Can infections generate cross-reactive antibodies without proving autoimmune disease?

Infections can sometimes elicit antibodies that recognize both microbial antigens and structurally similar host molecules—a biologically credible and directly demonstrated form of molecular mimicry. However, detecting such an antibody, whether after infection or on an autoantibody panel, is not equivalent to demonstrating autoimmune disease or antibody-mediated injury.

Clinical and mechanistic evidence

• In Campylobacter jejuni–associated Guillain–Barré syndrome, patient sera recognized bacterial lipooligosaccharides (LOS) and peripheral-nerve gangliosides GM1/GD1a. Relevant isolates contained GM1- or GD1a-like structures, and infecting-strain LOS inhibited 90–100% of anti-ganglioside reactivity in two patients—strong evidence that shared structural epitopes can drive cross-recognition.
• In rheumatic carditis, human monoclonal antibodies reacted with both Streptococcus pyogenes and human cardiac myosin; some also bound cardiac tissue and valve endothelium. Streptococcal M-protein-related determinants and group-A carbohydrate GlcNAc have been implicated.
• These examples support mimicry through structurally similar epitopes, not necessarily identical protein sequences. Clinical injury nevertheless depends on factors beyond binding, including microbial strain, host susceptibility, inflammatory context, target accessibility, and other immune mechanisms.

Interpreting antibody positivity

• Autoantibody results require a compatible syndrome, consideration of alternative diagnoses, and attention to the specific antibody, assay, specimen, and pretest probability.
• In a 2025 clinical cohort, only 38.2% of positive neural-autoantibody tests were adjudicated true positives, while 61.8% were false positives; 95.2% of false-positive results were serum-only.
• Pathogenic inference is stronger with phenotype concordance, affected-tissue or intrathecal relevance, functional effects, and confirmatory testing. Antibodies to accessible extracellular targets are generally more plausibly pathogenic than intracellular-target antibodies, which may primarily be biomarkers.

Bottom line

• Molecular mimicry can generate cross-reactive self-binding antibodies, but a positive antibody test alone does not prove pathogenic autoimmune disease; clinical and mechanistic corroboration are essential.