discovery for Diabetes research:

The hormone, called betatrophin, causes mice to produce insulin-secreting pancreatic beta cells up to thirty times the normal rate…it could eventually mean that instead of taking insulin injections three times a day, you might take an injection of this hormone once a week or once a month, or in the best case maybe even once a year.

The 2013 paper itself was also very bold with statements such as:

Thus, betatrophin treatment could augment or replace insulin injections by increasing the number of endogenous insulin-producing cells in diabetics.


They raised expectations sky high.

I blogged about that apparent blockbuster finding here back then in 2013 and I definitely was excited about it given how it sounded. Now there’s a major letdown. The new paper’s title alone pretty much says it all:

ANGPTL8/Betatrophin Does Not Control Pancreatic Beta Cell Expansion

You can see the graphical abstract from this paper at left in which the authors report that Betatrophin, which now should probably go by the more objective name ANGPTL8, does not substantially impact beta cell growth, but rather seems to have a notable role in mouse triglyceride metabolism. ANGPTL8 is probably a very interesting molecule, but it is not what it seemed to be.

In the perspectives piece, the Betatrophin authors say that boosting their N of mice has led their main conclusion to come into doubt:

In Yi et al. we reported an average beta cell replication rate of ∼4% in betatrophin-injected mice (n = 7); with five additional experiments (n = 52 mice in total), the average beta cell replication rate in betatrophin-injected mice drops to 1.2%. While still significantly above control levels (p = 0.016 for all experiments) of beta cell replication (0.6%), the conclusion from Yi et al. must be corrected and modified with respect to the magnitude of the effect.

It’s now unclear what the fate of the 2013 Betatrophin paper will be moving forward given that its central argument is incorrect and even the naming of the molecule “Betatrophin” is indeed perhaps not appropriate any more.

Top 10 Takeaways From Harvard Stem Cell Diabetes Paper

Harvard stem cells DiabetesThe idea of using stem cells to treat Type I Diabetes is very promising and could have huge practical impact. Real progress has been achieved toward this goal over the last decade. In perhaps another decade there might be a validated treatment.

A new stem cell paper has just come out in Cell from Harvard in this area. It reports making insulin-producing β cells from human embryonic stem cells (hESC) with the notion of some day using these cells to treat Diabetes. See graphical abstract at left from the paper. These β cells help Diabetic mice control their blood sugar. Very cool.

This is an important, exciting paper, but one that the media has for the most part totally hyped. For a balanced, well-researched newspaper article on this Harvard Diabetes research I recommend this piece by John Lauerman at Bloomberg. It’s also worth noting that the same Harvard team published another big Cell paper related to stem cells and Diabetes last year on a hormone they discovered called Betatrophin that stimulates cells to turn into β cells.

Here are my top 10 key take home messages for this new paper in the context of the larger field of stem cell-related Diabetes research.

  • There are many labs around the world working on research toward the goal of using stem cells as a basis to treat Diabetes. The Harvard work is notable, but should be discussed in this larger context.
  • For example, another entity, the privately held stem cell biotech ViaCyte, has an approved IND for an hESC-based Diabetes treatment and an approved device in the form of a capsule.
  • Harvard will also need some kind of capsulation technology/device, which it does not appear to have at present. Harvard team leader Doug Melton described such a needed device as a “tea bag” with the idea being that the cells inside sense blood sugar (which goes into the capsule from blood) and make insulin (which goes out into the blood), but no cells should be able to go in or out.
  • ViaCyte will start a clinical trial in as soon as a few months for their hESC-based Diabetes therapy.
  • Barring some major change in regulatory approach by the FDA, the Harvard stem cell Diabetes product probably won’t start to be tested in humans for at least about 4 years given the experiences of other groups using human embryonic stem cell-based products. It’s important to keep expectations realistic.
  • Going back to 2006 we find a Nature Biotechnology paper using hESC to make pancreatic progenitor cells that can turn into β cells and other kinds of pancreatic cells, work from a team led by Ed Baetge of Novocell (now ViaCyte). Only now is ViaCyte going to start their in-human trials. These things take time.
  • ViaCyte’s product is not pure β cells, while the Harvard product appears to be relatively pure β cells. This could prove to be an important advantage over ViaCyte’s approach in the long run.
  • A challenge for any stem cell-based product is potential immune rejection by the patient given that Type I Diabetes has an immune component as a disease and the fact that the product will be allogeneic.
  • What works in mice often doesn’t work in humans so caution is in order.
  • Whether talking about Harvard or ViaCyte or anyone else, it is way premature for the media to talk about “cures” for Diabetes. The most appropriate word to use is “treatment”.