Cross-reacting antibacterial auto-antibodies are produced within coronary atherosclerotic plaques of acute coronary syndrome patients

Abstract

Coronary atherosclerosis, the main condition predisposing to acute myocardial infarction, has an inflammatory component caused by stimuli that are yet unknown. We molecularly investigated the nature of the immune response within human coronary lesion in four coronary plaques obtained by endoluminal atherectomy from four patients. We constructed phage-display libraries containing the IgG1/kappa antibody fragments produced by B-lymphocytes present in each plaque. By immunoaffinity, we selected from these libraries a monoclonal antibody, arbitrarily named Fab7816, able to react both with coronary and carotid atherosclerotic tissue samples. We also demonstrated by confocal microscopy that this monoclonal antibody recognized human transgelin type 1, a cytoskeleton protein involved in atherogenesis, and that it co-localized with fibrocyte-like cells transgelin+, CD68+, CD45+ in human sections of coronary and carotid plaques. In vitro fibrocytes obtained by differentiating CD14+ cells isolated from peripheral blood mononuclear cells also interacted with Fab7816, thus supporting the hypothesis of a specific recognition of fibrocytes into the atherosclerotic lesions. Interestingly, the same antibody, cross-reacted with the outer membrane proteins of Proteus mirabilis and Klebsiella pneumoniae (and possibly with homologous proteins of other enterobacteriaceae present in the microbiota). From all the other three libraries, we were able to clone, by immunoaffinity selection, human monoclonal antibodies cross-reacting with bacterial outer membrane proteins and with transgelin. These findings demonstrated that in human atherosclerotic plaques a local cross-reactive immune response takes place.

Figure 1. Study flow-chart.

TAGLN  =  transgelin; OMPs  =  outer membrane proteins.

Figure 2. Biopanning selection with combinatorial IgG/k library ID-A.

Library ID-A was selected by immunoaffinity on atherosclerotic plaque lysate. Results of screening ELISA assays of 30 clones after four rounds of biopanning selection is shown. Sequence analysis of the positively selected clones (O.D.450 nm >0.25 above background) is shown next to the ELISA screening.

Figure 3. Identification of putative natural self antigen.

A) Bidimensional electrophoresis gel stained with colloidal Coomassie Brilliant Blue. Sixty five µg of purified carotid atherosclerotic plaque proteins were loaded on strip pH 3-10NL, 7 cm. The 2nd dimension was carried out using 12.5% SDS-PAGE. B) 2D electrophoresis Western Blotting. Similarly, 70 µg of proteins were loaded on strip pH 3-10NL, 7 cm. The 2nd dimension was carried out using 12.5% SDS-PAGE. After transfer, proteins were probed with Fab7816-FLAG (10 µg/mL). Protein spots of interest (red box) were excised from the gel, digested with trypsin and analysed by MALDI-ToF mass spectrometry. In both spots human TAGLN was identified with almost complete sequence coverage.

Figure 4. Immunofluorescence on human coronary plaqes and Immunohistochemistry on human carotid sections with Fab7816-FLAG.

a) Representative section stained by Movat’s pentchrome (left panel) showed the morphology of a portion of the coronary plaque tissue (plaque ID-A) displaying slightly damaged media rich in smooth muscle cells. Confocal microscopy (right panels) showed the presence of CD45⁺ cells (red) labelled by Fab7816-FLAG revealed by MAb M2 anti-FLAG-FITC (green) in the coronary plaque tissue region indicated by symbol (#). DAPI stained the nuclei (blue). b) Immunoperoxidase on carotid plaque samples demonstrated the presence of several cells reacting with Fab7816-FLAG in an area close to the lumen (left panels) as revealed by MAb M2 anti-FLAG-HRP developed with DAB (brown). The signal is absent in a serial section where the Fab7816-FLAG is omitted (ctrl-, right panels). Magnified images in the bottom panels evidence the spindle shape of Fab7816-FLAG+ cells and their localization in between a clusters of altered cells, possibly foam cell. Haematoxylin (blue) stains nuclei. Scale bars indicate the magnification.

Figure 5. Confocal microscopy on human atherosclerotic carotid sections.

Multiple staining of two carotid plaques are displayed to characterize the Fab7816-FLAG+ cells type. First plaque A-F panels, 2^nd plaque G-I to N. The reconstruction of a section from 1^st plaque stained with haematoxylin and eosin (HE) obtained by multiple images grabbing tool of Lucia-G software is in A; asterisk indicates the lumen in correspondence of Fab7816-FLAG+ cells, on the shoulder of the atheromasic lesion demonstrated in B by confocal lmicroscopy. Confocal microscopy images in B,C and G,H demonstrated by Fab7816-FLAG (green), goat-anti-human TAGLN (white), mouse-anti-human CD68 (red) the presence of triple positive cells in the intima, closely to the lumen. Single or double staining are shown in D-F and I-N. Squares and arrows indicated the enlarged areas. DAPI stains the nuclei (blue). Scale bars indicate the magnification.

Figure 6. Confocal microscopy on human atherosclerotic carotid sections.

The macroscopic aspect of a carotid plaque is reconstructed in A by stereomicroscope and displays a lipidic core and of luminal thrombus. Features are confirmed in B by histology (haematoxylin and eosin). Confocal microscopy in C shows a reconstruction of part of the area close to the lumen (L) in a section stained with Fab7816-FLAG (green), mouse-anti-human Collagen type I (white) and mouse anti-human CD45 (red), revealed by opportune secondary antibodies. In D is an enlargement of the area indicated by § symbol in C with several Fab7816-FLAG+/CD45+ cells in the neointima. indicate a regions magnified. Scale bars indicate the magnification.

Figure 7. Fibrocytes morphology and immunoreactivity with 7816Fab FLAG.

A) Confocal microscopy staining showing presence of 7816Fab FLAG +/CD45+/TAGLN+ cells in human carotid plaque lesion. B,C) Confocal microscopy image of non-confluent CD14+ fibrocytes grown on glass for 4 days in the absence of serum. Spindle shaped cells are stained with 7816Fab FLAG +/CD45+(green and red, respectively). DAPI stains the nuclei (blue). Scale bars indicate the magnification.

Figure 8. Western Blotting of bacterial lysates and on bacterial OMPs with Fab 7816.

A) Western blotting of bacterial lysates with Fab7816. 1 µg of each bacterial lysate was loaded in each lane. Fab 7816 clearly reacted with Proteus mirabilis and Klebsiella pneumoniae lysates. B) Western blotting on induced/non induced BL21(DE3) bacterial cells transformed with pET28b vector expressing OmpK36. Two pET vectors were constructed: the NcoI/XhoI wild type OmpK36 vector or the HIS-tagged Ompk36 BamHI/XhoI vector. In both cases Fab 7816 recognized only IPTG induced bacteria.

Figure 9. Biopanning selection with combinatorial IgG/k libraries ID-B, ID-C and ID-D.

Each library was independently selected by immunoaffinity on purified OmpK36. Results of screening ELISA assays of 30 clones after four rounds of biopanning selection is shown. Sequence analysis of the positively selected clones (O.D.450 nm >0.25 above background) is shown next to each ELISA screening.

Figure 10. Cross reactivity of representative Fabs from all patients with TAGLN and OMPs.

A) WB of purified human TAGLN (400 ng) with all representative Fabs (10 µg/mL). Unrelated human e8Fab-FLAGwas used as negative control. Anti-MYC-tag (C-terminal tag) and commercial anti-TAGLN were used as positive controls. While the commercial anti-TAGLN recognize selectively only one form of TAGLN, cloned human Fabs recognized both forms. B) WB of OmpK36 (500 ng) with all representative Fabs (10 ng/mL). C) ELISA with all representative Fabs on bacterial OMPs. Reactivity against bovine serum albumin (BSA) used as blocking antigen, is also shown.

Figure 11. Cross reactivity of commercial available monoclonal antibodies with TAGLN and OMPs.

A) WB of OmpK36 with five commercial monoclonal mouse anti-human TAGLN antibodies (used at 1 or 5 µg/ml). Three commercial mouse monoclonal antibodies were used as negative controls. Anti-6xHIS antibody (Roche) was used as positive control B) ELISA with all representative and five commercial monoclonal mouse anti-human TAGLN antibodies on purified human TAGLN or C) bacterial OMPs. Reactivity against bovine serum albumin (BSA) used as blocking antigen, is also shown.