JCI - Protective immunity against atherosclerosis carried by B cells of hypercholesterolemic mice

Animals. Male apoE° mice (8) were bred to a ten-generation backcross against C57BL/6J at M&B (Ry, Denmark). At 6 weeks of age, anesthetized mice were subjected to either sham operation or splenectomy (n = 4 to 7 mice per group; see Table 1 and figure legends). In certain experiments, splenectomy or sham operation was immediately followed by transfer through the tail vein of cells from syngenic donors (see below). Key experiments were repeated three times, and a total of 107 mice were used for analysis of atherosclerosis. After surgery, mice were maintained for 12 weeks in isolator units with filtered air, sterilized beddings, and sterilized Western diet with 0.15% cholesterol (29) and treated with a cocktail of antibiotics (sulfametoxine and trimethoprim, Borgal; Hoechst, Frankfurt, Germany). The perioperative mortality was 8% and no infections were registered during daily examinations by the staff, during weekly examinations by a veterinarian, or at autopsy. Mice were sacrificed under anesthesia, sera harvested, and cell suspensions obtained from the inguinal, axillary, cervical, and paraaortic lymph nodes. The heart and the aorta were mounted in OCT embedding medium, frozen in carbon dioxide, and stored at –85°C.

Isolation of spleen cells. Spleens were collected from donor mice and cells harvested through 100-μm nylon filters. Erythrocytes were lysed and nucleated cells were washed twice, counted, suspended in culture medium, and injected into recipient mice within 3 hours. The spleen cell preparations contained 16.7% ± 2.5% CD3⁺ T cells, 43.1% ± 1.1% CD19⁺ B cells, and 40.1% non-B, non-T cells. B and T cells were isolated by negative selection after an initial adherence step with cells in culture medium on tissue culture plastic dishes for 30 minutes at 37°C. Biotin-conjugated anti-CD3 and anti-CD19 antibodies were added to non-adherent cells, which were eliminated by using paramagnetic avidin-coated beads, and MidiMax columns (Becton Dickinson Immunocytometry Systems, Mountain View, California, USA). Immunomagnetic selection was repeated twice, and unbound cells were washed and injected into the recipient. The purity of isolated T and B cells was always >98% by FACS in the CD45⁺ × lymphocyte (FSC/SSC) gate. The T cell preparations contained 60% ± 3% CD4⁺ and 40% ± 4% CD8⁺ cells.

Transfer of cells. Whole (∼60 × 10⁶) or selected (∼20 × 10⁶) spleen cell suspensions in a fixed volume of 500 μl sterile PBS were slowly injected into the tail veins of anesthetized recipient mice using a 27-gauge needle mounted on a 1-ml syringe.

LDL preparations and serum assays. LDL (d = 1.019–1.063 g/ml) was obtained under sterile conditions by ultracentrifugation of human plasma collected from ten donors. Malondialdehyde (MDA) modification was performed as described (30), and the extent of modification was assessed by fluorimetry at 400 and 470 nm. The titers of specific anti–MDA-LDL antibodies were measured by ELISA using alkaline phosphatase–labeled anti-mouse IgG (30) and calculated as described (31). Serum levels of IgG isotypes, IFN-γ, and IL-4 were analyzed by ELISA (30) (PharMingen, San Diego, California, USA) and serum cholesterol was analyzed by a cholesterol oxidase kit (Boehringer Mannheim GmbH, Mannheim, Germany).

FACS analysis. Lymph node cells from individual mice were stained with the fluorescent antibodies CD45-Cy, CD3-FITC, CD19-PE, CD4-Cy, and CD8-PE (PharMingen) using a triple-staining immunofluorescence protocol and a FACSCalibur (Becton Dickinson Immunocytometry Systems) flow cytometer. Five thousand lymphocytes identified by light scatter and CD45 staining were analyzed in each sample.

Cell proliferation analysis. Lymph node cells (inguinal, axillary, and cervical from individual mice) were incubated for 3 days with concanavalin A (Con A) (2.5 μg/ml) followed by Con A and ³H-thymidine for 18 hours and analyzed by β counting (14).

Evaluation of spleen cell transfer and homing. Two different approaches were used to study homing of transferred cells. Spleen cells isolated from apoE° mice were labeled with the CellTracker CM-DiI kit (Molecular Probes, Europe BV, Leiden, The Netherlands). Six splenectomized and three sham-operated apoE° mice received 60 × 10⁶ DiI-labeled spleen cells by intravenous injection. Four weeks later, cell suspensions were obtained from the inguinal, mesenteric, and paraaortic lymph nodes and, when present, the spleen. Cells from individual mice were stained with anti–CD3-FITC and anti–CD45-Cy, and DiI⁺ cells (FL2 channel) were detected in the lymphocyte population gated on the FSC/SSC scatter and among the CD45⁺ (i.e., leukocyte) and the CD3⁺ populations. Six non-reconstituted splenectomized apoE° mice were used as controls.

Four apoE° splenectomized and four sham-operated mice were injected each with 70 × 10⁶ spleen cells from apoE° mice transgenic for the thymidine kinase (TK). apoE° mice were crossed with Ep TK mice (C57BL/6J background), in which the Herpes simplex TK gene is controlled by the human CD4 promoter. This results in TK expression exclusively in mature T cells (32) and offers the possibility of tracking the injected cells either by PCR amplification of genomic DNA, thus detecting all transferred cells, or by amplifying mRNA by reverse transcription–PCR (RT-PCR), which detects transferred T cells. Four weeks after reconstitution, samples were harvested from organs and blood. The same amounts of tissue (by weight) were used for extraction of genomic DNA and mRNA (DNeasy and RNeasy kits; QIAGEN GmbH, Hilden, Germany). The genesis and phenotype of the apoE° TK mice will be described in greater detail elsewhere.

Morphology and immunohistochemistry. Serial cryostat sections were cut from the proximal 1 mm of the aortic root (14). Hematoxylin/oil red O–stained sections were used for computer-assisted morphometry, and avidin-biotin-peroxidase visualization of sections reacted with cell-specific monoclonals to CD4, CD8, and I-A^b (PharMingen) was employed for cell counting (14). An alkaline phosphatase–conjugated monoclonal anti–α SM-actin antibody (Sigma Chemical Co., St. Louis, Missouri, USA) was used to detect smooth muscle cells. Fibrous cap formation was scored in hematoxylin/oil red O–stained sections (–, no smooth muscle cap; +, <50% of lesion covered by cap; ++, >50% of lesion covered by cap) based on αSM-actin staining. The mean lesion size was determined after measuring ten sections collected at every 100 μm over a 1-mm segment of the aortic root (14). For each section, images were captured in a microcomputer and the surface areas of the lesion(s) and of the entire vessel were measured. The fraction area of lesion (%) is the ratio between the cross-section area occupied by lesion and the total vessel cross-section area. This compensates for variations in the section angle between tissue sections (14). Immunohistochemical data were obtained by manual counting of total cells and all antibody-stained cells in all lesions of one section per mouse. The analysis was validated when another investigator recounted the slides. Both investigators were blinded with regard to the group and treatment from which the slides originated, and the inter-investigator difference was less than 15%.

Genomic DNA analysis. DNA was normalized between samples based on absorbance at 260 nm and amplified by PCR in a mastermix containing 67 mM Tris-HCl, 16 mM (NH₄)₂SO₄, 0.1% Tween-20, 0.75 mM MgCl₂, 0.2 mM dNTP, 1% vol/vol dimethylsulfoxide, 40 U/ml Euroblue Taq polymerase (Eurobio SA, Les Ulis, France), 2 μM primers (upper: 5′-GGA-CGC-GGC-GGT-GGT-AAT-GAC-AAG-3′; lower: 5′-GCG-CGG-CCG-GGT-AGC-ACA-GG-3′). The kinetics of the reaction were characterized by analyzing liver DNA after 30, 34, 38, 40, 42, 44, 46, and 48 PCR cycles. A 37-cycle PCR was chosen for the subsequent analyses since sufficient reaction product was produced to permit easy detection while the reaction was still in an exponential phase (see Results). In all these analyses, the first PCR cycle started with 2.5 minutes in a hot block at 92°C followed by 1 minute at 61°C and 2 minutes at 72°C. The ensuing 36 cycles consisted of 30 seconds at 92°C, 1 minute at 61°C, and 2 minutes at 72°C in a Robocycler (Stratagene, La Jolla, California, USA). Densitometric analysis of the amplified products (182 bp) provided the amount of the TK genomic DNA (arbitrary units), which is expected to be proportional to the number of apoE° TK cells present in each tissue sample.

RT-PCR. RNA was reverse-transcribed as described previously (30). cDNA was amplified with primers specific for GAPDH (2 μM; upper: 5′-GTG-AAG-GTC-GGA-GTC-AAC-G-3′; lower: 5′-GGT-GAA-GAC-GCC-AGT-GGA-CTC-3′) with the same PCR settings as described for genomic DNA. Densitometry of the amplified products was used to normalize the amount of DNA. New PCR reactions were performed with the primers for GAPDH and TK (2 μM), respectively, using the same PCR settings. The kinetics of the reactions were evaluated by amplifying thymus cDNA for 24, 26, 28, 32, and 36 cycles using GAPDH primers and for 30 to 48 cycles using TK primers. A 27-cycle PCR was selected for GAPDH, while 37 cycles were found to be optimal for TK. A densitometric analysis was performed on the amplified TK (182 bp) and GAPDH (299 bp) products, and results are expressed as TK/GAPDH ratio. This measure is expected to reflect the proportion of apoE° TK spleen T cells in the cell population of each tissue.

Statistical analysis. Results are expressed as means ± SEM. Continuous data were analyzed by one-factor ANOVA followed by Fisher’s test for comparisons between means. Data scores for fibrous caps were analyzed by Wilcoxon’s sign rank test. Differences between groups were considered significant if P was less than 0.05.