Regulation of hormone-sensitive lipase in islets

1. Introduction 

Hormone-sensitive lipase (HSL) is an intracellular neutral lipase that hydrolyzes stored fat in the form of triacylglycerol (TG), diacylglycerol (DAG) and monoacyglycerol, as well as cholesteryl ester, and releases free fatty acids (FFA) [1]. HSL is predominantly expressed in adipose tissue, and it responds to fast acting hormones through changes in its phosphorylation state, its oligomeric status and its subcellular localization, resulting in a rapid regulation of its hydrolytic activity [1]. For instance, ACTH and catecholamines stimulate its activity, whereas insulin decreases its activity. Therefore, HSL is considered an important regulator of cellular lipid homeostasis and a major modulator of plasma FFA. Although the mechanism is not fully understood, FFA have been shown to be involved in the process of glucose stimulated insulin secretion (GSIS), probably through metabolites that could affect FFA partitioning, such as malonyl-CoA and long-chain acyl-CoA, or DAG, which is both a substrate and a product of HSL hydrolysis. Accumulating data suggest that these lipid derived signals are important for proper GSIS control [2], [3], [4], [5]. Malonyl-CoA has been shown to inhibit carnitine palmitoyltransferase I, thereby enhancing insulin secretion. Long-chain acyl-CoA acutely enhances exocytosis in insulin-secreting cells [6]. DAG has been shown to activate isoforms of PKC, can serve as a substrate for arachidonic acid production and can bind Munc-13, a protein involved in regulated exocytosis [7], [8], [9], and hence plays an important role in regulation of insulin secretion in β-cells [10], [11], [12]. Therefore, proper lipid homeostasis is needed for the normal control of insulin secretion; any disturbances in lipid metabolism in the β-cell induced by high glucose, increased lipid esterification, or protein acylation, might modulate KATP channel-independent insulin secretion.

HSL appears to be expressed as an unique isoform in pancreatic β-cells [13], and the expression of β-cell HSL has been shown to be induced by long-term exposure to high glucose [14]. However, conflicting data have been reported by several investigators in regard to the role of HSL in insulin secretion. Initial observations [15] showed disturbed GSIS in HSL−/− mice, but a recent report from another group [16] showed only impaired insulin sensitivity with intact insulin secretion. All of the HSL−/− mice generated are normoglycemic under basal conditions (normal chow feeding condition), but show a decrease in circulating FFA levels and intolerance to glucose when fasted. Islet TG content was also reported to be increased 2–2.5-fold in HSL−/− mice compared to wild type mice. Therefore, the ablation of the HSL gene has an impact on the normal function of islets in mice, however, whether insulin secretion is affected is debatable. Under basal condition (normal chow feeding condition), there is an increased serum insulin level in HSL−/− mice; after overnight fasting, the serum insulin level is lower in HSL−/− than control mice [17]. As to the role of HSL in insulin secretion, controversial data have been presented as well. Orlistat, a lipase inhibitor, has been shown to inhibit forsklin-stimulated insulin release from rat islets [18]. However, data obtained using isolated islets from 4-month-old HSL−/− mice showed that GLP-1 stimulated insulin release seemed to be intact. In the same publication, decreased GLP-1 stimulated insulin secretion in islets isolated from 7 months old HSL−/− was reported compared with wild type mice, suggesting a compensated insulintropic effect of GLP-1 in HSL−/− mice [19]. In addition to the results obtained with HSL−/− mice, treating isolated rat islets in vitro with a HSL specific inhibitor impaired GLP-1 as well as foskolin potentiated GSIS [20]. Taken together, these findings suggest that HSL could be important in maintaining lipid homeostasis in β-cells and, hence, the regulation of GSIS. In this report, we studied the regulation of HSL activity in β-cells in relation to GSIS under different conditions and analyzed the changes in gene expression profiles in the islets of HSL deficient mice.

3. Results 

3.1. FFA loading increases the expression and activity of HSL in clonal β-cells 

To determine whether increased cellular TG alters HSL expression in β-cells, INS-1 cells were incubated with 1mM oleic acid for different times in RPMI-1640 (with 10% fetal bovine serum, 11.1mM glucose). Cells were then harvested and protein extracted. Immunoblotting using anti-HSL antibody showed that both the 84 and 89kDa isoform of HSL are expressed in pancreatic β-cells and adipose tissue. There was an increase of the β-cell specific 89kDa HSL in the cells that were loaded with 1mM oleic acid overnight (Fig. 1(a) lane3) compared with no FFA loading (lane 2). Lane 1 is a sample from adipose tissue. Equal amounts of total protein were loaded in lanes 2 and 3; the expression of the 84kDa isoform of HSL did not change with treatment. The TG content in the cells that were loaded with FFA is 1.6-fold higher than cells not treated with FFA (data not shown). This result shows that FFA loading increases the cellular TG content of β-cells and, at the same time, the expression of the β-cell form of HSL. In separate sets of experiments, INS-1 cells were loaded with different concentrations of oleate for overnight before harvesting. To eliminate the possible inhibitory effect of FFA in the cell extract on HSL hydrolytic activity, the procedure for measuring HSL hydrolytic activity was modified to include an additional step of protein precipitation under acidic conditions (pH 5.4). Proteins were washed and resuspended in buffer for measurement of activity. Neutral cholesterol ester hydrolase (NCEH) activity was assayed to represent HSL activity, since the NCEH activity in islets of HSL−/− mice is less than 15% of that of wild type mice (see Fig. 3(a)). Results show that (Fig. 1(b)) there is a more than two-fold increase of HSL activity with increasing concentrations of oleate to 0.3mM in the loading media after overnight loading compared to no FFA loading (p<0.005). When the concentrations of oleate in the loading media increased to 0.4, 0.5 and 1mM, there are still significant increase of HSL activity as compare with no FFA loading (p<0.05), although higher concentrations of oleate in the loading media do not increase HSL activity further.

View Large Image Download to PowerPoint

Fig. 1. FFA loading increases the expression and activity of HSL in clonal β-cells: (a) INS-1 cells were incubated with (lane 3) or without (lane 2) 1mM oleic acid for overnight in RPMI 1640 media. Cell were extracted in TES and blotted for immunoreactive protein with anti-HSL antibody. Lane 1 is a sample from adipocyte tissue. The results are representation of three sets of independent experiments. (b) INS-1 cells were incubated with different concentrations of oleate in RPMI 1640 media for overnight. Cells were washed with PBS and extracted in TES buffer. Proteins were precipitated at pH 5.4 before assay for HSL activity, results are means and standard deviations of four independent assays. *p<0.05, ***p<0.005 by t test.

3.2. GLP-1 increases HSL hydrolytic activity in INS-1 cells 

Recently, it was shown that a specific inhibitor of HSL can block GLP-1 potentiated GSIS in isolated rat islets [20], which further supports the importance of HSL in insulin secretion. In order to determine whether GLP-1 has any direct effects on HSL activity in β-cells, INS-1 cells were grown overnight in high glucose (20mM) to increase the amount of HSL protein. Cells were then washed with PBS, incubated in Kreb's Ringer buffer with 0.2% BSA and 2.5mM glucose, and then treated with different amounts of GLP-1 in low (2.5mM) or high (20mM) glucose for 1h. Media and cells were harvested for assay of insulin secretion and HSL activity, respectively. As shown in Fig. 2, treatment with 50nM GLP-1 resulted in increased HSL activity under both low (2.5mM) and high (20mM) glucose (Fig. 2(a)). With 2.5mM glucose, 50nM of GLP-1 treatment increased HSL activity almost three-fold (p<0.005); while with 20mM glucose there is an 60% increase in HSL activity above the increase of HSL activity induced by high glucose (p=0.013). The addition of a cAMP analogue (CPT) along with 50nM GLP-1 did not further increase HSL activity. Measurement of insulin secretion show that treatment of INS-1 cell with 50nM GLP-1 in high glucose lead to almost four-fold (p<0.05) increase in insulin secretion (Fig. 2(b)). No changes in HSL protein expression were observed under these conditions (Fig. 2(c)).

View Large Image Download to PowerPoint

Fig. 2. GLP-1 increase HSL hydrolytic activity in INS-1 cells: INS-1 cells were grown to 80% confluence in RPMI-1640 media supplemented with 11.1mM glucose and then switched to 20mM glucose overnight. Cells were then washed twice with PBS and then incubated twice for 30min in Kreb's Ringer buffer with 0.2% BSA and 2.5mM glucose before treatment with different amounts of GLP-1 and cAMP in low (2.5mM) or high (20mM) glucose for 1h. Media were collected for measurement of insulin (b) using RIA kit from Linco Research. Cells were harvested for assay of HSL activity (a), and HSL immunoreactive protein (c). Results presented are means and standard deviations of three independent experiments. *p<0.05, **p<0.01, ***p<0.005 by t test.

3.3. Islets of HSL−/− mice have decreased neutral cholesterol ester hydrolase activity and show decreased GLP-1 potentiated insulin secretion 

To study the contribution of HSL to islet lipase activity and GLP-1 potentiated insulin secretion, islets of wild type and HSL−/− mice were isolated (Fig. 3). Measurement of HSL activity using CE as substrate showed that the islets from HSL−/− mice retained <20% activity of control mice (p<0.05). However, using TG as substrate, activity was reduced only 30–40% (p<0.05), suggesting the presence in the islets of one or more other neutral TG lipases. This is consistent with previous papers showing decreased NCEH activity and TG lipase activity [19], [26]. It should be noted that glycerol release in islets was unaffected by the absence of HSL [19], [26]. Treatment of wild type islets with GLP-1 in the presence of high glucose showed a more than five-fold increase in insulin secretion (p<0.05). Islets from HSL−/− mice have higher basal insulin secretion (p<0.01), but fail to increase insulin secretion further in response to high glucose treatment. Addition of GLP-1 to the high glucose condition for the islets from HSL−/− mice showed a trend of increased insulin secretion, but did not reach statistical significance; the increase failed to reach the same level as that of islets from wild type mice (p=0.05).

View Large Image Download to PowerPoint

Fig. 3. Islets of HSL−/− mice lack NCEH activity, have decreased TG hydrolase activity and show decreased GLP-1 potentiated insulin secretion. Islets were isolated from 6 months old female HSL−/− and wild type mice. After overnight recovery in RPMI media with 10% FBS, islets were harvested and sonicated in TES buffer for HSL activity using cholesterol oleate (a) or triolein as substrate (b). (c) Islets were washed with PBS with 0.2% BSA and transferred to individual 12 well plates in groups of 50 and treated with low (2.5mM) or high glucose (20mM) with and without GLP-1 (50nM) in Kreb's Ringer's buffer with 0.2% BSA. After 1h treatment, media were harvested for assay of insulin secretion. Results presented are means and standard deviations of three sets of experiments. *p<0.05, **p<0.01 by t test.

3.4. Histopathology of HSL−/− mice 

HSL−/− mice display hypertrophy of adipocytes and accumulation of lipid in adrenal glands [27], [28]. H&E-stained sections of pancreata from HSL−/− mice showed hyperplasia of pancreatic islets (Fig. 4). Under both normal chow and high fat feeding condition, the islets of HSL−/− mice were significantly larger than islets of WT mice (normal chow p<0.005; high fat p<0.05). Mice on a high fat diet had larger islets than mice of the same genotype on a normal chow diet; however, the size difference was only significant for the HSL−/− mice (p<0.05). There were no differences in islet size between male and female mice (data not shown). Moreover, there were no differences between groups in the percentage of islet cells that expressed insulin (data not shown).

View Large Image Download to PowerPoint

Fig. 4. Histological changes in pancreatic islets. (A–H) Sections of pancreas from 27 week-old male mice in the fed state are shown. (A–D) haematoxylin and eosin staining; (E–H) anti-insulin immunostaining; (A and E), Normal chow (NC)-fed HSL+/+; (B and F) NC-fed HSL−/−; (C and G) High fat (HF)-fed HSL+/+; (D and H) HF-fed HSL−/−. Original magnification, ×10. Scale bars, 100μm. (I) Morphometric analysis was used to determine the area of each islet on three non-sequential sections of H&E-stained pancreata section, at least 10 islets were evaluated for each mouse. Results are presented as mean±S.D. for four mice in each group. *p<0.05, ***p<0.001, t test.

There were no significant differences between groups in the mean area of an islet cell (WT normal chow=99.1±3.28μm2; WT high fat=98.6±2.75μm2; HSL−/− normal chow=98.6±4.93μm2; HSL−/− high fat=95.9±4.19μm2; mean±S.D.; n=4 mice/group). Thus, the increase in islet size in the HSL−/− mice (as compared to WT mice) and in mice on a high fat diet (as compared to mice on normal chow) was due to an increase in the number of cells within each islet (hyperplasia), rather than to an increase in size (hypertrophy) of individual islet cells.

3.5. Gene expression profile in islets of HSL−/− mice and wild type mice 

In view of the morphological and biochemical changes in islets from HSL−/− mice, the gene expression profiles in islets isolated from HSL−/− mice was compared with wild type mice using the PancChip 2.2 from NIDDK. The PancChip 2.2 microarray contains three copies of 3139 non-redundant mouse IMAGE clones that were chosen using a combination of expression analysis and database mining [29]. Eight sets of PancChip 2.2 array experiments were performed; results were analyzed using significance analysis of microarray (SAM) [30] and summarized in Table 1. Significant changes in expression were observed in genes that are involved in metabolism, such as isocitrate dehydrogenase and HMG-CoA lyase precursor, which are involved in fatty acid oxidation and are both up-regulated 2-fold in HSL−/− mice (p<0.05). β-hydroxysteroid dehydrogenase, squalene synthase, and acryl-acylamidase, which are involved in lipid metabolism, also showed significant regulation. β-hydroxysteroid dehydrogenase and squalene synthase were down regulated 2.3–2.5-fold (p<0.05). Acryl-acylamidase, which can function as a microsomal lipase, was up-regulated 2.7-fold (p<0.05). Twelve genes involved in basic cellular functions, which include heat shock protein 84, synexin, and lysyl hydrolase isoform 2, also showed significant changes in gene expression (p<0.05). Other genes that showed changes in expression include genes involved in signal transduction (casein kinase II β-subunit, increased two-fold; serine/threonine kinase, increased 2.5-fold, p<0.05), genes involved in insulin secretion and signaling (succinyl-CoA synthase β-subunit precursor, increased two-fold p<0.05), as well as immune related genes (transforming protein RFP, increased two-fold; IgE-binding factor, decreased 2.8-fold; p<0.05). Thus, of the 3139 non-redundant mouse IMAGE clones contained on the PancChip 2.2 array, the expression of 17 was increased more than two-fold, while six were decreased more than two-fold.

Table 1. Microarray comparison analysis of gene expression in islets between wild type and HSL−/− mice

	Category	Gene name	Fold (KO/WT) Δ	Accession number	Function
	Metabolism	Similar to isocitrate dehydrogenase	2.08	AI173556	Fatty acid oxidation
		Branched chain α-ketoacid decaboxylase E1a	2.01	AI62247	Ketoacid metabolism
		Similar to HMG-CoA lyase precursor	2.00	AA838929	Fatty acid oxidation
		Uricase	2.72	AA237345	Response to oxidative reaction
	Lipid metabolism	β-Hydroxysteroid dehydrogenase	−2.48	AA450665	Facilitate corticosteroid action
		Squalene synthase	−2.26	Ai195100	Cholesterogenic
		Acryl-acylamidase	2.66	A53856	Can function as microsomal lipase
	Insulin secretion and signaling	Succinyl-CoA synthase β-subunit precursor	2.11	AA253809	Possess histidine kinase activity, involved in insulin secretion
	Signal transduction	Similar to casein kinase II β subunit	1.98	AI323164	Kinase, involved in signaling transduction
		Similar to serine/threonine kinase	2.53	AA286018	Involved in signaling transduction
	Immune related	Similar to IgE-binding factor	−2.84	AA896119	Regulate the synthesis of B-cell-derived immunoglobulin
		Transforming protein RFP	1.99	AI462889	Associated with development
		Large prolin-rich protein BAT3	1.98	A1325425	Involved in immune response of the cell
	Basic cellular function	Synexin	−2.43	AA175243	Ca binding protein, secretory function of the cell
		Similar to B-myb	2.06	AA432951	Cell cycle regulated transcription factor
		UBF transcription factor	−2.46	AI893975	DNA binding protein, control gene transcription
		TGF-β binding protein	5.37	AI314962	TGF complex component, involved in TGF signaling pathway
		TGF receptor 1	2.34	AI119338	Involved in TGF signaling pathway
		Lysyl hydrolase isoform 2	3.50	AI605652	Involved in post-translational modification for collagen
		Similar to ZO-1	2.50	AI326488	Involved in cell proliferation and differentiation
		Heat-shock protein 84	2.40	AI875559	Regulation of cell response to stress
		Similar to P120	2.34	AA000495	Inhibitor of TGF signaling pathway
		Similar to GRB10	2.21	AA260248	Signaling through interaction with protein kinase B
		Similar to cathepsin precursoe	2.18	AI644153	Secretory vesicle, prohormone processing enzyme
		Similar to ubiquitin-conjugating enzyme 2	2.13	AA940461	Protein processing
	Cytoskeleton	Actin-like protein 3	−2.46	AW106357	Induce actin polymerization

For microarray expression analysis, six pancrea/group were pooled and islets isolated and total RNA isolation using the RNeasy kit. Eight set of PancChip 2.2 array experiments were performed (see Section 2 for detail), all changes listed here are with a p<0.05.

Some of the genes that are important for lipid metabolism were not present in the PancChip 2.2 array, therefore in addition to microarray analysis, we selected and analyzed some of the genes that are involved in lipid metabolism using Taqman quantitative real time PCR. Of the 15 genes we examined, seven showed significant changes (Table 2). Two of the enzymes that are involved in fatty acid and TG synthesis were up-regulated (SCD1 increased 10.6-fold, p<0.001; DGAT1, increased 2.7-fold; p<0.001). The expression of other enzymes for TG synthesis, such as glycerol-3-phosphate acyltransferase (GPAT), fatty acid synthase (FAS), and acetyl-CoA carboxylase-1 (ACC-1), however, were not changed. The expression of c/EBP-α tended to increase, but did not reach statistical significance (increased 5.2-fold, p=0.12). Interestingly, SREBP-1c and ATP-citrate lyase (ACLY) were both down-regulated (SREBP-1c, decreased 1.87-fold, p<0.001; ACLY, decreased 1.5-fold, p<0.05). As seen in some other tissues in HSL−/− mice, lipoprotein lipase expression in HSL−/− islets was also elevated (increased 15.7-fold, p<0.001).

Table 2. Quantitative PCR analysis of gene expression in islets of wt and HSL−/− mice

	Gene	Fold (KO/WT)
	c/EBP-α	+5.2±3.2
	SREBP-1C	−1.87±0.05**
	SCD1	+10.6±1.25**
	DGAT1	+2.7±0.2**
	LPL	+15.7±2.8**
	UCP-1	Not detected
	UCP-2	+5.8±1.76*
	ACLY	−1.5±0.045*
	c/EBP-β	NS
	GPR40	NS
	ACC-1	NS
	ACRP-30	NS
	FAS	NS
	GPAT	NS
	PPAR-γ	NS

*p<0.05, **p<0.001 by t-test; NS: not significantly changed.

Recently, UCP-2 has been shown to be an important modulator of insulin secretion and the up-regulation of UCP-2 has been linked with disturbed insulin secretion. The expression of UCP-2 in the islets of HSL−/− mice was up-regulated by more than five-fold (p<0.05).

4. Discussion 

HSL functions as a hydrolase for the metabolism of triacylglycerols, diacylglycerols and cholesterol esters [1]. The rapid modulation of HSL in response to hormones through changes in its phosphorylation, subcellular localization and subsequent change in hydrolytic activity makes HSL an important regulator of cellular lipid homeostasis. An isoform of HSL has been identified in pancreatic β-cells [13]. Although accumulating data have indicated a role for HSL in regulating GSIS [15], [31], a recent report suggested that there is impairment of insulin sensitivity but not disturbed GSIS in HSL−/− mice [16]. In order to further study the involvement of HSL in regulating β-cell function and specifically GSIS, we have studied HSL expression under several conditions, i.e. FFA loading and GLP-1 treatment. We have shown here that FFA loading, which results in higher cellular TG accumulation, induces the expression of the β-cell form of HSL in the clonal β-cell line INS-1. These findings are consistent with the increased HSL expression in islets observed in animals that were fed a short-term high fat diet [32]. This response of HSL to TG accumulation can be viewed as a potential adaptive mechanism, with adaptation working on the hydrolysis of excess TG accumulation to prevent lipotoxicity. Recent studies using adenoviral overexpression of adipose HSL in rat islets have shown that increased expression of HSL in islets can ameliorate palmitate induced lipotoxicity in pancreatic β-cells (N. Fujita, Personal communication). With higher levels or prolonged TG accumulation, the detrimental effects of TG accumulation might not be easily corrected by an increase of HSL, thus resulting in lipotoxicity to β-cells and β-cell apoptosis. Apparently, HSL expression in islets starts to decline in high fat fed animals after 4 months of feeding [32], and changes of HSL were suggested as a mechanism of defense against FFA induced apoptosis. Supporting evidence comes from a recent demonstration that unsaturated fatty acids were shown to protect non-adipose cells against lipotoxicity through promotion of triglyceride accumulation [33]. There are several possible mechanisms by which FFA loading could increase HSL activity, one of them could be through the FFA facilitated interaction of HSL and fatty acid binding protein [34], which result in increased HSL activity; and another possibility could be through the up-regulation of HSL transcription via PPAR-γ and Sp-1 [35].

GLP-1 is known to potentiate insulin secretion by binding to GLP-1 receptors on β-cells, resulting in increased cellular cAMP levels. Recently, GLP-1 was shown to increase lipolysis and FFA release in the pancreatic β-cell line HIT-15 [36]. Our current data show that GLP-1 increases HSL activity in parallel to the potentiation of insulin secretion. Increased HSL activity would result in an increase in FFA release from stored TG in β-cells. Although the mechanism is still not clear, it has been proposed that FFA are converted to long-chain fatty acyl-CoAs and enter mitochondria for oxidation, resulting in an increased ATP:ADP ratio, and an increased insulin secretion [37]. This result is further supported by our studies using islets isolated from HSL−/− mice, where GLP-1 fails to stimulate insulin secretion to the same level as that observed in islets of wild type mice. Taken together, our data support the involvement of HSL in the signaling pathway for GLP-1 potentiated insulin secretion. Potentiation of insulin secretion by GLP-1, however, does not occur solely through HSL, since GLP-1 can increase insulin secretion in the absence of HSL, but stimulation is subnormal compared to wild type islets. During the preparation of this manuscript, Peyot et al., reported that HSL is not essential for GLP-1 action based on their results from 4 month old mice HSL−/− [19]. However, similar to our current observation, GLP-1 potentiated insulin secretion was attenuated in islets from 7 month old HSL−/− mice, indicating the involvement of HSL in this incretin action of GLP-1 [19].

Leptin has been shown to stimulate a novel form of lipolysis in which glycerol is released without a proportional release of FFA [38]. In addition, although leptin can suppress insulin hypersecretion in high glucose fed-rats [39], leptin has no effect on GSIS [40]. We have also studied the effect of leptin, no changes in β-cell HSL expression or activity were seen with leptin treatment in INS-1 cells (data not shown). Therefore, the expression and/or activity of the β-cell form of HSL changes in parallel to changes in GSIS, and HSL does not appears to be altered when GSIS is not affected.

Several different lines of HSL−/− mice have been reported [15], [27], [16], with some phenotypic variations related to insulin secretion observed. Initial observations [15] showed disturbed GSIS in HSL−/− mice, but recent reports from a different group [26] observed only impaired insulin sensitivity with intact insulin secretion, while conflicting results with insulin secretion were reported with mice of 4 month old versus 7 month old [19]. The mice used in the studies of Roduit et al. were generated from BALB/c to C57/BL/6J mice, whereas the mice used in studies of Mulder et al. were SV129/C57/BL/6J hybrid mice. Our HSL−/− mice were 129/Sv-C57BL6 hybrid descendants [17], [27], [28]. Therefore, the differences observed in insulin secretion by these various groups could be due to different genetic backgrounds and/or breeding of the mice, as well as the age of the mice studied. In our studies, we observed an increased basal insulin secretion, that is accompanied by a decreased stimulation of insulin secretion by high glucose, from islets isolated from 6-month old HSL−/− mice. Furthermore, we see a decreased response to GLP-1 stimulation of insulin secretion in the presence of high glucose in islets of HSL−/− mice. Thus, our data are consistent with the initial observations supporting the involvement of HSL in the signaling pathway for GSIS.

HSL−/− mice had significantly larger islets than WT mice, this observation is in agreement with larger islets seen with another line of HSL−/− mice [16], [26]. Furthermore, we have shown that mice on a high fat diet had larger islets than mice of the same genotype on a normal chow diet, which is in agreement with the present view that in non-diabetic obesity islets tend to be larger to compensate for the increased metabolic load and obesity-associated insulin resistance [41]. HSL−/− mice also had larger islets than WT mice under high fat feeding. Regardless of the feeding condition, the increases in mean islet size were due to an increase in the number of cells per islet (i.e. islet hyperplasia). The larger islets seen in HSL−/− mice might also be viewed as a compensatory mechanism for increased basal serum insulin levels seen in HSL−/− mice where peripheral insulin resistance has been observed [15], [16], [17].

Using microarray analysis and PancChip arrays, we have shown that multiple genes are altered in the islets of HSL−/− mice, including genes involved in lipid metabolism, cell cycle, basic cell functions, as well as immune related functions. Follow-up quantitative PCR confirmed the changes in the genes that are related to lipid metabolism. The up-regulation of SCD1 and DGAT1 are associated with TG accumulation. Although SREBP-1c and PPAR-γ have been shown to be important regulators that promote lipogenesis, no change in SREBP-1c mRNA levels has also been seen with lipogenesis regulated by growth hormone and/or insulin [42]. Interestingly, UCP-2 is up-regulated in the islets of HSL−/− mice. It has been shown that UCP-2 gene transcription can be up-regulated by FFA both in vivo under high fat feeding condition and in vitro in isolated islets [43], [44], [45], [46]; therefore, the up-regulation of UCP-2 mRNA is consistent with the increased TG accumulation and up-regulation of SCD1 and LPL in the islets of HSL−/− mice. The observation that UCP-2 gene expression is up-regulated in islets of HSL−/− mice is consistent with several recent reports showing that increased expression of UCP-2 in islets is associated with impaired GSIS [43], [47], [48], [49]. C/EBPα has been shown to be involved in the cytokine-regulated cell-death of β-cells [50]. In view of the 2–2.5-fold increase of TG in the islets of HSL−/− mice [51], the potential up-regulation of c/EBPα could be a response to over-accumulation of TG and a link between lipotoxicity and β-cell dysfunction. A recent study of transgenic mice over-expressing adipose HSL in pancreatic β-cells shows lipotoxicity of the β-cells and also an increase of UCP-2 expression in the islets [52]. The authors propose that increased expression of HSL in the islets of transgenic mice promotes the release of FFA, which is then used as a ligand for PPAR-γ and results in up-regulation of UCP-2, and hence decreased GSIS. On the one hand, it is interesting that both ablated expression of HSL and over-expression of HSL in islets result in up-regulation of UCP-2 and decreased GSIS. This might indicate that disturbed lipid homeostasis in islets causes changes in energy, altering UCP-2 levels. On the other hand, it is noteworthy that the pancreatic β-cell form of HSL is different from the adipose form of HSL and has a N-terminal extension. Therefore, the function and regulation of the β-cell isoform of HSL could be different from the adipose isoform.

In this report, we have shown that HSL activity and expression changes parallel changes in GSIS. When HSL expression is ablated in HSL−/− mice, there are changes in gene expression of a variety of different genes that appear to be associated with TG accumulation and disturbed insulin secretion.