We are convinced that one of the drivers of the onset of type 2 diabetes is a slow inflammatory process. If this is true, wouldn’t an anti-inflammatory drug like aspirin slow or even stop this process? Since I do not know the answer, this column provides me the perfect excuse to spend the time to find out!
In doing a literature search one of the first things I found was that data was generated over a century ago showing that aspirin helps with blood glucose levels in diabetic patients. Papers were published as early as the 1870s (Ebstein W. Zur therapie des diabetes mellitus insbesondere uber die anwendeng der salicylauren natron bei demselben. Berliner Klinische Wochenschrift 1876; 13:337-340 is one example). However, no one at that time had inkling as to how aspirin worked and the discovery of insulin in the early 20th century completely overshadowed these results.
So, let’s review inflammation and how drugs like aspirin inhibit it. Inflammation is the body’s response to danger. Danger could mean an infection (you’ve been invaded!) or it could mean trauma (whether car accident or cut finger…) or it could mean the presence of a chronic problem (like a metabolic imbalance). Sentinel cells called macrophages detect that a problem exists. How they “know” that they are dealing with a pathogen versus just trauma is a whole article in and of itself. Some of these cells leave the tissue and migrate to the nearest lymph node. This functions like a military base and is the home for countless T cells and B cells. In the lymph node, the macrophage gets queried and if a T cell happens to recognize one of the antigens the macrophage brought with it, then a full immune response is initiated.
Meanwhile, back at the site of the infection, other cells are busily getting ready for the immune response. They are sending out signals so that the immune cell army knows where to go. This involves making subtle changes to the capillaries (more accurately, the post capillary veinules) in that region of tissue so that they become leaky. The tissue swells and feels painful as fluid enters. This swelling (which doctors call edema) will make it much easier for the immune cells to move around and search for pathogens to destroy.
Some of these chemical signals being secreted by the tissue are long lived and will slowly diffuse out to provide a trail that the lymph node cells will use to find their way to the site of infection. Indeed one of the reasons we think that diabetes involves an inflammatory process is that we can measure increased amounts of these proteins in the blood of patients with either diabetes or pre-diabetes. Other chemical signals are quite short lived and are designed to act locally to organize the tissue for the upcoming war against the invader. It is this family of local signaling molecules that get inhibited by aspirin at standard doses that can be tolerated for long periods of time. They go by the name: prostaglandins. NSAIDs like aspirin or ibuprofen (Advil) or celecoxib (Celebrex) all work by blocking the enzyme that synthesizes the precursor for this family of signaling molecules. In doing so they shut down the ability of the tissue to undergo the changes necessary to support the process of inflammation.
At high doses NSAIDs do other things as well. Drugs are funny that way. They don’t always bind to just one molecular target. In this case we are lucky in that some NSAIDs bind to one of the master switches of inflammation and the immune response. The protein (called NF-kappa B) binds to DNA and turns on all sorts of immune and inflammation genes in many different cell types including all of the ones we have just talked about.
So to sum up what I have told you, so far, inflammation involves a number of separate events that need to take place including activation of sentinel cells, migration of some of these cells to a nearby lymph node to notify the T cells, secretion of signaling molecules into the blood, and remodeling of the damaged or invaded tissue (especially blood vessels) to prepare for an immune response. At moderate doses all NSAIDs inhibit local signals (prostaglandins) which drive the remodeling of the tissue. At high doses, some NSAIDs inhibit NF-kappa B, a key regulator of inflammation.
Returning, finally, to the initial question that started this piece; clinicians have not forgotten the work of Ebstein and his colleagues. Clinical studies have moved forward slowly over the decades. In the beginning there was only aspirin so the early clinical studies looked at that. A host of small studies (meaning less than 20 participants) looked at aspirin administered to diabetic patients for a short period of time (1 – 2 weeks). The results pretty much reproduced the much earlier work of the 1800s. Blood glucose levels were decreased as well as glucose in the urine. When measured, patients’ responses during the glucose tolerance test were improved. Importantly, in healthy patients taking the same aspirin doses there was no effect on glucose tolerance. These sorts of controls are essential to interpret data. What it says is that aspirin does not lower blood glucose directly but rather, only works when the patient has diabetes (suggesting that it is working via decreasing the underlying inflammation). A study in support of aspirin decreasing diabetes risk was one performed in 2004 by Hu and colleagues. They followed 32,826 nurses and compared inflammatory markers (specifically C reactive protein), the development of diabetes, and aspirin usage. They found that women who reported using aspirin regularly were less likely to develop diabetes and had lower levels of C-reactive protein in their blood as compared to the women who did not use aspirin. This was followed by a similarly large study in 2009 by Pradhan and colleagues. They examined 38,176 non-diabetic women, half of whom were taking low dose aspirin (100 mg every other day) for 10 years. There was no difference in diabetes risk between the two groups (i.e. equal numbers became diabetic over the 10 year observational period). Yet another study was published in 2009 by Hayashino and colleagues. This group examined 22,071 non-diabetic males who took 325 mg aspirin every other day for 22 years. There was a correlation with decreased risk of developing diabetes that measured about 14%. So how do we interpret these studies?
The thing to keep in mind is that inflammation involves many independent processes. As I discussed a couple of articles ago, only platelets are strongly affected at this dose. This has an important effect on clotting but only weakly affects some subset of inflammation pathways. High dose aspirin seems to show good effects on restoring insulin sensitivity in small clinical trials but these high doses almost invariably caused intolerable side effects such as tinnitus or headache in some participants. Interestingly, no stomach pain was reported which was surprising (to me at least). This would never work as a long term therapy. Perhaps the mechanism of action does not so much involve blockade of prostaglandins but rather, inhibition of NF-kappa B. The pharmaceutical industry was and perhaps still is trying to get NF-kappa B inhibitors into the clinic. Unfortunately they are plagued with toxicity issues.
Nevertheless, the study by Hayashino et al shows some promise. The decreased risk for diabetes was somewhat similar to the decreased risk for cancer. What is missing in all of these studies of tens of thousands of people is a way to address our genetic diversity. My personal feeling (and the feeling of most of the scientific community) is that our genetics plays a huge role in how we respond to therapy. Luckily for us we now have the technology to sequence entire genomes. It costs a bit (around $15,000 per genome) but the price is coming down. Our goal is to sequence your genome for $50! How we can do this and what it means for diabetes is a worthy subject for the next series of posts.