Insulitis: what physically happens in the attacked pancreas

Insulitis is the inflammation of the islets of Langerhans, the small clusters of cells in your pancreas that contain the insulin-producing beta cells. The term describes the phenomenon by which some cells of the immune system (white blood cells) gather around an islet and penetrate inside it to attack the beta cells. Insulitis is considered the hallmark of T1D — a sign visible under the microscope that your immune system has turned against your own insulin-producing cells [1].

According to the international consensus criteria, we speak of insulitis when at least three islets each contain a threshold of a minimum of 15 immune cells (marked with CD45), and the inflammatory infiltrate is made up predominantly of lymphocytes [1]. The number of cells involved is small, because insulitis in T1D is a discreet inflammation, hard to observe, which is why this lesion was for a long time difficult to define and to find.

An islet attacked by the immune system looks under the microscope like a cluster surrounded by small, dark cells (the immune system cells), which crowd at the edge of the islet and sometimes penetrate inside it. Doctors distinguish the situation in which the immune cells stay mainly at the periphery of the islet (mild attack, peri-insulitis) from the one in which they spread among the functional cells inside (advanced attack). Most immune cells remain at the edge of the islet and relatively few directly touch the beta cells. The attacked islet stains progressively more weakly for insulin, a sign that it is losing its beta cells [2].

A characteristic feature is the islet emptied of insulin-producing cells, which nonetheless keeps its other types of cells, such as the alpha cells (which produce glucagon) and the delta cells (which produce somatostatin). These islets are smaller and no longer stain for insulin. The immune attack withdraws after the beta cells have been completely destroyed, so the emptied islets look quiet, without inflammation. In a pancreas with T1D there is a mosaic: some islets look normal and full of insulin, others are actively inflamed, and many are emptied of beta cells [2].

The infiltrate is dominated by T lymphocytes, the most numerous being the cytotoxic T lymphocytes (CD8+), which directly attack the beta cells. Also taking part, in smaller numbers, are macrophages (CD68+), helper T lymphocytes (CD4+), B lymphocytes (CD20+), plasma cells (CD138+), rare regulatory T lymphocytes (FoxP3+) and occasionally natural killer cells (CD56+) [3]. B lymphocytes are rare in early insulitis and become more numerous as the destruction of the beta cells progresses, being more abundant in children diagnosed under 7 years of age.

A surprising aspect is that human insulitis is an inflammation of low intensity. The actual number of infiltrating cells is modest, much smaller than the dramatic inflammation observed in laboratory animals or in other autoimmune diseases [3]. This discreetness is one of the reasons why insulitis was for a long time hard to recognize and was considered an elusive lesion, the attack being so mild that it can pass unnoticed on an ordinary examination.

The insulitis process is not spread uniformly throughout the pancreas, but appears in a “dalmatian” pattern, in the form of small spots. The pancreas is anatomically organized into lobules, and the disease tends to affect a whole lobule at once: a lobule full of islets emptied of beta cells can sit right next to a lobule whose islets look almost normal, full of insulin. At clinical onset (stage 3), the loss of beta cells is severe in most lobules, but neighbouring lobules with an almost normal number of beta cells can persist [4].

It is not fully clarified why this happens. The main hypothesis is that the autoimmune process advances lobule by lobule, possibly because of the way the blood vessels, nerves and ducts are organized inside the pancreas [4]. This distribution in the form of scattered patches has an important practical consequence: a small sample (for example a biopsy) from a single area could hit a lobule untouched by inflammation and completely miss the diagnosis of insulitis.

The basic principle is that the immune system attacks the islets that still contain beta cells. Studies on the pancreas show that insulitis affects predominantly the islets that contain insulin (approximately one third of them), while the islets emptied of beta cells are almost completely spared [5]. Once an islet has lost its beta cells, the immune system cells lose their target and leave, leaving that islet untouched from then on.

An islet without beta cells becomes practically invisible to the immune attack and can persist that way indefinitely. In addition, there are also islets that keep their beta cells without having yet been found by the immune system. These mechanisms explain the diversity of the appearance of the islets within a single pancreas: the inflammation associated with the immune attack moves unevenly, avoiding the islets emptied of beta cells [5].

Most of the information about human pancreatic tissue comes from fragments taken from deceased organ donors, because the pancreas of a living person is very hard to examine. There are official international networks that collect pancreases from deceased donors who had T1D or who only had autoantibodies, without having yet developed the disease (stages 1 and 2), and distribute them free of charge to researchers all over the world [6]. A historical collection also keeps tissue obtained especially from children who died close to the moment of diagnosis.

These tissue collections have brought to light the fact that insulitis is real, discreet and distributed in the form of many uneven patches throughout the pancreas. It was possible to observe directly how the loss of beta cells is lobular, with the emptied islets coexisting with islets full of insulin, and the attack targets mainly the islets that still contain insulin [6]. It should be noted, however, that the samples studied internationally come from a relatively small number of donors.

The landmark study is a research programme from Norway (DiViD), in which the team took small fragments from the tail of the pancreas from a few young adults, just a few weeks after the diagnosis of T1D. There was a suspicion that tissue from deceased donors, taken a few hours after death, might not faithfully reflect the living organ, so these biopsies offered a unique image of the living pancreas, right at the onset of the disease [7].

The study confirmed that the criteria for insulitis were met in all patients, with the infiltrating cells being mainly T lymphocytes located around the islets (peri-insulitis), and an important reserve of islets full of insulin was still present, some completely normal and capable of secreting insulin [7]. The programme also showed the risks, however: it was stopped because of serious complications (bleeding and leakage of pancreatic juice), which explains why such biopsies cannot become a routine procedure.

No. The age at which the disease appears strongly influences the way insulitis looks, with two patterns existing in general. In children diagnosed before about 7 years of age, insulitis is more aggressive, affects a larger proportion of islets, is rich in B lymphocytes and is associated with a rapid, almost complete loss of the beta cells. In those diagnosed at 13 years or later, insulitis is milder, contains few B lymphocytes and leaves many more islets full of insulin intact [8]. Those diagnosed between 7 and 12 years can have the features of either of the two patterns.

An important practical consequence is that patients with onset at older ages keep at diagnosis a significant part of the insulin-producing islets, which suggests that in them the functional deficit matters more than the absolute loss of beta cells [8]. This difference has implications for treatment, because therapies that target B lymphocytes could act differently in the two groups. Observations from the pediatric population also show that a younger age at onset is associated with a more severe form of the disease and with more pronounced pancreatic inflammation [9].

Yes. Although insulitis is most intense and most frequent close to diagnosis, especially in young people, today it is clear that the autoimmune attack is chronic and can continue, at reduced intensity, for years. In tissue from donors, insulitis and surviving islets full of insulin have been found many years after diagnosis, and the presence of beta cells together with insulitis a few years after onset shows that islet autoimmunity extends well into the period after diagnosis [5].

This persistence changes the way we understand T1D: it is not a single destructive event that ends at diagnosis, but a continuous and dynamic process. As long as islets containing insulin remain, they remain possible targets, so the immune attack continues, admittedly at very low intensity [4]. This chronic character is, however, encouraging, because it means that there may be a prolonged window in which the surviving beta cells could be protected or saved through a treatment.

Remarkably, yes. A study dedicated to people who have been living with insulin-dependent diabetes for 50 years or more (the Joslin Medalist Study) showed that a large proportion of them still produce a detectable amount of C-peptide, the molecule released together with insulin — proof that the body still makes a very small amount of its own insulin. Some of these people even respond to a meal with a rise in C-peptide, which overturned the conviction that the beta cells are destroyed completely and uniformly [10].

Examination of the pancreas in these people confirmed the observation at the tissue level too: all the pancreases examined kept insulin-producing beta cells, isolated or in small groups [10] [11]. Some cells showed both signs of cell death and signs of division, which suggests a continuous balance between the disappearance and the renewal of beta cells, even after decades of disease. The implication is profound: if even a few beta cells survive and renew themselves over tens of years, strategies to protect or to multiply them could be useful even in those with long-standing diabetes.

There are three strong reasons. Firstly, the pancreas is a dangerous organ to biopsy, being full of digestive enzymes, so an incision risks leakage of pancreatic juice, with inflammation of the pancreas and significant bleeding. Secondly, the patchy and lobular distribution of insulitis means that a small sample can easily hit an untouched area and give a false negative result, which makes a single biopsy inconclusive.

Thirdly, the biopsy is quite simply useless for diagnosis, because T1D is diagnosed clinically and confirmed through blood tests, such as the autoantibodies and the measurement of C-peptide. Current guidelines build the diagnosis and staging on the clinical signs, on blood glucose and HbA1c, on autoantibodies and never on examination of the pancreatic tissue [12]. Subjecting a patient to a risky surgical procedure to confirm something that a blood test has already answered would be unjustified, so the biopsy remains reserved for research.