Human cytomegalovirus secretome contains factors that induce angiogenesis and wound healing

Abstract

Human cytomegalovirus (HCMV) is implicated in the acceleration of a number of vascular diseases including transplant vascular sclerosis (TVS), the lesion associated with chronic rejection (CR) of solid organ transplants. Although the virus persists in the allograft throughout the course of disease, few cells are directly infected by CMV. This observation is in contrast to the global effects that CMV has on the acceleration of TVS/CR, suggesting that CMV infection indirectly promotes the vascular disease process. Recent transcriptome analysis of CMV-infected heart allografts indicates that the virus induces cytokines and growth factors associated with angiogenesis (AG) and wound healing (WH), suggesting that CMV may accelerate TVS/CR through the induction and secretion of AG/WH factors from infected cells. We analyzed virus-free supernatants from HCMV-infected cells (HCMV secretomes) for growth factors, by mass spectrometry and immunoassays, and found that the HCMV secretome contains over 1,000 cellular proteins, many of which are involved in AG/WH. Importantly, functional assays demonstrated that CMV but not herpes simplex virus secretomes not only induce AG/WH but also promote neovessel stabilization and endothelial cell survival for 2 weeks. These findings suggest that CMV acceleration of TVS occurs through virus-induced growth factors and cytokines in the CMV secretome.

FIG. 1.

HCMV supernatants induce angiogenesis at 24 h. Primary HUVEC were serum starved for 3 h in SF-SFM before they were harvested and resuspended in this medium (10⁵ cells/ml). Cells (10⁴ cells/well) were introduced into 24-well plates containing polymerized plugs of growth factor-reduced Matrigel in the presence of 300 μl of control and test supernatants. Test supernatants were concentrated secretomes generated from mock- and HCMV strain AD169-infected cells, using Amicon Ultra-4 filters. Control supernatants included those cultured in SF-SFM (10× concentrated) and 1× complete SFM. Each supernatant was tested in quadruplicate. (A) Quantitative measures of angiogenesis consisted of the numbers of lumens and branch points. (B) Representative examples of each culture condition are shown as a low-power image (magnification, ×10). (C) High-power images emphasize differences in vessel integrity between conditions (magnification, ×20). The data presented in this figure are a representation of three individual experiments.

FIG. 2.

HCMV supernatants stabilize tubule formation at 48 h and at 2 weeks. An angiogenesis assay was performed using primary HUVEC, as previously described, at 48 h postplating. (A) Low-power (magnification, ×10) images show samples grown in the presence of control (SF-SFM alone or complete SFM containing human serum and ECGS) and test (mock or HCMV AD169) secretomes. (B) High-power (magnification, ×20) images under complete SFM and HCMV AD169 secretome conditions are shown. (C) Tubule survival is shown after 2 weeks on Matrigel in the presence of complete SFM or HCMV AD169 secretome (magnification, ×20). The data presented in this figure are a representation of three individual experiments.

FIG. 3.

Wound healing activity of the HCMV secretome. EC were grown to confluence on 8W1E ECIS arrays and exposed to test supernatants before they were electrically wounded. (A) Low-magnification images of ECIS electrodes containing unwounded and wounded EC grown in the presence of complete SFM at the time of wounding (0 h) and at 3 and 20 h postwounding. (B) Wound healing, as indicated by increasing resistance, is plotted as a function of time. Healing traces for duplicate wells are shown for the mock secretome (Mock, green trace) and HCMV strain AD169 secretome (HCMV, red trace). Also shown is the wound healing activity of secretomes from HF infected with UV-inactivated HCMV (UV-HCMV, pink trace), or infected (HCMV/Foscarnet, orange trace) or mock-infected (Mock/Foscarnet, light blue trace) in the presence of foscarnet. Control traces include a negative control (SF-SFM, yellow), a positive control (Complete SFM, dark blue), and a constant resistance trace measured from a confluent unwounded monolayer (Unwounded, black). Cells exposed to the HCMV secretome show wound repopulation within 8 to 10 h, whereas cells exposed to the mock secretome repopulate the wound inefficiently. Secretomes from UV-inactivated HCMV or HCMV plus foscarnet similarly fail to repopulate the wound, indicating that infectious virus and late gene expression are required for the production of wound healing factors. (C) Wound healing in response to the mock secretome (Mock, green), the HCMV AD169 secretome (HCMV, red), and the HSV-1 secretome (HSV-1, yellow). Also shown is the wound healing activity of secretomes from HF infected with UV-inactivated HCMV AD169 (UV-HCMV, light blue) and negative control (SF-SFM, pink), a positive control (Complete SFM, dark blue), and a constant resistance trace measured from a confluent unwounded monolayer (Unwounded, black). Cells exposed to the HCMV secretome show wound repair within 6 to 10 h, whereas cells exposed to the HSV-1 or mock secretomes repopulate the wound inefficiently, indicating that the production of wound healing factors is specific for HCMV infections. The data presented in this figure are a representation of three individual experiments.