Research ArticleOBSTRUCTIVE SLEEP APNEA

Increased internalization of complement inhibitor CD59 may contribute to endothelial inflammation in obstructive sleep apnea

See allHide authors and affiliations

Science Translational Medicine  06 Jan 2016:
Vol. 8, Issue 320, pp. 320ra1
DOI: 10.1126/scitranslmed.aad0634
  • Fig. 1. EC plasma membrane proteins are differentially expressed in OSA patients and OSA-free controls.

    (A) Harvesting of venous ECs from study participants (n = 76 OSA patients and n = 52 controls, demographics in table S1). (B) Frequency of peptide F binding to ECs harvested from OSA patients [n = 18: 45% male; age, 44.6 ± 2.4 years; body mass index (BMI), 38.1 ± 2.1 kg/m2; apnea-hypopnea index (AHI), 26.2 ± 6.2 events/hour; oxygen desaturation index (ODI), 11.4 ± 2.8 events/hour] and controls (n = 20: 55% male; age, 40.9 ± 2.7 years; BMI, 27.9 ± 1.3 kg/m2; AHI, 1.3 ± 0.4; ODI, 0.37 ± 0.3 events/hour) expressed as a percentage of the total number of bound peptides (mean ± SE; two-sided Student’s t test adjusted for age, gender, and BMI shown in fig. S2, P = 0.001). (C) Representative confocal image of peptide F binding to ECs harvested from OSA patients (n = 10: 30% male; age, 50.3 ± 3.2 years; BMI, 34.9 ± 1.8 kg/m2; AHI, 19.4 ± 5.8 events/hour; ODI, 12.3 ± 3.3 events/hour) and controls (n = 10: 60% male; age, 38.4 ± 3.4 years; BMI, 29.4 ± 1.6 kg/m2; AHI, 1.9 ± 0.7 events/hour; ODI, 1.1 ± 0.6 events/hour); ECs are identified by immunofluorescence for von Willebrand factor. Scale bars, 10 μm. (D) Native polyacrylamide gel electrophoresis (PAGE) and Western blotting with anti–peptide F antibodies (Ab) of human umbilical vein endothelial cell (HUVEC) lysate incubated with or without peptide F. The region of bound peptide F (a single major band at molecular weight of 250 kD) was cut out from the gel (red rectangle), subjected to trypsin digestion, and analyzed by mass spectrometry. Among proteins identified in this region, CD59 was the only one that is expressed on the plasma membrane. (E) In vitro binding of recombinant CD59 (rCD59; molecular weight, 18 to 20 kD) to peptide F (molecular weight, 700 to 800 daltons) after cross-linking but not to control peptide (Western blots probed with anti–peptide F, anti–control peptide, and anti-CD59 antibodies; n = 3). (F) Immunoprecipitation with beads coated with anti-DDK antibodies of human embryonic kidney (HEK) 293 cells transfected with Myc-DDK–tagged human CD59 plasmid incubated with peptide F, followed by SDS-PAGE and Western blotting with anti-CD59, anti-DDK, and anti–peptide F antibodies, showing specific binding of peptide F to CD59 (n = 3).

  • Fig. 2. Cellular distribution of CD59 is altered in OSA.

    (A) Mean fluorescence intensity of total cellular CD59 (flow cytometry, shaded area represents isotype control; the quantity of cells on the y axis is expressed as percentage of total specimens). Bar graph quantitates total CD59 fluorescence (n = 8 OSA patients: 62% male; age, 42.3 ± 5.2 years; BMI, 35.9 ± 2.0 kg/m2; AHI, 21.3 ± 8.7 events/hour; ODI, 12.9 ± 4.5 events/hour; n = 9 controls: 33% male; age, 36.7 ± 4.6 years; BMI, 35.8 ± 3.2 kg/m2; AHI, 3.2 ± 0.7 events/hour; ODI, 2.0 ± 0.5 events/hour) (mean ± SE; two-sided exact permutation test). (B) Quantitation of the CD59 mRNA expression in ECs harvested from OSA patients (n = 21: 67% male; age, 45.5 ± 2.7 years; BMI, 35.6 ± 2.1 kg/m2; AHI, 23.5 ± 5.5 events/hour; ODI, 19.3 ± 6.1 events/hour) expressed as a fold change over controls (n = 15: 27% male; age, 34.1 ± 3.4 years; BMI, 36.6 ± 3.0 kg/m2; AHI, 1.3 ± 0.4 events/hour; ODI, 0.3 ± 0.2 events/hour) (two-sided Student’s t test adjusted for age, gender, and BMI, shown in fig. S2). (C) Representative confocal images of cellular distribution of CD59 in ECs harvested from OSA patients (n = 10: 30% male; age, 50.3 ± 3.2 years; BMI, 34.9 ± 1.8 kg/m2; AHI, 19.4 ± 5.8 events/hour; ODI, 12.3 ± 3.3 events/hour) and controls (n = 10: 40% male; age, 34.5 ± 3.7 years; BMI, 34.1 ± 3.0 kg/m2; AHI, 2.2 ± 0.6 events/hour; ODI, 1.4 ± 0.6 events/hour); EC plasma membrane is identified by immunofluorescence for vascular endothelial (VE)–cadherin (CD144). Scale bars, 10 μm. (D) Histogram representing the percentage of total endothelial CD59 located on the plasma membrane for individual patients [n = 10 OSA patients; n = 10 controls; demographics in the legend for (C)]. (E) Quantitation of the percentage of total endothelial CD59 located on the plasma membrane [n = 10 OSA patients; n = 10 controls; demographics in the legend for (C)]. Linear regression confirmed that the difference in plasma membrane expression of CD59 is not confounded by age, gender, or BMI (fig. S2) (mean ± SE; two-sided exact permutation test). FITC, fluorescein isothiocyanate.

  • Fig. 3. IH increases endocytosis of endothelial CD59 and consequent MAC deposition.

    (A) Representative confocal images of CD59 endocytosis in HUVECs in normoxia, IH, and continuous hypoxia (CH). Arrows indicate endocytosed CD59 with flotillin (colocalized area in μm2). Endocytosis mediated by transferrin is shown on the bottom panel. Scale bars, 10 μm. (B) Quantitation of colocalized area of endocytosed CD59 with flotillin-1 and transferrin (n = 8). (C) Quantitation of the percentage of HUVECs positive for MAC deposition in normoxia, IH, and continuous hypoxia (flow cytometry), indicating the percentage of dead cells in each condition below the graph (identified by propidium iodide staining) (n = 8). (D) Representative confocal images of MAC deposition on ECs harvested from OSA patients and controls. Scale bars, 10 μm. Bar graph quantitates MAC deposition area (μm2) on ECs of OSA patients (n = 8: 62% male; age, 39.0 ± 4.2 years; BMI, 37.2 ± 3.5 kg/m2; AHI, 13.7 ± 2.4 events/hour; ODI, 11.5 ± 2.8 events/hour) and controls (n = 10: 33% male; age, 39.4 ± 4.4 years; BMI, 34.1 ± 3.4 kg/m2; AHI, 1.9 ± 0.9 events/hour; ODI, 1.7 ± 1.0 events/hour). Linear regression confirmed that the difference in MAC deposition on ECs among groups is not confounded by age, gender, or BMI (fig. S2). All data throughout the figure are shown as means ± SE (two-sided exact permutation test). NS, not significant.

  • Fig. 4. IH promotes endothelial inflammation in a CD59-dependent manner.

    (A) Representative confocal images showing nuclear translocation of NFκB in HUVECs exposed to normoxia or IH and incubated with 20% normal serum before and after addition of recombinant CD59 and with heat-inactivated serum. Scale bars, 10 μm; inset, 5 μm. Bar graph quantitates NFκB fluorescence intensity in HUVECs (n = 9). (B) Quantitation of the mean fluorescence of MCP-1 and ICAM-1 in permeabilized HUVECs incubated with 20% normal serum with or without recombinant CD59, or with heat-inactivated serum in normoxia and IH (flow cytometry; n = 5). (C) Representative confocal images of NFκB fluorescence intensity in ECs of OSA patients and controls. Scale bars, 10 μm. Bar graph quantitates NFκB fluorescence intensity in ECs from OSA patients (n = 9: 55% male; age, 42.4 ± 4.5 years; BMI, 34.5 ± 3.4 kg/m2; AHI, 8.2 ± 1.6 events/hour; ODI, 5.4 ± 1.8 events/hour) and controls (n = 7: 14% male; age, 43.9 ± 5.0 years; BMI, 36.0 ± 4.9 kg/m2; AHI, 1.4 ± 0.7 events/hour; ODI, 0.9 ± 0.9 events/hour). Linear regression confirmed that the difference in NFκB fluorescence intensity in ECs among groups is not confounded by age, gender, or BMI (fig. S2). All data throughout the figure are shown as means ± SE (two-sided exact permutation test).

  • Fig. 5. Statins prevent IH-induced internalization of CD59 and MAC deposition on ECs in CD59-dependent manner.

    (A) Quantitation of the percentage of total endothelial CD59 located on the plasma membrane in OSA patients taking statins (n = 10 controls not taking statins: 60% male; age, 38.4 ± 3.4 years; BMI, 29.4 ± 1.6 kg/m2; AHI, 1.9 ± 0.7 events/hour; ODI, 1.1 ± 0.6 events/hour; n = 9 OSA patients not taking statins: 33% male; age, 48.7 ± 3.0 years; BMI, 34.9 ± 2.0 kg/m2; AHI, 19.4 ± 5.8 events/hour; ODI, 12.3 ± 3.3 events/hour; n = 5 OSA patients taking statins: 60% male; age, 56.4 ± 4.6 years; BMI, 32.6 ± 2.2 kg/m2; AHI, 16.2 ± 2.3 events/hour; ODI, 8.5 ± 4.3 events/hour). Linear regression confirmed that the difference in plasma membrane expression of CD59 among groups is not confounded by age, gender, or BMI (fig. S2). (B) Representative confocal images of CD59 endocytosis in HUVECs in normoxia, IH, and continuous hypoxia without and with atorvastatin (n = 4). Arrows indicate endocytosed CD59 with flotillin (colocalized area in μm2). Scale bars, 10 μm. Bar graph quantitates colocalized area of endocytosed CD59 with flotillin-1. (C) Quantitation of the percentage of HUVECs positive for MAC deposition in normoxia and IH after incubation with 20% normal serum and atorvastatin (n = 4; flow cytometry). (D) Quantitation of the percentage of HUVECs positive for MAC deposition in IH after incubation with 20% normal serum with and without atorvastatin, and with or without transfection with CD59 siRNA to achieve CD59 knockdown (n = 4; flow cytometry). (E) Representative confocal images of nuclear translocation of NFκB in HUVECs exposed to normoxia or IH and incubated with 20% normal serum with or without atorvastatin. Bar graph quantitates NFκB fluorescence intensity in HUVECs in normoxia and IH with and without atorvastatin (n = 4). Scale bars, 20 μm; inset, 10 μm. All data throughout the figure are shown as means ± SE (two-sided exact permutation test).

  • Fig. 6. IH increases cellular cholesterol by decreasing gene expression of cholesterol efflux mediators.

    (A) Quantitation of the ABCA1 and ABCG1 mRNA expression in HUVECs in normoxia, IH, and continuous hypoxia expressed as a fold change over normoxia (n = 4). (B) Representative confocal images of cholera toxin B staining of the plasma membrane (depicting lipid rafts containing free cholesterol) in HUVECs in normoxia, IH, and continuous hypoxia. Scale bars, 10 μm. Bar graph quantitates cholera toxin B–positive area (μm2) (n = 4). All data throughout the figure are shown as means ± SE (two-sided exact permutation test).

Supplementary Materials

  • www.sciencetranslationalmedicine.org/cgi/content/full/8/320/320ra1/DC1

    Fig. S1. Expression of CD59 in leukocytes and platelets.

    Fig. S2. Linear regression analysis.

    Fig. S3. Endocytosis of CD59 with transferrin.

    Fig. S4. Confirmation of HUVEC transfection with CD59 siRNA.

    Fig. S5. Cholesterol biosynthesis in IH.

    Table S1. Baseline characteristics of patients with OSA and control subjects.

    Table S2. Phages isolated after panning without target cells and with bovine serum albumin in polystyrene wells.

  • Supplementary Material for:

    Increased internalization of complement inhibitor CD59 may contribute to endothelial inflammation in obstructive sleep apnea

    Memet Emin, Gang Wang, Francesco Castagna, Josanna Rodriguez-Lopez, Romina Wahab, Jing Wang, Tessa Adams, Ying Wei, Sanja Jelic*

    *Corresponding author. E-mail: sj366{at}cumc.columbia.edu

    Published 6 January 2016, Sci. Transl. Med. 8, 320ra1 (2016)
    DOI: 10.1126/scitranslmed.aad0634

    This PDF file includes:

    • Fig. S1. Expression of CD59 in leukocytes and platelets.
    • Fig. S2. Linear regression analysis.
    • Fig. S3. Endocytosis of CD59 with transferrin.
    • Fig. S4. Confirmation of HUVEC transfection with CD59 siRNA.
    • Fig. S5. Cholesterol biosynthesis in IH.
    • Table S1. Baseline characteristics of patients with OSA and control subjects.
    • Table S2. Phages isolated after panning without target cells and with bovine serum albumin in polystyrene wells.

    [Download PDF]

Stay Connected to Science Translational Medicine

Navigate This Article