L-4F was synthesized similar to the methods previously described 3031. Pravastatin sodium was purchased from LKT Laboratories, Inc. (St. Paul, MN, USA).

ApoE^-/-Fas^-/- B6 mice were originally produced by breeding apoE^-/- and Fas^-/- single knockout mice purchased from the Jackson Laboratories (Bar Harbor, ME, USA) and then further maintained in a colony 16. At eight to nine weeks of age, female apoE^-/-Fas^-/- mice were randomly grouped to receive one of four different treatment regimens: 1) pravastatin (5 mg/kg body weight in drinking water, n = 14), 2) L-4F (15 mg/kg in 50 mM ammonium bicarbonate buffer, pH 7.0, containing 0.1 mg/ml Tween-20 (ABCT) i.p., five days/week, n = 25), 3) L-4F plus pravastatin (administered as described for groups 1 and 2, n = 9), and 4) vehicle control (ABCT i.p., five days/week, n = 23) (Figure 1b). After 27 weeks, mice were fasted overnight and euthanized. At time of death, blood samples were collected via cardiac puncture. The mice were profused using phosphate buffered saline (PBS) (9.5 mM phosphate, pH 7.4, 2.7 mM KCl and 137 mM NaCl) prior to harvest of the spleen, lymph nodes, and kidneys (Figure 1b). All mice were treated in conformity with Public Health Service Policy. The mice were fed normal chow diet and maintained in a temperature-controlled room with a 12-hour light/dark cycle according to the approved protocol by the University of California, Los Angeles Animal Research Committee.

Serum and plasma samples were collected from each mouse at euthanasia. An ELISA kit was used to test relative levels of total IgG antibodies. Serum samples were diluted 1:200 to measure relative levels of IgG anti-dsDNA using a streptavidin-biotin method of ELISA, and an IgG anti-cardiolipin ELISA was used to measure levels of IgG antibodies to oxidized phospholipids (oxPLs) -1-palmitoyl-2(5-oxovaleroyl)-sn-glycero-3-phosphorylcholine (POVPC) and 1-palmitoyl-2-glutaroyl-sn-glycero-3-phosphorylcholine (PGPC) - as previously described 16. A standard curve was generated using serially diluted pooled sera from mice with known high concentrations of the desired antibody. Samples were measured using a goat anti-mouse IgG Fc antibody conjugated with either alkaline phosphatase enzyme or horseradish peroxidase (Bethyl Laboratories, Inc.; Montgomery, TX, USA).

Following euthanasia, one kidney from each mouse was fixed in 10% formalin. The samples were embedded in paraffin, sectioned at 3 μm, and stained using either periodic acid-Schiff (PAS) or hematoxylin and eosin (H&E). Stained sections were photographed electronically with a microscope fitted with a digital camera (Nikon Eclipse 600, Melville, NY, USA), assigned anonymous identification numbers, and analyzed using computer-assisted imaging software (Image ProPlus; Media Cybernetics, Bethesda, MD, USA) by a blinded observer. Twenty-five to thirty glomeruli for each sample were observed in representative fields on duplicate slides and were measured to calculate the mean glomerular tuft size for each mouse.

Morning urine was regularly collected from each mouse throughout the duration of the treatment protocol. Albustix strips (Bayer; Elkhart, IN, USA) were used to estimate proteinuria levels from fresh urine samples. Levels of proteinuria were expressed as follows: 0 = none, 1 = trace, 2 = approximately 30 mg/dl, 3 = approximately 100 mg/dl, 4 = approximately 300 mg/dl, and 5 = >2,000 mg/dl.

Following euthanasia, female mice were subsequently scanned using dual-energy X-ray absorptiometry (DEXA) with a Lunar PIXImus2 Densitometer (GE Medical Systems; Madison, WI, USA). BMD was measured for the whole skeleton excluding the skull, the lumbar spine (L2 to L6), and the femurs. Femoral BMD was calculated by averaging the BMD measurements for both femurs; in cases in which the left femur was used for bone marrow RNA extraction, femoral BMD was based on the BMD of the right femur alone.

L5 vertebrae were extracted from a random sample of mice and fixed in formalin. The vertebrae were packed in 1× PBS for evaluation using three-dimensional microtomography (μCT) (μCT 40, Scanco Medical; Bassordorf, Switzerland) in 12 μm slices at a threshold of 275 nm. Bone volume density, trabecular number, connectivity density, trabecular thickness, and trabecular separation were measured.

The basal portion of the heart and the proximal aorta were harvested to assess atherosclerotic manifestations, embedded in Tissue-Tec OCT medium, frozen in liquid nitrogen, and stored at -80°C. Tissue from the aortic root was selected for evaluation since most studies involving mouse models of atherosclerosis use it as reference tissue for plaque evaluation. Serial 10 μm thick cryosections were stained with Oil Red O and hematoxylin, counterstained with fast green, and analyzed via light microscope for atheromatous lesions 16.

Serial 10 μm thick cryosections of aortic root were individually immunohistochemically stained for either 1) macrophages (rat anti-mouse CD68; Vector Labs, Burlingame, CA, USA), 2) α-actin (alkaline phosphatase-conjugated monoclonal anti-α-smooth muscle actin; Sigma) 32, 3) T-cells (rat anti-mouse CD4; Vector Labs), or 4) VCAM-1 (rat anti-mouse VCAM-1; AbD Serotec; Raleigh, NC, USA). Slides were treated as previously described by Roque et al. using a biotinylated anti-rat IgG secondary antibody and Avidin/Biotinylated Enzyme Complexes (ABC Elite; Vector Labs) and visualized using VECTOR Red (P-nitrophenyl phosphate; VECTOR Red substrate kit; Vector Labs) 32. Negative controls were prepared by omission of the primary antibody.

The slides were analyzed using similar methodology listed under Kidney histology. Images were taken of three to six samples from duplicate slides, which were analyzed by a blinded observer to calculate a mean stained area per lesion area for each mouse. Additional slides were stained for various tissue components (elastic fibers, ground substance, muscle, collagen, and fibrinoid and fibrin) using a Movat pentachrome stain.

Plasma samples collected during euthanasia were analyzed for lipid levels (triglycerides, total cholesterol, HDL cholesterol, non-HDL cholesterol, unesterified cholesterol, and free fatty acids) using enzymatic colorimetric assays as previously described 33. Mouse HDL was isolated from pooled plasma samples before and six hours after injection of L-4F or scrambled L-4F (that is, identical amino acids as contained in L-4F but arranged in a random sequence that markedly reduces lipid binding) using fast-protein liquid chromatography (FPLC) fractionation 34. In order to assess the anti-inflammatory properties of L-4F, 10 mice from both the control group and the L-4F-treated group were randomly selected, totaling 20 mice, and combined to form three pools (with three to four mice per pool) for each group. Mouse LDL was isolated by FPLC from pooled plasma samples from both groups and tested for its ability to induce monocyte chemotactic activity in cultures of human aortic endothelial cells as previously described 34. Plasma samples were pooled for this assay in order to isolate sufficient concentrations of LDL and HDL particles; sample volumes obtained from individual mice did not provide adequate lipoprotein levels to determine monocyte chemotactic activity.

Luminex-based beadarray (RodentMap version 1.6; Rules Based Medicine, Inc., Austin, TX, USA) was used to simultaneously assess for 69 different antigens in plasma samples from 8 to 16 randomly selected mice per group. Fifteen of the 69 assays were not present at detectable levels (calbindin, EGF, endothelin-1, FGF-9, GM-CSF, GST-α, GST-μ, INF-γ, IL-11, IL-12p70, IL-17, IL-2, IL-3, IL-4, and NGAL) (See Supplemental table S1 in Additional file 1 for a complete list of chemokines/cytokines included in the Luminex assay).

Fluorescence-activated cell sorting (FACS) analysis was performed on spleen samples from the four different treatment groups to identify potential changes in immune cell subsets. Multi-color flow cytometry analysis was used to characterize populations of B cells (CD19, T1, T2, FO, MZ, and plasma cells), T cells (CD4 and CD8), and NK, CD11c, and CD11b cells. After standard Fc blocking, the fluorochrome-conjugated anti-mouse antibodies that were used for staining included FITC-, PE-, PerCP-, and APC-conjugated antibodies to CD19 (MB19-1), IgM (II/41), IgD (11-26c 11121314151617181920212223242526), CD21 (eBio8D9 (8D9)), CD23 (B3B4), B220 (RA3-6B2), CD93 (AA4.1), CD62L (MEL-14), CD4 (GK1.5), CD8 (H35-17.2), NK1.1 (PK136), CD11c (N418), Ly6C (HK1.4), CD11b (M1/70) (all eBioscience; San Diego, CA, USA), CD138 (281-2) (BD Biosciences; San Jose, CA, USA). Samples were acquired on a FACSCalibur flow cytometer (BD Biosciences) and analyzed using FloJo software (Tree Star, Ashland, OR, USA).

Data was collected and analyzed using Excel (Microsoft Office) or Prism 3.0 (Graphpad, La Jolla, CA, USA). For comparisons between two groups, unpaired student's t-test was used if the variance was normally distributed; Mann-Whitney U test was used for comparisons with a variance that was not normally distributed. Comparisons made between three or more groups were performed using one-way ANOVA. All results are presented as mean ± SD; P < 0.05 was considered significant. For Luminex-based beadarray of 69 plasma antigens, Bonferroni correction was applied for detectable antigens (n = 54); as a result, P < 0.0009, as calculated by (P < 0.05)/(n = 54) = (P < 0.0009), was considered significant.

In apoE^-/-Fas^-/- B6 mice that develop accelerated atherosclerosis and autoimmunity, we used a dose of Apo-A1 mimetic peptide twice as much as previously used in apoE^-/- B6 mice 2335. To determine an effective form of L-4F peptide, two groups of eight-week old double knockout mice (n = 10 per group) were fasted overnight, bled the following morning (0 h), injected with either 15 mg/kg i.p. L-4F or scrambled L-4F peptide, and harvested for blood samples six hours later. Compared to 0 h time point, two out of three blood sample pools from the L-4F group (three to four mice per pool), but none of the five sample pools from the scrambled L-4F group (two mice per pool), showed significant reduction in monocyte chemotactic activity after six hours (Figure 1a). These data suggest that 15 mg/kg of i.p. L-4F could improve HDL anti-inflammatory activity in young apoE^-/-Fas^-/- mice. Suboptimal dosage of pravastatin was determined as previously described by Navab et al. 23. This suboptimal dose was administered in order to prevent masking potential additive synergistic effects contributed by L-4F.

After 26 to 27 weeks of treatment with 1) pravastatin, 2) L-4F, 3) L-4F plus pravastatin, or 4) vehicle control, mice treated with L-4F or L-4F plus pravastatin showed improved lupus-like autoimmune manifestations compared to vehicle controls.

There was no significant difference in total IgG levels among the four groups, suggestive of no general immune suppression (Figure 2a). Serum levels of IgG anti-dsDNA antibodies and IgG anti-cardiolipin were significantly reduced in mice treated with L-4F (Figure 2b, c). Similarly, mice treated with L-4F, with or without pravastatin, had significantly lower serum levels of IgG autoantibodies to oxPLs--PGPC and POVPC compared to vehicle controls (Figure 2d). Although it appeared that pravastatin caused a mild canceling effect in combination treated mice, there was no significant difference in circulating levels of IgG anti-dsDNA and IgG anti-cardiolipin found between L-4F-treated mice and combination treatment mice.

Significantly smaller lymph nodes were present in both L-4F and L-4F plus pravastatin-treated mice compared to vehicle controls (0.17 ± 0.17 g and 0.16 ± 0.10 g vs. 0.40 ± 0.22 g; P = 0.001 and 0.004, respectively) (Figure 2e). However, upon comparison between treatment groups and vehicle controls, there was no significant difference in spleen size or splenocyte populations of B-cells, CD4+, CD8+ T-cells, NK, CD11c, CD11b cells as determined by multi-color flow analysis (data not shown).

Kidney disease was followed non-invasively via analysis of proteinuria levels during the course of treatment. L-4F treatment was associated with lower proteinuria levels than in vehicle controls starting at Week 20 of the treatment protocol (Figure 3d). Upon histological analysis, the controls also had increased glomerular cell infiltration, analogous to diffuse proliferative glomerulonephritis (DPGN) in SLE patients (Figure 3a) 16. L-4F or L-4F plus pravastatin-treated mice had decreased glomerular tuft size compared to vehicle controls (6,846 ± 1,062 μm² and 6,227 ± 1,007 μm² vs. 7,645 ± 1,201 μm²; P = 0.02 and 0.004, respectively), and combination treatment proved to be the most successful in preventing enlarged glomerular tufts (Figure 3a, b). Finally, immunofluorescence staining showed decreased amounts of IgG and C3 deposition in kidney sections of L-4F-treated mice compared to control mice (Figure 3c).

Compared to vehicle controls, total skeletal BMD (excluding the skull) and lumbar BMD, measured using DEXA, were significantly higher in female mice treated with pravastatin, L-4F, or L-4F plus pravastatin (total: 0.041 ± 0.002 vs. 0.043 ± 0.002 and 0.044 ± 0.002 and 0.044 ± 0.002 g/cm³, respectively and vertebral: 0.036 ± 0.004 vs. 0.051 ± 0.005 and 0.051 ± 0.005 and 0.053 ± 0.003 g/cm³, respectively), with no significant difference between the pravastatin, L-4F, and L-4F plus pravastatin-treated groups (Figure 4a). Additionally, there were no apparent treatment-dependent effects on femoral BMD. Concurrent μCT analysis showed that mice treated with L-4F had significantly higher bone volume density (P = 0.023), trabecular number (P = 0.019), and connectivity density (P = 0.00054) and significantly lower trabecular separation compared to vehicle controls (P = 0.04) (Figure 4b, c). In contrast, treatment with pravastatin alone was associated with a borderline reduction in bone volume density, and treatment with L-4F plus pravastatin did not show significant improvements in any of these trabecular characteristics.

Following 27 weeks of treatment then euthanasia, the basal portion of the heart and the proximal aorta showed enlarged aortic lesions in mice treated with pravastatin, L-4F, or L-4F plus pravastatin compared to controls (Figure 5a). Analysis of local plaque environment composition at the aortic root demonstrated significantly decreased CD68+ macrophage infiltration, when comparing the average total stained area per mean lesion area, in L-4F plus pravastatin-treated mice compared to age-matched vehicle controls (6.2 ± 1.2% vs. 9.8 ± 0.8%; P = 0.002) (Figure 5b, c). L-4F plus pravastatin-treated mice also showed increased α-actin smooth muscle content in aortic lesions compared to controls (7.8 ± 0.5% vs. 4.9 ± 2.3%; P = 0.04) (Figure 5b, c). Mice treated with pravastatin or L-4F alone did not show any significant improvements in aortic lesion cellular composition compared to control mice. Analysis of Movat, CD4+ T-cell, or VCAM-1 stained lesions did not show any significant differences in atheroma composition of elastic fibers, ground substance, muscle, collagen, fibrinoid and fibrin (See Supplemental figure S1 in Additional file 2 for Movat staining of aortic root atheromas), CD4+ T-cells or VCAM-1 distribution between any of the treatment groups and the control group (data not shown).

Plasma lipid profiles for apoE^-/-Fas^-/- L-4F-treated mice and vehicle control mice did not show any significant differences in triglyceride, total cholesterol, HDL cholesterol, non-HDL cholesterol, unesterified cholesterol, and free fatty acid levels (Figure 6d). L-4F improved the anti-inflammatory function of plasma HDL and decreased the proinflammatory effects of LDL from mice injected with L-4F as determined in cultures of human aortic endothelial cells compared to LDL from vehicle control mice (Figure 6e).

To explore potential biomarkers associated with treatment response, plasma from female apoE^-/-Fas^-/- mice was analyzed for 69 chemokines and cytokines using Luminex-based beadarray. L-4F treatment resulted in a trend toward decreased levels of tissue damage and inflammation indicators, including CRP (C-reactive protein), fibrinogen, TNF-α (tumor necrosis factor-alpha), and CCL12 (monocyte chemotactic protein 5 (MCP-5)), when compared to control mice (data not shown).

After Bonferroni correction for multiple testing (54 detectable antigens), plasma levels of IL-10 (interleukin-10) - a cytokine secreted in response to damaged tissue through growth and differentiation of NK and B cells and CCL9 (macrophage inflammatory protein-1γ (MIP-1γ)) - a chemoattractant for monocytes, neutrophils, and macrophages that contributes to monocyte infiltration in renal disease, were significantly lower in L-4F-treated mice (Figure 6a, b). Decreased levels of CCL19 -- a homeostatic interferon-regulated chemokine that binds to CCR7 and plays a role in recruiting T-cells and dendritic cells to target organs, promoting inflammatory responses, and unstable plaque formation in atherosclerosis 36 -- were present in mice treated with L-4F compared to control mice (Figure 6c). Similarly, the endothelial receptor VCAM-1, commonly associated with the recruitment of monocytes and lymphocytes during atherosclerotic plaque formation 37, was significantly decreased in plasma of mice treated with L-4F, as compared to control mice (Figure 6c). Pravastatin monotherapy alone did not significantly affect any of the levels of these circulating chemokines and cytokines.