Proteins translated from tau circular RNAs as drivers for Alzheimer’s disease and frontotemporal dementia

Background Neurofibrillary tangles formed by the microtubule protein tau ( MAPT ) are the pathologic hallmark of tauopathies, including Alzheimer’s disease and ∼1/2 of frontotemporal lobar degeneration (FTLD‐Tau) cases. We reported that MAPT forms circular RNAs through back‐splicing of exon 12 to either exon 10 or exon 7 (BBA Dis 1864:2753; PMID: 29729314). The MAPT circRNAs require primate‐specific Alu elements to form, which might explain their absence in mice. Both MAPT circRNAs lack stop codons and could form multimers of microtubule binding repeats upon translation. Method In cell culture studies, we used transfected reporter genes that express tau circRNAs with a 3xFlag tag. We tested the effect of FTLD‐Tau mutants (K317M and V337M), junction‐specific siRNAs and adenosine deaminase acting on RNA (ADAR) enzymes on circRNA translation, evaluated by RNAse protection, Western blots and validated by mass spectrometry. Result Circular MAPT constructs were readily translated and regulated by endogenous factors including ADAR and customized siRNAs. The circ12–>7 RNA contained a single natural occurring start codon. The circRNA is formed and translated when exons 7 and 12 were flanked by authentic tau or heterologous Alu elements. The circ12–>10 RNA lacks a natural start codon, but introduction of the FTLD‐Tau mutations K317M and V337M induced translation of the circular RNA. The presence of ADAR activity strongly promoted translation of both circRNAs, even in the absence of start codons, possibly by changing an AUA to an AUI start codon (I‐inosine). Despite the lack of a stop codon, translation of the circular MAPT RNA products stopped at defined sites, indicating that circular RNAs use stop signals distinct from canonical ones. Conclusion Our data indicate that MAPT circular RNAs are translated into proteins that could cause tau aggregations relevant to tauopathies. As most FTLD‐tau causing mutations are in the regions near the circularizing part of the gene, they could act through regulating circular RNA formation and/or translation. ADAR‐dependent upregulation of circTau translations caused by inflammation and/or brain injury could trigger tau pathology via this mechanism. Finally, tau circRNA can be selectively removed using siRNAs. We are assessing the impact of MAPT circRNAs on proteinaceous aggregations.