Chapter 6 Bone Marrow-Derived JNK2-Positive Cells in Perivascular Adipose Tissue Blunt Microvascular Insulin-Induced Vasodilation Early During Nutrient Excess Rick I. Meijer, Erik H. Serné, John S. Yudkin, Tom J.A. Kokhuis, Ester M. Weijers, Victor W.M. van Hinsbergh, Yvo M. Smulders and Etto C. Eringa

Manuscript in preparation

143

Abstract Rationale Obesity is associated with impaired delivery of insulin

and glucose by the microcirculation in skeletal muscle. Perivascular adipose tissue (PVAT) has anticontractile properties that are lost in obesity, thereby blunting insulin-induced vasodilation. We hypothesized that microvascular dysfunction develops during a two-week Western Diet (WD) and that PVAT is involved in this development, both in- and ex-vivo. We further hypothesized that c-Jun N-terminal kinase 2 (JNK2) expressing cells from the bone marrow in PVAT are causative for the change in PVAT function.

Methods and Results In C57Bl/6 mice after two weeks of WD, PVAT weight around the gracilis arteries increased compared to chow-fed controls. Mice on WD had reduced insulin-induced microvascular recruitment, along with insulin resistance, as shown by Contrast Enhanced Ultrasound (CEU) during a hyperinsulinemic euglycemic clamp. To study whether perfusion defects were caused by altered effects of PVAT on muscle resistance arteries (RA), insulin responses of RA from chow or WD mice were studied by pressure myography, in the absence, or presence of PVAT. Chow RA showed insulin-induced vasodilation after incubation with chow PVAT, but not when incubated alone or with WD PVAT. Moreover, insulin induced vasoconstriction in WD RA with WD PVAT, but not chow PVAT. Transplantation with JNK2-/- bone marrow restored the insulin-induced vasodilation by WD-PVAT.

Conclusions Impaired insulin-induced muscle perfusion is an early phenomenon during exposure to nutrient excess and is caused by endothelial as well as PVAT dysfunction. The latter is caused by infiltration of JNK2 positive cells from bone marrow in PVAT.

144

Introduction Increased adiposity increases the risk for insulin resistance and

type 2 diabetes, as well as cardiovascular disease,1 but our knowledge of the underlying mechanisms is incomplete. Skeletal muscle is the major site of action for insulin-stimulated glucose uptake, and insulin causes microvascular vasodilation in skeletal muscle, thereby enhancing access of nutrients by increasing the microvascular surface area available for exchange. These actions occur through the IRS-1, PI3k, Akt, eNOS, NO pathway.2, 3 Impairment of endothelial vasoregulation, particularly in the microcirculation, is seen in obesity and contributes to metabolic insulin resistance through a decrease in insulin-induced microvascular recruitment.4, 5 Microvascular endothelial dysfunction may occur early during nutrient excess as it is suggested that insulin signaling is downregulated in larger vessels after two weeks of increased nutrient intake.6 The mechanisms underlying microvascular dysfunction have been extensively studied, but remain poorly defined.

6Because of the control of perivascular adipose tissue (PVAT) over

endothelial vasomotor responses to various vasoactive substances,7-10 PVAT may be an early mediator in diet-induced microvascular dysfunction. We recently found an important regulatory role for PVAT in insulin-induced vasodilation.10 Despite pronounced effects of PVAT on vasoreactivity, interventions to improve PVAT function are scarce. Furthermore, the role of PVAT in regulation of skeletal muscle perfusion is not well studied. PVAT consists of adipocytes, but can also contain inflammatory cells such as macrophages, mast cells and T-cells.11 During nutrient excess the number of macrophages in PVAT increases, and their phenotype becomes pro-inflammatory, decreasing adiponectin production and increasing production of pro-inflammatory adipokines,12 a process involving c-Jun NH2-terminal kinase (JNK).13 The increase in the quantity of PVAT and the change in the adipokine profile in obesity raises the question whether this results from pre-existent cells, or newly recruited cells.

JNK in PVAT may be a regulator of insulin’s microvascular effects. JNK is involved in the pathogenesis of both insulin resistance and microvascular dysfunction.14-17 We have previously shown involvement of JNK in PVAT dysfunction.10 JNK promotes M1 polarization of macrophages,16 and decreases insulin sensitivity.14 JNK

145

is encoded by three genetic loci, JNK1, JNK2 and JNK3, of which both JNK1 and JNK2 are involved in the development of insulin resistance,18,

19 and JNK2 is implicated in vascular pathologies that are strongly linked to dysfunctional PVAT.20, 21 Whether JNK2 affects insulin resistance at least through microvascular dysfunction or PVAT dysfunction is unknown.

We investigated whether microvascular insulin resistance is already apparent after two weeks Western Diet (WD) and whether PVAT is involved in this process. We further investigated whether deletion of JNK2 in the bone marrow affects the functional phenotype of PVAT and resistance arteries. Research design and methods Animals

All animal work was approved by the local ethics committee for animal experiments of the VU University Amsterdam and complied with Dutch government guidelines. Male C57Bl/6NCrl mice (Bl/6) (The Jackson Laboratory, Amsterdam, The Netherlands) and male B6.129S2-Mapk9tm1Flv/J (JNK2-/-) (The Jackson Laboratory, Amsterdam, The Netherlands) were ordered at 4 weeks of age. Bl/6 mice were used as JNK2+/+ controls. Ubiquitin Green Fluorescent Protein+/+ mice were supplied by R. Mebius. Bl/6 mice were fed normal chow or Western Diet (WD) (21% fat, D12079B, Research Diets, New Brunswick) ad libitum for two weeks, starting at 6 weeks of age and studied at 8 weeks. Surgical procedure

Mice were anaesthetized with Fentanyl (0.31 mg/kg), Midazolam (6.25 mg/kg) and Acepromazine (6.25 mg/kg) intraperitoneally. The right jugular vein was cannulated with polyethylene tubing (PE-10). Microvascular perfusion imaging

The proximal adductor muscle group of the upper hindleg was imaged in cross-section with Contrast-Enhanced Ultrasonography (CEU). Phospholipid-coated microbubbles with a perfluorobutane (C4F10) gas core were made by sonication with a 1,2-distearyol-sn-glycero-3-phosphocholine (DSPC), polyoxyethylene-40 (PEG 40) stearate and 1,2-distearyol-sn-glycero-3-phosphoethanolamine-N-

146

[carboxy(polyethylene glycol)2000] (DSPE-PEG2000) coating (Erasmus Medical Center, Rotterdam, The Netherlands).22 Microbubbles were infused at 5 μl/min. Three consecutive replenishment curves were imaged at a mechanical index of 0.12, each following a one-second, 1.9 mechanical index pulse destroying all microbubbles in the cross-section. Video-intensities were analyzed with the Image Processing toolbox in MATLAB. Inflow curves were fitted to VI = MBV(1-e-MFV(t-0.5)) VI is the video intensity at a given time (t in secs), MBV is Microvascular Blood Volume, MFV is Microvascular Flow Velocity and e the natural logarithm. VI in skeletal muscle was normalized on VI in the femoral artery, correcting for differences in microbubble density in the circulation.

6 Euglycemic hyperinsulinemic clamp

Insulin sensitivity was assessed using a primed euglycemic hyperinsulinemic clamp; 200 mU/kg; Actrapid, Novo Nordisk, France, followed by continuous infusion (3 mU/kg/min) for 60 minutes, together with variable infusion of 20% D-glucose to maintain euglycemia. Blood glucose was assessed from the tail vein with a Freestyle Precision Xceed (Abbott, Hoofddorp, The Netherlands), and maintained at 5 mmol/l (90 mg/dl) by adjusting the variable glucose infusion rate. Bone marrow transplantation

To study the effect of the bone marrow on PVAT, we created chimeras of Bl/6 mice with bone marrow from GFP+/+, JNK2-/-, or JNK2+/+ mice. Recipient Bl/6, as well as donor mice were put on acidified drinking water with neomycin (100 μg/ml) and polymyxin B (8 μg/ml), and had chow diet, one week before bone marrow transplantation. Recipient-mice were irradiated one day before transplantation (2*6 Gy). Bone marrow was suspended in RPMI-1640, 2% fetal calf serum and heparin (5 U/ml). Bl/6 mice were injected intravenously with 5*106 cells. Mice are indicated by the type of diet they received, (Chow for chow diet and WD for Western Diet) and by the bone marrow transplantation. Thus:

mice without bone marrow transplantation are only indicated by their diet

147

chimeras with JNK2-/- bone marrow are indicated first by their diet, and then the bone marrow. Thus chimeras with JNK2-/- bone marrow: chowJNK2-/- BMResistance Artery (RA) and WDJNK2-/-

BMPVAT and chimeras with JNK2+/+ bone marrow chowJNK2+/+ BMRA and WDJNK2+/+ BMPVAT.

Ex-vivo vasoreactivity

Ex vivo insulin-induced vasoreactivity of isolated RAs that feed the gracilis muscle in the mouse hindlimb was studied as described previously.10 Resistance arteries were randomly assigned to be pre-incubated for 30 minutes without PVAT, with chow PVAT, or WD PVAT. All PVAT adjacent to the RA was isolated from the origin of the gracilis artery at the femoral artery, until the first major side branch.

After preconstriction with KCl (25 mM), the inner diameter of resistance arteries was recorded to determine baseline diameter and diameter changes induced by four concentrations of insulin (0.02, 0.2, 2.0 and 20 nM), each exposure being for 30 minutes. Endothelial integrity was determined by measuring responses to the endothelium-dependent vasodilator acetylcholine 1*10-7M (ACh) after each experiment, an RA failing to achieve at least 10% vasodilation to Ach was excluded from all analyses. Statistics

Normally distributed data are presented as mean ± SE, and as median and range when not normally distributed. Differences in weights of mice and of epididymal adipose tissue, insulin-sensitivity, infiltration of GFP-positive cells, as well as baseline vasoreactivity parameters were tested with unpaired Student’s t-tests, and the difference in PVAT weight was tested with a Mann-Whitney U test. Changes in microvascular blood volume and microvascular flow velocity during hyperinsulinemia were analyzed with a paired t-test. Ex-vivo vasoreactivity was tested for differences with a one-way ANOVA with Bonferroni post-hoc correction. Diameter-values and p-values in the legends represent analyses of reactivity to the 2.0 nM insulin concentration. Results

148

Two weeks WD reduces insulin sensitivity and insulin-induced microvascular recruitment

Metabolic insulin sensitivity was measured with a hyperinsulinemic euglycemic clamp in mice fed either chow or WD for two weeks. Insulin sensitivity was reduced by 48% in WD mice as compared to chow mice, as indicated by lower glucose infusion rates at similar levels of glycemia (figure 1A, B). Fasting glucose was unaffected by the WD (figure 1C). During the clamp, microvascular blood volume (MBV) increased in chow mice, (figure 1D, supplementary figure S1), but not in WD mice (figure 1E). This indicates a decrease in insulin-induced microvascular recruitment by WD. Microvascular flow velocities (MFV) decreased insignificantly in both the chow and the WD mice (Supplementary figure S1).

Increased PVAT weight and epididymal adipose tissue weight precede increased body weight during short-term WD

6After two weeks WD, epididymal adipose tissue weight was 70%

higher in WD mice than in chow mice, while overall body weight did not significantly differ between the two groups (figure 2A, B). Figures 2D and 2E show the location and extent of PVAT in the hindlimb and a cross-section of PVAT around the gracilis resistance artery respectively. Interestingly, PVAT surrounding the resistance artery (RA) feeding the gracilis muscle in WD mice weighed three times more than in chow mice despite a large interindividual variation (P<0.01; figure 2C). These data show that hypertrophy of PVAT is more pronounced than that of other adipose tissue depots.

149

Figure 1. Insulin sensitivity and microvascular recruitment are reduced in WD mice A. Chow mice (n=11) are more insulin sensitive than WD mice (n=9) during the last 30 minutes of a 60 minute hyperinsulinemic euglycemic clamp, as evidenced by higher glucose infusion rates during this period, *p<0.05. B. Mean blood glucose concentrations during the clamp, there was no significant difference between the two groups, p=0.29. C. Fasting blood glucose concentration was not significantly different in chow and WD, p=0.85. D. Pictures at different moments during the replenishment curve. First (black and white) picture during high mechanical index microbubble destruction, the middle picture depicts the first frame after microbubble destruction and shows very low video intensity within the region of interest in the muscle. The femoral artery can be seen at the top just above the region of interest. The latest picture shows a high video intensity in the region of interest. A corresponding typical example replenishment curve is shown as well. E. Microvascular Blood Volume (MBV) measured as video-intensity, increased during the clamp in chow mice (n=11) by approximately 38% (*p<0.05, compared to baseline). In WD mice (n=9), video-intensity did not increase from baseline to hyperinsulinemia, p=0.49.

150

6Figure 2. Baseline characteristics of chow and WD mice A. Body weights at the end of the two-week diet intervention were similar in chow (n=11) and WD mice (n=10), p=0.17. B. Epididymal adipose tissue mass increased by more than 70%, *p<0.02 between chow and WD mice. C. PVAT weight from the major RA supplying the gracilis muscle increases more than threefold during WD, **p<0.01. D. Pictures of the gracilis artery (left), gracilis vein (right) and adjacent PVAT in chow and WD mice. E. Hematoxylin eosin staining of the hindleg. A=Artery, LA=Lamina Adventitia, LI =Lamina Intima, LM=Lamina Media, M=muscle, PVAT=Perivascular Adipose Tissue, V=vein

Two weeks WD blunts the capability of PVAT to reveal insulin-induced vasodilation and also alters the response of resistance arteries to insulin Because insulin-induced microvascular recruitment was diminished by WD, and weight of PVAT had increased, we explored whether WD affected the vasodilator potential of insulin, as well as the influence of PVAT on resistance arteries (RAs). To this end, we combined chow and WD RAs with both chow and WD PVAT. Insulin induced a concentration-dependent vasodilation of chow RAs with chow PVAT, but not without PVAT (figure 3A), corroborating previous findings.10 In contrast, vasodilation was not seen in chow RA with WD PVAT (figure 3A). Moreover, WD RA did not show insulin-induced vasodilation, regardless of PVAT-origin (figure 3B). WD did not affect

151

maximum diameters, vascular tone or acetylcholine-mediated vasodilation (figures 3C-3E; supplemental table S1). Nevertheless, together, these findings indicate that two weeks WD affected both the PVAT-phenotype and endothelial insulin-signaling in RAs.

Figure 3. PVAT from WD mice does not uncover insulin-induced vasodilation and WD RAs become insensitive to the effects of chow PVAT A. Maximum RA diameters did not differ between chow (n=25) and WD RA (n=24), p=0.59. B. Tone development in RAs was also not different between chow and WD RA, p=0.73. C. Acetylcholine-mediated vasodilation remained unaffected by the WD, being similar in the two groups, p=0.93. D. Chow RAs pre-incubated with chow PVAT (n=11) exhibit insulin-induced vasodilation, compared to those without PVAT (n=8), *p<0.05, ***p<0.001. The capability to uncover insulin-induced vasodilation is restricted to chow PVAT, as vasodilation with WD PVAT (n=7) is restricted, p=0.82 vs. chow PVAT. E. WD RA without PVAT (n=9) did not vasodilate and neither did nor WD RA with chow PVAT (n=8), p=0.63. WD RA with WD PVAT showed a trend towards vasoconstriction compared to no PVAT (n=7), p=0.06

Short-term WD induces adiponectin resistance in RAs

We hypothesized that the WD RA did not exhibit insulin-induced vasodilation despite presence of chow PVAT due to adiponectin resistance, because of adiponectin’s vasodilator effects. Indeed, while

152

adiponectin enabled insulin-induced vasodilation in chow RA, this effect was lost in WD RA (figure 4).

Figure 4. WD is accompanied by adiponectin resistance in WD RAs WD RA incubated with adiponectin did not vasodilate (n=5) when insulin was added to the organ-bath, in contrast to the chow RA (n=6), *p<0.05.

6 Bone marrow-derived cells infiltrate PVAT during WD

To study to what extent influx of inflammatory cells contributes to PVAT expansion, we transplanted bone marrow from Bl/6 mice expressing Green Fluorescent Protein (GFP) to Bl/6 mice before the start of the diet. Indeed, after two weeks WD, we observed an increase of GFP-positive cells in PVAT (figure 5B). In WD PVAT T-lymphocytes, mast cells and monocytes/macrophages were encountered (Supplementary figure S3). Unexpectedly, the GFP expressing cells in WD PVAT not only consisted of small inflammatory cells, but, in contrast to chow PVAT, a portion of adipocytes was also GFP-positive (Figure 5A).

153

Figure 5. Bone marrow derived cells infiltrate PVAT during WD A. GFP expression is seen in leucocytes (yellow arrows), but also adipocytes (white arrow) of WD mice. Adipocytes in chow mice did however not express GFP. As a comparison a picture of a whole body GFP+/+ mouse is given, as well as a negative control. B. Total GFP expression is higher in WD PVAT (n=5) compared to chow PVAT (n=3). Column height is shown in arbitrary units, p<0.05.

Transplantation with JNK2-deficient bone marrow salvages the vasodilator phenotype of PVAT during exposure to WD

In absence of PVAT, transplantation with either JNK2+/+ bone marrow, or JNK2-/- bone marrow did not affect RA responses to insulin (figure 6A, compare figures 3A, B). Moreover, both WDJNK2+/+BMRAs and WDJNK2-/-BMRAs did not exhibit vasodilation in the presence of PVAT,

154

independent of its origin (chow vs. WD; JNK2+/+ vs. JNK2-/-) (figure 6B). The latter shows that deletion of JNK2 in the bone marrow does not protect the endothelium from the detrimental effects of WD.

Similarly to the RAs of non-chimera mice, chowJNK2+/+BMRAs with chowJNK2+/+BMPVAT and chowJNK2-/-BMRAs with chowJNK2-/-BMPVAT dilated normally in response to insulin (figure 6C). However, in contrast to WDJNK2+/+BMPVAT, which did not exhibit vasodilation in chowJNK2+/+BMRAs, PVAT from WDJNK2-/-BMmice restored insulin induced vasodilation in chowJNK2-/-BMRAs (figure 6D). This indicates that infiltration of JNK2 positive cells into PVAT during WD play a pivotal in the loss of its vasodilator properties.

6

Figure 6. PVAT phenotype during WD can be restored by deletion of JNK2-expressing cells in the bone marrow A. Without PVAT, insulin does not induce vasodilation in either chow JNK2+/+BMRA (n=5), p=0.07, WD JNK2+/+BMRA (n=3), p=0.35, chow JNK2-/-BMRA (n=5), p=0.30 or WD JNK2-

/-BMRA (n=7), p=0.36. Lines are overlapping, therefore, not all lines can be easily distinguished. B. In the presence of chow PVAT from chimera mice WD JNK2+/+BMRA (n=5), p=0.55 and WD JNK2-/-BMRA (n=5), p=0.21 do not exhibit insulin-induced vasodilation. With WD

155

PVAT from chimera mice WD JNK2+/+BMRA (n=5), p=0.55 and WD JNK2-/-BMRA (n=5), p=0.55, also showing a neutral response. C. Chow JNK2+/+BMRAs with chow JNK2+/+BMPVAT (n=5) and chowJNK2-/-BMRAs with chow JNK2-/-BMPVAT (n=6) vasodilated after addition of insulin, p<0.05 for both compared to baseline, p=0.36 between these two conditions. D. Chow JNK2-/-BMRAs show insulin-induced vasodilation with WD JNK2-/-BMPVAT (n=7), whereas chowJNK2+/+BMRAs did not show insulin-induced vasodilation with WD JNK2+/+BMPVAT (n=5). Absence of JNK2 in the RA does not influence the responsiveness to PVAT *p<0.05.

Maximum RA diameters (figure 7A) nor tone development (figure

7B) were affected by JNK2 status. Although acetylcholine induced vasodilation was not significantly affected by JNK2 genotype (Figure 7C), there was a trend towards higher acetylcholine-induced vasodilation in both chowJNK2-/-BMRAs and WDJNK2-/-BMRAs, compared with their JNK2+/+BMcontrols.

156

6

Figure 7. Baseline characteristics of RAs are not affected by bone marrow transplantation A. Maximum diameters of gracilis RAs are not affected by diet or JNK2 genotype, p=0.63. B. Tone development is not affected by JNK2 deletion in the bone marrow, neither does the diet in the vessels with JNK2+/+ bone marrow, but surprisingly, tone was lower in WD JNK2-/-BMvessels, compared to Chow JNK2-/-BMvessels, *=p<0.05. C. Acetylcholine mediated vasodilation is not significantly different between chow or WD chimeras with bone marrow from either JNK2-/- mice or JNK2+/+ mice,, despite there being a trend towards higher acetylcholine mediated vasodilation in JNK2-/-BMRAs (p=0.32 in the chow RAs and p=0.13 in WD RAs).

157

Discussion In the present study, we have shown that PVAT in skeletal

muscle loses its vasodilator properties after two weeks exposure to a Western diet. Moreover, endothelial insulin responses to PVAT are also blunted, due to adiponectin-resistance. During these two weeks cells from the bone marrow infiltrate PVAT, altering its vasodilator phenotype. JNK2-expressing cells are pivotal in this change in PVAT, as deletion of JNK2 in the bone marrow prior to the start of WD salvages the PVAT phenotype. JNK2-deficient bone marrow does not restore the phenotype of the endothelium.

It has previously been described that metabolic insulin resistance develops in muscle after short periods of high fat feeding.23 Whether microvascular insulin resistance also develops soon in the course of nutrient excess was not known. This study has shown that insulin-induced microvascular recruitment is indeed blunted early during Western Diet. This bears importance as microvascular insulin resistance decreases insulin transport and delivery, and is also implicated in the development of microvascular complications.24

PVAT from chow mice permits insulin-induced vasodilation in chow RA, indicating an important role of PVAT in microvascular function. In contrast, WD PVAT did not permit vasodilation. Importantly, without PVAT, no altered responses were observed in WD RAs. Strikingly, the increase in PVAT weight was subject to large variation, nevertheless, the lost potential to uncover insulin-induced vasodilation, i.e. the vasodilator phenotype of PVAT, was shared between all WD PVAT samples. This suggests that exposure to WD either reduces production of adipokines that permit insulin-induced vasodilation in PVAT, or increases production of vasoconstrictor adipokines.

WD RAs become insensitive to the anticontractile effects of chow PVAT. When we exposed WD RAs to globular adiponectin, they also did not show vasodilation on insulin. Therefore, the endothelium remains the most important determinant for the endothelium-PVAT interaction. This finding corroborates earlier findings in the Zucker Diabetic Fatty (ZDF) rat, that ZDF RA do not elicit NO-dependent vasodilation in

158

response to globular and full-length adiponectin, despite similar expression of adiponectin receptors as in RAs of Zucker lean rats.25 Adiponectin also regulates insulin-induced increases in microvascular blood volume in vivo through the Akt-eNOS-NO pathway.26 Together, this suggests that either adiponectin coming from PVAT, or other proteins which act through the AdipoR1-receptor, are important regulators of vasomotor responses. Once endothelial damage during a diet has occurred, signaling through the AdipoR1-receptor can not acutely restore this. Downregulation of AMPKα2 offers an elegant explanation for this effect. PVAT has been implicated in affecting (micro-)vascular vasodilation and vasoconstriction,7, 10 but its quantity is also associated with an increased risk for atherosclerosis.27 Whether the functional changes of PVAT observed here also affect large arteries is nevertheless unknown. 6

Transplantation with GFP+/+ bone marrow showed that not only inflammatory cells in PVAT are bone-marrow-derived, but a peculiar finding was that perivascular adipocytes express GFP in WD mice, whereas adipocytes from chow mice did not express GFP. The contribution of the adipocytes apparently derived from the bone marrow is unclear. We could show that the new inflammatory cells in PVAT were at least macrophages, mast cells and T-lymphocytes (supplemental figure S3). Through the specific deletion of JNK2 in the bone marrow, we showed that deficiency of JNK2 in bone marrow-derived cells prevents PVAT from being affected by WD. We did not observe an effect of this transplantation on the endothelium. However, cross-talk from bone marrow to the endothelium has previously been described.28-30 Whereas these studies targeted eNOS production, and thereby specifically targeted the endothelium. Nevertheless, this does not exclude an important role for JNK2 in endothelium, especially because pharmacological inhibition of JNK restored the vasodilator interaction between db/db PVAT and control RAs.10 The protective effect on WDJNK2-

/-BMPVAT was not due to irradiation and bone marrow transplantation per se, RAs from chimera mice showed essentially the same results as RAs from normal mice (figures 6G and 3D). It has been reported that JNK1 almost completely compensates for JNK2 deficiency in whole-body knockout mice during a diet.18 Our results partially contradict those

159

results, as deletion solely of JNK2 in bone marrow restored insulin-induced vasodilation in the presence of WD-PVAT. Possible explanations for this finding are that JNK2 is more important for PVAT properties than JNK1, and that we used a much shorter duration diet than others.18, 20 Because JNK1 and JNK2 are differentially activated, their activation-profile at the commencement of inflammation may differ.20 The current results with JNK2 deficiency in PVAT confirm and extend our previous findings using pharmacological inhibition of JNK to restore db/db PVAT and prevent TNF-α mediated blunting of insulin-induced vasodilator signalling.10, 17 Together these findings suggest that JNK2 in PVAT is a potential target for pharmacological interventions to treat obesity-dependent microvascular dysfunction and insulin resistance.

A limitation of the study is that we were unable to study in-vivo insulin sensitivity and microvascular responses in the chimeras, as mice did not tolerate the combination of bone marrow transplantation, hyperinsulinemic clamp and microvascular perfusion measurements. Another limitation is that we do not know yet which JNK2-expressing cell type determines the PVAT-phenotype during WD exposure. These questions will be addressed in follow-up studies.

In conclusion, the property of PVAT to synergize with insulin, evoking vasodilation in muscle resistance arteries is lost rapidly during Western diet exposure, in parallel to impairment of hyperinsulinemia-induced microvascular recruitment in vivo. The loss of anticontractile effect of PVAT during WD is critically dependent on JNK2 in bone marrow-derived cells. Acknowledgments Lona Kroese and Anne Komulainen provided excellent technical assistance. We thank Stan Heukelom and Phil W. Koken, department of Radiotherapy, VU University Medical Center, Amsterdam, The Netherlands for performing irradiation of the mice. We acknowledge Reina E. Mebius, department of Molecular Cell Biology and Immunology, VU University Medical Center, Amsterdam, The

160

Netherlands for kindly providing the ubiquitin Green Fluorescent Protein mice. Sources of Funding This research is sponsored by the Netherlands Heart Foundation (Grant 2009B098) and The Netherlands Organization for Scientific Research (Grant 916.76.179).

6

161

References 1. Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ,

Singh GM, Gutierrez HR, Lu Y, Bahalim AN, Farzadfar F, Riley LM, Ezzati M. National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet 2011;377:557-567.

2. Premilovac D, Bradley EA, Ng HL, Richards SM, Rattigan S, Keske MA. Muscle insulin resistance resulting from impaired microvascular insulin sensitivity in Sprague Dawley rats. Cardiovascular Research 2013;98:28-36.

3. de Boer MP, Meijer RI, Wijnstok NJ, Jonk AM, Houben AJ, Stehouwer CD, Smulders YM, Eringa EC, Serné EH. Microvascular dysfunction: a potential mechanism in the pathogenesis of obesity-associated insulin resistance and hypertension. Microcirculation 2012;19:5-18.

4. Meijer RI, de Boer MP, Groen MR, Eringa EC, Rattigan S, Barrett EJ, Smulders YM, Serné EH. Insulin-induced microvascular recruitment in skin and muscle are related and both are associated with whole body glucose uptake. Microcirculation 2012;10-8719.

5. Kubota T, Kubota N, Kumagai H, Yamaguchi S, Kozono H, Takahashi T, Inoue M, Itoh S, Takamoto I, Sasako T, Kumagai K, Kawai T, Hashimoto S, Kobayashi T, Sato M, Tokuyama K, Nishimura S, Tsunoda M, Ide T, Murakami K, Yamazaki T, Ezaki O, Kawamura K, Masuda H, Moroi M, Sugi K, Oike Y, Shimokawa H, Yanagihara N, Tsutsui M, Terauchi Y, Tobe K, Nagai R, Kamata K, Inoue K, Kodama T, Ueki K, Kadowaki T. Impaired insulin signaling in endothelial cells reduces insulin-induced glucose uptake by skeletal muscle. Cell Metab 2011;13:294-307.

6. Kim F, Pham M, Maloney E, Rizzo NO, Morton GJ, Wisse BE, Kirk EA, Chait A, Schwartz MW. Vascular inflammation, insulin resistance, and reduced nitric oxide production precede the onset of peripheral insulin resistance. Arterioscler Thromb Vasc Biol 2008;28:1982-1988.

7. Greenstein AS, Khavandi K, Withers SB, Sonoyama K, Clancy O, Jeziorska M, Laing I, Yates AP, Pemberton PW, Malik RA, Heagerty AM. Local inflammation and hypoxia abolish the protective anticontractile properties of perivascular fat in obese patients. Circulation 2009;119:1661-1670.

8. Rittig K, Staib K, Machann J, Bottcher M, Peter A, Schick F, Claussen C, Stefan N, Fritsche A, Haring HU, Balletshofer B. Perivascular fatty tissue at the brachial artery is linked to insulin resistance but not to local endothelial dysfunction. Diabetologia 2008;51:2093-2099.

162

9. Wang H, Luo W, Wang J, Guo C, Wang X, Wolffe SL, Bodary PF, Eitzman DT. Obesity-induced endothelial dysfunction is prevented by deficiency of P-selectin glycoprotein ligand-1. Diabetes 2012;61:3219-3227.

10. Meijer RI, Bakker W, Alta CL, Sipkema P, Yudkin JS, Viollet B, Richter EA, Smulders YM, van Hinsbergh, VWM, Serné EH, Eringa EC. Perivascular Adipose Tissue Control of Insulin-Induced Vasoreactivity in Muscle Is Impaired in db/db Mice. Diabetes 2013;62:590-598.

11. Takaoka M, Suzuki H, Shioda S, Sekikawa K, Saito Y, Nagai R, Sata M. Endovascular injury induces rapid phenotypic changes in perivascular adipose tissue. Arterioscler Thromb Vasc Biol 2010;30:1576-1582.

12. Chatterjee TK, Stoll LL, Denning GM, Harrelson A, Blomkalns AL, Idelman G, Rothenberg FG, Neltner B, Romig-Martin SA, Dickson EW, Rudich S, Weintraub NL. Proinflammatory Phenotype of Perivascular Adipocytes Influence of High-Fat Feeding. Circ Res 2009;104:541-549.

13. Withers SB, gabiti-Rosei C, Livingstone DM, Little MC, Aslam R, Malik RA, Heagerty AM. Macrophage activation is responsible for loss of anticontractile function in inflamed perivascular fat. Arterioscler Thromb Vasc Biol 2011;31:908-913.

614. Hirosumi J, Tuncman G, Chang L, Gorgun CZ, Uysal KT, Maeda K, Karin

M, Hotamisligil GS. A central role for JNK in obesity and insulin resistance. Nature 2002;420:333-336.

15. Solinas G, Vilcu C, Neels JG, Bandyopadhyay GK, Luo JL, Naugler W, Grivennikov S, Wynshaw-Boris A, Scadeng M, Olefsky JM, Karin M. JNK1 in hematopoietically derived cells contributes to diet-induced inflammation and insulin resistance without affecting obesity. Cell Metab 2007;6:386-397.

16. Han MS, Jung DY, Morel C, Lakhani SA, Kim JK, Flavell RA, Davis RJ. JNK expression by macrophages promotes obesity-induced insulin resistance and inflammation. Science 2013;339:218-222.

17. Eringa EC, Stehouwer CD, Walburg K, Clark AD, van Nieuw Amerongen GP, Westerhof N, Sipkema P. Physiological concentrations of insulin induce endothelin-dependent vasoconstriction of skeletal muscle resistance arteries in the presence of tumor necrosis factor-alpha dependence on c-Jun N-terminal kinase. Arterioscler Thromb Vasc Biol 2006;26:274-280.

18. Tuncman G, Hirosumi J, Solinas G, Chang L, Karin M, Hotamisligil GS. Functional in vivo interactions between JNK1 and JNK2 isoforms in obesity and insulin resistance. Proc Natl Acad Sci U S A 2006;103:10741-10746.

19. Singh R, Wang Y, Xiang Y, Tanaka KE, Gaarde WA, Czaja MJ. Differential effects of JNK1 and JNK2 inhibition on murine steatohepatitis and insulin resistance. Hepatology 2009;49:87-96.

163

20. Bogoyevitch MA. The isoform-specific functions of the c-Jun N-terminal Kinases (JNKs): differences revealed by gene targeting. Bioessays 2006;28:923-934.

21. Osto E, Matter CM, Kouroedov A, Malinski T, Bachschmid M, Camici GG, Kilic U, Stallmach T, Boren J, Iliceto S, Luscher TF, Cosentino F. c-Jun N-terminal kinase 2 deficiency protects against hypercholesterolemia-induced endothelial dysfunction and oxidative stress. Circulation 2008;118:2073-2080.

22. Klibanov AL, Rasche PT, Hughes MS, Wojdyla JK, Galen KP, Wible JH, Jr., Brandenburger GH. Detection of individual microbubbles of ultrasound contrast agents: imaging of free-floating and targeted bubbles. Invest Radiol 2004;39:187-195.

23. Park SY, Cho YR, Kim HJ, Higashimori T, Danton C, Lee MK, Dey A, Rothermel B, Kim YB, Kalinowski A, Russell KS, Kim JK. Unraveling the temporal pattern of diet-induced insulin resistance in individual organs and cardiac dysfunction in C57BL/6 mice. Diabetes 2005;54:3530-3540.

24. Mather KJ, Steinberg HO, Baron AD. Insulin resistance in the vasculature. J Clin Invest 2013;123:1003-1004.

25. Schmid PM, Resch M, Steege A, Fredersdorf-Hahn S, Stoelcker B, Birner C, Schach C, Buechler C, Riegger GA, Luchner A, Endemann DH. Globular and full-length adiponectin induce NO-dependent vasodilation in resistance arteries of Zucker lean but not Zucker diabetic fatty rats. Am J Hypertens 2011;24:270-277.

26. Zhao L, Chai W, Fu Z, Dong Z, Aylor KW, Barrett EJ, Cao W, Liu Z. Globular Adiponectin Enhances Muscle Insulin Action via Microvascular Recruitment and Increased Insulin Delivery. Circ Res 2013;112:1263-1271.

27. Meijer RI, Serné EH, Smulders YM, Van Hinsbergh VWM, Yudkin JS, Eringa EC. Perivascular adipose tissue and its role in type 2 diabetes and cardiovascular disease. Curr Diab Rep 2011;11:211-217.

28. Nakajima S, Ohashi J, Sawada A, Noda K, Fukumoto Y, Shimokawa H. Essential role of bone marrow for microvascular endothelial and metabolic functions in mice. Circ Res 2012;111:87-96.

29. Furuno Y, Morishita T, Toyohira Y, Yamada S, Ueno S, Morisada N, Sugita K, Noguchi K, Sakanashi M, Miyata H, Tanimoto A, Sasaguri Y, Shimokawa H, Otsuji Y, Yanagihara N, Tamura M, Tsutsui M. Crucial vasculoprotective role of the whole nitric oxide synthase system in vascular lesion formation in mice: Involvement of bone marrow-derived cells. Nitric Oxide 2011;25:350-359.

30. Aicher A, Heeschen C, Mildner-Rihm C, Urbich C, Ihling C, Technau-Ihling K, Zeiher AM, Dimmeler S. Essential role of endothelial nitric oxide

164

synthase for mobilization of stem and progenitor cells. Nat Med 2003;9:1370-1376.

6

165

Supplementary figures

1A Graphic example of a real-time replenishment curve at baseline and hyperinsulinemia. The plateau reached after approximately 15 seconds is well recognizable and represents MBV. The steepness of the curve is the β or MFV. The lower panels show how the microbubbles in the skeletal muscle microcirculation refill after being destructed (during first panel), directly after the destruction the skeletal muscle in the region of interest is almost completely devoid of contrast. After 30 seconds (last panel), the microbubbles have refilled the microcirculation. 1B Microvascular flow velocity decreases non-significantly in both chow and WD mice during hyperinsulinemia.

166

Supplemental Table S1

Chow

RA, no

PVAT

WD

RA, no

PVAT

Chow

RA +

Chow

PVAT

Chow

RA +

WD

PVAT

WD RA

+ Chow

PVAT

WD RA

+ WD

PVAT

P-

value

Acetylcholine-

induced

vasodilation

(%)

67±14 57±9 78±2 34±7 51±11 47±9 NS

Tone (% of

Dmax) 36±4 37±6 38±6 37±5 47±2 33±5 NS

Dmax (μm) 126±22 146±6 131±8 121±7 135±8 132±10 NS

6

Supplemental figure 2 Upper left panel and the middle panel reveal the presence of T-lymphocytes in PVAT. The upper right panel shows mast cells in PVAT Macrophages are also present in PVAT as shown in the bottom panel.

167

Supplemental Table S2

Acetylcholine-induced

vasodilation (%)

Tone (% of Dmax) Dmax (μm)

JNK2-/- BMChow RA, no

PVAT 56±15 44±6 122±8

JNK2-/-BMWD, no PVAT 65±9 41±4 118±12

JNK2-/- BMChow RA +

JNK2-/- BMChow PVAT 54±23 52±5 132±8

JNK2-/- BMChow RA +

JNK2-/-WD PVAT 59±13 47±7 129±9

JNK2-/- BMWD+JNK2-/-

Chow PVAT 47±19 25±8 124±8

JNK2-/- BMWD + JNK2-

/-WD PVA 51±18 31±6 120±16

JNK2+/+BMChow, no

PVAT 31±7 42±7

116±10

JNK2+/+BMWD, no

PVAT 30±6 35±15 143±7

JNK2+/+BMChow +

JNK2+/+ ChowPVAT 44±6 43±3 138±8

JNK2+/+BMChow +

JNK2+/+ WDPVAT 61±12 34±1 122±8

JNK2+/+BMWD +

JNK2+/+ ChowPVAT 39±8 30±7

130±2

JNK2+/+BMWD +

JNK2+/+ WDPVAT 50±12 35±10 129±9

p-value NS NS NS

168