A new chromogranin A-dependent angiogenic switch activated by thrombin

Abstract

Angiogenesis, the formation of blood vessels from pre-existing vasculature, is regulated by a complex interplay of anti and proangiogenic factors. We found that physiologic levels of circulating chromogranin A (CgA), a protein secreted by the neuroendocrine system, can inhibit angiogenesis in various in vitro and in vivo experimental models. Structure-activity studies showed that a functional anti-angiogenic site is located in the C-terminal region, whereas a latent anti-angiogenic site, activated by cleavage of Q76-K77 bond, is present in the N-terminal domain. Cleavage of CgA by thrombin abrogated its anti-angiogenic activity and generated fragments (lacking the C-terminal region) endowed of potent proangiogenic activity. Hematologic studies showed that biologically relevant levels of forms of full-length CgA and CgA1-76 (anti-angiogenic) and lower levels of fragments lacking the C-terminal region (proangiogenic) are present in circulation in healthy subjects. Blood coagulation caused, in a thrombin-dependent manner, almost complete conversion of CgA into fragments lacking the C-terminal region. These results suggest that the CgA-related circulating polypeptides form a balance of anti and proangiogenic factors tightly regulated by proteolysis. Thrombin-induced alteration of this balance could provide a novel mechanism for triggering angiogenesis in pathophysiologic conditions characterized by prothrombin activation.

Figure 1

Detection of CgA and CgA fragments by 76, 78, 439, 436/439, and 436/439+FRs-ELISAs in plasma samples obtained from healthy subjects. (A-D) Binding curves of recombinant CgA and CgA fragments obtained by 76, 78, 439, and 436/439+FRs-ELISA. Insets: schematic representation of the assays. (E) Circulating levels of natural CgA and CgA fragments, as measured with the indicated assays in plasma samples obtained from 6 healthy human subjects.

Figure 2

Full-length CgA inhibits spontaneous, bFGF and VEGF-induced angiogenesis. (A) Effect of the indicated doses of CgA1-439 and suramin on angiogenesis in the CAM assay. (Left) Stereomicroscope photographs of vessels recruited by the 6-mm (diameter) filter-paper disks (arrows; bar 1 mm). (Right) Quantitative evaluation of angiogenesis obtained with the indicated number of eggs (n). Blood vessels recruited by the disk, but not vessels that did not touch the filter, were counted in a blind manner by 2 observers. Circles correspond to vessel recruited by half-disk in each egg (box plots with median, interquartile, and 5%-95% values; **P < .01, ***P < .001; 2-tailed t test). (B) RAR assay. Effect of CgA1-439 on spontaneous, bFGF, or VEGF-induced angiogenesis. (Top panels) Bars represent the number of capillary-like structures emerging from the aorta rings treated as indicated, expressed as percentage of the untreated control (mean ± SD). The number of rings used is indicated in each panel (n). (Bottom panels) Microscopy photographs of aorta rings (white arrowheads) showing capillary-like structures (black arrows, 5× magnification, bar 150 μm; *P < .05, **P < .01, ***P < .001; 2-tailed t test).

Figure 3

The C-terminal domain of CgA is crucial for its anti-angiogenic activity. Effect of CgA fragments lacking the C-terminal region (CgA1-373, 1-400, and 1-409; top panels) or corresponding to the C-terminal region (CgA410-436, pGlu-serpinin, and CgA410-439; bottom panels) on bFGF and VEGF-induced angiogenesis in the RAR assay. Bars represent the number of capillary-like structures emerging from the aorta rings expressed as percentage of the untreated control (mean ± SD; *P < .05, **P < .01, ***P < .001; 2-tailed t test).

Figure 4

N-terminal fragments of CgA inhibit angiogenesis in vitro and in vivo. (A) Effect of N-terminal fragments of CgA (CgA1-78 and CgA1-76) on bFGF and VEGF-induced angiogenesis in the RAR assay. Bars represent the number of capillary-like structures emerging from the aorta rings expressed as percentage of the untreated control (mean ± SD). (B) Effect of CgA1-76 on vessel density and tumor growth in the murine WEHI-164 fibrosarcoma model. BALB/c mice (n = 6 per group) were treated intraperitoneally at the indicated time (arrows) after tumor implantation, with the indicated doses of CgA1-76. Tumors were excised and stained with anti-CD31 antibody (mAb MEC 13.3, BD Pharmingen) and AlexaFluor 546 goat anti–rat IgG (arrows, endothelial staining) and with 4,6-diamidino-2-phenylindole (DAPI; nuclear staining). Vessel density and tumor volumes are shown. Vessel density was quantified by counting the number of red spots (CD31⁺) in each field analyzed by fluorescence microscopy (5 fields/section, 3 sections/tumor, 6 tumors/group) using the ImageJ 1.47d software (National Institutes of Health). Each circle represents the average number of vessels/field/section (n = 18). Representative images of CD31 staining (corresponding to 60% area of original fields) are also shown (10× magnification; bar 200 μm). (C-D) Effect of CgA1-76 on tumor growth in the RMA lymphoma and TS/A adenocarcinoma models. Tumor-bearing mice (n = 8-14 per group) were treated intraperitoneally at the indicated time (arrows) after tumor implantation, with the indicated doses of CgA1-76. (A-C) *P < .05; **P < .01; ***P < .001. (A) Two-tailed t test; (B-C) 2-tailed Mann-Whitney test (treated vs untreated).

Figure 5

Thrombin cleaves the C-terminal region of CgA. (A) SDS-PAGE of recombinant CgA (6μM) before and after incubation with thrombin-sepharose (Thr-Seph) or sepharose alone (Seph). Thrombin (60 U, Sigma-Aldrich) was coupled to 200 μL of activated-CH-sepharose (GE Amersham), according to the manu-facturer's instructions. CgA (6μM in PBS) was mixed with the thrombin-CH-sepharose (1:10 suspension) and left to digest for 10 hours at 30°C under gentle agitation. The supernatant was then recovered and stored at −20°C until analysis. (B) Gel filtration chromatography of thrombin-digested CgA. Thrombin-digested CgA was loaded onto a Superdex 75 column and eluted with PBS. Fractions corresponding to the main peaks were collected and pooled (pools 1, 2, and 3). (C) 436/439 and 436/439+FRs-ELISAs of recombinant CgA, thrombin-digested CgA, pools 1, 2, and 3. (D) Molecular weight (Dalton) of fragments present in thrombin-digested CgA as measured by ESI-MS (Obs). The corresponding fragments and their expected molecular weight (Exp) are also shown. (E) Primary sequence of human CgA. Dibasic sites are indicated in bold. Arrows indicate the cleavage sites of thrombin as detected by ESI-MS.

Figure 6

The C-terminal region of CgA is cleaved by thrombin in blood during fibrin clot formation. (A) Analytical recovery of recombinant CgA in serum and plasma samples obtained from murine blood (n = 3) spiked with 10nM CgA, as measured by the indicated ELISAs. The analytical recovery of FRs was obtained by calculating the difference between 436/439+FRs and 436/439-ELISA. (B) Detection of natural circulating CgA in plasma (P) and serum (S) samples obtained from the same human subjects (n = 6), as measured by the indicated assays (mean ± SD; **P < .01, ***P < .001; 2-tailed t test).

Figure 7

Thrombin abrogates the anti-angiogenic activity of full-length CgA and generates proangiogenic fragments. (A) Effect of CgA, thrombin-digested CgA, pool 1, and recombinant CgA1-373 on spontaneous angiogenesis in the RAR assay. Bars represent the number of capillary-like structures emerging from the aorta rings, treated as indicated, expressed as percentage of the untreated control. The number of aorta rings tested is indicated (n). Open circles correspond to number of capillaries sprouting from each aortic ring (box plots with median, interquartile, and 5%-95% values). (Right panels) Microphotographs of aortic rings (white arrowheads) and capillaries (black arrows) obtained after 6 days of incubation with CgA or thrombin-digested CgA (5× magnification, bar, 250 μm). (B) Effect of CgA1-373 on the secretion of bFGF from endothelial cells. bFGF levels in the supernatant of endothelial cells, incubated for 1 hour with the indicated compounds, was analyzed by ELISA. Circles correspond to bFGF levels in each cell culture (box plots with median, interquartile, and 5-95 percentile values (*P < .05, **P < .01, ***P < .001; 2-tailed t test).