Evolution from one autoimmune disorder to another. Epitope spreading?

A patient in whom organs or molecules are affected sequentially by an autoimmune response prompted us to propose a hypothesis that epitope spreading may account for the sequential involvement of different cells. A 27-year-old male patient was diagnosed in January 2005 with idiopathic thrombocytopenic purpura (ITP), as all autoantibodies relevant to autoimmune diseases were negative. He was treated with immunosuppression and splenectomy. In 2014, the patient developed transient blurred vision, dizziness and walking instability and autoimmune chylomicronemia was diagnosed, based on the presence of autoantibodies to lipoprotein lipase (LPL).1 He was treated with intravenous methylprednisolone (1 mg/kg body weight), plasmapheresis, as well as ezetimibe-statin combination. In July 2016, the patient presented with anaemia (haemoglobin 11.2 g/dL) and thrombocytopenia (83 000/μL), as well as laboratory features of haemolysis: increased reticulocyte count (6.5%, normal value n.v. <2.5%), serum lactate dehydrogenase (620 U/L, n.v. 140-280 U/L), unconjugated bilirubin (2.1 mg/dL), strongly positive direct Coombs test against IgG, IgA, IgM and C3, and very low serum haptoglobins (0 mg/dL, n.v. 36-195 mg/dL). Evan's syndrome was diagnosed, and rituximab 375 mg/m2 body surface/wk was administered for 4 weeks. Thereafter, danazol 200 mg/d/po was added. A year later, he was on complete remission. Using the BLAST and SMART tools,2, 3 we searched for platelet/erythrocyte proteins with sequence or structural similarity with LPL. LPL contains a lipoxygenase homology LH2 domain, also found in arachidonate 12- and 15-lipoxygenases (12-LOX, 15-LOX). Both have been detected in human platelets and the latter has also been found in reticulocytes and rabbit erythrocytes. Autoimmune reaction towards platelet 12- or 15-LOX could be followed by secondary generation of autoantibodies to LPL and expansion of the reaction to the 15-LOX on erythrocytes, causing the observed in our patient thrombocytopenia, hyperchylomicronemia and haemolysis. An alternative explanation is that the observed sequential autoimmune reaction towards different targets could be attributed to the similarity between CD59, a surface molecule present on platelets and erythrocytes, and the Glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1 (GPIHBP1), an endothelial protein binding to LPL and required for lipolysis in the capillary lumen.4 GPIHBP1 contains a LU domain, a 3-fingered structure also found in CD59. Thus, the sequential autoimmune reactions in the presented case could start against CD59 and expand to GPIHBP1 and LPL. In a third hypothetical scenario, the sequence of autoimmune responses could start with autoantibodies against platelet 12- or 15-LOX, followed by reaction to LPL and its associated GPIHBP1 and then to the structurally related CD59. Interestingly, sequence similarity with the above-mentioned proteins was found in the CD59 protein homolog of Herpesvirus strains. Thus, we can speculate that a Herpes clinical or subclinical infection could initiate in a genetically prone individual a cascade of immune responses towards different proteins sharing sequence similarities with the invading agent. Although the discussed models are purely hypothetical and remain to be tested experimentally, this case represents a paradigm of sequential involvement of different organs/molecules by an autoimmune response possibly due to epitope similarities with an exogenous agent.