ABSTRACT The conversion of white adipocytes to thermogenic beige adipocytes represents a potential mechanism to treat obesity and related metabolic disorders. However, the mechanisms

involved in converting white to beige adipose tissue remain incompletely understood. Here we show profound beiging in a genetic mouse model lacking the transcriptional repressor Krüppel-like

factor 3 (KLF3). Bone marrow transplants from these animals confer the beige phenotype on wild type recipients. Analysis of the cellular and molecular changes reveal an accumulation of

eosinophils in adipose tissue. We examine the transcriptomic profile of adipose-resident eosinophils and posit that KLF3 regulates adipose tissue function via transcriptional control of

INGUINAL SUBCUTANEOUS WHITE ADIPOSE TISSUE Article Open access 28 September 2020 AGE-DEPENDENT PDGFRΒ SIGNALING DRIVES ADIPOCYTE PROGENITOR DYSFUNCTION TO ALTER THE BEIGE ADIPOGENIC NICHE IN

MALE MICE Article Open access 31 March 2023 GPR180 IS A COMPONENT OF TGFΒ SIGNALLING THAT PROMOTES THERMOGENIC ADIPOCYTE FUNCTION AND MEDIATES THE METABOLIC EFFECTS OF THE

ADIPOCYTE-SECRETED FACTOR CTHRC1 Article Open access 08 December 2021 INTRODUCTION White adipose tissue (AT), typically regarded as an energy storage site, can acquire the thermogenic

properties of brown AT to become ‘beige’ and drive energy expenditure1,2. The discovery that AT can be activated in this way may be important in combatting obesity and associated

cardiometabolic disorders. It is now apparent that resident immune cells, particularly mediators of type 2 immunity, are involved in beige AT activation and energy expenditure. Eosinophils

have been implicated in beiging3,4,5 but the underlying cellular and molecular mechanisms orchestrating their contributions remain incompletely understood. Obese mice have fewer eosinophils

in their AT than lean counterparts, and the ΔdblGATA transgenic mouse, which lacks eosinophils altogether, displays exaggerated weight gain on a high-calorie diet4. Using the ΔdblGATA model,

it has also been proposed that AT eosinophils facilitate adipocyte maturation to ameliorate diabetic complications of diet-induced obesity6. Conversely, mice genetically engineered to

overexpress interleukin 5 (IL-5) have supraphysiological levels of eosinophils and are protected from diet-induced obesity4. However, a recent study reported that artificially increasing

eosinophils in AT with recombinant IL-5 did not confer the expected metabolic benefits, leading the authors to suggest that the functional activities of eosinophils may be more crucial than

their abundance7. Furthermore, the principal contribution of macrophages to AT energy expenditure—their production of catecholamines8—is currently under debate, with several groups finding

that rather than synthesising catecholamines, macrophages are instead responsible for catecholamine uptake and degradation9,10,11. Thus, interest in the key cellular and molecular drivers of

beiging remains intense. We have previously shown that mice lacking the transcriptional repressor Krüppel-like factor 3 (KLF3) have reduced adiposity and are protected from diet-induced

obesity12,13. Here we report enhanced beige AT activation in _Klf3__−/−_ mice and that bone marrow (BM) transplants from these mice confer the lean, beige phenotype on recipients. In the

absence of KLF3, AT-resident eosinophils are more abundant and exhibit significant deregulation of important secreted molecules, including meteorin-like and IL-33, both of which influence

beiging5,14,15,16. We also report that co-culture of eosinophils with primary adipocytes increases thermogenic gene expression. These findings identify KLF3 as an important regulator of AT

eosinophil gene expression and function, advancing our understanding of how these little-understood immune cells may lead to improved strategies for therapeutically driving energy

expenditure. RESULTS REDUCED ADIPOSITY AND ENHANCED BEIGING IN _KLF3_ −/− MICE Disruption of the gene encoding the transcriptional repressor KLF3 results in mice that are smaller than their

wild type (WT) littermates with less white AT13 (Supplementary Fig. 1a, b), and confers protection from diet-induced obesity12. Body mass composition analysis of chow-fed WT and _Klf3__−/−_

animals housed at room temperature (22 °C) showed that _Klf3__−/−_ mice exhibit reduced total fat mass compared to WT littermates (Fig. 1a and Supplementary Fig. 1a), in addition to

differences in lean body mass, which constitute their reduced body weight (Supplementary Fig. 1b). This is reflected in the reduced size of white AT depots in _Klf3__−/−_ mice seen

previously12 (Fig. 1b and Supplementary Fig. 1c). Visual examination of subcutaneous (subcut) AT depots, the depots most prone to beiging17, revealed a browner complexion and smaller size in

_Klf3__−/−_ mice (Fig. 1c). Furthermore, H&E staining revealed that in the absence of KLF3, adipose cellular architecture is notably altered, with enrichment of multilocular adipocytes

evident that was not seen in the subcut AT of WT mice (Supplementary Fig. 2a, b). These observations also confirm the previous finding that _Klf3__−/−_ mice have smaller-sized

adipocytes12,13. Given that thermogenic energy expenditure via activation of beige AT may influence adiposity18,19,20, we examined the expression of archetypal thermogenic genes. We observed

upregulation of numerous thermogenic genes in the subcut AT of _Klf3__−/−_ mice – most notably _Ucp1_, _Cpt1b_, _Fatp1_ and the beige-specific marker21 _Tbx1_ (Fig. 1d). We next

investigated the levels of mitochondrial proteins by Western blotting of whole-cell extracts (WCE) from WT and _Klf3__−/−_ subcut AT. Uncoupling protein 1 (UCP1) protein levels were higher

in the subcut AT of _Klf3__−/−_ mice (Fig. 1e), as were mitochondrial oxidative phosphorylation (oxphos) complexes I–V (Fig. 1f). We also observed increased levels of the mitochondrial outer

membrane protein voltage-dependent anion channel (VDAC) in _Klf3__−/−_ subcut AT (Fig. 1g), suggesting higher mitochondrial number. Levels of multiple thermogenic genes were also increased

in the gonadal AT of _Klf3__−/−_ mice (Supplementary Fig. 1d), as were UCP1 and mitochondrial oxphos proteins (Supplementary Fig. 1e, f). Several genes were modestly increased in

interscapular brown AT of _Klf3__−/−_ mice while beige-specific markers1 _Tbx1_, _Tmem26_ and _Cd137_ were undetectable (Supplementary Fig. 1g). UCP1 protein content was mildly decreased in

_Klf3__−/−_ brown AT (Supplementary Fig. 1h, i). While this suggests that brown AT is unlikely to play a major role in the thermogenic phenotype of _Klf3__−/−_ mice, we cannot wholly rule

out its contribution given the existence of UCP1-independent thermogenic mechanisms in beige and brown fat22,23,24,25. Together, these results show that _Klf3__−/−_ mice exhibit reduced fat

mass that may result from enhanced AT beiging, as demonstrated by the widespread de-repression of thermogenic genes and mitochondrial proteins. KLF3 DEFICIENCY ENHANCES THE THERMOGENIC

RESPONSE Cold exposure is a recognised means of promoting adaptive thermogenesis in AT26,27,28. To study whether the absence of KLF3 enhances beige AT activation and influences body

temperature both in the cold and at thermoneutrality, we acutely exposed WT and _Klf3__−/−_ mice to ambient temperatures of 4 °C or 30 °C (Fig. 2a). No statistically significant difference

was seen in the body temperatures of WT and _Klf3__−/−_ mice during cold exposure (Fig. 2b). As expected, there was little disparity in body temperature between WT and _Klf3__−/−_ mice

housed at 30 °C. Subcut (Fig. 2c and Supplementary Fig. 3a) and gonadal AT (Supplementary Fig. 3b, c) from _Klf3__−/−_ mice were smaller than WT depots under both conditions, and

interestingly, interscapular brown AT was smaller in _Klf3__−/−_ mice acutely exposed to 4 °C than in WT mice at 4 °C or in _Klf3__−/−_ mice at thermoneutrality (Supplementary Figs. 3d, e).

Visual inspection of subcut AT depots from mice housed at 30 °C and 4 °C showed that _Klf3__−/−_ mice exhibited distinct, enhanced evidence of browning at 4 °C compared to WT mice, and this

disparity was retained at 30 °C though to a lesser extent (Fig. 2d). This striking change in macroscopic appearance was surprising, given the short period of cold exposure (5 h), and may in

part be strain-dependent and due to the elevated levels of basal thermogenesis in _Klf3__−/−_ mice. To examine beige AT activation at the transcriptional level, we assessed expression of

thermogenic genes in subcut AT from mice housed at 30 °C and 4 °C. At 4 °C, we found that multiple genes were increased in the absence of KLF3, including _Ucp1_, _Pgc1a, Cidea_, _Elovl3_ and

_Cpt1b_ (Fig. 2e), with similar deregulation seen in gonadal AT (Supplementary Fig. 3f) and to a lesser extent, brown AT (Supplementary Fig. 3g). _Ucp1_ mRNA expression was elevated in

subcut AT from _Klf3__−/−_ mice compared to WT mice housed at 30 °C. Western blotting of subcut UCP1 and mitochondrial oxphos complexes confirmed upregulation at the protein level at 4 °C

(Fig. 2f, g). Thus, under thermal stress at 4 °C, activation of beige AT is amplified in _Klf3__−/−_ mice. At thermoneutrality, KLF3 deficiency still augments beige AT activation, suggesting

that the effects seen are not solely due to cold temperature, but that in this model beiging of white AT can occur independently of thermal stress. REDUCED WEIGHT GAIN IN WT MICE

TRANSPLANTED WITH KLF3−/− BM We have previously shown that mice lacking KLF3 are protected from obesity when fed a high-fat diet, with no difference exhibited in food intake on either a chow

or high-fat diet compared to WT mice12. While the mechanisms underlying had remained unclear, evidence of enhanced beiging in _Klf3__−/−_ mice now provides a possible explanation. The role

of haematopoietic cells in AT function and thermogenesis is now widely appreciated, and KLF3 is both active in haematopoietic lineages29,30,31,32,33 and highly expressed in a range of

haematopoietic cell types according to the Immgen34 and Haemopedia35,36 databases. We thus focused on a possible role for KLF3 in BM haematopoietic cells that could give rise to cells

occupying the AT SVF niche and regulate adipose function. We irradiated WT mice and transplanted them with BM from WT or _Klf3__−/−_ mice to produce WTWT and WT_Klf3−/−_ chimeras (Fig. 3a).

Chimeric mice and a control cohort of untransplanted WT and _Klf3__−/−_ animals were fed a high-fat, high-sugar Western diet for 11 weeks in order to investigate resistance to diet-induced

obesity and adiposity. Successful haematopoietic reconstitution was assessed by genotyping tail biopsies (recipient and donor cells) and BM (recipient cells only) (Supplementary Fig. 4a).

Prior to commencing the Western diet regimen, WT_Klf3−/−_ chimeras were heavier than WTWT counterparts, then over the 11-week Western diet period, WT_Klf3−/−_ chimeras gained modestly but

significantly less weight than WTWT counterparts (Fig. 3b and Supplementary Fig. 4b, c) and had less body fat (Fig. 3c). Importantly, WT_Klf3−/−_ chimeras gained less subcut AT mass compared

to WTWT animals (Fig. 3d and Supplementary Fig. 4d) and more prominent browning in WT_Klf3−/−_ subcut AT was detected by visual inspection (Fig. 3e). No such differences were observed in

the weights or appearance of gonadal or brown AT (Supplementary Fig. 4e–i). The expression of thermogenic genes in the subcut AT of chimeric mice was tested and _Ucp1_, _Cpt1b_ and _Cd137_

were found to be upregulated in animals that received _Klf3__−/−_ BM compared to mice that received WT BM (Fig. 3f). Several genes were also upregulated in WT_Klf3−/−_ gonadal and brown AT

(Supplementary Fig. 4j, k). We performed biochemical analysis on the peripheral blood of transplanted mice and showed that plasma triglycerides were significantly lower in WT_Klf3−/−_ mice

compared to WTWT, while no difference existed in total cholesterol (Supplementary Fig. 5a, b). We also detected reduced hepatic triglycerides in untransplanted KLF3-deficient mice

(Supplementary Fig. 5c), and observed no difference in liver cholesterol levels (Supplementary Fig. 5d). Together these results demonstrate that transplantation of irradiated WT mice with

KLF3-deficient BM was sufficient to confer metabolic benefits on recipient mice. This suggests that haematopoietic cells may be instrumental in driving beiging and conferring resistance to

obesity in this system. AT-RESIDENT EOSINOPHILS ARE ALTERED IN THE ABSENCE OF KLF3 Given that transplantation with KLF3-deficient haematopoietic cells promoted resistance to obesity, we

sought to identify the important cell types in this process. _KLF3_ is expressed in multiple tissues and cell lineages, including various haematopoietic cells34,35, and the FANTOM5 SSTAR

database that measures transcript levels indicates it is more abundant in human eosinophils than any other cell type37, consistent with its observed presence in murine eosinophils38

(Supplementary Fig. 6a). Given their proposed role in adiposity and thermogenesis3,4,5,39,40,41, we sought to assess the contribution of eosinophils in driving the thermogenic phenotype of

_Klf3__−/−_ mice. We first examined the abundance of eosinophils in subcut AT of WT and _Klf3__−/−_ mice. SVF cells were collected and analysed by flow cytometry (according to the eosinophil

gating strategy in Supplementary Fig. 7a). AT-resident eosinophils were 3-fold more prevalent in _Klf3__−/−_ subcut AT (Fig. 4a). We also sorted eosinophils from whole subcut AT to

determine the number of cells per gram and confirmed that eosinophil abundance is greater in _Klf3__−/−_ subcut AT (Fig. 4b). mRNA levels of the gene encoding the eosinophil surface marker

Siglec-F (_Siglecf_) were also 3-fold higher in _Klf3__−/−_ subcut AT (Fig. 4c). It is noteworthy that while eosinophils are more than 2-fold more abundant in _Klf3__−/−_ gonadal AT

(Supplementary Fig. 7b), no differences in the number of lung or spleen eosinophils were detected (Supplementary Fig. 7c, d). Eosinophils were also found to be over 2-fold more prevalent in

the subcut AT of WT_Klf3−/−_ chimeric mice than in WTWT counterparts on a Western diet (Supplementary Fig. 7e, f). We next compared the genome-wide expression profiles of eosinophils

isolated from subcut SVF by FACS and observed broad deregulation of gene expression in the absence of KLF3 (Fig. 4d), resulting in changes to important cellular pathways including

signalling, localisation, response to stimulus and immune system processes (Supplementary Fig. 6b). Further breakdown of the deregulated immune system processes in _Klf3__−/−_ eosinophils

revealed striking enrichment of various biological pathways such as leukocyte migration and myeloid cell homeostasis (Supplementary Fig. 6c). To explore the transcriptional network further

we interrogated the transcriptional activator _Klf1_, which operates in a regulatory system with _Klf3_ in erythroid cells32,42. _Klf1_ was very lowly expressed in AT eosinophils and showed

no difference between WT and _Klf3__−/−_ (Supplementary Fig. 6d, e), with no expression detected in whole AT or SVF (Supplementary Fig. 6f), suggesting that KLF1 does not play an important

opposing role to KLF3 in AT eosinophils. Inspection of the most significantly deregulated genes in KLF3-deficient adipose eosinophils revealed three genes encoding proteins with reported

roles in beige AT activation—meteorin-like (_Metrnl_)5, met-enkephalin (_Penk_)14,16,43 and IL-33 (_Il33_)14,15 (Fig. 4e). We have previously reported that levels of galectin-3, a proposed

eosinophil chemoattractant44,45, are elevated in _Klf3__−/−_ AT29. Here we found that the eosinophil chemoattractants eotaxin-1 and -2 (_Ccl11_ and _Ccl24_) were upregulated in

KLF3-deficient AT eosinophils (Fig. 4f), providing a possible explanation for the increased eosinophil abundance in _Klf3__−/−_ AT. Given that eosinophil abundance is elevated in _Klf3__−/−_

AT, we sought to quantify other immune cells in the stromal vascular compartment. We did not observe changes to macrophage abundance (M1 or M2), dendritic cells, CD19+ B cells or CD3+ T

cells (Fig. 4g, h and Supplementary Fig. 8a–e). T regulatory lymphocytes (Tregs) were more abundant in KLF3-deficient subcut AT (Fig. 4i and Supplementary Fig. 8a)—unsurprising given the

elevated expression of IL-3346,47,48,49. While no significant differences to the overall lineage− CD127+ CD25+ group 2 innate lymphoid cell (ILC2) population were observed, we did detect a

significant deficit of ILC2s expressing the IL-33 receptor (ST2) in _Klf3__−/−_ subcut AT (Fig. 4j and Supplementary Fig. 8a, f, g). To assess whether changes in the AT immune profile

affected cytokine signalling, we measured expression of various inflammatory and anti-inflammatory genes in the SVF of WT and _Klf3__−/−_ subcut AT. Expression of the classical inflammatory

marker _Tnf_ (encoding the pro-inflammatory molecule TNF-α) was reduced in the absence of KLF3, while the M2 macrophage marker _Arg1_ was increased (Supplementary Fig. 7g). Unexpectedly, we

detected increased expression of _Ifng_ (encoding the inflammatory cytokine interferon-γ) in _Klf3__−/−_, which has been shown to block themogenesis50. On the other hand, levels of _Il10_

were reduced in the absence of KLF3, in line with recent findings that IL-10 blockade protects against diet-induced obesity and elicits browning of AT51. Together these findings demonstrate

that resident eosinophils are more abundant in _Klf3__−/−_ AT and that these cells exhibit widespread deregulated gene expression, including higher levels of genes implicated in beige AT

activation. These changes are accompanied by modifications to the immune and cytokine landscape of subcut AT. KLF3 DIRECTLY REGULATES AT EOSINOPHIL GENE EXPRESSION Our finding that AT

eosinophils lacking KLF3 express higher levels of _Metrnl_, _Penk_ and _Il33_ led us to examine these genes more closely. Using qPCR analysis, we measured higher expression of _Metrnl_ and

_Il33_ in subcut SVF from _Klf3__−/−_ compared to WT mice (Fig. 5a). Likewise, we found that levels _Metrnl_ and _Il33_ were higher in WT_Klf3−/−_ subcut SVF than in WTWT (Fig. 5b).

Interestingly, _Penk_ expression was only elevated in subcut SVF from _Klf3__−/−_ mice fed a Western diet, with no differences seen in whole subcut AT or SVF from transplanted mice or from

mice fed a chow diet (Fig. 5a, b and Supplementary Fig. 8a, b). To assess levels of secreted meteorin-like and IL-33 in AT, we quantified their concentration in supernatant from AT explants

cultured for 2 h, using ELISA. We found that both meteorin-like and IL-33 were more highly concentrated in _Klf3__−/−_ AT supernatant, suggesting greater secretion of these proteins (Fig. 

5c, d). However, when we measured meteorin-like and IL-33 in peripheral blood plasma we found no difference (Supplementary Fig. 8c, d), which may suggest that these proteins are elevated

locally and have a paracrine effect confined to the AT microenvironment. To determine whether AT eosinophils are able to directly drive beiging of adipocytes, we performed co-culture

experiments. FACS-sorted eosinophils from WT and _Klf3__−/−_ subcut SVF were co-cultured with mature primary adipocytes for 4 h, and the expression of thermogenic genes measured by qPCR

(Fig. 5e). We observed modest but significant increases in the expression of several thermogenic genes in adipocytes co-cultured with WT eosinophils compared to media alone. Additionally,

several genes, including _Ucp1_, _Ptgs2_ and _Cpt2_, were more highly expressed in adipocytes co-cultured with _Klf3__−/−_ AT eosinophils than WT eosinophils. Together these results suggest

that AT eosinophils alone are able to drive beiging of adipocytes, at least in vitro, and that KLF3 may directly regulate gene expression in eosinophils. Given that in the absence of the

transcriptional repressor KLF3, _Metrnl_ and _Il33_ are upregulated and secreted in higher abundance within AT, we sought to determine whether KLF3 directly binds and regulates these genes.

Inspection of ChIP-Seq data from mouse embryonic fibroblasts (Accession No. GSE44748)52 showed distinct KLF3 binding at the _Metrnl_ promoter region, but little enrichment at _Penk_ or

_Il33_ (Fig. 5f). To assess whether KLF3 binds these genes in a relevant cellular setting we performed ChIP in WT and _KLF3__−/−_ EoL-1 cells, a human eosinophilic cell line that highly

expresses _KLF3_ (Supplementary Fig. 8e). _KLF3__−/−_ cells were generated by CRISPR/Cas9 non-homologous end joining following the introduction of a sgRNA targeting exon 3 of _KLF3_ (Fig. 

5g). After undertaking immunoprecipitations with an anti-KLF3 antibody or normal goat IgG, we performed qPCR using primers designed to amplify the promoter regions of _METRNL_, _PENK_ and

_IL33_, and to positive control regions based on known ChIP-Seq peaks (_SP1_, _METRNL_ −4.6 kb_, IL33_ −24.5 kb) and a negative control locus (_VEGFA_). We found KLF3 binding was evident at

the _METRNL_ and _IL33_ promoters compared to the negative control region _VEGFA_, and that signals in _KLF3__−/−_ samples were negligible, confirming the specificity of the anti-KLF3

immunoprecipitation (Fig. 5h). This indicates that KLF3 directly binds and regulates the expression of the genes encoding important secreted molecules in adipose-resident eosinophils.

DISCUSSION Type 2 immune cells are now recognised as key players in AT homeostasis and beiging. Much attention has been given to the contribution of macrophages4,8,53,54, however recent

findings suggest that AT macrophages may not perform their previously-defined direct contribution to adaptive thermogenesis via catecholamine production9,10,11. This realisation provides us

with an opportunity to re-evaluate the current model of type 2 immunity in adiposity and beiging by defining previously unknown or under-appreciated immune cells and the mechanisms

underpinning their contribution to AT homeostasis. A recent study found that AT immune cells secrete acetylcholine that elicits beiging, highlighting the importance of haematopoietic cells

in AT homeostasis55. Interestingly though, acetylcholine-producing cells were predominantly B cells, T cells and macrophages, leaving the contribution of other lineages, such as eosinophils,

unresolved. In this study, using a combination of in vitro approaches, transgenic mice and BM transplantation models, we have shown that the transcriptional repressor KLF3 directly

regulates genes encoding secreted factors in AT eosinophils. Our findings from co-culture experiments suggest that eosinophils may be able to directly drive activation of beiging in

adipocytes through paracrine signalling. This highlights the importance of eosinophils in our model of type 2 immunity in AT beiging and suggests that eosinophils may be able to activate

beige fat in an organismal setting, providing protection from obesity. Future work will seek to further characterise the in vivo contribution of eosinophils to adipose homeostasis and

thermogenesis, and to understand the relevance of these cells in human setting. BM transplantation studies have been utilised to explore the contribution of haematopoietic cells in adiposity

and beiging, including the metabolic response to caloric restriction56 and hypermetabolism following burn injury57. We cannot exclude the influence of other haematopoietic cells to the

reduced weight gain and improved metabolic parameters in WT_Klf3−/−_ mice, and while this work explores the contribution of eosinophils, considerable effort in the field is currently being

directed towards understanding the diverse roles of various immune cell populations in beiging. Although eosinophil infiltration has been implicated in beiging3,40, it is also apparent that

artificially increasing their abundance by administration of IL-5 does not appear to significantly improve major metabolic parameters in diet-induced obese mice7. The authors proposed that

eosinophil activity may be more important than total numbers in AT. Significantly, as well as observing increased eosinophil abundance in _Klf3__−/−_ AT, we found considerable deregulation

of biological pathways in these cells and upregulation of genes encoding key secreted factors that activate beige AT. Meteorin-like is secreted by exercised muscle and induced in AT

following cold exposure5, however, a comprehensive understanding of the AT-specific source has remained elusive. Recent studies have reported adipose stromal cells as the primary source of

IL-33, with only a small contribution potentially originating from hematopoietic cells48,49. Here we provide evidence that eosinophils may contribute a portion of local adipose IL-33, as

well as meteorin-like. While the absence of _Penk_ deregulation in chow-fed _Klf3__−/−_ mice and lack of evidence for direct binding by chromatin immunoprecipitation led us to emphasize

other putative target genes, we cannot rule out a direct role for KLF3 in the regulation of _Penk_, or a role for Met-Enk in contributing to the phenotype observed—potentially in other cell

types. Overall, our findings demonstrate a mechanism by which type 2 immune cells may influence beiging and adiposity—via KLF3-driven regulation of important eosinophil-derived factors that

we term ‘eosinokines’40, as shown in our proposed model (Fig. 6). Our results not only suggest that eosinophils may play an under-appreciated role as signalling cells in metabolism, but also

reveal that KLF3 is a key regulator of secreted molecules in eosinophils. Multiple KLF family members have been implicated in the control of metabolism58,59,60,61,62 but few clear unifying

models of action have emerged. KLF3, for instance, does not appear to regulate meteorin-like or IL-33 in other cell types, and indeed its role in eosinophils has previously been

unrecognised. Targeting molecules secreted by immune cells and their receptors may have the potential to therapeutically drive energy expenditure via AT beiging to combat obesity. Our

findings advance the search for secreted factors produced by eosinophils and other resident haematopoietic cells, and support the hypothesis that eosinophils and their secreted ‘eosinokines’

are relevant to the type 2 immune network in AT. METHODS ANIMAL HUSBANDRY All animal work was carried out in accordance with approval from the UNSW Animal Care and Ethics Committee

(Approval Nos. 12/150A, 16/5B and 16/141B), the Murdoch Children’s Research Institute Animal Ethics Committee (Approval No. A760) and the University of Sydney Animal Care and Ethics

Committee (Approval No. L02/7-2009/3/5054). Animals were housed in a specific pathogen-free environment at a constant ambient temperature of 22 °C and 50% humidity on a 12 h light–dark

cycle, with _ad libitum_ access to standard chow food and water, unless otherwise specified. Generation of global _Klf3__−/−_ mice on an FVB/NJ background has been previously reported13.

Age-matched WT and _Klf3__−/−_ male littermates derived from _Klf3__+/_− × _Klf3__+/_− crosses were used for all animal studies. Male mice were housed in cages containing bedding, nesting

material and enrichment with up to five individual animals, except for cold and thermoneutral experiments which were undertaken in empty cages containing singly housed mice. ACUTE

THERMONEUTRAL AND COLD TEMPERATURE EXPOSURE For acute temperature exposure experiments, WT and _Klf3__−/−_ mice aged between 12 and 14 weeks were assessed at 30 °C or 4 °C for 5 h. Mice

housed at thermoneutrality (30 °C) were acclimatised at this temperature for 20 h prior to commencement of experiments. Food, water and bedding were removed from cold-exposed mice during the

5 h period. Body temperatures were obtained by rectal probe (BAT-12 microprobe thermometer) over a 5 h period between the times of 0800 h and 1400 h. Temperatures were measured at 0, 30,

60, 90, 120, 180, 240 and 300 min. Body weights were recorded before and after the 5 h temperature measurement period. At the conclusion of the 5 h assessment, all mice were euthanised and

tissues and blood were collected for further analysis. BONE MARROW TRANSPLANTATION STUDIES WT mice aged 7 weeks old were irradiated using an X-RAD 320 with two doses of 500 cGy, 6 h apart.

The following day, 7-week old WT and _Klf3__−/−_ donor mice were euthanised and femora and tibiae harvested. BM cells were flushed with sterile RPMI medium and red blood cells lysed using

distilled and deionised water. Cells were resuspended in phosphate-buffered saline (PBS) and adjusted to 5 × 107/mL. Recipient mice were warmed using infrared heat lamps then 1 × 106 cells

were injected into the tail vein. Transplantation recipients were administered with antibiotics containing 200 mg/L sulfamethoxazole and 40 mg/L trimethoprim via drinking water for 2 weeks.

Following this, recipient mice and a control cohort of WT and _Klf3__−/−_ mice were transferred to a high-fat, high-sugar Western diet for 11 weeks (Supplementary Table 1). Mice were weighed

and assessed for body mass composition using an EchoMRI before and throughout the study, and at the conclusion tissues and blood were collected for further analyses. Genotyping of tail

biopsies and BM was performed to confirm reconstitution of WT or _Klf3__−/−_ cells in the BM of chimeric mice. Genotyping primers can be found in Supplementary Table 2, and gel images were

analysed using Bio-Rad Image Lab software v6.0.1. BLOOD AND PLASMA ANALYSIS Peripheral blood was collected by cardiac bleed using 50 U/mL heparin sulphate as an anti-coagulant, and stored in

K2EDTA collection tubes. Whole blood was sent for biochemical analysis of triglycerides and total cholesterol at the University of Sydney Veterinary Pathology Diagnostic Services. Plasma

was isolated by centrifugation (2000 × _g_ for 15 min) of whole blood and stored for further analysis. MOUSE TISSUE PROCESSING AT was harvested from the posterior subcutaneous white AT depot

(composed of dorsolumbar, inguinal and gluteal portions), gonadal white AT depot and the interscapular brown AT depot. To obtain the AT stromal vascular fraction (SVF), depots were minced

then digested using 1 mg/mL type II collagenase and cells were passed through a 40 μm sieve to remove undigested particles. The same digestion procedure was applied for lung tissue with the

addition of 1 μg/mL DNaseI. Spleens were homogenised in cold PBS. Centrifugation at 500 × _g_ for 10 min was then performed to separate the SVF pellet from floating adipocytes, for

downstream usage. AT explants were performed25, with supernatants harvested after 2 h of culture before being frozen and stored for further analysis. ADIPOSE H&E STAINING Sections of

subcut AT were mounted and stained with haematoxylin and eosin according to routine protocols. H&E sections were imaged on an Aperio XT Slide Scanner and processing was undertaken using

Aperio ImageScope software. LIVER BIOCHEMICAL ASSAYS Liver was snap-frozen in liquid nitrogen and stored at −80 °C until lipid extraction. Lipids were extracted from powdered tissue using a

modified Folch method63. The lipid extract was dried under a steady-stream of nitrogen. Extracts were resuspended in 0.4 mL 95% ethanol and heated to 37 °C prior to lipid assays.

Colorimetric assays were used to measure triglycerides (Point Scientific) and cholesterol (Thermo Fisher). All assays were conducted according to the manufacturer’s protocols and results

were normalised to liver weight. CELL CULTURE All cell lines were incubated in a 37 °C 5% CO2 water-jacketed incubator. COS-7 cells were a gift from Stuart Orkin (Harvard), and were cultured

in DMEM supplemented with 10% foetal bovine serum (FBS) and 1% penicillin–streptomycin–glutamine (PSG). During passaging, adherent cells were lifted after a 5 min incubation at 37 °C with 2

 mM PBS-EDTA. EoL-1 cells were grown in RPMI 1640 supplemented with 10% FBS and 1% PSG. The EoL-1 cell line was supplied by the European Collection of Cell Cultures (ECACC; Salisbury, United

Kingdom) as catalogue number 94042252, and purchased from Sigma Aldrich. The cell line has been reported before64. GENOME EDITING For CRISPR-Cas9 genome editing, a plasmid encoding both the

Cas9 protein and the sgRNA was used to delete _KLF3_ in EoL-1 cells via double-strand breakage and non-homologous end joining. pSpCas9(BB)-2A-GFP (PX458) was a gift from Feng Zhang (Addgene

plasmid 48138)65. We designed sgRNA sequences using the Optimized CRISPR-Cas9 design v1 online tool provided by the Zhang laboratory from the Massachusetts Institute of Technology. EoL-1

cells were transfected by nucleofection using a Neon Transfection System (Life Technologies). Cells (5 × 105) were resuspended in nucleofection buffer R (Neon Transfection Kit) and given one

pulse of 1350 V for 30 ms. Cells were then cultured for 72 h in RPMI 1640 with 10% FBS but lacking antibiotics. Transfected cells were enriched by FACS, and clonal populations were

established by sorting single cells into 96-well culture plates. To screen clones for the desired _KLF3_ disruption, PCR was carried out on genomic DNA using Q5 polymerase (New England

BioLabs), before confirmation via Sanger sequencing of PCR products. sgRNA, PCR and sequencing oligonucleotides can be found in Supplementary Table 2. FLOW CYTOMETRY Flow cytometry was

performed using a BD LSRFortessa and BD LSRFortessa X-20. Sorting by FACS was performed using a BD Influx and BD FACS Aria II. All cells were pre-blocked with anti-CD16/32 Fc block to reduce

non-specific binding, and UltraComp eBeads (Invitrogen) were used for single-stained compensation controls. For identifying eosinophils, cells were stained with a combination of the

following antibodies: CD45-biotin (BD Pharmingen), Streptavidin-BV711 (BD Horizon), CD11b-FITC (BD Pharmingen), F4/80-PE/Cy7 (Biolegend) and SiglecF-BV421 (BD Horizon). Live cells were

identified by the addition of TO-PRO-3 viability dye. Eosinophils were defined as live CD45+ CD11b+ F4/80+ Siglec-F+ SSChi cells. For delayed flow cytometric analysis, fully stained cells

were fixed with 1% paraformaldehyde. In these instances, ZombieNIR (Biolegend), a fixable viability dye, was used for identifying live cells, and F4/80 was conjugated to PE/Cy5

(eBioscience). To analyse other AT immune cell populations, various fluorescently-conjugated antibodies were used and detailed antibody information can be found in Supplementary Table 3,

with gating strategies for immune cell populations available in the legends for Fig. 4 and Supplementary Fig. 8. Flow cytometry analysis was performed using FlowJo software v10. CO-CULTURE

EXPERIMENTS Culture and differentiation of primary mouse adipocytes was performed66. Briefly, isolated SVF cells from the inguinal subcutaneous fat of WT mice were differentiated in 12-well

plates for two days using an induction medium of DMEM-F12 GlutaMAX with 10% FBS and 1X PSG, 1 μg/mL insulin, 0.5 mM IBMX, 0.25 μM dexamethasone, 1 μM rosiglitazone and 1 nM triiodothyronine

(T3). Two days after induction, cells were cultured in maintenance medium (DMEM-F12 GlutaMAX containing 10% FBS, 1X PSG, 1 μM rosiglitazone, 1 nM T3 and 1 μg/mL insulin) which was refreshed

every two days. Co-culture experiments were performed in a bicompartmental system using 0.4 μm pore polycarbonate transwell inserts (Corning). A standardised number of adipose eosinophils

were sorted from WT and _Klf3__−/−_ subcutaneous SVF using a BD FACS Aria III then resuspended in RPMI containing 10% FBS, 1X PSG and 10 ng/mL IL-5. Eosinophils (or media alone) were then

added to the upper compartment of the transwell and incubated at 37 °C in 5% CO2 for 4 h. Adipocytes from the lower compartment were then collected for gene expression analysis by qPCR. GENE

EXPRESSION ANALYSIS To assess mRNA expression, total RNA was isolated from cells and tissues then subjected to cDNA synthesis29. Quantitative real-time PCR (qPCR) reactions were set up with

Power SYBR Green PCR Master Mix and were run with default cycle parameters on the Applied Biosystems 7500 Fast Real-Time PCR System (for 96-well plate format) or the Applied Biosystems

ViiA7 Real-Time PCR System (for 384-well plate format). Applied Biosystems 7500 software v2.3 and Applied Biosystems QuantStudio Real-Time PCR software v1.3 were used for qPCR data analysis.

Gene expression was quantified using the 2−ΔΔCT method and relative mRNA expression was normalised to _18__S_ rRNA levels which have been shown to display consistent expression across the

cells and conditions studied. All qPCR primers were designed using primer3 (http://primer3.ut.ee/) and can be found in Supplementary Table 2. For microarrays on eosinophils sorted from

subcut SVF by FACS, RNA was isolated using the RNeasy Micro Kit (Qiagen) then subjected to quality control using an Agilent 2100 Bioanalyzer, following preparation with an Agilent RNA 6000

Pico Kit. An Affymetrix 3’ IVT Pico Kit was used prior to microarrays which were performed on an Affymetrix GeneChip HT MG-430PM Array Plate. Partek Genomics Suite v7 software was used for

data analysis, and microarray datasets are available from GEO (Accession No. GSE117445). PROTEIN EXTRACTION AND QUANTIFICATION Whole-cell protein extracts (WCE) were prepared by homogenising

mouse tissues with a glass dounce, in the presence of radioimmunoprecipitation (RIPA) buffer (50 mM HEPES, pH 7.5; 500 mM LiCl; 1 mM EDTA; 1% NP-40; 0.7% sodium deoxycholate) containing

protease inhibitors (cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail; one tablet dissolved in 10 mL RIPA buffer). Homogenates were rotated at 4 °C for 1 h then centrifuged at 21,000 × 

_g_ for 20 min at 4 °C to obtain the WCE. Protein extracts were subjected to a bicinchoninic acid (BCA) assay using a BCA Protein Assay Kit (Pierce) to determine concentration according to

the manufacturer’s protocol. For detection of UCP1, 25 μg of WCE was loaded onto Novex NuPAGE 10% Bis-Tris gels, following denaturation and boiling (15 μg for brown AT WCE). After blocking,

nitrocellulose membranes were incubated overnight with anti-UCP1 antibody (Abcam). Membranes were then probed with HRP-linked anti-rabbit antibody (GE Healthcare) prior to exposure using the

GE ImageQuant LAS 500 in the presence of Immobilon Western Chemiluminescent HRP Substrate (Millipore). To detect expression of VDAC (voltage-dependent anion channel), 20 μg of WCE was used

and incubation took place overnight with anti-VDAC antibody (Cell Signaling) before probing with HRP-linked anti-rabbit secondary antibody. To detect mitochondrial electron transport chain

complexes, several deviations from the above protocol were performed. Briefly, WCE were boiled for 5 min at 50 °C before SDS-polyacrylamide gel electrophoresis (PAGE) and transferral to a

0.45 μm polyvinyl difluoride (PVDF) membrane. Blocking took place overnight at 4 °C using 5% (w/v) skim milk powder in PBS (pH 7.4) with gentle agitation, followed by washing in PBST (PBS

with 0.1% Tween-20). The membrane was incubated with the primary Total OXPHOS Rodent WB Antibody Cocktail (Abcam) in 1% skim (w/v)/PBS. This optimised primary antibody mix contains five

mouse antibodies, one each against the mitochondrial oxidative phosphorylation (oxphos) complexes (Complex I: NADH dehydrogenase; Complex II: succinate dehydrogenase; Complex III:

CoQH2-cytochrome _c_ reductase; Complex IV: cytochrome _c_ oxidase; Complex V: ATP synthase). The secondary HRP-linked anti-mouse antibody labelling was performed in 1% skim milk (w/v)/PBS

before imaging. All membranes were stripped with 0.5 M NaOH then re-blocked with skim milk before probing with an anti-β-actin antibody (Sigma). Following this, HRP-linked anti-mouse

antibody (GE Healthcare) was incubated on membranes prior to imaging. For measurement of meteorin-like and IL-33 levels in plasma and AT explant supernatant, mouse meteorin-like/METRNL

DuoSet and mouse IL-33 DuoSet ELISA kits were utilised according to manufacturer’s instructions (R&D Systems). Detailed antibody information can be found in Supplementary Table 3.

CHROMATIN IMMUNOPRECIPITATION ChIP experiments were performed67. Briefly, ~7 × 107 cells were used for each immunoprecipitation (IP). Cells were crosslinked with 1% formaldehyde for 10 min

at room temperature, and the reaction was quenched with glycine at a final concentration of 125 mM. Crosslinked cells were then lysed and sonicated to obtain ~200–400 bp fragments of

chromatin. Cross-linked DNA was pulled down at 4 °C overnight using 15 μg anti-KLF3 antibody (Thermo Fisher) or 15 μg control normal goat IgG (Santa Cruz Biotechnology) per IP. qPCR was

performed on ChIP material using an Applied Biosystems ViiA7 Real-Time PCR System. For analysis, IPs were first normalised to the relative amount of input DNA, then to IgG controls. All qPCR

primers for ChIP were designed using primer3 (http://primer3.ut.ee/) and can be found in Supplementary Table 2. A ChIP-Seq data set for V5-tagged KLF3 produced from murine embryonic

fibroblasts was obtained from GEO (Accession No. GSE44748)52 and used to assess genome-wide KLF3 binding using Integrative Genomics Viewer (Broad Institute)68,69. Detailed antibody

information can be found in Supplementary Table 3. SOFTWARE GraphPad Prism v8 was used for generation of graphs and statistical analysis. Adobe Photoshop and Illustrator CS5.1 were used for

compiling and creating figures. FlowJo v10 was used for flow cytometry and cell sorting analysis. Applied Biosystems 7500 v2.3 and QuantStudio Real-Time PCR v1.3 were used for running and

analysing qPCR experiments. Oligonucleotides were designed using the online primer3 tool v4.1.0, and the Optimized CRISPR design v1 online tool was used for designing sgRNAs. Integrative

Genomics Viewer v2.4.9 (Broad Institute) was used for viewing ChIP-Seq data. Bio-Rad Image Lab v6.0.1 software was used for capturing and analysing agarose gels. Aperio ImageScope software

was used for analysing H&E sections. Licensed BioRender.com online software was used for generating cartoons and figures. STATISTICAL ANALYSIS The mean is shown for data in each figure

with individual data points shown, and SEM is shown as error bars. For flow cytometry plots, the median representative plot for each genotype/condition is shown, accompanied by the means ± 

SEM. D’Agostino-Pearson normality tests were performed to determine whether data followed a Gaussian distribution. Given the sample sizes, non-normality was determined, and thus

non-parametric one-tailed Mann–Whitney _U_ tests were used to test specific, directional hypotheses. *_P_ < 0.05, **_P_ < 0.01, ***_P_ < 0.001 (or #_P_ < 0.05 where stated).

Two-way ANOVA followed by post-hoc Tukey testing was carried out to assess significance between genotype, body temperature and time in acute cold and thermoneutral experiments. Two-way ANOVA

followed by False Discovery Rate to test for multiple comparisons was used for significant differentially-expressed genes in microarray studies. Fisher’s Exact test was used to define

significantly enriched GO terms in microarray Forest Plots. All statistical analyses were performed in GraphPad Prism v8. REPORTING SUMMARY Further information on research design is

available in the Nature Research Reporting Summary linked to this article. DATA AVAILABILITY Data supporting the findings of this study can be found within the paper and its Supplementary

Information files, and a Source data file has been provided for data underlying figures. Further information is available upon reasonable request. Full scans of western blots are available

in Supplementary Fig. 10. Datasets are publically available from the NCBI Gene Expression Omnibus (GEO) using Accession Nos. GSE117445 and GSE44748, FANTOM5 SSTAR

https://fantom.gsc.riken.jp/5/sstar/Main_Page) and Haemosphere (haemosphere.org/). A reporting summary has also been provided as part of the Supplementary Information files. Source data are

provided with this paper. REFERENCES * Wu, J. et al. Beige adipocytes are a distinct type of thermogenic fat cell in mouse and human. _Cell_ 150, 366–376 (2012). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Wu, J., Cohen, P. & Spiegelman, B. M. Adaptive thermogenesis in adipocytes: is beige the new brown? _Genes Dev._ 27, 234–250 (2013). Article  CAS  PubMed 

PubMed Central  Google Scholar  * Qiu, Y. et al. Eosinophils and type 2 cytokine signaling in macrophages orchestrate development of functional beige fat. _Cell_ 157, 1292–1308 (2014).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Wu, D. et al. Eosinophils sustain adipose alternatively activated macrophages associated with glucose homeostasis. _Science_ 332,

243–247 (2011). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Rao, R. R. et al. Meteorin-like is a hormone that regulates immune-adipose interactions to increase beige fat

thermogenesis. _Cell_ 157, 1279–1291 (2014). Article  CAS  PubMed  PubMed Central  Google Scholar  * Lee, E.-H. et al. Eosinophils support adipocyte maturation and promote glucose tolerance

in obesity. _Sci. Rep._ 8, 9894 (2018). Article  ADS  PubMed  PubMed Central  CAS  Google Scholar  * Bolus, W. R. et al. Elevating adipose eosinophils in obese mice to physiologically normal

levels does not rescue metabolic impairments. _Mol. Metab._ 8, 86–95 (2018). Article  CAS  PubMed  Google Scholar  * Nguyen, K. D. et al. Alternatively activated macrophages produce

catecholamines to sustain adaptive thermogenesis. _Nature_ 480, 104–108 (2011). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Pirzgalska, R. M. et al. Sympathetic

neuron−associated macrophages contribute to obesity by importing and metabolizing norepinephrine. _Nat. Med_ 23, 1309–1318 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  *

Camell, C. D. et al. Inflammasome-driven catecholamine catabolism in macrophages blunts lipolysis during ageing. _Nature_ 550, 119–123 (2017). Article  ADS  CAS  PubMed  PubMed Central 

Google Scholar  * Fischer, K. et al. Alternatively activated macrophages do not synthesize catecholamines or contribute to adipose tissue adaptive thermogenesis. _Nat. Med_ 23, 623–630

(2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Bell-Anderson, K. S. et al. Loss of Kruppel-like factor 3 (KLF3/BKLF) leads to upregulation of the insulin-sensitizing factor

adipolin (FAM132A/CTRP12/C1qdc2). _Diabetes_ 62, 2728–2737 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Sue, N. et al. Targeted disruption of the basic Kruppel-like factor

gene (Klf3) reveals a role in adipogenesis. _Mol. Cell Biol._ 28, 3967–3978 (2008). Article  CAS  PubMed  PubMed Central  Google Scholar  * Brestoff, J. R. et al. Group 2 innate lymphoid

cells promote beiging of white adipose tissue and limit obesity. _Nature_ 519, 242–246 (2015). Article  ADS  CAS  PubMed  Google Scholar  * Lee, M. W. et al. Activated type 2 innate lymphoid

cells regulate beige fat biogenesis. _Cell_ 160, 74–87 (2015). Article  CAS  PubMed  Google Scholar  * Min, S. Y. et al. Human ‘brite/beige’ adipocytes develop from capillary networks, and

their implantation improves metabolic homeostasis in mice. _Nat. Med._ 22, 312–318 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Jia, R. et al. Characterization of

cold-induced remodelling reveals depot-specific differences across and within brown and white adipose tissues in mice. _Acta Physiol._ 217, 311–324 (2016). Article  CAS  Google Scholar  *

Ghorbani, M. & Himms-Hagen, J. Appearance of brown adipocytes in white adipose tissue during CL 316,243-induced reversal of obesity and diabetes in Zucker fa/fa rats. _Int. J. Obes.

(Lond.)_ 21, 465–475 (1997). Article  CAS  Google Scholar  * Harms, M. & Seale, P. Brown and beige fat: development, function and therapeutic potential. _Nat. Med._ 19, 1252–1263 (2013).

Article  CAS  PubMed  Google Scholar  * Seale, P. et al. Prdm16 determines the thermogenic program of subcutaneous white adipose tissue in mice. _J. Clin. Investig._ 121, 96–105 (2011).

Article  CAS  PubMed  Google Scholar  * Roh, H. C. et al. Warming Induces Significant Reprogramming of Beige, but Not Brown, Adipocyte Cellular Identity. _Cell Metab._ 27, 1121–1137. e5

(2018). Article  CAS  PubMed  PubMed Central  Google Scholar  * Pollard, A. E. et al. AMPK activation protects against diet-induced obesity through Ucp1-independent thermogenesis in

subcutaneous white adipose tissue. _Nat. Metab._ 1, 340–349 (2019). Article  CAS  PubMed  PubMed Central  Google Scholar  * Ikeda, K. et al. UCP1-independent signaling involving

SERCA2b-mediated calcium cycling regulates beige fat thermogenesis and systemic glucose homeostasis. _Nat. Med_. 23, 1454–1465 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  *

Kazak, L. et al. Ablation of adipocyte creatine transport impairs thermogenesis and causes diet-induced obesity. _Nat. Metab._ 1, 360–370 (2019). Article  CAS  PubMed  PubMed Central 

Google Scholar  * Kazak, L. et al. A creatine-driven substrate cycle enhances energy expenditure and thermogenesis in beige fat. _Cell_ 163, 643–655 (2015). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Ye, L. et al. Fat cells directly sense temperature to activate thermogenesis. _Proc. Natl Acad. Sci. USA_ 110, 12480–12485 (2013). Article  ADS  CAS  PubMed 

PubMed Central  Google Scholar  * Davis, T. R. Chamber cold acclimatization in man. _J. Appl. Physiol._ 16, 1011–1015 (1961). Article  CAS  PubMed  Google Scholar  * Passmore, R. &

Durnin, J. V. Human energy expenditure. _Physiol. Rev._ 35, 801–840 (1955). Article  CAS  PubMed  Google Scholar  * Knights, A. J. et al. Kruppel-like factor 3 (KLF3/BKLF) is required for

widespread repression of the inflammatory modulator galectin-3 (Lgals3). _J. Biol. Chem._ 291, 16048–16058 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Knights, A. J. et

al. Krüppel-like factor 3 (KLF3) suppresses NF-kB–driven inflammation in mice. _J. Biol. Chem._ https://doi.org/10.1074/jbc.RA120.013114 (2020). * Funnell, A. P. et al. Generation of mice

deficient in both KLF3/BKLF and KLF8 reveals a genetic interaction and a role for these factors in embryonic globin gene silencing. _Mol. Cell Biol._ 33, 2976–2987 (2013). Article  CAS 

PubMed  PubMed Central  Google Scholar  * Funnell, A. P. et al. The CACCC-binding protein KLF3/BKLF represses a subset of KLF1/EKLF target genes and is required for proper erythroid

maturation in vivo. _Mol. Cell Biol._ 32, 3281–3292 (2012). Article  CAS  PubMed  PubMed Central  Google Scholar  * Vu, T. T. et al. Impaired B cell development in the absence of

Kruppel-like factor 3. _J. Immunol._ 187, 5032–5042 (2011). Article  CAS  PubMed  Google Scholar  * Heng, T. S. & Painter, M. W. The Immunological Genome Project: networks of gene

expression in immune cells. _Nat. Immunol._ 9, 1091–1094 (2008). Article  CAS  PubMed  Google Scholar  * Lucas, E. C. et al. Haemopedia RNA-seq: a database of gene expression during

haematopoiesis in mice and humans. _Nucleic Acids Res_. 47, D780–D785 (2018). PubMed Central  Google Scholar  * de Graaf, C. A. et al. Haemopedia: An Expression Atlas of Murine Hematopoietic

Cells. _Stem Cell Rep._ 7, 571–582 (2016). Article  Google Scholar  * Abugessaisa, I. et al. FANTOM5 transcriptome catalog of cellular states based on Semantic MediaWiki. _Database_ 2016,

baw105 (2016). * Bouffi, C. et al. Transcription factor repertoire of homeostatic eosinophilopoiesis. _J. Immunol._ 195, 2683–2695 (2015). Article  CAS  PubMed  Google Scholar  * Lee, S. D.

& Tontonoz, P. Eosinophils in fat: pink is the new brown. _Cell_ 157, 1249–1250 (2014). Article  CAS  PubMed  Google Scholar  * Knights, A. J., Vohralik, E. J., Hoehn, K. L., Crossley,

M. & Quinlan, K. G. R. Defining eosinophil function in adiposity and weight loss. _Bioessays_ 40, e1800098 (2018). Article  PubMed  Google Scholar  * Vohralik, E. J., Psaila, A. M.,

Knights, A. J. & Quinlan, K. G. R. EoTHINophils: eosinophils as key players in adipose tissue homeostasis. _Clin. Exp. Pharmacol. Physiol_. https://doi.org/10.1111/1440-1681.13304

(2020). * Eaton, S. A. et al. A network of Kruppel-like factors (Klfs). Klf8 is repressed by Klf3 and activated by Klf1 in vivo. _J. Biol. Chem._ 283, 26937–26947 (2008). Article  CAS 

PubMed  PubMed Central  Google Scholar  * Suo, J., Zhao, X., Guo, X. & Zhao, X. Met-enkephalin improves metabolic syndrome in high fat diet challenged mice through promotion of adipose

tissue browning. _Toxicol. Appl. Pharm._ 359, 12–23 (2018). Article  CAS  Google Scholar  * Zuberi, R. I. et al. Critical role for galectin-3 in airway inflammation and bronchial

hyperresponsiveness in a murine model of asthma. _Am. J. Pathol._ 165, 2045–2053 (2004). Article  CAS  PubMed  PubMed Central  Google Scholar  * Ge, X. N., Ha, S. G., Liu, F. T., Rao, S. P.

& Sriramarao, P. Eosinophil-expressed galectin-3 regulates cell trafficking and migration. _Front Pharm._ 4, 37 (2013). Article  CAS  Google Scholar  * Vasanthakumar, A. et al. The

transcriptional regulators IRF4, BATF and IL-33 orchestrate development and maintenance of adipose tissue-resident regulatory T cells. _Nat. Immunol._ 16, 276–285 (2015). Article  CAS 

PubMed  Google Scholar  * Han, J. M. et al. IL-33 Reverses an obesity-induced deficit in visceral adipose tissue ST2+ T regulatory cells and ameliorates adipose tissue inflammation and

insulin resistance. _J. Immunol._ 194, 4777–4783 (2015). Article  CAS  PubMed  Google Scholar  * Spallanzani, R. G. et al. Distinct immunocyte-promoting and adipocyte-generating stromal

components coordinate adipose tissue immune and metabolic tenors. _Sci Immunol._ 4, eaaw3658 (2019). * Mahlakoiv, T. et al. Stromal cells maintain immune cell homeostasis in adipose tissue

via production of interleukin-33. _Sci. Immunol._ 4, eaax0416 (2019). Article  CAS  PubMed  PubMed Central  Google Scholar  * Moysidou, M. et al. CD8+ T cells in beige adipogenesis and

energy homeostasis. _JCI Insight_ 3, e95456 (2018). Article  PubMed Central  Google Scholar  * Rajbhandari, P. et al. IL-10 signaling remodels adipose chromatin architecture to limit

thermogenesis and energy expenditure. _Cell_ 172, 218–233. e217 (2018). Article  CAS  PubMed  Google Scholar  * Burdach, J. et al. Regions outside the DNA-binding domain are critical for

proper in vivo specificity of an archetypal zinc finger transcription factor. _Nucleic Acids Res._ 42, 276–289 (2014). Article  CAS  PubMed  Google Scholar  * Coats, B. R. et al.

Metabolically activated adipose tissue macrophages perform detrimental and beneficial functions during diet-induced obesity. _Cell Rep._ 20, 3149–3161 (2017). Article  ADS  CAS  PubMed 

PubMed Central  Google Scholar  * Lumeng, C. N., Bodzin, J. L. & Saltiel, A. R. Obesity induces a phenotypic switch in adipose tissue macrophage polarization. _J. Clin. Investig._ 117,

175–184 (2007). Article  CAS  PubMed  PubMed Central  Google Scholar  * Jun, H. et al. An immune-beige adipocyte communication via nicotinic acetylcholine receptor signaling. _Nat. Med_. 24,

814–822 (2018). Article  CAS  PubMed  PubMed Central  Google Scholar  * Fabbiano, S. et al. Caloric restriction leads to browning of white adipose tissue through type 2 immune signaling.

_Cell Metab._ 24, 434–446 (2016). Article  CAS  PubMed  Google Scholar  * Abdullahi, A. et al. IL-6 signal from the bone marrow is required for the browning of white adipose tissue post burn

injury. _Shock_ 47, 33–39 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Haldar, S. M. et al. Kruppel-like factor 15 regulates skeletal muscle lipid flux and exercise

adaptation. _Proc. Natl Acad. Sci. USA_ 109, 6739–6744 (2012). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Liao, X. et al. Kruppel-like factor 4 is critical for

transcriptional control of cardiac mitochondrial homeostasis. _J. Clin. Investig._ 125, 3461–3476 (2015). Article  PubMed  PubMed Central  Google Scholar  * Oishi, Y. et al. SUMOylation of

Kruppel-like transcription factor 5 acts as a molecular switch in transcriptional programs of lipid metabolism involving PPAR-delta. _Nat. Med_ 14, 656–666 (2008). Article  CAS  PubMed 

Google Scholar  * Loft, A. et al. Browning of human adipocytes requires KLF11 and reprogramming of PPARgamma superenhancers. _Genes Dev._ 29, 7–22 (2015). Article  PubMed  PubMed Central 

CAS  Google Scholar  * Guillaumond, F. et al. Kruppel-like factor KLF10 is a link between the circadian clock and metabolism in liver. _Mol. Cell Biol._ 30, 3059–3070 (2010). Article  CAS 

PubMed  PubMed Central  Google Scholar  * Folch, J., Lees, M. & Sloane Stanley, G. H. A simple method for the isolation and purification of total lipides from animal tissues. _J. Biol.

Chem._ 226, 497–509 (1957). CAS  PubMed  Google Scholar  * Saito, H. et al. Establishment and characterization of a new human eosinophilic leukemia cell line. _Blood_ 66, 1233–1240 (1985).

Article  CAS  PubMed  Google Scholar  * Ran, F. A. et al. Genome engineering using the CRISPR-Cas9 system. _Nat. Protoc._ 8, 2281–2308 (2013). Article  CAS  PubMed  PubMed Central  Google

Scholar  * Chung, K.-J. et al. A self-sustained loop of inflammation-driven inhibition of beige adipogenesis in obesity. _Nat. Immunol._ 18, 654–664 (2017). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Schmidt, D. et al. ChIP-seq: using high-throughput sequencing to discover protein-DNA interactions. _Methods_ 48, 240–248 (2009). Article  CAS  PubMed  PubMed

Central  Google Scholar  * Robinson, J. T. et al. Integrative genomics viewer. _Nat. Biotechnol._ 29, 24–26 (2011). Article  CAS  PubMed  PubMed Central  Google Scholar  * Thorvaldsdottir,

H., Robinson, J. T. & Mesirov, J. P. Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. _Brief. Bioinform._ 14, 178–192 (2013). Article  CAS

  PubMed  Google Scholar  Download references ACKNOWLEDGEMENTS We would like to thank Chris Brownlee and Emma Johansson at the UNSW BRIL Flow Cytometry Facility for their assistance with

flow cytometry and sorting experiments, along with Brendan Lee (UNSW BRIL) and Debbie Burnett (Garvan Institute) for their help with animal experiments. We would like to acknowledge the work

of the Garvan Histopathology Service for histological sample preparation and staining, and the Biomedical Imaging Facility (BMIF) at UNSW Sydney. We are also very grateful for the

assistance of Dr. Axel Kallies, Dr. Ajith Vasanthakumar and Dr. Jon Brestoff with analysing immune cell populations by flow cytometry. We would also like to acknowledge the assistance of the

Ramaciotti Centre for Genomics (UNSW) for their sequencing and microarray assistance. This work was supported by funding from the Australian National Health and Medical Research Council to

M.C. (APP1025873 and APP1025877). A.J.K., J.J.Y. and E.S.S. were supported by Australian Postgraduate Awards. H.M.J. was supported by an Academic Training Scheme (SLAB) scholarship from the

Malaysian Ministry of Higher Education. K.G.R.Q. was supported by a Scientia Fellowship and S.J.A. and E.J.V. were supported by Scientia Scholarships. AUTHOR INFORMATION AUTHORS AND

AFFILIATIONS * School of Biotechnology and Biomolecular Sciences, UNSW Sydney, Sydney, NSW, 2052, Australia Alexander J. Knights, Emily J. Vohralik, Elizabeth S. Stout, Laura J. Norton, 

Stephanie J. Alexopoulos, Jinfen. J. Yik, Ellen M. Olzomer, Kyle L. Hoehn, Merlin Crossley & Kate G. R. Quinlan * Murdoch Children’s Research Institute, The Royal Children’s Hospital,

Melbourne, VIC, 3052, Australia Peter J. Houweling & Kathryn N. North * Department of Paediatrics, University of Melbourne, Melbourne, VIC, 3052, Australia Peter J. Houweling & 

Kathryn N. North * Charles Perkins Centre, School of Life and Environmental Sciences, University of Sydney, Sydney, NSW, 2006, Australia Hanapi Mat Jusoh & Kim S. Bell-Anderson Authors *

Alexander J. Knights View author publications You can also search for this author inPubMed Google Scholar * Emily J. Vohralik View author publications You can also search for this author

inPubMed Google Scholar * Peter J. Houweling View author publications You can also search for this author inPubMed Google Scholar * Elizabeth S. Stout View author publications You can also

search for this author inPubMed Google Scholar * Laura J. Norton View author publications You can also search for this author inPubMed Google Scholar * Stephanie J. Alexopoulos View author

publications You can also search for this author inPubMed Google Scholar * Jinfen. J. Yik View author publications You can also search for this author inPubMed Google Scholar * Hanapi Mat

Jusoh View author publications You can also search for this author inPubMed Google Scholar * Ellen M. Olzomer View author publications You can also search for this author inPubMed Google

Scholar * Kim S. Bell-Anderson View author publications You can also search for this author inPubMed Google Scholar * Kathryn N. North View author publications You can also search for this

author inPubMed Google Scholar * Kyle L. Hoehn View author publications You can also search for this author inPubMed Google Scholar * Merlin Crossley View author publications You can also

search for this author inPubMed Google Scholar * Kate G. R. Quinlan View author publications You can also search for this author inPubMed Google Scholar CONTRIBUTIONS A.J.K., K.G.R.Q. and

M.C. wrote the paper. A.J.K. performed most of the experiments; E.J.V., P.J.H. and A.J.K. performed the acute cold and thermoneutral experiments under the supervision of K.G.R.Q. and K.N.N.;

S.J.A. performed liver biochemical assays; J.J.Y. generated the _KLF3_ CRISPR construct; H.M.J. and K.S.B.A. performed the AT explant experiments; E.S.S., L.J.N. and E.M.O. performed

preliminary studies and provided technical assistance with experiments. A.J.K., K.G.R.Q. and M.C. designed the study, with intellectual input provided by K.L.H. All authors reviewed the

results and approved the final version of the paper. CORRESPONDING AUTHOR Correspondence to Kate G. R. Quinlan. ETHICS DECLARATIONS COMPETING INTERESTS The authors declare no competing

interests. ADDITIONAL INFORMATION PEER REVIEW INFORMATION _Nature Communications_ thanks Konstantinos Drosatos and the other, anonymous, reviewer(s) for their contribution to the peer review

of this work. PUBLISHER’S NOTE Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. SUPPLEMENTARY INFORMATION SUPPLEMENTARY

INFORMATION REPORTING SUMMARY SOURCE DATA SOURCE DATA RIGHTS AND PERMISSIONS OPEN ACCESS This article is licensed under a Creative Commons Attribution 4.0 International License, which

permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to

the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless

indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or

exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Reprints

and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Knights, A.J., Vohralik, E.J., Houweling, P.J. _et al._ Eosinophil function in adipose tissue is regulated by Krüppel-like factor 3

(KLF3). _Nat Commun_ 11, 2922 (2020). https://doi.org/10.1038/s41467-020-16758-9 Download citation * Received: 07 November 2018 * Accepted: 20 May 2020 * Published: 10 June 2020 * DOI:

https://doi.org/10.1038/s41467-020-16758-9 SHARE THIS ARTICLE Anyone you share the following link with will be able to read this content: Get shareable link Sorry, a shareable link is not

currently available for this article. Copy to clipboard Provided by the Springer Nature SharedIt content-sharing initiative