Spreading of the autoimmune response in type 1 diabetes

The diversification of the autoimmune response is known in the international medical literature as "epitope spreading". Much like the dietary diversification of an infant, it is a process by which the immune system starts by recognising a single fragment of the beta cell and then gradually extends its attack. First the attack also covers other portions of the same protein initially recognised as the enemy (intramolecular spreading), then it covers new, completely different proteins (intermolecular spreading). This explains why, after months or years, several autoantibodies can be detected in the blood, added gradually [1].

This process is not a simple coincidence, but a direct consequence of the continuous destruction of the beta cells. When cells die, they release intracellular proteins that normally do not come into contact with the immune system. These "hidden" proteins are very irritating to the immune system and become new targets, amplifying the autoimmune response. The presence of several autoantibodies indicates a more mature and more advanced autoimmune process, not just a coincidence of positive tests [2].

The number of autoantibodies increases over time because the destruction of the beta cells is a slow and progressive process. At each stage, the immune system encounters new antigens released from the affected cells. In addition, cellular stress produces chemical modifications of proteins, which can end up looking very different from their initial state, irritating the immune system (neo-antigens). The immune system then treats them as foreign targets and begins to produce more and more autoantibodies against them [2].

This mechanism explains why a first test may identify a single autoantibody, while repeated tests at 6, 12 or 24 months may identify one or two more. Once started, the process is self-sustaining: the more the beta cells are affected, the more antigens (areas that irritate the immune system) become visible, and the autoimmune repertoire expands. This is why international guidelines recommend periodic reassessment of the autoantibodies [3].

International cohort studies, such as TEDDY, BABYDIAB, DAISY and DIPP, have established a certain order, but one that is only encountered somewhat more frequently. Seroconversion, the first appearance of an autoantibody, is rare under the age of 6 months. In children at high genetic risk, the first antibody usually appears as early as one year of age. In the young child, the first antibody to appear is usually the anti-insulin one (IAA), while in adolescents and young adults the most frequently encountered is GADA (autoantibodies against glutamic acid decarboxylase, GAD65) [4].

Anti-tyrosine phosphatase autoantibodies (IA2A) and anti-zinc transporter 8 autoantibodies (ZnT8A) usually appear after another autoantibody already exists (they rarely appear alone). They tend to be added at later stages, signalling an advanced autoimmune process and an approaching clinical onset, especially in the case of IA2A. The average interval from the first to the second autoantibody is generally short, a few months, after which diversification slows down (the third appears only with difficulty) [4].

In the young child, anti-insulin antibodies (IAA) often appear first because of a special combination of immunological and genetic factors. Children who inherit the HLA-DR4-DQ8 haplotype have an increased risk of developing IAA as the first autoantibody, because fragments derived from proinsulin are presented more easily to autoreactive T lymphocytes. In practice, the HLA "platform" shows the immune system fragments of (pro)insulin, and the immune response subsequently focuses precisely on insulin [5].

In addition to genetics, the immune system of the young child is still developing and is learning to distinguish between "self" and "non-self". Insulin is actively produced by the beta cells, and exposing the immune system to fragments of this protein in a vulnerable immunological context can trigger a specific autoimmune response. This is why a very early appearance of autoantibodies (under 3 years) is correlated with a higher risk and a more rapid progression towards clinical type 1 diabetes (stage 3) [5].

In adolescents and adults, GADA more often appears first because it is influenced by a different genetic and immunological context. The HLA-DR3-DQ2 haplotype is preferentially associated with an autoimmune profile dominated by anti-GAD65 autoantibodies (GADA). GAD65 is the 65 kDa (kilodaltons, a unit that shows the "weight" of the molecule) isoform of glutamate decarboxylase, an enzyme expressed both in pancreatic beta cells and in some neurons in the brain (the GABAergic ones). This genetic variant stimulates the preferential presentation of fragments derived from GAD65 to T lymphocytes [6].

GADA is the least specific of the islet autoantibodies, being associated also with other autoimmune diseases, such as autoimmune thyroiditis or coeliac disease. The speed of progression of the anti-beta-cell autoimmune response in adults is usually slower and less aggressive than in the young child. GADA can persist as the sole autoantibody for years, and progression towards clinical type 1 diabetes (stage 3) can stretch over decades. This slow course explains latent autoimmune diabetes of adults (LADA), where GADA is often the only positive marker [6].

The number of autoantibodies that are positive at the same time has a very clear prognostic meaning. A single positive autoantibody represents a heterogeneous group, which includes both people whose autoimmunity will progress and people with transient autoimmunity, who will never develop type 1 diabetes. The presence of two or more confirmed autoantibodies defines preclinical stages 1 (with normal blood glucose) and 2 (with prediabetes) of type 1 diabetes. Three positive autoantibodies at the same time mean a mature autoimmune process, with more rapid progression towards stage 3 [7].

An important nuance is that the isolated presence of IA-2A (anti-tyrosine phosphatase antibodies) is currently treated as a risk equivalent to the presence of 2 or more autoantibodies. This reflects the fact that IA-2A usually appears late in the autoimmune cascade and signals advanced destruction of the beta cells. In general, progression from stage 1 to stage 3 is estimated at approximately 30% at 5 years and 85% at 15 years [3].

The presence of two or more confirmed autoantibodies predicts type 1 diabetes with remarkable accuracy (approximately 85% risk over 15 years) because it marks the transition from a potentially reversible autoimmunity (a single autoantibody) to a consolidated autoimmune disease. The risk of progression to stage 3 type 1 diabetes in the presence of two autoantibodies is close to 100% over the whole course of life. This risk curve makes the number of autoantibodies the most powerful predictor of type 1 diabetes available in the clinic [8].

The jump from 1 to 2 or more autoantibodies is the most important prognostic transition in preclinical type 1 diabetes. It reflects the diversification of the autoimmune response and a subsequent progressive loss of beta-cell mass. Precisely for this reason, people with two or more autoantibodies must be monitored closely. As access becomes ever easier, these people will be able to benefit from therapies that can delay progression, such as teplizumab. Identifying them early also reduces the risk of diabetic ketoacidosis at the moment of the clinical diagnosis of type 1 diabetes (stage 3) [8].

Yes, it is perfectly possible to remain for years with a single positive autoantibody, without progressing towards clinical type 1 diabetes. This scenario is more frequent in older children, adolescents and especially in adults. Factors that predict a lower risk include a more advanced age at seroconversion, a low titre and a neutral HLA background [9].

Precisely because this situation is heterogeneous, periodic reassessment is recommended, at intervals from 6 months to 3 years, depending on age and on the type of autoantibody. The important exception is IA-2A which, even alone, is now considered equivalent to the risk brought by two autoantibodies. Close monitoring of those with stable autoimmunity provides time for education and for the prevention of severe complications at onset [3].

Yes, the autoantibodies can disappear after they have appeared. This phenomenon is called seroreversion and is encountered especially in people with a single autoantibody, with a low titre and a more advanced age at seroconversion. Most seroreversions occur in the first two years after the first appearance of an autoantibody (seroconversion). By contrast, the disappearance of autoantibodies in people with two or more autoantibodies is very rare. Once preclinical stage 1 is reached, the autoimmune process is generally irreversible [10].

Seroreversion reduces, but does not completely remove, the risk of type 1 diabetes. People who have had positive autoantibodies remain with a somewhat higher risk than those who have never had autoantibodies. A negative test after a previously positive one is obviously good news, but not a guarantee. Continued monitoring remains necessary, and a confirmed autoimmunity must be regarded as a reason for long-term surveillance [10].