Active Diabetes Management Sat, 10 Mar 2018 15:03:35 +0000 en-US hourly 1 Saturday Morning Research Review – March 10, 2018 Sat, 10 Mar 2018 15:03:35 +0000 Presence and activation status of insulin-specific T cells is related to insulin autoantibodies
by Adam Burrack, PhD

In today͛s edition, I will describe a project that I was involved with, in a minor role. It has been an on-going
debate in the field whether immune responses against the insulin molecule itself are the driving force behind beta cell destruction. George Eisenbarth͛s laboratory put this question to the test in a 2005 report which
demonstrated that one key mutation to the insulin protein abrogated autoimmunity in NOD mice. Whether
this paradigm applies to human T1D has remained an open question. Addressing this question in human
samples has been a major challenge, because while determining individual risk of developing T1D is possible,
predicting when an individual will develop T1D is not straight-forward.
In addition, CD4 T cells are relatively rare in the peripheral blood. To address this question in human subjects therefore requires relatively large volumes. As a caveat, obtaining pancreas or pancreatic lymph node biopsies is not feasible in most cases from human patients, limiting our analyses to peripheral blood. These logistic
challenges restrict the number of clinical institutions capable of executing these experiments very short. Such institutions include the Barbara Davis Center in Denver, the Benaroya Institute in Seattle, the Joslin Center in
Boston, and the University of Florida. Perhaps a handful of others.
Lastly, the human HLA DQ8 molecule, which is very unstable chemically, is the HLA molecule most closely
related to the mouse MHC molecule I-Ag7. These MHC alleles are most closely associated with T1D onset –
presumably through their ability to load unique peptide sequences and activate CD4 T cell responses. Since
insulin is the hypothesized key target of CD4 T cells during diabetes pathogenesis, making a peptide-MHC
tetramer reagent for insulin loaded in DQ8 is a key road block. Enter into this conversation the laboratory of Dr. Brian Fife at the University of Minnesota and his post-
doctoral trainee Dr Justin Spanier. Dr Fife has been interested in the peripheral regulation of autoreactive CD4 T cells since his post-doctoral work with Dr Jeff Bluestone at
UCSF and has been on the faculty of the University of Minnesota since 2008. Dr Spanier is a trained biochemist and learned to make peptide-MHC tetramers for this study, including the DQ8-insulin reagent described
above. These authors quantified and characterized autoreactive insulin-specific CD4 T cells in the peripheral
blood of people with T1D or control subjects. To summarize, Dr Fife and Spanier found that shorter disease
duration correlated with the presence of more insulin-specific CD4 T cells in the peripheral blood, which suggests that insulin-specific CD4 T cell number ͞peaks͟ around the time of diabetes onset, and then decreases as beta cells are destroyed and progressively fewer target cells remain. Dr Spanier also found that
insulin-specific CD4 T cells tended to have an ͚effector memory͛ phenotype in these individuals, suggesting
recent antigen experience, or that these CD4 T cells had recently fulfilled their function in promoting beta cell destruction.
Dr͛s Fife and Spanier then collaborated with Dr Aaron Michels at the Barbara Davis Center to determine
insulin-specific autoantibody production in these individuals. Reminder that insulin-specific autoantibody is a marker of disease but does not appear to be directly pathogenic to beta cells. In contrast, the current
understanding in the field is that T cells are the perpetrator of beta cell death (see figure 1 of this recent
review about beta cell destruction from the Fife lab. Regardless, this quantification is key,
because self-tolerance to insulin is lost among both T cells and B cells during the development of T1D. The
question of which response occurs first and ͚helps the other along͛ is key to the future design of therapies to
perturb this disease process and prevent T1D onset. Work with Dr Michels demonstrated a direct correlation between the number of insulin-specific CD4 T cells in the blood and insulin autoantibody titers. This means the more insulin-specific CD4 T cells detected, the higher than anti-insulin antibody levels. This suggests a
functional relationship between autoreactive T cells and B cells. Since CD4 T cells ͞help͟ B cells to mature,
class-switch, and produce antibodies within germinal centers of lymph nodes, these results suggest CD4 T cell responses to insulin are
critical for promoting B cell responses and subsequent beta cell destruction. An important note is that CD4 T
cells are also key for activating CD8 T cell (or ͞killer͟ T cell) responses. As such, due to this critical role for CD4
T cells in activating both cell-mediated and humoral immunity, an interpretation of these results is that
depletion or inhibition of CD4 T cells specific for insulin might delay or prevent T1D in at-risk individuals. In
other words, without insulin-specific CD4 T cell help it is unlikely that human T1D would develop. This report
demonstrates the ability to detect insulin-specific CD4 T cells in the blood of recent-onset T1D patients and
that the number of these cells correlates with insulin autoantibody levels.
Perturbing the development and function of insulin-reactive CD4 T cells in at-risk individuals is the next
frontier of this type of immune-profiling research. There are a number of on-going clinical trials which have
the potential to affect these cells at If you or a loved one live with T1D, please contact your
Congressional representatives and express your strong support for the Special Diabetes Program. To continue to move clinical
care forward, as a society we need to make evidence-based decisions, decisions based on scientific research.

Saturday Morning Research Review – February 3, 2018 Sat, 03 Feb 2018 14:30:26 +0000 Gran Fondos and delayed hypoglycemia
by Adam Burrack, PhD
I had the opportunity last summer to race in two duathlons and two longer-distance cycling events. In this post I͛ll share
how my diabetes management worked – or didn͛t work as the case may be – for these events. I raced the regional ͚
championship͛ race for the Duathlon on Memorial Day weekend, as well as another duathlon race in late July. This
summer I added a couple of ~65 mile cycling events in the rolling hills of southwest Wisconsin via the Wisconsin Gran
Fondo series. All four of these events provided unique challenges,
described below.
As I͛ve written about previously in our series, I enjoy the duathlon (run-bike-run) race format and manage my diabetes for these 80-100 minute events with
a pretty straight-forward method: I eat a normal breakfast the day of the event (50-80 g CHO) 2-4 hours before race
start, perform the normal bolus for current BG level and CHO in the meal, and suspend my insulin pump for the duration
of the event. Given there at least 2 hours between breakfast and the exercise, and that I suspend the pump for the
event, I͛ve never had problems with low BG levels during these events. Upon completion of the event – when I am
uniformly in the physiologic normal BG range – I add-back 50% of the basal insulin I lost as soon as possible (as a quick
bolus). This approach works well for me for races of 1-2 hours duration. This includes my sweet spot of 10 mile to half-
marathon running races, most duathlons, cycling time trials, and shorter road cycling races. This approach was
successful at the USAT regional duathlon race Memorial Day weekend where I took 2nd in my age group
and the Minnesota Duathlon the final Sunday in July, where
I took 7th overall.
The challenge with these events, for my diabetes management, occurs 2-6 hours after finishing the event: there have
been several occasions when I͛ve gotten low blood sugars 4-ish hours after finishing the event, presumably due to
enhanced insulin reception expression – and therefore enhanced insulin sensitivity – following these events. Enhanced
insulin sensitivity following these events is expected, and is something I should be able to trouble-shoot. From my
experience, this becomes a particular challenge for me as events increase in duration beyond 90 minutes. Also from my
experience, as liver glycogen is replenished the following overnight, I seem to be at higher risk of low blood sugar levels
for several nights afterward.
A particular challenge I͛ve run into (technically biked into, pardon the pun) this summer is bumping the duration of a
hard aerobic effort up to 4 hours for the ~65 mile cycling event with rolling hills. Happily, I was well-trained for these
events – the Tour de Coulee and the Tour de
Circus in July – and experienced no problems during
the events or in the 4-6 hour window following completion of these events. Briefly, my in-race strategy for diabetes
management for these races was the same as I͛ve done for longer-duration hikes in the mountains: 50% temporary
basal rate for the duration of activity, along with 30-50 g CHO (via solid food) per hour of exercise (with 25-50% of the
normal bolus for the CHO in the food). I finished both of these events in 4 hours +/- 5 minutes and finished strong.
However, I had severe overnight low blood sugars following both events. These low events precipitated insulin-resistant
rebound high BG events the following days. This can be quite frustrating in the days following a successful long-distance
endurance event; it can feel as if the purpose of the exercise – improving diabetes management – has been defeated. In
addition, in the 4-7 days following these events, I felt as if I was getting low BG levels on my (8-mile) bike commute to
work, which again would be defeating the purpose of the exercise to a degree.
In light of these challenges, I under-took a literature search. I was looking for relationships between improved aerobic
fitness with T1D and the potential for increased frequency of severe hypoglycemia (ie, really bad low blood sugar level
events). Interestingly, I found an article from the journal Diabetes Medicine, published in 2016, which concluded a
similar point: that better VO2max (proxy for aerobic fitness) correlates with more frequent low blood sugar level events
during exercise for people with T1D. One limitation of applying this
study to my own situation is the study was not following competitive athletes with relatively high aerobic capacities, this study was designed to find relationships in fit, healthy, non-competitive individuals. In addition, this was a static study; these authors did not test whether this tendency becomes more pronounced throughout a training season, which was
specifically what I thought I might be experiencing. I know from previous Dexcom use that I do not dip into low BG levels during exercise in general, including mile-specific running training and last year͛s duathlon season.
To test this, I have resumed my Dexcom device. Initial results confirm my suspicion that I am not getting low more often and missing the signs and symptoms. In contrast, it appears I am ͞floating͟ around 150-200 mg/dl throughout the day.
Therefore, if anything I need to increase my basal rate, decrease my CHO ratio, and inject more insulin overall. This is
more in line with my expectations than the paper cited above. In previous cross country/half marathon training
experience, as I͛ve worked my way into better aerobic shape I͛ve had to do inject more insulin. My experience following the 4-hour cycling events with challenging hills is more in line with an hypothesis that I was entirely glycogen-depleted
following these efforts, and that it took several days to replenish, and that during those 4-7 day windows following the events I probably was getting low BG levels more often. I conclude that I should have been more mindful to fully recover from these events, rather than plowing forward with my training plans. This is another example of why hypotheses are great to test; you can find out what͛s really happening through a series of logical tests, and act to improve the outcome in the future 

Saturday Morning Research Review – December 30, 2017 Sat, 30 Dec 2017 13:52:21 +0000 Advancing our understanding of transplant recognition by the immune system

​by Adam Burrack, PhD

Today I have the privilege of describing research I performed. A fundamental problem facing scientists seeking to cure type 1 diabetes (T1D) through beta cell replacement is the immune response to “foreign” tissues following transplantation. Ironically, the genes with the highest diversity in the human genome (“polymorphic”) are related to the activation of T cells of the immune system. Long story short, T cells recognize these genetic differences between proteins very effectively, and these differences promote very strong immune responses. Unfortunately, T cell responses against these differences resist tolerance-promoted therapies and must be suppressed for the lifetime of transplant recipients. These lifelong immune suppressive therapies limit wide application of organ transplantation to cure kidney failure or hypoglycemia unawareness in people with T1D.

Here is where details matter, a lot. The molecule we think T cells are targeting is the Major Histocompatibility Complex – MHC – which shows or presents T cells peptides derived from viruses, bacteria, or other pathogens. MHC is the most polymorphic gene in the human genome, so any transplant from anyone other than your identical twin (if you have one) will be targeted by your immune system. But it gets worse! In theory, if we understood precisely what the immune system is targeting during these responses, scientists would have a chance to develop target-specific therapies to promote specific tolerance, leaving the rest of the immune system intact. We think, as immunology researchers, that MHC is the key target of the immune response following transplantation. But MHC only gets to the surface of cells if it’s loaded with a peptide. So the question for transplant response is this; what are the peptides loaded in donor MHC? Is there a common peptide that loaded in any donor MHC would promote an immune response in transplant recipients?

Enter into this context the laboratory of Dr. Marc Jenkins. For the past 25 years, Dr Jenkins’ laboratory has pioneered and popularized the use of a research tool called peptide-MHC tetramers to study CD4 T cell biology also called helper T cells, in a variety of biological settings including vaccination studies, infectious disease, autoimmunity, and now transplantation. Peptide MHC tetramers are a method to study very specific set of T cells which are specific for particular peptides. For example, a vaccine is intended to expand virus-specific T cells and peptide-MHC tetramers are a great method to track that expansion. Using a machine called a flow cytometer immunologists can quantify both the expansion of peptide-specific T cells (using peptide-MHC tetramers) as well as determine the behavior of the cells by analyzing cell surface proteins characteristic of various types of activation. These tools and techniques give researchers useful information to help determine (a) the presence, (b) the expansion, and (c) the effector type of T cells in various normal biology and pathologic conditions.

So, given this context, I was trying to determine ‘what CD4 T cells see’ which precipitates transplant rejection, in general. For this study we used a skin transplantation model to study organ rejection in general. We were not addressing autoimmunity against the transplant, which would also be in play in a T1D recipient of beta cells. The reagents we used in this study were peptide-MHC tetramers specific for the MHC of the transplant donor – which would be foreign to the recipient – loaded with peptides derived from cells called dendritic cells. Dendritic cells are key antigen-presenting cells which interact with T cells to influence immune responses. A long-standing hypothesis in basic immunology is that “passenger leukocytes” from the transplant – ie, dendritic cells – are a key target promoting recipient T cell responses. We tested this premise in our paper. Following transplantation of skin, donor dendritic cells travel to lymph nodes or the spleen of the recipient and interact with recipient T cells. This promotes a massive immune response leading to transplant rejection. We made four key discoveries. First, several peptides that are expressed on the surface of dendritic cells and loaded in donor MHC promote CD4 T cell responses in transplant recipients; no individual response dominated, it was a lot of separate T cell responses all happening in parallel. Second, both the MHC of the donor and the specific peptides are required as targets for a specific set of T cells. We showed this using MHC knockout or over-expression models, as well as peptide knockout and over-expression molecules. Third, these individual sets of transplant-reactive T cells are much smaller than sets of T cells elicited by immunization protocols; the very large overall immune response to transplants is a consequence of many of these responses happening in parallel. Fourth, transplant-reactive T cells appear to have a characterize effector type, or phenotype, called Th1. This effector type is also associated with autoimmunity and responses against viruses. The other possibility was an effector type called Th17 which are critical for responses against fungal infections. Our results clearly demonstrated transplant-reactive T cells were not of the Th17 type.

Understanding targets of the immune response which leads to transplant rejection gets us, as a field, one step closer to being able to rationally develop therapies to modulate these immune responses. Combined with our ever-increasing knowledge of autoantigens targeted in T1D onset, we now have two separate, parallel immune responses to modulate in future attempts to prevent rejection of replacement beta cells in individuals with long-standing T1D.

Saturday Morning Research Review – November 25, 2017 Sat, 25 Nov 2017 15:04:56 +0000 Position Statement on Standardizing Clinically Meaningful Outcome Measures Beyond HbA1c for Type 1 Diabetes
by Daniel Schneider, MSc
When I head to the endocrinologist to check in with my doctor, I’m usually curious and waiting for the results of that benchmark test by which we are all sometimes guilty of celebrating or harshly judging ourselves. Our Hemoglobin A1c (HbA1c), a surrogate measure of blood glucose averaged over 2-3 months, is used as the primary clinical outcome to assess diabetes control. However, using HbA1c alone leads to an incomplete picture of diabetes management and limits an outsider’s perspective on the true impact diabetes can have on a patient’s life.
A doctor cannot look at HbA1c and assume to know about waking up sweating with a low blood sugar in the middle of the night, or to know that you’ve successfully treated hyperglycemia with strategic
exercise, reducing the amount of time with a high blood sugar. Thankfully, as diabetes management technology has improved, continuous glucose monitors (CGMs) lead to a more complete picture of the short term changes associated with blood glucose management. With these changes in mind, a group made up of representatives from the major diabetes organizations in the United States attempted to identify outcomes other than HbA1c that are meaningful for patients and their physical, mental and social well-being. The groups involved included American Association of Clinical Endocrinologists, the American Association of Diabetes Educators, the American Diabetes Association, the Endocrine Society, JDRF International, The Leona M. and Harry B. Helmsley Charitable Trust, the Pediatric Endocrine
Society, and the T1D Exchange.
The committee identified 4 further clinical outcomes that impact the life and health of patients. Those outcomes are hypoglycemia (including definitions of severity), hyperglycemia (also including definitions of severity), time in range (>70 and <180mg/dl; >3.9 and <10 mmol/L), and Diabetic Ketoacidosis.

These measures create a more complete clinical picture of diabetes management, but there remains work to be done. If these outcomes are to be applied consistently to patient management, their
measurement needs to be standardized by physicians. Further, more research is needed to establish how these ͞new͟ measures affect the health outcomes of the patient.
The reality is that managing type 1 diabetes is challenging and to summarize disease management with a single outcome/number does a disservice to the patients and the doctor. Establishing these measures as appropriate clinical outcomes will serve to broaden the understanding of managing a complex and dynamic disease.
Read the whole consensus statement here.

Saturday Morning Research Review – October 28, 2017 Sat, 28 Oct 2017 05:02:01 +0000 Another target of autoreactive T cells: defective protein production

by Adam Burrack, PhD

One reason insulin may become a target of autoreactive T cells is because it is associated with inflammation in the pancreas. As I’ve previously described, inflammation is one signal T cells that from their environment which indicates that it’s time to respond to a threat (like a virus or bacteria). Another signal that helps to “activate” T cells to respond to threats are new peptide targets. Normally, T cells that have the potential to cause autoimmune disease are deleted in the thymus during development. In a system where these cells weren’t purged, humans would develop multiple types of autoimmunity at very young ages. In fact, the safe-guard goes a full step further. T cells which have moderately strong reactivity to self (think of it as a spectrum of strength of response) are not deleted, but rather are diverted into a different lineage of cells called regulatory T cells. The sole purpose of this sub-set of T cells is to prevent autoimmunity during times of inflammation and moderate organ and tissue damage. For example, part of the training response to exercise involves moderate amounts of muscle damage – even the release of some damaged proteins – so that the muscles can respond to the stress and become stronger. If we did not have regulatory T cells – as well as different ‘flavors’ of inflammation, the type of inflammation associated with exercise (mediate by the cytokine IL-6) is slightly different than the type of inflammation resulting from viral infection (mediated by type 1 interferon) – we would develop autoimmunity to our own tissues. Therefore, very few – if any – T cells with the potential to cause autoreactivity make it into the circulation of a normal adult human. So the question remains: how does autoimmunity develop?

One answer is genetic susceptibility. For unknown reasons, individuals with certain gene types related to how T cells are activated develop autoimmunity – including type 1 diabetes – more frequently than other individuals. A second answer is slight perturbations in T cell development. For reasons that are not completely clear, individuals with the propensity to develop autoimmunity seem to have a higher-than-normal number of autoreactive T cells. This likely has something to do with how T cells develop in the thymus, but is not understood well enough by basic science researchers in order to manipulate it to prevent autoimmunity. A third answer might be gut microbiome. Emerging evidence has begun to suggest that a more narrow range of bacteria in the gut is associated with autoimmunity, including type 1 diabetes. Whether this is cause, effect, or unrelated to disease onset is not clear at this time.

A fourth potential answer is the development of new peptides associated with beta cell destruction, to which there are no pre-existing T cells. In this case, there would be no regulatory T cells to prevent attack, and there would have been no selection in the thymus during T cell development against the survival of T cells responding to these peptides. From the perspective of the T cells, they would be doing what they should: attacking the source of a new peptide that is associated with inflammation. This is how the immune system has evolved to operate in mammals. The recent observations of hybrid peptides, which contain one portion of the insulin molecule and a second portion from a different protein found in the insulin secretory granule was the first demonstration of this principle in type 1 diabetes.

Today’s article establishes a second example of this “new peptide” explanation for the origins of autoimmunity. In a recent article in the journal Nature Medicine from Bart Roep’s group at City of Hope in Los Angeles, researchers characterized a ‘defective ribosomal insulin gene product’ as a target of both CD4 T cells and CD8 T cells. What this means in English (or at least non-basic science-speak) is the following. When beta cells are stressed out by inflammation and high-demand for insulin production (as in when neighboring insulin molecules are getting killed and there is more demand for each remaining beta cell to produce more-and-more insulin), sometimes the production of insulin gets screwed up in the rush to keep up. This ‘screwed-up’ insulin could then be seen as a ‘new protein’ being produced by cells under the attack. These cells under attack could then be mistaken by the immune system, logically enough, as infected with a bacteria or virus. The immune system then re-doubles its efforts to eradicate the cells responsible for production of this foreign peptide.

From the perspective of the beta cell, that sucks. You’re doing job as best as you can, a couple mistakes get made along the way because you’re just that busy, and that gets you killed – and your neighboring beta cells killed – even faster. From the perspective of the immune system, job well done. And not only that, but memory T cells will be developed against this mis-made insulin target. So that if an insulin-producing transplant is ever performed, the T cells will remember that weird-insulin target.

The good news is, having discovered this basic principal that T cells appear to really be doing their job quite well when they are destroying beta cells, researchers can now begin to think about ways to leverage these ‘new targets’ in attempts to re-create immune tolerance to insulin-producing transplanted cells. Understanding what the targets of autoreactive T cells are – especially the ‘new targets’ like hybrid peptides and defective ribosomal products – gives researchers rational approaches to try to reverse autoimmunity. Most likely this will require a two-step process: (1) get rid of the memory autoreactive T cells (using as-yet-to-be-invented technology), (2) induce immune tolerance to ‘new targetes’ (using unknown future method). So as a reality check, the strategy is clear, the methods; however, are not.


Saturday Morning Research Review – September 30, 2017 Sat, 30 Sep 2017 14:36:57 +0000 Peptide immunotherapy to stop new-onset type 1 diabetes

by Adam Burrack, PhD

The recent observation that so-called hybrid peptides may represent new targets of the immune system which are targeted in the build-up to developing type 1 diabetes has opened new vistas of possibilities for trying to re-establish immune tolerance to pancreatic beta cells. This paradigm-shifting observation may answer the long-standing conundrum of ‘why are beta cells specifically targeted’? If a ‘new’ target (as far as the immune system is concerned, if the target was not shown to T cells in the thymus, then if it is found in the periphery it’s new) is associated with (1) inflammation and (2) dead cells, the immune system will interpret these signals as a call to action to clear an infection. Hybrid peptides produced from dead and dying beta cells fit these two criteria. Viewed from this lens, the immune system is logically attacking beta cells.

The question for basic researchers and clinicians was how to leverage this knowledge to try to prevent or stop type 1 diabetes development in at-risk human patients. Whom do you treat? With which potential hybrid peptide? For how long? What do you measure to track efficacy of treatments?

In a report recently published in Science Translational Medicine, researchers attempt to leverage this new knowledge in a phase 1 (ie, safety) clinical trial to stop new-onset type 1 diabetes in human patients. Several clinical research groups were involved in this work, including Universities and Hospitals in Cardiff (Cardiff School of Medicine), London (King’s College), Newcastle-upon-Thyne, University of Swansea, and biostatisticians at Western Michigan University. This paper has >20 authors, representing a ton of work by many different people at different institutions. This type of large collaborative effort is a good example of how successful, impactful, clinical trials must be conducted.

Some answers to several of the above questions were developed. As with all good research, more questions were raised. Firstly, researchers treated new-onset type 1 diabetic patients, within 100 days of diabetes onset. These subjects were well within the ‘honeymoon’ phase of the disease, meaning they had residual C-peptide production: this is indicative of moderate to significant remaining beta cell mass. Therefore, if a treatment could stop beta cell destruction in these individuals, it might facilitate diabetes reversal. This treatment was not intended to reverse long-term diabetes or replace lost beta cell mass. The goal was to preserve residual beta cell mass in new-onset type 1 diabetic patients.

Second, the researchers went after the most logical target hybrid peptide; proinsulin. As a reminder, insulin is made as an inactive precursor within beta cells, preproinsulin. Various portions of this protein are cleaved (cut off) within beta cells before the active hormone, insulin, is released. The final portion of proinsulin cut off is called C-peptide (for its shape) and can be measured in the circulation. In fact, C-peptide level is one way to measure small amounts of insulin production by remaining beta cells in long-term type 1 diabetic patients. One of the hybrid peptides recently characterized contains portions of C-peptide and portions of the insulin protein. Therefore, to try to re-establish tolerance to this hybrid peptide, or other pro-insulin-derived peptides, patients were treated with pro-insulin. Other recently-characterized targets of autoreactive T cells map to a different portion of the proinsulin molecule. Therefore, proinsulin is a good choice for initial attempts to try to re-establish tolerance to beta cells.

Long story short, these results show promise – this is why it’s a Science Translational Medicine publication. Some individuals responded to the treatment more effectively than others, for reasons which are not fully understood. For responders, the treatment appeared to maintain beta cell mass – as shown by preserved C-peptide levels – and may have promoted the activity of regulatory T cells (though this was not conclusively shown by this study). As I’ve discussed before, regulatory T cells are a key cell type which typically protects our own tissues from attack by the immune system. Adding back these cells in large numbers is a separate potential approach to re-establish tolerance to pancreatic beta cells and reverse type 1 diabetes. These two outcomes, preserving beta cell mass and promoting regulatory T cells, are a great starting point for a new clinical therapy.

Despite these promising early returns for peptide immunotherapy (PIT) in individuals with new-onset T1D, this was only a phase 1 clinical trial and several critical questions remain. First and foremost, this study demonstrated that the ‘honeymoon phase’ was prolonged, not that the T1D disease process was halted. Future phase 2 or phase 3 studies, conducted over longer time frames, will presumably address these questions. Second, while this report suggests enhancement in regulatory T cells, it does not demonstrate insulin (or proinsulin)-specific regulatory T cells, which are the most likely to help protect beta cells. As we have previously described in this series, adding back regulatory T cells specific to insulin peptides is a separate strategy which may synergize with the PIT approach. Presumably phase 2 or 3 trials would more specifically address regulatory T cells. Having stated these limitations, this phase 1 trial is a very promising result, in a high-tier journal, and many researchers (in particular in my sub-field) will be highly anticipating phase 2 results of this and similar studies in the USA. This is a great start toward re-establishing insulin-specific immune system tolerance in new-onset type 1 diabetic individuals.

The full-length manuscript can be accessed here.

Saturday Morning Research Review – September 2, 2017 Sat, 02 Sep 2017 13:43:12 +0000 Further evidence of a role for the gut microbiome in autoimmunity

by Adam Burrack, PhD

I’ve written a couple articles about the interplay between the bacteria living in our gut and our immune system, and how this interaction can influence the subsequent development of autoimmunity. The first time I touched upon this subject, it was to describe work from Mark Atkinson’s laboratory at the University of Florida. In this work, the authors described a relative lack of diversity in gut-resident bacteria in infants and toddlers who went on to develop type 1 diabetes compared to control subjects who did not develop diabetes. This outcome suggests that gut bacteria diversity may somehow protect against autoimmunity. In another highlighted paper later in our series, I described a study from Yale in which researchers demonstrated that a beta cell toxin used to induce diabetes for mouse experiments promotes movement of gut bacteria into the pancreas-draining lymph node following treatment with a beta cell toxin. This result suggested that beta cell death and inflammation induced by this chemical promoted movement of bacteria as well, supporting a role for bacterial products in activation of the immune system following treatment with this beta cell toxin. Thirdly, an article I described earlier this year demonstrated that a diverse diet promotes a diverse range of gut-resident bacteria, and the reverse: an “American” high fat-high protein diet promotes both a narrow range of gut-resident bacteria along with high levels of systemic inflammation. This suggests that our diet has a clear ability to affect our gut-resident bacteria; the bacteria that thrive in our intestine will be those that are best-adapted to utilize the fuel available. The fuel available in our gut is determined by the foods we eat.

Enter into this conversation three recent papers studying the relationship between our gut bacteria and the timing of type 1 diabetes (T1D) onset in at-risk individuals.

First, a study from Finland and published in the journal Pediatric Diabetes in November 2016 characterized both the bacteria and the viruses in the gut of individuals at-risk of developing T1D. In particular, the authors were interested in the prevalence of a type of gut-resident bacteria called Bacteroides, which is a common component of the gut microbiome but could be considered an opportunistic pathogen. An opportunistic pathogen means Bacteroides might mediate harmful effects if its population were out of proportion to other gut microbes. No previous study had established a pathogenic role for any specific gut bacteria in subsequent autoimmunity. Results from this study suggest that a particular bacteriophage (bacteriophages are viruses which attack bacteria), which targets Bacteroides may be more prevalent in the gut of toddlers who go on to develop T1D. These results should be interpreted with caution; this study demonstrates a correlation between changes in the prevalence of these gut microbes and later development of T1D. This study does not establish any of these changes as causative of beta cell destruction.

Second, in a publication in the March 2017 issue of Diabetologia a partially overlapping set of authors from Finland examined the presence of enterovirus and whether the presence of this category of gut microbe at higher-than-normal levels might predict later beta cell destruction. Enteroviruses are named for their life-style; they live within the intestine. These authors examined stool samples from 129 ‘case children’ (who went on to develop islet autoantibodies) and 282 control children who did not develop islet autoantibodies. This was a prospective study, the authors’ aim was to determine if infection with any particular enterovirus serotypes correlated with later development of islet-specific autoantibodies. These authors found that children who developed autoantibodies experienced more frequent enterovirus infections than control subjects. Coxsackie virus is a sub-type of enterovirus which has been hypothesized to play a role in beta cell destruction. The presence of several types of coxsackie virus were associated with later development of islet autoantibodies: A4, A2, and A16. Importantly, these infections occurred more than 1 year prior to development of autoantibodies, suggesting a slowly-evolving autoimmune process. These results suggest these types of coxsackie virus may help to precipitate later development of islet autoantibodies. Caution assigning causation to these serotypes, however; this was an observational study and this work does not describe how these viruses promote autoreactivity against insulin-producing beta cells. Future studies are required to determine how these viruses might promote autoimmunity specifically against beta cells in the pancreas and not against other hormone-producing cell types within pancreatic islets.

Third and most recently, scientists at Washington University in St Louis and collaborators characterized the stool microbiome, beginning at birth, of 11 at-risk children. Five of these 11 children have subsequently developed clinical T1D, compared to none of the control subjects in this study. In general, these authors observed less diverse ‘gut viromes’ in subjects who subsequently developed T1D than in controls. In other words, the smaller the range of gut microbes, the more likely the subject was to develop T1D. In addition, the authors demonstrate that bacteriophage diversity was significantly lower in T1D at-risk subjects than controls – in agreement with the study cited above. Here’s why this paper was published in Proceedings of the National Academy: moving toward causation, since the authors collected stool samples over the lifespan of these subjects they could test how gut microbe diversity related to T1D development. The authors found that decreases in gut virome diversity in at-risk subjects who developed T1D were different from control subjects who did not develop T1D. In other words, these changes in gut microbe diversity may be causative of beta cell-specific autoimmunity rather than a consequence of the disease process.

This is exciting news. Next steps include determining which bacteria and bacteriophage might be protective versus which might be pathogenic, or whether it’s ‘just maintain microbe diversity’ to stave off T1D. Then, how and why do some microbes promote beta cell-specific autoimmunity? It’s not clear to me whether coxsackie virus infection of beta cells matters, or whether its beta cell-specific inflammation in general that promotes autoimmunity in at-risk individuals. Whatever the answer to these questions, clinical translation of these results – harnessing the diversity of the gut microbiome to prevent or reverse beta cell-specific autoimmunity – is a long way off. But these initial observations establishing that there might be a causative relationship between decreased gut microbe diversity and subsequent beta cell-specific autoimmunity is a very interesting start.

In other words, these results might represent the “end of the beginning” of this sub-field. It looks, to me, like the gut microbiome ‘matters’ for developing autoimmune diabetes. The question of mechanism (eg, decreased diversity of gut microbes precedes T1D development) and what we do about this change to prevent diabetes (eg, maintain gut microbe diversity, somehow) are more challenging to address, but represent the next key steps in leveraging this knowledge to prevent T1D in some at-risk individuals.


Saturday Morning Research Review – August 12, 2017 Sat, 12 Aug 2017 05:53:44 +0000 Glucagon dynamics during exercise with type 1 diabetes               

by Adam Burrack, PhD

Previously in our series, we have described the exercise science of maximal performance in terms of oxygen consumption (basically 2-mile race pace) and how imperfect, artificial, control of insulin levels during exercise may negatively affect exercise performance. Too much insulin in the circulation during intense exercise could lead to low blood sugar levels, which can be dangerous in fast-moving sports. Too little insulin in the circulation can lead to high blood sugars – and potentially diabetic ketoacidosis – in long-duration endurance events.

But what about the other side of the blood sugar level thermostat? Glucagon is the antagonist and mirror image of insulin. Since insulin levels can be so out-of-whack in the exercising T1D, how does this effect glucagon levels and the efficacy of glucagon that is produced? Since the main function of glucagon is to increase blood sugar levels, during prolonged high-intensity exercise (think 5k, 10k, or half marathon running races), it would be useful to have a little boost from our internal glucagon to ‘make sure’ our blood sugar levels don’t crash. That is, in fact, the normal sequence of events in non-diabetic individuals in these situations: levels of insulin in the plasma go way down, and levels of glucagon to up. At running paces approaching 75% of VO2 max (approximate half marathon pace for most of us), or faster, this relationship between insulin and glucagon is key for mobilizing fuel to get us to finish line without bonking. In other words, in this situation glucagon is your friend, big time.

It is known that some people with type 1 diabetes will eventually develop hypoglycemia unawareness, a potentially life-threatening condition. This condition is the current clinical prerequisite to go onto the pancreas transplant waiting list. We also know that at least some of the autonomic nerves going into pancreatic islets are destroyed as part of the disease process of type 1 diabetes. These nerves – which likely carry signals both from the pancreatic islets to the brain stem and back from the brain stem to the islets – do not appear to be destroyed during the development of type 2 diabetes. This is one of several key differences between type 1 and type2 diabetes. Today, we will delve into the biology of glucose metabolism during exercise in the individual with type 1 diabetes through the lens of glucagon, the antagonist of insulin.

The lab of Ananda and Rita at the Mayo Clinic in Rochester MN has been studying glucose metabolism for >10 years. This husband-and-wife team have published on a range of topics, including glucagon dynamics in individuals with type 1 diabetes during exercise. Long story short, type 1 diabetic athletes have more insulin and less glucagon in their circulation during steady-state exercise (60% VO2 max for 60-75 minutes) than do non-diabetic individuals. This represents a sort of “double jeopardy” for the T1D athlete at low-to-moderate exercise intensity. Due to injections of insulin (reminder that normally insulin goes essentially straight from the pancreas to the liver, where it directly promotes glycogen production and energy storage), they have way more insulin in their blood during exercise and also significantly less glucagon, creating a situation where they are at very high risk for hypoglycemia. This mirrors empirical results referenced in the JDRF exercise guidelines for T1D athletes. One method around this situation is to consume 30 grams of carbohydrate (or more) per hour of moderate intensity exercise. Another option is to tend toward higher-intensity exercise – with the attendant risk of low blood sugar levels after high-intensity exercise. Regardless of approach, the athlete with T1D being aware of the fundamental problem – and not blaming themselves for poorly managing their during-exercise blood glucose levels – is the point.

Another clinically applicable area the Basu lab is engaged in is development better methods to track glucagon levels and to deliver glucagon in tandem with insulin, in a stable form. To track glucagon this research group is using of labeled-water methods to track metabolism over the course of an in-hospital study. This group is one of several working – in collaboration with industry partners – to determine more stable formulations of glucagon that would be remain useable for up to 5 days in an insulin pump-type device. Finally, the Basu lab is exploring methods of glucagon delivery (ie subcutaneous as insulin is currently delivered via pumps compared to intravenous delivery) that was optimize its physiologic function. Overall, the field has much more experience with making and delivering insulin than glucagon. There is some homework to be done to get glucagon up to speed for a widely applicable dual-hormone replacement system. And there are some clever, hard-working folks working on these challenges.


Saturday Morning Research Review – July 1, 2017 Sat, 01 Jul 2017 05:49:52 +0000 Top options for beta cell replacement

by Josh Boyer, MS

The one thing that any cure for type 1 diabetes needs to accomplish is to replace the beta cells that have been destroyed by the autoimmune destruction of islets.  Beta cell replacement strategies are an important issue that applies to both the development of technology to manage diabetes (encapsulated islets) and potential cures (transplantation of islets; reviewed by Adam here).  Along with protection from immune attack and nutrient/oxygen delivery this is a very important aspect for the success of encapsulation or any cell based diabetes therapy.  Because of a limited supply of cadaverous donors for islets and difficulty in obtaining viable cells from donated organs, beta cells will have to come from alternative sources.  Two proposed options are stem cell-derived human beta cells or pig xenograft transplants.

The pig option is an interesting one that I will briefly touch on.  Pigs have a long history in the treatment of diabetes.  The first insulin developed to treat type 1 diabetics was sourced from pigs (first identified by the Nobel Laureate Frederick Banting in 1922) and used until recombinant DNA technology developed by Genentech replaced it in the 1970s. Pigs maintain similar physiologic levels of blood glucose to humans and transplanted porcine islets have shown success in non-human primate.  This indicates that pig insulin can function promote glucose uptake by human cells and that pig beta cells have the potential to maintain glucose levels at an optimal target.  In fact pigs have been used as a model animal to study the functionality of both closed-loop insulin pump systems and encapsulated islets.  The main drawback to this option is that the transplanted beta cells are foreign cells and have a high probability of being rejected by the immune system.  One interesting recent development is the work on genetically engineering donor pigs to develop ‘humanized’ cells.  This is a research area being explored not just for diabetes but multiple diseases that require transplanted organs or tissues.  While it seems fairly science fiction now and there have been recent setbacks, science can also sometimes move rapidly and this may be an option that warrant further investigation.

My current favorite choice for expanding beta cells is to use induced pluripotent stem cells (iPSCs).  A stem cell has the ability for both self-renewal (generating more of itself) and differentiation into functional cell types.  A pluripotent stem cell means that it can differentiate into any cell type in the body.  Induced pluripotent stem cells were first described in 2007 through simultaneous work by Shinya Yamanaka at Tokyo University and James Thomson at UW-Madison.  iPSCs are stem cells derived from cells that have already differentiated into specific functional cells.  They are returned to a pluripotent state through addition of signaling molecules and growth factors and from there can be coaxed to develop into a different cell type.  The expansion and differentiation of beta cells from stem cell progenitors has been well described and Adam has reviewed this strategy previously.  The main advantage that I see from this technique is that in theory you could take a diabetic’s skin cells, turn them into iPSCs and develop beta cells from those that will be a perfect match for transplanting as they are from the patient.  This doesn’t block the autoimmune response but avoids the chronic transplant rejection issues that allogeneic or xenograft transplantations face.

Within the beta cell replacement field, something I’ve come across regarding encapsulation are data showing that the ideal cells to transplant are not fully differentiated beta cells but rather beta cell progenitors.  Beta cells develop from the endoderm through a multipotent pancreatic progenitor cell that progresses to an endocrine progenitor.  It is these pancreatic progenitors that are being used within encapsulated devices.  One advantage to this design is that circumvents (ie avoids) the hypoxic stress that newly transplanted islets experience.  Following implantation, the progenitor cells appear to be able to expand and maintain a larger number of beta cells than transplanted whole islets.  Following transplantation into a diabetic recipient, these cell clusters differentiate into ‘islet-like cells’ that include insulin-producing cells.  Current work in several groups is trying to increase the number of beta-like cells within these progenitor populations.  One concern with this method is the risk of a tumorigenic cell population developing because of stem cells’ self renewal properties.

Beta cell replacement is a very important area of research for type 1 diabetes therapies and cures.  However, there is also a chance that these transplantations may not even be necessary.  Recent work has shown that undifferentiated beta cells persist in the islets even in diabetic patients.  This work demonstrates that there may be trans-differentiation from alpha cells (that produce glucagon) to beta cells.  These cell populations may not exist for all diabetics (there is variability in degree of beta cell destruction) and even then these cells may not be able to expand to form a functional beta cell mass.  However, it is worth investigating whether they may be driven to produce sufficient mass as this would be an ideal fix that requires no transplantation.  Once we fix that autoimmune issue, of course.

Saturday Morning Research Review – June 24, 2017 Sat, 24 Jun 2017 12:05:45 +0000 Dietary strategies for management of Type 1 Diabetes – The Paleo Diet

by Rachel Fenske, BS

I am excited to bring you my second post in a special series for YOUglycemia regarding the complex interactions of type 1 diabetes pathophysiology, diet, and exercise.

Today, I will be reviewing what’s been termed the “Paleo” diet, or “cave-man” diet. The anecdotal indications of the benefits of a paleo diet for T1D management are growing. The paleo diet encourages essentially unlimited consumption of fresh vegetables, lean meats, fish and eggs, and more limited consumption of fruits, nuts, and seeds. Most notably, dairy and grains are prohibited. Prior to the very recent development of commercially available “paleo” snack products, like crackers, tortillas, baked goods, etc., all processed foods were also prohibited on the diet. Thus, a type 1 diabetic following this diet would be consuming a diet high in protein and fat, low in simple starches, while not void of carbohydrates entirely as vegetables would provide a variety of complex starches. These dietary conditions would result in less blood glucose spikes and a general reduction in the volume of daily insulin required, meaning the validity of the anecdotal evidence is somewhat substantiated.

The efficacy of low-carbohydrate diets for individuals with T1D has been evaluated briefly, with a consensus from most studies showing an improvement in blood glucose control and lowering of HbA1c. These improvements are extremely dependent on the level of adherence to the diet. Adherence to the diet is very important due to the dramatic reduction in required insulin. Blood glucose management would become extremely difficult if for one day you decided to switch back to a more standard diet that contained a higher carbohydrate load.

Additional considerations for athletes need to be taken into account as carbohydrates are a crucial source of fuel and are limited on the paleo diet. Primary among these considerations are enhanced insulin sensitivity following exercise and the replenishment of carbohydrate stores on a diet which limits carbohydrate intake. Other studies have attempted to evaluate these concerns for athletes on low carbohydrate diets and have found mixed results dependent on type and duration of exercise. Importantly, results also vary based on how the remaining calories are divided among fats and protein. Low carbohydrate diets themselves are inherently higher in fats and/or protein and this can have a large impact on fuel source availability for an athlete. The complexity of high protein and high fat diets elicits a more thorough evaluation and will be the focus of future posts in this series.

Studies focusing specifically on the “paleo diet” began building in the early 2000s, not all that long ago. One of the earliest proponents and foremost researcher of all things Paleo diet is Dr. Loren Cordain. His interest and research have notably culminated in the popular N.Y. Times Best Seller, “The Paleo Diet”, and more recently “The Paleo Diet – Revised” for athletes. He has worked diligently to synthesize and elaborate on the current body of evidence into a comprehendible, but extensive read. His peer-reviewed work encompasses Paleolithic nutrition and fitness, and has recently focused on the role of milk protein in age-related metabolic diseases.

In addition to the work by Dr. Cordain, some of earliest studies investigating the diet are in healthy volunteers, those with impaired glucose metabolism, or those with type 2 diabetes. Unfortunately, more studies have not strayed from these populations. The data emerging from these studies is mixed (reviewed thoroughly here ); in some cases showing improved blood pressure , insulin sensitivity , and HbA1c , while showing no improvement in others . Markers of inflammation and intestinal permeability, which have been discussed previously in the “Science of Diabetes” series, were shown to be no different, in one study of 24 patients with metabolic syndrome .

Although not directly relevant to type 1 diabetics, these studies of healthy subjects are necessary to establish a baseline effect of the diet and studies of individuals with other metabolic conditions are therefore important to developing a breadth of understanding. From this baseline, researchers could then layer-in studies with T1D subjects and note key differences from healthy controls. To further the research, studies which specifically look at metabolic measures and inflammatory markers in type 1 diabetic patients, as well as subjects with other autoimmune conditions, are necessary to provide a more direct implication of paleo diet on T1D glucose control and other outcomes.


Based on the current body of evidence, it cannot be determined if the paleo diet would be an appropriate therapeutic dietary intervention for individuals with T1D. Far more research is warranted to determine the potential influence such an extreme dietary shift would have on blood glucose management for those with T1D and even more work is required to discern if a paleo diet strategy would be appropriate for an athlete with T1D.