BDNF is expressed primarily in adipose SVF
We previously demonstrated that BDNF secretion increases from scWAT in response to noradrenergic stimulation, after administration of the β-3 adrenergic receptor (ADRβ3) agonist CL316,243 (8). We then endeavored to determine which adipose compartment and cell type was the source of local BDNF secretion in scWAT. Adipose tissue is a heterogeneous organ and although adipocytes are the main cell type, numerous other cell types are contained within the SVF of adipose tissue. Adipose SVF consists predominantly of hematopoietic lineage cells including adipose tissue macrophages (ATMs (41, 42)), but also contains numerous other immune cell types, preadipocytes, vascular endothelial cells, and pericytes. To determine the compartmental source of adipose BDNF, adult male C57BL/6 mice were cold exposed and whole SVF was isolated from mature adipocytes of inguinal scWAT by collagenase dispersion. Gene expression analysis revealed that Bdnf was almost exclusively expressed in the SVF, while other common growth factors (nerve growth factor; NGF, and vascular endothelial growth factor, VEGFa) showed no difference of expression between SVF and mature adipocyte fractions (p=<0.0001, p=0.1572, p=0.0767) (Fig. 1A).
Given the known expression of BDNF in immune cells, we sought to determine the contribution of myeloid-lineage BDNF to adipose innervation by creating a KO mouse model using Cre-Lox technology. LysMCre+/- mice were bred to BDNFfl/fl mice to generate LysMCre+/-::BDNF-/- (KO) animals, which lacked BDNF in myeloid lineage cells (Fig. 1B). (For the genotyping strategy, see Suppl. Fig. S1A.) Compared to littermate controls, KO animals exhibited a significant, although not complete, decrease of Bdnf in adipose SVF as measured by gene expression (p=0.0001) (Fig. 1C). Likely additional SVF cell types also contribute to tissue BDNF levels. Since myeloid lineage cells are also expressed in the brain, and BDNF has been shown to play a role in energy balance via CNS control of satiety, we also investigated whether our KO model affected BDNF expression in the hypothalamus. Gene expression of Bdnf in the hypothalamus did not differ between KO animals and their littermate controls (p=0.6132) (Fig. 1D). Physiological assessment using metabolic cages performed on KO animals and their littermate controls in the basal state revealed that KO mice had significantly lower energy expenditure (p=<0.0001) (Fig. 1E) despite showing no difference in body weight or adiposity (p=0.2735, p=0.6364) (Suppl. Fig. S1B).
KO mice have a blunted response to cold stimulation
Since a decrease in energy expenditure could be indicative of impaired sympathetic drive, we stimulated sympathetic nerve activity via cold exposure. When adult (22-23 week old) male mice were cold challenged at 5°C for 4 days, KO and control animals maintained similar body weight (p=0.0750) Fig. 2A, left panel). However, KO animals had a trend for increased adiposity (p=0.1107) (Fig. 2A, right panel). As cold exposure stimulates catecholamine-induced lipolysis mediated through sympathetic nerves, we next investigated innervation of the inguinal scWAT depot following a 7-day cold challenge. Protein expression of the pan-neuronal marker PGP9.5 was markedly reduced in inguinal scWAT of male KO mice compared to littermate controls (p=0.0025) (Fig. 2B, left panel). Protein expression of tyrosine hydroxylase (TH), a marker of sympathetic innervation and activation, was also drastically reduced in inguinal scWAT of KO animals compared to controls (p=0.0200) (Fig. 2B, right panel). Gene expression of synaptic/innervation markers (Psd95, Sox10, Synapsin I, Synapsin II, Synaptopysisn) in inguinal scWAT also showed a coordinated trend to be decreased in KO animals (p=0.386, p=0.126, p=0.155, p=0.130, p=0.728) (Suppl. Fig S1C), further indicating perturbations in scWAT innervation of KO animals. Together, these data indicated a ‘genetic denervation’ in this model.
Interestingly, gene expression of lipolytic markers Atgl and Hsl showed an increased trend in scWAT of KO animals (p=0.081, p=0.078) (Suppl. Fig S1C). Loss of scWAT innervation potentially results in decreased lipolysis due to lack of SNS mediated release of norepinephrine. However, due to the critical role of lipolysis in facilitating thermoregulation during cold challenge, we suspect that lipolysis in KO animals is achieved by alternative, nerve-independent, pathways. For example, glucocorticoid-mediate lipolysis has been attributed to increased β-adrenergic responsiveness via direct signaling or indirectly by inducing secretion of angiopoietin-like 4 (Angptl4) (49, 50).
Cold exposure induces expression of uncoupling protein 1 (UCP1) as required for adaptive thermogenesis, due to its ability to uncouple the mitochondrial respiratory chain resulting in a proton leak and heat production. UCP1 is therefore a unique marker of BAT and cold-induced ‘browning’ in scWAT. Gene expression of Ucp1 in inguinal scWAT of 7-day cold exposed mice was reduced in KO animals compared to littermate controls (p=0.0573) (Fig. 2C). Thyroid hormone potentiates sympathetic nervous system (SNS) activation of thermogenesis in BAT, but is produced via a distinct neuronal pathway from adipose sympathetic drive. We observed no changes in circulating thyroid hormones between KO and control (CON) animals (p=0.9526, p=0.3471) (Fig. 2D). Histological assessment of inguinal scWAT revealed what appeared like increased multilocularity suggestive of browning in KO animals (Fig. 2E, Typogen Black staining). However, there was a striking lack of UCP1 expression in these areas in KO tissues (Fig. 2E). Typogen Black (used here to reduce tissue autofluorescence) may also mark immune cells, thus there is potential for increased immune cell infiltration in the KO tissues as well. However, perilipin (a lipid droplet associated protein) staining of scWAT from KO animals clearly showed that the multilocular cells lacking UCP1 expression contained high amounts of lipid droplets (Fig. 2F, UCP1 immunofluorescence in red, perilipin immunofluorescence in green on serial sections). Preadipocytes maintain multilocular lipid droplets prior to differentiation, and considering that denervation of WAT increases hyperplasia, the observed multilocularity could more likely be areas of increased preadipocytes. In total, these findings demonstrate reduced thermogenic potential in denervated scWAT.
Genetic denervation of scWAT in LysMCre+/-::BDNF-/- (KO) mice is depot specific
We next sought to determine whether genetic denervation of scWAT in KO animals was restricted to this adipose tissue or extended to other organs. BDNF is a known myokine, and muscle is an energy expending tissue. We assessed innervation of fast twitch (gastrocnemius) and slow twitch (soleus) muscle in CON and KO, by investigating occupancy of neuromuscular junctions (NMJs) at basal conditions in adult male mice. Immunostaining of the presynaptic nerve and vesicles (neurofilament and SV2 markers, respectively) with postsynaptic acetylcholine receptors was performed to allow visualization of NMJ (Suppl. Fig. S2A, left panel). Following counts of occupied, partially occupied, and unoccupied NMJs it was determined that there was no evidence of neurodegeneration in the NMJ of KO animals (Suppl. Fig. S2A, right panel). In the same animals we also assessed axon numbers of spinal (L5 ventral root), motor, and sensory nerves through cross-section imaging (p=0.7344, p=0.3363, p=0.1469) (Suppl. Fig. S2B). A lower axon count could reflect neuronal death in these larger PNS nerves, but no difference was observed between CON and KO animals.
Myeloid cells are also present in BAT tissue. We therefore wanted to evaluate whether a lack of BDNF in BAT myeloid cells would have an effect on this tissue’s function. BAT of 7-day cold exposed 12-25 week old male CON and KO mice was evaluated. Protein expression of neuronal markers TH and PGP9.5 did not differ between KO and control mice (p=0.7747, p=0.7680) (Suppl. Fig. S3A), indicating that BDNF may not have an important neurotrophic role in BAT. Consistent with that, histological assessment of BAT revealed no difference in cellular morphology nor UCP1 expression (Suppl. Fig. S3B). Indeed, when adult male C57BL/6 mice were cold exposed or treated with the pharmacological ADRβ3 agonist, CL316,243, no difference in BDNF secretion from BAT was observed when compared to basal conditions (not analyzed due to low N of control animals) (Suppl. Fig. S3C). Together, these data supported scWAT depot specificity of our genetic denervation model, and BAT neuronal function may simply be due to increased sympathetic nerve activity and not changes in neural plasticity and neurite outgrowth in the tissue. Alternatively, a different neurotrophic factor, like NGF, may be more important for BAT than WAT.
HFD feeding exacerbates fat mass accumulation in LysMCre+/-::BDNF-/- (KO) Mice
Loss of sympathetic innervation to inguinal scWAT has been shown to increase depot mass (51), however, in our genetic model of scWAT denervation, no difference in adiposity was observed under basal conditions (Suppl. Fig. S1B) despite the observed decrease in energy expenditure (Fig. 1E). We next metabolically challenged CON and KO mice with a 45% high fat diet (HFD). Adult (25 week old) male CON and KO mice were placed on a HFD for 3-11 weeks, to assess adipose integrity and energy balance. At 3 weeks of HFD feeding animals were characterized in metabolic cages. HFD resulted in only a slight decrease in energy expenditure in KO mice compared to littermate controls (p=< 0.05) (Fig. 3A) as analyzed via waveform analysis over multiple 24hr periods. However, KO mice showed a higher respiratory exchange ratio (RER) than CON animals during the light cycle, indicative of preferential metabolism of carbohydrates over lipids for fuel (p=< 0.05) (Fig. 3B). These data fits with studies demonstrating that adipose nerves are important for lipolysis (52) and that denervation would shift fuel preference to carbohydrates. These physiological differences between CON and KO animals were observed despite no difference in food intake or change in body weight, (p=0.4755, p=0.2772) (Fig. 3C-D) at early time points. After 6 weeks of HFD, KO mice displayed better glucose control compared to CON animals (p=0.0165, p=0.0469) (Fig. 3E). Considering the higher RER displayed by KO mice, we attribute the favorable glucose control to a shift in fuel utilization toward carbohydrates instead of lipids (as nerve-induced lipolysis is likely blunted). By 11 weeks of HFD feeding, KO animals displayed greater adiposity than littermate controls, (p=0.0408) (Fig. 3F), potentially due to the reduced innervation and loss of neural control of certain metabolic processes.
Cold-induced neuroimmune cells (CINCs) are recruited to scWAT and express BDNF
After demonstrating the significance of myeloid derived BDNF to scWAT innervation, we sought to determine which myeloid cells were the source of tissue BDNF. Given their multifaceted role in adipose tissue, being a source of BDNF in the brain (microglia), and the phenotype observed in a myeloid-lineage KO, we hypothesized that monocytes/macrophages were the leading source of BDNF in scWAT. Since previous studies indicated that BDNF is increased in scWAT with noradrenergic stimulation (8), CD11b+ F4/80+ macrophages were isolated from SVF of inguinal scWAT of room temperature and 5-day cold exposed C57BL/6 adult (12 week old) male mice. Bdnf gene expression did not differ between room temperature and cold exposed CD11b+ F4/80+ macrophages (p=0.5796) (Fig. 4A). F4/80+ is considered a pan-macrophage marker, as such we considered it too broad to reveal phenotypic changes in the spectrum of macrophage populations in scWAT. Furthermore, F4/80+ is not the ideal marker of monocytes (macrophage precursors), which could be infiltrating the tissue in response to cold exposure. Based on these findings we applied a different approach to determining which myeloid cells are the source of scWAT BDNF. Adult (12 week old) female control animals were maintained at room temperature or cold exposed for 10 days. Inguinal scWAT SVF was isolated and flow cytometrically analyzed using a custom antibody cocktail against immune cells. Surprisingly, cold exposure did not have an effect on either M1 or M2 ATMs (p=0.8485, p=0.3387) (Fig 4B). Instead, the greatest increase was in Ly6C+CCR2+ monocytes. Both Ly6C+ and CCR2+ are markers of inflammatory monocyte migration (53). Following cold exposure, both Ly6C+CCR2+Cx3CR1- and Ly6C+CCR2+Cx3CR1+ populations increased in inguinal scWAT of female mice (p=0.0401, p=0.0354), however, only the Ly6C+CCR2+Cx3CR1+ population showed a strong trend for increase in males (p=0.9525 and p=0.0566) (Fig. 4C). In a separate cohort of adult male and female C57BL/6 mice, unbiased assessment using t-distributed Stochastic Neighbor Embedding (tSNE) revealed subpopulations of immune cells changing in propensity in scWAT with cold, and between male and female mice (cold and room temperature). Interestingly, the Ly6C+CCR2+Cx3CR1+ population, which we call cold-induced neuroimmune cells (CINCs) again increased in both male and female scWAT with cold (Fig. 4D).
To confirm that Ly6C+CCR2+Cx3CR1+ monocytes/macrophages were the source of BDNF, we FACS sorted out Ly6C+CCR2+Cx3CR1+ cells from inguinal scWAT SVF of 14 day cold exposed adult (12-13 week old) male C57BL/6 mice. We measured Bdnf gene expression in cold induced Ly6C+CCR2+Cx3CR1+ cells and found that they showed a trend for increased expression of Bdnf compared to cold exposed non-myeloid cells (p=0.1529) (Fig. 4E). We believe the previously observed increase in BDNF in adipose SVF ((8) and Fig. 1A) was due to infiltration of CINCs homing to scWAT upon cold exposure, and not an increase in per-cell BDNF levels.
Adrβ3 gene expression in Ly6C+CCR2+Cx3CR1+ cells confirmed the presence of norepinephrine (NE) receptor on these cells, indicating the potential to be responsive to SNS stimulation (p=0.7788) (Suppl. Fig. S4A), a likely mode for promoting BDNF release to the tissue after cold exposure. Taken together, these data indicated that Ly6C+CCR2+Cx3CR1+ cells are cold-induced neuroimmune cells (CINCs) that increase in number in scWAT after cold, have the potential to be stimulated by sympathetic nerves, and express BDNF.
Adipose lymph node and adipose lymphatics may play a role in recruitment of CINCs
CINCs express the transient markers Ly6C+ and CCR2+, as well as Cx3CR1+, indicating that they are recruited to adipose tissue following cold stimulation as opposed to being tissue resident immune cells. Immune cells are generally recruited to tissue through vasculature upon chemoattract release from target tissue. To further investigate dynamics of CINC recruitment to adipose tissue we utilized a Cx3CR1-EGFP reporter mouse. Whole mount imagining of axillary adipose tissue clearly demonstrated that Cx3CR1+ macrophages were present at the surface of and within the adipose lymph node (LN; Fig. 5A-B). 3D reconstruction of adipose LN imaging with depth coding allowed us to visualize Cx3CR1+ macrophages present on the capsule as well as within the cortex of the LN (Fig. 5B, middle and right panel). Furthermore, Cx3CR1+ macrophages were enriched in adipose lymphatics (Fig. 5C left panel, yellow arrow) compared to blood vasculature (Fig. 5C left panel, red arrow). We took advantage of innate autofluorescence to visualize vasculature in these samples, and although blood and lymph vasculature are somewhat morphologically similar, the bulbous sacs found at capillary terminals are a unique feature of lymphatic vasculature only (Fig. C middle and right panels, bulbous sacs outlined in white). These lymphatics were the vessels that predominantly contained Cx3CR1+ cells, and not the blood vasculature.
Confocal imaging of Cx3CR1+ macrophages on lymphatic endothelium clearly demonstrated that these cells have a distinct morphology (Fig. 5D left panel). The cell bodies are elongated with cytoplastic extensions creating a spindeloid shape associated with alternatively activated macrophages (54). 3D reconstruction with depth coding following confocal imaging of adipose lymph vessels revealed that Cx3CR1+ macrophages did not merely interact with lymph endothelial cells but were present within the lumen of the vessels (Fig. 5D, middle and right panels). Cx3CR1+ macrophages were present in greater numbers around and within the lymphatics of adipose (Fig. 5E), and as previously reported were also associated with adipose nerves (Fig. 5F) (55). Taken together, these data indicated that Cx3CR1+ macrophages are transported by the lymphatic vasculature to scWAT.