Identification of a Cytochrome P4502E1/Bid/C1q-dependent Axis Mediating Inflammation in Adipose Tissue after Chronic Ethanol Feeding to Mice*

Background: Chronic alcohol consumption leads to inflammation in adipose tissue, disrupting normal metabolic activity of adipocytes. Results: Expression of an alcohol metabolizing enzyme, cytochrome P4502E1, initiates inflammation in adipose. Bid-dependent apoptosis and activation of complement then exacerbate this initial response. Conclusion: Adipose inflammation during alcohol feeding develops in response to cytochrome P450 expression. Significance: Preventing adipose inflammation may prevent the pathphysiological effects of ethanol. Chronic, heavy alcohol exposure results in inflammation in adipose tissue, insulin resistance, and liver injury. Here we have identified a CYP2E1/Bid/C1q-dependent pathway that is activated in response to chronic ethanol and is required for the development of inflammation in adipose tissue. Ethanol feeding for 25 days to wild-type (C57BL/6J) mice increased expression of multiple markers of adipose tissue inflammation relative to pair-fed controls independent of increased body weight or adipocyte size. Ethanol feeding increased the expression of CYP2E1 in adipocytes, but not stromal vascular cells, in adipose tissue and Cyp2e1−/− mice were protected from adipose tissue inflammation in response to ethanol. Ethanol feeding also increased the number of TUNEL-positive nuclei in adipose tissue of wild-type mice but not in Cyp2e1−/− or Bid −/− mice. Apoptosis contributed to adipose inflammation, as the expression of multiple inflammatory markers was decreased in mice lacking the Bid-dependent apoptotic pathway. The complement protein C1q binds to apoptotic cells, facilitating their clearance and activating complement. Making use of C1q-deficient mice, we found that activation of complement via C1q provided the critical link between CYP2E1/Bid-dependent apoptosis and onset of adipose tissue inflammation in response to chronic ethanol. In summary, chronic ethanol increases CYP2E1 activity in adipose, leading to Bid-mediated apoptosis and activation of complement via C1q, finally resulting in adipose tissue inflammation. Taken together, these data identify a novel mechanism for the development of adipose tissue inflammation that likely contributes to the pathophysiological effects of ethanol.

Adipose tissue, long considered to be a relatively inert tissue that functioned solely as an energy store, is emerging as a key regulatory organ, acting as a component of the immune system as well as secreting a number of proteins regulating metabolism and behavior (1). These adipose tissue-derived bioactive substances place adipose tissue at an important nexus in the coordinated regulation of metabolism and immunity. Importantly, many of these cytokines, such as TNF␣ and IL-6, chemokines, such as monocyte/macrophage chemoattractant protein 1 (MCP-1), 2 and adipokines, such as adiponectin, also have central roles in the regulation of insulin sensitivity (1) and can contribute to chronic inflammatory diseases in other organs, such as alcoholic and non-alcoholic (steatohepatitis) liver disease (2).
After chronic ethanol feeding, indicators of oxidative stress, such as 4-hydroxynonenal, are detected in adipose tissue (3,4). Expression of inflammatory cytokines and macrophage infiltration into adipose tissue are also increased, and adipocytes exhibit insulin resistance (3). Both insulin-stimulated glucose transport and the ability of insulin to inhibit lipolysis are reduced after chronic ethanol feeding (3) associated with the development of systemic insulin resistance and the development of hepatic steatosis (5). Chronic ethanol feeding also decreases circulating adiponectin (6,7), leptin, and resistin (8) in animal models.
These effects of ethanol on the metabolic and innate immune activity of adipose tissue likely contribute to the ethanol-induced tissue injury (9) and may also influence behaviors regu-lating alcohol consumption (10,11). However, the mechanisms by which chronic ethanol feeding leads to adipose tissue inflammation and disrupts adipose metabolism have not been identified. The initiating signals for adipose inflammation in response to ethanol is likely distinct from the early signals leading to adipose tissue inflammation in obesity, as ethanol feeding does not increase adiposity or body weight compared with controls (5). Many of the pathophysiological effects of ethanol are dependent on ethanol metabolism, particularly via cytochrome P4502E1 (CYP2E1). CYP2E1 is induced in response to chronic exposure to high concentrations of ethanol; expression of CYP2E1 is greatest in the liver, but expression is also increased in adipose tissue of rats after ethanol feeding (4).
Induction of CYP2E1 is associated with increased cellular death by apoptosis and necrosis in the liver during ethanol exposure (12). Apoptosis is a complex, caspase-dependent cell death mechanism mediated by two distinct pathways called the extrinsic (death receptor-mediated) and the intrinsic (mitochondria-mediated) pathways. Interaction of the death receptors (TNF␣ receptor or Fas) with their respective ligands initiates the extrinsic pathway, whereas mitochondrial dysfunction triggers the intrinsic pathway. Bid is a key pro-apoptotic Bcl-2 family member that is crucial for the induction of apoptotic death and serves as a bridge between these two fundamental cell death pathways. Bid-dependent apoptosis in adipose tissue is important in the development of inflammation in models of non-alcoholic steatohepatitis and is associated with the subsequent development of insulin resistance and hepatic steatosis (13). Although early studies in cultured cells suggested that apoptosis does not elicit strong inflammatory responses, instead mediating a "silent clean-up" of dead cells (14), more recent data from in vivo models implicates apoptotic cells as strong inducers of inflammatory responses in a number of pathophysiological conditions (15).
The mechanistic link between apoptosis and inflammation is not well understood, although there is a growing body of evidence indicating that complement may be involved in inflammatory responses evoked by apoptotic cells (16,17). Complement is an ancient component of the innate immune response and is involved in protection from bacterial infections as well as tissue repair in response to injury (18). Complement is activated via three pathways, the classical, lectin, and alternative. The classical pathway of complement is activated upon the binding of C1q, the recognition subunit of C1, to immune complexes or via interactions with cell surface markers on apoptotic cell (16,17). The recognition molecules for C1q on the surface of the apoptotic cells are not clearly defined but likely include phosphatidylserine, surface blebs, and/or nucleic acids on the cell surface. C1q binds apoptotic cells with its globular head domain and can then serve as a bridging molecule, interacting with receptors on the surface of macrophages with its collagenous tail and stimulating phagocytosis of apoptotic cells (16,17).
Adipose tissue produces C3, factor B, and factor D, essential components for the alternative pathway of activation, as well as properdin, several complement receptors, and regulatory proteins (19). The C3a receptor was recently identified as a key determinant of insulin resistance and adipose tissue inflamma-tion in diet-induced obesity in mice (20). Here we have tested the hypothesis that ethanol metabolism via CYP2E1 is linked to the development adipose tissue inflammation via Bid-dependent apoptosis and the subsequent activation of complement. Making use of genetically manipulated mice, we have identified the pathophysiologically relevant sequence of events in adipose tissue during chronic ethanol exposure that leads to inflammation; ethanol induces CYP2E1 expression, leading to increased Bid-dependent apoptosis in adipose tissue, activation of complement via C1q in the classical pathway, and increased production of inflammatory cytokines and chemokines in adipose tissue.
Measures of ethanol metabolism were determined in response to an acute challenge with ethanol via gavage. Ethanol-naïve mice 6 -8 weeks old were fasted 16 -18 h and then exposed to 6 g/kg ethanol (in saline) via an intragastric gavage (acute gavage). After 90 min, blood samples were taken from the saphenous vein into heparinized hematocrit tubes. Plasma was isolated and then stored at Ϫ80°C. Plasma ethanol concentrations (Table 1) were determined using an Ethanol L3K Assay kit (Genzyme Diagnostics; Framingham, MA) following the manufacturer's instructions.
Characteristics of the mice (initial and final body weights, food intake, and measures of ethanol metabolism) are shown in Table 1. At the time of euthanasia, mice were anesthetized, gonadal adipose tissue, liver, and blood were harvested and either frozen in liquid nitrogen for protein analysis, fixed in formalin for histology, placed in RNA Later for mRNA analysis, or used for the isolation of adipocytes and stromal vascular cells.
Complement Activation in Plasma-Blood was drawn from mice via the saphenous vein (without anesthesia) using Microvette CB300 EDTA-coated capillary collection tubes and placed on ice. Plasma was collected by centrifugation for 2.5 min at 10,000 ϫ g and transferred into a 1.5-ml microcentrifuge tube containing 1.2 l of 0.5 M EDTA (for a final EDTA concentration of 20 mM). Samples were immediately denatured in Laemmli buffer for Western blot analysis and then stored at Ϫ80°C. Plasma collected without added EDTA was treated with 2 mg/ml zymosan (28) and used as a positive control for complement activation.
Immunohistochemistry and TUNEL Staining-Formalinfixed paraffin-embedded adipose tissue samples were stained with hematoxylin and eosin or de-paraffinized and used for immunohistochemistry and TUNEL staining. Immunoreactive C3b/iC3b/C3c (collectively termed "C3b" here) was detected with an antibody against neo-epitopes revealed upon C3 cleavage (Hycult Biotech; Canton, MA), as described (29). TUNEL was visualized using the ApopTag plus In Situ Apoptosis Detection kit (S7165 for rhodamine staining; Chemicon International, Temecula, CA) following the manufacturer's instructions.
Isolation of Stromal Vascular Cells and Adipocytes and Preparation of Adipose and Liver Tissue Samples-Adipose tissue harvested from pair-and ethanol-fed mice was digested in collagenase (30). The resulting cell suspension was filtered through coarse mesh and centrifuged at 233 ϫ g for 1 min to separate adipocytes from stromal vascular cells. Frozen adipose tissue (0.2 g of fat/ml), or cell fractions were lysed and processed for Western blot analysis, as previously described (4). Microsomes were isolated for frozen adipose tissue for CYP2E1 activity assays, as previously described (4).
Western Blot Analysis-Lysate (20 g protein) or plasma (0.5 l) samples were loaded onto 8% reducing SDS-polyacrylamide gels. Antibodies C3 (used for the detection of C3d at ϳ40 kDa) were from MP Biomedicals, LLC; Solon, OH, CYP2E1 was from Abcam, and HSC70 (used as a loading control for Western blots of tissue samples) was from Santa Cruz Biotechnology, Inc). Immunoreactive proteins were visualized using enhanced chemiluminescence, images were collected, and signal intensities were quantified using Eastman Kodak Co. Image Station 4000R.
Activity of CYP2E1-One hundred microliters of microsomal protein isolated from adipose tissue or liver was used to measure CYP2E1 activity (31).
Quantitative Real-time-PCR-Total RNA was isolated from adipose tissue using the RNeasy lipid tissue mini kit following the manufacturer's instructions. Total RNA was reverse-transcribed using the RETROscript kit (Ambion). Real-time PCR amplification was performed using the SYBR Green PCR Master Mix (Applied Biosystems) in an Mx3000p PCR machine, as previously described (4). The primer sequences used were:

CYP2E1-Bid-C1q Axis in Adipose Inflammation
Statistical Analysis-Values reported are the means Ϯ S.E. Data were analyzed using Student's t test or general linear models procedure followed by least square means testing (SAS Institute, Cary, IN).

RESULTS
After chronic ethanol feeding to mice, multiple markers characteristic of adipose tissue inflammation were apparent after 18 -25 days of ethanol feeding. First, crown-like structures were observed in ethanol-fed but not pair-fed mice (Fig. 1A). Second, expression of mRNA for TNF␣, IL-6, and MCP-1 (Fig.  1B) and CD11c (Fig. 1C) was increased in ethanol-fed relative to pair-fed controls after 25 days, but not 18 days, of ethanol feeding. Expression of CD11c by adipose macrophages is associated with adipose inflammation in models of metabolic syndrome (32). Finally, expression of multiple members of the complement system, C1q, C3, C3aR, and C5aR, was increased in adipose tissue in response to 25 days of ethanol feeding (Fig. 1D). mRNA expression for cytokines and complement proteins pro-FIGURE 2. Cytochrome P450 2E1 and adipose tissue inflammation and apoptosis after chronic ethanol feeding. Wild-type and Cyp2e1 Ϫ/Ϫ mice were allowed free access to ethanol containing diets for up to 25 days or pair-fed control diets. A, quantity of immunoreactive CYP2E1 protein was measured by Western blot in adipose tissue. HSC70 was used as a loading control. Values with different superscripts are significantly different from each other, p Ͻ 0.05, n ϭ 4 -7. B, the enzymatic activity of CYP2E1 was measured in microsomes isolated from adipose tissue (n ϭ 4). *, p Ͻ 0.05 compared with pair-fed. C, the quantity of immunoreactive CYP2E1 protein was measured by Western blot in adipocytes and stromal vascular cells (SVC) isolated from adipose tissue. Images are representative of at least four mice per treatment group. P, pair; E, ethanol. D, expression of mRNA for inflammatory markers was measured as in Fig. 1. Data are expressed as -fold over pair-fed of the same genotype; there were no differences in expression between genotypes in pair-fed mice (data not shown). Values represent the means Ϯ S.E. n ϭ 3-7. *, p Ͻ 0.05 compared with pair-fed. E, TUNEL-positive nuclei (red) were detected in adipose tissue from wild-type and Cyp2e1 Ϫ/Ϫ mice. Nuclei were labeled with DAPI (blue). TUNEL-positive nuclei were counted and expressed as the percent of DAPI positive nuclei (WT pair-fed, 4.2 Ϯ 0.4; ethanol-fed, 21.2 Ϯ 5.4 (p Ͻ 0.05 compared with wild-type pair-fed mice); Cyp2e1 Ϫ/Ϫ pair-fed, 1.4 Ϯ 0.6; ethanol-fed, 1.5 Ϯ 1.0, values represent the means Ϯ S.E., n ϭ 3-6).

CYP2E1-Bid-C1q Axis in Adipose Inflammation
vide a measure of local production by adipose tissue independent of possible systemic changes in the quantity of these secreted proteins.
The mechanisms for adipose tissue inflammation after chronic ethanol feeding are likely distinct from adipose tissue inflammation in models of obesity, as chronic ethanol feeding did not increase body weight over pair-fed controls (see Table  1) or affect adipocyte size when compared with pair-fed control mice (Fig. 1A). Instead, we hypothesized that ethanol metabolism via inducible CYP2E1 contributes to chronic ethanol-induced inflammation in adipose tissue. CYP2E1 protein was increased in adipose tissue from days 11 to 25 of ethanol feeding ( Fig. 2A) before the appearance of inflammatory markers in adipose. This increase in protein was associated with increased CYP2E1 enzymatic activity, assayed in microsomes isolated from adipose tissue (Fig. 2B). Activity of CYP2E1 in liver microsomes was 10 -15-fold higher than that observed in adipose tissue (data not shown). CYP2E1 was predominantly expressed in adipocytes and not stromal vascular cells (SVC) in the adipose tissue (Fig. 2C).
If CYP2E1 was important in the development of adipose tissue inflammation during ethanol feeding, then mice deficient in CYP2E1 should be protected from ethanol-induced inflammation in adipose tissue. Wild-type and Cyp2e1 Ϫ/Ϫ mice were allowed free access to the ethanol containing diet for 25 days or pair-fed control diets. Although ethanol feeding increased the expression of mRNA for TNF␣, IL-6, MCP-1, C1q, C3aR, and C5aR, these responses were ameliorated in Cyp2e1 Ϫ/Ϫ mice (Fig. 2D).
Elevated expression of CYP2E1 is associated with oxidative stress and apoptotic cell death in a number of model systems (12). Chronic ethanol feeding increased the number of TUNELpositive cells in the adipose tissue of wild-type mice but not Cyp2e1 Ϫ/Ϫ mice (Fig. 2E).
In a high fat diet induced model of obesity in mice, Bid-dependent apoptosis mediates adipose tissue inflammation (13). Therefore, we hypothesized that chronic ethanol feeding could lead to Bid-dependent apoptosis in adipose tissue and, thus, contribute to adipose tissue inflammation. Chronic ethanol feeding increased the number of TUNEL-positive nuclei in adi-  OCTOBER 14, 2011 • VOLUME 286 • NUMBER 41 pose tissue from wild-type, but not Bid Ϫ/Ϫ mice, at 25 days (Fig.  3A). Multiple markers of inflammation, including expression of IL-6, MCP-1, and CD11c mRNA, in adipose tissue were normalized in Bid Ϫ/Ϫ mice (Fig. 3B). In contrast, expression of mRNA for C3aR and C5aR was only partially attenuated, and the expression of mRNA for TNF␣ remained elevated (Fig. 3B). Chronic ethanol feeding increased CYP2E1 protein in the Bid Ϫ/Ϫ mice (data not shown). Taken together, these data suggest that CYP2E1 was sufficient for increased expression of TNF␣, C3aR, and C5aR mRNA in adipose tissue in response to chronic ethanol but that Bid-dependent apoptosis contributed to an exacerbation of adipose tissue inflammation.

CYP2E1-Bid-C1q Axis in Adipose Inflammation
We next determined whether chronic ethanol feeding increased complement activation in adipose tissue. Ethanol feeding increased the quantity of immunoreactive C3d, an indicator of complement activation, by d18 and d25 of ethanol feeding in adipose tissue lysates (Fig. 4A) and plasma (Fig. 4B) compared with pair-fed control mice. C3b deposition was also detected in adipose tissue by immunohistochemistry at d18 and d25 of ethanol feeding (Fig. 4C). Importantly, C3b deposition in adipose tissue was not observed in Cyp2e1 Ϫ/Ϫ or Bid Ϫ/Ϫ mice (Fig. 4D), indicating that ethanol-induced apoptosis occurred up-stream of complement activation in response to ethanol feeding. Expression of mRNA for factors in the complement pathway, including C1q and C3aR, were not affected among genotypes (data not shown). Furthermore, complement activation in adipose tissue, evident by d18, preceded the increase in inflammatory cytokines and chemokines at d25.
Making use of mice deficient in C1q (C1qa Ϫ/Ϫ , lacking the classical pathway of complement activation), C4 (C4 Ϫ/Ϫ , lacking classical and lectin pathways), or complement factor D (Cfd Ϫ/Ϫ , lacking the alternative pathway), we investigated whether complement activation was required for ethanol-induced adipose tissue inflammation as well as which pathway(s) of complement activation was involved in the response to chronic ethanol in adipose tissue. Expression of TNF␣, IL-6, and MCP-1 mRNA was increased in adipose tissue from wildtype mice after ethanol feeding compared with pair-fed con-trols (Fig. 5A). Mice deficient in C1q or C4 were protected from ethanol-induced increases in IL-6 and MCP-1 mRNA in adipose tissue (Fig. 5A). In contrast, expression of IL-6 and MCP-1 did not differ between wild-type and Cfd Ϫ/Ϫ mice after ethanol feeding (Fig. 5A). Ethanol feeding to C1q-deficient mice increased CYP2E1 protein expression (data not shown) and TUNEL-positive nuclei (Fig. 5B) in adipose tissue as well as TNF␣ (Fig. 5A) and C3aR and C5aR mRNA expression (Fig.  5C).
If C1q was sufficient to mediate ethanol-induced inflammation in response to ethanol, then mice deficient in both MBLa/c and complement factor D (MBLa/c/Cfd Ϫ/Ϫ ), lacking both functional lectin and alternative pathways, should be sensitive to ethanol exposure. MBLa/c/Cfd Ϫ/Ϫ mice were not protected from ethanol-induced increases in TNF␣, IL-6 or MCP-1 mRNA (Fig. 5D). Taken together, these data suggest a necessary and sufficient role for the classical pathway in ethanol-induced expression of IL-6 and MCP-1 mRNA in adipose tissue.

DISCUSSION
Although studies of the pathophysiological effects of obesity reveal that adipose tissue inflammation plays a role in the development of non-alcoholic fatty liver and steatohepatitis and metabolic syndrome, the interaction between adipose tissue and alcoholic liver disease has not been well studied. Here we find that multiple markers of inflammation are detected in adipose tissue after chronic ethanol feeding. The development of adipose tissue inflammation in obese animals is associated with an increase in body weight and adipocyte size (33). In contrast, ethanol feeding is not associated with increased body weight, adiposity, or adipocyte size (3,5), suggesting that adipose inflammation in response to ethanol is initiated via a mechanism(s) distinct from those involved in obesity-induced adipose inflammation. Ethanol metabolism via CYP2E1 increases oxidative stress and is a well studied contributor to alcoholic liver disease (12). CYP2E1 expression was increased in adipose tissue during ethanol feeding and was critical to the development of inflammation in adipose tissue in mice. Increased expression FIGURE 4. Activation of complement in adipose tissue during ethanol feeding to mice. A, quantity of immunoreactive C3d in adipose tissue lysates was measured by Western blot analysis in wild-type mice fed ethanol diets for up to 25 days or pair-fed control diets for 25 days. HSC70 was used as a loading control. Values represent the means Ϯ S.E., n ϭ 4 -8. *, p Ͻ 0.02 compared with pair-fed. B, quantity of immunoreactive C3d in plasma was measured by Western blot from pair-or ethanol-fed wild-type mice at 18 and 25 days. Immunoreactive C3d is expressed relative to the quantity of plasma loaded onto the gels. Values represent the means Ϯ S.E., n ϭ 8 -10. *, p Ͻ 0.01, compared with pair-fed. C and D, immunoreactive C3b (red) was detected by immunohistochemistry in paraffin-embedded section of adipose from pair (P)-and ethanol (E)-fed wild-type (C) or cyp2e1 Ϫ/Ϫ or Bid Ϫ/Ϫ mice (D). DAPI (blue) was used to visualize nuclei. Chow-fed C3 Ϫ/Ϫ mice (C) were used as a negative control. Images are representative of at least four mice per diet group. and activity of CYP2E1 in adipose tissue led to an increase in Bid-dependent apoptosis, activation of complement via C1q, and increased inflammation during ethanol feeding. Taken together, these data elucidate a novel mechanism for adipose tissue inflammation (summarized in Fig. 6).
Ethanol exposure is associated with the development of oxidant stress in a number of tissues (12) including adipose tissue (3,4). In the liver, ethanol-induced oxidant stress is due at least in part to ethanol metabolism via CYP2E1 (12). Although white adipose tissue has negligible alcohol dehydrogenase activity (34), CYP2E1 mRNA and protein are expressed in adipose tissue (35) and can be induced by chronic ethanol exposure (4).
Here we also show that chronic ethanol feeding increases CYP2E1 activity in adipose tissue (Fig. 2). CYP2E1 protein was detected primarily in adipocytes and not stromal vascular cells. Making use of Cyp2e1 Ϫ/Ϫ mice, we found that CYP2E1 expres-sion was required for ethanol-induced apoptosis as well as increased expression of inflammatory cytokines and chemokines in adipose tissue. These data suggest that ethanol metabolism via CYP2E1 expressed in adipocytes was the initiating event in adipose tissue inflammation during ethanol feeding.
Increased production of reactive oxygen species during ethanol metabolism via CYP2E1 could directly lead to increased inflammatory cytokine expression via activation of redox-sensitive transcription factors (36). Alternatively, apoptosis of adipocytes in response to activation of cell death pathways could indirectly link CYP2E1 activity to inflammatory responses (13). Mice deficient in Bid were protected from both ethanol-induced apoptosis and increased expression of most inflammatory markers in adipose tissue. Only expression of TNF-␣ mRNA remained elevated in Bid-deficient mice, suggesting that, of the markers of inflammation measured, only TNF-␣ FIGURE 5. C1qa, but not MBLa/c or Cfd, contributes to chronic ethanol-induced adipose inflammation. Shown is expression of mRNA after 25 days of ethanol or pair feeding in adipose tissue from wild-type or knock-out mice deficient in specific pathways of complement activation (A and C) or triple-knock out (MBLa/c/Cfd Ϫ/Ϫ ) mice expressing only the classical pathway of complement activation (D). A, C, and D, mRNA was measured by quantitative real-time-PCR and normalized to 18 S. Data are expressed as -fold over pair-fed of the same genotype; there were no differences in expression between genotypes in pair-fed mice (data not shown). Values represent the means Ϯ S.E., n ϭ 3-6. *, p Ͻ 0.05 compared with pair-fed in each genotype. B, TUNEL-positive nuclei (red) were detected in adipose tissue from wild-type and C1qa Ϫ/Ϫ mice. Nuclei were labeled with DAPI (blue). TUNEL-positive nuclei were counted and expressed as percent of DAPI positive nuclei (wild-type pair-fed, 0 Ϯ 0; ethanol-fed. 21.0 Ϯ 5.1 (values represent means Ϯ S.E., n ϭ 4); C1qa Ϫ/Ϫ pair-fed, 4.9 Ϯ 1.0; ethanol-fed, 29.0 Ϯ 7.1 (values represent means Ϯ S.E., n ϭ 4).
mRNA expression was directly influenced by CYP2E1. These data are similar to the role for Bid-dependent apoptosis in mediating adipose tissue inflammation in the high fat diet-induced obesity model of non-alcoholic steatohepatitis (13). However, it is not known whether CYP2E1 is also involved in Bid-dependent adipose tissue inflammation in models of obesity.
Clearance of apoptotic cells can be a strong signal for induction of inflammatory responses (13,15). Crown-like structures are thought to be composed of dead/dying adipocytes surrounded by macrophages and other inflammatory cells; these inflammatory cells scavenge lipids from the adipocyte in a process analogous to formation of atherosclerotic foam cells (37). Complement proteins C1q and C3b act as opsonins to facilitate the clearance of apoptotic cell bodies, thus activating the complement cascade and initiating inflammatory responses (16,17). Our recent work has identified a critical link between ethanol-induced apoptosis in liver and complement activation via C1q (27); similar links between neuronal apoptosis, C1q, and inflammation are found in some murine models of Alzheimer disease (38). Interestingly, C1q can also play a protective role in models of Alzheimer disease by inhibiting LPS-stimulated cytokine expression (39). This differential response appears to be related both to the stage of apoptosis, with late apoptotic cells generating a stronger inflammatory signal as well as the differentiation state of the phagocytic cell interacting with the C1qopsonized cell (39). Binding of C1q to calreticulin is also linked to the induction of an immunogenic response to apoptotic cells (16).
In models of obesity-induced adipose tissue inflammation, macrophages exhibit a characteristic pattern of increased expression of F4/80, CD11b, and CD11c (40), shifting the macrophage polarization from alternatively activated (M2) to classically activated (M1) (41). M1 macrophages expressing CD11c are particularly important for the development of metabolic syndrome (32). After chronic ethanol feeding, CD11c expression, but not F4/80 or CD11b, was increased. These data suggest a strong proinflammatory milieu after ethanol feeding but clearly indicate that there are differences in the immune mileau in adipose tissue after chronic ethanol compared with diet-induced obesity. Given the influence of macrophage phenotype on the interaction between C1q and apoptotic cells observed in mouse brain (39), it will be interesting in future studies to compare the role of C1q in mediating obesity-induced versus ethanol-induced adipose tissue inflammation, particularly in relation to the phenotypic characteristics of the innate immune cells present in adipose tissue.
Although adipose tissue expresses a number of proteins in the complement pathway, the exact role of complement in adipose tissue and its relationship to innate immune responses continues to unfold. Complement appears to make fundamental contributions to the maintenance of normal biological function of adipose as well as contribute to local and systemic immune responses (19). Here we find for the first time that complement was activated in adipose tissue in response to ethanol feeding, evidenced by the accumulation of two C3 cleavage products, C3b and C3d, in adipose tissue from mice after ethanol feeding. Complement activation was dependent on the expression of CYP2E1 and Bid but preceded the appearance of most markers of inflammation in adipose tissue during ethanol exposure. Importantly, C1q and the classical pathway of complement activation were found to have a necessary and suffi- FIGURE 6. Schematic diagram illustrating the role of CYP2E1, Bid-dependent apoptosis, and complement activation in the development of adipose tissue inflammation during chronic ethanol feeding to mice. Chronic ethanol feeding induces the expression of CYP2E1 in adipocytes. CYP2E1 then leads to increased expression of TNF-␣, which in turn activates Bid-dependent apoptosis of adipocytes. C1q, a component of the classical pathway of complement, binds to cell surface markers on apoptotic adipocytes, resulting in complement activation. Complement activation products, interacting with the anaphylatoxin receptors, C3a receptor and C5a receptor, further increase cytokine and chemokine production in adipose. The cumulative result of this CYP2E1 3 Bid-dependent apoptosis 3 C1q-mediated complement activation pathway is the development of adipose tissue inflammation. cient role in adipose tissue inflammation in response to ethanol, thus providing a critical link between adipocyte apoptosis in response to ethanol and increased inflammatory responses. C1q is implicated in several chronic inflammatory diseases, typically associated with the clearance of dying cells, such as systemic lupus erythematosus (16), Alzheimer disease (38), and chronic ethanol-induced steatohepatitis (27). Previous studies have shown that both C1qa Ϫ/Ϫ (27) and Cyp2e1 Ϫ/Ϫ (42,43) mice are protected from chronic ethanol-induced liver injury, indicating that adipose tissue inflammation is associated with the severity of liver injury during chronic ethanol exposure.
Although multiple markers of ethanol-induced inflammation in adipose tissue were dependent on Bid-C1q, increased expression of TNF␣ and the anaphylatoxin receptors C3aR and C5aR was solely dependent on the expression of CYP2E1. These data suggest that this CYP2E1-dependent increase in C3aR and C5aR would act to enhance the sensitivity of adipose to complement activation and, thus, exacerbate inflammatory responses to complement activation.
Pathophysiological changes in the metabolic and innate immune activity of adipose tissue can lead to the development of insulin resistance and hepatic steatosis in alcoholic liver disease and metabolic syndrome/non-alcoholic steatohepatitis (2) and are associated with the regulation of appetitive behaviors, such as alcohol craving and dependence (10,11). Here we have identified a novel pathway for the development of adipose tissue inflammation in response to chronic ethanol. This pathway links the critical ethanol-metabolizing enzyme, CYP2E1, with apoptosis in adipose tissue. Apoptotic cells in adipose are immunogenic, leading to the activation of complement via C1q and the classical pathway and subsequent production of inflammatory cytokines and chemokines. Elucidation of this pathway in adipose tissue suggests novel therapeutic approaches to preventing and/or reversing the pathophysiological effects of ethanol.