Mesenteric Resistance Arteries in Type 2 Diabetic db/db Mice Undergo Outward Remodeling

Objective Resistance vessel remodeling is controlled by myriad of hemodynamic and neurohormonal factors. This study characterized structural and molecular remodeling in mesenteric resistance arteries (MRAs) in diabetic (db/db) and control (Db/db) mice. Methods Structural properties were assessed in isolated MRAs from 12 and 16 wk-old db/db and Db/db mice by pressure myography. Matrix regulatory proteins were measured by Western blot analysis. Mean arterial pressure and superior mesenteric blood flow were measured in 12 wk-old mice by telemetry and a Doppler flow nanoprobe, respectively. Results Blood pressure was similar between groups. Lumen diameter and medial cross-sectional area were significantly increased in 16 wk-old db/db MRA compared to control, indicating outward hypertrophic remodeling. Moreover, wall stress and cross-sectional compliance were significantly larger in diabetic arteries. These remodeling indices were associated with increased expression of matrix regulatory proteins matrix metalloproteinase (MMP)-9, MMP-12, tissue inhibitors of matrix metalloproteinase (TIMP)-1, TIMP-2, and plasminogen activator inhibitor-1 (PAI-1) in db/db arteries. Finally, superior mesenteric artery blood flow was increased by 46% in 12 wk-old db/db mice, a finding that preceded mesenteric resistance artery remodeling. Conclusions These data suggest that flow-induced hemodynamic changes may supersede the local neurohormonal and metabolic milieu to culminate in hypertrophic outward remodeling of type 2 DM mesenteric resistance arteries.


Introduction
Diabetes mellitus (DM) is the fifth leading cause of death in the United States, and type 2 DM accounts for 90-95% of all diabetic cases for which vascular complications are a leading cause of morbidity, mortality and economic burden [1]. Chronic hyperglycemia, coupled with insulin resistance, is a hallmark of type 2 DM and underlies many of the observed pathophysiological complications in the cardiovascular system [2,3]. Moreover, type 2 DM is associated with a cluster of risk factors including obesity, hypercholesterolemia, hyperlipidemia and hypertension. Together, these factors lead to increased accumulation of advanced glycation end products, alterations in the renin-angiotensin system (RAS), oxidative stress, and reduced nitric oxide bioavailability [4,5,6] culminating in endothelial dysfunction, altered vasoreactivity, and vessel wall remodeling [7,8].
Although the regulation of resistance artery function has been extensively studied within the context of type 2 DM, much less is known about the mechanisms that dictate diabetes-associated structural remodeling of the microvasculature. Vascular remodeling involves the reorganization of existing cells and extracellular matrix (ECM) or changes in vascular smooth muscle cell (VSMC) growth and migration [9], and can be characterized as hypertrophic (increased cross-sectional area), eutrophic (no change in cross-sectional area) or hypotrophic remodeling (reduced cross-sectional area). These processes have been associated with either inward remodeling (smaller lumen diameter) or outward remodeling (increased lumen diameter) [10]. In small resistance arteries from type 2 diabetic patients, hypertrophic remodeling and reduced fibrosis has been reported [11]. In contrast, hypertensive patients with pre-existing type 2 DM had predominantly eutrophic remodeling and increased fibrosis [12].
Dynamic regulation of the ECM is an essential component of the remodeling process. The ECM not only dictates in part the integrity of the vessel wall, but is also an important determinant of vessel compliance; however, the molecular mechanisms that dictate ECM accumulation and turnover in diabetic microvessels have not been extensively studied. A balance between matrix metalloproteinases (MMPs) and tissue inhibitors of matrix metalloproteinases (TIMPs) is thought to play a central role in vessel remodeling, and an imbalance between MMPs and TIMPs may lead to excess ECM accumulation or degradation [13,14]. For example, diabetic nephropathy and cardiomyopathy involves excessive ECM deposition due to a decrease in the MMP/TIMP ratio [15,16]. The pro-fibrotic plasminogen activator inhibitor-1 (PAI-1) is increased in DM, insulin resistance, and hypertension [17].
In addition to neurohormonal regulation of ECM and VSMC growth, hemodynamics also dictate the pattern of microvascular remodeling. For example, outward remodeling occurs in response to chronic increases in blood flow in rat mesenteric resistance arteries [18]. Thus, the pattern of vascular remodeling in diabetes may be tissue-specific, reflecting a dynamic interaction between hyperglycemia, hemodynamic stimuli and locally generated angiotensin II, as well as oxidative stress and inflammation [19]. The aim of the present study was to examine mesenteric resistance arteriole (MRA) structure in leptin-receptor deficient db/db mice, a model of type 2 DM, insulin-resistance, hyperlipidemia and obesity.

Ethics Statement
This study conformed with the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication No. 85-23, revised 1996) and was approved by the LSU Health Sciences Center Institutional Animal Care and Use Committee (IACUC protocol number 2498, Institutional Animal Assurance number A3094-01) and the Nationwide Children's Hospital Institutional Animal Care and Use Committee (IACUC protocol number AR08-00023, Institutional Animal Assurance Number A3544-01).

Animals
All experiments used 12 and 16 wk-old male type 2 DM db/db (C57BL/KsJ-db-/db-) and control heterozygous Db/db mice (C57BL/KsJ-db+/db2, The Jackson Laboratory, Bar Harbor, ME, USA). By 4 wks of age db/db mice are obese, develop hyperglycemia by 8 wks of age, overt DM by 12 wks of age, and exhibit many common features of type 2 DM, including hyperlipidemia, obesity and insulin resistance. Heterozygous mice were used as control, which cannot be distinguished morphologically or physiologically from wild type mice (data not shown). Mice were housed under a 12-hour light/dark cycle at 22uC and 60% humidity. They were allowed ad libitum access to water and standard laboratory mouse chow.

Blood Pressure and Blood Glucose Measurements
Radiotelemetry transmitter catheters were implanted into the right carotid arteries of 16 wk-old Db/db and db/db mice using aseptic surgical techniques in animals anesthetized with 2% isoflurane. A subcutaneous space was bluntly dissected in the right flank region to house the radiotelemetry transmitter (Model PhysioTel PA-C10; Data Sciences International, St. Paul, MN, USA). Baseline mean arterial blood pressure was recorded for 4 weeks in conscious, freely moving mice using Dataquest A.R.T (Data Sciences, St. Paul, MN, USA) acquisition software after the return of diurnal cycles (10 days post surgery). Following the acquisition of baseline blood pressure and heart rate, data were collected for one hour at the same time each day and averaged into 2-second bins over 7 days. Prior to the terminal vessel experiment and 12 hours after fasting, blood glucose was measured with an Accu-Chek II portable glucose meter (Roche, Indianapolis, IN, USA).

Mesenteric Isolation
Db/db and db/db mice were anesthetized using ketamine (200 mg kg 21 , ip) and xylazine (10 mg kg 21 , ip). Mesenteric resistance arteries at the fourth to fifth branch of the superior mesenteric artery were dissected and isolated from surrounding adipose and connective tissue using a dissecting microscope and mounted onto two glass micropipettes in a video-monitored perfusion system as previously described [20]. The intravascular pressure was controlled by a pressure servo control (Living Systems Instrumentation, Burlington, VT, USA). Additional MRAs were isolated from separate sets of Db/db and db/db mice and either formalin fixed (10% phosphate buffered formalin) or frozen in liquid N 2 for histological or biochemical experiments, respectively.

Mesenteric Resistant Artery Measurements
The following structural and mechanical parameters were determined as previously described [20,21]: Media cross-sectional area (CSA) = p (D e 2 2D i 2 ) /4 Cross-sectional compliance (Ccs, mm 2 mmHg 21 ) = DAi/DP, in which DAi is the change in internal lumen cross-sectional area induced by a pressure change (DP).
Cross-sectional distensibility (Dcs, kPa 21 ) as Dcs = Ccs/Ai Circumferential wall strain (e) = (D-D 0 ) /D 0 , where D was the observed lumen diameter for a given intraluminal pressure and D 0 was the original diameter measured at 10 mmHg intraluminal pressure.
Circumferential wall stress (s) = (PD)/(2WT), where P is the intraluminal pressure and D is the lumen diameter. Pressure is converted to dynes per cm 2 (1 mmHg = 1.3343610 5 dynes/cm 2 ) when stress is examined as a function of strain.
Incremental Elastic Modulus (E inc ) =Ds/De, which is used when the stress-strain curve is curvilinear rather than linear.

Vascular Morphology
Formalin fixed, paraffin-embedded, fourth to fifth order isolated mesenteric resistance arteries from 16 wk-old mice were sectioned (6 mm), stained with picrosirius red (PSR) for collagen and quantified using picrosirius polarization methods [22]. The images were digitized using Olympus MicroSuite 5. Two sections per animal (n = 8) were used in each group. Data are expressed as collagen area percent and was normalized to medial cross sectional area.

Blood Flow Measurements
Direct hemodynamic measures of superior mesenteric blood flow were performed in 12 wk-old male Db/db (n = 5) and db/db (n = 6) mice. Mice were maintained on isoflurane anesthesia (2.25%) at a body temperature of 37uC. The superior mesenteric artery was carefully isolated and a Doppler flow nanoprobe (Model 0.7PSB, Transonic Systems, Inc., Ithaca, NY) connected to a TS420 perivascular flowmeter (Transonic Systems, Inc., Ithaca, NY, USA) was gently placed around the artery. A 3-minute average of mesenteric blood flow was recorded and analyzed after a 25-minute equilibration period using LabScribe 2 data acquisition and analysis software (iWORX Systems, Inc., Dover, NH, USA).

Data Analysis
Unpaired Student's t-test or two-way repeated measures ANOVA was used when appropriate. Data are presented as mean 6 SEM, with p,0.05 representing significance.

Blood Pressure and Blood Glucose
At 16 wks of age, there was no difference in systolic blood pressure between groups (Table 1). Diabetic db/db mice displayed significantly higher blood glucose concentrations and body weight compared to Db/db mice (Table 1).

Mesenteric Resistance Artery Structure in 16 wk-old Mice
MRAs from 16 wk-old db/db mice had larger internal lumen diameter and external diameter measured at each pressure compared to MRAs isolated from control mice (Fig. 1a,b). Media cross-sectional area was also significantly increased in db/db mice compared to control mice (Fig. 1c); however, wall thickness was not different between groups (data not shown). As a result, the wall-to-lumen ratio was significantly decreased in db/db compared to control mice (Fig. 1d).
MRAs from db/db mice were more compliant at low pressures than control mice (Fig. 2a). Neither the diabetic nor control   mesenteric arteries displayed significant alterations in crosssectional distensibility, except for a minor increase at 25 mmHg of pressure (Fig. 2b). Circumferential wall stress was significantly greater in db/db mesenteric arteries compared with control arteries at 75, 100, and 125 mmHg (data not shown). Finally, there was a rightward shift in the stress-strain relationship in db/db MRAs compared to control (Fig. 2c). Incremental elastic modulus, calculated from the incremental slopes of the stress-strain curve, was generally not different between normal and diabetic MRAs, although the diabetic MRAs were stiffer at the highest pressure (125 mmHg, Fig. 2d).

Expression of Matrix Regulatory Proteins in Mesenteric Resistance Arteries in 16 wk-old Mice
Western blot analysis was used to determine the relationship between pro-and anti-fibrotic regulatory protein expression and ECM accumulation. MMP-9 and MMP-12 protein expression were significantly increased in MRAs isolated from diabetic db/db mice compared with the controls (Fig. 3a,b,c).
The expression of PAI-1 (Fig. 4a,b), and TIMP-1 and TIMP-2 (Fig. 4c,d,e), were significantly increased in db/db MRAs compared with control vessels. PSR staining indicated that these dynamic changes in pro-and anti-fibrotic protein expression did not result in a net change in collagen accumulation in db/db MRAs within the vascular wall (Fig. 5a,b).

MRA remodeling and Superior Mesenteric Blood Flow in 12 wk-old Mice
We next determined whether changes in superior mesenteric blood flow preceded the observed hypertrophic outward remodeling in 16 wk-old db/db mice. At 12 wks of age, there was no observed MRA structural remodeling in db/db mice as evidenced by no differences in lumen diameter, external diameter, medial cross-sectional area, and wall:lumen ratio (Fig. 6a-d). MRA mechanics between normal and diabetic mice were also unchanged (data not shown). Conversely, superior mesenteric blood flow in db/db mice was significantly increased by 46% when compared with control mice (Fig. 7).

Discussion
This study shows for the first time that MRAs in 16 wk-old type 2 DM db/db mice undergo hypertrophic outward remodeling associated with increased compliance at lower pressures and augmented stiffness at higher pressures. This remodeling process was accompanied by dynamic changes in the expression of MMP-9 and MMP-12, TIMP-1 and TIMP-2, and PAI-1 but no net change in medial collagen. These structural and biochemical findings were preceded by a 46% increase in superior mesenteric artery blood flow in 12 wk-old diabetic mice -a time point at which no MRA remodeling was present -suggesting that increased hemodynamic forces may precede and chronically influence downstream resistance artery remodeling.
Structural alterations of small arterial walls are the most potent predictors of adverse cardiovascular events in high-risk populations, including type 2 diabetic patients [23]. Arterial wall hypertrophy is present in resistance arteries (from gluteal biopsies) from patients with type 2 DM, with little to no change in intraluminal diameter [24]. Other studies in MRAs from streptozotocin-treated rats demonstrated either vascular hypertrophy [25] or eutrophic (i.e. no change in medial cross sectional area) outward remodeling as little as 3 weeks after experimentally-induced DM [26]. In contrast, our data reveal that MRAs in type 2 DM db/db mice had increased luminal and external diameters and medial cross-sectional areas, indicating hypertrophic outward remodeling. These structural changes were blood pressure-independent since we saw no significant differences in systolic blood pressure between groups ( Table 1).
The reasons for these disparate remodeling processes are unclear but may reflect differences in vascular beds, insulin sensitivity, obesity and/or chronic exposure to hyperglycemia. Previous studies have shown that chronic changes in blood flow are one of the major stimuli dictating arterial diameter and reorganization of the vascular wall [27]. One possible explanation for our observations is that flow-induced hemodynamic changes may supersede the local neurohormonal and metabolic milieu that favor inward remodeling. Chronic hyperphagia in this model causes increased blood flow to accommodate increased nutrient uptake [28]. The thickness, composition and architecture of the artery wall are determined in part by stresses imposed by pressure and flow [29]. Both chronic increases in blood flow and wall stress trigger a combined response of hypertrophic and outward remodeling [18]. As shown by us and others, blood flow to the small intestine is also increased in the diabetic animals secondary to prolonged hyperphagia and may act as a stimulus for outward remodeling [30]. This is consistent with data showing outward hypertrophic remodeling of resistance arteries is the result of a chronic increase in blood flow [18,31,32,33]. Our data support the notion that increased blood flow resulting from prolonged feeding in obese db/db mice outweigh the influences of hyperglycemia, inflammation, and oxidative stress that dictate vessel wall remodeling in DM. Given that each vascular bed is exposed to a unique pattern of hemodynamic, neurohormonal and inflammatory stimuli, the effects of type 2 DM on vascular remodeling cannot be globally applied to all resistance beds.
Vessel wall remodeling also leads to changes in the structural mechanical properties. Arterial compliance, defined as the change in lumen area for a given change in pressure, depends on the size of the vascular lumen as a hollow structure and cushions changes in arterial pulse pressure. Distensibility represents compliance corrected for vascular lumen size and depends on the medial mass and intrinsic elastic properties of the vascular wall [26,34]. E inc reflects the mechanical properties of the vascular wall material (collagen, vascular smooth muscle, elastin and other ECM components) independent of geometry. Our results show increased MRA cross-sectional compliance in db/db mice, resulting in a less stiff vessel at lower physiological pressures. Conversely, there were no differences in cross-sectional distensibility between normal and diabetic mice, except for an increase at 25 mmHg. These data are consistent with previous reports that increased arterial compliance fails to correlate with an increase in arterial distensibility [26,34], which are likely due to the combination of increased lumen diameter and wall stress without alterations in wall thickness. The stress-strain relationship is an additional geometry-dependent measure of vascular compliance. In general, leftward deviations in this parameter functionally describe a stiffer, less compliant vessel [35]. In contrast, our results show a rightward deviation in the stress-strain curve in diabetic MRAs, indicating a less stiff, more compliant vessel compared to non-diabetic controls in the face of an unchanged incremental modulus, except for a transient increase at higher physiological pressure. These data are the first to report increased MRA cross-sectional compliance in type 2 DM mice, since most studies have reported either a leftward shift or no differences in the stress-strain relationship [36,37]. The current study also demonstrates that as intraluminal pressure increases, wall stress increases in diabetic mice compared to controls, which is a function of and directly proportional to increased luminal diameter. Previous studies have shown that changes in wall stress result in either impaired myogenic response (stimulus for vascular hypertrophy) or in altered vessel geometry to reduce wall stress [14,18,24,38,39].
The molecular mechanisms that govern ECM remodeling in diabetic resistance arteries have not been extensively studied. Here we show that type 2 DM MRA structural remodeling was accompanied by no net matrix accumulation. This was not surprising, since matrix accumulation is usually associated with vessel stiffness and inward remodeling [40]. However, the concomitant increase in both anti-fibrotic (MMP-9 and MMP-12) and pro-fibrotic (TIMP-1, TIMP-2 and PAI-1) molecules is indicative of dynamic ECM turnover. We speculate that the ECM is in a constant state of flux as the vessel undergoes outward remodeling in an attempt to maintain wall integrity. The mechanisms involved in the induction of matrix regulatory proteins are unclear but may reflect both hemodynamic and humoral responses. Increased MMP-9 and MMP-12, which can hydrolyze elastin, may be secondary to reactive oxidative species and inflammation in diabetic vessels. Uemura et al. demonstrated an increase in MMP-9 activity, mRNA and protein expression in endothelial cells exposed to hyperglycemia, which may facilitate the pro-inflammatory response in the vessel wall [41]. On the other hand, the increased expression of pro-fibrotic factors, such as PAI-1 and ECM accumulation is consistent with DM-induced nephropathy and vasculopathy [17,42]. Similarly, other groups report an imbalance between MMP and TIMP levels that lead to increased synthesis and accumulation of matrix proteins in the heart and kidneys of experimental models of DM [15,16]. Plasma and urine samples from human diabetic patients show marked increases in MMP and TIMP concentrations, reflecting abnormal ECM turnover [43,44]. It is also important to note that it is the overall organization of the ECM into a 3-dimensional network, as well as the fluid properties of the non-collagenous components of the ECM (elastin, fibrin) that will ultimately impart elasticity and compliance to the tissue [45]. Thus the overall balance between anti-fibrotic and pro-fibrotic factors at this time point maintains ECM integrity without accumulation in an attempt limit outward vessel remodeling. Future studies examining the temporal and spatial profile of MMPs and TIMPs in DM-induced microvascular remodeling are warranted.
In summary, our study shows significant hypertrophic outward remodeling of MRAs from obese, type 2 diabetic db/db mice that is accompanied by increased compliance and wall stress. This remodeling likely reflects an integrated response of a combination of increased regional hemodynamics that precedes structural adaptation, type 2 DM-dependent insulin resistance, inflammation, obesity, and oxidative stress. Changes in the balance between pro-and anti-fibrotic factors resulted in no net ECM accumulation that work in concert to reduce wall stress and limit outward remodeling.