Overweight in Mice and Enhanced Adipogenesis In Vitro Are Associated With Lack of the Hedgehog Coreceptor Boc

The prevalence of obesity and associated metabolic pathologies has attracted attention to the identification of etiological factors (1,2). Obesity arises from genetic, environmental, and behavioral factors that influence energy balance. The principal feature of obesity is excessive accumulation of white adipose tissue (WAT) (1,3–5). Although obesity is understood to be a centrally regulated consequence of overnutrition and/or reduced expenditure of energy, the processes that regulate WAT expansion at the level of the adipocyte are not well characterized.

WAT expansion occurs through adipocyte hypertrophy and hyperplasia, so such processes presumably engage at some level the process of adipogenesis (3,4). Adipogenesis has been analyzed with preadipocyte cell lines such as 3T3-L1 and various mutant mouse lines (6). These studies elucidated a transcriptional network centered around PPARγ and C/EBP family members, which drive differentiation of preadipocytes into lipid-accumulating adipocytes (6). Recent studies have identified cells in the stromal vascular fraction (SVF) of WAT that are bona fide adipocyte progenitor cells (7,8). Such cells in WAT must respond to hormonal and local cues that regulate homeostatic maintenance of this tissue. The cues that act on adipocyte progenitor cells are not well understood, but among those implicated is the Hedgehog (HH) signaling pathway.

HH proteins regulate developmental events in organisms as diverse as insects and mammals. Among mammalian HH proteins, Sonic Hedgehog (SHH) plays the broadest role and is involved in the growth and/or morphogenesis of many body structures (9). HH proteins activate a conserved signal transduction pathway (10–12). In the absence of ligand, the primary HH receptor PTCH1 functions to inhibit signaling by a second membrane protein, SMO. Binding of HH to PTCH1 relieves inhibition of SMO, and SMO signals to activate pathway target genes via GLI transcription factors. Among such target genes are Gli1 and Ptch1 themselves, and they are often used as readouts of pathway activity (10–12).

CDO (also called CDON), BOC, and GAS1 are cell surface proteins that promote HH pathway activity as ligand-binding coreceptors with PTCH1 (13–20). CDO and BOC are related transmembrane proteins (21,22), whereas GAS1 is a GPI-anchored protein unrelated to CDO and BOC (23). Analysis of mice with targeted mutations in Cdon, Boc, and Gas1 revealed that although none is essential for HH pathway activity, they are collectively required for pathway function in the early mouse embryo (13,14,16,19). A ternary complex of HH ligand, PTCH1, and at least one of these coreceptors appears to be required for successful signal transduction (15,16,24). Boc-null mice, the subject of this study, are viable and fertile but display defects in SHH-dependent neural patterning and axon guidance (19,25–28).

The HH pathway negatively regulates adipogenesis. Activation of the HH pathway with either SHH or the SMO agonist purmorphamine inhibited adipogenesis of 3T3-L1 and additional cell lines (29–32). Furthermore, blockade of HH signaling, with a dominant-negativeform of GLI2 or the SMO antagonist cyclopamine, enhanced adipogenesis of these cells in response to inducing factors (31,32). HH signals blocked early steps of adipogenesis, upstream of the expression of PPARγ (29,31,32). These studies suggest that the HH pathway functions to block differentiation of preadipocytes in vitro. Experiments in mice are consistent with this notion. Adult mice homozygous for a hypomorphic allele of Ptch1 showed reduced WAT mass, which had lower levels of adipose markers and elevated levels of HH target genes (33). Adipose tissue–specific mutation of another HH pathway inhibitor, Sufu, led to a loss of WAT (34). Finally, mice with diet-induced or genetic (ob/ob) obesity displayed decreased expression of HH pathway components (31).

Although mice with a genetic gain of function in HH pathway activity showed loss of WAT (33,34), it has not been demonstrated that mice with a loss of HH activity have enhanced adiposity. We report that mice with a germline mutation in Boc display age-dependent overweight due to an increase in WAT. These mice also show an exaggerated response to a high-fat diet (HFD), and Boc^−/− embryo fibroblasts differentiate into adipocytes more efficiently than wild-type cells. These results reveal that BOC, presumably acting as an SHH coreceptor, is required for maintenance of normal weight in vivo and appropriate regulation of adipogenic differentiation in vitro.

Boc^AP-1 mice (19) were backcrossed onto a C57BL/6N background for at least six generations. For age-related effects, male Boc^+/+ or Boc^−/− mice backcrossed for six generations were used and weighed every 4 weeks over 32 weeks. To assess food intake, food was weighed and replenished every other day for individually caged mice for 2 weeks. For HFD-induced effects, male Boc^+/+ or Boc^−/− mice backcrossed for 10 generations were used. Four-week-old mice were fed with HFD (60% fat content; Harlan) and weighed weekly over 8 weeks. For blood glucose levels, tail vein blood was used after animals were fasted for 16 h, with free access to water. For the glucose tolerance test (GTT), mice were fasted for 16 h and injected intraperitoneally with 1.5 g 20% d-glucose/kg body weight (Sigma-Aldrich), followed by measurement of blood glucose levels. For the insulin tolerance test (ITT), mice were fasted for 6 h and injected intraperitoneally with 1 IU insulin/kg body weight (Sigma-Aldrich), followed by measurement of glucose levels. Mouse tissues were immediately frozen in liquid nitrogen and stored at −70°C or fixed in 4% paraformaldehyde (PFA). Plasma insulin was measured with the Mouse Insulin ELISA Kit (U-type; Shibayagi Corp.), and serum triglycerides and nonesterified fatty acids were measured by colorimetric assay kits (Wako).

For assessment of metabolic parameters (Fig. 2), 5-month-old Boc^+/+ and Boc^−/− mice were analyzed with metabolic cages (Harvard Apparatus; Panlab). Measurements were performed for 48 h, during which animals had access to food and water. Core body temperature was measured with an HB-101 homeothermic blanket/pad equipment (Harvard Apparatus).

For hematoxylin-eosin (H-E) staining, WAT and liver tissues were fixed and processed for 20-μm paraffin sections. For Oil Red O staining, livers were fixed in 4% PFA and dehydrated by sucrose gradient followed by an optimal cutting temperature compound-cryosection (10 μm) and Oil Red O staining. For Oil Red O staining of differentiated mouse embryonic fibroblasts (MEFs) and 3T3-L1 cells, cultures were fixed in 10% formaldehyde for 10 min and stained with Oil Red O. Stained plates were treated with 1 mL isopropanol/4% NP-40 for 10 min by a gentle shaking, and extracted Oil Red O was quantified by measuring optical density at 520 nm. For placental alkaline phosphatase (PLAP) staining, tissues were fixed with 2% PFA plus 0.2% glutaraldehyde for 2 h at 4°C, washed in PBS, permeabilized in 0.3% Triton overnight at 4°C, washed in saline, heat inactivated for 30 min at 72°C, washed in NTM (100 mmol/L NaCl, 100 mmol/L Tris-HCl, pH 9.5, 50 mmol/L MgCl₂) buffer, and stained with BM purple (Roche).

Preparation of the SVF from WAT was performed as previously described (8). In brief, WAT was minced with scissors and digested with 2 mg/mL Collagenase I (Sigma-Aldrich) in DMEM at 37°C for 1 h. DMEM plus 10% FBS was added to double the volume, floating adipocytes were removed, and the digest was filtered through a 100-μm mesh that retained vessels of the stromal-particulate fraction. The filtrate was centrifuged at 800g for 5 min, and the SVF pellet was resuspended in DMEM containing 10% FBS and cultured for 2 days.

Isolation of primary MEFs was carried out as previously described (35). 3T3-L1 cells and MEFs were cultured in growth medium (DMEM plus 10% FBS) at subconfluence. For adipocyte differentiation, cells were grown to confluence and switched into differentiation medium I (growth medium plus 0.5 μmol/L IBMX, 1 μg/μL insulin, 0.25 μmol/L dexamethazone, 2 μmol/L rosiglitazone; Sigma-Aldrich) for 2 days. The culture medium was then changed to differentiation medium II (growth medium plus 2 μmol/L rosiglitazone), and the medium was changed every 2 days. To activate SHH signaling, MEFs were treated with 250 ng/mL SHH (R&D Systems) or 5.2 μmol/L purmorphamine (Calbiochem) in differentiation media from differentiation day 2 onward.

Immunoblotting was performed as previously described (15). Antibodies used are listed in Table 1.

Tissues were homogenized with FastPrep-24 (MP Biomedicals) and extracted with the RNA Extraction Kit (iNtRON). cDNA was generated with PrimeScript RT Reagent Kit (Takara) and amplified with nTaq polymerase (Enzynomics). Quantitative RT-PCR (qRT-PCR) was performed with SYBR Green (Takara) and analyzed with a TP8000 System (Takara). All data are normalized to the expression of ribosomal protein–encoding L32. Primers used in this study are listed in Table 2.

Microscopy was performed on a Nikon ECLIPSE TE2000-U with NIS-Elements F software (Nikon), Zeiss AxioPlan 2, and Zeiss AxioPlan 2 IE microscopes. Adipocytes were traced and their area measured using ImageJ software.

Values are means ± SEM or SD as noted. Statistical significance was calculated by paired or unpaired two-tailed Student t test; differences were considered significant at P < 0.05.

Body weights of Boc^+/+ and Boc^−/− mice fed normal chow were measured on a monthly basis from 4 to 32 weeks of age. Boc^−/− mice were significantly heavier than Boc^+/+ mice by 16 weeks of age and weighed 16% more than control animals at 32 weeks of age (Fig. 1A). Therefore, BOC-deficient mice displayed age-dependent overweight. To determine whether this phenotype was due to increased adiposity, the weight of WAT and other tissues from 32-week-old Boc^+/+ and Boc^−/− mice was analyzed. Boc^−/− mice had significantly more subcutaneous WAT than Boc^+/+ mice; mutants also had somewhat more epididymal (visceral) WAT, although this was not statistically significant (Fig. 1B). Weights of other tissues, including liver, spleen, kidney, heart, lungs, and diaphragm, were not different from those of Boc^+/+ mice (Fig. 1B and C). Boc^−/− mice exhibited a greater amount of brown adipose tissue (BAT) (Fig. 1C), which had a pale appearance with areas of obvious whitening (Fig. 1D) and enlarged lipid droplets (Fig. 1E). The mean area of WAT adipocytes was similar between Boc^+/+ and Boc^−/− mice; the mutants had more adipocytes of the largest size, but this difference was not statistically significant (Fig. 1F–H). Finally, Boc^+/+ and Boc^−/− animals consumed similar amounts of food per day (Fig. 1I). These data suggest that BOC-deficient mice became overweight from a modest but progressive increase in adiposity.

Since Boc^−/− mice were not hyperphagic, we assessed whether decreased energy expenditure contributed to their increased adiposity. Normal chow-fed 5-month-old Boc^+/+ and Boc^−/− mice (Fig. 2A) were used to assess whole-body metabolic rate. Boc^−/− mice exhibited slightly lower body temperature than Boc^+/+ mice (Fig. 2B). Spontaneous locomotive activity in Boc^−/− mice averaged about half that in control mice but this difference was not statistically significant, and there was no difference in rearing activity (Fig. 2C and D). Boc^−/− mice exhibited slightly decreased oxygen consumption and carbon dioxide production during the light phase (Fig. 2E and F). Respiratory quotient was slightly lower in Boc^−/− mice in both dark and light phases but was within normal range (Fig. 2G). Energy expenditure by Boc^−/− mice was also slightly decreased during the light phase (Fig. 2H), but total energy expenditure (dark plus light) was not different. These results suggest Boc^−/− mice have alterations of whole-body metabolism that may contribute to their overweight.

As BOC deficiency predisposed mice to greater adiposity, we challenged 6-week-old mice with HFD for a total of 8 weeks. Beginning at the third week, Boc^−/− mice gained more body weight than Boc^+/+ mice under both diets, but the difference between mutant and control mice was greater on HFD than normal chow (Fig. 3A and B). To determine whether enhanced weight gain in Boc^−/− mice was accompanied by abnormal glucose homeostasis, we measured steady-state blood glucose levels and performed tolerance tests for glucose and insulin after 7 weeks of HFD. Prior to challenge, blood glucose levels were similar between control and mutant mice (Fig. 3C). Under glucose challenge, a spike in blood glucose levels was observed at 15 min, followed by a slower further increase and progressive return to near baseline in both Boc^+/+ and Boc^−/− mice. However, Boc^−/− mice had higher blood glucose levels at the 15- and 30-min time points than Boc^+/+ mice (Fig. 3D). In the ITT, prior to insulin injection, Boc^−/− mice had slightly, but not statistically significantly, lower blood insulin levels than Boc^+/+ mice (Fig. 3F). Serum glucose levels decreased sharply 15–45 min after insulin injection in both Boc^+/+ and Boc^−/− mice, but Boc^−/− mice displayed significantly less sensitivity to insulin (Fig. 3E). Therefore, Boc^−/− mice on HFD had perturbations in glucose homeostasis. Furthermore, when on HFD, mutant animals also showed elevated plasma levels of triglycerides and nonesterified fatty acids as compared with Boc^+/+ mice (Fig. 3G and H).

We next assessed livers and WAT of Boc^+/+ and Boc^−/− mice after 8 weeks of HFD. Boc^−/− livers displayed more and larger lipid droplets than Boc^+/+ livers upon Oil Red O staining (Fig. 4A,). Hepatic lipid accumulation is often associated with enhanced expression of genes involved in lipid metabolism (e.g., Acc, Fas, Scd1, and aP2) and that promote adipogenesis (e.g., Pgc1α, Cebpα, Srebp1c, and Pparγ) (6). We therefore assessed by qRT-PCR and Western blot analyses the expression of these genes in livers of Boc^+/+ and Boc^−/− mice on HFD. Expression of all these genes was enhanced in Boc^−/− livers relative to controls (Fig. 4C–E). H-E staining of WAT revealed that adipocytes from Boc^−/− mice on HFD were larger than those of Boc^+/+ mice on this diet (Fig. 4A and B). We then assessed the expression of lipid metabolism, adipogenic, and adipokine-encoding genes (Ccl5 and TNFα) in WAT from Boc^+/+ and Boc^−/− mice on HFD. Expression of all these genes was significantly higher in Boc^−/− WAT than Boc^+/+ WAT (Fig. 4F and G). Finally, expression of uncoupling protein (Ucp) 1, 2, and 3 genes was assessed by qRT-PCR in BAT. Ucp1 and Ucp3 expression was significantly reduced in Boc^−/− BAT (Fig. 4H). Taken together, it is concluded that Boc^−/− mice showed an exaggerated response when challenged with HFD.

To gain information on potential sites of action relevant to the phenotypes of overweight and HFD response, Boc expression was examined in adult tissues by qRT-PCR. Highest levels of Boc expression were seen in brain, lung, and WAT; expression in liver was barely detectable (Fig. 5A). The mutant Boc allele contains a PLAP reporter gene, which faithfully reflects endogenous Boc expression (19,36). The PLAP reporter was expressed at high levels in WAT from Boc^+/PLAP mice (Fig. 5B). To investigate which cell types in WAT express Boc, mature adipocytes were separated from the SVF, which contains blood vessels and preadipocytes (7,8). The SVF was plated to separate nonadherent blood vessels from preadipocytes, which attached to the dish. PLAP activity was seen in floating adipocytes and strongly in mural-like cells that lined vessels and that resemble adipocyte precursors in their location, appearance, and frequency (8). Upon brief culture of the adherent SVF fraction, small PLAP-positive colonies formed with a frequency and morphology similar to that of previously described adipocyte progenitors (Fig. 5B). Consistent with this, Boc was expressed in total WAT and the SVF but much less in mature adipocytes (Fig. 5C). As expected, Pparγ expression was detected at the highest levels in WAT and adipocytes, whereas the SVF expressed lower levels (Fig. 5C). These data suggest that Boc is expressed in adipocyte progenitors.

BOC expression was next analyzed by Western blotting during differentiation of 3T3-L1 preadipocytes. BOC levels were relatively low during growth conditions and strongly induced when cells reached confluence, just prior to switching cultures to adipogenic induction medium (Fig. 6A). BOC levels were partially diminished when the cells were induced to differentiate, prior even to expression of the adipogenic regulators PPARγ and C/EBPα, but were restored to their peak during the later stages of differentiation (Fig. 6A). GAS1, another SHH coreceptor, was expressed with a pattern similar to that of BOC, although the effect of cell confluence was not observed (Fig. 6A). In contrast to BOC and GAS1, the BOC paralog CDO was barely detected in 3T3-L1 cells. CDO was expressed robustly by C2C12 myoblasts, whereas BOC levels were low in this cell line, relative to 3T3-L1 cells (Fig. 6A).

To investigate BOC’s role in adipogenesis, we isolated MEFs from E13.5 Boc^+/+ and Boc^−/− embryos and cultured them in adipogenic induction medium for 15 days (D15). Boc^−/− MEFs had more Oil Red O–positive colonies at D15 than did Boc^+/+ MEFs (Fig. 6B and C). The timing of induction of lipogenic genes was similar between Boc^+/+ and Boc^−/− MEFs, but expression of each was higher in the Boc^−/− MEFs, as analyzed by semiquantitative RT-PCR (Fig. 6D). Additionally, levels of aP2, FAS, and C/EBPα proteins were elevated in Boc^−/− MEFs at D3 and D6, relative to control MEFs (Fig. 6E). Therefore, similar to results in Boc^−/− mice, Boc deficiency resulted in enhanced adipogenesis of MEFs in vitro.

SHH signaling suppresses adipogenesis, and BOC promotes SHH signaling as a coreceptor (16,18,20). The effects of BOC deficiency on adipogenesis may therefore be related to effects on SHH signaling. We initially addressed this question by analyzing adult Boc^+/+ and Boc^−/− WAT for mRNA levels of various SHH signaling components by qRT-PCR. We examined expression of Cdo, Gas1, Ptch1, and Gli1; these genes not only encode components of the SHH pathway but their expression is also regulated by SHH pathway activity (Ptch1 and Gli1 expression is induced; expression of Cdo and Gas1 is repressed, at least in early mouse embryos [11,18]). Expression of Cdo, Gas1, and Ptch1 was significantly lower in Boc^−/− WAT, compared with the Boc^−/− WAT, whereas Gli1 expression was similar (Fig. 7A). These results do not immediately suggest a major alteration of SHH signaling in Boc^−/− WAT, but it may be simplistic to assume that expression of these genes is only regulated by SHH activity in a complex tissue that is also perturbed by loss of BOC.

To address this question further, expression of these genes in Boc^+/+ and Boc^−/− MEFs during adipogenic differentiation was analyzed by semiquantitative RT-PCR. Expression of Gli1, Ptch1, and Boc was enhanced at day 2 of adipogenic differentiation in Boc^+/+ cells (Fig. 7B). Induction of Gli1 and Ptch1 was diminished in Boc^−/− MEFs and remained lower than in control cells throughout the differentiation time course (Fig. 7B). Western blot analyses revealed a similar pattern for GLI1 and PTCH1 protein levels (Fig. 7C). As Gli1 and Ptch1 are both direct SHH pathway target genes, these data suggest that BOC deficiency impairs SHH signaling activation during adipogenesis. Interestingly, Cdo was not expressed in Boc^+/+ cells but was abnormally induced in Boc^−/− MEFs at D6 (Fig. 7B); however, CDO protein levels were too low for detection by Western blot (not shown). Gas1 expression was not altered through the adipogenic differentiation time course and was similar in Boc^+/+ and Boc^−/− MEFs (Fig. 7B and C).

BOC’s major function in SHH signaling is as a coreceptor with PTCH1 (16). It would therefore be predicted that Boc^−/− MEFs would remain sensitive to inhibition of adipogenic differentiation by activation of the SHH pathway at a point downstream of ligand reception. To test this notion, Boc^+/+ and Boc^−/− MEFs were induced to differentiate in the presence of the SMO agonist purmorphamine, followed by Oil Red O staining at D12. Purmorphamine treatment resulted in strong reduction of adipocyte differentiation in both Boc^+/+ and Boc^−/− MEFs (Fig. 7D and E). Expression of aP2, Fas, and Scd1 was also analyzed in purmorphamine-treated Boc^+/+ and Boc^−/− MEFs at D6 by qRT-PCR. In agreement with the Oil Red O staining results in Fig. 7C and D, aP2, Fas, and Scd1 expression was higher in vehicle-treated Boc^−/− MEFs than vehicle-treated Boc^+/+ MEFs but was decreased by purmorphamine treatment in both Boc^+/+ and Boc^−/− MEFs (Fig. 7F). Western blot analyses revealed a similar pattern of aP2, FAS, and C/EBPα protein levels (Fig. 7G). These data indicate that activation of HH signaling downstream of BOC coreceptor function is still sufficient to inhibit adipogenesis. BOC plays redundant roles with CDO and GAS1 as SHH coreceptors (13,16,19). Our findings here suggest that BOC plays a partially rate-limiting role in SHH signaling during adipogenic differentiation of MEFs. To assess whether Boc^−/− MEFs were still responsive to the SHH ligand itself, Boc^+/+ and Boc^−/− MEFs were induced to differentiate into adipocytes in the presence of exogenous, recombinant SHH for 12 days and analyzed by Oil Red O staining and qRT-PCR for adipogenic markers. Adipogenic differentiation of both Boc^+/+ and Boc^−/− MEFs was inhibited to a similar degree (Fig. 7H–J). These results suggest that although loss of BOC may diminish endogenous SHH signaling during adipogenesis, expression of other coreceptors is likely sufficient to override the lack of BOC in Boc^−/− MEFs and confer responsiveness to exogenous SHH.

Genetic factors that contribute to obesity-related WAT accumulation are poorly understood. The HH pathway plays a conserved role in adipogenesis, inhibiting fat formation (29–32). Mice with elevated HH signaling globally due to a hypomorphic Ptch1 mutation, or in the adipocyte lineage via conditional targeted mutagenesis of Sufu, have sharply diminished WAT (33,34). Although these studies indicate that a genetic gain of function in HH pathway activity blocks adipogenesis, the converse has not been shown, i.e., that reduced HH pathway function results in WAT accumulation.

We report that mice lacking the HH coreceptor BOC displayed age-related overweight, with an increase in WAT, but not in the weight of internal organs. Furthermore, they had an exacerbated response to HFD, including enhanced weight gain and adipocyte hypertrophy, livers with greater fat accumulation, and elevated expression of genes related to adipogenesis, lipid metabolism, and adipokine production. Boc^−/− MEFs showed enhanced adipogenesis and had reduced expression of the direct HH pathway target genes, Gli1 and Ptch1, during adipogenic differentiation. Therefore, loss of BOC, and an associated decrease in HH signaling, in WAT precursor cells may underlie the age-related overweight and enhanced response to HFD seen in Boc^−/− mice. Consistent with this possibility, Boc is prominently expressed in WAT, particularly in cells of the SVF that may function as adipocyte precursors. However, adipogenesis per se is not thought to be the prime driver of obesity (4,5), so it is not clear that moderate loss of HH function would be sufficient to override regulatory mechanisms of energy balance to drive enhanced adiposity. In fact, a role for loss of BOC outside the adipose lineage may be a major contributor to the enhanced adiposity of Boc^−/− mice, as these animals had alterations in some metabolic parameters, including decreased body temperature and locomotive activity. Boc is expressed in the developing and adult central nervous system, albeit at low levels in the hypothalamus, a key structure for energy balance regulation (36). Construction of a Boc conditional mutant mouse line will be required to address this point. It is worth noting that the global Ptch1 and adipose-specific Sufu mutants did not display substantial alterations in metabolic parameters or in the GTT (33,34); however, these animals had reduced WAT, rather than overweight or obesity.

The effects of Boc mutation are relatively modest. Boc^−/− mice became overweight by 4 months of age but were not obese. They showed elevated levels of serum triglycerides and nonesterified fatty acids, and defects in GTT and ITT, when on the HFD, but their fasted blood glucose and insulin levels were similar to those of control mice. This may reflect redundancy of BOC with GAS1 and CDO, additional HH coreceptors. These coreceptors have overlapping functions and are collectively required for HH signaling in the mouse embryo (13,16,19). Gas1 was expressed in 3T3-L1 cells and MEFs, whereas Cdo was expressed at low levels in these cells. Furthermore, Cdo mutant mice do not have an overweight phenotype (unpublished data). GAS1 may therefore function in the absence of BOC to provide a level of HH coreceptor activity that limits the effects of loss of BOC to overweight rather than full-blown obesity. As expected, activation of the HH pathway downstream of BOC with the SMO agonist purmorphamine inhibited adipogenesis of both Boc^+/+ and Boc^−/− MEFs. However, adipogenesis of both cell types was also inhibited by recombinant SHH. We hypothesize that the presence of GAS1, and perhaps CDO, was sufficient to allow a strong response to exogenous SHH, but that loss of BOC was enough to cause a diminished response to endogenously produced HH ligand. It seems likely, therefore, that genetic removal of Gas1, and perhaps Cdo, from Boc^−/− mice would result in further loss of HH coreceptor activity and worsen the overweight phenotype; this is what is seen in neural tube patterning (13). Testing this notion will also require a conditional mutagenesis approach, as double mutants for any of these three factors have severe phenotypes and are not viable (13,19).

Genetic variants in the HH pathway may play a role in WAT accumulation in human obesity. Loss-of-function mutations in human genes encoding several HH pathway components, including CDO and GAS1, are associated with holoprosencephaly, a common and often devastating developmental defect of the forebrain (15,37,38). Mice with mutations in these genes are good models for holoprosencephaly (39). However, Boc^−/− mice do not have holoprosencephaly and are viable (19). Therefore, it is possible that “weak” BOC alleles may exist in the human population and contribute to WAT accumulation in individuals with additional genetic or lifestyle-based predisposition to obesity. It is interesting that the effects of Boc mutation were seen with age and diet, two variables involved with weight gain. Genome-wide association studies have not linked BOC to obesity, although that is not surprising given that Boc^−/− mice are overweight, not obese. Many variant BOC alleles have been documented in the 1000 Genomes data, including many with likely deleterious changes. Some of these could act as modifier genes in individuals with obesity or metabolic disorders. In summary, the HH pathway is a conserved inhibitor of fat formation, and we report for the first time a loss-of-function mutation in the HH pathway associated with WAT accumulation and overweight.

Funding. This research was supported by National Institutes of Healthhttp://dx.doi.org/10.13039/100000002 grant AR-46207 (R.S.K.) and a National Research Foundation of Koreahttp://dx.doi.org/10.13039/501100003725> grant funded by the Korea Government (NRF-2011-0017315) (J.-S.K.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. H.-J.L., S.-B.J., A.I.R., H.-J.L., M.-J.K., and S.-H.K. contributed to the experimental design, research, data analysis, and review of the manuscript. R.S.K. and J.-S.K. contributed to experimental design, data analysis, and writing of the manuscript. R.S.K. and J.-S.K. are the guarantors of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. This work was presented as a poster at the International Conference of the Korean Society for Molecular and Cellular Biology, Seoul, Korea, 9–11 October 2013.