Targeted deletions of cyclooxygenase-2 and atherogenesis in mice

Abstract

Background: Although the dominant product of vascular Cyclooxygenase-2 (COX-2), prostacyclin (PGI(2)), restrains atherogenesis, inhibition and deletion of COX-2 have yielded conflicting results in mouse models of atherosclerosis. Floxed mice were used to parse distinct cellular contributions of COX-2 in macrophages and T cells (TCs) to atherogenesis.

Methods and results: Deletion of macrophage-COX-2 (Mac-COX-2KOs) was attained with LysMCre mice and completely suppressed lipopolysaccharide-stimulated macrophage prostaglandin (PG) formation and lipopolysaccharide-evoked systemic PG biosynthesis by approximately 30%. Lipopolysaccharide-stimulated COX-2 expression was suppressed in polymorphonuclear leukocytes isolated from MacKOs, but PG formation was not even detected in polymorphonuclear leukocyte supernatants from control mice. Atherogenesis was attenuated when MacKOs were crossed into hyperlipidemic low-density lipoprotein receptor knockouts. Deletion of Mac-COX-2 appeared to remove a restraint on COX-2 expression in lesional nonleukocyte (CD45- and CD11b-negative) vascular cells that express vascular cell adhesion molecule and variably alpha-smooth muscle actin and vimentin, portending a shift in PG profile and consequent atheroprotection. Basal expression of COX-2 was minimal in TCs, but use of CD4Cre to generate TC knockouts depressed its modest upregulation by anti-CD3epsilon. However, biosynthesis of PGs, TC composition in lymphatic organs, and atherogenesis in low-density lipoprotein receptor knockouts were unaltered in TC knockouts.

Conclusions: Macrophage-COX-2, primarily a source of thromboxane A(2) and prostaglandin (PG)E(2), promotes atherogenesis and exerts a restraint on enzyme expression by lesional cells suggestive of vascular smooth muscle cells, a prominent source of atheroprotective prostacyclin. TC COX-2 does not detectably influence TC development or function or atherogenesis in mice.

Figure 1. Macrophage-specific COX-2 deletion characterization

A. COX-2 mRNA levels determined by real-time PCR. All samples were normalized with 18s RNA. COX-2 mRNA was significantly reduced in stimulated Mac-COX-2 KO peritoneal macrophages (p=0.001). **, p<0.01; n=5-6. B. COX protein levels in peritoneal macrophages. Samples were collected from a pull of 3 mice in each genotype. Data are presented using medians with first and third quartiles.

Figure 2. Impact of macrophage COX-2 deletion in prostanoid production measured by mass spectrometry

A. Prostanoid profiles in cultured macrophage supernatants. PGE₂, TxA₂, PGI₂ and PGD₂, was determined by quantitation of PGE₂, TxA₂, 6-keto PGF_1α, and PGD₂. All data were normalized with total proteins. PG levels were significantly lower in stimulated Mac-COX-2 KO supernatants (p=0.002 for PGE₂, TxA₂, PGI₂ and PGD₂). **p<0.01; n=5-6. B. Profiles of urinary prostanoid metabolites in LPS endotoxemia. Twenty-four hr urine was collected before and after a bolus of LPS 1mg/kg intraperitoneal injection. Biosynthesis of PGE₂, TxA₂, PGI₂ and PGD₂ was assessed by quantitation of their major urinary metabolites: 7-hydroxy-5,11-diketotetranorprostane-1,16-dioic acid (PGE-M), 2,3 dinot TxB₂ (Tx-M), 2,3 dinor 6 keto PGF1α (PGI-M) and 11,15-dioxo-9α-hydroxy-2,3,4,5-tetranorprostan-1,20-dioic acid (tetranor PGD-M), respectively. All urine data were normalized with creatinine. Analysis of variance revealed a significant reduction in LPS stimulated prostanoid metabolite excretion in the Mac-COX-2 KO mice (p=0.04, p=0.02, p=0.02, and p=0.012 for PGE-M, Tx-M, PGI-M, and PGD-M respectively). *, p<0.05; n=25-27. Significant depression of Tx-M (p=0.038), and PGI-M (p=0.044) were observed in male KOs and of PGE-M (p=0.024), Tx-M (p=0.031), and PGI-M (p=0.039) and PGD-M (p=0.009) were observed in female KOs (^#, p<0.05; ^##, p<0.01; n=14-16). Data are presented using medians with first and third quartiles.

Figure 3. Effects of cell specific COX-2 deletion in atherogenesis

All mice were backcrossed onto a low density lipoprotein receptor deficient (LdlR^-/-) background and fed a high fat diet (HFD) from 8 wks of age. Aortic atherosclerotic lesion formation, represented by the ratio of lesion area to total aortic area, in mice either after 3 or 6 months on a HFD, was analyzed by en face. A. Analysis of variance revealed a significant reduction in lesion area in the Mac-COX-2 KO/ LdlR^-/- mice on 6 months of HFD (p=0.007; n=28-31). **, p<0.01. Subsequent comparison demonstrated same trend in both females (p=0.029, n=14-16) and males (p=0.014, n=12-17). ^#, p<0.05. B. T cell COX-2 deletion failed to impact lesion formation. p=0.23, n=8-12 at 3 months and p=0.15, 16-20 at 6 months. Data are presented using medians with first and third quartiles.

Figure 4. Impact of macrophage COX-2 deletion in vascular smooth muscle COX-2 expression in atherosclerotic lesions by immunohistochemistry

Lesion morphology in aortic roots from mice on HFD for 3 months was analysed. Upper left panel: Representative COX-2 staining in WT/LdlR^-/- (a and b) and Mac-COX-2 KO/ LdlR^-/- (c and d) aortic root sections. Upper right panel: COX-2 (e and f) and CD11b (g and h) staining on Mac-COX-2 KO/LdlR^-/- aortic sections. COX-2 expressing cells are not CD11b+ macrophages. Arrows indicate corresponding areas on serial sections. a and c, and e and g, are magnifications of boxed areas in b and d, and f and h, respectively. Lower left panel: Cells expressing COX-2 in Mac-COX-2 KO/LdlR^-/- aortic sections (i) are negative for or express only low levels of α-smooth muscle actin (SMA) (j) and are not CD45 positive lymphocytes (k). Arrows indicate corresponding areas on serial sections. Lower right panel: Representative COX-2 staining in WT/LdlR^-/- (l) and in Mac-COX-2 KO/LdlR^-/- aortic sections (m). Cells expressing COX-2 in Mac-COX-2 KO/LdlR^-/- aortic sections are also positive for VCAM (n) in neointima. L: lumen, NI: neointima, M: media, A: adventitia. Parentheses show NI.

Figure 5. Impact of macrophage COX-2 deletion on apoptosis in atherosclerotic lesions

Apoptosis was measured by active caspase 3 staining. Mac-COX-2 KO/LdlR^-/- aortic sections demonstrated an increased incidence of apoptotic cells.