Interaction between cytokines and inflammatory cells in islet dysfunction, insulin resistance, and vascular disease

Abstract

Inflammation is an established pathogenic player in insulin resistance, islet demise, and atherosclerosis. The complex interactions between cytokines, immune cells, and affected tissues result in sustained inflammation in diabetes and atherosclerosis. 12- and 15-lipoxygenase (LO), such as 12/15LO, produces a variety of metabolites through peroxidation of fatty acids and potentially contributes to the complex molecular crosstalk at the site of inflammation. 12- and 15-LO pathways are frequently activated in tissues affected by diabetes and atherosclerosis including adipose tissue, islets, and the vasculature. Moreover, mice with whole body and tissue-specific knockout of 12/15-LO are protected against insulin resistance, hyperglycemia, and atherosclerosis supporting functional contribution of 12- and 15-LO pathways in diabetes and atherosclerosis. Recently it has emerged that there is a temporal regulation of the particular isoforms of 12- and 15-LO in human adipose tissue and islets during the development of type 1 and type 2 diabetes and obesity. Analyses of tissues affected by diabetes and atherosclerosis also implied the roles of interleukin (IL)-12 and NADPH oxidase 1 (NOX1) in islets and IL-17A in atherosclerosis. Future studies should aim to test the efficacy of inhibitions of these mediators for treatment of diabetes and atherosclerosis.

Introduction

Inflammation, one of the body's basic responses to “harmful stimuli” has gained recognition as a key player in the pathogenesis of insulin resistance and atherosclerosis in recent years [1,2]. In addition, the contribution of cytokines and immune cells in islet demise has been implicated not only in type 1 diabetes (T1D) but also in type 2 diabetes (T2D, [3]). Current world-wide epidemic of obesity is expected to be followed by an increase in the incidence of T2D and cardiovascular disease for years to come. T1D, an autoimmune form of diabetes classically considered to be distinct from T2D, is also increasing coinciding with obesity epidemics [3]. Therefore, there is a heightened interest in understanding the mechanisms by which metabolic stress initiates inflammation in islets, insulin target tissues, and arteries. Unresolved inflammation is considered to impair function and integrity of tissues in diabetes and cardiovascular disease through complex interactions between cytokines, migratory immune cells, and cells that are normal constituents of each tissue. 12- and 15-lipoxygenases (LO) are oxidoreductases for arachidonic acid (AA) and other fatty acids with wide distribution and broad functions including regulation of inflammation (Figure 1). Here, we review accumulating evidence indicating that dysregulation of 12- and 15-LO pathways have an active role in the inflammatory processes associated with diabetes and cardiovascular disease in a wide range of tissues. We will similarly discuss additional pathways that can potentially crosstalk with the 12- and 15-LO pathway, and thus, may contribute to the development of diabetes and atherosclerosis.

Figure 1. Generation of lipid metabolites by 12- and 15-lipoxygenase.

A: arachidonic acid, ω-6 polyunsaturated fatty acid (20:4) is a substrate, 12- and 15-lipoxygenase (LO) produce 12(S)-hydroperoxyeicosatetraenoic acid (H(p)ETE) and 15(S)-H(p)ETE through oxidation at carbon-12 (9^th carbon from ω tail) and carbon-15 (6^th carbon from ω tail) respectively. H(p)ETEs are unstable intermediate metabolites and quickly converted to hydroxyeicosatetraenoic acids (HETEs). B: When docosahexaenoic acid (DHA), ω-3 polyunsaturated fatty acid (22:6) is a substrate, 12-LO activity produces 14(S)-hydroperoxy-docosahexaenoic-acid (H(p)DHA), followed by the formation of 14(S)-hydroxy-docosahexaenoic-acid (HDHA) through catalysis on carbon-14 (9^th carbon from ω tail). 17(S)-H(p)DHA followed by 17(S)-HDHA will be produced from DHA through catalysis on carbon-17 (6^th carbon from ω tail) by 15-LO activity. Different isoforms of 12- and 15-LO (Table 1) possess variable degree of combination of 12- and 15-LO activities. Also, 12- and 15-LO are active toward wide array of fatty acids 4. Here, the major metabolites relevant to the review are shown for clarity.

12- and 15-LO pathway in adipose tissue (AT): Possible contribution to insulin resistance and AT inflammation

The role of 12- and 15-LO pathway in tissue inflammation

12-LO and 15-LO, and their products, play important physiological and pathological roles in many tissues and organs, including AT, pancreatic islet, vasculature, brain, and kidney. The complex array of metabolites formed as a result of 12- and 15-LO catalytic activity are tissue and species-specific and can have both pro- and anti-inflammatory effects, although dysregulation of 12/15 LO pathway is considered to result in uncontrolled inflammation. Targeted deletion studies in mouse models have helped to identify the potential roles of these pathways in tissue inflammation. They are involved in the pathogenesis of various human diseases, including T1D and T2D and their complications, cardiovascular disease, renal disease, and the neurological conditions Parkinson's and Alzheimer's disease. An important barrier in the LO research has been substantial species-differences in substrate preference, lipid mediators, cellular expression and functional roles of the different LO isoforms, which complicates the extrapolation of animal studies to human pathophysiology [4]. For example, the mouse has six different 12- and 15-ALOX genes (note that the LOX genes are termed by convention “ALOX”, for arachidonic acid lipoxygenase) (Table 1). In humans, five 12- and 15-LO isoforms were identified: 12(S)-LO (also known as platelet-12LO or 12-LO-1); 12(R)-LO (or 12-LO-2); epidermis-type LO-3 (eLO-3); 15-LO-1 and 15-LO-2. The lipid products generated by the different isoforms may also differ between species [4,5]. For example, the mouse leukocyte 12-LO (12/15-LO, Table 1) generates both 12- and 15- hydroperoxyeicosatetraenoic acids (H(p)ETEs), while the human isoform (12(S)-LO, Table 1) only generates the 12-H(p)ETE product [4,6]. In humans, the two 15-LO isoforms have different substrate specificity and generate different lipid products. 15-LO-2 can only act on AA, while 15-LO-1 has AA, linoleic acid (LA) and docosahexaenoic acid (DHA) as substrates. When utilizing AA as substrate, 15-LO-1 produces 90% 15-H(p)ETE and 10% 12-H(p)ETE, while 15-LO-2 produces exclusively 15-H(p)ETE (Figure 1, [7,8]). Therefore, it is crucial to mechanistically dissociate the effects of the 12- and 15-LO pathways and their respective metabolites in order to understand their contributions to diabetes and cardiovascular disease

Table 1.

Human and mouse 12- and 15-lipoxygenase genes.

Lipoxygenase protein	Human gene	Mouse gene
Leukocyte-type 12-LO (15-LO-1 for human, 12/15-LO for mouse)	ALOX15	alox15
15-LO (15-LO-2)	ALOX15B	alox15b*
Platelet-type 12-LO (12(S)-LO)	ALOX12	alox12
12(R)-LO	ALOX12B	alox12b
Epidermal-type 12-LO	Not expressed	alox12e
Epidermis-type LO3 (eLO-3)	ALOXE3	aloxe3

^*mouse alox15b product functions as 8-LO (5).

The role of the 12- and 15-LO pathway in AT inflammation

AT inflammation is considered a hallmark of obesity and insulin resistance. The remarkable cellular complexity and plasticity of this tissue are reflected in multi-faceted roles for various inflammatory pathways that may be operational in more than one cell type and associated with both escalation and resolution of inflammation that accompany physiological or pathological tissue remodeling. There is emerging evidence that the various 12- and 15-LO isoforms are expressed in multiple cell types in both white (WAT) and brown fat: adipocytes, vascular cells, macrophages (MΦs) and pre-adipocytes.

The presence and roles of 12/15-LO in adipocytes has been extensively investigated by our group and by others in mouse models of obesity, insulin resistance, and T2D. Comparison of 12/15-LO knockout mice (12/15-LO KO) with C57BL/6J wild type mice fed either a standard chow or high-fat “western” type diet revealed that 12/15-LO is the primary enzyme generating the 12(S)-hydroxyeicosatetraenoic acids (HETEs) under obese conditions (Figure 1, [9]). A pro-inflammatory role of 12(S)-HETE is implicated since the increased 12/15-LO activity coincides with increased inflammation both systemically and within epididymal AT [9,10]. The increased expression of 12/15-LO was seen in isolated white adipocytes in C57BL/6J mice on a high-fat diet for 8-16 weeks supporting the contribution of adipocyte 12/15-LO [11]. Zucker obese rats, a genetically-induced rodent model of obesity and insulin resistance, also exhibit increased expression of 12/15-LO in isolated white adipocytes compared to lean controls [12]. More recent data from our lab showed that db/db mice undergo expressional increases of both 12/15-LO and platelet-12-LO in AT and isolated adipocytes that coincide with the metabolic decline (Figure 2A and B). To confirm a direct role of 12- and 15-LO activity in mediating inflammation and insulin resistance in adipocytes, the effect of 12-HETE, the major 12/15-LO metabolite in mouse, was examined in 3T3-L1 adipocytes by Chakrabarti et al.[11]. Addition of both the 12(S)-HETE and its precursor 12(S)-HpETE, to fully differentiated 3T3-L1 adipocyte cultures significantly induced pro-inflammatory gene expression and secretion of many pro-inflammatory cytokines, including TNFα, Chemokine (C-C motif) ligand 2 (CCL2), IL-6, and IL-12p40. Also, the anti-inflammatory adiponectin was significantly decreased under these conditions. Importantly, addition of the same metabolites led to an increase in activation of c-Jun N-terminal kinase 1 while insulin-mediated activation of key signaling proteins such as Akt and insulin receptor substrate (IRS)-1 was decreased. Addition of palmitic acid to 3T3L1 adipocytes increased 12/15-LO expression with concomitant increase in cytokine expression. Taken together, the activation of 12/15-LO in adipocytes seems to provoke a pro-inflammatory response and to impair insulin signaling.

Figure 2. Variation in adipose tissue expression of leukocyte-12/15-LO and platelet-12-LO and lipid metabolite profile indb/dbmice.

A: Phenotype of db/db mice (gray bar) and heterozygous +/db controls (black bar) at 5, 8 and 10 weeks of age (n=7-10 mice/group); at week 10 the db/db mice have full expression of obesity and are severely glucose intolerant; they also show a decline in the average islet number. B: Longitudinal expression of leukocyte-12/15-LO (L-12-LO) and platelet-12-LO (P-12-LO) in adipose tissue, isolated adipocytes and stromal vascular fraction (SVF); the two LO isoforms are abundantly expressed in SVF and show different patterns of expression. L-12-LO is increased early and diminishes with the decline in the metabolic phenotype; P-12-LO increases in parallel with the decline in the glucose tolerance and with beginning of islet loss. Gene expression was measured by real-time PCR using SYBR green primers and 18S as a housekeeping control (n=7-10/group). db/db mice in gray bars and controls in black bars. C: Lipid metabolite profile in the SVF of db/db mice (gray bar) and controls (black bar) at 8 and 10 weeks of age. Lipids were measured using LC-MS/MS on a Kinetec C18 column after extraction with methyl-tert-butyl- ether; chiral analysis was performed on the same system using a Chiralpack AD-LH column (n=7-10 mice/group); Data are mean ± s.e.m. Statistical analysis was performed using one-way ANOVA. HETE = hydroxyeicosatetraenoic acid; HDHA = hydroxy-docosahexanoic-acid.

Adipocyte dysfunction is not only marked by chronic inflammation and insulin resistance of WAT, but also by endoplasmic reticulum (ER) stress [13]. Obesity-associated ER stress has been shown to increase inflammation and insulin resistance [13]. Recent evidence from our lab demonstrates that 12/15-LO is a novel inflammatory pathway that mediates ER stress in the adipocyte [14]. 3T3-L1 adipocytes treated with 12(S)-HETE and 12(S)-HpETE displayed activation of the ER stress markers associated with each unfolding protein response pathway [14] and pre-treatment of cells with the 12/15-LO inhibitor, CDC (cinnamyl1-3, 4-dihdroxy-α-cyanocinnamate), led to a significant reduction of the tunicamycin-induced ER stress response [14]. Additionally, the comparison of isolated epididymal white adipocytes from C57BL/6J wild type or 12/15-LO KO treated with tunicamycin revealed that ER stress induction was significantly impaired in the absence of 12/15-LO [14].

To confirm that adipocyte-derived 12/15-LO plays a key role in obesity-induced insulin resistance, our lab has generated a fat-specific conditional-12/15-LO knockout mouse driven by the aP2-Cre transgene (aP2-12/15-LO KO). Control and aP2-12/15-LO KO were placed on either a chow or 60% high-fat diet for 16 weeks. 12/15-LO deletion from WAT was able to reduce inflammation and MΦ infiltration into the epididymal WAT [15]. Also, aP2-12/15-LO KO had reduced fasting blood glucose levels and non-fasting serum insulin levels compared to control mice on a high-fat diet. In addition these same mice exhibited improvements in insulin sensitivity and secretion as measured by glucose and insulin tolerance tests [15]. Inflammation in the pancreatic islet was also reduced in the high-fat diet-fed aP2-12/15-LO KO compared to controls [15]. These data suggest a crosstalk between 12/15-LO expression in WAT and inflammation in pancreatic tissue, revealing a considerable systemic impact of chronic 12/15-LO activity in fat in diet-induced obesity.

A lot less is known about the contribution of AT MΦ-derived 12/15-LO or the endothelial/smooth muscle enzymes on inflammation and insulin resistance. Also, less investigation has been devoted to the role of epidermal- and platelet-12-LO in obesity-induced adipocyte dysfunction. However, a recent paper from our lab has demonstrated that platelet-12-LO is upregulated in adipocytes from C57BL6/J mice fed a western diet (42 %kcal fat and 42.7% kcal carbohydrate) for 12 weeks. Interestingly, treatment with an angiotensin type 1 receptor (AT1R) blocker, valsartan, can abolish this effect [16]. It would be of interest to follow-up whether 12/15-LO is also regulated by the renin-angiotensin system (RAS) in AT, as much evidence reveals that 12- and 15-LO products upregulate RAS components, and in turn can be regulated by the RAS in several cell types [16].

Recent data collected in db/db mice showed that both 12/15-LO and platelet-12-LO are abundantly expressed in the AT stromal vascular fraction (SVF) containing vascular, immune, and progenitor cells (Figure 2B). Interestingly, 12/15-LO is significantly elevated in db/db mice compared to age-matched control heterozygotes even prior to development of the full metabolic phenotype (week 5). Also, 12/15-LO expression is reduced significantly at week 10 compared to the younger mice, which coincides with the onset of islet loss and a fully expressed metabolic phenotype (Figure 2A and B). A different pattern of expression was found for the platelet-12-LO which showed a significant increase at week 10 compared to younger animals and the control heterozygotes. This suggests a potentially critical pathogenic role of 12/15-LO that precedes the decline in glucose homeostasis, and a more prominent pro-inflammatory role of the platelet-12-LO isoform at the time of the T2D phenotype onset suggesting a potential contribution to the islet functional decline.

While abundant evidence emphasizes the pro-inflammatory role of the 12- and 15-LO pathway in AT in rodents [4], there is evidence that the 12- and 15-LO pathway may also generate metabolites that are key in the resolution of inflammation [17-19]. In particular, 12/15-LO may generate ω-6 polyunsaturated fatty acids (PUFAs), such as lipoxins, or ω-3 PUFAs, such as maresins, resolvins and protectins, in conjunction with the 5-lipoxygenase and cyclooxygenase (COX) enzymes [20,21]. AT dysfunction in obesity is likely the result of an inappropriate inflammatory response that remains uncontrolled due to intrinsic inability of the tissue to completely resolve inflammation. The pro-resolving metabolites formed by different LOs may therefore play beneficial effects for limiting inflammation in the AT provoked by nutritional overload. A recent publication showed that Resolvin D1 promotes resolution of AT inflammation in diet-induced obese mice by eliciting MΦ polarization toward an M2-like phenotype [22]. Also, a recent paper suggests that alternatively activated MΦ expressing 12/15-LO may act as a sink for distinct soluble receptors for apoptotic cells via controlled phagocytosis by the resident tissue MΦ [23].

The types and relative amounts of lipid metabolites generated in a tissue that expresses different 12- and15-LO isoforms are dependent not only on types and abundance of 12- and 15-LO isoforms expressed, but also on substrate availability, local metabolite concentration, or partial oxygen pressure. Increased availability of the ω-3 versus ω-6 fatty acids should lead to the predominant formation of lipid metabolites with anti-inflammatory properties. Interestingly, AT is the major storage site for PUFAs in obese individuals [24]. The 17-hydroxydocosahexaenoic acid (17HDHA) metabolite (Figure 1) along with protectin D1 and resolvin D1 were identified in AT of obese ob/ob mice [25]. Recent data from our laboratory also showed a variety of both pro- and anti-inflammatory lipid mediators in the SVF of db/db mice that showed a change in abundance as the mice became more glucose intolerant and insulin resistant (Figure 2A and C). Both the pro-inflammatory mediators derived from AA (12-HETE and 15-HETE) and the anti-inflammatory ones derived from the ω-3 PUFAs (EPA and DHA) were significantly higher in the db/db mice at all time points compared to heterozygous controls. This suggests that a number of anti-inflammatory resolving processes are associated with the stromal vascular component of the AT concomitant with the pro-inflammatory process, and it would be critical to determine the specific roles and contribution of these mediators to the overall inflammation and metabolic status, as well as the major cellular sources.

Important differences were emphasized in the 12- and15-LO pathway between rodents and humans [26]. The human 12(S)-LO was originally reported to be expressed in the platelets, endothelial and smooth muscle cells of large arteries, as well as in monocytes [27,28]. In a recent publication, we reported 12(S)-LO mRNA and protein expression in human AT with exclusive localization in the SVF both in the subcutaneous (SC) and in the omental (OM) fat [26]. This result does not recapitulate the localization in rodents, where adipocytes are an abundant source of 12/15-LO [11,12]. We have shown selective expression of the 15-LO isoforms in human visceral AT. Expression of 15-LO-2 was found in both SC and OM visceral human AT, exclusively localized in the SVF [26]. Intriguingly, while 15-LO-1 mRNA and protein expression were undetectable in the SC AT, they did exhibit robust expression in OM AT and only in cells of the SVF [26]. We also showed that all of the isoforms are expressed both in the CD34⁺ fraction of the SVF and in the CD34^- fraction containing monocytes and various lymphocytes [26]. Also, by immunohistochemistry we found robust expression of the 15-LO-1 in the AT vasculature. Increased expression of all of the 12- and15-LO enzyme isoforms was found in OM versus SC AT indicating the pathway may contribute to the pro-inflammatory milieu prominently associated with visceral fat in obesity [26]. In support of a pro-inflammatory role of 12- and 15-LO pathway in human visceral fat, gene array analysis of AT showed that AA metabolism is the second most significantly upregulated pathway in human OM compared to SC AT in human obese subjects with a 7.6-fold higher expression of 15-LO-1 in OM fat [29]. Importantly, we recently found using a lipidomic approach that both the 12- and 15-HETEs are significantly higher in OM compared to SC fat (Figure 3). At the same time, we and others [30] found 12- and 15-LO metabolites with anti-inflammatory properties in human fat (Figure 3). Taken together, the overall increase in the activity of the 12- and 15-LO pathway is seen in AT from obese human subjects. However, functional roles of the 12- and 15-LO pathway, and changes with different pathological conditions within human AT warrants future studies to identify the roles of different isoforms and the lipid mediators that are key for regulation of inflammation in human obesity and T2D.

Figure 3. Lipid profiling of ALOX12 and ALOX15 metabolites in subcutaneous (SC) and omental (OM) human adipose tissue.

Paired biopsies were collected during bariatric surgery from 12 morbidly obese subjects (BMI= 44.3±3.8). Total adipose tissue was subjected to lipid extraction using methanol/NaOH under reducing conditions and the lipids were measured by LC-ESI-MS/MS using a Triplequad instrument Agilent 6460/1200SL equipped with a Phenomenex Kinetex Column. Statistical analysis was performed using paired Student's t-test. HETE = hydroxyeicosatetraenoic acid; HDHA: hydroxy-docosahexanoic- acid; HODE = 13-hydroxy-octadecadienoic acid.

While understanding the mechanisms that regulates the 12/15-LO pathway require substantial future efforts, especially to translate rodent studies into humans, the contribution of this pathway to inflammation in AT in obesity is highly likely and may constitute a valuable therapeutic target.

12- and 15-LO pathway in pancreatic islets: Its contribution to islet dysfunction and β cell loss

Elevated inflammatory cytokines have been reported for both T1D and T2D (reviewed [3]). Ex vivo studies on human or mouse primary islets show that a brief (6 h) exposure to three inflammatory cytokines (TNFα, IL-1β, IFNγ) that are elevated in diabetes, results in loss of glucose-stimulated insulin secretion, increased expression of an array of genes, and induction of apoptosis [31,32]. Interestingly, the same inflammatory cytokine stimulation induces the activity of 12/15-LO (mouse) and 12(S)-LO (human) in islets and β-cells. An active lipid product of 12/15-LO and 12(S)-LO activity, 12(S)-HETE, can mimic, in part, cytokine effects on β-cells and induction of islet/β-cell dysfunction [33] (Taylor-Fishwick and Nadler, unpublished results). Conversely, inhibition of 12/15-LO and 12(S)-LO with novel selective molecules [34] protects islets/β-cells from inflammatory cytokine induced β-cell dysfunction (Taylor-Fishwick, unpublished results). These data strongly suggest that the 12- and 15-LO pathway is one mediator of inflammatory cytokine-induced β-cell dysfunction. Here, we discuss evidence supporting the contribution of the 12- and 15-LO pathway in islets in the development of both T1D and T2D along with emerging pathways that collaborate with 12- and 15-LO under cytokine-induced stress in islets.

The role of the 12- and 15-LO pathway in the development of T1D

We have appreciated for some time that the 12- and 15-LO pathway is involved in the pathogenesis of T1D. Initial studies showed that upon treatment with multiple low-dose streptozotocin, C57BL/6 mice deficient for 12/15-LO were protected from developing type 1-like diabetes [35]. These results spurred the development of a congenic line deficient in 12/15-LO on the non-obese diabetic (NOD) background [36]. Upon generation of this novel strain of mice with a global deficiency of 12/15-LO (NOD-Alox15^null congenic mice), we found that diabetes incidence was significantly reduced to less than 3% in females. Normally, NOD females develop diabetes at a rate of 60-90% following progressive leukocytic infiltration that begins around 4 weeks of age, depending upon the environmental conditions. It appears that 12/15-LO produced by cells within the islets contributes to the pathogenesis of T1D. Our initial findings indicated that both MΦ and islets of NOD mice express 12/15-LO [36]. Although studies are underway to more definitively tease out the independent roles for MΦ versus islet 12/15-LO, we recently reported on experiments that were designed to separate the function of MΦ from islet 12/15-LO. These recent studies indicated that 12/15-LO protein expression in islets peaked at 8 weeks of age [37]. Additionally, as early as 4 weeks of age, NOD-Alox15^null mice had significantly increased β-cell mass compared to NOD mice, suggesting that the 12- and 15-LO pathway contributes directly to islet dysfunction prior to lymphocytic infiltration. Downstream of 12/15-LO, we found evidence for increasing levels of pro-inflammatory cytokine mRNA expressed in the islets of NOD mice as they aged. Increases were seen in IL-12, IL-1β, IFNγ, and TNFα, as well as mediators of IL-12 signaling, including IL-12Rβ1, IL-12Rβ2, and signal transducer and activator of transcription 4 (STAT4). While it is certain that this is due in part to increased levels of infiltrating lymphocytes, there is also strong evidence that the β-cells themselves are producing some of the cytokines, as β-cell lines devoid of lymphocytes have been shown to generate cytokines, such as IL-12 [31,37]. This will be discussed in greater detail later in this review.

Concomitant with the increase in pro-inflammatory cytokines, we found significantly altered levels of islet health indicators in NOD islets when compared with age-matched NOD-Alox15^null islets. Specifically, NOD islets exhibited a significant loss of Pdx1 and Glut2 mRNA expression (important for islet function), and a significant increase in UCP2, indicating increased mitochondrial stress.

Furthermore, it is likely that expression of 12/15-LO in NOD mice increases ER stress, which has been shown to cause diminished islet health in NOD mice [38]. Although the link has not been directly made in the NOD strain, it has been shown that 12/15-LO activity can lead to increased ER stress in islets from C57BL/6 mice on a high-fat (HF) diet as we discussed [14]. In human islets, 12(S)-HETE, 12-LO products leads to decreased insulin secretion and reduced β-cell viability along with the activation of stress kinase p38-MAPK [33]. Decreased insulin secretion and reduced viability may occur through an increase in oxidative stress. 12- and 15-LO has also been shown to play a role in oxidative stress, as metabolism of AA by 12- and 15-LO can lead to an increase in NADPH oxidase-1 (NOX-1) as discussed below [32]. Taken together, the activation of the 12- and 15-LO pathway in islets likely contributes to autoimmune destruction of β-cells through multiple pathways, including pro-inflammatory cytokine production, oxidative stress, and ER stress (Figure 4).

Figure 4. Crosstalk between 12- and 15-lipoxygenase, cytokines, oxidative stress, and endoplasmic reticulum stress in cytokine-induced β-cell demise and dysfunction.

It has been well established that the combination of cytokines including TNFα, IL-1β, and INFγ results in β-cell demise and dysfunction. 12- and 15-lipoxygenase (LO) plays a role in the process through its crosstalk with cytokines, oxidative stress, and endoplasmic reticulum (ER) stress pathways. 12/15-LO plays the major role in mouse islets, while 12(S)-LO is important in human islets. Both are collectively represented as 12-LO in the figure. The cytokine mixture upregulates 12-LO and increases its proinflammatory metabolite 12(S)-HETE. 12(S)-HETE stimulates production of IL-12, activates p38-MAPK, and induces NADPH oxidase -1 (NOX-1) in islets. Although not directly shown in islets, there is evidence implicating the upregulation of ER stress by 12(S)-HETE in islets. 12-LO also increases IL-12 directly. Once induced, IL-12 increases the production of cytokines in islets and creates a feed-forward loop to amplify islet inflammation. The induction of NOX-1 will increase reactive oxygen species (ROS) in islets, cause oxidative stress, induce CCL2 expression in islets, and further augment islet inflammation. The cytokine mixture also causes ER stress directly or through the activation of 12-LO, which is shown as the increase in Bip, Xbp1, spliced form of Xbp1 (Xbp1s), and CHOP. Importantly, ER stress itself will activate 12-LO, creating another feed-forward loop. Insulin secretion will be impaired as a result of cytokine induction, ROS production and ER stress. Also, the activation of p38-MAPK, oxidative stress, and ER stress results in apoptosis. Therefore, 12-LO plays a significant role in amplifying cytokine effects by creating positive feedback loops at multiple pathways including oxidative stress, ER stress, and inflammatory responses.

One of the major challenges in understanding the pathogenesis of T1D has been difficulty in assessing alteration in human islets during the progression of disease. This is especially important for dissecting the roles of 12- and 15-LO pathways, since isoforms with different substrate and product specificities are expressed in mouse and human (Table 1). We have obtained human pancreatic sections from donors with established T1D, as well as those are positive for antibodies against islets (from Juvenile Diabetes Foundation, Network for Pancreatic Organ Donors with Diabetes (nPOD), University of Florida, Gainesville FL, USA). We have found that under inflammatory conditions, islets show strong positivity for 12(S)-LO (Morris, Nadler, unpublished results).

The role of the 12- and 15-LO pathway in the development of islet dysfunction in T2D

Given that the loss of islet mass and function is central to the development of T2D, factors that protect islets in the insulin resistant state are logical targets for treatment and prevention of T2D. There is substantial evidence that the 12- and 15-LO pathway serves as a promising target as one of the major class of enzymes involved in islet demise in T2D by provoking islet inflammation.

The expression of 12/15-LO in islets is increased in several animal models of T2D. 12/15-LO is one of the islet genes increased in Sprague-Dawley rats after 90% partial pancreatectomy [39]. Although the model is not associated with insulin resistance, rats become hyperglycaemic and show insulin secretory defects similar to those seen in diabetic humans [39]. The model suggests that hyperglycemia or reduction of islets mass, both commonly associated with the development of T2D, may be sufficient to increase the expression of 12/15-LO in islets in the absence of insulin resistance [39]. The expression of 12/15LO in islets is also increased in C57BL/6J mice fed western diet [9] and prediabetic ZDF rats [40], indicating 12/15-LO activation in islets in the presence of insulin resistance as well. What triggers the activation of 12/15-LO in islets in the insulin resistant state? One potential mediator is ER stress that is associated with T2D islets [3] and is known to upregulate 12/15-LO [14]. Cytokines, including IL-1β, TNFα, and IFNγ, are other potential mediators of the 12- and 15-LO pathway activation in islets as we discussed.

The functional contribution of 12/15-LO activation for the development of T2D was first shown when global 12/15-LO KO mice fed a western diet were protected against hyperglycaemia [9]. The improvement in glucose tolerance and insulin secretion seen in 12/15-LO KO may be partly due to reduced insulin resistance from the reduction of 12/15-LO activity in insulin target organs, especially from AT as we discussed. It has been proposed that the reduction of insulin resistance aids in the preservation of islet function through decreasing lipotoxicity, lowering demand for insulin, and reducing systemic inflammation [3]. To address the contribution of 12/15-LO in various tissues on islet dysfunction in obesity, we have created a series of mice with tissue-specific deletions of 12/15-LO. As we discussed, the targeted deletion of 12/15-LO in AT mediated by aP2-Cre (aP2-12/15-LO KO) not only reduced AT inflammation on a HF diet (58% kcal fat, 25.5% kcal carbohydrate) but also increased islet mass and percentage of β-cells in islets despite significant improvements in insulin sensitivity, implicating that 12/15-LO activity in AT may crosstalk with islets to influence their size. 12/15-LO and IL-6 mRNA expression was significantly reduced in islets of aP2-12/15LO KO on a HF diet indicating that local inflammation is reduced in islets after the deletion of 12/15-LO in AT. As 12/15-LO expression in islets from mice on regular chow diet was not altered in aP2-12/15LO KO, it is unlikely that the reduction in 12/15-LO is an aberrant silencing of 12/15-LO in β-cells. Also, peritoneal MΦ isolated from aP2-12/15-LO KO maintained 12/15-LO expression, arguing against leak of aP2-Cre in MΦ. In addition, islet-specific deletion of 12/15-LO by Pdx-Cre (Pdx-12/15-LO KO) protected mice against the development of hyperglycaemia on a HF diet despite a weight gain similar to that seen in wild-type mice on a HF diet and continued insulin resistance. The size and number of islets were increased in Pdx-12/15-LO KO (Tersey and Mirmira, University of Indiana, Nadler, unpublished results). Taken together, the suppression of 12/15-LO activity both in islets and AT seems to allow the expansion of islet mass and improve glucose homeostasis when challenged with a HF diet.

As discussed above, the activation of the 12- and 15-LO pathway in islets increases the production of pro-inflammatory cytokines and oxidative stress, resulting in islet demise. ER stress, another pathway known to contribute to islet dysfunction in T2D, is also strongly associated with 12- and 15-LO pathway activation (Figure 4). Islets from C57BL/6J placed on a HF diet exhibit signs of ER stress, including up-regulation of Bip, Chop, XBP-1 spliced variant, and ATF3. These responses were attenuated in 12/15-LO KO placed on a HF diet [14]. Since ER stress itself increases 12/15-LO expression, 12/15-LO activity and ER stress can create a feed-forward loop that results in a downward spiral of cellular dysfunction [14].

Emerging evidence supporting the potential of 12- and 15- LO targeted therapy in islet protection in T2D is that 12(S)-LO expression is increased in a certain group of human islets from donors with T2D (Imai, Nadler, unpublished results). When human islets from T2D donors are classified based on glucose-stimulated insulin secretion ex vivo, the expression of 12(S)-LO is increased in those with moderate impairment in insulin secretion. On the other hand, CCL2, a cytokine known to be increased in T2D islets [41] is expressed at higher levels in T2D islets that show markedly impaired insulin secretion (Imai, Nadler, unpublished results). There is an intriguing possibility that 12(S)-LO may be involved in an early stage of development of islet dysfunction at which patients still have a potential to reverse the disease. As discussed in the T1D section, the studies of pancreatic sections from humans with T1D and antibody positive donors show that islet cells strongly positive for 12(S)-LO are seen more frequently in antibody positive donors, the population enriched with preclinical T1D. However, 12(S)-LO positive cells are less prominent in donors with established T1D. The unique pattern of expression of 12(S)-LO in preclinical T1D and early T2D islets indicates the involvement of this enzyme in pathogenesis of both forms of diabetes at a certain stage. Interestingly, cells strongly positive for 12(S)-LO in islets of antibody positive donors are those that stain with pancreatic polypeptide (PP) (Morris, Nadler, unpublished results). It has been previously reported that the expression of PP is also increased in human T2D islets, albeit no clear increase in immunohistologically positive PP cells were noted in their group of donors [42]. As PP producing cells are developmentally situated as precursors of β-cells, at least in the mouse [43], and co-expression of PP genes with insulin genes are widely observed in both adult mice and mouse embryos [44], islet cells positive for 12(S)-LO and PP expression may define a special population of cells that are under stress during the development of both T1D and T2D. Of note recently, de-differentiation of β-cells defined as loss of insulin gene expression in previously insulin expressing cells is highlighted as an additional cause of functional β-cell loss under metabolic stress [45]. Given that 12/15-LO-deficiency in mice increases β-cell mass and improves glucose tolerance, the inhibition of 12-LO activity may serve as a novel therapy to prevent the loss of functional islet mass, especially in early stages of both T1D and T2D. We are collaborating with investigators at the University of California at Santa Cruz and National Institutes of Health to identify novel inhibitors of 12(S)-LO for possible testing in vivo.

Emerging pathways that mediate 12- and 15-LO actions in islets

Sensitivity to elevated intracellular reactive oxygen species (ROS) is a recognized fragility of β-cells [46]. This is due in part to pancreatic islets having low activity of free-radical detoxifying enzymes (e.g., catalase, superoxide dismutase, glutathione peroxidase) when compared to other tissues (reviewed, [3]). Islets are also very poor in rectifying oxidative damage to DNA [47]. Thus, under conditions of sustained activation of intracellular ROS, islets are readily overwhelmed and undergo oxidative stress [46]. Serum conditions associated with diabetes, which include increased inflammatory cytokines, high FFA, and elevated glucose levels are all potent inducers of elevated cellular ROS [48-55]. Candidate cellular sources of ROS in the β-cell have been identified and include induced mitochondrial stress (reviewed, [56]) and ER stress (reviewed, [57]). However, identification of NADPH oxidase complexes in β-cells has raised the possibility of a third contributing source for elevated ROS in β-cells. The NOX family of NADPH oxidases is comprised of proteins that transfer electrons across biological membranes (plasma or organelle). Their function is the generation of ROS, superoxide, and hydrogen peroxide (H₂O₂). This contrasts with mitochondrial or ER stress where ROS generation is a byproduct. To date, five NOX complexes have been identified termed NOX-1 through -5. In the β-cell, phagocyte NADPH oxidase (NOX-2) activity has been associated with mitochondrial dysregulation, although a physiological role for NOX-2 has been indicated in the transient ROS elevation driving insulin secretion [58-60]. A pathological role for NOX-derived ROS in β-cells is suggested with a linkage of NOX-2 activity to β-cell dysfunction induced by very low-density lipoprotein or FFA [61,62]. Detection of NOX-1 and NOX-4 has additionally been reported in β-cells (reviewed, [63]). Exposure of β-cells to the triple inflammatory cytokines that induce β-cell dysfunction leads to activation of NOX, and induction of intracellular ROS. Inflammatory cytokine stimulation (TNFα, IL-1β, IFNγ) of human or mouse primary islets, and mouse or rat β-cell lines, identified a preferential induction in the expression of NOX-1 in β-cells [32]. Thus, NOX-1 may also contribute to β-cell dysfunction (Figure 4). Intriguingly, efforts to modulate NOX-1 activity, in order to evaluate the role of the molecule in β-cell dysfunction mediated by inflammatory cytokines, has exposed a potential key role for NOX-1 in β-cell pathology. These studies evidence a model whereby a component of NOX-1 expression derives from a self-sustained upregulation of NOX-1, mediated by elevated ROS and second messengers [64]. The primary clue for a feed-forward regulation of NOX-1 in β-cells arose from the observation that an induced expression of NOX-1 in β-cells was abrogated with inhibitors of NADPH oxidase activity. Since NOX activity is secondary to induced NOX-1 gene expression, the most straightforward explanation of the data is that NOX activity upregulates NOX-1 gene expression, a feed-forward control. NOX-1 induction of ROS was further implicated since anti-oxidants (which neutralize cellular ROS) inhibited NOX-1 expression and pro-oxidants (that directly elevate cellular ROS) induced NOX-1 expression [64]. Redox-sensitive kinases, including Src-kinase, are shown by inhibitor studies to mediate feed-forward regulation of NOX-1 [64]. Importantly, markers of β-cell dysfunction following inflammatory cytokine stimulation, loss of glucose stimulated-insulin secretion, increased gene expression and induction of apoptosis, are all preserved by disruption of the NOX-1 feed-forward regulation ([64], Taylor-Fishwick, unpublished results). In β-cells, which have a low anti-oxidant capacity relative to metabolic activity, a feed-forward control of ROS production could rapidly result in a pathological state. Consequently, identifying and inhibiting such regulation presents an attractive target in developing new strategies for preservation and protection of functional β-cell mass in diabetes. With regards to mediators of inflammatory cytokine-induced NOX-1, common outcomes from manipulation of 12- and 15- LO activity and NOX-1 activity have been directly investigated [32]. These studies have identified an association between the 12- and 15- LO pathway and NOX-1 pathway [32]. Direct addition of 12-HETE to β-cells induced NOX-1 expression. Conversely, inhibition of 12(S)-LO activity, with selective small molecules [34], blocked the induction of NOX-1 by inflammatory cytokines [32]. These data integrate inflammatory cytokine stimulation with 12- and 15- LO activity to upregulation of NOX-1 in β-cells: a pathway that is linked to β-cell dysfunction.

An additional new piece to the molecular jigsaw stems from our recent description of interleukin-12 (IL-12) production and function in β-cells [31]. An upregulation in the IL-12 gene and protein expression is observed following inflammatory cytokine stimulation (TNFα, IL-1β, IFNγ) of primary human and mouse islets, or homogeneous β-cell lines (Figure 4). IL-12 is classically produced by immune cells. However, these β-cell lines are devoid of any immune cell ‘contamination’. While this is a concept that challenges established immune-based sources of IL-12, a local β-cell production of IL-12 could play a significant role in targeting additional immune mediator recruitment to an inflammatory β-cell microenvironment. Consequentially, a paracrine function for IL-12 in β-cells is also supported [31]. β-cells, including human β-receptor and are responsive to IL-12 ligand/IL-12 receptor ligation. Both IL-12 ligand and receptor are upregulated in β-cells exposed to inflammatory cytokine stimulation (TNFα, IL-1β, IFNγ). IL-12 induced a dose-dependent expression of IFNγ in β-cell lines suggesting a functional IL-12-STAT4-IFNγ axis. Previous studies have evidenced an active STAT4 signaling pathway in islet β-cells [65,66]. New inhibitors of STAT4 help to preserve β-cell function in the presence of inflammatory cytokines (Taylor-Fishwick, unpublished results). IL-12 directly mediated β-cell dysfunction, including induction of apoptosis and disruption of glucose-stimulated insulin secretion [31]. These functional defects corresponded to those seen with inflammatory cytokine stimulation. Importantly, neutralization of IL-12 with an IL-12 antibody blocked β-cell dysfunction induced by inflammatory cytokine stimulation [31]. These data suggest that induction of IL-12 by inflammatory cytokines may be a mediator of inflammatory cytokine-induced β-cell dysfunction. Lastly, novel selective inhibitors of 12(S)-LO [34], suppress induction of IL-12 ligand in islets and β-cells exposed to inflammatory cytokine stimulation (Taylor-Fishwick and Nadler, unpublished). Thus, inflammatory cytokine stimulation of 12(S)-LO activity is implied in the regulation of β-cell IL-12 expression.

Taken together, assembly of the pieces of this puzzle of cytokine-induced β-cell damage is emboldened by the interrelated observations. Rather than a single jigsaw piece, the integration of several pieces to form an actual picture is encouraging. The deep literature supporting each component is pertinent, as are the promise of new inhibitors being studied. Combining a strategy to preserve and protect β-cells in an inflammatory environment with a strategy to enhance β-cell mass holds promise for an effective diabetic therapy.

Inflammatory mediators in atherosclerosis

One of the common features of the metabolic syndrome and cardiovascular disease is the accumulation of toxic lipid metabolites within adipocytes, β-cells, hepatocytes, and vascular endothelial and smooth muscle cells; which results in the induction and persistence of a low-grade inflammation within AT, islets, liver, and the vasculature. Multiple lines of evidence have shown that the immune system does play an integral role in the pathology of atherosclerosis. Indeed, several pro-inflammatory cytokines may act at the level of the vasculature to promote inflammation in diabetes; therefore, the functions of pro-inflammatory cytokines in diabetes-accelerated atherosclerosis are worth considering. Moreover, from a therapeutic point of view, an anti-inflammatory regimen for diabetes which has cardioprotective features in addition to improving insulin sensitivity and islet function is highly desirable. Therefore, it is important to delineate the involvement of cytokines and immune cells in the cascade of events that results in atheroma formation.

The role of the 12- and 15-LO pathway in vascular pathophysiology is complex. 12- and 15- LO catalytic activity results in the formation of a complex array of pro- and/or anti-inflammatory metabolites depending on vessel location, vascular cell type, and intracellular redox state. The complex roles of 12- and 15-LO metabolites on vascular reactivity, atherosclerosis, and angiogenesis have been discussed in a recent review [4]. Overall, mouse studies indicate that inhibition of the 12/15-LO enzyme is protective against the development of atherosclerosis. High-fat diet-fed apoE^-/-, LDLr^-/- and apobec^-/-/LDLr^-/- mice lacking the alox15 gene that encodes the 12/15-LO enzyme consistently showed reduction of atherosclerosis compared to controls (reviewed, [4]). A recent publication demonstrated that LDLr^-/- and alox15^-/- mice fed an atherogenic diet rich in PUFAs led to decreased atherosclerotic lesion development compared to single LDLr^-/- mice. The atheroprotective function of 12/15-LO deficiency is proposed to be through loss of 12/15-LO activity in MΦ and consequent reduction of 12/15-LO pro-inflammatory metabolites [67]. If proven to be atheroprotective in humans, inhibitors of 12- and 15-LO activity for the treatment of diabetes may also confer additional cardiovascular protective effects. However, further studies are required before extrapolations from mouse studies to humans can be made as human isoforms for 12-and 15-LO are not identical in mice (Table 1). Recent studies of human atheroma plaques within coronary arteries of diabetics reveal that polymorphisms in ALOX12 are associated with atherosclerosis and T2D [68]. While studies suggest that 12- and 15-LO may be atherogenic in the context of inflammation, the precise role of 12- and 15-LO during atherogenesis in humans remains to be determined.

The validity of anti-inflammatory therapy in atherosclerosis has been pursued through multiple potential targets. The role of Interleukin 17A (IL-17A) in atherosclerosis has been proposed by several laboratories [69]. IL-17A, a hallmark cytokine of T helper 17 (Th17) cells, is an important mediator of inflammation in multiple models of autoimmune and inflammatory disorders. Interestingly, several lines of evidence suggest that Th17 cells may be involved in the pathology of T2D, hypertension, and atherosclerosis. Plasma IL-17A and peripheral blood Th17 numbers positively correlate with both obesity and hemoglobin A1C levels within type 2 diabetic individuals [70]. In addition, neutralization of IL-17A in KK-Ay- and angiotensin 1-infused C57BL/6J mice, two mouse models of insulin resistance, significantly increased glucose uptake within skeletal muscle, attenuated the increase in blood glucose levels after an oral glucose challenge, and decreased plasma TNFα levels [71]. Similarly, in an angiotensin II-dependent model of hypertension, neutralization of IL-17A lowered the blood pressure and restored acetylcholine-dependent vasorelaxation in infused C57BL/6J mice, and AT-II-infused Il17a^-/- mice displayed a reduction in aortic T cell infiltration [72], suggesting that Th17 cells are involved in the pathology of hypertension. Similarly, several groups have examined the involvement of IL-17A in atherosclerosis; however, the results have been somehow controversial with multiple studies suggesting a pro-atherogenic role for IL-17A and others proposing a protective role or no effect on atherosclerosis (reviewed,[69]). To directly examine the role of IL-17A and its receptor IL-17RA in Apoe^-/- mice, our group recently generated Il17a^-/-Apoe^-/- and Il17ra^-/-Apoe^-/- mice [73]. Interestingly, Il17a^-/-Apoe^-/- and Il17ra^-/-Apoe^-/- mice on a 12-15 week western diet displayed a ∼20% reduction in aortic root lesions and a ∼35% reduction in aortic lesions, specifically within the aortic arch. Examination of the mechanism behind the site-specific effects of IL-17A revealed that while IL-17RA was present throughout the aorta, IL-17A-producing cells were only detectable within the aortic arch of 12-week western diet-fed Apoe^-/- mice. In addition, as Il17a^-/-Apoe^-/- mice displayed reduced expression of multiple pro-inflammatory cytokines and chemokines and possessed a lower number of aortic infiltrating macrophages and neutrophils, we hypothesized that IL-17A may promote monocyte and neutrophil recruitment to the aorta through the production of aortic chemokines. To test our hypothesis, we adoptively transferred Apoe^-/- peripheral blood leukocytes to western diet-fed Apoe^-/- and Il17ra^-/-Apoe^-/- mice and examined the migration of transferred cells to the aorta. In this system, Apoe^-/- recipients were able to recruit significantly more monocytes and neutrophils than their Il17ra^-/-Apoe^-/- counterparts, suggesting that IL-17A promotes atherogenesis through the production of aortic pro-inflammatory chemokines and the recruitment of monocytes and neutrophils to atherosclerotic plaques. Thus IL-17A, which is elevated in T2D patients, may serve to help accelerate atherosclerosis through the activation of arterial endothelial and smooth muscle cells and the recruitment of monocytes and neutrophils to the atherosclerotic sites and possibly other sites of inflammation as well.

Given that both the immune system and the 12- and 15-LO pathway play significant roles in the pathophysiology of vascular disease and diabetes, understanding the convergence of these pathways will provide better understanding for designing future therapeutic interventions for both diseases.

Concluding remarks

Data from both humans and rodents support that the activation of the 12- and 15- LO pathway, especially that of 12/15-LO and 12(S)-LO, is a common phenomenon in AT under nutritional overload in T2D and islets in both forms of diabetes. Moreover, various T1D and T2D rodent models of 12/15-LO deficiency show improvement in glucose homeostasis providing a rationale to target the 12- and 15-LO pathway for reduction of insulin resistance and preservation of functional islets in diabetes. The atheroprotective phenotype implicated in 12/15-LO-deficient mice provides an added benefit to targeting the 12- and 15- LO pathway in diabetes and associated cardiovascular disease. There has been progress into clarifying the roles of particular 12- and 15-LO isoforms in tissues affected with diabetes and obesity in humans. This has paved a path to formulate specific pharmacologic inhibitors for each of the LO isoforms in humans affected by insulin resistance and diabetes. Current intensive investigation is underway with an imperative goal to identify effective, small molecule inhibitors [74]. Finally, the investigation of pathways that crosstalk with the 12- and 15-LO pathway continues to be important for improving our understanding of complex interaction between cytokines, immune cells, and tissues during inflammatory responses associated with insulin resistance, islet demise, and atherosclerosis.