In this paper Ding et al. describe the generation of a mouse mutant null for one of the snoRNA’s present in the Prader-Willi Syndrome (PWS) imprinted cluster. This snoRNA (snord116, formerly HBII-85) has been suggested to play an important role in the development of PWS. The authors have done extensive analysis of their mouse model, and discovered abnormalities in a number of endophenotypes of relevance to PWS. The most important of these is the evidence of hyperphagia and altered feeding behaviour.

The evidence for hyperphagia is seen to develop over time in the mouse model, and is apparent with relation to feeding on standard rodent chow from age 3 months onwards in males, and 6 months onwards in females (Figure 4C). The authors were first alerted to this difference when analysing feeding behaviour in the metabolic chamber studies (Figure 4A). As the mutant mice are growth retarded, the authors have necessarily normalised food intake to body-weight (BW). However, it is not entirely clear how this was done, and the indication from the Methods is that the intake of food was simply divided by BW. If this is the case, it is most certainly incorrect. It is well known that metabolic analyses such as food consumption, vary with BW but that this follows a characteristic scaling value, the Kleiber scale, which is somewhere between BW 0.6-power and BW 0.75-power. A recent meta-analysis (1) of intra-specific scaling values points to a figure of BW 0.72-power for mice.

Using the mean BW values (Figure 2A) and food intake data provided in this paper we have re-calculated the food-intake normalised to BW 0.75-power (the least conservative) for the 4 age points presented in Figure 4C and the metabolic chamber study in Figure 4A. Overall this obviously brings all the data points closer together. Furthermore we would suggest that with this re-calculation there is no difference in male consumption at age 3 months (WT v Del; 0.31 v 0.33) and that the only difference in the data from the metabolic chamber study is on the first 3 days of testing; on the last 3 days both WT and Del animals have values of 0.30. Of course this re-analysis is slightly crude, and ideally each data point needs to be re-calculated with the individuals BW scalar for proper statistical analysis. However, it does not detract from the fact that the difference in amount of food consumed between WT and Del mice is not as striking as first thought, and in the metabolic chambers may statistically interact with day of testing, suggesting some biological interaction with the novel environment. More generally, this may also point to a possible explanation as to why these animals do not become obese – their difference in food consumption is in fact not that great.

Nevertheless this does not detract from the other findings in the paper, in particular the exciting suggestion of altered eating bout latencies. Furthermore, it may be that the authors have in fact taken metabolic scaling into account. However, this is certainly not apparent from the text, and needs to be clarified.

Anthony R. Isles & Dinko Relkovic
Behavioural Genetics Group
Cardiff University
U.K.

1. D. S. Glazier, Biol Rev Camb Philos Soc 80, 611 (Nov, 2005).