Loss of SNORD115 mitigates SNORD116-driven sleep abnormalities in mouse models of Prader-Willi syndrome

Prader-Willi syndrome (PWS) is a neurodevelopmental disorder caused by the loss of paternally expressed genes within the imprinted 15q11-q13 locus, which includes clusters of box C/D small nucleolar RNAs (SNORDs), notably the SNORD115 and SNORD116 gene families. Although paternally inherited SNORD116 deletions have been associated with sleep disturbances in patients and mouse models, the respective and combined contributions of SNORD115 and SNORD116 to sleep regulation remain unclear. Here, we combined polysomnographic recordings with quantitative analysis of hypothalamic neuronal populations to assess sleep-wake architecture, sleep homeostasis, and underlying circuit alterations in mice carrying paternal deletions of SNORD115, SNORD116, or both clusters. SNORD116 deletion was associated with a selective increase in REM sleep, particularly during the light phase and during recovery following sleep deprivation, without affecting slow-wave sleep or REM-associated theta activity. In contrast, SNORD115 deletion did not alter sleep. Unexpectedly, combined deletion of SNORD115 and SNORD116 did not reproduce the REM sleep phenotype observed in SNORD116-deficient mice, indicating that SNORD115 loss attenuates SNORD116-dependent REM sleep alterations. At the cellular level, melanin-concentrating hormone (MCH) neuron density was reduced in both SNORD116-KO and double SNORD116/115-KO mice, whereas hypocretin (Hcrt) neurons were preserved across genotypes. Notably, REM sleep alterations did not parallel MCH loss, as increased REM sleep was absent in double-KO animals despite comparable reduction in MCH neuron density. Transcriptomic analyses at ZT0 further revealed only limited changes in hypothalamic gene expression across models. Together, these findings reveal an unanticipated interaction between SNORD115 and SNORD116 in the regulation of REM sleep and uncover a dissociation between genetic alterations, neuronal circuit organization, and sleep phenotype. More broadly, they caution against inferring physiological functions or disease-relevant mechanisms from single SNORD deletions within complex imprinted loci, and indicate that current mouse models may not faithfully capture sleep alterations associated with PWS.