Thursday, November 21, 2013

Successful Case Study of New Encapsulated Islet Cell Device

Research topics ebb and flow over time.  I remember when my daughter was first diagnosed 10 year ago (was it really that long ago?) that encapsulated beta cell research was all the rage.  Then it seemed to die off, just leaving Living Cell Technology (LCT) in the field.  Now it seems to be heating up again.

Successful Case Study of New Encapsulated Islet Cell Device


Encapsulated  beta cells are one potential cure for type-1 diabetes.  Beta cells are wrapped in a protective coating and are put in a person.  The coating allows blood sugar in, and insulin out, but does not allow the body's immune system to attack the beta cells. It also allows nutrients in and waste products out. This allows the beta cells to naturally grow and to react to the body's sugar by generating insulin which goes into the body's blood system. Meanwhile, the body's autoimmune attack can not target these beta cells, and you don't need to take any immunosuppression drugs (as you would for a normal beta cell transplantation).

Encapsulation research started in people in the 1990s.  The current leaders in the field are LCT, in phase-II human trials, and University clinical Hospital Saint-Luc and The Sydney Project who have both done phase-I trials.  Plus there is Sernova, the Islet Sheet Project, Viacyte (and probably several more) in animal research.

This Study

These researchers (who I will call "the Dresden group") have published a case study.  A case study report on a single person, so it's not a clinical trial.   For this person, they implanted some human beta cells (from a cadaver) encapsulated using their own alginate technology [d1], and added an oxygen infuser (which had to be reloaded every day) [d2].  The researchers implanted 2k cells / kg, which is about 1/5 of what they expect to need to cure someone.  It is very common in phase-I trials to give a very small amount of whatever you are testing, in order to check for safety.  They monitored insulin usage, and measured C-peptide levels every three months.  After 10 months, they removed the capsule and ran some tests on it.


C-peptides in Response to Carbs
One way to test the effectiveness of this device, is to give the patient some sugar, and see if they generate insulin (by measuring C-peptide levels) in response to that sugar.  These researchers tested the patient once every three months, and the data is shown in the graph below:

Notice that all three response lines are pretty much the same [d3].  That is very positive news, because it means that the implanted beta cells worked just as well after 9 months as after 3 months. Previous encapsulated beta cell treatments have degraded over time, and the results after 9 months would be noticeably worse than after 3 months.

Insulin Usage Before and After
Prior to implantation, this patient used about 52 units of insulin per day.  After the implantation he used 43 units.  That's a drop of 17%.  For the first human trial, with a dose much smaller than expected for a cure, that's a very promising result.  It is especially interesting that by transplanting about 1/5 as many cells as they expect to need, the patient required almost 1/5th less insulin.  That might be coincidence, or it might foreshadow more good news to come.

Fasting C-peptides
These measured about 0.04 nmol/L on average.  The MedScape normal range (ie. non-diabetics) is about 0.26-1.03 nmol/L, so these guys were well below normal.  But again, for a first case study, with a small dose of islets, this has promise.

Moving Forward

This research needs to move forward in three ways, and hopefully they can move forward in all at the same time.  First, they need to run an actual clinical trial (ie. get more people treated).  Case studies just don't attract the same attention as clinical trials, nor should they.  One important next step for this research program is to implant this in 6-20 people and see what happens.  The second step is to implant more beta cells, and the third is to find a better way to supply oxygen to the cells.

For this case study, they implanted about 2k/kg cells.  They believe that about 10k/kg cells will need to be implanted to cure type-1 diabetes.  If you look at the insulin requirements for the patient, they dropped just below 20% after implantation.  So the next big question is, if you give more beta cells, do they generate more insulin, in proportion?  If so, that is very important.


These researchers are using human cadaver beta cells and a proprietary alginate encapsulation technology, but of which have been done before by others.  However, they are also doing something unique, which is oxygenating the cells.  They hope that providing extra oxygen will make the cells work better and survive longer.  The trade off is that the oxygen generation system requires daily maintenance.

The obvious research to compare this to is LCT's "Diabcell".  In LCT's most recent clinical trial (a Phase-II trial), they implanted  a total of 10k/kg cells in 4 patients and 20k/kg cells in 4 patients.  The higher dose group saw a 20% drop in insulin usage. LCT did not report on the durability of this result.  So we don't know if they continued to use 20% less for months, or if the effect went away relatively quickly.  So the most recent LCT results and the Dresden results are very similar, except that these researchers reported on durability and the LCT researchers have not, and the Dresden team used far fewer cells.

In addition to the possibility of being a cure by itself (especially if the oxygen supply issue can be solved), this research is also important for what it tells us about encapsulated beta cell longevity in general.  If bigger trials show the same durability (that the cells continue to work for a long period of time), that strongly suggests that the problem that causes encapsulated beta cells to die over time is a lack of oxygen.  So that means that everyone trying to create encapsulated beta cell cures will know to concentrate on the oxygen supply.  Although this might not be new news to researchers: I remember  David King (Islet Sheet developer) talking about the importance of oxygen supply years ago.

News articles:
General link on encapsulation:

Extra Discussion

[d1] Alginate is a general term that covers all (or almost all) of the recent encapsulation techniques.

[d2] One important question is: how long is the oxygen generator used?  The paper makes it sound like it was used for the whole 10 months that the device was in use.  However, one of the researchers was quoted as follows:
For that reason, the current version of the device had an oxygen port on the outside of the body attached via tubing that had to be refilled daily by the patient for as long as a month or two, Block explained.
and that implies the oxygen generated is only needed for a short time.  Since the generator needs to be refilled each day, and that is a hassle, this duration is important.  I did ask the researchers how long the oxygen generated was used for, but had not heard back at the time I finished this posting.

[d3] Different people view this data differently.  I tend to see them as all being about the same, but at least one person (Celsus, on thinks that the 3 month is stronger, because it ends stronger.  However, when I look at the whole graph, I think they are all similar.  The official way to make the comparison is AUC "area under the curve", and I don't think the researchers did that math.

Joshua Levy --
publicjoshualevy at gmail dot com
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF, JDCA, or Tidepool news, views, policies or opinions. My daughter has type-1 diabetes and participates in clinical trials, which might be discussed here. My blog contains a more complete non-conflict of interest statement. Thanks to everyone who helps with the blog.

Friday, November 8, 2013

Carnitine As Newborn Marker for Type-1 and Speculative Cure

Carnitine As Newborn Marker for Type-1 and Speculative Cure

A team in Italy has been storing blood samples taken the day a baby is born.  Another team performed the following experiment:  If a child was diagnosed with type-1 diabetes at age 6 or less, they tested the stored blood for Carnitine and a variety of closely related chemicals.  These chemicals occur naturally and are biologically active, in your body.  They are vaguely related to B vitamins, but are not technically vitamins themselves.  The team then compared the levels of these chemicals to the levels in the newborn blood taken from babies who had not been diagnosed with type-1 diabetes, but were born within one day of those that were.  (These children formed a control group.)

The results were striking: the children who would later get type-1 diabetes had statistically significantly lower levels of many of these Carnitine related chemicals.  Now, they did not find a clear line, where carnitine levels below X singled type-1 diabetes, and over Y was safe.  There was significant overlap.  What they found was that on average lower levels predicted higher type-1 rates. [d1]

This study could be the first step towards predicting type-1 diabetes.  Obviously, being able to predict type-1 diabetes at birth, even if not perfect, would still be very helpful, both in terms of extra vigilance for earlier diagnoses [d2], and for early intervention and prevention.

Speculation About A Cure

This brings up the issue of "cause vs. association".   Specifically, the study did not show that low levels of Carnitine cause type-1 diabetes, but only that low levels are associated with type-1.   From an early detection point of view, this doesn't matter.  As long as something can be measured early, it doesn't really matter if it is a cause or an association.  But for prevention the "cause vs. association" question is critical.

The researchers [d3] who ran this experiment, think that low levels of Carnitine (and/or related compounds) during the first day or two of life, causes type-1 diabetes later in life.  It's a unique theory.  I've never heard any other researcher who believes anything similar.  The theory is low Carnitine triggers a failure when the body is starting up it's immune system, that it never recovers from, and that manifests itself as type-1 diabetes later in life.  Their full paper is on-line (link below), so you can read their explanation yourself.

If low Carnitine during the first few days of life does cause type-1 diabetes, then perhaps it could be prevented by giving Carnitine supplements to pregnant women, or to newborns.  Neither of these interventions is impossible.  Today it is recommended that pregnant women (and even those trying to get pregnant) take folic acid, to prevent a specific type of birth defect.  And also, newborns are often given vitamin K (and a Hep B vaccine) within hours of birth.  So similar could be done for Carnitine, if it mattered.

Obviously, the next step would be an intervention study, where people are given Carnitine, and then tracked to see if it lowers the chance they will get type-1 diabetes.  Such studies require a lot of people, and a long time, so I would not expect a quick answer.  The study reported on here, is a retrospective, population based study, and so is much less persuasive than an intervention study. The good news is that L-carnitine is available now as a "dietary supplement" so setting up an intervention study should not be that complicated.

News article:
Full paper:

Extra Discussion

[d1] In more detail: what they found was that, for a whole group of Carnitine related chemicals, people who later got type-1 diabetes had -- on average -- lower levels, than those who would not get type-1.  None of these chemicals could be used alone to predict type-1 diabetes.    None of them had a line which separated type-1 diabetics from everyone else.  Instead, at higher levels very few people got type-1 diabetes, at the middle levels it gradually shifted from few type-1s to more and more of them.  And then at lower levels the number of type-1 diabetics were higher.  But you could never say "level X means type-1" or "level Y means you will not get type-1".

This brings up a critical question: even if none of the chemicals gave a yes/no answer to the question of future type-1 diabetes, maybe by combining data for all the chemicals we could give a yes/no answer.  There is no discussion in the paper of the authors' attempt to use mathematical analysis to see if a combination of different chemical levels could be used to predict type-1 diabetes.

[d2] For example, diabetic ketoacidosis at the time of diagnosis might become much more rare, than it is now, if people were on the lookout for type-1 from birth.  Obviously, there is a downside in terms of hypervigilance, and over protectiveness, etc.  But the only way we can get to prevention is through early intervention, and the only way to get to early intervention is to predict who is going to get the disease.

[d3] This group of researchers was led by Gian Franco Bottazzo, who is generally credited with discovering (in 1974) that type-1 diabetes was an autoimmune disease.  Obviously, this was a huge breakthrough in type-1 research.  He has won the Banting medal, and several other international honors.

Joshua Levy --
publicjoshualevy at gmail dot com 
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF, JDCA, or Tidepool news, views, policies or opinions. My daughter has type-1 diabetes and participates in clinical trials, which might be discussed here. My blog contains a more complete non-conflict of interest statement. Thanks to everyone who helps with the blog.