Dual Role of Caspase 8 in Adipocyte Apoptosis and Metabolic Inflammation

Obesity-associated metabolic inflammation is a fundamental characteristic of type 2 diabetes where adipose tissue has increasingly been found to play an important role (1). Humans have white and brown or beige adipose tissue; the latter two are characterized by high metabolic activity and expression of thermogenic genes, such as for uncoupling protein 1 (UCP1) (2). On the other hand, white adipose tissue contributes to metabolic inflammation by secreting modulatory adipokines and serves as a site for macrophage accumulation and immune cell activation (1,3). Unhealthy adipose tissue expansion that leads to adipocyte death is hypothesized to be the initiating factor in this inflammatory process (4). However, the signaling mechanisms and molecular regulations of this cell death remain unclear.

In this context, caspase 8 is an intriguing target of study because of its fundamental apoptotic and nonapoptotic roles. Caspases are best known as major effectors of apoptosis, or programmed cell death. Caspase 8 is the major effector of the death receptor (DR)–mediated extrinsic apoptosis pathway (5,6). Prototypical DR Fas, upon binding of its ligand FasL, recruits the Fas-associated death domain–containing adaptor protein (FADD) and caspase 8, forming the death-inducing signaling complex (DISC). Caspase 8 becomes activated through proximity-induced autocleavage, which can then cleave caspase 3 for activation and subsequent execution of apoptosis (7).

Emerging evidence has demonstrated additional context-specific roles of caspases in cellular survival and inflammation. In particular, caspase 8 has a range of diverse functions in different cells types (8–14). For example, we previously showed that disrupting caspase 8 in pancreatic β-cells protected both models of type 1 and type 2 diabetes from dysglycemia, but caspase 8 was required for maintenance of β-cell mass with aging (9).

Given the importance of adipose tissue in obesity-associated insulin resistance, including the key role of inflammation and adipocyte death, our objective was to determine the role of caspase 8–mediated signaling in adipocytes. In this study, we found that caspase 8 expression increases in adipose tissue of both mouse models and humans with obesity and type 2 diabetes. Inhibition of caspase 8 in 3T3-L1 adipocytes decreased Fas-mediated signaling. We then generated novel, adipose tissue–specific and adipocyte-specific caspase 8 knockout mice, which demonstrated improved glucose homeostasis and increased energy expenditure compared with littermate controls. These changes were associated with decreased adipose tissue inflammation and increased thermogenic UCP1 expression, suggesting that caspase 8 regulates adipose tissue inflammation and energy expenditure via Fas and nuclear factor κ-light-chain-enhancer of activated B (NF-κB) mediated pathways.

Mice with adipose tissue–specific or adipocyte-specific deletion of caspase 8 were generated by breeding aP2Cre⁺ (15) or AdipoqCre⁺ (16) mice (strains #005069 and #010803; The Jackson Laboratory) with Casp8^fl/fl (17) to generate aP2Casp8^+/− or adipoqCasp8^+/− mice, which were then intercrossed to generate aP2Casp8^−/− or adipoqCasp8^−/− mice, respectively. Genotypes were identified by PCR as previously described (13). Littermate aP2Casp8^+/+, adipoqCasp8^+/+, or Cre⁻adipoq⁻ mice were primarily used as controls as indicated. Mice were maintained on a C57BL/6 background, housed in a pathogen-free animal facility with a 12-h light-dark cycle, and fed standard irradiated rodent chow ad libitum (5% fat; Harlan Teklad). Male and female mice were used as indicated. Sample size was estimated based on previous studies using mouse models (18–20). Animals were randomly assigned to groups by the experimenter. A cohort of mice was fed a high-fat diet (HFD) (60% fat, 24% carbohydrates, and 16% protein based on caloric content, F3282; Bio-Serv) for up to 20 weeks starting at 6–8 weeks of age. All animal experimental protocols were approved by the St. Michael’s Hospital animal ethics committee and conformed to Canadian Council on Animal Care guidelines.

Glucose tolerance tests were performed on animals fasted overnight (14–16 h) using glucose 1 g/kg body weight injected intraperitoneally. Insulin tolerance tests were performed on animals fasted for 4 h using insulin lispro (Eli Lilly) 1.0 units/kg body weight for chow or HFD-fed mice. Measurements of blood glucose were taken at 0, 15, 30, 45, 60, and 120 min after the injection (21). For energy expenditure measurements, mice were individually housed in metabolic cages with free access to food and water. After 24-h acclimation to the apparatus, data for 24 h were collected and analyzed using a comprehensive laboratory animal monitoring system (Columbus Instruments). Food and water consumption were determined by weighing food or measuring water volume before and after 24 h. Insulin levels were measured by a mouse insulin ELISA kit (Crystal Chem). Serum adipokines were measured by Luminex technology using a mouse serum adipokine kit (Millipore).

Total protein extracts from cultured human adipocytes were obtained from Zen-Bio (TCE-OAD10, TCE-OA10). Samples were from healthy women with type 2 diabetes (n = 5, average BMI 49 kg/m² [range 39.5–57.4 kg/m²], average age 47 years [range 40–59 years]) or without diabetes (n = 4, average BMI 33.1 kg/m² [range 26.0–52.1 kg/m²], average age 45 years [range 37–63 years]).

The 3T3-L1 preadipocyte cells and cell culture reagents were obtained from Zen-Bio. The cell line was authenticated by quality control staining and tested for mycoplasma by the vendor. Cells were cultured and differentiated per manufacturer instructions. Experiments were performed after 7–10 days’ differentiation when mature adipocytes were formed. For caspase 8 inhibition, cells were serum starved for 4 h, followed by treatment for 24 h with adipocyte media containing 50 μmol/L Z-IETD-FMK (Millipore), which has previously been shown to be a relatively specific inhibitor of caspase 8 (22). For inflammatory stimulus, 20 ng/mL tumor necrosis factor-α (TNF-α) (Thermo Fisher Scientific) was added for 24 h. In vitro experiments were repeated three to four times.

Protein lysates were separated by SDS-PAGE and immunoblotted with antibodies for caspase 8 (NB100-56116 [Novus Biologicals] for mouse and 13423-1-AP [Proteintech] for human), Fas (sc-1023), cleaved caspase 3 (a-277) (Santa Cruz Biotechnology), NF-κB p65 (D12E12), phosphorylated NF-κB p65 (Ser536) (3033), perilipin (3470), and GAPDH (2118) (Cell Signaling). Band intensities were quantified using ImageJ software.

Adipocyte RNA was isolated using TRIzol reagent (Invitrogen). RNA was reverse transcribed by random primers using Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen), and quantitative real-time PCR was performed with a 7900HT Fast Real-Time PCR System (Applied Biosystems) with SYBR Green Master Mix reagent (Applied Biosystems). Primers used are listed in Supplementary Table 1. The expression level of each test gene was normalized to the internal control 18s (Rn18S). Each sample was run in triplicate.

Adipose and liver tissues were harvested after an overnight fast and fixed in 4% paraformaldehyde in 0.1 mol/L PBS (pH 7.4). Sections were stained with hematoxylin-eosin (H-E). Perigonadal adipose tissue sections were used for TUNEL (Roche Biochemicals) and immunofluorescence for perilipin (Cell Signaling) and F4/80 (Santa Cruz Biotechnology). For cell size and TUNEL analysis, at least 100 cells were counted per mouse. TUNEL-positive, perilipin-positive cells as a percentage of perilipin-positive cells were quantified. Adipocyte size was measured using ImageJ software. Adipocyte number was calculated as previously described (18,23). Macrophages were excluded from adipocyte size and number calculations by appearance in crown-like structures or costaining for F4/80 as previously described (24). Immunohistochemistry for CD3 and B220 were performed by The Centre for Phenogenomics (Toronto, ON, Canada).

Bulk RNA sequencing data of adipose tissue biopsies were obtained from Gene Expression Omnibus accession as indicated (Supplementary Table 2). When available, prenormalized data were used as provided, while raw count data were normalized using DESeq2 prior to further analysis. Data sets were examined for the relationship between clinical parameters and normalized gene expression. For two groups, unpaired t test was used to compare differences in gene expression distributions. For three groups, Dunnett test was used. Visualization and statistical analysis were performed using the ggplot2 version 3.4.2 and ggpubr version 0.6.0 packages in R version 4.1.1 (25,26).

Data are presented as mean ± SEM and were analyzed by two-tailed independent-sample Student t test for comparisons between two groups after testing for normality of the data. Two-way ANOVA was used for comparison among multiple measurements as indicated. GraphPad Prism 8 software was used. P ≤ 0.05 was considered as statistically significant.

The RNA sequencing data sets used are available from the National Center for Biotechnology Information’s Gene Expression Omnibus database under accession numbers GSE162653, GSE165932, GSE104674, GSE166047, and GSE135134. Other data are available from the corresponding author upon request.

To determine whether caspase 8 could potentially play a role in the setting of obesity and type 2 diabetes, we first examined caspase 8 in mouse models of obesity-associated insulin resistance and in humans with type 2 diabetes. Caspase 8 and cleaved caspase 8 protein were increased in perigonadal adipose tissue from mice fed a western-style HFD compared with chow diet–fed controls (Fig. 1A and Supplementary Fig. 1A). Moreover, caspase 8 protein was increased in cultured omental adipocytes obtained from humans with type 2 diabetes compared with humans without diabetes (Fig. 1B and Supplementary Fig. 1B). In humans, differences in caspase 8 may occur more so at the protein level. We examined adipocyte transcriptomic data from publicly available data sets, which did not suggest a trend toward increased CASP8 mRNA expression with either elevated BMI or insulin resistance (27–32) (Supplementary Fig. 2). Together, these data suggest that caspase 8 protein is increased and could potentially play a role in adipose tissue in the setting of obesity and insulin resistance.

To further study the role of caspase 8 in adipose tissue in vivo, we generated whole–adipose tissue caspase 8 knockout mice using an aP2 promoter–driven Cre-loxP system and adipocyte-specific caspase 8 knockout mice using an adiponectin-Cre-loxP system (15,16,33). Adiponectin-Cre is relatively specific for mature adipocytes, whereas aP2-Cre can be more widely expressed in adipocytes, preadipocytes, and other tissues. Use of both models reinforce an adipocyte-mediated role for common findings. Resulting knockout (aP2Casp8^−/−, adipoqCasp8^−/−) mice had caspase 8 deleted specifically in adipose tissue, with no change in liver, muscle, or other tissues important for glucose homeostasis compared with littermate controls (Supplementary Fig. 3A and B). Data were similar between Cre⁺ and Cre⁻ controls, similar to that previously described (20) (Supplementary Fig. 3C–F).

Interestingly, both aP2Casp8^−/− and adipoqCasp8^−/− mice were protected from HFD-induced metabolic abnormalities. After 8 weeks of HFD feeding, both male and female aP2Casp8^−/− knockout mice had significantly lower fasting blood glucose, lower serum insulin, and improved glucose tolerance on glucose tolerance testing compared with littermate controls (Fig. 2A, B, D, and E and Supplementary Fig. 4A and B). Male and female adipoqCasp8^−/− mice fed HFD also showed improved glucose tolerance compared with littermate controls (Fig. 2C, F, and G and Supplementary Fig. 4C and D). Under basal conditions, on standard chow diet, aP2Casp8^−/− and adipoqCasp8^−/− mice showed no significant difference in glucose tolerance or insulin tolerance (Fig. 3 and Supplementary Fig. 4E–H). Thus, disruption of caspase 8 in adipose tissue improves dysglycemia and glucose homeostasis with HFD. Overall, this finding suggests that caspase 8 plays a role in HFD-induced metabolic dysregulation.

We further monitored body weight in mice with adipose tissue- and adipocyte-specific disruption of caspase 8 and found that male and female aP2Casp8^−/− and adipoqCasp8^−/− mice were protected from weight gain on HFD compared with littermate controls (Fig. 4A–E). No difference in body weight was seen on chow diet (Fig. 4A, B, D, and E). Decreased weight gain was associated with decreased inguinal and perigonadal fat pad weights in both aP2Casp8^−/− and adipoqCasp8^−/− males, as well as in aP2Casp8^−/− females (Fig. 4F–I). The aP2Casp8^−/− mice on chow diet, but not adipoqCasp8^−/− mice, had a small decrease in fat pad weight (Supplementary Fig. 5).

Consistent with reduced adiposity, male and female aP2Casp8^−/− mice and male adipoqCasp8^−/− mice fed HFD had decreased adipocyte size in perigonadal fat pads as demonstrated by H-E staining and analysis of adipocyte cell size distribution (Fig. 5). Estimated total numbers of adipocytes were not different between aP2Casp8^−/− or adipoqCasp8^−/− mice and littermate controls (Fig. 5D, G, and K). Similar changes were seen in interscapular brown adipose tissue, and aP2Casp8^−/− mice were markedly protected from hepatic steatosis (Fig. 5A). Overall, these findings were consistent with smaller fat pad size associated with decreased weight gain in the setting of HFD-induced metabolic stress.

To characterize changes in caspase 8 signaling in adipocytes regulating glucose and energy homeostasis, we characterized activation of Fas-mediated apoptosis in adipose tissue of aP2Casp8^−/− mice. Decreased Fas, FasL, and cleaved caspase 3 were seen by Western blotting in perigonadal adipose tissue of aP2Casp8^−/− mice compared with controls, which were consistent with decreased activation of the extrinsic apoptosis pathway (Fig. 6A). Similar results were seen in differentiated 3T3-L1 adipocytes treated with caspase 8 inhibitor Z-IETD-FMK. Decreased levels of cleaved caspase 3 within this in vitro environment indicates a direct effect of caspase 8 inhibition in adipocytes rather than changes secondary to systemic differences (Fig. 6B). Furthermore, disruption of caspase 8 was associated with decreased levels of TUNEL for apoptosis in perilipin-stained cells and decreased F4/80 staining by immunofluorescence in adipose tissue (Fig. 6C and D). Immunohistochemical staining for CD3 and B220 did not show differences in perigonadal fat from aP2Casp8^−/− mice fed HFD compared with controls (Supplementary Fig. 6A). The adipoqCasp8^−/− mice were found to have increased CD3⁺ cells in perigonadal fat (Supplementary Fig. 6B). While the disruption of caspase 8 in vivo decreased activation of Fas signaling, in vitro experiments provided further insight that caspase 8 directly affects adipocyte signaling. Altogether, the data suggest that disrupting caspase 8 reduces Fas signaling and subsequent apoptosis.

We analyzed existing transcriptomic data sets from human adipocytes. Some data sets showed an association between insulin resistance and elevated BMI with increased expression of downstream targets of NF-κB signaling (27,28,32) (Supplementary Fig. 7). Caspase 8 has also been implicated in regulating inflammation by modulating NF-κB signaling (34). To determine whether caspase 8 regulation of NF-κB may be a potential mechanism whereby caspase 8 regulates adipose tissue inflammation and obesity-associated insulin resistance, we measured levels of phosphorylated p65 in perigonadal adipose tissue from aP2Casp8^−/− mice and found decreased levels (Fig. 7A). Disruption of caspase 8 in adipocytes also resulted in decreased inflammation as shown by decreased MCP-1 gene expression (Fig. 7B). Serum leptin was also decreased in HFD-fed aP2Casp8^−/− mice (Fig. 7C and Supplementary Fig. 6C and D). No difference was seen in serum adiponectin levels (Supplementary Fig. 6E and F). Overall, these data are consistent with decreased activation of NF-κB signaling and decreased adipose tissue inflammation following disruption of caspase 8, resulting in improved insulin sensitivity.

To further test whether caspase 8 directly regulates inflammation in adipocytes, we stimulated 3T3-L1 adipocytes with TNF-α in the presence or absence of caspase 8 inhibitor Z-IETD-FMK. Inhibition of caspase 8 abolished increases in TNF-α gene expression from inflammatory stimulation (Fig. 7D and Supplementary Fig. 8A–C). Z-IETD-FMK treatment also attenuated increases in gene expression for noncanonical IκB kinases IKKε and TANK binding kinase 1 (TBK1) (Fig. 7E and F). This finding supports an adipocyte-specific role for caspase 8 in regulating NF-κB signaling.

To test the extent to which targeting caspase 8 in adipocytes improves glucose homeostasis by downregulating NF-κB signaling, we generated adipoqCasp8^−/−adipoqIKK2^−/− mice with concomitant adipocyte-specific caspase 8 deletion and adipocyte-specific expression of an activated form of inhibitor of NF-κB kinase subunit β (Ikbkb), resulting in constitutive activation of the NF-κB pathway in adipocytes (35). Since improved glucose tolerance was seen in both aP2Casp8^−/− and adipoqCasp8^−/− mice, adiponectin-Cre–mediated deletion of caspase 8 and expression of activated Ikbkb was used because of its greater specificity for mature adipocytes. When mice were fed an HFD, adipocyte-specific activation of NF-κB partially attenuated the improved glucose tolerance seen with caspase 8 knockdown (Fig. 7G and Supplementary Fig. 8D), showing that caspase 8 mediates adipose tissue inflammation and glucose intolerance at least in part via regulation of adipocyte NF-κB.

Finally, because adipose tissue caspase 8 knockout mice were protected from weight gain on HFD, we sought to determine how caspase 8 regulates energy homeostasis. We measured energy expenditure in these animals using a metabolic caging apparatus and indirect calorimetry. Lean aP2Casp8^−/− mice had no difference in total body weight compared with controls but had increased oxygen consumption, indicative of increased energy expenditure (Fig. 8A). No differences in respiratory exchange ratio, activity levels, and food or water intake were seen compared with controls (Fig. 8B–E). Intriguingly, both aP2Casp8^−/− and adipoqCasp8^−/− mice had increased core body temperature compared with littermate controls (Fig. 8F and Supplementary Fig. 9A).

To determine the mechanism for this altered thermogenesis and energy expenditure, we measured expression of adipogenic and thermogenic genes (Fig. 8G and H). UCP1 was increased significantly in perigonadal white adipose tissue from aP2Casp8^−/− male mice (Fig. 8H). Noncanonical IκB kinases have been implicated in the negative regulation of UCP1 expression in white adipose tissue (36,37). Given the decreased phosphorylated p65 seen in perigonadal white adipose tissue, we also measured gene expression of IKKε and TBK1 and found these to be decreased in perigonadal and inguinal white adipose tissue from aP2Casp8^−/− male mice, which was consistent with decreased activation of NF-κB signaling and upregulation of UCP1 (Fig. 8H and Supplementary Fig. 9B). Similar trends in thermogenic gene regulation were seen in perigonadal and inguinal white adipose tissue from aP2Casp8^−/− male mice and perigonadal white adipose tissue from adipoqCasp8^−/− male mice (Supplementary Fig. 9B–D).

To test whether induction of UCP1 resulted in differences in weight gain and glucose tolerance, a cohort of HFD-fed adipoqCasp8^−/− and control mice were housed at thermoneutrality for 8 weeks, which minimized induction of thermogenesis in adipose tissue (Supplementary Fig. 9A). The adipoqCasp8^−/− rather than aP2Casp8^−/− mice were used to more specifically study the role of adipocyte thermogenesis and avoid confounding effects from other glucoregulatory tissues, such as the central nervous system (38). At thermoneutrality, adipoqCasp8^−/− mice gained weight similar to littermate controls (Fig. 8I). Moreover, knockout mice showed no difference in glucose tolerance (Fig. 8J), demonstrating that upregulation of thermogenesis with caspase 8 knockdown in adipocytes contributes to decreased weight gain and improved glucose tolerance with metabolic stress. These data identify an important role for caspase 8 in regulating energy homeostasis, inflammation, and insulin sensitivity via its role in regulating Fas-mediated and NF-κB signaling in adipose tissue (Supplementary Fig. 10).

In this work, we delineate the role of caspase 8–related signaling in adipose tissue and metabolism. Caspase 8 is a major mediator of the extrinsic pathway of apoptosis but has also been found to play a number of roles outside apoptosis, such as in regulating the major inflammatory NF-κB pathway. Here, we find that caspase 8 is upregulated in adipose tissue with obesity and type 2 diabetes and plays an important role in adipose tissue inflammation, glucose intolerance, and energy expenditure. To definitively study the role of caspase 8 in whole-body physiology, we generated mice with adipose tissue–specific and adipocyte-specific knockdown of caspase 8. These knockout mice were protected from HFD-induced obesity and glucose intolerance. Both sexes had improved glucose tolerance when fed HFD and decreased weight gain. The aP2Casp8^−/− mice also had increased energy expenditure, decreased adiposity, smaller adipocytes, and decreased hepatic steatosis in this setting. Major findings were similar in adipoqCasp8^−/− mice, including differences in glucose tolerance and weight gain, showing that results were mediated through knockdown, specifically in adipocytes. Similarly, treatment of 3T3-L1 adipocytes with caspase 8 inhibitor Z-IETD-FMK showed a direct reduction in DR-mediated signaling via Fas and attenuation of inflammation-induced TNF-α and noncanonical IκB kinases. In vivo, disruption of caspase 8 in adipocytes resulted in decreased adipose tissue inflammation and levels of activated NF-κB. Activation of Ikbkb in adipose tissue attenuated improvements in glucose tolerance seen with caspase 8 knockdown. Additionally, disruption of caspase 8 in vivo was associated with increased energy expenditure, core body temperature, and UCP1 expression in white adipose tissue. Improvements in glucose tolerance were abolished by housing mice at thermoneutrality. Collectively, our findings show that caspase 8 regulates adipose tissue inflammation, glucose tolerance, and energy expenditure in adipose tissue. Alongside the canonical role of caspase 8 in apoptosis, our data identify a novel role for adipocyte caspase 8 in mediating inflammation signaling while also impairing whole-body glucose and energy homeostasis.

By demonstrating increased caspase 8 expression within adipose tissue during obesity in mice and humans, our results are consistent with previous studies reporting a correlation between circulating caspase 8 and waist-to-hip ratio in humans (39), as well as increased caspase 8 and 9 gene expression in visceral and subcutaneous adipose tissue during obesity (40). We further demonstrate a context-specific role for caspase 8 in dysglycemia and energy metabolism through our animal models. Caspase 8 knockdown in vivo improves glucose tolerance and decreases weight gain during HFD feeding. Notably, adipose tissue knockdown of caspase 8 does not alter glucose metabolism or weight under basal, chow diet–fed conditions, implicating caspase 8 specifically in the pathogenesis of obesity-associated metabolic dysfunction.

Our data are strengthened by the use of two different Cre-mediated models of caspase 8 knockdown. Adiponectin-Cre is specific to mature adipocytes, whereas aP2 can be expressed in adipocytes, preadipocytes, and other tissues (33). For example, aP2 can be expressed in the nervous system (38). The normal morphology in aP2-Cre mice at birth and on chow diet suggest a major central nervous system contribution may be less likely, but effects from this or other tissues cannot be ruled out in the aP2-Cre mice alone. Thus, the similarity in glucose tolerance and body composition with aP2-Cre– and adiponectin-Cre–mediated knockdown support that the results seen are primarily adipocyte mediated, but this is an important consideration when interpreting these results.

Cinti et al. (4) are credited with the hypothesis that adipocyte cell death is the initiating factor leading to metabolic inflammation and insulin resistance. This association is now well recognized; however, the precise nature and role of this cell death remains unclear. While many studies report cell death and apoptosis to play an important role in diabetes (41–44), some studies suggest that more inflammatory forms of cell death, such as necrosis and associated inflammasome activation, may be more important in metabolic inflammation (45,46). On the other hand, other studies have been able to dissociate inflammation from insulin resistance (41,42,47). Insulin resistance in adipose tissue may promote inflammation, and inflammation can enable expansion of white adipose tissue in some animal models (48–50). In this work, we demonstrate that disruption of caspase 8, a major effector of apoptosis, is sufficient to confer significant protection from adipose tissue inflammation and insulin resistance in the setting of metabolic stress. Overall, this suggests that, at least in some contexts, cell death plays an important role in adipose tissue, and inhibition of apoptosis may represent a novel approach to improving metabolic homeostasis.

Notably, activation of caspase 8 in adipocytes has been used as a tool to induce apoptosis and loss of adipose tissue in FAT-ATTAC mice (43,51). These animals are glucose intolerant, and adipocyte apoptosis appears to trigger massive infiltration of macrophages. Our work complements these findings, as caspase 8 deletion led to improved glucose tolerance and decreased macrophage infiltration. A diversity of macrophages can exist in adipose tissue with distinct inflammatory or physiological roles (52–55). Further work is needed to determine whether caspase 8 signaling also regulates immune cell types in adipose tissue.

Caspases have increasingly been found to be important in roles beyond apoptosis. Caspases are known to target and cleave molecules involved in other cellular functions, such as cell cycle regulation (56,57). Caspase 8, in particular, has been found to have multiple roles dependent on tissue type and conditions. Caspase 8 was initially thought to be important for endothelial cell survival and cardiac development (11,58). Studies examining caspase 8 in T cells have shown its important role in proliferation and activation, including in patients with autoimmune lymphoproliferative-like syndrome (11–13). Caspase 8 is also crucial in B-cell proliferation in response to toll-like receptor stimulation, hematopoietic progenitor differentiation, and monocyte cell differentiation (11,34,59). Further roles for caspase 8 can include the regulation of autophagy via inhibition of autophagic proteins or degradation of catalase and suppression of tumor metastasis (60,61).

Within the context of diabetes, Liadis et al. (9) previously showed that deleting caspase 8 specifically in pancreatic β-cells can confer protection against apoptosis and improve glucose homeostasis in a multiple low-dose streptozotocin injection model of type 1 diabetes and in an HFD-induced model of obesity and type 2 diabetes. Their findings are consistent with our data demonstrating that deletion of caspase 8 can reduce levels of adipocyte death, similarly conferring protection against metabolic stress. Thus, caspase 8 can have context-specific functions with major implications for whole-body glucose homeostasis.

The present study also shows that deletion of caspase 8 reduces activation of NF-κB signaling in adipose tissue, particularly in the setting of metabolic inflammation. In T cells, caspase 8 has been reported to activate NF-κB following T-cell receptor stimulation by directly interacting with TNF receptor–associated factor 6 (TRAF6) and can cleave cellular FLIP long isoform protein (c-FLIPL) (14,62–65). In B cells with TLR4 stimulation, caspase 8 is recruited to an IKKαβ-containing complex, and the deletion of caspase 8 delays NF-κB translocation and impairs transcriptional activity. Our data support a similar role for caspase 8 in adipocytes, perpetuating inflammation in metabolic stress and ultimately dysglycemia.

Deletion of caspase 8 also protects mice from weight gain on HFD. The noncanonical IκB kinases IKKε and TBK1 can downregulate the expression of thermogenic proteins in adipose tissue (36,37). Consistent with decreased activation of NF-κB signaling, we also found lower expression of IKKε and TBK1 in the adipose tissue of our caspase 8 knockout mice. Expression of IKKε and TBK1 genes are also attenuated in 3T3-L1 adipocytes with caspase 8 inhibitor treatment in the setting of inflammatory stimulation. This suggests a direct mechanism regulating adipocyte thermogenesis, although other contributory mechanisms, such as via the central nervous system in aP2-Cre mice, cannot be ruled out (33,38). These data are also consistent with prior work demonstrating that inflammation in 3T3-L1 adipocytes induces caspase-mediated cleavage of the major adipogenic transcription factor peroxisome proliferator–activated receptor-γ (PPARγ) (66,67). TNF-related apoptosis-inducing ligand (TRAIL) is another member of the TNF family that may activate caspase 8 in adipocytes and promote PPARγ cleavage (67–69). Accordingly, with caspase 8 deletion, we see higher expression of PPARγ in adipose tissue. These data expand on previous studies implicating caspase 8 as a major regulator of inflammatory signaling and identify further repercussions for these roles in the setting of obesity and diabetes.

These data build on previous work reporting that upstream deletion of Fas can reduce adipose tissue inflammation and glucose intolerance (70). In this work, we similarly show that deletion of caspase 8 can confer protection against obesity-associated metabolic dysfunction. We additionally show that caspase 8 knockdown protects from weight gain on HFD and increases energy expenditure, which are findings that were not reported with Fas knockdown. Overall, these data further elucidate the complex roles of caspase 8 and identify new potential regulators of metabolism and therapeutic targets in adipose tissue.

Altogether, our findings suggest that caspase 8 plays an important role in adipocytes in the setting of metabolic stress. Caspase 8 increases in adipose tissue with obesity and insulin resistance, and disruption of caspase 8 decreases signaling through the extrinsic apoptosis pathway, as well as via canonical and noncanonical IκB kinases to decrease adipose tissue inflammation, increase energy expenditure, and protect from glucose intolerance and weight gain. Overall, these findings identify an important pathway in the pathogenesis of obesity-associated insulin resistance and suggest potential new therapeutic approaches to obesity and type 2 diabetes.

Acknowledgments. The authors thank the Keenan Research Centre for Biomedical Science Core Facilities at St. Michael’s Hospital, including Pamela J. Plant, Xiaofeng Lu, and Caterina Di Ciano-Oliveira, for technical advice and training.

Funding. This work was supported by Canadian Institutes of Health Research (CIHR) operating grant MOP-93707 to M.W. The study was also supported by Natural Sciences and Engineering Research Council of Canada discovery grant RGPIN-2018-05671, a J.P. Bickell Foundation grant, a Banting and Best Diabetes Centre new investigator award, a Heart and Stroke Richard Lewar Centres of Excellence in Cardiovascular Research and BI-LILLY new investigator award, CIHR project grant PJT-168996, and CIHR early career investigator award ARI-170743 to C.T.L. M.W. is supported by the Canada Research Chair in Signal Transduction in Diabetes Pathogenesis. C.T.L. was supported by a Heart and Stroke Canada new investigator award. D.J.H. was supported by a Banting and Best Diabetes Center-Novo Nordisk graduate studentship and an Ontario Graduate Scholarship. C.K.C. is supported by a Keenan Research Centre for Biomedical Science Research Trainee Centre top-up award.

Duality of Interest. C.T.L. has received meeting participant honoraria from Novartis. D.A.Y. is a consultant for Vivace Therapeutics and received speaker honoraria from GSK and Merck. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. C.T.L., C.K.C., F.C., S.Y.S., P.S.M., Y.Z.L., E.P.-T., S.A.S., H.R.D., T.S., E.P.C., M.K., D.J.H., A.C., and R.A. conducted experiments and analyzed data. C.T.L. and M.W. designed and oversaw the study and drafted and revised the manuscript. D.A.Y. advised and partially oversaw the study. A.H. and R.H. generated animal models and advised on the study. All authors reviewed the manuscript. C.T.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.