Environmental Enrichment Normalizes Metabolic Function in the Murine Model of Prader–Willi Syndrome Magel2-Null Mice

Abstract

Prader–Willi syndrome (PWS) is a rare genetic disease that causes developmental delays, intellectual impairment, constant hunger, obesity, endocrine dysfunction, and various behavioral and neuropsychiatric abnormalities. Standard care of PWS is limited to strict supervision of food intake and GH therapy, highlighting the unmet need for new therapeutic strategies. Environmental enrichment (EE), a housing environment providing physical, social, and cognitive stimulations, exerts broad benefits on mental and physical health. Here we assessed the metabolic and behavioral effects of EE in the Magel2-null mouse model of PWS. EE initiated after the occurrence of metabolic abnormality was sufficient to normalize body weight and body composition, reverse hyperleptinemia, and improve glucose metabolism in the male Magel2-null mice. These metabolic improvements induced by EE were comparable to those achieved by a hypothalamic brain-derived neurotrophic factor gene therapy although the underlying mechanisms remain to be determined. These data suggest biobehavioral interventions such as EE could be effective in the treatment of PWS-related metabolic abnormalities.

Prader–Willi syndrome (PWS) is a rare genetic disorder that occurs in approximately 1 in 15 000 individuals (1). PWS is caused by mutation/deletion of a paternally inherited copy of chromosome 15q11-q13 (2, 3). Newborns with PWS exhibit poor feeding and development delay, while older children and adults display intellectual impairment; behavioral and neuropsychiatric abnormalities such as obsessive-compulsive behavior, anxiety, and temper tantrums; endocrine dysfunction such as GH deficiency; hypothalamic hypogonadism; and constant hunger (hyperphagia) that often leads to obesity and type 2 diabetes (4-6).

PWS has no cure. To date, GH therapy and strict supervision of daily food intake have been the standard of care for PWS (7, 8). GH therapy has several shortcomings including patient compliance; lack of efficacy data in older adults; inability to target the hypothalamic origins of PWS; and exclusion criteria including common comorbidities of PWS such as severe obesity, uncontrolled diabetes, and active psychosis (9). Strict supervision and management of individuals with PWS lead to high levels of caregiver burden (10-13), which may exceed that of caregivers for patients with Alzheimer's and traumatic brain injury (12, 13). Recent pharmacological development has focused on targeting the hypothalamic leptin-proopiomelanocortin (POMC) pathway such as melanocortin-4 receptor (MC4R) agonist melanotan II and setmelanotide that have known roles in food intake reduction (14, 15). Other categories of therapeutics have been tested in preclinical models and clinical trials at various stages. However, there are no Food and Drug Administration-approved therapies to treat compulsive eating and emotional reactivity that cause the day-to-day challenges PWS patients and families face (16), highlighting the urgent need for new therapeutic strategies.

MAGE family member L2 (MAGEL2) is 1 of the 5 protein-coding, maternally imprinted, paternally expressed genes in the PWS-critical domain on chromosome 15q11-q13. Loss of MAGEL2 function contributes to several aspects of PWS pathophysiology (17), including disturbance in the hypothalamic leptin-POMC pathway, which plays a critical role in regulating feeding and energy expenditure. Studies in a mouse model of Magel2 deficiency have revealed that Magel2 is required for leptin-mediated depolarization of hypothalamic anorexigenic POMC neurons (14). Moreover, the loss of Magel2 results in reduced anorexigenic α-melanocyte-stimulating hormone axons (18, 19), while orexigenic agouti-related peptide fibers are unaffected (19, 20). These defects in the PWS-driven absence of Magel2 result in disrupted feeding behavior and energy homeostasis. In addition, further characterizations have demonstrated that the Magel2 mutant mice recapitulate many of the phenotypes of PWS and are a valid animal model for preclinical interventional studies (16).

Brain-derived neurotrophic factor (BDNF) is a potential therapeutic target for PWS because BDNF functions downstream in the leptin-POMC pathway and regulates energy homeostasis and behavior (21-28) In fact, reductions in peripheral BDNF have been reported in individuals with PWS (29), and reduced BDNF expression is thought to cause PWS-related transcriptomic alterations in the hypothalamus (23). However, therapeutic use of BDNF protein has been largely unsuccessful due to unfavorable pharmacokinetics and the need for fine delivery localization and repeated dosing (30-32) Our previous work has shown that hypothalamic injection of an autoregulatory adeno-associated virus (AAV)-BDNF vector efficiently alleviates deficits in the leptin-POMC signaling pathway, thereby mitigating obesity; impaired glycemic control; hepatic steatosis; and other metabolic syndromes in genetic, diet-induced, and aging-related models of obesity (33-35) We recently published that hypothalamic AAV-BDNF gene therapy resulted in metabolic improvements including reduced adiposity, increased lean mass, elevated energy expenditure, improved glycemic control, reduced circulating leptin level, and amelioration of neuroinflammation in Magel2-null mice without adverse behavioral effects (36).

Although preclinical testing has demonstrated a high level of efficacy and safety, the translation of hypothalamic BDNF gene therapy remains challenging. Thus, alternative approaches to stimulating hypothalamic BDNF are worthy of investigation. Previous work from our lab has demonstrated that environmental enrichment (EE), a housing environment providing complex stimulations, exerts a wide range of benefits on body composition, energy balance, immunity, cancer, behavior, and healthy aging (37-44) We have identified hypothalamic BDNF as a key mediator in the brain coordinating the diverse health benefits induced by EE (37, 40, 43-46) However, it was unknown whether the Magel2-null mice were responsive to EE. Here, we investigated whether an environmental approach such as EE could mimic the metabolic and behavioral effects of hypothalamic BDNF gene therapy (36) in the Magel2-null murine model of PWS.

Materials and Methods

Mice

Magel2-null mice harbor a maternally inherited imprinted/silenced wild-type (WT) allele and a paternally inherited Magel2-lacZ knockin allele that abolishes endogenous Magel2 gene function. Male mice containing the Magel2-lacZ allele (Jackson Labs #009062, RRID:IMSR_JAX:009062) on C57BL/6J background were bred with female C57BL/6J mice to produce both WT mice and Magel2-null littermates. Mice were genotyped from ear notch biopsies. Identification of mutant offspring was performed by PCR genotyping with Magel2 and LacZ oligonucleotide primers (common forward, 5′-ATGGCTCCATCAGGAGAAC; Magel2 reverse, 5′-GATGGAAAGACCCTTGAGGT; and LacZ reverse, RW4237, 5′-GGGATAGGTCACGTTGGTGT). Genotype was further confirmed by quantitative RT-PCR (qRT-PCR) detecting Magel2 transcript in the hypothalamus (primers see Table 1).

Table 1.

Primer sequences used for quantitative PCR

Gene	Sequence
ActinB	ACCCGCGAGCACAGCTT ATATCGTCATCCATGGCGAACT
Adrb3	GGACGCTGTTCCTTTAAAAGCA TCCATCTCACCCCCCATGT
Adipoq	CCCTCCACCCAAGGGAACT CCATTGTGGCCAGGATGTC
Aif1	GGAGATTTCAAAAGCTGATGTGGA CCTCAGACGCTGGTTGTCTT
Aim2	CTGGCCGCATAGTCATCCTT AGTCCCAGGATCAGCCTAGA
Bdnf	CCATAAGGACGCGGACTTGT AGGCTCCAAAGGCACTTGACT
Cfd	GTGCAAGTGAACGGCACA GTCGTCATCCGTCACTCCATC
Crh	TGGCCCCAAGGAGGAAA CCACTGCAGCTCCAAATAAAAA
Cxcl10	AAGTGCTGCCGTCATTTTCT CTTCCCTATGGCCCTCATTC
Cx3cr1	TCACCGTCATCAGCATCGAC CGCCCAGACTAATGGTGACA
Fh1	CGCGAATGGCAAGCCAAAAT CGTTCCGTAGCACCTCCAAT
Gpat	CAACACCATCCCCGACATC TGACCTTCGATTATGCGATCAT
Hsl	GCGCCAGGACTGGAAAGAAT TGAGAACGCTGAGGCTTTGAT
Lep	ATTTCACACACGCAGTCGGTAT AGCCCAGGAATGAAGTCCAA
Lpl	TCGTCATCGAGAGGATCCGA TGTTTGTCCAGTGTCAGCCA
Nfkbia	TGCCTGGCCAGTGTAGCAGTCTT CAAAGTCACCAAGTGCTCCACGAT
Obrb	AATGACGCAGGGCTGTATGT TCAGGCTCCAGAAGAAGAGG
Parp1	AAGGCGGAGAAGACATTGGG CGCATCTGGCCCTTCTCTAT
Pparg	ATGGGTGAAACTCTGGGAGATTCA CTTGGAGCTTCAGGTCATATTTGTA
Ppargc1a	AAGTGTGGAACTCTCTGGAACTG GGGTTATCTTGGTTGGCTTTATG
Pten	TGGATTCGACTTAGACTTGACCT GCGGTGTCATAATGTCTCTCAG
Retn	AAGAACCTTTCATTTCCCCTCCT GTCCAGCAATTTAAGCCAATGTT
Socs3	ATGGTCACCCACAGCAAGTTT TCCAGTAGAATCCGCTCTCCT
Trem2	ACAGCACCTCCAGGAATCAAG AACTTGCTCAGGAGAACGCA
TrkB-FL	GACAATGCACGCAAGGACTT AGTAGTCGGTGCTGTACACA
Vegfa	TACCTCCACCATGCCAAGTG CATGGGACTTCTGCTCTCCTTCT
Hprt1	TGTTGTTGGATATGCCCTTG GCGCTCATCTTAGGCTTTGT
Magel2	ACAGGCAAATCCCGTTCACT CTGGTTCTCGTAGAGTGCGG

EE Protocol

Magel2-null mice develop metabolic deficiencies over time (18, 19); systemic manifestation occurs by 16 weeks of age (47). Adult male mice, 20 to 40 weeks of age, were randomized to live in standard environment (SE) or EE housing for 16 weeks. SE mice were group housed (5 mice) in standard laboratory cages (19.4 cm × 18.1 cm × 39.8 cm). EE mice were group housed (5 mice) in larger cages (63 cm × 49 cm × 44 cm) supplemented with running wheels, igloos, toys, tunnels, a maze, and nesting material (48). Mice were housed within temperature (22-23 °C) and humidity (30-70%) controlled rooms under a 12:12 light:dark cycle. All mice had ad libitum access to food (normal chow diet, 11% fat, caloric density 3.4 kcal/g, Teklad) and water. All animal experiments were in accordance with the regulations of The Ohio State University's Institutional Animal Care and Use Committee, protocol approval 2012A00000060-R3.

Body Composition Assessment

Body composition was assessed with a 3-in-1 Analyzer (EchoMRI LLC, Houston, TX) prior to initiation of EE and at 4 weeks and 12 weeks posthousing. Fat and lean mass were measured in live mice without anesthesia according to manufacturer instructions. Mice were subjected to a 5-Gauss magnetic field, and whole-body masses of fat, lean, free water, and total water were determined during separate cycles by manufacturer software comparison to a canola oil standard.

Body Weight and Food Intake Measurement

Body weight was recorded weekly throughout the experiment. Food intake was measured at the cage level on a weekly basis throughout the experiment. Food intake data collected in weeks during which invasive profiling occurred (eg, overnight fast for glucose tolerance test) were excluded. Average food intake was calculated on a per-mouse per-day basis.

Glucose Tolerance Test

At 5 weeks posthousing, mice were injected intraperitoneally with glucose solution (2.0 g glucose per kg body weight) after a 16-hour overnight fast. Blood was obtained from the tail at 0, 15, 30, 60, 90, and 120 minutes after glucose injection. Blood glucose concentrations were measured with a portable glucose meter using default manufacturer settings (Bayer Contour Next). Following blood collection, styptic powder was used to stop bleeding.

Open Field Test

At 9 weeks posthousing, mice were individually placed into the center of an open square arena (60 cm × 60 cm, enclosed by walls of 48 cm). Each mouse was allowed to explore the arena for 10 minutes, during which time and activity—in the center and the periphery of the open field—was recorded and analyzed using TopScan (Clever Sys, Inc.) software. Between each trial, the arena was cleaned with Opticide to remove odor cues. Behavioral test was performed at 12 Pm and later.

3-Chamber Sociability Test

At 10 weeks posthousing, mice were placed in an apparatus consisting of 3 connected plexiglass chambers (18 cm × 41 cm × 20 cm each) with removable dividers between each chamber. Each test subject was individually placed in the center plexiglass chamber for 5 minutes of habituation. In the first phase—testing social affiliation—another, unfamiliar mouse was placed in either the right or left chamber in a small wire cage, while another wire cage remained empty in the opposite chamber. The wire cage restricted social or aggressive interactions between the 2 mice beyond nose contact. Chamber dividers were lifted after the habituation period to allow the test subject to move freely about all 3 chambers for a 10-minute observation period. A second 10-minute test—assessing novel social engagement—was performed immediately afterward, using the conspecific from the first test (now denoted as a familiar mouse) and a novel unfamiliar mouse in the opposite chamber. Between each trial, the arena was cleaned with Opticide to remove odor cues. Trials were video recorded. An observer unaware of genotype/treatment used an open-source event-logging software, BORIS (49), to create a user-scored ethogram with timestamps and length calculations for behavioral activities (ie, chamber entry, mouse investigation). Once the behavioral coding process was completed, observation data were exported and analyzed. The social preference index was calculated by dividing the amount of time spent with the novel mouse by the amount of time spent in the empty chamber during the first test phase. The social novelty index was calculated by dividing the amount of time spent with the novel unfamiliar mouse by the amount of time spent with the familiar mouse during the second test phase. A behavioral test was performed at 12 Pm and later.

Novel Object Recognition Test

At 11 weeks posthousing, mice were subjected to the novel object recognition test (50). For the test familiarization period, mice were placed in an open arena (60 cm × 60 cm, enclosed by walls of 48 cm) with 2 identical objects (either 2 Falcon tubes filled with water or 2 stacks of large Legos; habituation objects were varied between mice). Mice were allowed to explore the identical objects until 20 seconds of cumulative exploration time was achieved. Mice were returned to their home cage while the arena/objects were cleaned with 70% ethanol. During the novel object recognition test session, the 2 training objects were replaced with 1 matched item from the training session and a novel item. Mice were placed in the arena and allowed to explore both the novel and the learned object until 20 seconds of cumulative exploration time was achieved, with a maximal time of 10 minutes to reach the criteria; those that did not meet the criteria were excluded from statistical analyses. Time spent exploring each respective object was recorded. Mice were returned to their home cages, and the arena/objects were cleaned with 70% ethanol. Exploration activity was defined as “directing the nose toward the object at a distance less than or equal to 2 cm,” and time spent climbing on or chewing on objects was not deemed exploration activity (50). The discrimination index was calculated by dividing the novel object exploration time by the total amount of object exploration during the test. A behavioral test was performed at 12 Pm and later.

Tissue Collection

At 16 weeks posthousing, mice were euthanized following a 4-hour fast. Mice were anesthetized with 2.5% isoflurane (1.0 L/min) and then decapitated to collect trunk blood. Tissues to be used for mRNA and protein analyses were flash-frozen on dry ice and stored at −80 °C until further analysis. Hypothalamus was collected under a dissection microscope at sacrifice, with the left and right side being collected separately to allow for RNA and protein isolation from each mouse.

Serum Harvest and Analysis

Trunk blood was collected at euthanasia, clotted on ice, and centrifuged at 10 000 rpm for 10 minutes at 4 °C. The serum component was collected and stored at −20 °C until further analysis. R&D Systems ELISA kits were used to assay serum leptin (R and D Systems Cat# DY498, RRID: AB_3668792) and adiponectin (R and D Systems Cat# DY1119, RRID:AB_3668794). Additional ELISAs were performed for insulin (Alpco Diagnostics Cat# 80-INSMSU-E01, RRID: AB_2792981) and corticosterone (Enzo Life Sciences Cat# ADI-900-097, RRID:AB_2307314). The Caymen Chemical colorimetric assay kit was used to assay glucose (#10009582). The homeostatic model assessment for insulin resistance index was calculated as [fasting serum glucose (mmol/L) × fasting serum insulin (pmol/L) / 22.5] as described elsewhere (51).

qRT-PCR

Following tissue sonication, RNA was isolated using the QIAGEN RNeasy Mini kit (#74804) with RNase-free DNase treatment. cDNA was reverse transcribed using Taqman Reverse Transcription Reagents (Applied Biosystems #N8080234). qRT-PCR was conducted on the StepOnePlus Real-Time PCR System using Power SYBR Green PCR Master Mix (Applied Biosystems #A25742). Primer sequences are available in Table 1. Data were calibrated to endogenous controls—Hprt1 for hypothalamus and Actinb for adipose tissues. The relative gene expression was quantified using the 2 ^−ΔΔCT method.

Western Blotting

One lobe of hypothalamus per mouse was homogenized in ice-cold radioimmunoprecipitation assay buffer (Pierce #89901) containing 1 × PhosSTOP (Roche #4906845001) and protease inhibitor cocktail III (Calbiochem #539134). Tissues lysates were spun at 13 000 rpm for 15 minutes at 4 °C. The supernatant was collected, and the protein concentration was determined with a bicinchoninic acid protein assay kit (Pierce). Protein from each sample was loaded (15 µg) and separated by gradient gel (4-20%, Mini-PROTEAN TGX, Bio-Rad), then transferred to a nitrocellulose membrane (Bio-Rad). Blots were incubated overnight at 4 °C with the following primary antibodies: Vinculin (Cell Signaling Technology Cat# 13901, RRID:AB_2728768, 1:1000), Phospho-AKT-Ser473 (Cell Signaling Technology Cat# 9271, RRID:AB_329825, 1:1000), total AKT (Cell Signaling Technology Cat# 9272, RRID:AB_329827, 1:1000), Phospho-p44/42 MAPK (T202/Y204) (Cell Signaling Technology Cat# 9101, RRID:AB_331646, 1:1000), Total p44/42 MAPK (Cell Signaling Technology Cat# 4695, RRID:AB_390779, 1:1000), Phospho-p38 MAPK (T180/Y182) (Cell Signaling Technology Cat# 9211, RRID: AB_331641, 1:1000), Total p38 MAPK (Cell Signaling Technology Cat# 9212, RRID:AB_330713, 1:1000), Ras (Cell Signaling Technology Cat# 3965, RRID:AB_2180216, 1:1000), TrkB (Cell Signaling Technology Cat# 4603, RRID:AB_2155125, 1:1000), and Iba1 (FUJIFILM Wako Pure Chemical Corporation Cat# 016-20001, RRID:AB_839506, 1:500). Blots were rinsed and incubated with horseradish peroxidase-conjugated secondary antibodies (Bio-Rad, 1:3000). Chemiluminescence signal was detected and visualized by the LI-COR Odyssey Fc imaging system (LI-COR Biotechnology, Lincoln, NE). Quantification analysis was carried out with Image Studio software version 5.2 (LI-COR Biotechnology).

Quantification and Statistical Analysis

Statistical analyses were performed using GraphPad Prism 10 software (GraphPad, San Diego, CA). One-way ANOVA with Tukey's post hoc test was utilized for comparisons between 3 groups. Time course data (body weight, glucose tolerance test) were analyzed using a mixed ANOVA, and area under the curve calculations were performed where applicable. Normality was tested using the Shapiro–Wilk method. Outliers were determined and removed using the ROUT method. Significance was defined as P < .05. Data are reported as means ± SEM. Sample sizes (n) can be found in each figure's respective legend.

Results

EE Normalizes Metabolic Function in Magel2-null Mice

Male Magel2-null mice and WT littermates, 20 to 40 weeks of age, were randomized to live in EE housing or SE housing for a total of 16 weeks. A battery of metabolic and behavioral assessments was conducted as shown in Fig. 1A. As occasional genotyping error occurred in our previous studies, the Magel2 mRNA was determined by qRT-PCR in the hypothalamus samples at termination of the experiment. Among the 30 mice enrolled in the study, only 1 mouse originally assigned to the Magel2-null SE group was in fact WT, and therefore the in vivo and postmortem data of this animal were grouped to the WT SE.

Figure 1.

EE alters body weight and food intake. (A) Experimental timeline of metabolic and behavioral assessments. (B) Body weight. (C) Body weight change (percentage of baseline). (D) Average daily food intake. Data are means ± SEM. Sample size: WT SE n = 11, Magel2-null SE n = 9, Magel2-null EE n = 10. * or # P < .05, ** or ## P < .01, *** or +++ P < .001, **** or ++++ P < .0001. For time course measurement in B and C, # denotes WT SE vs Magle2-null SE, * denotes Magel2-null SE vs Magel2-null EE, and + denotes WT SE vs Magel2-null EE.

Abbreviations: EE, environmental enrichment; SE, standard environment; WT, wild-type.

Body weight and food intake were monitored weekly throughout the 16-week experiment. Magel2-null mice living in SE (Magel2-null SE) exhibited significantly higher body weight compared to WT mice living in SE (WT SE) (Fig. 1B). Magel2-null SE mice displayed slightly smaller weight gain compared to WT SE but did not reach significance (Fig. 1C). In contrast, Magel2-null mice living in EE (Magel2-null EE) exhibited a significant weight loss as early as 1 week posthousing, and the weight loss was maintained over the 16-week experiment. As a result, EE reversed the genotype effect on body weight (Fig. 1B and C). Consistent with previous reports that Magel2-null mice do not exhibit hyperphagia (47, 52), food intake in Magel2-null SE mice was not different from WT SE mice (Fig. 1D). Interestingly, Magel2-null EE mice displayed a significant increase in food intake compared to mice living in SE regardless of the genotype (Fig. 1D).

Prior to the initiation of EE, echo magnetic resonance imaging was performed to assess body composition to allow for experimental group randomization. Consistent with genotype, the adiposity of Magel2-null mice was approximately 3 times that of the adiposity of WT mice at baseline associated with significantly decreased lean mass (Fig. 2A and B). Four-weeks EE housing completely reversed the excessive adiposity and low lean mass in Magel2-null mice, and the normalized body composition sustained to 12 weeks posthousing (Fig. 2A and B).

Figure 2.

EE improves body composition and glucose tolerance. (A) Relative fat mass measured by echoMRI. (B) Relative lean mass measured by echoMRI. Data are means ± SEM. Sample size: WT SE n = 11, Magel2-null SE n = 9, Magel2-null EE n = 10. (C) Glucose tolerance test at 5 weeks posthousing. (D) Area under the curve of the glucose tolerance test. Data are means ± SEM. Sample size: WT SE n = 10, Magel2-null SE n = 8, Magel2-null EE n = 8. * or # or + P < .05, ** or ++ P < .01, *** or +++ P < .001, **** P < .0001. For time course measurement in C, # denotes WT SE vs Magle2-null SE, * denotes Magel2-null SE vs Magel2-null EE, and + denotes WT SE vs Magel2-null EE.

Abbreviations: EE, environmental enrichment; MRI, magnetic resonance imaging; SE, standard environment; WT, wild-type.

At 5 weeks posthousing, a glucose tolerance test was performed. Magel2-null SE mice exhibited impaired glucose tolerance compared to WT SE mice. This genotypic deficit was rescued by EE. Moreover, Magel2-null EE mice displayed improved glycemic control over WT mice (Fig. 2C and D). In addition, the elevated fasting glucose level in Magel2-null SE mice was reversed by EE housing (Fig. 2C).

EE Exerts Minor Behavioral Effects in Magel2-null Mice

To assess whether EE could improve aberrant behaviors in Magel2-null mice, behavioral profiling was performed following metabolic assessments. At 9 weeks posthousing, an open field test (OFT) (53) was performed to assess exploratory activity and anxiety-like behavior (54, 55). No genotype-driven effect was observed in total distance traveled, a parameter of exploratory activity and locomotion (Fig. 3A), whereas Magel2-null EE mice exhibited significantly reduced locomotion compared to both Magel2-null SE and WT SE mice (Fig. 3A). No changes were observed in percent distance travelled in the center (Fig. 3B) or periphery (Fig. 3C) of the open field arena, indicating no changes in anxiety-like behavior.

Figure 3.

EE exerts minor behavioral effects. (A) Total distance traveled in the open field arena. (B) Percentage distance traveled in the center of the open field arena. (C) Percentage distance traveled in the periphery of the open field arena. Data are means ± SEM. Sample size: WT SE n = 11, Magel2-null SE n = 9, Magel2-null EE n = 9. (D) Social preference index as measured during the 3-chamber sociability test. (E) Social novelty index as measured during the 3-chamber sociability test. Data are means ± SEM. Sample size: WT SE n = 11, Magel2-null SE n = 8, Magel2-null EE n = 10. (F) Novel object discrimination index. Data are means ± SEM. Sample size: WT SE n = 11, Magel2-null SE n = 9, Magel2-null EE n = 10. Data are means ± SEM *P < .05, ***P < .001.

Abbreviations: EE, environmental enrichment; SE, standard environment; WT, wild-type.

It has been reported that Magel2-null mice manifest altered social phenotypes (55, 56). At 10 weeks posthousing, the 3-chamber sociability test (57) was performed to assess social affiliation and social novelty preference. Mice were given the opportunity to wander 3 chambers while investigating social stimuli. In the first test phase, mice were exposed to a novel confined peer and an opposing empty chamber to examine social affiliation. In the second test phase, mice were exposed to both familiar and novel confined peers in opposing chambers to assess social novelty engagement (57). No genotype- or housing-induced alterations were observed in social preference during the first test phase (Fig. 3D) or in social novelty-seeking behavior during the second test phase (Fig. 3E).

At 11 weeks posthousing, the novel object recognition test was performed to assess recognition memory. Female Magel2-null mice have been shown to be averse to novel objects and environments (54). However, a reduced novel object discrimination index was not observed in the male Magel2-null SE mice over WT SE mice (Fig. 3F). The novel object discrimination index in Magel2-null EE mice was slightly higher but not significant (Fig. 3F).

EE Reduces Adipose Tissue Depot Size and Normalizes Circulating Biomarkers in Magel2-null Mice

Mice were euthanized at 16 weeks posthousing after a 4-hour fast, and tissues were harvested. Consistent with body composition measurement by echo magnetic resonance imaging, Magel2-null SE mice exhibited increased relative mass of 3 white adipose tissue (WAT) depots—the inguinal white adipose tissue (iWAT), epidydimal white adipose tissue, and retroperitoneal white adipose tissue (rWAT)— compared to WT SE mice (Fig. 4A). The enlarged WAT depots in Magel2-null mice were normalized with EE housing (Fig. 4A). The relative brown adipose tissue weight in Magel2-null SE mice was trending higher compared to WT SE mice although not significant (P = .09). EE significantly decreased the relative brown adipose tissue mass compared to Magel2-null SE mice (Fig. 4A). No genotype- or housing-driven changes were observed in relative liver weight or relative pancreas weight (Fig. 4A).

Figure 4.

EE reduces adipose tissue mass and normalizes hyperleptinemia. (A) Relative tissue weight at euthanasia 16 weeks posthousing. (B) Serum leptin. (C) Serum adiponectin. (D) Serum corticosterone. (E) Serum insulin. (F) Serum glucose. (G) Homeostatic model assessment for insulin resistance index. Data are means ± SEM. Sample size: WT SE n = 11, Magel2-null SE n = 9, Magel2-null EE n = 10. *P < .05, **P < .01, ***P < .001, ****P < .0001.

Abbreviations: EE, environmental enrichment; SE, standard environment; WT, wild-type.

Serum samples were profiled to assess systemic changes following EE housing. The adipokine leptin plays a key role in regulating energy homeostasis, and its level is positively correlated with adipose tissue mass (58). Previous study reports that loss of Magel2 results in leptin resistance in the hypothalamic POMC neurons contributing to the metabolic deficits in Magel2-null mice (14). Hyperleptinemia is a prominent feature of Magel2 knockout. Consistent with literature, Magel2-null SE mice exhibited a drastic increase in serum leptin levels over WT counterparts (Fig. 4B). This hyperleptinemia in Magel2-null mice was ameliorated with EE housing (Fig. 4B). Adiponectin, an abundant adipokine, plays roles in insulin sensitivity, glucose homeostasis, and other systemic metabolic function (59). Contrary to leptin, circulating adiponectin levels are negatively correlated with fat mass and impaired glucose tolerance in the general population. However, adiponectin levels were significantly higher in subjects with PWS than in non-PWS obese controls (60). Magel2-null SE mice exhibited a significant increase in serum level of adiponectin as compared to WT counterparts (Fig. 4C). EE reversed the elevated adiponectin level in Magel2-null mice (Fig. 4C). Our previous studies have shown a significant but mild increase in serum level of corticosterone in WT mice after exposure to EE (43, 46). Interestingly, Magel2-null SE mice displayed a significant increase in serum corticosterone as compared to WT SE mice (Fig. 4D), consistent with a previous report of this model (61). This genotype-driven effect was reversed by EE (Fig. 4D). No significant genotype- or housing-induced changes were observed in serum insulin or glucose levels (Fig. 4E and F). The homeostatic model assessment for insulin resistance index was lower in the Magel2-null EE mice as compared to Magel2-null SE mice but did not reach significance (P = .088), suggesting a trend toward improved insulin sensitivity following EE (Fig. 4G).

EE Modulates WAT Gene Expression in Magel2-null Mice

Our previous work has described a profound adipose remodeling induced by EE mediated by a specific brain-adipose axis—the hypothalamic-sympathoneural-adipocyte axis (40). Environmental and genetic stimulation of hypothalamic BDNF confers a white-to-brown adipose phenotypic shift via increased sympathetic tone. Accordingly, the gene expression signature induced by EE observed in our previous studies was profiled. In the visceral WAT depot, the rWAT (Fig. 5A), a genotype-induced reduction in Adrb3 (encoding adrenoceptor β3) was observed, which was rescued by EE. Similarly, the genotype-induced reduction in Cfd (encoding complement factor D, or adipsin) was normalized by EE. In contrast, Magel2-null SE mice showed a sharp upregulation of Lep (encoding leptin) in rWAT compared to WT counterparts, which was corrected by EE (Fig. 5A). EE-induced upregulation was observed in Hsl (encoding hormone-sensitive lipase), crucial enzyme in lipolysis; Gpat (encoding glycerol-3-phsphate acyltransferase), the rate-limiting enzyme in the de novo glycerolipid synthesis; Ppargc1a (encoding peroxisome proliferator-activated receptor gamma coactivator 1-alpha), a transcriptional coactivator regulating genes involved in energy metabolism (62); Pten (encoding phosphatase and tensin homolog), which plays a role in regulating the size of adipocytes (63); Fh1 (encoding mitochondrial fumarate hydratase 1), involved in thermogenesis; Parp1 (encoding poly ADP-ribosyl transferase-1), associated with metabolic adaption; and Vegfa (encoding vascular endothelial growth factor A), which regulates EE-induced beige cells (45). No significant differences in Adipoq (encoding adiponectin), Lpl (encoding lipoprotein lipase), Pparg (encoding peroxisome proliferator-activated receptor gamma), or Resn (encoding resistin) were observed (Fig. 5A).

Figure 5.

EE alters white adipose tissue gene expression. (A) Relative gene expression in the retroperitoneal white adipose tissue. (B) Relative gene expression in the inguinal white adipose tissue. Data are means ± SEM. Sample size: WT SE n = 5, Magel2-null SE n = 5, Magel2-null EE n = 5. *P < .05, **P < .01, ***P < .001.

Abbreviations: EE, environmental enrichment; SE, standard environment; WT, wild-type.

The panel of gene expression was profiled in the subcutaneous WAT depot, iWAT (Fig. 5B). Consistent with observations in the rWAT, Magel2-null mice exhibited downregulation of Cfd and upregulation of Lep as to WT counterparts; both were corrected by EE (Fig. 5B). Contrary to the EE-induced upregulation in the rWAT (Fig. 5A), downregulation of Adrb3 and Hsl was observed in the iWAT (Fig. 5B). Overall, fewer gene expression changes were observed in the iWAT.

EE Alters Hypothalamic Gene Expression and Signaling Pathways in Magel2-null Mice

Previous work from our lab has identified hypothalamic BDNF as the key mediator in the brain coordinating the metabolic and immune adaptions induced by EE (37, 40, 43-46) Hypothalamic BDNF gene therapy altered hypothalamic gene expression and signaling pathways in Magel2-null mice (36, 64). Accordingly, gene expression was profiled in 1 hemisphere of hypothalamus by qRT-PCR (Fig. 6A), and signal molecules were probed in the other hemisphere of hypothalamus by Western blotting (64) (Fig. 6B). Unexpectedly, Bdnf was upregulated in Magel2-null SE mice as compared to WT counterparts, which was reversed by EE (Fig. 6A). BDNF receptor TrkB-FL (encoding tropomyosin receptor kinase B full length) was upregulated in Magel2-null mice regardless of housing (Fig. 6A). Crh (encoding corticotrophin-releasing hormone) was downregulated in Magel2-null mice in both housing conditions. Consistent with previous findings (64), microglial-related genes Aif1 (encoding allograft inflammatory factor 1, also known as Iba1) and Trem2 (encoding triggering receptor expressed on myeloid cells 2) were upregulated in Magel2-null SE compared to WT counterparts. Hypothalamic BDNF gene therapy was reported to correct this genotype-driven upregulation of microglial markers (64). In contrast, EE had no significant effects on Aif1 or Trem2. Socs3 (encoding suppressor of cytokine signaling 3) was downregulated in Magel2-null mice regardless of housing. Nfkbia (encoding nuclear factor-kappa-B-inhibitor alpha) was upregulated in EE while no genotype effect was observed (Fig. 6A). No significant differences in Obrb (encoding leptin receptor long form), Cx3cr1 (encoding CX3C motif chemokine receptor 1), Cxcl10 (encoding C-X-C motif chemokine ligand 10), or Aim2 (encoding absent in melanoma 2) were observed (Fig. 6A).

Figure 6.

EE affects hypothalamic gene expression and signaling pathways. (A) Relative gene expression in the hypothalamus. Data are means ± SEM. Sample size: WT SE n = 5, Magel2-null SE n = 5, Magel2-null EE n = 5. (B) Western blotting and quantification. Data are means ± SEM. Sample size: WT SE n = 4, Magel2-null SE n = 5, Magel2-null EE n = 5. # P < .06, *P < .05, **P < .01, ***P < .001.

Abbreviations: EE, environmental enrichment; SE, standard environment; WT, wild-type.

Immunoblotting was performed to probe signaling molecules downstream of the TrkB receptor in the hypothalamus. A previous study reported a significant reduction of Ras level in Magel2-null mice treated with AAV-YFP as compared to WT mice treated with AAV-YFP. This genotype-driven reduction of Ras was normalized by AAV-BDNF treatment (64). Here, EE exerted a similar effect although did not reach significance (P = .057, Fig. 6B). Contrary to a previous report (64), no genotype-related decrease in phosphorylation of ERK was observed in Magel2-null SE mice. Instead, EE was associated with a significant reduction of ERK phosphorylation (Fig. 6B). No significant differences in phosphorylation of AKT or p38 MAPK were observed (Fig. 6B). The protein levels of Iba1/AIF1 (Fig. 6B) were inconsistent with mRNA changes (Fig. 6A). Inconsistency between mRNA and protein was also observed in TrkB-FL. Magel2-null mice displayed a significantly lower level of truncated TrkB in both SE and EE as compared to WT while the ratio of full length to truncated TrkB was not altered (Fig. 6B).

Discussion

EE refers to a housing with increased space, physical activity, and social interactions that facilitate enhanced sensory, cognitive, motor, and social stimulation (65, 66). A large body of work has revealed a wide range of cerebral changes of EE including enhanced neurogenesis, improved learning and memory, neuroprotection against brain insult, and modulation of neuroinflammation (65, 67). Beneficial effects of EE have been demonstrated in many neural and psychiatric disease models such as Alzheimer's disease, Huntington's disease, Parkinson's disease, autism, stroke, depression, anxiety, and drug addiction (68-71) Relative to the extensive investigations on the neurobehavioral effects, metabolic and other peripheral outcomes of EE have not been the focus in the field. Our work in the past decade has demonstrated that EE has robust effects on body composition, energy balance, immunity, cancer, and healthy aging (72-75) One mechanism is a brain-fat axis—the hypothalamic-sympathoneural-adipocyte axis. The complex stimuli provided by EE induces BDNF in the hypothalamus and thereby elevates the sympathetic tone to the adipose tissue. The resulting WAT remodeling, including induction of beige cells and the suppression of leptin, leads to leanness and anticancer phenotype. Furthermore, EE finetunes immunity based on context, enhancing natural killer cell and T cell immunity in cancer models (43) while suppressing autoimmunity in the multiple sclerosis model (46), which is mediated also by the hypothalamic BDNF. A recent meta-analysis reports that SE increases morbidity and mortality in research rodents as compared to EE (76), supporting the effort of developing EE mimetics for pharmacological use (77, 78).

Targeting hypothalamic BDNF was our first attempt to develop EE mimetics. Our prior studies found Bdnf in the ventromedial hypothalamus (VMH) and arcuate nucleus (ARC) was an immediate early gene responding to EE. Knockdown Bdnf expression in VMH/ARC completely abolished EE-induced anticancer, antiobesity, and regulation of natural killer and T cell immunity. Injection of AAV-BDNF to VMH/ARC reproduced all these effects that were examined (37, 40, 43, 46). As such, enhancing VMH/ARC BDNF signaling would be the most effective mimetic of EE. To improve the safety of the BDNF gene therapy, we developed an autoregulatory system to control therapeutic gene expression mimicking the body's natural feedback systems (34). Using this system, we demonstrated the long-term therapeutic efficacy and safety in the MC4R deficient mice model of the most common monogenic human obesity (33) and in the Magel2-null mouse model PWS (36). Although hypothalamic AAV-BDNF gene therapy might be viable for rare genetic forms of obesity such as PWS, its potential for clinical translation remains less clear given the difficulty and risk of delivering AAV-based gene therapy to the hypothalamus. Thus, in the present study, we investigated whether a biobehavioral intervention such as EE could mimic the therapeutic effects of hypothalamic BDNF gene therapy (36).

EE initiated at older age (>20 weeks) after the occurrence of metabolic defects (61) was highly effective to normalize the metabolic abnormalities including excessive adiposity, reduced lean mass, and impaired glycemic control in the Magel2-null mice (Fig. 1, Fig. 2, Fig.4A). These metabolic improvements are comparable to the metabolic effects via the hypothalamic AAV-BDNF gene therapy in male Magel2-null mice (36). One distinction was the average food intake. Magel2-null mice do not recapitulate human hyperphagia when fed normal chow diet (47, 52). Consistent with the literature, food intake was not different in Magel2-null SE mice compared to WT counterparts. EE significantly increased the food intake while body weight and adiposity were reduced, indicating elevated energy expenditure mediating the correction of body weight and adiposity (Fig. 1D). Previous studies have shown EE increased food intake while it decreased adiposity in WT C57BL/6 mice at varying ages and in the BTBR mouse model of autism (39, 40, 42) when fed normal chow diet. The food intake increase observed in Magel2-null EE mice could reflect a compensatory response to weight loss and/or fat loss. In contrast, hypothalamic BDNF gene therapy resulted in a decrease in food intake in both male and female Magel2-null mice (36). Of note, the effects on food intake were influenced by diet and genetic background. Both EE and hypothalamic AAV-BDNF injection ameliorated obesity in the high fat diet-induced obesity models without changes in food intake (40). Hypothalamic AAV-BDNF injection increased food intake in WT mice (40), while it rescued hyperphagia in the MC4R-deficient mice (33) being fed a normal diet. Nevertheless, the discrepancy between EE and hypothalamic BDNF overexpression regarding food intake indicates mechanisms in addition to BDNF could be involved in EE-induced metabolic adaptions in the absence of Magel2.

Hyperleptinemia is a hallmark of the Magel2-null model, which was corrected by EE in the male Magel2-null mice (Fig. 4B). The normalization of hyperleptinemia was observed in female Magel2-null mice treated with AAV-BDNF, although data in male mice were incomplete (36). Sex-dependent changes in the serum level of adiponectin were observed: there was no difference in the female mice (36) but a genotype-driven increase in the male mice, which was reversed by EE (Fig. 4C). Magel2-null mice are reported to manifest abnormalities of hypothalamic pituitary adrenal (HPA) axis including elevated basal corticosterone levels in both sexes and blunted response to insulin and dexamethasone only in female mice (61). In the present study, an elevated basal level of corticosterone was observed in the male Magel2-null SE mice, which was normalized by EE (Fig. 4D). Notably, our previous work demonstrates an EE-induced increase in the basal level of corticosterone in WT mice of normal weight, and the modulation of the HPA axis plays critical roles in EE-induced immunoregulations (37, 43, 46). Thus, the impact of EE on the HPA axis is likely context dependent. Immune profiling of the Magel2-null mice remains unknown. Future research will fill in this gap and further explore the stress responses and immune outcomes induced by EE in the Magel2-null model of PWS.

Consistent with previous findings in WT mice (40), EE induced more gene expression changes in the visceral fat depot rWAT than that in the subcutaneous fat depot iWAT (Fig. 5). In the rWAT, genotype-driven gene expression changes were reversed by EE. Adrb3, conveying the sympathetic nervous system signaling, was downregulated in Magel2-null mice concurrent with downregulation of Hsl and upregulation of Lep, both downstream targets of β-adrenergic signaling. Cfd encoding complement factor-D, also known as adipsin, was significantly downregulated in both rWAT and iWAT of Magel2-null mice compared to WT mice (Fig. 5). Adipsin is an adipokine predominantly produced in adipose tissue and plays a critical role in the activation of the complement system and thereby is involved in metabolic and immune regulations. Adipsin deficiency is rare in humans and is associated with a heightened risk for infection (79, 80). It is reported that adults with PWS exhibit more low-grade systemic inflammation associated with a higher plasma level of complement component C3 compared to matched non-PWS obese subjects (81). However, data on adipsin remains unavailable. In the Magel2-null mice, EE normalized the adipsin expression in both rWAT and iWAT, whose functional impact is worthy of further investigation.

One unexpected finding of the study was the upregulations of hypothalamic Bdnf and its receptors mRNAs in male Magel2-null mice compared to male WT mice (Fig. 6A). However, the protein level of TrkB was not consistent with the mRNA levels (Fig. 6B). The implication of this genotypic alteration remains unclear. Moreover, in our transcriptomic profiling of the hypothalamus of females, neither Bdnf nor TrkB was among the 567 differentially expressed genes identified by comparing female Magel2-null to WT counterparts (64). qRT-PCR analysis validated the mRNA sequencing results (64). Long-term gene therapy study also found no genotypic effect of Bdnf mRNA in female Magel2-null mice (36), suggesting genotypic alteration of Bdnf expression is likely sex dependent. It is worthy to note that Magel2-null mice are reported to be insensitive to the anorexic effect of peripherally administered leptin but are hypersensitive to the anorexigenic effects of the MC4R agonist melanotan II, suggesting MC4R agonist can bypass the leptin insensitivity in mice lacking Magel2 (14, 15). As BDNF acts downstream of MC4R (21), overexpressing BDNF is highly effective in this model of PWS even though a genotypic upregulation of Bdnf was observed in the male Magel2-null mice. Admittedly, although therapeutic effects are evident, the mechanisms of action, either hypothalamic BDNF gene therapy or EE, remain unclear and require further investigation.

Another unexpected finding of the study is that EE reversed the genotype-associated upregulation of hypothalamic Bdnf expression (Fig. 6A). Previous studies have consistently demonstrated that EE induced upregulation of Bdnf in the hypothalamus of both male and female WT C57BL/6 mice (37, 39, 40, 43, 45, 46). Of interest, EE stimulated hypothalamic Bdnf expression in the male but not the female BTBR mice, which was associated with fewer metabolic responses to EE in female BTBR mice (42). The downregulation of hypothalamic Bdnf expression induced by EE in male Magel2-null mice may indicate mechanisms other than hypothalamic Bdnf that mediate the EE benefits. In fact, Magel2-null mice can serve as a unique model to elucidate the unknown central pathways in the context of EE, as well as provide an understanding of the intertwined regulatory network among leptin, MC4R, and BDNF.

Regarding additional hypothalamic changes, EE was unable to correct the genotype-driven downregulation of Crh in the hypothalamus (Fig. 6A), although the circulating corticosterone level was normalized by EE (Fig. 4D). Inconsistency between mRNA levels and protein levels was observed in TrkB receptors and Iba1, which might be partially explained by the temporal difference in response to fast. We previously published that AAV-BDNF gene therapy reserved a hypothalamic neuroinflammatory signature in the Magel2-null mice of both sexes (64). This gene expression signature was not reproduced in the present study with fewer genotype-driven gene expression changes in mice not being subjected to stereotaxic surgery (Fig. 6A). This raises the possibility that stereotaxic surgery to the hypothalamus could contribute to the robust upregulations of microglial and proinflammatory markers in the Magel2-null mice receiving the control AAV. In other words, Magel2-null mice might show increased glial scar in response to surgery compared to WT mice receiving the same control AAV vector, which was alleviated by overexpression of BDNF (64).

In addition to an improvement of metabolic functions, hypothalamic AAV-BDNF gene therapy led to behavioral amelioration in female Magel2-null mice including recognition memory and depression-like behavior (36). In the present study, no difference in cognition, social affiliation, or anxiety-like behavior was observed in Magel2-null EE mice compared to Magel2-null SE mice (Fig. 3). It is worthy to note that genotype-driven behavioral alterations were not found in the male mice of the present study as opposed to the female mice of the gene therapy study (36). Whether sex dimorphism causes this discrepancy requires further investigation. Another caveat is the larger age range in the present study: 20 to 40 weeks of age at the start of the experiment due to a difficulty of accumulating enough Magel2-null mice. On another note, strain and genetic background may contribute to varying behavioral outcomes. We previously reported that EE resulted in decreased anxiety-like behavior in the OFT and increased social affiliation in the three chamber sociability test only in the male BTBR mice (42). These EE-induced behavioral alterations were not observed in Magel2-null mice. Consistent with the literature (41, 82), Magel2-null EE mice exhibited significantly less locomotion in the OFT as compared to SE mice regardless of genotype (Fig. 3A), which may reflect an enhanced habituation derived from greater experience dealing with a changing environment. The present study limited behavioral assessments to those that were examined in the AAV-BDNF gene therapy study. As a wide range of behavioral effects induced by EE have been reported in normal animals and various disease models, expanding the behavioral assessments in the Magel2-null mice is worthy of consideration.

The present study demonstrated that EE initiated after the occurrence of metabolic dysfunction was sufficient to normalize the metabolic defects in the Magel2-null mice, and the metabolic improvements by EE were comparable to those achieved by hypothalamic BDNF gene therapy. However, several distinctive features were observed, suggesting additional mechanisms. Recently it has been reported that EE prevented gut dysbiosis progression in the diet-induced obesity model, thereby enhancing intestinal barrier integrity, attenuating colon and adipose inflammation, reducing hepatosteatosis, and normalizing glucose metabolism (83). It remains to be seen whether central mechanisms other than the hypothalamic BDNF and/or peripheral mechanisms contribute to the EE-induced improvement of metabolic health in the Magel2-null model of PWS. In addition, future study is needed to assess whether initiating EE at an earlier age can prevent the manifestation of metabolic dysfunctions in the Magel2-null mice.

Whether physical activity accounts for EE phenotypes is one of the most frequently asked questions. Our previous research has demonstrated that the EE-induced anticancer and antiobesity phenotypes could not be accounted for by physical activity alone (37, 40, 45). Moreover, in the aging study, voluntary wheel running preferentially affected muscle while EE regulated adipose tissue and liver (41). EE and voluntary wheel running induced distinct gene expression signatures in the hypothalamus, in the adipose tissue, and in other peripheral tissues. These data obtained in WT mice collectively indicate that physical exercise is a major component of EE, but it alone is insufficient to account for the EE phenotypes. Even though some phenotypes are shared between physical exercise and EE such as reduced adiposity and increased muscle mass, the underlying mechanisms can be distinct. However, current study in Magel2-null mice designed to assess the “therapeutic” effects is unable to address the role of elevated exercise in EE-induced phenotypic changes.

Another limitation lies with the animal modeling of PWS. Given the high lethality in murine models with large deletions at the PWS locus, single-gene knockout models targeting various PWS genes are commonly used in preclinical studies including Magel2, Snord116, and Ndn (16, 84). These preclinical models recapitulate some, but not all, PWS phenotypes (16, 84). One major limitation of the model used in the present study is that Magel2-null mice do not mirror the hallmark symptom of human PWS—hyperphagia (47, 52)—despite broad systemic metabolic and endocrine dysfunction. Future studies will be needed to assess whether EE is effective in other models of PWS.

EE is a laboratory condition that recapitulates some aspects of an active lifestyle and therefore can facilitate mechanistic studies (85). Physical and behavioral therapies are fundamental components of the management of PWS symptoms (86-88) As such, studies on EE in PWS-relevant animal models may elucidate fundamental mechanisms that are likely conserved in humans and identify biomarkers to assess the therapeutic efficacy of these biobehavioral interventions. On the other hand, translating EE-mediated therapy originated from animal models has been attempted in clinical trials in autism spectrum disorder (89-91) and Rett syndrome (92). These randomized clinical trials have shown that biobehavioral intervention informed by EE research is capable of ameliorating symptoms of children and adults with autism spectrum disorder (91) and improving the symptoms of children with Rett syndrome (92). This study provides preclinical data of EE in a mouse model of PWS that may facilitate translational efforts in individuals with PWS.

Funding

This work was supported by a grant from the Foundation for Prader-Willi Research, National Institutes of Health grant CA166590, and internal funding from The Ohio State University Comprehensive Cancer Center-Arthur G. James Cancer Hospital and Richard J Solove Research Institute.

Author Contributions

Conceptualization, N.J.Q. and L.C.; Methodology, N.J.Q., W.H., and L.C.; Validation, N.J.Q. and L.C.; Investigation, N.J.Q. X.Z., W.H., T.M., and L.C.; Data Curation, N.J.Q. and L.C.; Writing, L.C.; Funding acquisition, L.C.

Disclosures

All authors declare no conflicts of interest.

Data Availability

Original data generated and analyzed during this study are included in this published article.