In mammals, two broad categories of ATs i.e., WAT and BAT, function antagonistically to control systemic energy homeostasis (29, 30). The anabolic WAT collects, stores, and releases energy in the form of lipids (31), whereas the catabolic BAT oxidizes lipids to produce heat – a process known as adaptive thermogenesis (32–34). Functionally distinct from WAT and BAT, the bone marrow adipose tissue [also referred to as “yellow adipose tissue” or “marrow adipose tissue” (MAT)] is a key regulator of bone metabolism (35, 36). Although less well-documented than WAT and BAT, MAT has also been acknowledged as contributing to systemic energy metabolism regulation (37–40).

WAT is the most abundant form of ATs. It can be subdivided into subcutaneous (SCAT) and visceral (VAT) adipose tissue (4, 41, 42), although smaller depots are scattered throughout the body, surrounding organs and lymph nodes ( Figure 1A ). SCAT and VAT differ significantly with regard to their cellular, molecular, and physiological characteristics, and oppositely contribute to the metabolic syndrome (2, 47, 48). Indeed, while increased amount of SCAT is associated with improved insulin sensitivity (49), VAT accumulation is commonly associated with increased insulin resistance, high-risk of type 2 diabetes, and high mortality (50, 51). SCAT is located beneath the skin, and should be distinguished from the less-characterized dermal WAT (DAT), which resides directly below the reticular dermis (i.e., above the SCAT), and is involved in thermal insulation, hair regeneration, wound healing and protection against skin infections (17, 18, 52–55). VAT includes intrathoracic fat depots (surrounding the heart: epicardial and pericardial fat, situated within the mediastinum: mediastinal fat), and intraabdominal fat depots (mesenteric, omental, perigonadal, perirenal and retroperitoneal fat). In addition, under certain physiological (e.g., aging) or pathological (e.g., metabolic syndrome) conditions, fat depots can also develop around or within skeletal muscles (56) and in liver (57, 58).

Compared to WAT, BAT is located in more specific areas (41, 44, 45). In rodents, the largest and most investigated BAT depot is the subcutaneous interscapular BAT, but BAT depots can also be found in the supraclavicular region of the neck (46), and in perirenal region (45) ( Figure 1B ). In humans, the supraclavicular region contains the highest proportion of total body BAT volume, followed by the mediastinal, thoracic paravertebral, perinephric, and adrenal loci (59). While previously believed to be nonexistent or nonfunctional in adult humans, several reports provide unequivocal evidence of the presence and activity of BAT in this population (60, 61).

In stark contrast to WAT and BAT, MAT has a specific location in the bone marrow of certain long bones and vertebrae, where it resides side-by-side with hematopoietic and bone cells (36, 40, 62, 63). MAT includes two distinct subtypes that have different locations in bones – the “constitutive MAT”, concentrated in the distal skeletal bones, and the “regulated MAT” that is diffusely distributed in the spine and proximal limb bones, and is regulated by several environmental factors (40, 64).

A momentous property of ATs is their high degree of plasticity. To meet the organism’s needs, ATs’ size, metabolism, structure, and phenotype can change rapidly (25–27, 65).

The potential for ATs to grow and regress in size is substantial. WAT can quickly expand its volume within days after initiation of an obesogenic diet through an increase in the size (hypertrophy, due to enhanced lipid storage capacity) and/or the number (hyperplasia) of adipocytes, whereas it rapidly shrinks by lipolysis upon fasting or cold exposure (66–69). This illustrates (i) the metabolic plasticity of WAT i.e., the ability to switch between two opposing metabolic programs: nutrient storage (lipogenesis) vs. nutrient release (lipolysis), and (ii) the cellular plasticity of WAT since newly generated adipocytes can be recruited during obesity development and relapse, as well as during cold exposure (65, 70–73).

As observed in obesity, aging is also associated with expanded WAT mass (resulting from decreased SCAT and DAT mass, and marked increased VAT mass), and reduced BAT mass and activity (74–77). At the molecular level, transcription control through forkhead box protein A3 (FOXA3) has been proposed as a potential regulatory factor for WAT accretion and BAT decline during aging and obesity (78). In addition, it has been reported that aging and obesity are also associated with increased MAT mass (79, 80). Most importantly, the age- and obesity-related changes in ATs’ mass and location are associated with progressive dysfunction of these tissues, ultimately leading to systemic inflammation and occurrence of metabolic disorders; this defines the concept of “adipaging” (77, 81, 82).

Another striking example of ATs’ remarkable metabolic plasticity is that, under various physiological, pathological and pharmacological conditions, WAT can acquire oxidative BAT-like features and, conversely, that BAT can convert into WAT-like tissue, in processes respectively termed “WAT browning” (or “beiging”) (6, 43) and “BAT whitening” (83). Thermogenic adaptation of ATs has been initially described during environmental cold exposure: BAT mass increases together with elevated expression of thermogenic genes (84). In WAT, especially in rodents, cold exposure induces the development of thermogenic beige adipocytes within the tissue (84, 85). Results from several studies further support that “inducible” brown adipocytes (aka “beige” or “brite”) can emerge within classical WAT depots in response to cold acclimatation and other stimuli such as β 3-adrenergic stimulation (28, 86). The exact mechanism is currently unclear (41, 87). Some have proposed that the browning/beiging of WAT is a result of de novo production of beige adipocytes (88), while other studies support that beige adipocytes derive from preexisting adipocytes (89). It is noteworthy that both aging and obesity are associated with impaired WAT browning (75, 90), and enhanced BAT whitening (91). Very few studies have focused on mechanisms involved in BAT whitening. For optimal thermogenic activity, BAT requires an extensive vascularization to ensure efficient supplies of oxygen needed to support the high energy consumption (92). It has been proposed that BAT whitening may partly rely on decreased vascularity of the tissue, leading to functional hypoxia and decreased thermogenic activity (93). Moreover, beige adipocytes can transform into energy-storing white adipocytes within days after external stimuli are withdrawn (94, 95).

The biomedical interest in beige AT – now viewed as the third type of ATs, is currently centered on the capacity of this oxidative tissue to counteract obesity via induction of energy expenditure, and to mitigate the vast array of obesity-associated diseases, including type 2 diabetes, heart disease, insulin resistance, hyperglycemia, dyslipidemia, hypertension, and many types of cancer (96, 97).

The lymphatic system is distributed throughout the body and consists of lymphoid nodes and organs, and lymphatic vessels (105, 106). The lymphatic system forms a one-direction transit pathway from the extracellular space toward the venous circulation to maintain fluid homeostasis by removing the protein-rich lymph from the extracellular space among tissues and returning it to the bloodstream (107). In addition to its fluid homeostasis function, the lymphatic system is also important for transport of pathogens [bacteria, viruses, prions (108)], antigens, exosomes and immune cells (including antigen-presenting cells) to regional lymph nodes and lymphoid structures, and management of immune cell trafficking and inflammation (107, 109–114). Besides, it has been reported that lymphatic endothelial cells regulate immune responses more directly by controlling entry of immune cells into lymphatic capillaries, presenting antigens on major histocompatibility complex proteins, and modulating antigen presenting cells (115).

In humans, lymphatic density varies greatly from AT depot to depot, but local lymphatic vessel function may still impact local adipose health (116, 117). In mice, lymphatics are absent in BAT and present in WAT (rare in gonadal VAT, sparse in SCAT). For instance, SCAT lies in proximity to the dermal lymphatic vasculature, and VAT surrounds the collecting lymphatic vessels of the mesentery, cisterna chyli and thoracic duct, as well as the efferent and afferent lymphatic vessels of intra-abdominal lymph nodes (118).

In addition, a close relationship exists between ATs and lymph nodes – the organizing centers of immune surveillance and response. Indeed, even in the leanest animals, lymph nodes are always found surrounded by WAT (119). Extensive work by Pond and colleagues further revealed the bidirectional functional partnership connecting WAT and lymph nodes: through increasing its rate of lipolysis, perinodal WAT serves as a reservoir of energy that is deployed to power local immune responses while, vice-versa, chronic lymph node activation results in increased WAT mass (119–124).

Despite a rather simple histological appearance, ATs’ cellular composition is complex (125). Although contributing to more than 90% of ATs’ volume, mature, lipid-filled adipocytes represent less than 50% of adipose cells (126). Other cell types in ATs (collectively referred to as stromal-vascular cells) include e.g., multipotent mesenchymal progenitor cells, preadipocytes, and a broad spectrum of innate and adaptive immune cells – which all spatially and functionally interact. Interest in adipose immune cells was significantly accelerated by the discovery that ATs’ immune cell composition is highly sensitive to metabolic and nutritional states (7, 8); this exemplifies another aspect of ATs’ cellular adaptability. Although mainly described for WAT, in which they contribute to the regulation of systemic metabolism by modulating the inflammatory tone of the tissue, immune cells are also present in thermogenic ATs (i.e., BAT and beige adipose tissue) where they have been proposed to support and regulate tissue remodeling and thermogenic function (12–14, 127, 128).

AT’s innate and adaptive immune cell types and functions have been extensively documented, mainly for WAT and predominantly in the context of obesity (7–10, 129). In the lean state, WAT contains anti-inflammatory M2-like macrophages, regulatory T cells (T_reg cells), type 2 innate lymphoid cells (ILC2s), invariant natural killer cells (iNKT cells), natural killer cells (NK cells), eosinophils, dendritic cells (DCs), and γδ T cells, which all cooperate to prevent inflammation and coordinate metabolic responses. In the obese state, the profile of adipose immune cells shifts towards a proinflammatory-type of cells: the number of neutrophils, inflammatory M1 macrophages, mast cells, B cells, DCs, CD8⁺ T cells, T helper (Th) 1 cells and Th17 cells increases, while the number of eosinophils, iNKT cells, ILC2s and T_reg cells decreases (9, 129). Interestingly, in the context of mouse and human obesity it has been reported that immune cell composition differ between SCAT and VAT (130–132).

A summary of immune cells identified in ATs is presented in Table 1 .

In addition, it has to be mentioned that some intraabdominal and intrathoracic VAT depots are rich in immune cell clusters called fat-associated lymphoid clusters (FALCs). FALCs correspond to inducible, atypical lymphoid tissues that were first identified in the omentum where they were termed “milky spots” (158). Later, the presence of FALCs was reported in other VAT depots (i.e., mesenteric, mediastinal, gonadal and pericardial) (159–161). FALCs are in direct contact with adipocytes, lack fibrous capsule (unlike lymph nodes), and contain B cells, T cells, macrophages, dendritic cells, NKT cells, iNKT cells and ILC2s (159). Importantly, FALCs can give rise to germinal centers under certain conditions (162). These small clusters of immune cells behave as a secondary lymphoid organ and are responsible for modulating both innate and adaptive immune responses (163). The presence of FALCs, together with the aforementioned resident adipose immune cells, have led some authors to propose that ATs can be considered as a tertiary lymphoid organ, with hallmarks of innate and adaptive immune responses (164).

The mechanisms by which ATs contribute to immune responses may be via direct effects of adipose immune cells, and/or via indirect effects whereby adipocytes modulate immunity and inflammation through endocrine, paracrine, autocrine or juxtacrine mechanisms of action. As endocrine organs, ATs communicate with other organs (including immunological organs) via the synthesis and secretion of a multitude (>600) of molecules typically referred to as “adipokines”, such as TNFα, IL-6, chemerin, resistin, visfatin, vaspin, irisin, omentin-1, lipocalin-2, apelin, adiponectin and leptin, to name but a few (15, 50, 165–167). In ATs, adipokines can be released by adipocytes, preadipocytes, adipose resident and -infiltrated immune cells, or other cell types. In addition to their regulation of ATs’ metabolic homeostasis, some adipokines have immunoregulatory and inflammatory functions. For example, adiponectin has anti-inflammatory properties (168) and can negatively regulate macrophage function (169). In contrast, leptin is a proinflammatory factor (170) with broad actions on both the innate (activation of monocytes/macrophages, neutrophils and NK cells) and adaptive (promotion of CD4⁺ T cell proliferation and IL-2 secretion) immunity (171, 172).

In addition, accumulating evidence indicates that adipocytes can behave as immune cells and sense inflammatory cues; thereby playing an important role in shaping immune responses. Indeed, adipocytes have the ability to express innate immune receptors (173), produce proinflammatory cytokines (50), express chemokines (174), and present antigens (175–177). Most recently, Caputa et al., reported that adipocytes can be licensed by adipose innate immune cells (i.e., NK cells and iNKT cells) to acquire anti-bacterial functions during Listeria monocytogenes infection; this highlights the extra-metabolic capacity of adipocytes to actively participate in the immune response to bacterial infection (178). Such fascinating findings highlight the intricacy and potential for adipocytes to robustly modulate ATs and systemic inflammation that in turn impact global immune responsiveness under homeostatic and disease state conditions.

Pathogens that have been reported to infect ATs, and eventually persist within, are listed in Table 2 .

Tissue-resident memory CD8⁺ T cells (T_RM) are a distinct memory population that is generated and persists at the infection site (275–277). Upon exposure to the same (or similar) pathogen, T_RM cells provide a first line of adaptive cellular defense and are crucial in lethal challenge models (278). In mice, non-human primates and humans, it has been shown that ATs (only reported for WAT) harbor many T_RM cells (279). Interestingly, WAT’s T_RM cells metabolically adapt to the tissue through up-regulating genes involved in lipid metabolism, indicating that they use fatty acid metabolism for their homeostasis (280).

So far, only few reports have underscored that ATs act as reservoirs for CD8⁺ T_RM cells: after infection (VAT) (193, 214), dietary restriction (MAT) (281), and obesity (VAT) (214, 282, 283), thereby likely participating to adaptive immune responses.

Han and colleagues reported that, following T. gondii or Y. pseudotuberculosis infection, nearly half of all CD8⁺ T cells within WAT are specific for the administered pathogen, suggesting that CD8⁺ T cells home to and persist (as CD8⁺ T_RM cells) within WAT after infection (193). Importantly, the authors also demonstrated that the induction of memory T cell responses in WAT results in the remodeling of WAT’s physiology in favor of the activation of antimicrobial response at the expense of lipid metabolism (193). These observations were the first to identify WAT as immune reservoirs sustaining pool of memory T cells and facilitating their reactivation, thus potentially contributing to immunological memory after infection. Whether WAT’s T_RM cells are important for recall responses to T. gondii and/or Y. pseudotuberculosis reinfections in humans remains to be seen.

Misumi and colleagues brought the second demonstration of an active anti-infectious role of ATs’ memory T cells (214). Following lymphocytic choriomeningitis virus infection, the authors showed that more memory IFNγ⁺ CD8⁺ T cells accumulate in the WAT of obese mice than in the WAT of lean mice. Strikingly, WAT memory T cells of obese mice rapidly caused lethal immunopathology upon re-challenge infection, whereas lean mice remained unaffected and could efficiently control the challenge (214). This demonstrates that obesity gives rise to an unusual form of T-cell-mediated pathogenesis during viral infection that leads to lethality.

Altogether, these major findings illustrate the importance of AT’s T cells to infection and in memory/recall responses and may have implications for the management of infections in individuals having alterations in AT’s immune cell composition and function, such as individuals with obesity and aged adults, who display impaired immune responses to pathogens.

In addition, unlike stereotypical lymphoid organs that primarily contain adaptive immune cells, ATs are enriched with several types of innate immune cells (e.g., macrophages, DCs), as well as with ILCs and innate-like T cells (e.g., iNKT cells, γδ T cells) ( Table 1 ), which are likely contributing to host defense against infection. Indeed, adipose iNKT cells and γδ T cells have been reported to mediate host defense by modulating the number and/or function of other adipose immune cells (e.g., Tregs, DCs, NK cells and macrophages) (284), and DCs were shown to control adaptive immune responses to AT’s infection (133, 134, 285, 286).