Loss of Hepatic Carcinoembryonic Antigen-Related Cell Adhesion Molecule 1 Links Nonalcoholic Steatohepatitis to Atherosclerosis

Generation of Liver-Specific AlbCre⁺Cc1^fl/fl Null Mice Propagated on an Ldlr^−/− Background

C57BL/6J.AlbCre⁺Cc1^fl/fl mice were generated, as described,⁽¹⁷⁾ and backcrossed >6 times with Ldlr^−/−;B6.129S7-Ldlrtm1Her/J (JAX#002207; Jackson Laboratory, Bar Harbor, ME). Offspring were genotyped by polymerase chain reaction (PCR) analysis of ear DNA, using Ceacam1-specific primers (Supporting Fig. S1). We used homozygotes with the wild-type Ceacam1 allele with AlbCre (Ldlr^−/−AlbCre⁺Cc1^+/+ or Alb⁺Cc1^+/+) or without AlbCre (Ldlr^−/−AlbCre^–Cc1^+/+ or Alb^–Cc1^+/+) and homozygotes with the Ceacam1-floxed allele with AlbCre (Ldlr^−/−AlbCre⁺Cc1^fl/fl or Alb⁺Cc1^fl/fl) or without AlbCre (Ldlr^−/−AlbCre^–Cc1^fl/fl or Alb^–Cc1^fl/fl); all homozygotes from the same breeding were used to mitigate potential confounding effects of floxing and introducing AlbCre.

Mice Maintenance

All animals were housed in a 12-hour dark–light cycle at the Division of Laboratory Animal Resources at each institution. Starting at 4 months of age, male mice were fed a high-cholesterol (HC) atherogenic diet ad libitum (Harlan Teklad, TD.88137; Harlan, Haslett, MI) containing 0.2% total cholesterol (42% kcal from fat; 42.7% kcal from carbohydrate [high sucrose 34% by weight]) for 2 months (unless otherwise noted). All procedures were approved by the institutional animal care and use committees at each institution.

Metabolic Phenotyping

Body composition was assessed by nuclear magnetic resonance (Bruker Minispec, Billerica, MA), as described.⁽¹⁹⁾ For insulin or glucose tolerance, awake mice were fasted for 6 hours and injected intraperitoneally with either insulin (0.75 units/kg body weight [BW] human regular insulin [Novo Nordisk, Princeton, NJ] or 1.5 g/kg BW dextrose solution). Blood glucose was measured from the tail at 0-180 minutes. Retro-orbital venous blood was drawn at 1100 hours from overnight-fasted mice into heparinized microhematocrit capillary tubes (Fisherbrand, Waltham, MA). Plasma and tissue biochemistry parameters were assessed as detailed in the Supporting Information.

Enzymatic Activity Assays in Liver

Liver homogenates were centrifuged and fatty acid synthase (FASN) activity was assayed in a reaction mix containing 0.1 μCi [¹⁴C]malonyl-coenzyme A (CoA; Perkin-Elmer) and 25 nmol malonyl-CoA in the absence or presence of 500 μM nicotinamide adenine dinucleotide phosphate, reduced form (NADPH; Sigma-Aldrich, MO), as originally described.⁽²⁴⁾ Activity was calculated as counts per minute of [¹⁴C]-incorporated Bq/μg cell lysates. β-Hydroxy β-methylglutaryl (HMG)-CoA reductase activity was carried out following the manufacturer’s instructions (HMG-CoA Reductase Activity Colorimetric Assay Kit, ab204701; Abcam, Cambridge, MA).

Ex-Vivo Palmitate Oxidation

This assay was carried out in the presence of [^1-14C]palmitate (0.5 mCi/mL; American Radiolabeled Chemicals Inc., St Louis, MO)-2 mM adenosine triphosphate, terminated with perchloric acid to recover trapped CO₂ radioactivity. The partial oxidation products were then measured by liquid scintillation using CytoCint (MP Biomedicals, Solon, OH). The oxidation rate was expressed as the sum of total and partial fatty acid oxidation (nmoles/g/minute).⁽¹⁸⁾

Liver Histology

Formalin-fixed paraffin-embedded sections were stained with hematoxylin and eosin (H&E). Sections were deparaffinized and rehydrated before being stained with 0.1% sirius red stain (Direct Red80; Sigma-Aldrich), as described.⁽²³⁾

En-Face and Lesion Analysis

Aortae were dissected from the root to the abdominal area and then formalin fixed. Connective tissues were removed from the longitudinally opened aortae, stained with Oil-Red-O (ORO) (#O0625; Sigma-Aldrich), fixed on a coverslip, and photographed with an Olympus CKX41 (Tokyo, Japan). The total surface and ORO-positive areas were determined using CellSens Standard software in area of pixel². The extent of atherosclerotic lesions was defined as the percentage of total ORO-positive lesion area/total surface area.

Aortic Root Sectioning and Plaque Analysis

Hearts were perfused through the left ventricle with 1X phosphate-buffered saline (PBS; Thermo Scientific, Waltham, MA), followed by 4% paraformaldehyde (Sigma-Aldrich) and cut and embedded in optimal cutting temperature (OCT) compound (Tissue-Tek 4583). This was followed by frozen sectioning on a microtome-cryostat (10 µm) starting from where the aorta exits the ventricle and moving toward the aortic sinus. Sections were stained with ORO or trichrome (Gomori's Trichrome Stain Kit, 87020; Thermo Scientific) or immunostained with monocyte and macrophage 2 (MOMA-2) antibody (1:500; Abcam, Cambridge, United Kindom) overnight at 4°C before incubating with secondary antibody for 1 hour (immunohistochemistry-labeled streptavidin biotin kit; Dako). Images were taken at 4× using an Olympus SZX7-TR30 microscope and quantified with the CellSens Standard program.

Intravital Microscopy of Leukocyte Adhesion on the Carotid Artery

Mice were anesthetized with ketamine and xylazine (100/10 mg/kg) and fixed on a 15-cm cell-culture lid in a supine position. The right jugular vein and left carotid artery were exposed through a middle incision, as described.⁽²⁵⁾ We injected 100 µL of 0.5 mg/mL rhodamine 6G (R4127; Sigma-Aldrich) through a jugular vein puncture to label cells having mitochondria, including leukocytes. The carotid artery was carefully isolated from the surrounding tissue, and one piece of small, U-shaped, black plastic was placed under the vessel to block background fluorescence. The carotid artery (~4-5 mm length) was observed in real time using an intravital microscope (Leica DM6 FS), and video images were captured with a 14-bit RetigaR1 charge-coupled device color digital camera (Teledyne QImaging, Surrey, Canada) and STP7-S-STDT Streampix7 software (Norpix, Montreal, Canada). Video images were analyzed offline for leukocyte adhesion. Cells that adhered to the vessel wall without rolling or moving for at least 3 seconds were counted over the vessel observed. Total numbers were used for statistical analysis.

Isolation of Peritoneal Macrophages

Mice were injected intraperitoneally with 1 mL thioglycollate (T9032; Sigma-Aldrich). Four days later, 6-8 mL of cold 1X PBS was injected into the peritoneal cavity; the peritoneal liquid was aspirated into conical centrifuge tubes and centrifuged to collect and culture the pellet in Roswell Park Memorial Institute (RPMI) medium (Thermo Scientific). Cells were treated overnight with 100 µg/mL of native low-density lipoprotein (LDL) (human plasma, 99% #J65039; Alfa Aesar, United Kingdom) or oxidized LDL (ox-LDL) (human plasma, Hi-TBAR #J65261; Alfa Aesar), fixed with 10% formalin for 10 minutes, and rinsed in PBS once (1 minute) and then with 60% isopropanol for 15 seconds. Cells were stained with filtered ORO (#O0625; Sigma-Aldrich) at 37°C for 1 minute in the dark, washed with 60% isopropanol for 15 seconds, then rinsed with PBS 3 times for 3 minutes each. Images were taken at 20× magnification.

Bone Marrow Isolation From Tibia and Femur

Mice were euthanized and their entire leg dissected. Skin was peeled off, and the tibia and femur were separated and placed in RPMI. The ends of both tibias and femurs were cut to flush out the bone marrow by using a syringe. Cells were collected after centrifugation and cultured in RPMI supplemented with 10-20 ng/mL recombinant macrophage colony-stimulating factor (Thermo Scientific).

Total RNA was isolated from cells using NucleoSpin RNA (740955.50; Macherey-Nagel, Bethlehem, PA). Complementary DNA (cDNA) was synthesized with the iScript cDNA Synthesis Kit (Bio-Rad), using 1 μg of total RNA and oligo deoxythymine primers. cDNA for total and Ceacam1-long (L) and short (S) isoforms were evaluated by quantitative reverse-transcription (qRT)-PCR (StepOne Plus; Applied Biosystems, Foster City, CA) and normalized against 18S, using primers listed in Supporting Table S1.

Western and qRT-PCR Analyses

Western blots and qRT-PCR analyses were carried out as routinely done and as detailed in the Supporting Information.

Statistical Analysis

Data were analyzed using one-way analysis of variance (ANOVA) with Tukey’s test for multiple comparisons, using GraphPad Prism6 software. Data were presented as mean ± SEM. P < 0.05 was considered statistically significant.

Ldlr^−/−AlbCre⁺Cc1^fl/fl Mice Manifested Insulin Resistance

Starting at 5 weeks of HC intake (Table 1; Supporting Fig. S3), Ldlr^−/−AlbCre⁺Cc1^fl/fl mice manifested a higher body weight gain and exhibited an increase in fat and visceral mass with a reciprocal decrease in lean mass relative to the three control groups after 2 months of HC (Table 1). They maintained fasting hyperinsulinemia relative to their littermate controls (Table 1), with impaired insulin clearance, which was measured by steady-state C-peptide/insulin molar ratio (Table 1). Null mice exhibited higher postprandial blood glucose levels (Table 1) and intolerance to exogenous insulin and glucose relative to controls (Fig. 1A,B). Consistent with hyperinsulinemia repressing insulin receptor (IR) expression,⁽²⁶⁾ immunoblotting with α-IR_β and normalizing against α-tubulin showed a ~50% reduction in IR level in liver and WAT of mutants (Fig. 1Cb,c). Insulin release (Fig. 1Ca) in control but not null mice refed (RF) for 7 hours following an overnight fast induced IR_β phosphorylation relative to fasted (F) mice (Fig. 1Cb,c). Together, this demonstrated systemic insulin resistance with elevated fasting hyperglycemia in Ldlr^−/−AlbCre⁺Cc1^fl/fl mice (Table 1).

Ldlr^−/−AlbCre⁺Cc1^fl/fl Mice Exhibited Altered Hepatic Lipid Metabolism With Hypertriglyceridemia and Hypercholesterolemia

Histologic evaluation of H&E-stained liver sections revealed diffused macrovesicular fat infiltration in Ldlr^−/−AlbCre⁺Cc1^fl/flmice compared to controls, which displayed predominantly microsteatosis (Fig. 2A), in agreement with their higher hepatic triacylglycerol content (Table 1). This likely resulted from increased lipid transport into hepatocytes (higher messenger RNA [mRNA] levels of cluster of differentiation [Cd]36, fatty acid transport protein 1 [Fatp1], and Fatp4 [Supporting Table S2]) and from de novo lipogenesis (higher mRNA [Fig. 3Aa; Supporting Table S2] and protein [Fig. 3Ab] levels of FASN stemming from hyperinsulinemia-driven transcriptional activation of sterol regulatory element binding protein 1c [Srebp1c] activity in null livers [Supporting Table S2]). Moreover, insulin release in RF (Fig. 1Ca) caused CEACAM1 phosphorylation (pCC1) and association with FASN in control but not in null livers, as shown by immunodetection of pCC1 in the FASN immunopellet (Fig. 3Ab). As shown by Najjar et al.,⁽²⁴⁾ this lowers hepatic FASN enzymatic activity in controls (Fig. 3Ac), whereas in null mice, de novo lipogenesis increased followed by redistribution of substrates to WAT to cause hypertriglyceridemia, visceral adiposity, and lipolysis (increased plasma NEFA levels) (Table 1). Compromised fatty acid β-oxidation contributed to fat accumulation in Ldlr^−/−AlbCre⁺Cc1^fl/fl livers (Fig. 3B).

Ldlr^−/−AlbCre⁺Cc1^fl/fl mice exhibited higher levels of hepatic and plasma total and free cholesterol. Plasma LDL cholesterol (LDL-C) and very low-density lipoprotein cholesterol (VLDL-C) were increased with a reciprocal decrease in high-density lipoprotein cholesterol (HDL-C) content in null mice (Table 1). This is consistent with increased hepatic cholesterol production in liver-specific inactivation of CEACAM1 (L-SACC1) mice with liver-specific CEACAM1 inactivation.⁽²⁷⁾ Accordingly, hepatic Srebp2 mRNA levels (Supporting Table S2), a master regulator of genes involved in cholesterol synthesis, including HMG-CoA reductase (Hmgcr) (Supporting Table S2), were elevated in parallel to its increased enzymatic activity (Fig. 3C) and an increase in plasma apolipoprotein B (ApoB) levels in mutants (Table 1). The increase in hepatic Srebp1c and Srebp2 mRNA was not associated with changes in hepatic insulin-induced gene 1 (Insig-1) and Insig-2a/b mRNA levels (Supporting Table S2). Whether the regulatory effect of these endoplasmic reticulum membrane proteins on cholesterol synthesis⁽²⁸⁾ were altered by liver-specific Ceacam1 deletion remains to be examined. Nevertheless, it is possible that, like Srebp1c, increased transcriptional activity of Srebp2 in Ldlr^−/−AlbCre⁺Cc1^fl/fl mice was driven by chronic hyperinsulinemia.⁽²⁹⁾

Elevated LDL-C and non-HDL-C in Ldlr^−/−AlbCre⁺Cc1^fl/fl mice could also be caused by reduced clearance, as supported by elevated plasma levels of ApoB and proprotein convertase subtilisin/kexin type 9 (PCSK9) (Table 1). Also supporting this finding is reduced expression of LDL receptor-related protein 1 (Supporting Table S2), which is involved in hepatocytic clearance of chylomicron remnants. Whether CEACAM1 modulates LDL clearance is unclear, but mRNA levels of hepatic Niemann-Pick type C1 protein (Npc1), a late endosomal protein involved in exporting LDL-C to other cellular compartments, was reduced by ~3-fold in nulls (Supporting Table S2). This suggests partitioning of free cholesterol to mitochondria where it could contribute to the ~3-fold to 4-fold reduction in plasma glutathione (GSH) in nulls (Fig. 4Aa).⁽³⁰⁾ Together, this demonstrates that Ldlr^−/−AlbCre⁺Cc1^fl/fl mice developed the dyslipidemia profile of NAFLD and atherosclerosis.^{(8, 31)}

Ldlr^−/−AlbCre⁺Cc1^fl/fl Mice Exhibited Inflammation in Liver

H&E staining revealed few inflammatory islands with no change in the hepatocellular architecture of control livers (Fig. 2A). In contrast, Ldlr^−/−AlbCre⁺Cc1^fl/fl mice exhibited multiple and extended foci of mononuclear inflammatory infiltrates in hepatic lobules and periportal areas (Fig. 2A). qRT-PCR analysis (Supporting Table S2) showed activated macrophages (F4/80; Cd68) and T cells (Cd4/Cd8) without significance changes in anti-inflammatory regulatory T-cell (Treg) pools, such as forkhead box protein P3 (FoxP3) and interleukin (Il)-10 (Supporting Table S2). Elevated hepatic mRNA of interferon gamma (Ifnγ) but not Il-4/Il-13 suggests a CD4⁺T helper 1 (Th1) but not CD4⁺Th2 response in mutant livers.

Consistent with nuclear factor kappa B (NF-κB) activation (phosphorylation) in nulls (Fig. 2C), hepatic mRNA of its targets tumor necrosis factor alpha (Tnfα) and Il-6 (Supporting Table S2) and their plasma levels (Figs. 4Ab,c) were elevated in nulls.

Hepatic mRNA (Supporting Table S2) and plasma levels (Fig. 4Ac) of IL-6 were higher in nulls. This could activate (phosphorylate) signal transducer and activator of transcription (STAT3) (Fig. 2C) to induce monocyte chemoattractant protein 1 (Mcp-1)/chemokine (C-C motif) ligand 2 (Ccl2) in monocytes/macrophages and their migration (Supporting Table S2) in addition to repressing IFN regulatory factor 8 (Irf-8) (Supporting Table S2), which could in turn induce Cd11b macrophage expression (Supporting Table S2).

WAT of null mice exhibited more macrophage recruitment, as demonstrated by higher mRNA levels of F4/80, Cd68, Tnfα, Il-1β, and Il-6 (Supporting Table S3). This could significantly contribute to elevated plasma IL-6 and TNFα levels and to systemic proinflammatory and insulin resistance in Ldlr^−/−AlbCre⁺Cc1^fl/fl mice.⁽³²⁾

Ldlr^−/−AlbCre⁺Cc1^fl/fl Mice Exhibited Oxidative Stress and Apoptosis in Liver

Hepatic mRNA levels of oxidative stress markers (NADPH oxidase 1 [Nox1], Nox4, glycosylated 91-kDa glycoprotein [Gp91]) were elevated by ~2-fold to 3-fold in nulls (Supporting Table S2). Consistently, plasma NO level was lower (Fig. 4Ad) and NAD/NADH was higher (Fig. 4Ae) in nulls. Together with increased plasma 8-isoprostane (Fig. 4Af), this demonstrated oxidative stress and lipid peroxidation in nulls.

Consistent with oxidative stress causing apoptosis in the presence of high TNFα levels,⁽¹⁰⁾ CCAAT/enhancer binding protein homologous protein (CHOP) levels were elevated in null livers (Fig. 2C). This was supported by a ~2-fold increase in mRNA levels of markers of hepatocyte injury (Supporting Table S2), without noticeable ballooning (Fig. 2A). Consistently, Ldlr^−/−AlbCre⁺Cc1^fl/fl mice developed liver dysfunction (elevated plasma alanine aminotransferase [ALT] and aspartate aminotransferase [AST] activities) (Table 1).

Ldlr^−/−AlbCre⁺Cc1^fl/fl Mice Exhibited Hepatic Fibrosis

Sirius red staining (Fig. 2B) revealed a perivenular and/or pericellular bridging “chicken-wire” pattern of collagen deposition in mutant livers, despite elevated levels of the proinflammatory antifibrotic IFNγ cytokine. This was supported by higher protein levels of alpha smooth muscle actin (α-SMA) (Fig. 2C) and mRNA levels of profibrogenic genes (α-Sma, collagen type VI alpha 3 chain [Col6α3], and transforming growth factor beta [Tgfβ]) together with activation of the TGFβ-mothers against decapentaplegic homolog 3 (Smad3) pathway (Fig. 2C) that could result from TNFα suppression of Smad7, an inhibitor of TGFβ activation, in null livers (Supporting Table S2).

Fibrosis could stem from increased plasma and hepatic levels of ET-1 (Fig. 4Ag) that could activate the profibrogenic ET-1 receptor A (Etar) with ~15-fold higher expression than receptor B (Etbr) (Fig. 4Ba). Elevated plasma ET-1 levels likely resulted from lower plasma prostaglandin E2 (PGE2)⁽³³⁾ (Fig. 4Ah) and hyperinsulinemia.⁽²⁰⁾

Development of Atherogenic Lesions in Ldlr^−/−Cre⁺Cc1^fl/fl Aortae

En-face analysis revealed a 2-fold to 3-fold increase in the plaque area of mutants fed HC for 2 months (Fig. 5A) but not 1 month (not shown). ORO staining of aortic root cross-sections showed lipid deposition in mutant intima (Fig. 5Ba and graph). Consistently, mRNA levels of genes involved in fat (Cd36, Srepb-1c, Fasn) and cholesterol (Srebp2, Hmgcr, and Npc1) transport and synthesis were higher in null aortae (Supporting Table S4).

Immunostaining of aortic root cross-sections with α-MOMA-2 showed inflammatory infiltration in Ldlr^−/−AlbCre⁺Cc1^fl/fl aortic walls (Fig. 5Bb and graph). Intravital microscopy demonstrated ~5-fold increase in the adhesiveness of circulating leukocytes to the carotid artery endothelium of mutants (Fig. 6A and graph). Accordingly, mRNA levels of vascular cell adhesion protein 1 (Vcam-1), intercellular adhesion molecule 1 (Icam-1), and P-selectin, which play an important role in attracting leukocytes to endothelium, were higher in Ldlr^−/−AlbCre⁺Cc1^fl/fl aortae (Supporting Table S4).

mRNA levels of cadherin 5, type 2 (VE-cadherin) and β-catenin, which preserve endothelial integrity and regulate leukocyte adhesion, were reduced by ~2-fold to 3-fold in Ldlr^−/−AlbCre⁺Cc1^fl/fl aortae (Supporting Table S4). Contributing to junctional permeability and transendothelial migration of immune cells in Ldlr^−/−AlbCre⁺Cc1^fl/fl aortae is the 2-fold to 3-fold decrease in mRNA levels of vascular endothelial growth factor A (Vegf-A) and its receptor (Vegfr-2), Vegfr-1, and angiopoietins (Ang-1 and Ang-2) in mutant aortae (Supporting Table S4). Intact Ceacam1 mRNA (total and L and S isoforms) in aortae (Fig. 6C) and macrophages (Fig. 6D) ruled out changes in Ceacam1 expression and its potential effect on the transendothelial migration of leukocytes in nulls.

Consistent with monocyte infiltration through permeable endothelium, the pool of activated macrophages (F4/80, Cd68, Il-1β, TNFα, toll-like receptor (Tlr)2/4) (Supporting Table S4) increased in nulls. Accordingly, NF-κB phosphorylation was elevated (Fig. 6B) to induce TNFα and IL-6 transcription in Ldlr^−/−AlbCre⁺Cc1^fl/fl (Supporting Table S4) macrophages in response to free-cholesterol accumulation,⁽³⁴⁾ as suggested by lower Npc1 mRNA levels.

Elevated IL-6 could activate (phosphorylate) STAT3 (Fig. 6B) to induce Mcp-1/Ccl2 (Supporting Table S4), which mediates transendothelial migration of monocytes/macrophages and represses Irf-8 (Supporting Table S4) to increase expression of CD11b, a receptor for ICAM-1 involved in leukocyte adhesion, in mutant aortae (Supporting Table S4).

Additionally, mRNA of markers of activated T cells (Cd4, Cd8) but not the atheroprotective Treg (FoxP3 and Il-10) were elevated in null aortae (Supporting Table S4). This induced an atherogenic CD4⁺Th1 response rather than the atheroprotective CD4⁺Th2 type, as suggested by high Ifnγ but not Il-4/Il-13 mRNA levels in mutant aortae (Supporting Table S4).

ORO staining of peritoneal macrophages treated with native or ox-LDL indicated a shared susceptibility to Ox-LDL for lipid-loaded macrophage formation by all mouse groups (Fig. 6E). This ruled out differential macrophage preactivation (priming), assigning a major role for systemic factors in the development of atherosclerosis in Ldlr^−/−AlbCre⁺Cc1^fl/fl mice.

Ldlr^−/−Cre⁺Cc1^fl/fl Aortae Exhibited Fibrosis

Trichrome staining of aortic root sections showed ~5-fold to 6-fold increase in collagen deposition in mutant mice (Fig. 5Bc and graph). This was supported by higher mRNA levels of profibrogenic genes (connective tissue growth factor [Ctgf], Col6α3, α-Sma, fibronectin, Tgfβ) (Supporting Table S4) and by the activation of the TGFβ-Smad3 pathway in mutant aortae (Fig. 6B), resulting, in part, from lower Smad7 mRNA (Supporting Table S4). Fibrosis in Ldlr^−/−AlbCre⁺Cc1^fl/fl aortae could be mediated by the higher profibrogenic activity of plasma and aortic ET-1, as supported by the ~115-fold rise in the Etar/Etbr ratio (Fig. 4Bb).

Mechanisms Underlying Altered Vascular Morphogenesis in Ldlr^−/−Cre⁺Cc1^fl/fl Mice

Consistent with lower plasma NO levels (Fig. 4Ad), aortic NO (Fig. 4C) and mRNA levels of Gp91, Nox1, and Nox4 were lower in mutants (Supporting Table S4). Because hyperinsulinemia suppresses NO and induces ET-1,⁽²⁰⁾ we examined whether insulin signaling was altered in mutant aortae. Immunoblotting with α-IR_βrevealed a ~50% decrease in IR protein content in mutants (Fig. 7A), likely driven by hyperinsulinemia.⁽²⁶⁾ Immunoblotting with α-IR_β (Fig. 7A) demonstrated blunted insulin-stimulated IR_β phosphorylation in mutants (Fig. 7Ba; RF). Consistently, insulin failed to induce the phosphorylation of downstream signaling molecules (insulin receptor substrate 1 [IRS1], Akt, and endothelial nitric oxide synthase [eNOS]) in mutant aortae (Fig. 7B-D). Dysregulated insulin-dependent eNOS activation yielded compromised NO production in Ldlr^−/−AlbCre⁺Cc1^fl/fl aortae.

Restricted IR_β phosphorylation failed to phosphorylate CEACAM1 (Fig. 7E; immunoprecipitation with α-pCC1), thus diminishing Shc sequestration and its rerouting to IR_β, as shown by the higher IR_β level in the Shc immunopellet (Fig. 7E, middle gel). This was facilitated by the reciprocal decrease in IRS1 binding to IR_β (Fig. 7B) because Shc and IRS1 compete for binding to the Asn-Pro-Glu-Tyr960 sequence in IR_β.⁽¹⁶⁾ Higher Shc/IR_β association amplified insulin-stimulated mitogen-activated protein kinase (MAPK) phosphorylation to a larger extent in mutant aortae (Fig. 7F), inducing ET-1 (Et-1)⁽²⁰⁾ by 4-fold to 5-fold in mutant aortae (Supporting Table S4).