Transcriptional control of brown adipocyte differentiation and function by NFIA: recent perspectives on deciphering metabolic diseases

Abstract

Brown adipocytes dissipate chemical energy as heat and confer protection against type 2 diabetes and obesity. Nuclear factor I-A (NFIA) is a transcription factor that orchestrates the brown fat gene programme by activating cell type-specific enhancers and facilitating the genomic binding of PPARγ, the master regulator of adipogenesis, to these enhancers. NFIA promotes mitochondrial oxidative phosphorylation and thermogenesis, while reciprocally suppressing adipose tissue inflammation, thereby contributing to the maintenance of glucose and body weight homeostasis in mice. Here the author provides an overview of the identification of NFIA as a pivotal regulator of brown adipocyte biology, elucidates its underlying mechanisms of action, examines its implications for systemic metabolism and outlines future perspectives for research in this field.

Graphical Abstract

Graphical Abstract.

Abbreviations

BMI

body mass index

BAT

brown adipose tissue

ChIP-seq

Chromatin immunoprecipitation coupled with high-throughput sequencing

CI

confidence interval

DALY

disease-adjusted life years

EBF2

early B cell factor 2

EHMT1

euchromatic histone-lysine N-methyltransferase 1

FDG-PET

fluorodeoxyglucose-positron emission tomography

GLP-1

glucagon-like peptide-1

GWAS

genome-wide association studies

MCP-1

monocyte chemoattractant protein-1

NFIA

Nuclear factor I-A

Ox-Phos

oxidative phosphorylation

PPARγ

peroxisome proliferator activated receptor gamma

PRDM16

PRD1-BF1-RIZ1 homologous-domain-containing protein 16

SNPs

single nucleotide polymorphisms

Ucp1

uncoupling protein 1

WAT

white adipose tissue

Obesity is defined as a condition resulting from a chronic excess of energy intake and insufficient energy expenditure. A global epidemiological study examining the impact of 88 risk factors on their associated health outcomes across 631 risk–outcome pairs measured in disease-adjusted life years (DALYs) revealed that high body mass index (BMI) was found to have the sixth strongest negative impact on the global burden of disease (1). Unfortunately, the percentage of the disease burden attributed to high BMI significantly increased from 2000 (2.5%, 95% confidence interval (CI): 1.1–3.9%) to 2021 (4.5%, 95% CI: 1.9–6.8%), indicating an urgent need to identify and implement effective interventions. Consistently, multiple large-scale biobank studies using polygenic risk scores—which indicate the cumulative effect of disease susceptibility variants identified by genome-wide association studies (GWAS)—have reported that obesity has one of the strongest negative impacts on human lifespan (2, 3).

Recently, glucagon-like peptide-1 (GLP-1) receptor agonist has been widely prescribed to obese individuals with type 2 diabetes, and their use is expected to increase for obese individuals without diabetes. However, not everyone is responsive to GLP-1 receptor agonist, and substantial number of patients need to discontinue the treatment due to adverse effects such as nausea (4). While number of bariatric surgeries performed is increasing both in Japan and globally, the absolute number of cases is still limited, at least in Japan (5). Furthermore, all the currently approved anti-obesity treatments including GLP-1 receptor agonist and bariatric surgeries act to repress energy intake; however, at least in principle, enhancing energy dissipation would also be a reasonable approach to treat obesity.

Brown adipocytes are anticipated as a potential target of anti-obesity and anti-diabetes therapy because they are able to dissipate energy as heat via mitochondrial uncoupling protein 1 (Ucp1), while white adipocytes generally store energy as lipid (6). ‘Classical’ brown adipocytes, which are reported to share a common progenitor with skeletal myocytes but not with white adipocytes (7), are located in distinct brown adipose tissue (BAT) depots such as the interscapular and perirenal regions. On the other hand, ‘inducible’ brown adipocytes—often referred to as ‘beige’ or ‘brite’ adipocytes—are sporadically found in white adipose tissue (WAT) depots, especially in response to cold or β-adrenergic stimulations (8, 9). Both trans differentiation from white adipocytes into beige adipocytes and de novo beige adipogenesis have been proposed (10, 11). Although transcriptional regulators of classical brown and beige adipocytes are reported to substantially overlap, the specific functions of beige adipocytes remain under active investigation (Fig. 1). Over the last two decades, the existence of functional brown adipocytes in human adults is established by multiple independent evidences using fluorodeoxyglucose-positron emission tomography and computed tomography (FDG-PET) (12–15). Human brown adipocyte activity is inversely correlated with BMI (14), and studies have shown that cold exposure (16, 17) and β3-adrenergic receptor activation (18, 19) are able to activate brown adipocytes and increase systemic energy expenditure. Mechanistically, while Ucp1 is the best known effector protein driving thermogenesis in brown and beige adipocytes, recent studies have identified Ucp1-independent thermogenic pathways, including futile creatine cycle and Ca²⁺ cycling (20). Additionally, the roles of thermogenic adipocytes extend beyond heat generation (9), with increasing attention given to their endocrine functions (21, 22) and contributions to adipose tissue remodelling (23). Collectively, these findings warrant further investigation of brown adipocytes towards development of novel anti-obesity and anti-diabetes therapy.

Fig. 1.

Cellular lineage of brown, beige and white adipocytes. Brown adipocytes dissipate energy as heat via mitochondrial Ucp1, while white adipocytes generally store energy as lipid. Classical brown adipocytes, which are reported to share a common progenitor with skeletal myocytes but not with white adipocytes, are located in distinct brown adipose tissue depots. Inducible brown, or beige adipocytes are sporadically found in white adipose tissue depots, especially in response to cold or β-adrenergic stimulations.

GWAS have identified hundreds of single nucleotide polymorphisms (SNPs) associated with BMI (24). Among these, SNPs located between introns 1 and 2 of the FTO gene show the largest effect on BMI among common variants (minor allele frequency ~1%). Carrying the risk allele is associated with an increase in BMI of 0.25–0.41 kg/m² and increased an odds ratio for obesity of 1.18–1.27 (25). Intriguingly, risk alleles of the FTO variants have been shown to engage in long-range chromatin interactions that modulate IRX3/5 expression, rather than FTO itself, in both the brain (26) and adipocytes (27). Moreover, these risk alleles have been reported to impair the function of brown and beige adipocytes (27, 28). Given that gene–environment interactions between FTO variants and physical activity have been consistently reported in relation to both cross-sectional BMI (29–31) and longitudinal changes in BMI (32), the interrplay between human brown adipocyte function and modifiable lifestyle factors, such as physical activity, would be of considerable interest for further investigation. Most recently, an association study was conducted in the Japanese population to examine the relationship between genetic variants near β1-, β2- and β3-adrenergic receptor genes (ADRB1, ADRB2 and ADRB3) and human brown adipocyte activity under mild cold exposure, as measured by FDG-PET (33). In that study, a significant association was identified between rs1042718, located near ADRB2, and brown adipocyte activity. Genome-wide approaches are anticipated to provide a more comprehensive understanding of how genetic variation contributes to the heterogeneity of human brown adipocyte activity.

Transcriptional Regulators of Brown Adipogenesis

The transcription factor peroxisome proliferator activated receptor gamma (PPARγ) is the master regulator of adipocyte differentiation (34). PPARγ works in concert with the transcription factor C/EBPα, and the two factors engage in a positive feedback loop by mutually enhancing each other’s expression, thereby promoting and maintaining adipocyte differentiation (35). However, PPARγ alone is not sufficient to determine brown versus white adipocyte identity. In this context, peroxisome proliferator-activated receptor gamma coactivator 1 (PGC-1) was identified as a cold-inducible coactivator of PPARγ that enhances its transcriptional activity on the Ucp1 promoter (36). More recently, a transcriptional co-regulator called PRD1-BF1-RIZ1 homologous-domain-containing protein 16 (PRDM16) was identified as one of the most highly enriched transcriptional regulators in brown adipocytes relative to white adipocytes (37). Subsequently, PRDM16 was shown to control cell fate switch between brown adipocytes and skeletal myocytes (7). Transgenic expression of PRDM16 in adipocytes is sufficient to display increased energy expenditure, limited weight gain and improved glucose tolerance in response to a high-fat diet (38). On the other hand, adipocyte-specific deletion of PRDM16 resulted in obesity and insulin resistance (39). Intriguingly, classical brown adipocyte-specific deletion of PRDM16 revealed that while PRDM16 is dispensable for embryonic brown fat development, it is required for maintenance of its function in aged mice (40). PRDM16 physically interacts with PPARγ (41), C/EBPβ (42) and euchromatic histone-lysine N-methyltransferase 1 (EHMT1) (43) to control brown and beige adipocyte cell fate and function. Intriguingly, expression levels of the brown fat-specific genes including UCP1 in human prerenal adipose tissue were highly correlated with the expression of the PRDM16-EHMT1 complex (44). Motif analysis of brown adipose tissue-specific PPARγ binding sites identified early B cell factor 2 (EBF2) as a regular of brown adipocyte identity (45). EBF2 was one of the most selectively expressed genes in brown adipogenic precursor cells (46). EBF2 physically interacts with DPF3, a brown adipocyte-selective component of the BAF chromatin remodelling complex that was required for brown fat gene programming and mitochondrial function (47). Transgenic expression of EBF2 in adipocytes induces browning of white adipose tissue and confers resistance to high-fat-diet-induced obesity (48). Adipocyte-specific deletion of EBF2 results in impaired thermogenic gene expression in classical brown adipocytes and leads to cold intolerance (49).

Transcriptional Control of Brown Fat Development by NFIA

It is very well known that regulatory elements controlling gene expression, including enhancers and promoters, are characterized by accessible or open chromatin structures that allow binding of transcription factors and co-factors to the genome. Cell fate and lineage relationships can be inferred from the accessible chromatin landscape. In fact, unbiased clustering of the accessible chromatin landscape has been shown to more accurately recapitulate known cellular lineage relationships compared to clustering based on gene expression patterns (50). Therefore, genome-wide open chromatin analysis (51, 52) is a promising strategy to identify novel transcriptional regulator in an unbiased manner. In a genome-wide open chromatin analysis of BAT, inguinal WAT and epididymal WAT followed by motif analysis of the brown fat-specific open chromatin regions in mice, a binding motif for the NFI transcription factor was the most highly enriched. Of the four isoforms of the NFI family, NFIA was highly expressed in BAT compared with WAT or skeletal muscle, and the expression level of NFIA was robustly induced during brown adipocyte differentiation. These results strongly indicated that NFIA is a candidate regulator that defines brown adipocyte identity. Indeed, both gain- and loss-of-function experiments demonstrated that NFIA is a crucial regulator of brown adipocyte differentiation. Chromatin immunoprecipitation coupled with high-throughput sequencing (ChIP-seq) revealed that NFIA binds to and activates the brown fat-specific enhancers even before differentiation. During the differentiation, NFIA facilitates the binding of PPARγ and other transcription factors including C/EBPα, C/EBPβ and EBF2, to the brown fat-specific enhancers and promoters to control the brown fat gene programme (Fig. 2). Of note, shRNA-mediated knockdown of PPARγ almost completely abolished the effects of NFIA. On the other hand, PRDM16 was dispensable for the effect of NFIA. Consistently, co-immunoprecipitation experiments showed that NFIA does not physically interact with PRDM16. In human perirenal brown adipose tissue, expression of NFIA and the brown fat-specific genes is positively correlated (53).

Fig. 2.

Transcriptional control of brown adipocyte differentiation by NFIA. NFIA binds to and activates the brown fat-specific enhancers even before differentiation. During the differentiation, NFIA facilitates the binding of PPARγ and other transcription factors, including C/EBPα, C/EBPβ and EBF2, to the brown fat-specific enhancers and promoters to control the brown fat gene programme.

For the differentiation of mesodermal precursor cells into brown adipocytes, not only activation of the brown fat gene programme but also inactivation of the myogenic gene programme is required. The site-directed mutagenesis experiments revealed that a C-terminal proline-rich domain of NFIA, especially the C-terminal 17 amino acid residues (aa 493–509), is indispensable for inducing adipogenesis. The 17 amino acid residues are indispensable in order to rescue impaired adipogenesis caused by NFIA-KO. However, these residues are dispensable for suppressing myogenesis, indicating dual ways of action of NFIA to ensure optimal regulation of brown adipocyte differentiation (54).

Cis-Effect of Genetic Variation on Ucp1 Expression between B6 and 129 Inbred Mouse Strains Are Regulated by NFIA

Human brown adipocyte activity is inversely correlated with BMI (14) and age (55); however, its activity is highly variable even among lean and young individuals (17). Intriguingly, inbred mouse strains also show differential vulnerability to diet-induced obesity and exhibit concordant adipocyte browning capability (56–59), suggesting the effect of genetic variation on thermogenic adipocyte activity in mice and humans. While C57BL/6 J (B6) is the most commonly used inbred mouse strain in metabolism research, this strain is more prone to diet-induced obesity than other strains such as 129X1/SvJ (129). Consistently, the 129 strain is known to exhibit higher browning capability compared to B6. When stromal vascular fraction (SVF) from inguinal WAT of B6 and 129 mice were isolated and induced to differentiate into thermogenic adipocytes by established protocols (60), B6- and 129-derived cells exhibited a comparable degree of adipocyte differentiation. Nevertheless, 129-derived cells showed significantly higher Ucp1 expression. Comprehensive genomics, epigenomics and transcriptomics of adipocytes from B6, 129 and F1 offspring of these two strains were performed to identify cis- and trans-acting effects of genetic variation on Ucp1 expression that underlie phenotypic diversity. Allelic imbalance analysis of ChIP-seq with optimized protocol (61) in F1 cells revealed that a cis-regulatory variant rs47238345 is responsible for differential Ucp1 expression. The alternative T allele of rs47238345 at the Ucp1 -12 kb enhancer in 129 facilitates the allele-specific binding of NFIA to mediate allele-specific enhancer–promoter interaction and Ucp1 transcription. Moreover, CRISPR-Cas9/Cpf1-mediated SNP editing of rs47238345 from reference C (B6) to alternative T (129) in the B6 background resulted in increased Ucp1 expression (62).

Transcriptional Control of Brown Adipocyte Function by NFIA

Mice with transgenic expression of 3xFLAG-tagged NFIA in adipocytes under the control of the −5.4 kb promoter/enhancer of the Fabp4 gene (NFIA-Tg) exhibited limited weight gain, increased whole-body energy expenditure and improved glucose tolerance relative to WT mice under high-fat diet. Cell-based respirometry analysis revealed that both Ucp1-dependent and independent respiration was significantly upregulated in NFIA-Tg adipocytes than WT adipocytes. Notably, both Ucp1-dependent and independent adaptive thermogenesis relies on mitochondrial oxidative phosphorylation (Ox-Phos). Indeed, gene set enrichment analysis of RNA-seq dataset revealed that the genes most significantly and coordinately upregulated by NFIA-Tg were associated with gene sets such as Ox-Phos, adipogenesis and fatty acid metabolism. NFIA upregulates genes involved in Ox-Phos and brown fat-specific genes by enhancer activation that involves facilitated genomic binding of PPARγ. On the other hand, genes that were the most significantly and coordinately downregulated by NFIA-Tg were associated with gene sets such as allograft rejection and inflammatory response. NFIA in adipocytes, but not in macrophages, downregulates inflammatory cytokine gene expression both at the basal state and upon LPS treatment, in a cell-autonomous manner. ChIP-seq and luciferase reporter assay revealed that NFIA binds to the regulatory region of the Ccl2 gene, which encodes proinflammatory cytokine monocyte chemoattractant protein-1 (MCP-1), to downregulate its transcription and to ameliorate adipose tissue inflammation. Collectively, NFIA-mediated reciprocal regulation of thermogenesis and adipose tissue inflammation together contribute to improvement of glucose homeostasis and obesity (63) (Fig. 3).

Fig. 3.

Transcriptional control of brown adipocyte function by NFIA. Mice with transgenic expression of NFIA in adipocytes exhibited improved glucose tolerance and limited weight gain. NFIA upregulates Ox-Phos and thermogenesis by enhancer activation that involves facilitated genomic binding of PPARγ. In contrast, NFIA downregulates proinflammatory cytokine genes to ameliorate adipose tissue inflammation.

Perivascular adipocytes, which account for the majority of the outer matrix surrounding systemic blood vessels, undergo anti-inflammatory changes in response to vascular injury and thus fine-tune vascular remodelling as a protective mechanism. In a mouse model of endovascular injury, perivascular adipocytes exhibit browning-like phenotypic changes to attenuate inflammation and pathological vascular remodelling. While Ucp1 is dispensable for anti-inflammatory phenotype of perivascular adipocytes, a humoral factor neuregulin 4 (Nrg4), which critically regulates alternative macrophage activation, is indispensable for the observed phenotype. Notably, NFIA is required for optimal expression of Nrg4 in perivascular adipocytes (64). These results further support the importance of anti-inflammatory function of NFIA in maintaining systemic homeostasis. The effects of NFIA and other representative transcriptional regulators on the differentiation and function of white, beige and brown adipocytes are summarized in Fig. 4.

Fig. 4.

Effect of representative transcriptional regulators on white, beige and classical brown adipocytes. The effects of NFIA and other representative transcriptional regulators on the differentiation and function of white, beige and classical brown adipocytes.

Future Perspectives

The significance of brown adipocytes in adaptive thermogenesis has been well appreciated. While NFIA has shown to be sufficient to upregulate Ox-Phos and brown fat-specific genes by enhancer activation that involves facilitated genomic binding of PPARγ, the upstream signals that activate NFIA remain elusive. Identification of such signals might be a key to improve protocols for differentiating mesenchymal precursors or induced pluripotent stem cells into brown and beige adipocytes (65–68) towards reprogramming-based therapies and drug discovery. In this regard, it has been reported that Ebf2-positive precursors can be isolated from the murine dorsal anterior region as early as embryonic day 14.5, and these cells differentiated into brown adipocytes in vitro (46). Temporal expression patterns and hierarchical relationships among NFIA, EBF2 and other transcription factors during embryonic brown fat development remain to be elucidated. Moreover, although ChIP-seq analyses have clearly demonstrated that NFIA facilitates the recruitment of PPARγ to brown fat-specific enhancers, further studies—particularly through biochemical approaches—are required to elucidate the underlying molecular mechanisms. NFIA may function by recruiting chromatin remodelling complexes (69) or by acting as a pioneer factor that promotes chromatin structural changes to enhance accessibility (70). Furthermore, NFIA’s role in suppressing adipose tissue inflammation highlights a novel function of brown and beige adipocytes in maintaining glucose homeostasis. In this regard, PRDM16 has also reported to repress the type I interferon response pathway in adipocytes (71). The interplay between NFIA, PRDM16 and other transcriptional regulators of brown and beige adipocytes in mediating the suppression of adipose tissue inflammation remains to be elucidated. Given the suspected and substantial effect of genetic variation on human brown adipocyte activity, a cis-effect of genetic variation on UCP1 and other genes—mediated by allele-specific binding of NFIA—would be a compelling area for future investigation. Ongoing biochemical, genetic and pathophysiological investigations into the mechanisms of NFIA are anticipated to advance the field of adipocyte biology and facilitate the development of novel therapeutic strategies for metabolic diseases.