PREreview of Age-associated sleep-wake patterns are altered with Prdm13 signaling in the dorsomedial hypothalamus and dietary restriction in mice

This review has been formulated based on my brief comment posted on Pubpeer.

https://pubpeer.com/publications/030D11C6AA3F20092C564B6599BFE1#

In this preprint, Tsuji et al. deal with an interesting topic of aging and sleep and also present a great deal of data. However, in the current version, there seem clear discrepancies between the data presented and the authors' claims. I also found it very difficult to draw a logical conclusion from the data presented in this study due to inconsistent results and issues with experimental design. 

Major concerns:

1.   No clear evidence that DMH-Prdm13-KO mice recapitulate the age-associated alterations of sleep observed in the wild-type mice

Tsuji et al. generated a dorsomedial hypothalamus (DMH)-specific Prdm13 KO mice and measured various sleep indices to test the hypothesis that the Prdm13 function in the DMH may be causally involved in age-associated sleep alterations. However, although the authors claim that they found a commonality in sleep phenotypes between DMH-Prdm13-KO mice and aged C57BL/6J mice, the presented data do not support the claim: Specifically, the data in Fig. 3 show that sleep measures of DMH-Prdm13-KO mice were mostly similar to those of controls. Looking more closely, DMH-Prdm13-KO mice showed significantly decreased episode duration of N-REM sleep in the dark phase and significantly increased episode duration of REM sleep in the dark phase (Fig. 3c, d, h, and g), which are completely different from the age-associated changes of sleep measures observed in C57BL/6J mice (Fig. 1a, b, c, and d).

The same problem can be seen in the comparison between the sleep phenotypes of aged C57BL/6J mice and aged DMH-Prdm13-KO mice: Although aged DMH-Prdm13-KO mice exhibited changes in some of the sleep indices compared to their age-matched controls such as significantly increased wakefulness bouts and N-REM sleep bouts in the light phase, and significantly decreased wakefulness episode duration in the light phase and N-REM sleep episode duration in the dark phase (Fig. 4a, b, and c, Supplementary Fig. 4e). However, again, these changes have no commonality to the age-associated changes in sleep measures observed in C57BL/6J mice (Fig. 1a, b, c, and d).

Thus, because both young and aged DMH-Prdm13-KO mice did not share changes in sleep parameters with aged C57BL/6J mice, the authors' key assertion that the DMH-Prdm13-KO mice recapitulated age-associated sleep alterations is not supported by actual data.

2.    Problematic experimental design of sleep deprivation (SD) assay

The only consistent phenotype among old C57BL/6J mice, young DMH-Prdm13-KO mice, and old DMH-Prdm13-KO mice seems the increased number of sleep attempts shown in Fig. 1g, 3e, and 4d. However, given the design of the authors' SD experiments, it is highly doubtful that this increase in sleep attempts specifically reflects an increase in sleepiness as the authors interpret.

According to the description in the methods section, the authors first removed food from cages and then kept mice awake for 6 hours by stimulating them with long Q-tips, simultaneously counting sleep attempts relying only on behavioral observation. However, in this experimental design, changes in the number of sleep attempts could be caused by various sensory and emotional changes, such as habituation or hypersensitivity to Q-tip stimulation and the presence of the experimenter, and changes in stress and anxiety levels. In addition, as shown by Đukanović et al, mice deprived of sleep are also deprived of food, even when they have access to food (Dukanovic et al., 2022). In the present study, the authors completely deprived the mice of food during the 6 hours of SD manipulation, which must have caused a greater effect of food deprivation on the mice.

Therefore, it is difficult to interpret the cFos expression in Prdm13+ cells shown in Fig. 2 or Fig. 7 as evidence that Prdm13+ cells were activated by sleep deprivation. As mentioned above, there are various alternative possibilities; Prdm13+ cells might have been activated by Q-tip stimulation itself, stress, changes in emotional states due to repeated tactile stimulation, changes in internal states due to lack of food, a complex combination of these factors. Similarly, it is not possible to conclude from the presented data that changes in the number of sleep attempts directly reflect sleep-related changes. Indeed, despite the significant increase in the number of sleep attempts, there was no alteration in EEG SWA after SD in young DMH-Prdm13-KO animals (Fig. 3e, f).

In addition to the above problems, the authors' sleep attempt counts rely solely on subjective behavioral observations and lack the use of EEG/EMG measurements that would have allowed more objective determinations of sleep onset. Moreover, there is no description of the strength and location of the Q-tip stimuli, as well as the time between the presentation of the Q-tip and the judgment of whether the mouse reacted or not. Although the authors wrote that they performed their SD procedure as previously described by Franken et al., the cited study actually performed a very different SD procedure, in which rats were kept awakened by presenting novel objects, acoustic stimuli, and/or gentle tactile stimuli with having undisturbed access to food, and EEG/EMG measurements were also used to determine sleep onsets (Franken et al., 1991). Such an inappropriate citation and inadequate description of the method raise a concern that the changes in the number of sleep attempts that the authors report in this study may have been partly due to artifacts caused by the authors’ arbitrary and variable experimental procedure.

More minor points:

-The authors show that many sleep measures changed after dietary restriction (DR) (Fig. 5), but these could easily be indirect or non-specific effects on sleep measures and not reflect changes in sleep per se. Indeed, the authors discuss that the further reduced EEG power during N-REM and REM sleep in old-DR mice (Fig. 5d) might be due to lower body temperature and not necessarily reflect changes in sleep pressure. The authors should use the same level of caution on the possibility of such indirect or non-specific effects throughout the manuscript.

-The authors extracted more than 20 sleep indices from the same EEG traces and performed statistical comparisons of each index individually (Figures 1, 3-5, Supplementary figures 1, 3-5). In such cases, the multiplicity of hypothesis testing becomes an issue. For example, when 20 hypotheses are simultaneously tested at an alpha level of 0.05, the probability that at least one non-significant trait is determined to be significant by chance is more than 64% (actual alpha level > 0.64). The p-values should be appropriately adjusted for the multiple hypothesis testing.

-The authors previously reported that DMH-specific Prdm13 knockdown causes low delta power in N-REM sleep (Satoh et al., 2015), but do not discuss whether this was reproduced by the DMH-specific Prdm13 KO in the current study. This is a very important point and should be discussed explicitly.

-The authors counted and compared the number of cFos+, CCK+, GRP+ cells, in Fig. 2a, and Supplementary Fig. 7b, d. However, the number of positive cells is greatly influenced by the difference in the number of total cells examined. The percentage of positive cells in the total number of examined cells should be reported to ensure fair comparisons.

-The authors suggest that Prdm13+ cells are "electrically active cells such as a neuron" using a whole-cell patch technique (Supplementary Fig. 2d-f). However, immunostaining with neuronal markers such as NeuN (and glial markers) should be more informative to estimate the identities of the Prdm13+ cell population.

References:

Dukanovic, N., La Spada, F., Emmenegger, Y., Niederhauser, G., Preitner, F., and Franken, P. (2022). Depriving Mice of Sleep also Deprives of Food. Clocks & sleep 4, 37-51.

Franken, P., Dijk, D.J., Tobler, I., and Borbely, A.A. (1991). Sleep deprivation in rats: effects on EEG power spectra, vigilance states, and cortical temperature. Am J Physiol 261, R198-208.

Satoh, A., Brace, C.S., Rensing, N., and Imai, S. (2015). Deficiency of Prdm13, a dorsomedial hypothalamus-enriched gene, mimics age-associated changes in sleep quality and adiposity. Aging Cell 14, 209-218.