Atherogenic LOX-1 signaling is controlled by SPPL2-mediated intramembrane proteolysis

Abstract

The lectin-like oxidized LDL receptor 1 (LOX-1) is a key player in the development of atherosclerosis. LOX-1 promotes endothelial activation and dysfunction by mediating uptake of oxidized LDL and inducing pro-atherogenic signaling. However, little is known about modulators of LOX-1-mediated responses. Here, we show that the function of LOX-1 is controlled proteolytically. Ectodomain shedding by the metalloprotease ADAM10 and lysosomal degradation generate membrane-bound N-terminal fragments (NTFs), which we identified as novel substrates of the intramembrane proteases signal peptide peptidase-like 2a and b (SPPL2a/b). SPPL2a/b control cellular LOX-1 NTF levels which, following self-association via their transmembrane domain, can activate MAP kinases in a ligand-independent manner. This leads to an up-regulation of several pro-atherogenic and pro-fibrotic targets including ICAM-1 and the connective tissue growth factor CTGF. Consequently, SPPL2a/b-deficient mice, which accumulate LOX-1 NTFs, develop larger and more advanced atherosclerotic plaques than controls. This identifies intramembrane proteolysis by SPPL2a/b as a novel atheroprotective mechanism via negative regulation of LOX-1 signaling.

Graphical abstract

Figure 1.

LOX-1 NTFs are generated by ADAM10 and lysosomal proteases. (A and B) Two LOX-1 NTFs were detected by Western blotting upon overexpression of HA-LOX-1 in either HeLa (A) or iMAEC (B) cells. (C) Usage of the potential glycosylation sites at N72 or N92 was analyzed in HeLa cells using respective mutants. (D) HEK cells deficient for ADAM10, ADAM17, or both and WT cells were transfected with HA-LOX-1-FLAG. sLOX-1 was recovered from conditioned media after 16 h by TCA precipitation. (E) Quantification of D. N = 2–3, n = 4–6. One-way ANOVA with Dunnett’s post hoc test. (F) iMAECs stably expressing HA-LOX-1-FLAG were preincubated for 2 h in serum-free DMEM containing 10 µM marimastat (Mari) or DMSO. Cells were treated for 30 min with 1 µM ionomycin (Iono) or 100 nM PMA. Cell lysates and conditioned media were analyzed as in D. (G) Lysosomal processing of HA-LOX-1 was blocked in HeLa cells by incubation with 100 nM bafilomycin a1 (Baf a1) for 24 h. (H) Quantification of G. N = 3, n = 3. Student’s t test. (I) Delivery of HA-LOX-1 to LAMP-2–positive compartments was visualized by indirect immunofluorescence in HeLa cells treated with either DMSO or 100 nM bafilomycin a1 for 6 h. Bars, 10 µm. (K) HA-LOX-1–expressing HeLa cells were treated for 4 h with 40 or 80 µg/ml oxLDL, and NTF formation was analyzed by Western blotting. (L) Quantification of NTF1+2/FL ratios of cells treated with 40 µg/ml oxLDL. N = 2, n = 4. Student’s t test. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 2.

LOX-1 NTFs undergo SPPL2a/b-dependent intramembrane proteolysis. (A) HeLa cells were transfected with HA-LOX-1 and WT or inactive (D/A) SPPL2 proteases. Where indicated, SPPL2a/b activity was inhibited with 20 µM ZLL for 6 h. (B) Quantification of A. N = 7, n = 7. Student’s t test. Norm., normalized. (C) Colocalization of HA-LOX-1 and SPPL2a-myc or SPPL2b-myc in transfected HeLa cells. (D and E) iMAECs transfected with HA-LOX-1 were treated with 40 µM ZLL for 24 h before Western blot analysis. LOX-1 (NTF1+2)/FL ratios are depicted in E. N = 2, n = 5. Student’s t test. (F and G) WT or SPPL2a/b dKO MEFs were stably transduced with HA-LOX-1-FLAG followed by Western blot analysis. LOX-1 (NTF1+2)/FL ratios are depicted in G. N = 2, n = 6. Student’s t test. (H) WT, SPPL2a- SPPL2b-, or double-deficient MEFs were transfected with HA-LOX-1 and analyzed by Western blotting. (I) Quantification of H. N = 3–4, n = 5–8. One-way ANOVA with Tukey’s post hoc testing. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 3.

The LOX-1 TMD is important for SPPL2-dependent intramembrane cleavage. (A–C) Mass-spectrometric determination of SPPL2a/b cleavage sites within the LOX-1 NTF. (A) Amino acid sequence of the employed model substrate. Determined cleavage sites in the TMD (gray) are marked. (B) Secreted C-terminal fragments were purified from conditioned media and analyzed by MS. Arrows indicate peptides increased by protease overexpression with the predominant peaks labeled in red. (C) Peptides assigned to the respective peaks shown in B. Peak 5 corresponds to a peptide with a potential N-terminal glutamate to pyroglutamate conversion. obs., observed; calc., calculated. (D) Scheme of LOX-1 NTF TMD mutants. (E) Subcellular sorting of LOX-1 TMD mutants was compared with the WT LOX-1 NTF by indirect immunofluorescence with anti-HA in HeLa cells. Bar, 10 µm. (F) Cleavage of LOX-1 NTF TMD mutants by SPPL2b was analyzed in transfected HeLa cells. (G) Quantification of F. N = 4, n = 4. One-way ANOVA with Dunnett’s post hoc test. ***, P ≤ 0.001; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 4.

Endogenous LOX-1 is processed by SPPL2a and SPPL2b. (A and B) Evaluation of functionality of the newly generated antibody against the N terminus of murine LOX-1 in transfected HeLa cells. (A) Blot probed with new antibody. (B) Same blot as in A probed with anti-HA. (C) Expression of SPPL2a and SPPL2b was demonstrated in iMAECs and aortic lysates by Western blotting. (D) iMAECs were treated for 24 h with either 40 µM ZLL or DMSO before Western blot analysis. (E) Quantification of D. N = 2, n = 6. Student’s t test. (F) LOX-1 NTF levels were analyzed in aortic lysates from WT, SPPL2a-, SPPL2b-, or dKO mice by Western blotting. (G) Quantification of F. N = 4, n = 4. One-way ANOVA with Tukey’s post hoc testing. **, P ≤ 0.01; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 5.

The LOX-1 NTF regulates MAP kinase activity. (A) T-REx FlipIn cells inducibly overexpressing LOX-1 were generated. LOX-1 expression after induction with 10 µg/ml doxycycline (dox) for 24 h was confirmed by Western blotting. (B) LOX-1 expression was induced in T-REx FlipIn cells (+dox) or cells were left uninduced (–dox). Prior to stimulation with 40 µg/ml oxLDL, cells were cultured in serum-free DMEM for 4 h in the presence of 40 µM ZLL or DMSO. pERK and total ERK1/2 (ERK) levels were determined. (C) Quantification of B. N = 4, n = 4. Student’s t test. (D) iMAECs were incubated with 40 µM ZLL or DMSO for 24 h. For the last 4 h, cells were transferred to serum-free DMEM before treatment with 40 µg/ml oxLDL and analysis of ERK1/2 activation. (E) Quantification of D. N = 4, n = 4. Student’s t test. (F) LOX-1 surface levels were analyzed by flow cytometry in iMAECs stably overexpressing LOX-1 after 24 h treatment with 40 µM ZLL. N = 2, n = 6. Student’s t test. (G) Stable expression of the LOX-1 NTF in transduced iMAECs was validated. (H) Transduced iMAECs were cultivated for 4 h under serum-free conditions before stimulation with 40 µg/ml oxLDL and assessment of MAP kinase activation. (I) Quantification of H. N = 3, n = 3. Student’s t test. (K) HeLa cells were transfected as indicated with differentially tagged full-length LOX-1 and NTF. Interaction of both proteins was analyzed by coimmunoprecipitation using anti-V5. (L) HeLa cells transfected with 3xFLAG-LOX-1 (3xFLAG-FL) and/or HA-LOX-1_1-88 (HA-NTF) were treated for 30 min with BS³ and subsequently analyzed by Western blotting. As described in Materials and methods, phosphorylated and total forms of ERK were detected from the same membranes. After detection of pERK, the respective membranes were stripped and reprobed to detect total ERK. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 6.

The LOX-1 NTF autonomously activates MAP kinases. (A and B) iMAECs were stably transduced with the pMSCV puro vector (–) or a LOX-1-NTF coding construct. Levels of phosphorylated as well as total ERK1/2, p38 (A), p65, and Akt (B) were analyzed by Western blotting. (C) Quantification of A and B. N = 3, n = 9. Student’s t test. (D) Control or NTF-transduced iMAECs were starved for 16 h in serum-free DMEM and subsequently analyzed for activation of MAP kinases. (E) Quantification of D. N = 2, n = 6. Student’s t test. (F) Vector (–) or LOX-1 NTF transfected HEK cells were selected with puromycin for 4 d. (G) Quantification of F. N = 2, n = 7. Student’s t test. (H) Empty vector (–) or LOX-1 NTF transduced iMAECs were treated for 3 h with 25 µM U0126 (MEK-Inh.), 1 µM Saracatinib (Src-Inh.), or 10 µM Y-27632 (ROCK-Inh.) and subsequently analyzed for ERK1/2 activation. (I) Quantification of H. N = 2, n = 6. One-way ANOVA with Tukey’s post hoc test. (K) Pathways upstream of p38 MAP kinase activation were assessed as depicted in H. (L) Quantification of K. n = 4, n = 12. One-way ANOVA with Tukey’s post hoc test. (M) Subcellular sorting of an unphosphorylatable LOX-1 NTF (NTFΔP) was compared with the WT NTF by indirect immunofluorescence in HeLa cells. Bars, 10 µm. (N) HEK cells were transfected and incubated for 4 d with 10 µg/ml puromycin. Activation of ERK1/2 was monitored by Western blotting. (O) Quantification of N. pERK/ERK levels were normalized (norm.) to LOX-1 NTF expression. N = 2, n = 6. Student’s t test. (P) Induction of pERK by LOX-1 TMD mutants was analyzed by Western blotting. (Q) pERK/ERK ratios were normalized to NTF expression. N = 3, n = 6. One-way ANOVA with Dunnett’s post hoc test. (R) Coimmunoprecipitation of the HA-tagged NTF mutants with a FLAG-tagged WT LOX-1 NTF from lysates of transfected HEK cells. (S) Quantification from N = 3, n = 3 experiments. One-way ANOVA with Dunnett’s post hoc test. As described in Materials and methods, phosphorylated and total forms of ERK, p38, p65, and Akt were detected from the same membranes. After detection of the phosphorylated forms, membranes were stripped and reprobed to detect the total proteins. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 7.

Accumulation of the LOX-1 NTF induces a pro-atherogenic state in endothelial cells. (A) Expression of ICAM-1 was analyzed in control (–) or LOX-1 NTF transduced iMAECs by Western blotting. (B) Quantification of A. N = 2, n = 6. Student’s t test. (C) Up-regulation of surface ICAM-1 in LOX-1 NTF transduced iMAECs was validated by flow cytometry. As a control, cells were treated with 5 ng/ml TNF. N = 2, n = 6. One-way ANOVA with Tukey’s post hoc test. (D) Up-regulation of Icam-1 was validated by qPCR. N = 2, n = 6. Student’s t test. (E) Candidate genes from endothelial cell biology and atherosclerosis RT² Profiler arrays with differential regulation between control and iMAEC NTF cells. (F) Differences in mRNA levels of the selected genes were validated by qPCR. N = 2–3, n = 6–9. Student’s t test. (G) Secretion of CTGF was blocked in iMAEC control (−) or NTF cells by incubation with Brefeldin A (1 µg/ml) for 6 h. Intracellular CTGF levels were analyzed by Western blotting. (H) Quantification of G. N = 2, n = 12. Student’s t test. (I–L) iMAEC control or NTF cells were treated for 3 h with 25 µM U0126 (MEK-Inh.), 1 µM Saracatinib (Src-Inh.), or 10 µM Y-27632 (ROCK-Inh.) or left untreated as indicated. Icam-1 (I, N = 3–4, n = 9–12), Ctgf (K, N = 3, n = 8–9), and Pdgfb (L, N = 3, n = 8–9) mRNA levels were quantified by qPCR. One-way ANOVA with Dunnett’s post hoc test. (M) Up-regulation of validated candidate genes was monitored in iMAECs treated for 16 h with either DMSO or 40 µM ZLL. N = 2–3, n = 6–9. Student’s t test. (N) Up-regulation of ICAM-1 upon ZLL administration was validated by flow cytometry. N = 3, n = 11. Student’s t test. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 8.

Enhanced atherosclerosis in SPPL2a/b double-deficient mice. (A–G) Hypercholesterolemia and atherosclerosis were induced in WT and SPPL2a/b-deficient (dKO) mice by adeno-associated viral expression of D377Y-mPCSK9 and an HCD for 9 wk following bone marrow transplantation (BMT) with WT bone marrow. (A) Proportions of transitional stage 1 (1) and 2 (T) as well as mature B cells (% of viable splenocytes) were analyzed by flow cytometry at the end of the atherosclerosis experiment. The dot plots depict viable B220⁺ splenocytes. (B) Plasma triglyceride (TG) and cholesterol (Chol) levels were determined 3 wk after HCD initiation. N = 2, n = 18–20. (C) Atherosclerotic plaque development was analyzed histologically in H&E-stained aortic root cross-sections. Bars, 400 µm. Plaque macrophage, smooth muscle cell, and collagen contents were quantified based on MAC3, αSMA (alpha smooth muscle actin), and Sirius Red staining, respectively. Bars, 200 µm. N = 2, n = 13–18 (αSMA); n = 13–17 (MAC3), n = 15–18 (size), n = 15–19 (Sirius Red). (D) Quantification of necrotic plaque area. N = 2, n = 16–18. (E) Scoring of atherosclerotic plaques. (F) Activation of ERK1/2 and p38 MAP kinases as well as ICAM-1 levels was compared in atherosclerotic aortae from WT and dKO mice. (G) Quantification of pERK/ERK (N = 2, n = 6), p-p38/p38 (N = 4, n = 11–12) and ICAM-1/actin (N = 2, n = 5–6) ratios. As described in Materials and methods, phosphorylated and total forms of ERK and p38 were detected from the same membranes. After detection of pERK and p38, the respective membranes were stripped and reprobed to detect the total proteins. Student’s t test. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 9.

The role of SPPL2a/b for processing of LOX-1 NTFs is conserved in humans. (A) HeLa cells were transfected with HA-tagged hLOX-1 alone or in combination with human SPPL2a (hSPPL2a) or SPPL2b (hSPPL2b-myc). Where indicated, cells were treated with 20 µM ZLL for 6 h before lysis and Western blot analysis. (B) Quantification of A. N = 3–4, n = 3–4. One-way ANOVA with Dunnett’s post hoc test. (C) Accumulation of hLOX-1 NTFs in HeLa cells upon inhibition of endogenous SPPL2 proteases with 10, 20, and 40 µM ZLL. (D) Quantification of C for 40 µM ZLL treatment. N = 4, n = 8. Student’s t test. (E–G) The cleavage sites of human SPPL2a and SPPL2b in the TMD of hLOX-1 were analyzed by MS using V5-hLOX-1 NTF N73A-FLAG as model substrate. Arrows indicate peptides up-regulated by protease overexpression (red, dominant peaks). Peak 4 corresponds to a peptide starting at Q59 with a potential glutamate to pyroglutamate conversion. Rel., relative. (H and I) SPPL2a and SPPL2b are expressed in human coronary artery endothelial cells (HCAECs) as revealed by RT-PCR (H) and qRT-PCR (I). (K) Presence of SPPL2a and SPPL2b in human atherosclerotic plaques was investigated by immunohistochemistry. Bars, 400 µm; close-ups, 50 µm. ***, P ≤ 0.001; **, P ≤ 0.01; *, P ≤ 0.05; ns, not significant. N, the number of independent experiments; n, the number of individual samples for quantification. All data are shown as mean ± SD.

Figure 10.

The intramembrane proteases SPPL2a/b control the development of atherosclerosis. SPPL2a and SPPL2b maintain endothelial homeostasis by clearing LOX-1 NTFs generated by ADAM10 and in lysosomes, thereby releasing the LOX-1 ICD. In the absence of SPPL2a/b, LOX-1 NTFs accumulate, enhance oxLDL-induced signaling by full-length LOX-1, and activate MAP kinases in a ligand-independent manner. This promotes endothelial dysfunction and causes a pro-atherogenic and pro-fibrotic phenotype in SPPL2a/b double-deficient mice when challenged with hypercholesterolemia.