The homeodomain transcription factor Six3 regulates hypothalamic Pomc expression and its absence from POMC neurons induces hyperphagia and mild obesity in male mice

Abstract

Objective

Proopiomelanocortin (POMC) neurons release potent anorexigenic neuropeptides, which suppress food intake and enhance energy expenditure via melanocortin receptors. Although the importance of central melanocortin in physiological regulation is well established, the underlying genetic mechanisms that define the functional identity of melanocortin neurons and maintain hypothalamic Pomc expression remain to be fully determined. In this study, we investigate the functional significance of Six3, a transcriptional regulator notably expressed in embryonic and adult mouse POMC neurons, in the regulation of hypothalamic Pomc expression and downstream physiological consequences.

Methods

We first evaluated the expression of Six3 in the developing and adult hypothalamus by double fluorescence in situ hybridization. Next, we assessed POMC immunoreactivity in mutant mice selectively lacking Six3 from Pomc-expressing neurons and quantified Pomc mRNA levels in a tamoxifen-inducible Six3 knockout mouse model activated at embryonic E9.5 days. We also determined glucose and insulin sensitivity, daily food intake, body composition and body weight in adult male and female mice lacking Six3 specifically from POMC neurons. Lastly, we assessed the physiological consequences of ablating Six3 from POMC neurons in adult mice.

Results

Six3 and Pomc were co-expressed in mouse hypothalamic neurons during development and adulthood. Mouse embryos deficient in Six3 showed reduced Pomc expression in the developing hypothalamus. Targeted deletion of Six3 specifically from POMC neurons resulted in decreased hypothalamic Pomc expression, increased daily food intake, enhanced glucose sensitivity and mild obesity in male but not in female mice. Finally, conditional removal of Six3 from POMC neurons in adult mice led to a reduction in hypothalamic POMC immunoreactivity with no significant effects in body weight or food intake.

Conclusions

Altogether, our results demonstrate that Six3 plays an essential role in the early establishment of POMC neuron identity and the maintenance of physiological levels of hypothalamic Pomc expression. In addition, our study demonstrates that the functional significance of Six3 expression in POMC neurons is sexually dimorphic and age-dependent.

Highlights

• •Six3 is co-expressed with Pomc in the developing and adult hypothalamus.
• •Conditional deletion of Six3 during embryogenesis disrupts the onset of Pomc expression.
• •The developmental absence of Six3 in POMC neurons results in mild obesity and improved glucose tolerance in adult male, but not in female mice.
• •Deletion of Six3 in POMC neurons in adulthood reduces Pomc expression but does not induce significant physiological changes.

1. Introduction

The regulation of body weight and energy balance is governed by neural brain circuits which orchestrate food intake behavior by integrating multiple central and peripheral signals. The arcuate nucleus of the hypothalamus (ARC) plays crucial roles in the regulation of appetite, energy balance, and metabolism. Specifically, proopiomelanocortin (POMC) neurons release potent anorexigenic melanocortins α-, β-, and γ-melanocyte-stimulating hormones, which suppress food intake and increase energy expenditure through melanocortin receptors [1]. Mice lacking POMC specifically from ARC neurons develop early-onset hyperphagia and severe obesity, whereas reinstating Pomc expression in these mice normalized their food intake and adiposity [2]. Similarly, humans carrying null-allele mutations in POMC exhibit early-onset obesity and significant adrenal insufficiency [3], whereas Labrador Retriever dogs carrying deletion mutations in POMC often manifest morbid obesity [4]. Of note, null mutations in the melanocortin receptor 4 (MC4R) gene have been identified as the most common monogenic cause of obesity [5,6].

Since the initial characterization of POMC in humans and bovines [7], a series of studies have delved into elucidating the function of POMC neurons and the regulatory mechanisms governing Pomc expression. Developmental investigations have revealed that Pomc expression commences in the tuberal portion of the prospective mouse hypothalamus as early as embryonic day 10.5 (E10.5) [8]. Key transcription factors such as Mash1 (Ascl1) [9], Isl1 [10], Nkx2-1 [11], Shh [12], Ngn3 [13], and Prdm12 [14] play pivotal roles in establishing early POMC neuron identity and maintaining normal Pomc expression throughout postnatal life. More recently, the transcription factor Tbx3 has also been shown to be expressed in POMC neurons from late embryonic stages through adulthood, suggesting its crucial role in maintaining terminal differentiation rather than in establishing the early identity of these neurons [15]. Current progress in single-cell technology (scRNA-seq) is greatly advancing these investigations, allowing for the simultaneous identification of multiple transcripts within individual POMC neurons. In fact, a recent study from our laboratory traced the origin and maturation of arcuate POMC neurons in the developing and early postnatal hypothalamus using scRNA-seq transcriptomics of fluorescently labeled POMC cells at critical embryonic (E11.5, E13.5, E15.5, and E17.5) and early postnatal (P5 and P12) stages [16]. This study unveiled eight distinct POMC clusters based on the differential expression of several transcription factor and neuropeptide genes [16]. Notably, the Pomc^(high)/Prdm12 cluster, which represents canonical POMC neurons with the greatest abundance of Pomc transcripts, also showed high level expression of several transcription factor genes known to play critical roles in the early establishment of POMC neuron identity, such as Isl1 [16], Nkx2.1 [16], Prdm12 [16] and Tbx3 [16]. Another transcription factor that stood out in this major Pomc^(high)/Prdm12 cluster for being expressed at high levels at all developmental time points was Six3 [16]. Thus, the current investigation focused on elucidating the role of Six3 in the regulation of hypothalamic Pomc expression and its significance in central melanocortin function.

The homeodomain transcription factor Six3 (Sine oculis 3) was first identified in 1995 as one of the most anterior homeobox genes reported at that time [17]. Six3 expression commences as early as embryonic day 6.5 (E6.5) in the anterior neural plate, with increased signaling around E8.5. At this stage, Six3 expression localizes predominantly in the anterior neural ridge and its derivatives, including the hypothalamic anlage, optic vesicles, ventral forebrain, and neurohypophysis. In humans, SIX3 mutations contribute to ∼1.3% of holoprosencephaly (HPE) cases [18]. In the teleost fish medaka, overexpression of Six3 induces expanded and ectopic retinal primordia while its absence results in forebrain and eye developmental defects [19]. Consistent with observations in humans and fish, Six3-null mouse embryos exhibit perinatal lethality due to the absence of most craniofacial structures [20]. Furthermore, Six3 expression is implicated in pituitary morphology and function, being considered one of the causative genes for Combined Pituitary Hormone Deficiency [21]. Haploinsufficiency of Six3 leads to pituitary dysmorphology, and conditional knockout of Six3 in the hypothalamus of mice results in failed expansion of Rathke's pouch. Recent investigations have also supported a role for Six3 in metabolism, as evidenced by the deletion of Six3 from mature neurons resulting in dwarfism with improved glucose tolerance, increased lean mass, and elevated metabolic rates [22].

Given the pivotal role of Six3 in early hypothalamic development and its abundant mRNA presence in POMC neurons [15], we sought to elucidate the significance of Six3 in the initial establishment of POMC neuronal identity. Furthermore, we aimed to assess the physiological consequences elicited by the targeted deletion of Six3 specifically from POMC neurons.

2. Methods and materials

2.1. Animal care

All animal experiments followed the Public Health Service guidelines for the humane care and use of experimental animals and were conducted according to the Institutional Animal Care and Use Committee (PRO00010383) at the University of Michigan. The mice were housed in ventilated cages under a controlled temperature (∼23 °C) and photoperiod (12 h light/dark cycle, lights on from 6:00 am to 6:00 pm). Mice had free access to tap water and laboratory chow (5L0D; LabDiet). Breeding mice were fed with the breeder chow diet (5008, LabDiet). The strains of mice used in these studies were transgenic Pomc-TdTomato [23], compound Pomc-Cre.Six3^fl/fl, compound Pomc-CreERT.Six3^fl/fl, and compound CAAG-CreERT.Six3^fl/fl. The Six3^fl/fl mice were generated by Dr. Guillermo Oliver from Northwestern University [19]. BAC transgenic Pomc-CreERT mice were a gift from Dr. Joel Elmquist at UT Southwestern University [24]. Pomc-Cre [BAC Tg(Pomc1-cre)1^Lowl] (JAX strain # 005965) and CAAG-CreERT [Tg(CAG-cre/Esr1∗)5^Amc/J] (JAX strain # 004682) were obtained from The Jackson Laboratory. Mice were ear tagged and tailed around postnatal day 14 and weaned after 21 days.

2.2. High fat diet (HFD) experiment

After 17 weeks, Pomc-CreERT.Six3^fl/fl mice were transitioned to a rodent diet containing 60 kcal% fat (Research Diets, D12492), whereas Pomc-Cre.Six3^fl/fl mice were introduced to the same diet after reaching 4 weeks of age. High-fat diet (HFD) pellets were placed in a small food hopper to prevent spillage. Cages were cleaned at least once a week, or twice a week for mice prone to food wasting. Food hoppers were refilled twice a week, or three times for food-wasting mice. Prior to transitioning to the high-fat diet, mice were weighed weekly on a specific day of the week. Once on the high-fat diet, weekly body weight measurements continued until the mice underwent metabolic tests conducted by the Animal Phenotyping Core at the University of Michigan just prior to euthanasia.

2.3. Tamoxifen injection

Tamoxifen or control vehicle was administered to adult Pomc-CreERT.Six3^fl/fl mice at 9 weeks of age for 5 consecutive days, at a dosage of 50 mg/kg (i.p.) (Sigma–Aldrich, T5648). Metabolic parameters, including body weight, were assessed the week following tamoxifen administration. In pregnancy experiments, breeding females were monitored daily for the presence of vaginal copulation plugs, marking the day of plug detection as embryonic day 0.5 (E0.5). Pregnant CAAG-CreERT females received a single dose of 150 mg/kg tamoxifen (i.p.) when embryos were E9.5 days old. Embryos were harvested for quantitative RT-PCR analysis at E11.5 and for histology at E12.5. Tamoxifen was prepared using a solution of 10% ethanol and 90% sesame oil (Sigma–Aldrich, S3547). A control vehicle consisting of 10% ethanol and 90% sesame oil was also prepared.

2.4. RNA isolation and quantification

Total RNA extraction was carried out from the entire embryonic heads at E11.5 (one day prior to the onset of Pomc expression in the embryonic pituitary at E12.5) and adult pituitary samples were isolated from the underlying pituitary. TRIzol™ Reagent (Sigma–Aldrich, 15596026) was employed for RNA extraction following the manufacturer's protocol. Subsequently, first-strand cDNA synthesis was performed using 1 μg of RNA along with random primers and M-MLV Reverse Transcriptase (Invitrogen, 28025013). Quantitative real-time polymerase chain reaction (qRT-PCR) analysis was conducted on all samples utilizing a CFX Real-Time PCR system (Bio-Rad) with SYBR Green Master Mix (Applied Biosystems). Primer details can be found in Supplementary Table 1, with all primers utilized at a final concentration of 250 nM. Data analysis was carried out using the 2^−ΔΔCT relative quantification method.

2.5. Glucose and insulin tolerance tests

Oral glucose tolerance tests (OGTT) and corresponding insulin measurements were performed by the Metabolic, Physiological and Behavioral Phenotyping Core at the University of Michigan. Briefly, the mice were fasted for 5 h (8:00 am-1:00 pm) and orally gavaged with 2 g/kg glucose in PBS. Blood was sampled at 0, 15, 30, 60, and 120 min for glucose (Accu-Check glucometer, Roche) and insulin measurements (Millipore rat/mouse insulin ELISA kit, EZRMI-13 K). The HOMA index was calculated using fasting glucose ∗ fasting insulin/22.5. For insulin tolerance tests (ITT), the animals were fasted for 5 h (9:30 am-2:30 pm) and 1 U/kg insulin per body weight was injected intraperitoneally. Blood glucose was measured at 0, 15, 30, 60, and 120 min after insulin injection.

2.6. Fluorescence in situ hybridization (RNAscope)

Embryos ranging from E10.5 to E17.5 were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) at 4 °C for varying durations depending on their developmental stage. Specifically, embryos at E10.5 to E12.5 were fixed for 1 h, while heads from embryos at E15.5 to E16.6 were fixed for 2 h, and those at E17.5 were fixed for 4 h. Tissues were stabilized in 10% sucrose/10% gelatin in PBS at 37 °C for 30 min prior to embedding in OCT compound as described previously [11]. The experiment was done according to the RNAscope Multiplex Fluorescent Reagent Kit V2 Assay (Advanced Cell Diagnostics) with some modifications. The adult brain and embryo samples were sectioned using a cryostat and mounted to SuperFrost Excell microscope slides (Fisher Scientific). Sections were fixed with 4% PFA, washed in PBS, dehydrated in ethanol series, and treated with RNAscope pretreatment agents. Then the hybridized probes were amplified, and fluorophores were attached according to the manufacturer's instructions. Slides were dried overnight and imaged the next day using a Nikon Instruments A1 Confocal Laser Microscope with NIS-Elements Software. Probes used for staining were Pomc (Mm-Pomc-C2) for native Pomc expression and Six3 (Mm-Six3-C3) for detection of endogenous Six3 expression.

2.7. Immunohistochemistry and image quantification

For DAB (3,3′-Diaminobenzidine) staining, adult Pomc-CreERT.Six3^fl/fl mice were transcardially perfused with 4% paraformaldehyde in PBS. Brains were removed, placed in 4% paraformaldehyde overnight, and dehydrated in 30% sucrose for several days. The brains were then cut into 30 μm sections using a freezing microtome (Leica). Sections were treated sequentially with 3% hydrogen peroxide/0.5% sodium hydroxide, 0.3% glycine, 0.03% sodium dodecyl sulfate, and blocking solution (PBS with 0.1% triton, 3% Normal Goat Serum) for 1 h and incubated with primary antibody (Anti-POMC, H-029-30, Phoenix Pharmaceuticals) overnight. After three times wash with PBS, the biotinylated secondary antibody (Goat Anti-Rabbit, BA-1000-1.5, Vector Lab) was applied to the sections followed by another three times wash with PBS and incubation in avidin-biotin complex (Thermo Scientific, 32050) for 30 min. Freshly made DAB substrate solution (Thermo Scientific, 34002) was used for color development. All slides were sent to the Department of Pathology Aperio Slide Scanning Core at the University of Michigan for image acquisition at 40× magnification. Quantitative analysis was manually performed in anatomically similar sections from 3 to 4 animals at −1.4 and −1.9 mm distance to bregma. The POMC positive cells per hemisection of the hypothalamus were quantified using Image J. The quantification of POMC immunofluorescence intensity in the Pomc-Cre.Six3^fl/fl embryonic brain was conducted using Image J. This analysis involved comparing three anatomically similar sections from three animals, with replicates for each measurement.

2.8. Daily and hourly food intake

Adult mice were individually housed, and a fresh cage was provided prior to measuring food intake. Hourly food consumption was quantified every 2 h over a 6-hour period following an 18-hour fast. Daily food intake was recorded for five consecutive days at 4:00 pm. Some of the Six3^fl/fl mice exhibited pronounced food wasting behavior, especially females, these mice were excluded from the food intake experiment.

2.9. Single-cell RNA sequencing and TRAP-seq analysis

The single-cell RNA-sequencing data were analyzed as previously described [16], and violin plots were generated with Seurat (version 2.0). The percentages of cells in each cluster were calculated by identifying the number of cells that had at least one unique molecular identifier (UMI) for Six3 against the total number of cells in each cluster. The TRAP-seq data [16], presented as counts per million (CPM), were normalized for the counts of Six3 and Pomc against the total counts.

2.10. Data analysis

The data presented are shown as means ± standard error of the mean (SEM) and sample sizes are indicated in the figure legends and the main text. All data were analyzed using GraphPad Prism 9.0 software (GraphPad Software, San Diego, California, USA). When two groups were compared, the unpaired two-tailed Student's t test was applied. One-way or two-way ANOVAs followed by Tukey's or Bonferroni's multiple comparison tests were conducted for pair-wise analyses. A detailed description of the statistical method used for each analysis is noted in each figure legend. P < 0.05 was considered significant.

3. Results

3.1. Six3 and Pomc co-express during hypothalamic development

By using single-cell RNA sequencing (scRNA-seq), we have recently characterized the transcriptomic profiles of mouse hypothalamic POMC neurons at various developmental and postnatal stages [16]. In particular, Six3 emerged as one of the transcription factor coding genes prominently expressed within clusters exhibiting high Pomc transcript abundance [16]. Reanalysis in this report of the datasets collected in that study indicate that Six3 is notably enriched in cluster 1 (Pomc^high/Prdm12); cluster 4 (Pomc^low/Otp) and cluster 7 (Pomc^low/Prdm13) with over 75% of cells within these three clusters expressing Six3 (Figure 1A). Six3 transcripts are less abundant in the other five clusters in which Pomc is also detected at much lower levels [16]. Further analysis of cluster 1 across four embryonic stages (E11.5, E13.5, E15.5 and E17.5) and two early postnatal periods (P5 and P12) revealed that Six3 transcripts are present in more than 75% of the highest Pomc-expressing neurons at all six developmental stages (Figure 1B). In addition, reanalysis of a different transcriptomic dataset collected from adult (P60) male mice hypothalamic POMC neurons obtained by performing a Translating Ribosome Affinity Purification (TRAP)-seq study [16] also showed that Six3 is significantly enriched in the ribosomal fraction associated with high Pomc expression (Figure 1C). To further verify that Six3 and Pomc coexpress in arcuate hypothalamic neurons we conducted double immunofluorescence on sagittal sections taken from Pomc-TdTomato transgenic E10.5 mouse embryos (Figure 1D, top panels). In addition, we performed double fluorescence in situ hybridization (RNAscope) on sagittal mouse brain sections taken at E12.5 and E16.5 (Figure 1E) and on coronal E14.5 and adult mouse hypothalamic sections (Supplementary Fig. 1). Our findings show that Six3 is expressed in POMC neurons since the onset of Pomc expression in the basal hypothalamic region at E10.5 and maintained throughout development and adulthood.

Figure 1.

Six3 is expressed in Pomc-expressing cells during development and adulthood. (A) Percentage of Six3-positive cells across the eight cell clusters from previous single-cell RNAseq study [16]; (B) Percentage of Six3-positive cells across six developmental stages in the Pomc^high/Prdm12 cluster [16]; (C) The expression levels of Six3 and Pomc are both significantly elevated in pulled-down samples from a previous TRAP-seq study performed on male mice at postnatal day 60 [16]; (D) Double immunofluorescence on sagittal cryosections of Pomc-TdTomato E10.5 embryos showing the co-localization of TdTomato (red) and SIX3 (green). Double fluorescence in situ hybridization showing the co-localization of Pomc (green) and Six3 (red) at E12.5 and at E16.5. CPM: Counts Per Million; TRAP-seq: Translating ribosome affinity purification with RNA-sequencing; Pit: Pituitary; 3V: third ventricle; RP: Rathke's pouch; Sagital image orientation: left, posterior (P); right, anterior (A). Scale Insets are magnified views of the indicated boxes. Arrows indicate co-expressing neurons in the merged panels. Scale bar: 50 μm.

3.2. Deletion of Six3 from POMC neurons reduced Pomc expression

To elucidate the potential molecular regulation of Six3 in hypothalamic Pomc expression, we generated mice specifically lacking Six3 from POMC cells (POMCSix3KO) by crossing POMC cell-specific BAC transgenic Pomc-Cre mice with conditional Six3 mutant (Six3^fl/fl) mice [19]. We found that immunofluorescence intensity in POMC+ hypothalamic neurons from E12.5 and E16.5 POMCSix3KO embryos decreased by 33% and 48% respectively, in comparison with same age wild-type embryos (Figure 2A), indicating a role for Six3 in hypothalamic POMC immuno-expression during development. To explore whether Six3 is involved in the early establishment of the hypothalamic POMC lineage, we generated a temporal-specific conditional mutant mouse strain allowing for Six3 ablation at different developmental stages by crossing Six3^fl/fl conditional mutants with mice expressing a tamoxifen-inducible Cre recombinase transgene (CAAG-CreERT) [20]. Tamoxifen (150 mg/kg, i.p.) was administered to compound CAAG-CreERT.Six3^fl/fl pregnant dams at post-coitum day 9.5, one day before the initiation of Pomc expression in the embryonic hypothalamus. The embryos were subsequently collected either at E11.5 for quantitative RT-PCR analysis or at E12.5 for immunofluorescence detection of POMC, respectively (Figure 2B). A great reduction (nearly 90%) of Six3 transcript levels was found in brain samples from E11.5 CAAG-CreERT.Six3^fl/fl embryos (Six3KO@9.5) confirming a successful tamoxifen-inducible deletion of Six3 in these compound mutant mice (Figure 2D). Interestingly, the significant drop in Six3 mRNA levels found in Six3KO@9.5 embryos was paralleled by ∼50% decrease in the level of Pomc transcripts at E11.5 (Figure 2C), underscoring the functional significance of Six3 for optimal arcuate Pomc expression in early development. This significant reduction in Pomc expression was observed in both sexes (Supplementary Fig. 2). We also assessed in E11.5 Six3KO@9.5 embryos the expression levels of several additional transcription factor genes known to regulate Pomc expression, including Nkx2-1 [11], Mash1 [9], Ngn3 [13], Prdm12 [14], Isl1 [10], and Shh [12] (Supplementary Fig. 3) and found that deletion of Six3 at E9.5 did not alter the expression of those genes, suggesting a direct independent effect of Six3 in the transcriptional regulation of arcuate Pomc expression. Given that Pomc is also abundantly expressed in corticotrophs and melanotrophs of the pituitary, we quantified Pomc transcript levels in the pituitary glands of adult POMCSix3KO mice. Our findings revealed a 36% reduction in Pomc expression with the deletion of Six3 (Supplementary Fig. 4). We also evaluated the expression levels of Pitx1, an important regulator in the development and maintenance of the normal function of the pituitary [25]. Our results showed deletion of Six3 from Pomc-expressing cells does not affect the expression of Pitx1 (Supplementary Fig. 4).

Figure 2.

Six3 ablation reduced Pomc expression during development. (A) Comparison of POMC (green) immunofluorescence intensity on sagittal cryosections of Six3^fl/fl and POMCSix3KO E12.5 and E16.5 embryos. Intensity (right) was quantified on anatomically similar sections from three embryos. (B) Schematic diagram showing the overall experimental procedure for panels C and D. (C) Reduced POMC (green) immunofluorescence intensity in Six3KO@E9.5 embryos at E12.5. (D) Quantitative PCR showing significantly decreased Six3 and Pomc transcripts abundance in Six3KO@E9.5 embryos analyzed at E11.5. Pit: Pituitary; 3V: third ventricle; RP: Rathke's pouch; Image orientation: left, posterior (P); right, anterior (A). Scale Insets are magnified views of the indicated boxes. Arrows indicate co-expressing neurons in the merged panels. Scale bar: 50 μm. Figure 2B was generated from Biorender.

3.3. Selective deletion of Six3 from POMC neurons induced hyperphagia and mild obesity in males

Given that Six3 and Pomc coexpress at all developmental and postnatal ages, we investigated whether deletion of Six3 specifically from POMC neurons throughout life utilizing the compound Pomc-Cre.Six3^fl/fl mouse strain alters melanocortin function. To this end, we evaluated several metabolic parameters, food intake and body weight in adult compound POMCSix3KO mice. When tested at 16–17 weeks of age, POMCSix3KO male mice showed a 13% increase (P = 0.0512) in daily food intake, with no further increase after 18 h fasting (Figure 3A). Consequently, body growth curves revealed that POMCSix3KO males are mildly obese from 14 weeks of age onward, relative to their wild-type siblings (Figure 3B). Consistent with body weight, body composition measured at 8 weeks of age showed no significant difference (Figure 3C) whereas at 16 weeks of age, POMCSix3KO males exhibited a 33% increase in body fat composition, although without reaching statistical significance (Six3^fl/fl: 12.80 ± 1.59 % vs. POMCSix3KO: 17.04 ± 1.58 %; P = 0.068) (Figure 3D). Given the association between obesity and impaired glucose metabolism, we conducted glucose tolerance tests on these mice. Our results indicated that POMCSix3KO males exhibited significantly improved glucose tolerance (Figure 3E) and no significant changes in circulating insulin levels (Figure 3F), HOMA-IR values (Figure 3G) or insulin tolerance (Figure 3H). Different from what we found in males, POMCSix3KO females displayed normal body weight curves (Supplementary Fig. 5A), normal body fat composition when measured at 8 and 16 weeks of age (Supplementary Figs. 5B and 5C) and no clear signs of hyperphagia when fed ad libitum or after an 18-h fasting (Supplementary Fig. 5D).

Figure 3.

Six3-specific deletion from POMC neurons during embryogenesis induced mild obesity in adult male mice. (A) Average daily food intake at 16–17 weeks of age and food intake measurement after 18 h fasting. (B) Body weight comparisons between POMCSix3KO and Six3^fl/fl mice from 3 to 24 weeks of age. (C–D) Body composition of POMCSix3KO and Six3^fl/fl mice at 8 (C) and 16 weeks of age (D). (E) Oral glucose tolerance test with corresponding blood insulin measurements (F) and HOMA-IR (G). (H) Insulin tolerance tests in POMCSix3KO and Six3^fl/fl mice. Two-tailed unpaired Student's t-tests were used to compare the genotype effects. Data shown are the means ± s.e.m of biologically independent samples. ∗P < 0.05, ∗∗P < 0.01.

3.4. High-fat diet does not exacerbate obesity in POMCSix3KO mice

Given that POMCSix3KO males developed mild obesity when fed ad libitum with normal chow, we investigated whether a high-fat diet (HFD; 60% fat content) initiated at 4 weeks of age and onward would exacerbate obesity. Similar to what we observed with the chow diet, POMCSix3KO males fed on a HFD exhibited increased body weight (∼5 g in average) after 14 weeks of age, in comparison to their wild-type brothers (Figure 4A) with no significant changes in body fat composition (Figure 4B) or daily food intake when either fed ad libitum (Figure 4C, left) or refed after an 18-h fast, respectively (Figure 4C, right). Glucose tolerance (Figure 4D), serum insulin levels (Figure 4E), HOMA-IR (Figure 4F) and insulin tolerance (Figure 4G) were also normal in POMCSix3KO males. Like with the chow diet, POMCSix3KO females fed a HFD did not develop obesity (Supplementary Fig. 6A) nor showed changes in body fat content (Supplementary Fig. 6B), glucose tolerance (Supplementary Fig. 6C), corresponding insulin levels during the oral GTT (Supplementary Fig. 6D), HOMA-IR (Supplementary Fig. 6E), and insulin sensitivity (Supplementary Fig. 6F).

Figure 4.

Deletion of Six3 specifically from POMC neurons led to modest obesity in male mice subjected to a high-fat diet. (A) Body weight comparisons from 4 to 21 weeks of age. (B) Body composition. (C) Average daily food intake and refeeding intake after 18 h fasting. (D) Glucose tolerance test and corresponding blood insulin measurements. (E) HOMA-IR (F) and insulin tolerance test (G) in POMCSix3KO and Six3^fl/fl mice subjected to HFD. Data shown are the means ± s.e.m of biologically independent samples. ∗P < 0.05.

3.5. Deletion of Six3 from POMC neurons in adult mice does not induce obesity

Given the co-expression of Six3 and Pomc in the adult arcuate nucleus (Supplementary Fig. 1, bottom panel), we investigated whether the specific ablation of Six3 from POMC neurons in adult mice affects Pomc expression and melanocortin function. To this end, we administered tamoxifen or vehicle to 12-week-old (postnatal day 84) Pomc-CreERT.Six3^fl/fl (POMCSix3KO@P84) mice and their corresponding Six3^fl/fl control littermates [24]. We performed immunohistochemistry in coronal brain sections from POMCSix3KO@P84 mice and found 23% fewer POMC immunolabeled hypothalamic neurons in comparison with vehicle treated control POMCSix3KO@P84 mice (Figure 5A).

Figure 5.

Conditional deletion of Six3 from POMC neurons in adult male mice did not induce obesity. (A) Immunostaining illustrates the reduction in the number of POMC neurons following tamoxifen injection compared to vehicle injection in POMCSix3KO@P84 mice. (B) Schematic diagram outlining the overall experimental design. (C) Body weight, (D) body composition, (E) daily food intake and refeeding intake after 18 h fasting, (F) glucose tolerance and (G) insulin tolerance in Pomc-CreERT.Six3^fl/fl and Six3^fl/fl mice treated either with tamoxifen or vehicle. Data shown are the means ± s.e.m of biologically independent samples. ∗P < 0.05, ∗∗P < 0.01. Figure 5B was generated from Biorender.

To investigate the physiological consequences of this reduction, we measured body growth curves in the POMCSix3KO@P84 receiving a chow diet until 19 weeks of age, then switched to a HFD from 19 to 27 weeks of age (Figure 5B). Body weight was recorded weekly before and after tamoxifen or vehicle injections. Two-way ANOVA mixed-effects analysis revealed no significant differences among all groups (Figure 5C). No differences were observed either in body composition or food intake at 25 weeks of age (Figure 5D,E). Glucose tolerance tests performed on 27-week-old POMCSix3KO@P84 males fasted for 5 h indicated similar glucose clearance from their blood compared to vehicle-treated Pomc-CreERT.Six3^fl/fl control mice (Figure 5F), but significantly worsened compared to vehicle-treated Six3^fl/fl mice. We found no changes in insulin sensitivity among the four groups (Figure 5G). When similar comparisons were conducted in females, we found no significant changes in body weight (Supplementary Fig. 7A), fat and lean mass composition (Supplementary Fig. 7B), food intake (Supplementary Fig. 7C), glucose or insulin tolerance (Supplementary Figs. 7D and 7E) in any of the experimentally tested groups. In summary, although Six3 ablation from mature POMC neurons led to a significant reduction in arcuate POMC expression, this deficit did not result in any overt physiological changes.

4. Discussion

In this study, we combined developmental, molecular, cellular, and functional genetics approaches to explore the participation of the transcriptional regulator Six3 in the establishment of melanocortin neuron identity and the initiation and maintenance of Pomc expression within the arcuate nucleus. Specifically, our findings indicate the following key observations: 1) Six3 and Pomc co-localize in arcuate hypothalamic neurons from the onset of Pomc expression at E10.5 and throughout the entire lifespan; 2) Selective deletion of Six3 from POMC neurons reduces hypothalamic Pomc expression during development; 3) Six3 expression in the developing ventral hypothalamus is necessary for the onset of arcuate Pomc expression as demonstrated in temporally conditioned Six3 null allele mutant mice; 4) Six3 early ablation does not impair transcript levels of other transcription factors known to regulate the onset of Pomc expression, suggesting a direct participation of Six3 as a transactivator of arcuate Pomc; 5) the expression pattern of Six3 in the hypothalamus extends beyond that of Pomc, suggesting that Six3 is necessary but not sufficient for arcuate Pomc expression; 6) Selective deletion of Six3 from POMC neurons induces mild obesity and improved glucose tolerance in adult male mice; 7) Conditional deletion of Six3 from POMC neurons of adult mice leads to a ∼23% reduction in hypothalamic Pomc expression without critically altering any measured metabolic parameter. Together, these results indicate that Six3 plays important roles in the regulation of hypothalamic Pomc expression in age dependent and sexually dimorphic patterns.

The Sine Oculis Homeobox Homolog (SIX) proteins represent a family of evolutionarily conserved transcription factors that play crucial roles in regulating multiple cellular processes, including proliferation, differentiation, apoptosis, adhesion, and migration [26]. The protein SIX3 was initially found at the anterior border of the developing neural plate and within the eye, underscoring its significance in forebrain and ocular development, respectively [17]. Six3 is also expressed early in the ventral hypothalamic anlage, and a few studies have documented its important role in the maturation of GnRH and kisspeptin expressing neurons [27,28].

Our recent scRNA-seq study performed with POMC neurons at different developmental ages revealed that Six3 is also present in arcuate POMC neurons [15]. To study the functional roles of Six3 in arcuate POMC neurons, we first generated Pomc-Cre.Six3^fl/fl mice and demonstrated that the selective deletion of Six3 from POMC neurons leads to a 40% reduction in Pomc expression, resulting in a mild obesity phenotype in adult males (Figure 2A-B and Figure 3). This finding supports the threshold theory, which suggests that below certain incrementally reduced levels of Pomc expression mice develop from mild to extreme hyperphagia and obesity, respectively. Evidence supporting this theory is based on the fact that a ∼20% reduction of Pomc mRNA found in mice lacking the Pomc neuronal enhancer 2 (nPE2) does not cause obesity, while a 70% reduction in Pomc mRNA resulting from the absence of Pomc neuronal enhancer 1 (nPE1) leads to an appreciable but mild increase in body weight, whereas the simultaneous ablation of nPE1 and nPE2 enhancers reduces Pomc levels by 90% resulting in extreme obesity [29]. These results have led us to propose that food intake and body weight are visibly impaired once hypothalamic Pomc levels drop beyond ∼40% [29].

Our results showed improved glucose tolerance when Six3 was developmentally absent in POMC neurons. Given that a subset of POMC neurons expressing the leptin receptor (Lepr) directly regulate glucose homeostasis [30], we hypothesized that the deletion of Six3 in POMC neurons could impact Lepr-expressing neurons and thus might contribute to the improved glucose tolerance observed in our animal model. To investigate this, we assessed the percentage of cells expressing both Pomc and Lepr cells that also express Six3 using single-cell RNAseq data from a previous study [27]. Our findings indicated that approximately 85% of cells expressing both Pomc and Lepr also express Six3, suggesting that the deletion of Six3 in POMC neurons could potentially influence the expression of Lepr and the function of cells co-expression of Pomc and Lepr. However, it is worth noting that this study [30] identified only a small population (9%) of POMC neurons expressing Lepr, which is consistent with another single-cell RNA-seq study showing that only 12% of POMC neurons express Lepr [31]. Another possible explanation for the improved glucose tolerance is the potential increase in glycosuria, although this was not examined in the current study. This phenotype has previously been reported by our lab in obese hypothalamic POMC-deficient mice with improved glucose tolerance [32].

Consistent with these findings, several studies have demonstrated that restoring Pomc expression in specific POMC neuronal subpopulations can mitigate some obesity-related phenotypes in Pomc-deficient mice. For example, Burke et al. found that Pomc re-expression selectively in 5-hydroxytryptamine 2c receptor neurons, which comprise 40% of all arcuate POMC neurons, prevents hyperphagia in both sexes but mitigates obesity only in male mice [33]. Similarly, Lam et al. (2015) reported that conditional Pomc re-expression in leptin receptor-positive POMC neurons restored POMC immunoreactivity by 67% and normalized energy expenditure, glucose levels, and locomotor activity [34]. A more recent study highlighted that inducing Pomc expression in approximately 23–25% of GABAergic POMC neurons was sufficient to normalize food intake and body weight [35]. Conversely, a nonspecific Pomc rescue in the same study, which increased Pomc expression by 32–35% in Pomc-deficient mice, did not improve food intake but resulted in a slight weight loss, suggesting that the optimal Pomc dose for maintaining satiety signals depends on the specific neuronal subpopulation [35]. It is worth noting that the above-mentioned studies utilized a reversible strain of Pomc-deficient mice, which are constitutively predisposed to hyperphagia and obesity [2]. Thus, it is possible that the threshold for restoring Pomc expression in these extremely obese mice differs from the threshold of Pomc mRNA levels below which obesity becomes apparent in mice. Overall, the comparison of our current study with prior research emphasizes that a threshold of ∼40% of arcuate Pomc expression is necessary to maintain proper satiety signals, at least in laboratory mice fed ad libitum. Thus, our study presents a novel role for SIX3, as a cell-autonomous transcription factor necessary to maintain high expression levels of Pomc in the developing and adult arcuate nucleus. SIX3 adds to the repertoire of transcription factors, which include ISL1 [9], NKX2.1 [10] and PRDM12 [13], that dictate the early specification of melanocortin neuron identity in the developing hypothalamus. Each of these four transcription factors plays a critical role -rather than partially redundant functions- in the regulation of hypothalamic Pomc expression since the individual conditional mutants of each of these four genes specifically in POMC cells showed severely impaired hypothalamic Pomc expression [9,10,13, and this study].

Our study found that male mice lacking Six3 specifically from POMC neurons develop mild obesity (Figure 3, Figure 4). Although this phenotype has not been matched in females (Supplementary Figs. 5 and 6) we did observe excessive food wasting behavior in female mice. Our qRT-PCR study separated by sex showed deletion of Six3 leads to a significant reduction in Pomc expression in both sexes (Supplementary Fig. 2) precluding a sexually dimorphic dependence of Six3 in arcuate Pomc expression (Figure 2A-D). Moreover, given the expression of estrogen receptor alpha (Esr1) in POMC neurons, we examined the Esr1 transcript levels in the absence of Six3. Our results indicate that the deletion of Six3 does not affect Esr1 expression (data not shown). However, sexual dimorphism in feeding behavior and energy homeostasis is well documented in several mutant mouse models. For example, deletion of the transcription factor Tap63 specifically from POMC neurons increased diet-induced obesity (DIO) in female but not in male mice [36]. Similar sex-specific regulatory mechanisms have been reported in male mice lacking Sirt3 (Sirtuin 3) in POMC neurons, which showed decreased body weight and adiposity together with an increase in energy expenditure, all effects that were not observed in females [37]. Another example has been found after restoring Pomc expression specifically in 5HT2cR+ (5-hydroxytryptamine 2c receptor) POMC neurons of male mice which, in contrast to females, normalized physical activity and energy expenditure [33]. Our own previous study reported that reactivating neuronal Pomc expression in arcuate Pomc-deficient mice normalized food intake in females whereas males maintained residual hyperphagia [2].

In summary, our findings demonstrate that Six3 plays a crucial role in the regulation of hypothalamic Pomc expression, and that its absence from Pomc-expressing cells leads to mild obesity in male mice. Therefore, SIX3 integrates a unique repertoire of TFs, including ISL1, NKX2.1 and PRDM12 which, together, dictate the early identity of central melanocortin neurons and control arcuate Pomc expression from development to adulthood.

CRediT authorship contribution statement

Hui Yu: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Angelika Chiang: Writing – original draft, Validation, Methodology, Formal analysis, Data curation. Marcelo Rubinstein: Writing – review & editing, Investigation, Funding acquisition, Formal analysis. Malcolm J. Low: Writing – review & editing, Supervision, Investigation, Funding acquisition, Formal analysis, Data curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Table 1

qRT-PCR primers used in this study.

Supplementary Figure 1

Double fluorescence in situ hybridization showing the co-localization of Pomc (green) and Six3 (red) mRNA at E14.5 and in adults in wild-type C57BL/6J mice. 3V: third ventricle; Scale Insets are magnified views of the indicated boxes. Arrows indicate co-expressing neurons in the merged panels. Scale bar: 50 μm.

Supplementary Figure 2

Quantitative PCR showing significant reduction of Six3 and Pomc transcripts abundance in Six3KO@E9.5 embryos analyzed at E11.5 in (A) males and (B) females. Data shown are the mean ± s.e.m of biologically independent samples. ∗∗∗∗P<0.0001and ∗∗P<0.01 indicate significance.

Supplementary Figure 3

Deletion of Six3 at E9.5 does not change the expression of transcription factor genes involved in early hypothalamic Pomc expression. Quantitative RT-PCR showing the relative transcript abundance in Six3KO@E9.5 and Six3^fl/fl embryos at E11.5. Data shown are the means ± s.e.m of biologically independent samples.

Supplementary Figure 4

Deletion of Six3 reduced Pomc expression in adult pituitary. Quantitative PCR reveals a significant decrease in the transcription abundance of Six3, Pitx1 and Pomc in the pituitary. Data shown are the mean s.e.m of biologically independent samples. ∗P<0.05 indicate significance. ∗∗P<0.01 and ∗∗∗∗P<0.0001 indicate significance.

Supplementary Figure 5

Deletion of Six3 specifically from POMC neurons does not induce obesity in female mice. (A), body weight comparison from 3 to 18 weeks of age, body composition analysis at 8 weeks of age (B) and 16 weeks of age (C) and refeed food intake after 18-hour fasting period (D). Data shown are the means ± s.e.m of biologically independent samples.

Supplementary Figure 6

The high-fat diet feeding does not induce obesity in female mice lacking Six3 expression in POMC neurons. (A) Body weight comparison from 4 to 21 weeks of age. (B)Body composition analysis. (C) Glucose tolerance test with (D) corresponding blood insulin measurements. (E) HOMA-IR and (F) insulin tolerance tests. Data shown are the means ± s.e.m of biologically independent samples.

Supplementary Figure 7

Deletion of Six3 from POMC neurons in adult female mice at age 7 wk (P84) does not lead to obesity phenotype. (A) Body weight comparisons from 9 to 25 weeks of age after tamoxifen was administrated at 12-weeks of age. (B) Body composition analysis. (C) Average daily food intake and refed food intake after 18-hour fasting period. (D) Glucose tolerance test and (E) Insulin tolerance test. Data shown are the means ± s.e.m of biologically independent samples.

Data availability

Data will be made available on request.