In Vivo Effects of Insulin and Free Fatty Acids on Matrix Metalloproteinases in Rat Aorta

Adult male Sprague-Dawley rats (300–350 g) were purchased from Charles River Laboratories (Wilmington, MA) and were housed in an environmentally controlled room with a 12-h light/dark cycle, where they had free access to standard rat diet (60% carbohydrate, 10% fat, and 30% protein) and water. One week before the studies, the animals were anesthesized (with 5% oxygen and 2% isoflurane). A polyvinyl catheter (internal diameter 0.02 in) was inserted into the right internal jugular vein and extended to the right atrium. Another catheter was advanced through the left carotid artery until its tip reached the aortic arch. The free ends of both catheters were attached to long segments of steel tubing and tunneled subcutaneously to the back of the neck where they were exteriorized and secured to the skin with clips. At the end of the procedure, the catheters were flushed with isotonic saline containing heparin (50 units/ml) and Ampicillin (5 mg/ml) and filled with a viscous solution of heparin (500 units/ml) and 50% dextrose to prevent refluxing of blood into the catheter lumen. All studies were performed in accordance with the guidelines for the use and care of laboratory animals of the Temple University Institutional Animal Care and Use Committee.

The rats were allowed 1 week to recover from the effects of surgery. At that time, they were within 3% of their preoperative weight. Euglycemic-hyperinsulinemic clamps were conducted in the morning after a 14-h overnight fast. Throughout the studies, the animals were allowed to move freely in their cages. All substrates were administered into the venous catheter, and blood samples were obtained from the arterial catheters. After the clamps, the rats were killed by an overdose of isoflurane and the aorta freeze clamped, excised, carefully cleaned of adventitia, and frozen at −80°C until assayed.

Euglycemic-hyperinsulinemic clamps with or without lipid/heparin infusions were performed in awake and unrestrained rats as described (15) with some modifications. Insulin (4.8 mU · kg⁻¹ · min⁻¹) was infused through the jugular vein catheter from 0 to 240 min. Glucose concentrations were clamped at euglycemic levels by a variable-rate infusion of 25% glucose. Blood glucose levels were monitored with an Elite Glucometer (Bayer, Elkhart, IN), and glucose infusion rates (GIRs) were adjusted every 5–10 min as needed.

Liposyn II, a 20% triglyceride emulsion (Abbott Labs, Chicago, IL), was infused at 0.618 ml/h with heparin (20 units/h). In the euglycemic-hyperinsulinemic clamps without lipid/heparin infusions, glycerol (143 μmol/h) instead of lipid/heparin was infused to match the glycerol content of Liposyn II.

In control experiments, glycerol/saline or lipid/heparin was infused without insulin and glucose was maintained at ∼5.5 mmol/l. Glycerol was infused to simulate the glycerol content in Liposyn II. During these studies, blood samples (∼200 μl) were obtained from the carotid artery at −30, 0, 60, 120, 180, 210, and 240 min. Blood was centrifuged immediately, and the red cells were reinfused into the animals.

Insulin was measured in plasma by radioimmunoassay using rat insulin as standard (Millipore, St. Charles, MO). FFAs were measured in plasma to which a lipoprotein lipase inhibitor (Paroxon; Sigma-Aldrich, St. Louis, MO) had been added with a kit from Wako Pure Chemicals (Richmond, VA).

Rat TIMP-1 was determined with an ELISA kit from R&D Systems (Minneapolis, MN) and rat TIMP-2 with an ELISA kit from Calbiochem (San Diego, CA) following instructions provided by the respective manufacturers. Absorbance at 450 nm was measured with a microplate reader (Labsystems, Franklin, MA).

A rabbit anti–IRS-1 serum (Upstate, Lake Placid, NY) and Protein A-Agarose beads were used to immune-precipitate IRS-1–associated PI 3-kinase (PI3K) from aortic extracts (100 μg).

Aortic tissues were extracted and protein content was measured using the Bio-Rad protein assay (BioRad, Richmond, CA). Samples were separated on 4–20% SDS gels using a Tris-glycine running buffer (0.2 mol/l Tris-base, 0.2 mol/l glycine, pH 6.8, and 0.1% SDS). The separated samples were then transferred to a nitrocellulose membrane in Tris-glycine transfer buffer supplemented with 20% methanol. The nitrocellulose membranes were blocked in 5% dried milk–Tris-buffered saline (TBS) containing 0.1% Tween 20 for 1 h, then incubated with primary antibodies diluted in 5% dried milk–TBS containing 0.1% Tween 20. The primary antibodies were: mouse antibodies (Calbiochem) against the 72-kDa latent and the 66-kDa active forms of MMP-2, the 92-kDa latent and the 68-kDa active forms of MMP-9, and the ∼60-kDa MTI-MMP; a rabbit anti-serum (Upstate) that recognizes the N-SH₂ region of PI3K and the regulatory p85 subunit of PI3K and a rabbit anti-serum that recognizes rat IRS-1; a rabbit antibody (Cell Signaling, Danvers, MA) that recognizes the active, dually phosphorylated (at threonine 202 and tyrosine 204) forms of ERK1/2 (44 and 42 kDa, respectively) and a rabbit anti-serum that detects ERK1/2; a rabbit antibody (Cell Signaling) that detects the human and rat phosphorylated (at threonine 180 and tyrosine 182) forms of p38-α, -β, and -γ mitogen-activated protein kinase (MAPK) (43 kDA); and a rabbit anti-serum that recognizes total p38 MAPK, a rabbit antibody (Cell Signaling) that detects endogenous levels of p46 (c-jun NH₂-terminal kinase [JNK]-1) and p54 (JNK-2 and -3) dually phosphorylated at threonine 183 and tyrosine 185, and a rabbit anti-serum that detects total JNK protein. Membranes were washed in TBS containing 0.1% Tween 20 and incubated with secondary antibodies for 1 h. Bands were visualized using an enhanced chemiluminescense detection kit from Amersham Life Sciences (Arlington Height, IL).

MMP-2 activities were measured by gelatin zymography. Aortic tissue extracts were loaded onto SDS-PAGE gels containing 1 mg/ml of gelatin under nonreducing conditions and were run at 100 V for 45 min with molecular weight standards (BioRad). Gels were then washed twice in 2.5% Triton X-100 and incubated overnight in zymogram development buffer (Bio-Rad, Hercules, CA). Gels were then stained with Coomassie blue R-250 followed by destaining with 55% methanol and 7% acetic acid.

Plasma glucose concentrations were clamped at ∼5.4 ± 0.1 mmol/l in all four studies (Fig. 1).

Insulin rose from 80 ± 20 before to means of 1,564 ± 30 and 1,659 ± 56 pmol/l during hyperinsulinemic clamps with and without lipid/heparin, respectively, and did not change during saline infusions (168 ± 92 vs. 237 ± 35 pmol/l, NS) or during lipid/heparin (69 ± 15 vs. 121 ± 43 pmol/l, NS).

Plasma FFA remained unchanged during saline infusions (623 ± 48 pre-saline vs. 524 ± 13 μmol/l post-saline), decreased from 736 ± 72 to 175 ± 20 μmol/l (P < 0.01) in response to insulin, and rose from 692 ± 56 to 2,138 ± 187 μmol/l (P < 0.001) in response to insulin plus lipid/heparin and from 636 ± 34 to 2,138 ± 284 μmol/l (P < 0.001) during lipid/heparin.

GIR (the rate of glucose infusion needed to maintain euglycemia during insulin infusions and a measure of insulin-stimulated glucose uptake) rose from 0 to 203 ± 9 and 208 ± 11 μmol · kg⁻¹ · min⁻¹, respectively, after 1 h in the two hyperinsulinemic clamps. After that, GIR remained elevated in the hyperinsulinemic clamps without lipid/heparin and decreased ∼30% (to 144 ± 10 μmol/kg min) at the end of the hyperinsulinemic clamps with lipid/heparin infusions, indicating the development of FFA-mediated insulin resistance and confirming previous results (14,15). GIR did not change during saline or lipid/heparin infusions.

MMPs are synthesized as inactive pro-zymogens and activated by proteolytic conversion (7,16). The bioactive (66 kDa) MMP-2 (presented as MMP-2–to–β-actin ratio) did not change during the 4-h saline infusions (0.5 ± 0.1 prestudy vs. 0.4 ± 0.1 poststudy) or during lipid/heparin infusion (0.4 ± 0.02) but increased ∼6-fold to 3.1 ± 0.3 after 4 h of hyperinsulinemia (P < 0.02) and increased further (∼11-fold) to 5.4 ± 0.8 (P < 0.002) with infusion of insulin and lipid/heparin (Fig. 2).

The bioactive (68 kDa) MMP-9 (MMP-9–to–β-actin ratio) also did not change during saline infusions (0.36 ± 0.04 vs. 0.4 ± 0.04) or during lipid/heparin infusion (0.4 ± 0.02) but increased ∼13-fold during hyperinsulinemia (to 4.7 ± 0.6, P < 0.001) and increased further (∼23-fold) during insulin and lipid/heparin infusion (to 8.2 ± 1.6, P < 0.05).

MT1-MMP, which activates MMP-2 from the pro- to the active form (7,16), similarly did not change during saline (MT1-MMP–to–β actin ratio pre- vs. postsaline, 0.13 ± 0.1 vs. 0.13 ± 0.05) or during lipid/heparin infusion (0.14 vs. 0.02) but increased ∼8-fold in response to insulin (to 1.08 ± 0.05, P < 0.001) and increased further (∼20-fold) in response to insulin plus lipid/heparin infusion (to 2.59 ± 0.18, P < 0.001).

We also examined the effects of insulin and insulin plus lipid/heparin on MMP-2 abundance in epidydimal fat. The bioactive (66 kDa) MMP-2–to–β-actin ratio did not change in response to saline (1.24 ± 0.1 vs. 1.36 ± 0.24, NS) or insulin plus lipid/heparin (0.13 ± 0.22, NS) but increased in response to insulin (1.95 ± 0.18, P < 0.01) (Fig. 3 lower panel).

We next examined protein abundance of TIMP-1 and TIMP-2, the physiological inhibitors of MMPs. TIMP-1 did not change in response to saline (pre- vs. postsaline, 130 ± 5 vs. 144 ± 6 pg/ml, NS), lipid/heparin (174 ± 64, NS), insulin (200 ± 66 pg/ml, NS), or insulin plus lipid/heparin (180 ± 52 pg/ml, NS). TIMP-2 similarly did not change in response to saline (pre- vs. postsaline, 32 ± 14 vs. 33 ± 24 pg/ml, NS), lipid/heparin (32 ± 23, NS), insulin (33 ± 12 pg/ml, NS), or insulin plus lipid/heparin (32 ± 16 pg/ml, NS). Thus, there were no significant changes in either TIMP1 or TIMP2 in response to saline, insulin, or insulin plus lipid infusions.

Effects of hyperinsulinemia (with and without lipids) on the gelatinolytic activity of MMP-2 was determined with gelatin zymography (Fig. 3). MMP-2 gelatinolytic activity did not change in response to saline (18.9 ± 3.4 vs. 19.9 ± 7.7 arbitrary units [AU]) but increased twofold in response to hyperinsulinemia (to 35.7 ± 5.4 AU, P < 0.025) and increased further (threefold) in response to insulin plus lipid/heparin (to 57.8 ± 7.9 AU, P < 0.001). Very similar changes were seen with pro–MMP-2.

To assess signaling pathways possibly involved in the observed FFA-induced enhancement of insulin-stimulated MMP-2, MMP-9, and MT1-MMP, we examined effects of insulin and insulin plus lipid/heparin on key components of the IRS/PI3K/Akt cascade and three major MAP kinase pathways. As seen in Fig. 4, both the IRS-1 association with the p85 regulatory unit of PI3K and ERK1/2 phosphorylation (indicating activation of the enzymes) increased under the influence of hyperinsulinemia but decreased in response to insulin plus lipid/heparin. This indicated that neither of these pathways was involved in the FFA-mediated increase in MMP-2, MMP-9, and MT1-MMP.

Insulin activated JNK, i.e., increased phospho-JNK, while FFA further enhanced insulin activation of JNK. p38 MAPK phosphorylation was not affected by insulin but was potently increased by FFA. Thus, either one or both of these pathways could have been involved in the FFA-mediated stimulation of MMPs (Fig. 5).

The main findings in this study were that 1) insulin stimulated the bioactive forms of MMP-2 in aortic tissue ∼6-fold, of MMP-9 ∼13-fold, and of MT1-MMP ∼8-fold and 2) FFAs further augmented these increases (to ∼11-fold, ∼23-fold, and ∼20-fold, respectively), while at the same time inhibiting insulin stimulation of total body glucose uptake, i.e., causing insulin resistance. Neither insulin nor FFA affected the two major MMP inhibitors, TIMP-1 and TIMP-2. The increase in the MMP-to-TIMP ratio indicated increased activity of these proteinases, and this was directly demonstrated for MMP-2 by gelatin zymography. These results indicate that physiologically high plasma insulin and FFA levels, two abnormalities characteristically seen in obese insulin-resistant patients, strongly activated several MMPs that have been strongly implicated in the development and progression of ASVD (9–13).

Previous reports on the effects of insulin on MMPs have produced discrepant results. On one hand, in vitro studies have shown that hyperinsulinemia increased MMP-9 and/or MMP-2 activity in human monocytes and cultured rat glomerular mesangial cells (17,18). On the other hand, Dandona et al. (19) have shown, in vivo, that 4 h of euglycemic hyperinsulinemia (from 78 to 144 pmol/l) lowered plasma MMP-9 by 18% in 10 obese nondiabetic subjects. The reason for these differences are not clear but could possibly be due to tissue versus plasma MMP level differences, species differences, and interorgan differences. In support of interorgan differences of the FFA effects on MMPs, we were able to show that FFA augmented insulin stimulation of MMP-2 activity in the aorta while decreasing it in epidydimal adipose tissue (Fig. 2).

We are not aware of previous reports on effects of FFA on MMPs except for a microarray screen of muscle biopsies obtained after 48 h of lipid infusion in normal volunteers, which showed increased MMP-2, -11, and -28 and TIMP-1 expression (20).

We have used heparin to promote lipolysis of the infused lipids and to prevent hypertriglyceridemia (21). Heparin has been reported to decrease phorbolester-stimulated MMP-9 expression while having no effect on MMP-2 or TIMP-1 (22,23). Thus, it is possible that we may have underestimated the effect of FFA on MMP-9.

To investigate pathways through which insulin and FFAs may have affected MMPs, we examined key terminal targets of several insulin signaling cascades. Previous reports have implicated insulin or IGF-1 signaling through the PI3K/Akt cascade as being responsible for stimulation of MMP activity (18,24). In vascular tissues, this pathway controls the metabolic actions of insulin (including glucose and amino acid uptake, glycogen synthesis, and nitric oxide production) (25) and is inhibited by FFA in major insulin target tissues (14). Our results confirmed in rat aorta that insulin stimulated PI3K and that FFA suppressed the insulin-stimulated PI3K activity. This makes it highly unlikely that the FFA-enhanced MMP activity observed in this study was mediated through this pathway.

Other reports have implicated MAPKs in the insulin stimulation of MMPs (24,26). MAPKs transmit effects of extracellular stimuli (growth factors including insulin, stress, and bacteria) from the cell membrane to the nucleus and mainly control cell growth and proliferation (27). Examining effects of insulin and FFA on ERK1/2, the terminal target of the canonical Ras/Raf/MEK/ERK1/2 cascade, we found that similar to its effect on PI3K, FFAs also inhibited insulin stimulation of ERK1/2.

There are several possible explanations for the simultaneous inhibition of the PI3K and ERK1/2 pathways by FFA. First, it has recently been shown that the PI3K/Akt and the Ras/Raf/MEK/ERK pathways can be interdependent by exerting regulatory control on each other through transphosphorylation (28). Second, ERK1/2 can bind to activated p38 MAPK, which prevents their activation by upstream MEK kinases and results in decreased ERK activation (29). Whatever the mechanism, the suppression by FFA of insulin stimulation of both PI3K and ERK1/2 makes it unlikely that either of these two pathways was involved in the FFA stimulation of MMPs.

We next examined JNK and p38 MAPK, the terminal targets of two other MAPK cascades, which can be activated by environmental and oxidative stress factors (27). We found that acute elevations of plasma FFA levels significantly increased insulin-stimulated JNK activity. JNK activity has been shown to be elevated in many animal models of obesity, by insulin, FFA, endoplasmic reticulum stress, inflammatory cytokines, and bacterial LPS, and has been implicated as a mediator of obesity-associated insulin resistance (30–36). Once activated, JNK upregulates expression of inflammatory genes by activating the AP-1 transcription factor complex.

We also found that p38 MAPK was strongly activated by FFA. In vitro, FFAs have been shown to activate p38 MAPK in hepatocytes, cardiac myocytes, and endothelial cells (37–40). p38 MAPK plays a central role in many inflammatory disorders, and p38 MAPK antagonists are currently being tested in clinical trials for treatment of rheumatoid arthritis, Alzheimer's disease, and inflammatory bowel disease (41).

Our findings support the concept of selective insulin resistance. This concept is important to explain why hyperinsulinemia, the usual consequence of metabolic insulin resistance, is associated with decreased insulin action on glucose uptake, glycogen synthesis, and lipolysis in skeletal muscle, liver, and adipose tissue, while remaining uninhibited in vascular tissue with respect to its action on vascular smooth muscle growth and endothelial cell matrix production. We have expanded the concept of selective insulin resistance by showing that it is not always characterized by suppressed of PI3K and uninhibited ERK1/2 activity, as was originally proposed for obesity-associated insulin resistance (42,43), but that, at least in the case of FFA-induced insulin resistance, some MAPKs are inhibited together with PI3K, while other MAP kinases are stimulated. Thus, our data suggest that selective insulin resistance is more variable than was believed earlier and that exactly which MAPKs are inhibited and which are stimulated may depend on the cause for the insulin resistance and the specific tissues involved.

The upstream events leading to activation of these kinases are not entirely clear, although there are several possibilities. First, an increase in plasma FFA has been shown to result in intramyocellular accumulation of diacylglycerol and activation of several protein kinase C (PKC) isoforms including PKC-β2 in human muscle (44), PKC-δ in rat liver (15), and PKC-θ in rat skeletal muscle (45) and in activation of inhibitor of IκB-α (IKK) and activation of nuclear factor κB (NFκB) in rat liver (15). Gao et al. (33) have recently shown that the FFA-mediated activation of JNK and IKK is PKC-dependent, suggesting PKC activation as an upstream event. Second, some recent evidence suggests that FFA-mediated activation of JNK, and perhaps other serine/threonine kinases, may be, at least partly, mediated by the Toll like receptor-4 (TLR-4) (46). TLR-4 is essential for the development of innate immunity to pathogens and triggers production of inflammatory cytokines (36). Third, several G protein–coupled receptors including GRP-40 and GRP-120 have been shown to bind FFA (47,48). There is as yet, however, no evidence that these receptors are involved in any of the FFA activities mentioned above.

While the precise mechanisms by which FFA activate these serine/threonine kinases remains to elucidated, some of the consequences have become clear. First, activation of PKC, JNK, and IKK can cause insulin resistance (15,30,44). This occurs mainly through serine phosphorylation of IRS1/2, which inhibits insulin signaling (49). Second, activation of IKK, JNK, or p38 MAPK can result in activation NFκB and release of proinflammatory cytokines and chemokines including tumor necrosis factor (TNF)-α, interleukin (IL)-1β, monocyte chemoattractant (MCP)-1 (15,50). These cytokines can also be induced in macrophages by insulin (51). Cytokines, in turn, promote the activation of MMPs (4,15). In fact, using the same experimental protocol as in the current study, we have shown in a previous study that acutely elevating plasma FFA levels results in a strong and continuous increase in circulating MCP-1 levels in rats (15). Hence, the effects of FFA and insulin on MMPs are likely to be indirect and mediated through cytokines.

In this study, we have shown that hyperinsulinemia increased the activities of three MMPs several-fold and that the combination of hyperinsulinemia and high plasma FFA levels further augmented these increases. All three of these MMPs have been implicated in playing important roles in the development and progression of ASVD such as heart attacks, strokes, and peripheral arterial diseases. We have also shown that hyperinsulinemia activated JNK and that the combination of hyperinsulinemia and elevated FFA levels further augmented the activation of JNK and, in addition, strongly activated p38 MAPK. Activation of these pathways are known to result in the synthesis and release of proinflammatory and proatherogenic cytokines. These findings are pertinent because the combination of high plasma insulin and FFA levels is common in obese insulin-resistant patients regardless of whether they have type 2 diabetes. This group of individuals also has a greatly increased risk for ASVD compared with nonobese and non–insulin-resistant subjects. Assuming that our results obtained in rat aortic tissue are applicable to human vascular tissue, we believe that the increased MMP activity and the activation of several MAP kinases shown here may over time increase the degradation of extracellular matrix and produce inflammatory changes, which together may lead to progression of atherosclerotic lesions and contribute to the increased risk for ASVD in obesity.

This work was supported by National Institutes of Health Grants R01-DK58895, HL-0733267, and R01-DK066003 and a Mentor-Based Training Grant from the American Diabetes Association (to G.B.).

We thank Constance Harris Crews for typing the manuscript and Maria Mozzoli, BS, for technical assistance.

The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.