Lasse K. Markussen, Maja W. Andersen, Susanne Mandrup, Center for Functional Genomics and Tissue Plasticity (ATLAS), Functional Genomics & Metabolism Research Unit, University of Southern Denmark. s.mandrup@bmb.sdu.dk

White and brown adipocytes play different roles in metabolism and physiology. However, both cell types appear to be plastic with an ability to switch between metabolic states in response to changes in the cellular environment. Hence, the separation between brown and white adipocytes is less clear than previously thought. 

Adipocytes are mesenchymal cells specialized in storing large quantities of lipids in the cytosol as lipid droplets. They are found in adipose tissues such as those found subcutaneously or intra-abdominally, as well as in many smaller depots within or surrounding organs. Examples include the adipose tissue in the bone marrow and the adipose tissue associated with the kidneys and the adrenal gland. In addition, adipocytes can be found interspersed in other tissues, e.g. in muscle fibers and thymus. While the general biochemical function of all adipocytes is related to lipid storage, the specific physiological functions of adipocytes in the different tissues and their cross-talk with other resident cell types are not well understood. Interestingly, an increasing body of evidence indicates that adipocytes display a remarkable plasticity in response to the specific tissue niche as well as physiological stresses such as under- and over-nutrition, environmental temperature, exercise, and age.

White and brown adipocytes

Traditionally, adipocytes have been classified as white or brown. White adipocytes are large unilocular cells specialized in storage and release of fatty acids. Endocrine signals derived from these adipocytes provide continuous information to the brain as well as peripheral tissues on the degree of filling of adipocyte stores. Brown adipocytes contain many smaller lipid droplets (multilocular) and many mitochondria that are partially uncoupled owing to high expression of the uncoupling protein 1 (UCP1). As a result, these adipocytes are specialized in oxidation rather than release of fatty acids and in conversion of metabolic energy into heat (i.e. non-shivering thermogenesis). The major activator of non-shivering thermogenesis is believed to be cold exposure, which leads to release of norepinephrine from the sympathetic nervous system (SNS) and activation of β-adrenergic receptors on brown adipocytes.

Brown-like adipocytes – a new cell type?

More recently a third type of adipocyte, the brown-like (or beige, brite) adipocyte, has been shown to be induced in certain white adipose tissues (WAT) in rodents in response to physiological stresses such as cold (Figure 1).Similar to brown adipocytes, brown-like adipocytes are rich in mitochondria that express UCP1 and can perform non-shivering thermogenesis. During the recent years, these adipocytes have attracted considerable research interest because of their potential to increase WAT energy metabolism, which may increase whole-body energy expenditure and promote clearance of glucose and fatty acids from the plasma. Notably, mice lacking brown adipose tissue (BAT) become obese, while BAT transplantation reverses obesity in obese mice, indicating that, at least in rodents, BAT contributes significantly to organismal energy expenditure. Thus, the hope is that by recruiting brown-like adipocytes in humans one could combat obesity or the metabolic consequences thereof. So far more than hundred signals/factors promoting browning of white adipocytes in rodents have been identified (Figure 1). In humans, browning of WAT has been demonstrated in patients with pheochromocytoma, i.e. with chronically high levels of norepinephrine, as well as in cancer cachexia. These examples show that browning of human WAT is possible; however, the extent to which browning of WAT occurs in humans during non-diseased physiological conditions is currently not clear.

One of the muchdebated questions in the field is whether brown-like adipocytes should be considered brown adipocytes, or whether they represent a distinct cell type. It has been argued that brown-like adipocytes express specific brown-like defining marker genes; however, none of these marker genes have been linked to important biochemical functions and many appear to derive from anatomical position. Furthermore, no markers are unique for any single adipocyte type (brown, white or brown-like) (1). Importantly, brown-like adipocytes appear to perform the same biochemical functions as classical brown adipocytes and to be activated by the same browning agents. In this respect, it is also worth noting that the sympathetic neuronal pathways stimulating WAT and BAT have identical origins in the brain. Thus, the distinction between brown and brown-like adipocytes appears somewhat artificial.

Another controversial question is which cells give rise to the brown-like adipocytes, i.e. do they primarily arise by phenotypic switching of existing white adipocytes into brown-like adipocytes (transdifferentiation), or are they recruited by de novo differentiation of progenitors? Lineage-tracing studies, where cells are permanently “labelled” due to their expression of a particular lineage-specific gene, have recently provided evidence that both pathways play a role in browning of rodent WAT and that the relative importance of these pathways depends on the tissue and the environmental challenge. Notably, however, such lineage-tracing studies are inherently biased towards cells that express one particular gene, and further insight into this question awaits future more unbiased approaches such as single-cell sequencing.

Adipocyte heterogeneity and plasticity           

A key question that to date remains unanswered is whether the ability to undergo browning is restricted to a subpopulation(s) of progenitor cells and mature adipocytes in WAT, or whether every white adipocyte has the ability to differentiate into a brown-like adipocyte. In support of adipocyte heterogeneity, a recent study identified two adipocyte populations in subcutaneous adipose tissue in mice – one brown-like subpopulation that expressed high levels of medium-chain acyl-CoA dehydrogenase (MCAD) and UCP1 but low levels of fatty acid synthase (FASN), and another white subpopulation that expressed high levels of FASN, and low levels of MCAD and UCP1 (2). Furthermore, another recent study observed that around 50% of former UCP1-positive brown-like adipocytes became UCP1-positive again when re-exposed to cold after an intermediate whitening period (3), suggesting that adipocytes that once were brown-like constitute a subpopulation with an epigenetic memory of being brown.

The above reports clearly support adipocyte heterogeneity of WAT that has been exposed to browning signals. The question is how this heterogeneity arises in the first place. Interestingly, after re-exposure to cold temperatures, brown-like adipocytes primarily localize to islands of brown-like adipocytes within the WAT, indicating that the microenvironment is a crucial factor for determining adipocyte fate. In particular, the proximity to norepinephrine fibers might be a determining factor, since these fibers correlate with the number of brown-like adipocytes, and since the lack of β₃-adrenoreceptors depressed browning (4). Thus, the apparent heterogeneity may arise as a result of proximity to norepinephrine fibers.

Interestingly, during conditions of thermoneutrality or chronic positive energy balance, brown and brown-like adipocytes convert to white adipocytes, a process called whitening. In addition, in several mammals as well as humans there is a general decrease in the amount of BAT following weaning, during aging, and in response to development of obesity. This indicates that the thermogenic phenotype of brown and brown-like adipocytes is reversible and requires sustained β-adrenergic stimulation to maintain the thermogenic signature.

Thus, depending on how one defines a cell type, one could view brown-like adipocytes as products of cellular plasticity of progenitors and mature adipocytes, and the interconversion of white to brown-like adipocytes as phenotypic transitions allowing the organism to adapt to physiological stresses. These emerging intricacies call for a revision of the classification of adipocytes.

Progenitor heterogeneity and plasticity

Intriguingly, progenitors isolated from different adipose depots retain their depot-specific characteristics in culture (5), suggesting that the microenvironment has preprogrammed the progenitors to a particular adipocyte sub-lineage and that this preprogramming is relatively stable. The question is whether there in a given WAT are distinct populations of white and brown-like progenitor cells, or whether progenitors are mostly bipotent with cellular fate decisions being determined stochastically or by conditional signals. Subclonal analyses of cells from the subcutaneous WAT suggested that there are distinct subpopulations of brown-like and white progenitors. Interestingly, however, there is also evidence of bipotent progenitor cells that have the potential to differentiate into brown-like adipocytes upon β-adrenergic stimulation and into white adipocytes following feeding with high-fat diet. These results indicate that despite differential preprogramming of progenitor cells in the different adipose tissues, at least some progenitor cells display plasticity in response to the appropriate signals.

Interplay between plasticity and heterogeneity 

The above discussion highlights that adipocyte phenotype is the result of the balance between fate-conserving and plasticity-inducing mechanisms. According to the more classical model (Figure 2A),dedicated white progenitor cells exclusively differentiate into white adipocytes, brown progenitor into brown adipocytes and brown-like progenitor cells into brown-like adipocytes. According to this model, brown-like adipocytes convert back to dormant “white” adipocytes when thermogenesis is no longer a priority. The dormant “white” adipocytes and classical white adipocytes are viewed as different cell types defined by specific marker genes, as discussed above. The second model (Figure 2B) proposes that although progenitor cells and mature adipocytes are preprogrammed toward a specific lineage, they still retain some degree of plasticity and can change this preprogramming in response to physiological signals. At this point none of the models can be excluded, but data to support the second model are accumulating.

Concluding remarks and future perspectives

Thermogenic adipocytes have been intensively studied during the recent years because of their alleged potential to counteract the development or consequences of obesity. However, there are still many open questions regarding the origin, plasticity and regulation of these cells. These include questions such as: What drives thermogenic adipocyte plasticity? What niche-specific signals regulate plasticity? And are there fundamental functional differences between the brown and brown-like thermogenic program, or do brown-like adipocytes represent shades of brown? Resolving these fundamental questions will require a combination of the use of intelligent lineage-tracing mouse models, single cell sequencing and advanced molecular studies.  

References

1.         Roh HC, Tsai LTY, Shao M, Tenen D, Shen Y, Kumari M, et al. Warming Induces Significant Reprogramming of Beige, but Not Brown, Adipocyte Cellular Identity. Cell metabolism. 2018;27(5):1121-37.e5.

2.         Lee YH, Kim SN, Kwon HJ, Granneman JG. Metabolic heterogeneity of activated beige/brite adipocytes in inguinal adipose tissue. Scientific reports. 2017;7:39794.

3.         Rosenwald M, Perdikari A, Rulicke T, Wolfrum C. Bi-directional interconversion of brite and white adipocytes. Nature cell biology. 2013;15(6):659-67.

4.         Jimenez M, Barbatelli G, Allevi R, Cinti S, Seydoux J, Giacobino JP, et al. Beta 3-adrenoceptor knockout in C57BL/6J mice depresses the occurrence of brown adipocytes in white fat. European journal of biochemistry. 2003;270(4):699-705.

5.         Siersbaek MS, Loft A, Aagaard MM, Nielsen R, Schmidt SF, Petrovic N, et al. Genome-wide profiling of peroxisome proliferator-activated receptor gamma in primary epididymal, inguinal, and brown adipocytes reveals depot-selective binding correlated with gene expression. Molecular and cellular biology. 2012;32(17):3452-63.