Paradoxical leanness in the imprinting-centre deletion mouse model for Prader–Willi syndrome

Prader–Willi syndrome (PWS), a neurodevelopmental disorder caused by loss of paternal gene expression from 15q11–q13, is characterised by growth retardation, hyperphagia and obesity. However, as single gene mutation mouse models for this condition display an incomplete spectrum of the PWS phenotype, we have characterised the metabolic impairment in a mouse model for ‘full’ PWS, in which deletion of the imprinting centre (IC) abolishes paternal gene expression from the entire PWS cluster. We show that PWS-ICdel mice displayed postnatal growth retardation, with reduced body weight, hyperghrelinaemia and marked abdominal leanness; proportionate retroperitoneal, epididymal/omental and inguinal white adipose tissue (WAT) weights being reduced by 82%, 84% and 67%, respectively. PWS-ICdel mice also displayed a 48% reduction in proportionate interscapular brown adipose tissue (isBAT) weight with significant ‘beiging’ of abdominal WAT, and a 2°C increase in interscapular surface body temperature. Maintenance of PWS-ICdel mice under thermoneutral conditions (30°C) suppressed the thermogenic activity in PWS-ICdel males, but failed to elevate the abdominal WAT weight, possibly due to a normalisation of caloric intake. Interestingly, PWS-ICdel mice also showed exaggerated food hoarding behaviour with standard and high-fat diets, but despite becoming hyperphagic when switched to a high-fat diet, PWS-ICdel mice failed to gain weight. This evidence indicates that, unlike humans with PWS, loss of paternal gene expression from the PWS cluster in mice results in abdominal leanness. Although reduced subcutaneous insulation may lead to exaggerated heat loss and thermogenesis, abdominal leanness is likely to arise from a reduced lipid storage capacity rather than increased energy utilisation in BAT.


Introduction
Prader-Willi syndrome (PWS) is caused by a lack of paternal gene expression from the 15q11-q13 imprinting cluster and results from large chromosomal deletions, chromosome 15 maternal uniparental disomy or imprinting-centre (IC) mutations. This disorder is associated with significant metabolic impairment.
After severe neonatal hypotonia and a failure to thrive in infancy, subsequent development of hyperphagia and reduced satiety responses result in obesity, unless managed carefully. Indeed, as PWS is the most common syndromal cause of morbid obesity (Butler 1990), elucidating the aetiology of this metabolic disturbance will not only benefit individuals with PWS but also may have wider implications for ameliorating obesity.
Insight into the relationship between the imprinted genes within the PWS cluster and impaired function has been gained by comparing the phenotype of mouse models bearing different PWS-associated mutations. For example, neonatal failure to thrive and early life metabolic impairment are seen in many models (Yang et al. 1998, Ge et al. 2002, Stefan et al. 2005, including a number of single gene knock-out mice (Ding et al. 2008, Schaller et al. 2010. In contrast, evidence for feeding and metabolic changes in adult mice are inconsistent. For instance, despite being hypophagic (Kozlov et al. 2007), mice null for Magel2 become obese (Bischof et al. 2007), whereas mice carrying a deletion of the small nucleolar (sno)RNA Snord116 show mild hyperphagia, and impaired meal termination, coupled with leanness, even on a highfat diet (Ding et al. 2008).
Given that single gene mutations constitute a rare subset of PWS individuals (Sahoo et al. 2008) and display an incomplete phenotypic spectrum (Peters 2008), we have characterised the metabolic impairment in a mouse model for 'full' PWS, the PWS-IC del mouse. IC deletion in this model results in the complete loss of paternal gene expression from the PWS interval (Chamberlain et al. 2004), including its expression in the central nervous system (Relkovic et al. 2010). Although PWS-IC del mice suffer considerable postnatal lethality, sufficient survivors have enabled us to demonstrate many of the behavioural aspects of PWS in this model, including reduced locomotor activity and deficits in cognitive function, such as attention and impulse control (Doe et al. 2009, Relkovic et al. 2010. However, although we have shown that hyperphagia in PWS-IC del mice is driven primarily by calorie seeking rather than hedonic processes (Davies et al. 2015), there has been no systematic characterisation of metabolism in this or any other mouse model for 'full' PWS (Bervini & Herzog 2013).
We have therefore characterised the metabolic phenotype of the PWS-IC del model in this study, quantifying appetitive and consummatory feeding behaviour, lipid profiling and storage, short-term responses to high-fat diet exposure and assessment of thermogenic function.

Animals
The PWS-IC +/− (referred throughout as PWS-IC del ) and wild-type (WT) mice used in this study were bred under the Authority of the Animals (scientific procedures) Act 1986 (UK), with subsequent procedures conforming with the institutional and national guidelines, including those for genetically modified animals, and specifically approved by local ethical review.
PWS-IC del mice and WT littermates were generated by crossing IC del -positive males with WT females. Given the nature of the epigenetic regulation of imprinted genes, paternally inherited IC deletion results in a lack of gene expression from the PWS interval. As PWS-IC del animals on a pure C57BL/6J background suffer severe postnatal lethality, we crossed IC del -positive males with CD1 females and selectively culled WT littermates (identified on the basis of their increased size 48 h after birth) leaving only 1 or 2 WT pups per litter, as previously described (Chamberlain et al. 2004, Davies et al. 2015. Animals were weaned at approximately 4 weeks of age and were single-sexed group housed with WT littermates (2-5 animals per cage). This procedure generated the six cohorts of adult mice used below.
All animals were maintained on a 12-h light/ darkness cycle (lights on 07:00 h), with ad libitum access to water and standard laboratory chow (Rat and Mouse No. 3 Breeding Diet, Special Diet Services Ltd., Witham, Essex, UK, containing 4.2% crude fat; 22.4% crude protein; 4.2% crude fibre; 7.6% crude ash), unless stated otherwise.

Study 1. Adiposity profile in PWS-IC del mice
After an overnight fast (with water available ad libitum), 18-month-old male WT and PWS-IC del littermate mice were weighed and killed by cervical dislocation. Inguinal, retroperitoneal and epididymal white adipose tissue (WAT) and interscapular brown adipose tissue (isBAT) fat depots and liver were excised, weighed, snap frozen and stored at −70°C for subsequent histological analysis (see below). Left tibiae were excised, and the length was determined with a hand-held micrometre.

Study 2. Plasma lipid profiles in PWS-IC del mice
Two groups of 16-to 17-month-old female WT and PWS-IC littermate mice were weighed and killed by cervical

Study 4. The effect of suppressing thermogenesis in PWS-IC del mice
To determine the impact of thermogenesis on intraabdominal adiposity, 6-to 15-month-old male and female PWS-IC del mice, and WT littermates, were group housed (2-3 mice/cage) at either standard room temperature (20-22°C) or at a thermoneutral ambient temperature (30°C) for 9 weeks. 30°C represents the lower critical temperature in mice, below which nonshivering thermogenesis is induced (Overton 2010). Mice maintained at thermoneutrality were housed in standard cages inside a warming chamber (Scanbur Type 48-VS-IV, Karslunde, Denmark), beginning at room temperature and increasing the temperature by no more than 2°C per day until a stable chamber temperature of 30°C was attained on day 8. Body weight was monitored 3 times/week. In weeks 7-8, mice were individually housed in metabolic cages under standard or thermoneutral ambient temperatures. After 3 days acclimatisation, ad libitum food intake was measured on two consecutive days and mean daily food intake calculated. The mice were then returned to their original group cages for the remainder of the 9 weeks, after which mice were weighed and killed by decapitation under isoflurane anaesthesia. Interscapular BAT, inguinal, epididymal/omental and retroperitoneal WAT and liver were excised, weighed, snap frozen and stored at −80°C for subsequent analysis.

Study 5. Feeding behaviour in PWS-IC del mice
Study 5a. Ad libitum feeding Groups of 5-to 9-month-old male and female WT and PWS-IC del mice were housed individually in metabolic cages. After 5 days acclimatisation, mice were subjected to ad libitum feeding with a standard diet (SDS 10% Atwater fuel energy (AFE) Fat, in which energy sources (3.68 kcal/g) were Crude Fat 10%; Crude Protein 20%; Carbohydrate 70% (Special Diet Services, Witham, Essex, UK)) for 1 week, followed by a further 1-week period with a high-fat diet (SDS 45% AFE Fat, in which energy sources (4.54 kcal/g) were: Crude Fat 45% Crude Protein 20%; Carbohydrate 35%). Body weight and food intake were monitored daily, food consumption being normalised to body weight. After the first day of exposure to each diet, the proportion of 'hoarded' food was determined by expressing the mass of food in the inner ('spill') compartment of the food hopper (Supplementary Fig. 1A and B, see section on supplementary data given at the end of this article) as a percentage of the total remaining diet. The proportion of diet consumed during the darkness and light phases was determined on the fifth day of exposure to each diet. At the end of the dietary manipulations, mice were killed by decapitation under isoflurane anaesthesia and trunk blood was collected. Separated plasma was stored at −20°C prior to subsequent analysis of plasma ghrelin (total) concentration (see below). Study 5b. Post-fast feeding Group housed 4-to 6-month-old male and female WT and PWS-IC del mice were subjected to a 16-h overnight fast (17:00-09:00) (with water provided ad libitum). Mice were placed individually in normal shoe-box cages under low lighting levels and presented with a pre-weighed pot of wet mash (1 part standard diet (SDS RM3):1 part water). Wet mash was used to overcome the propensity of PWS-IC del mice to hoard diet pellets and spill powdered diet ( Supplementary Fig. 1). Mice were allowed to feed freely for 60 min, the pots being weighed after 30 and 60 min to determine food consumption.

Tissue analysis
Adipocyte size and hepatic lipid content were quantified as previously described (Davies et al. 2009). Samples of retroperitoneal WAT were paraffin-embedded and 5 µm sections were stained with Masson's Trichrome. Cryostat sections (15 µm) of liver were stained with oil red-O and counterstained with Meyer's haematoxylin. Digital images (one image per section, three images per mouse) 232:1 were obtained under light microscopy. Adipocyte size was quantified in cells with transversely sectioned peripheral nuclei (minimum 9 cells per section/3 sections per sample) using ImageJ and hepatic lipid droplets measured using Leica Q-Win. The degree of 'beiging' was determined in three random full screen digital images of retroperitoneal WAT. Areas of 'beiging' were identified (see Fig. 2a and b), delineated, filled and quantified using ImageJ, the data being expressed as %-'beiging' of total field.
Uncoupling protein-1 (Ucp-1) mRNA expression in isBAT and inguinal WAT was quantified as previously described (Wells et al. 2012). In brief, total RNA from lysed tissue was purified using the Qiagen RNeasy Mini extraction kit before quantification and assessment of RNA purity. RNA was subsequently reverse transcribed using the Precision nanoScript Reverse Transcription kit (Primerdesign, Southampton, UK), 1 µg RNA being reverse transcribed in 20 µL reactions containing 1 µL oligoDT and 1 µL random decamers (Primerdesign) prior to freeze storing at −20°C. cDNA was diluted 1:10 in dd.H 2 O prior to quantification of Ucp-1 and β-actin cDNA by quantitative real-time PCR amplification using a Bio-Rad IQ5 thermal cycler and Precision 2× qPCR Mastermix reagents (Primerdesign). Oligonucleotide primers for mouse Ucp-1 were purchased from Primerdesign (Ucp-1 sense primer, GGTAAAAACAAGATTCATCAACTCTC: Ucp-1 anti-sense primer, TCGCAGAAAAGAAGCCACAA). PCR thermal cycling conditions included an initial enzyme activation for 10 min at 95°C (1 cycle), followed by 51 cycles of denaturation at 95°C for 15 s and real-time Fluorogenic Data Collection (FDC) at 60°C for 1 min. The final FDCs were acquired at 0.5°C temperature increments between 55°C and 95°C at 10-s intervals (81 cycles). The PCR products were quantified by incorporation of SYBR green into double stranded DNA. Single amplicon identity was verified by melt curve analysis. Individual samples were normalised to the expression of β-actin (ACTB Primer mix, Primer Design # ge-SY-6; Batch # 12474). The quantity of cDNA in each reaction was calculated by reference to a standard curve constructed from serial dilutions of cDNA from WT BAT (Wells et al. 2012).

Plasma analysis
Plasma ghrelin (total) concentrations were determined by RIA (Millipore/Linco (IAV: 3.28-8.05%)). Plasma lipids were extracted by the Folch method (Kates 1986). Classes of individual non-polar lipids (triacyglycerols, free fatty acids and sterol esters), and total polar lipids were separated by thin-layer chromatography and quantified by their fatty acid contents. Fatty acid methyl esters were separated on a Clarus 500 gas chromatograph (PerkinElmer) identified by comparing retention times with fatty acid standards (Nu-Chek Prep Inc., Elysian, MN, USA) and quantified against an internal pentadecanoate standard.

Statistical analysis
Results are expressed as mean ± s.e.m., and differences were established by 1-way and 2-way repeated measures ANOVA and Bonferroni post hoc test or unpaired Student's t-test (using GraphPad Prism, GraphPad Software), as indicated in the figure and table legends, with P < 0.05 considered significantly different.

Study 2. Plasma lipid profiles in PWS-IC del mice
We therefore investigated whether the reduction in WAT mass was due to impaired lipid handling. Although circulating total polar lipid in fed PWS-IC del mice was only 79% of that in WT littermates (P > 0.05; Table 1), none of the markers of dietary lipid intake (S1 and S2 sterol esters), hepatic lipid output (triacylglycerol: TAG) or WAT lipid output (free fatty acids: FFAs) were significantly different.
Similarly, the effect of fasting on circulating lipids was comparable in WT and PWS-IC del mice (Table 1). As shown previously (Davies et al. 2015), circulating ghrelin (total) in PWS-IC del females was 134% of that in female WT littermates (Table 1), but was regulated appropriately after the 24-h fast.

232:1 Study 3. Thermogenesis in PWS-IC del mice
We subsequently investigated the thermogenic activity in PWS-IC del mice. Proportionate isBAT weight was reduced by 48% in PWS-IC del mice ( Fig. 2A; P < 0.01), consistent with elevated thermogenesis. Similarly, retroperitoneal WAT ( Fig. 2 pictures a and b) showed a four-fold elevation in the degree of differentiation to 'beige' or 'brite' adipose tissue ( Fig. 2B; P < 0.05). Thermal imaging revealed that surface temperatures of the head and interscapular regions were increased in PWS-IC del mice by 1.4°C and 2.0°C, respectively ( Fig. 2C and pictures c and d; P < 0.05), whereas dorsolumbar tail root and radiant temperatures were not significantly affected (Fig. 2C).

Study 4. The effect of suppressing thermogenesis in PWS-IC del mice
To establish whether the increased energy demands of elevated thermogenesis account for the lean phenotype, PWS-IC del mice and WT littermates, were maintained under thermoneutral conditions (30°C). After 9 weeks, proportionate isBAT weight in WT and PWS-IC del males was elevated by 72% and 76%, respectively (Fig. 3A), indicating reduced lipid utilisation; these changes being less pronounced in WT and PWS-IC del females. Thermoneutrality halved Ucp-1 mRNA expression, the classic marker of thermogenic activity, in WT males (Fig. 3B), but this effect was not significant in PWS-IC del animals (P = 0.053 (males); P = 0.183 (females)). Despite suppressed thermogenesis, proportionate WAT mass (total: data not shown, or individual depots: Fig. 3C, E and F) was not increased. For example, although proportionate retroperitoneal WAT mass in WT males at thermoneutrality was 156% of that in corresponding mice at room temperature, these means were not significantly different ( Fig. 3E; P = 0.09), whereas for male PWS-IC del mice at thermoneutrality, retroperitoneal WAT mass was 97% of that in PWS-IC del mice at room temperature. In addition, thermoneutrality had no significant effect on WAT Ucp-1 mRNA expression (Fig. 3D). However, thermoneutrality reduced liver weight in WT males by 12% (data not shown) suggesting either reduced hepatic liver storage or increased lipid export. Interestingly,

232:1
suppression of thermogenesis failed to increase body weight gain in any of the groups (Fig. 3G), although WT females at thermoneutrality showed a transient increase in weight gain between days 11 and 15 (P < 0.05; data not shown). This lack of weight gain may arise from the suppressive influence of thermoneutrality on daily food intake (Fig. 3H), mean proportionate daily food intake in WT males, PWS-IC del males, WT females and PWS-IC del females at thermoneutrality being 73%, 50%, 74% and 62% of that in mice housed at room temperature. Proportionate hyperphagia in PWS-IC del mice was completely abolished at thermoneutrality (Fig. 3H).

Study 5. Feeding behaviour in PWS-IC del mice
Study 5a. Ad libitum feeding We subsequently examined feeding behaviour in PWS-IC del mice at normal ambient temperature. Food intake was measured in male and female WT and PWC-IC del mice, maintained for 1 week on standard and high-fat diets. Male and female WT mice ate roughly similar amounts of both diets (Fig. 4C). Although PWS-IC del females consumed 32% less of the high-fat diet (compared to WT females; P < 0.05; Fig. 4C), when corrected for body weight, PWS-IC del males consumed 76% more of the high-fat diet (compared to WT males; P < 0.01; Fig. 4D). Interestingly, although WT males maintained a constant proportionate caloric intake Thermoneutrality suppresses thermogenesis and hyperphagia in PWS-IC del mice. Quantification of interscapular BAT mass (A) and Ucp-1 mRNA expression (B), inguinal WAT mass (C) and Ucp-1 mRNA expression (D), epididymal/omental (E) and retroperitoneal (R-P: F) WAT mass, body weight gain (G) and daily food intake (H) in 6-to 14-month-old male and female WT and PWS-IC del littermate mice maintained at either standard room temperature (20-22°C; white bars) or thermoneutrality (30°C; black bars) for 9 weeks. Data shown are mean ± s.e.m. (n = 5 (WT and PWS-IC del males at 30°C) and 6 (all other groups)), with statistical comparisons made with 1-way ANOVA and Bonferroni's selected pairs post hoc test (*P < 0.05, ***P < 0.001 vs WT male (room temperature); # P < 0.05 vs PWS-IC del male (at room temperature); • P < 0.05 vs WT female (at thermoneutrality)).

232:1
irrespective of which diet they were eating, PWS-IC del males were susceptible to overconsumption when eating the high-fat diet (Fig. 4E). Similar patterns of caloric intake were observed in WT and PWS-IC del females (Fig. 4E). Adiposity is influenced by the timing of food consumption (Glad et al. 2011), but day:night food intake ratios were not significantly different between WT and PWS-IC del mice (Fig. 4G). During this experiment, a marked difference in appetitive behaviour became apparent. When presented with refilled diet hoppers PWS-IC del mice immediately began moving the diet from the outer 'fill' compartment to the inner 'spill' compartment ( Supplementary Fig. 1A and B). When quantifying the amount of diet in the 'spill' compartment as a percentage of the remaining diet, PWS-IC del males were found to 'hoard' more than 3-times as much diet as WT males (P < 0.001; Fig. 4F and Supplementary Fig. 1A). This behaviour was prominent with both standard and high-fat diets, PWS-IC del males hoarding 86-91% of the remaining diet. Interestingly, this pattern of hoarding in PWS-IC del males was broadly similar to that in both WT and PWS-IC del females (Fig. 4F).
Although constrained by the short period of monitoring, it is particularly striking that although PWS-IC del mice consumed most calories when maintained on the high-fat diet, they not only failed to gain weight but also gained markedly less weight than their respective WT littermates (P < 0.01; Fig. 4A). Analysis of terminal blood samples revealed that mean circulating ghrelin (total) was tripled in PWS-IC del mice in comparison to their sex-matched littermates (Fig. 4B), although this was not statistically significant in PWS-IC del males (P = 0.116).
Study 5b. Post-fast feeding As PWS-IC del mice show proportionate hyperphagia, we investigated individual meal consumption patterns after an overnight fast. When adjusted for body weight, PWS-IC del mice consumed 48% more diet than WT littermates ( Fig. 4H; repeated measures ANOVA, main effect of GENOTYPE F 1,11 = 6.01, P = 0.032). This was due mainly to continued consumption after 30 min as evidenced by a significant interaction with the time of measurement (F 1,11 = 13.43, P = 0.004).

Discussion
Loss of paternal gene expression from the imprinted gene cluster on human chromosome 15q11-q13 impairs neuroendocrine and metabolic function in PWS. However, investigating these impairments in mouse models for PWS has been hampered by high postnatal lethality. We have utilised a mixed genetic background to increase survival, enabling us to characterise metabolic status in a mouse model in which deletion of the homologous PWS-IC results in complete loss of paternal gene expression from the entire PWS locus (Chamberlain et al. 2004, Relkovic et al. 2010. We now report that, with one notable exception, the metabolic phenotype of PWS is replicated in adult PWS-IC del mice. The growth abnormalities normally associated with PWS are recapitulated in PWS-IC del mice, with reductions in body weight and limb bone length combined with a 40-50% reduction in circulating IGF-1 (unpublished data). As similar growth results have been reported in another 'full' genetic mouse model for PWS (Stefan et al. 2005, Ding et al. 2008, whereas mice with specific deletion of Magel2 exhibit normal skeletal growth (Bischof et al. 2007); this aspect of the PWS phenotype appears to be dependent upon lack of expression of one or more of the other genes within the PWS locus. Indeed, the reduced circulating IGF-1 and elevated GH-releasing hormone mRNA reported in Snord116del mice (Qi et al. 2016) indicate that loss of Snord116 expression causes a pituitary-specific deficiency in the GH axis.
Despite growth retardation, PWS-IC del mice displayed a remarkable reduction in adiposity, affecting intraabdominal, subcutaneous and intramedullary (data not presented) depots. Although reduced visceral adiposity has been reported in human females with PWS (Goldstone et al. 2001(Goldstone et al. , 2004, the profound leanness in the PWS-IC del mice is unlike the obesity usually associated with this condition in humans. Indeed, the degree of leanness seen in the PWS-IC del model is considerably greater than the transient reduction in triglycerides in TgPWS mice (Stefan et al. 2005), and the mild reduction in fat mass in Snord116del mice (Ding et al. 2008, Qi et al. 2016. Given this surprising result, which corroborates previous reports in younger mice with IC deletions (Yang et al. 1998, Chamberlain et al. 2004, we examined the aetiology of this leanness, quantifying indices of energy intake, storage and utilisation. Leanness in PWS-IC del mice is clearly not due to impaired energy input, with males in particular showing proportionate hyperphagia, especially with a high-fat diet. Similarly, the normal day:night consumption ratios indicate that the altered lipid storage efficiency associated with daytime feeding in nocturnal rodents (Glad et al. 2011) does not contribute to the lean phenotype. The hyperphagia seen in PWS-IC del mice, which is driven by calorie seeking rather than hedonic processes (Davies et al. 2015), was even more exaggerated after overnight fasting, where prolonged consumption indicates a failure in meal termination characteristic of PWS. These indices of increased consummatory behaviour are likely to reflect the sustained hyperghrelinaemia observed in both the current and previous studies (Davies et al. 2015), but as this is not accompanied by an increase in NPY mRNA expression (unpublished data), the hypothalamic mechanisms underlying proportionate hyperphagia remain to be elucidated. That said, as Magel2-null mice show hypophagia (Bischof et al. 2007), whereas specific deletion of Snord116 results in a similar combination of impaired meal termination and leanness (Ding et al. 2008), these aspects of the PWS-IC del mouse phenotype probably result from loss of Snord116 expression.
When monitoring food intake, it became rapidly apparent that PWS males also displayed a remarkable increase in appetitive behaviour; moving crushed diet from the fill compartment to the spill compartment ( Supplementary Fig. 1A and B). When supplied with pelleted diet, individual PWS-IC del mice from our Cardiff colony displayed increased diet distribution in both home and metabolic cages ( Supplementary Fig. 1C), whereas PWS-IC del mice in a Florida colony packed bedding around high-fat diet pellets in the food hopper ( Supplementary Fig. 1D). This hoarding behaviour, previously described in birds, rodents and, in certain circumstances, humans (Deacon 2006, Keen-Rinehart et al. 2010) is most prominent in species not readily storing energy in the form of accumulated WAT (Keen-Rinehart et al. 2010) and is exaggerated after surgical removal of epididymal WAT (Dailey & Bartness 2008). This implies that the lean phenotype of these mice may be a significant causal factor in this behaviour. Interestingly, our data also indicate that in WT mice, hoarding appears to be a predominantly female behaviour. Therefore, the hyperphagia and hoarding behaviour observed in PWS-IC del males suggest that loss of paternal gene expression in the PWS locus results in a feminisation of feeding behaviour. Given that the central circuits regulating both consummatory and appetitive feeding behaviours are activated by ghrelin (Nakazato et al. 2001, Lawrence et al. 2002, Keen-Rhinehart & Bartness 2005, the elevation in food intake and diet hoarding may result from the sustained hyperghrelinaemia seen in this model (Davies et al. 2015) and in humans with PWS (Cummings et al. 2002, DelParigi et al. 2002, Haqq et al. 2003, Erdie-Lalena et al. 2006, Bizzari et al. 2010. However, it is interesting to note that PWS-IC del mice show a normal ghrelin response to fasting (Table 1) and re-feeding (unpublished data), as reported recently in humans with this condition (Kuppens et al. 2016).
It is also possible that proportionate hyperphagia may arise from impairment of the anorexic oxytocin circuitry. Although we have yet to quantify oxytocin expression in PWS-IC del mice, a reduction in the population of parvocellular oxytocin neurons has previously been described in humans with PWS (Swaab et al. 1995) and in mice with a specific deletion of Magel2 (Schaller et al. 2010). Indeed, postnatal injections of oxytocin have been reported to restore suckling in Magel2-null mice, preventing neonatal mortality (Schaller et al. 2010) and ameliorating deficits in social and learning behaviour in adults (Meziane et al. 2015). As Magel2 expression is absent in PWS-IC del mice (Chamberlain et al. 2004, Relkovic et al. 2010, it is likely that correcting any imbalance in the oxytocin system in these animals may have similar beneficial effects, but whether this will prevent hyperphagia in adults and restore fat deposition remains to be determined. In the context of elevated proportionate caloric intake (which was seen despite correcting for hoarding behaviour), it is possible that leanness occurs in PWS-IC del mice as a result of impaired nutrient absorption. For the dietary lipids at least this does not appear to be the case, as circulating sterol esters, which reflect dietary lipid absorption (Jones & Glomset 1985), are unaffected (Table 1). In addition, although lipid processing in the liver may be disturbed (Fig. 1), the absence of any significant reduction in circulating TAG (Table 1) indicates that hepatic lipid output was not significantly curtailed.
The normal circulating lipid profile in the context of proportionate hyperphagia implies an increase in substrate utilisation. However, we have previously reported that locomotor activity in PWS-IC del mice is significantly reduced (Doe et al. 2009, Relkovic et al. 2010. We therefore examined whether the energy utilisation associated with BAT-mediated thermogenesis was elevated in this model. This appears to be the case. The halving of isBAT weight, which reflects reduced TAG storage, together with the increased transdifferentiation of white adipocytes to 'beige' or 'brite' adipocytes, is consistent with an increased demand for heat production (Barbatelli et al. 2010, Cinti 2011, Schultz et al. 2013. Similarly, the increased surface temperature in the head and interscapular regions (directly above the isBAT) in PWS-IC del mice indicates an elevation in non-shivering thermogenesis. In this context, it should be noted that although reduced locomotor activity in PWS-IC del mice may increase the infra-red signal measured, this does not result in an increase in surface temperature in those regions not associated with BAT activity, or in radiant temperature, which would be particularly susceptible to a reduction in movement. Thus, despite an absence of increased UCP1 expression, the classic marker of thermogenic activity, in PWS-IC del mice (Fig. 3B), our data indicate that lipid utilisation in BAT is elevated in this model.
Evidence of elevated thermogenesis in PWS is sketchy, but our data extend previous reports that Snord116del mice show elevated daytime subcutaneous temperature (Lassi et al. 2016) and an improved ability to maintain core body temperature on cold exposure (Ding et al. 2008). In humans, a small but statistically significant proportion of PWS children are reported to experience persistent (Ince et al. 2005) or episodic hypothermia (surface temperature <94°F) (Williams et al. 1994), whereas fingertip temperature is unaltered (DiMario & Burleson 2002). As BAT is now emerging as a potentially important tissue in adult humans, particularly in the context of obesity (Cannon & Nedergaard 2010, Cinti 2011, a comprehensive analysis of BAT function in individuals with PWS is now required. Elevated thermogenesis in PWS-IC del mice is potentially significant because elevating BAT activity is highly effective in sequestering circulating lipid (Bartelt et al. 2011), thereby 'starving' WAT of energy substrate. When compounded by the increased energy utilisation of the more abundant 'beige'/'brite' adipocytes within the WAT stores, this could result in the intraabdominal leanness observed. To test this hypothesis, we examined the adiposity in PWS-IC del mice after suppressing thermogenic activity in a thermoneutral 232:1 environment.
Although the thermoneutral set point may be higher in smaller PWS-IC del mice (due to their increased surface area to volume ratio (Speakman & Keijer 2012)), maintaining mice at 30°C was clearly effective in suppressing thermogenesis, BAT mass increasing in WT and in PWS-IC del males (Fig. 3A). However, this failed to induce a corresponding elevation in WAT mass, body weight gain only being elevated transiently in WT females. This surprising outcome is most likely due to the influence of ambient temperature on food intake, proportionate hyperphagia in PWS-IC del males being abolished at thermoneutrality. Although this may suggest that simple elevations in ambient temperature may reduce orexigenic drive in human PWS, it demonstrates that elevated thermogenesis is not the primary cause of abdominal leanness in PWS-IC del mice.
In the absence of reduced energy intake/absorption or a causal increase in energy expenditure, leanness may result from a primary failure of lipid storage in WAT. Storage capacity is determined by adipogenesis (the terminal differentiation of pre-adipocytes), whereas lipid accumulates in response to increased substrate supply/ uptake, de novo lipogenesis and/or decreased lipolysis or lipid export, resulting in hypertrophy (Wells 2009).
We did not quantify intra-abdominal adipogenesis directly, but the reduced adiposity in the tibial marrow of PWS-IC del mice was partly due to a 50% reduction in adipocyte number (data not shown). Evidence of reduced adipogenesis in subcutaneous WAT has recently been reported in human PWS, adipocyte progenitor number being halved (Cadoudal et al. 2014). These findings are consistent with a reduction in circulating GH, as GH promotes pre-adipocyte differentiation (Guller et al. 1989), and GH replacement in children with PWS restores pre-adipocyte numbers (Cadoudal et al. 2014).
Despite a reduction in lipid storage capacity, humans with PWS still develop obesity, whereas in PWS-IC del mice, intra-abdominal adipocyte size is reduced by 69% (Fig. 1E). With lipolysis and lipid export apparently unaffected (as indicated by normal circulating FFAs and normal lipolytic responses in cultured human adipocytes (Cadoudal et al. 2014)), the only option remaining is a failure of lipid uptake and/or synthesis. To test this, we monitored weight gain in mice maintained on standard and high-fat diets. Although we acknowledge the limitation of short time periods used, it is remarkable that despite increased caloric intake on the high-fat diet, PWS-IC del males, unlike WT littermates, failed to gain weight. When taken together with the reduced adipocyte size and the absence of elevated WAT mass at thermoneutrality, this implies that PWS-IC del adipocytes fail to import energy substrate.
In summary, when considered alongside our previous studies, the current data indicate that PWS-IC del mice display many of the endocrine and metabolic characteristics of human PWS, including marked hyperghrelinaemia and proportionate hyperphagia (Fig. 5). However, despite these obesogenic influences Mechanisms of leanness in PWS-IC del mice. PWS-IC del mice display marked abdominal leanness (1; current study) despite a preponderance of obesogenic signals (underactive GH axis (current study), proportionate hyperphagia (with pronounced food hoarding) (current study; Davies et al. 2015), hyperghrelinaemia (current study; Davies et al. 2015) and reduced physical activity (Doe et al. 2009, Relkovic et al. 2010)) (2). Suppression of elevated thermogenesis (current study) in brown adipose tissue (BAT) by maintenance in a thermoneutral environment (3) fails to elevate white adipose tissue (WAT) mass (current study). A full colour version of this figure is available at http://dx.doi.org/10.1530/JOE-16-0367.