Animal protocols were approved by the University of Virginia Animal Care and Use Committee. NFAT5^+/−ApoE^−/− mice were generated through breeding ApoE^−/− (The Jackson Laboratory) and NFAT5^+/− mice (gift from H. M. Kwon; Go et al., 2004). HFD studies: 22 male littermates were placed on a HFD (by weight 21% fat, 0.15% cholesterol, 19.5% casein, Harlan Labs) for 20 weeks. BMT studies: 27 male NFAT5^+/+ApoE^−/− mice underwent irradiation (600 RAD, twice). BM extracted from the femurs and tibias of three male NFAT5^+/+ApoE^−/− donor mice and three male NFAT5^+/−ApoE^−/− donor mice were spun down, resuspended in RPMI media + 5% FBS, and 3 × 10⁶ cells were injected via the tail vein into two groups of recipient NFAT5^+/+ApoE^−/− mice. For proof of radiation efficacy, 2 NFAT5^+/+ApoE^−/− recipients did not receive bone marrow and subsequently died. Following a 4-week recovery, recipient mice were placed on a HFD for 16 weeks.

Thoracic aortas were fixed in 4% paraformaldehyde, cleaned of all adipose and connective tissue, and stained with Sudan IV for 5 min. The tissues were then longitudinally cut, pinned flat, and imaged. Two different cameras had to be used to image aortas for the first and second in vivo studies, which accounts for the slight difference in images observed from one study to the next. Because different cameras were used, we did not cross-analyze quantitative en face data between the two studies. ImageJ software was used for staining quantification.

Aortic arches were fixed with 4% paraformaldehyde and embedded in paraffin. Five micrometer sections were prepared for staining with primary antibodies (mouse anti-SMαA [Santa Cruz], mouse anti-MAC2 [Cedarlane]) or hematoxylin and eosin (H&E). Sections cut at the center of the aortic arch (where most lesions develop in the lesser curvature, around 200 um into the tissue) were used for immunohistochemical staining and analysis. In brief, after microwave Antigen Retrieval (Vector Laboratories; with the exception of MAC2), primary antibodies were detected using a Vectastain Elite Kit (Vector Laboratories) and DAB (Dako Corp) for visualization. Harris Hematoxylin (Richard-Allen Scientific) served as a counterstain. Negative controls were run with the omission of the primary antibody.

Blood samples were collected for all mice at the conclusion of HFD feeding. Samples were incubated with the Infinity Cholesterol Reagent (Thermo Scientific) for 5 min at 37°C, and absorbance was measured at 500 nm. All unknown samples were calibrated to a sample of known concentration.

Image-Pro Plus 7.0 software was used to quantify SMαA and MAC2 staining in the lesion. ImageJ software was used to determine medial area (=external elastic lamina area-internal elastic lamina [IEL] area) and lesion area (=IEL area-luminal area) from H&E-stained aortic arch sections.

BM was harvested from the femurs and tibias of 11–13 week-old WT and NFAT5^+/− mice. Bones were kept sterile, cut at the ends, and placed into a needle-punctured 500 μl centrifuge tube set within a larger 1500 μl tube. Following high-speed centrifugation, extracted BM cells were treated with 0.83% ammonium chloride to lyse red blood cells. The remaining BM cells were plated onto 100 mm dishes in RPMI Media 1640 (containing 1% antibiotic/antimycotic, 20 mM Hepes, & 10% FBS) plus the addition of macrophage colony stimulating factor (M-CSF, Life, 30 ng/mL). Cells were kept in fresh media containing M-CSF for a period of 7 days to differentiate the cells into BMDMs. Adhesion assay: Human umbilical vein endothelial cells (HUVECs) were plated in a 12-well and stimulated with vehicle (diH₂O, 10 ng/mL) or TNFα (Millipore, 10 ng/mL) in M199 Media for 4 h. BMDMs pre-incubated with Calcein AM (Life, 5 μg) were added to wells containing stimulated HUVECs for a 30-min incubation at 37°. HUVECs were washed thoroughly with PBS and fluorescence was quantified in a FLUOstar plate reader to identify the amount of adherent BMDMs. Boyden chamber migration assay: BMDMs in RPMI Media 1640 were plated in the wells of a Costar Transwell (Fisher) and given 24 h to adhere to the membrane. Cells were washed, and RPMI Media 1640 containing M-CSF (Life, 30 ng/mL) or vehicle (diH₂O, 30 ng/mL) was added to the bottom of the Transwell. Media without M-CSF was added to the top. Cells were incubated for 24 h, washed, and fixed in 4% paraformaldehyde. Cells on the top of the membrane were scraped off with a sterile swab, and migrated cells on the bottom of the membrane were stained with crystal violet, imaged, and quantified using ImageJ software. Control wells were not scraped to verify that BMDMs did originally adhere to the membrane. BrdU proliferation assay: BMDMs in RPMI Media 1640 were plated in a 24-well and allowed 24 h to adhere. Cells were washed and RPMI Media 1640 containing M-CSF (Life, 30 ng/mL) or vehicle (diH₂O, 30 ng/mL) and BrdU (Roche, 10 μm) was added for 24 h. Cells were then washed, fixed in 4% paraformaldehyde, incubated with an anti-BrdU-POD antibody + TMB substrate (Pierce), and absorbance was measured on a FLUOstar plate reader to quantify cellular proliferation. Engulfment assay: BMDMs in RPMI Media 1640 were plated in a 24-well and allowed 24 h to adhere. RPMI Media 1640 containing FluoSphere Carboxylate Microspheres (Life, 4.5 × 10⁶ beads per well) was added to wells containing BMDMs for 30 min. BMDMs were thoroughly washed with PBS and fluorescence was quantified to test for engulfment of FluoSpheres. Wells were imaged under a light microscope to verify that no FluoSpheres had simply adhered to the culture dish.

Statistical significance was confirmed through a t-test or One-Way ANOVA (p < 0.05).

Due to the perinatal lethality of our NFAT5^−/− mice, NFAT5^+/−ApoE^−/− mice were generated for atherosclerosis studies. Aortic arches were harvested for immunohistochemical analysis, and descending thoracic aortas were stained with Sudan IV and laid en face for lesion identification (Figure 1). En face analysis identified that NFAT5^+/−ApoE^−/− thoracic aortas displayed a 73% reduction in atherosclerotic lesion coverage compared to NFAT5^+/+ApoE^−/− controls (Figure 2A). Similarly, H&E staining of lesions in the lesser curvature of the aortic arch identified a 43% reduction in lesion/media area in NFAT5^+/−ApoE^−/− aortas (Figure 2B). Following 20 weeks of HFD feeding, no difference was recorded in total cholesterol or body weight between NFAT5^+/−ApoE^−/− and control NFAT5^+/+ApoE^−/− mice (Figure 2C).

To examine whether NFAT5 levels had an effect on lesion composition or cellular content, we performed immunohistochemistry to quantify SMC and macrophage investment of aortic arch lesions (Figure 3). Smooth muscle alpha actin (SMαA) and macrophage 2 (MAC2) protein identified SMCs and macrophages, respectively, that had migrated into the lesion. With this method of analysis, we were unable to record a significant difference in SMC or macrophage investment of the lesion (Figure 3). Further cell-specific NFAT5 haploinsufficient in vivo experiments were, therefore, required to determine how NFAT5 deficiency in the cells of the lesion affected atherosclerosis progression.

We sought to establish if NFAT5 levels in BM-derived cells alone could account for the reduction in atherosclerosis observed in NFAT5 haploinsufficient mice. To test this hypothesis, we performed a BMT of NFAT5^+/−ApoE^−/− BM into NFAT5^+/+ApoE^−/− mice. Following a 4-week recovery and 16-week HFD, mouse aortas were prepared for analysis (Figure 1). Sudan IV staining and en face preparation of the thoracic aorta identified an 86% reduction in atherosclerotic lesion coverage in mice transplanted with NFAT5^+/−ApoE^−/− BM (Figure 4A). Unlike our genome-wide NFAT5 haploinsufficient atherosclerosis studies, mice transplanted with NFAT5^+/−ApoE^−/− BM developed little to no lesions in the aortic arch and therefore could not be used for immunohistochemistry-based lesion composition analysis. After 16 weeks of HFD feeding, no difference was recorded in total cholesterol or body weight between both BMT groups (Figure 4B).

Results from our BMT studies demonstrate that NFAT5 expression in BM-derived cells is required for atherosclerosis development in the mouse aorta (Figure 4A). Although the BM is comprised of many different cell types (i.e., macrophages, T-cells, B-cells, dendritic cells, mast cells, and neutrophils), these cells have varying degrees of involvement in atherosclerosis (Hansson and Hermansson, 2011). Macrophages play the most dominant role in enhancing lesion size, instability, and vascular inflammation (Moore and Tabas, 2011), which led us to question if NFAT5 expression in macrophages was required for their normal function in atherosclerosis. These functions are: (1) adhesion to ECs in atherosclerosis-prone regions of the vasculature, (2) chemotactic migration toward cytokines produced in the lesion, (3) cytokine-stimulated proliferation, and (4) engulfment of particles. BM harvested from WT and NFAT5^+/− mice was differentiated into BMDMs for in vitro analysis. Our results show that NFAT5 is not required for macrophage adhesion (Figure 5A), proliferation (Figure 5C), or engulfment (Figure 5D), but NFAT5 is required for macrophage migration (Figure 5B). NFAT5^+/− BMDMs exhibited significantly decreased chemotactic migration in response to M-CSF stimulation compared to control WT BMDMs (Figure 5B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.