Abstract The amyloid β peptide (Aβ) starting with pyroglutamate (pE) at position 3 and ending at position 42 (Aβ3pE-42) is a dominant species that accumulates in the Alzheimer’s disease (AD) brain. Consistently, a therapeutic antibody raised against this species, donanemab, has been shown to be effective in recent clinical trials. While the primary Aβ species produced physiologically is Aβ1-40/42, an explanation for how and why this physiological Aβ is converted to the pathological form has remained elusive. The conversion of Aβ1-42 to Aβ3pE-42 is likely to take place after deposition of Aβ1-42 given that Aβ3pE-42 plaques arise significantly later than Aβ1-42 deposition in the brains of single App knock-in and APP-transgenic mice. Here, we present experimental evidence that accounts for the aging-associated Aβ3pE-42 deposition: [1] Aβ3pE-42 is metabolically more stable than other AβX-42 species; [2] Deficiency of neprilysin (NEP), the major Aβ-degrading enzyme, induces a relatively selective deposition of Aβ3pE-42 in APP-Tg mice. [3] Aβ3pE-42 deposition always colocalizes with cored plaques in both APP-Tg and App knock-in mouse brains; [4] Aβ3E-42, an immediate precursor of Aβ3pE-42, as well as Aβ2A-42 and Aβ4F-42 are more short-lived than Aβ1-42 in vivo , indicating that simple N-terminal truncation that can arise enzymatically or spontaneously makes AβX-42 easier to catabolize. Consistently, newly generated knock-in mice, App NL-(ΔDA)-F and App NL-(ΔDA)-Q-F , showed no detectable Aβ pathology even after aging, indicating that the Aβ3E-42 and Aβ3Q-42 species are extremely labile to the in vivo catabolic system and that the E/Q cyclase activity present in mouse brain is insufficient for Aβ3pE-42 generation. In addition, a deficiency of NEP facilitated Aβ3pE-42 deposition. Of note, we identified a trace amount of Aβ3pE-42 and its immediate precursor, Aβ3E-42, in the insoluble fraction of NEP-deficient APP-Tg mouse brains. Aβ3pE-42 is thus likely to be a probabilistic by-product of Aβ1-42 metabolism that selectively accumulates over a long-time range of brain aging. It is likely produced in the solid state or at the solid-liquid interface. Our findings suggest that anti-Aβ therapies will probably be most effective if given before Aβ3pE-42 deposition takes place.