My Body Is Not Spam: Retraining the Diabetic Immune System

One of the conceptual problems I have with the state of Type 1 diabetes treatment is that, as with many complex and uncured diseases, all we can do is try to lessen the effects of the symptoms of the disease. Unmodulated blood glucose– the hallmark of diabetes– is only an incidental result of the real problem: my body’s immune system has labeled the insulin-producing beta cells in my pancreas as dangerous, and has killed them off. Without these insulin-producing cells, my body can’t regulate blood glucose levels on its own.

The problem is, the immune system’s attack is hard to reverse and hard to untangle; the best solution we have so far is to inject insulin, and manually control my blood glucose levels so that elevated blood sugar doesn’t overtime lead to worse symptoms– organ death first, and, ultimately, just death.

Insulin treatment, then, is necessary, and it is a God-send for diabetics like me (have you read the story of the first use of insulin in the 1920s? Ridiculously Epic, with a capital E; why isn’t that a movie yet?) But, it’s not a solution, and it doesn’t address the root of the problem: the immune system.

So what is the immune system, anyway? And what does it have against my beta cells? These are the questions that I have been pondering and researching for the past week or so. I’ve only just dipped my toes in the vast body of scientific literature about the immune system, but here’s what I’ve determined so far:

The immune system is my body’s spam detector.

Basically, the immune system is a biological classification system. In the simplest terms, its job is to label entities found in the body as “pathogen” or “safe”– much like your email system labels email as “spam” or “eh, let it through.” The trick is, as with any good classification system, the world is not black and white to my immune system; every decision is a question of weights and probabilities, of likelihoods and comparisons. Is entity X a pathogen? Well, this molecular compound indicates maybe, but only if it co-occurs with another particular molecule that I don’t see here; let it by for now.

As with a spam detector, certain rules are innate, built in to my immune system. Any email with no subject and only a link is spam, no matter how many times I tell my spam detector it’s not. Likewise, certain bacterial mechanisms and molecular structures are known at birth by my immune system to be unwanted, and it doesn’t have to develop immunity over time.

Other rules, though, are learned; spammers are smart, and they keep changing to try to trick spam detectors. When all emails with “Viagra” and a link were marked as spam, spammers started sending “V1agra” and paragraphs of Old English text. Enough of those emails get labeled as “spam” by their recipients, and the spam detectors learn– both “Viagra” and “V1agra” are bad, and so is random text from old books. Similarly, my immune system might let certain pathogens through the first time around, but, seeing the ill-effects on other parts of the body, and seeing that the pathogen causes bad reactions, will train T-cells to look out for that particular pathogen, and to lie in wait if it ever comes back.

The problem is that sometimes classification systems get a bit overzealous and learn rules they shouldn’t. In spam detection, this means all my boss’s emails keep getting caught in the spam filter, and I miss important pieces of information. In immunology, this means I have an autoimmune disease, and my immune system keeps labeling my precious beta cells for destruction.

Why does this happen? Somewhere during training, the classification system learns rules that it applies to each new scenario, and it layers these rules to come up with a likelihood that the entity at hand is either good or bad. Is the email short? Does it have links? Is it addressed by name? Is the recipient in the to-line or BCCed? No single feature here gives a certain positive, but merely adds a weighting to the likelihood overall that the email is spam.

Similarly, in the immune system, maybe at some point I got a virus that had a certain peptide– GAD, let’s say, for the sake of example– and my body learned to look for GAD, and to assign protein sequences that had the peptide GAD with a certain weighting towards “pathogen.” Well, it just so happens that beta cells also have that peptide sequence. And it also just so happens that perhaps genetically, perhaps by the luck of the draw, perhaps by the individual variation that is possible in any biological system, I have too few NKT cells in my immune system. Or too weak regulatory T-cells. Or some minor imbalance of another kind. And the weighting against that one peptide doesn’t get out-weighed by some other factor of my beta cells. And my immune system labels my beta cells as pathogens.

And the rest is history. Once the beta cells are initially labeled, the attack begins. To complicate matters, over time my immune system, out of balance and unaware that it’s doing damage to its own body, begins to think it’s doing the right thing by killing beta cells. And so it starts to train itself based on other features of the beta cells. Then it’s not just the initial peptide that indicates pathogenicity, but others associated with beta cells and insulin generation likewise get weighted as likely indicators of an unwanted entity. And pretty soon every beta cell in my body has been tossed into the spam folder, and my body is incapable of producing the insulin it needs to regulate blood sugar, and I am a Type 1 diabetic.

So. What do I do? For now, I treat the symptoms with insulin. But what about going forward? How do we address the root of the problem?

Well, what do I do when my spam detector gets overzealous, and starts marking safe emails as spam? I retrain it, of course– first, I make sure to watch the spam filter, and un-label every incorrectly labeled email. I pull them out of the spam folder. I create a separate, explicit rule that tries to force certain emails into accepted folders. I stamp my feet and gnash my teeth and tell the computer, “No, I need these emails! MY BODY IS NOT SPAM.”

Oops, I think I just lost track of my metaphors there. But with a purpose: how do I tell my immune system that my beta cells– and all their molecular markers– are not spam? How do I retrain my immune system? Is such a thing possible? What if I started to expose it to molecular compounds mixing GAD, insulin, and the other beta cell biomarkers co-occurring with molecules my body knows to be good? The FDA just approved a prostate cancer drug that trains the immune system to know cancerous cells as pathogenic, and we have centuries of successful training in the form of vaccines; what has been found in the realm of reverse-vaccines and un-teaching the immune system rules it has learned?

If you know the answers to those questions, please let me know. Otherwise, that’s the plan for the next few weeks of pondering and researching.

Interested in knowing more about the immune system and Type 1 diabetes? This is a nice primer. With pictures!

Karmel Allison

Karmel was born in Southern California, diagnosed with Type 1 Diabetes at the age of nine, and educated at UC Berkeley. Karmel now lives in San Diego with her husband, where she is loving the sunshine, working in computational biology at the University of California, San Diego, and learning to use the active voice when talking about her diabetes.

Karmel Allison

Karmel was born in Southern California, diagnosed with Type 1 Diabetes at the age of nine, and educated at UC Berkeley. Karmel now lives in San Diego with her husband, where she is loving the sunshine, working in computational biology at the University of California, San Diego, and learning to use the active voice when talking about her diabetes.