Myxoma and vaccinia viruses bind differentially to human leukocytes

Abstract

Myxoma virus (MYXV) and vaccinia virus (VACV), two distinct members of the family Poxviridae, are both currently being developed as oncolytic virotherapeutic agents. Recent studies have demonstrated that ex vivo treatment with MYXV can selectively recognize and kill contaminating cancerous cells from autologous bone marrow transplants without perturbing the engraftment of normal CD34(+) hematopoietic stem and progenitor cells. However, the mechanism(s) by which MYXV specifically recognizes and eliminates the cancer cells in the autografts is not understood. While little is known about the cellular attachment factor(s) exploited by MYXV for entry into any target cells, VACV has been shown to utilize cell surface glycosaminoglycans such as heparan sulfate (HS), the extracellular matrix protein laminin, and/or integrin β1. We have constructed MYXV and VACV virions tagged with the Venus fluorescent protein and compared their characteristics of binding to various human cancer cell lines as well as to primary human leukocytes. We report that the binding of MYXV or VACV to some adherent cell lines could be partially inhibited by heparin, but laminin blocked only VACV binding. In contrast to cultured fibroblasts, the binding of MYXV and VACV to a wide spectrum of primary human leukocytes could not be competed by either HS or laminin. Additionally, MYXV and VACV exhibited very different binding characteristics against certain select human leukocytes, suggesting that the two poxviruses utilize different cell surface determinants for the attachment to these cells. These results indicate that VACV and MYXV can exhibit very different oncolytic tropisms against some cancerous human leukocytes.

Fig 1

vMyx-Venus/M093 produces fluorescent foci that are similar in size to those produced by the parental viruses. (A) Construction of a recombinant vMyx-Venus/M093. M093 fused with Venus fluorescent protein at the amino terminus (Venus/M093) was generated by inserting the coding sequence of Venus in frame immediately following the start codon of the M093L gene. The arrow indicates the direction of transcription of M093L. aa, amino acids. (B) Characterization of fluorescent foci formed by vMyx-Venus/M093. BSC-40 cell monolayers were infected with the indicated viruses at 37°C for 1 h. At 1 h after infection, inoculum was removed and cell monolayers were overlaid with liquid medium. At 3 days postinfection, fluorescence and phase-contrast images of foci were captured.

Fig 2

Venus/M093 fusion protein is incorporated into MYXV virions. (A) Temporal synthesis of Venus/M093 protein. BSC-40 cells were mock infected or infected with vMyx-Venus/M093 at an MOI of 10.0 in the absence or presence of cytosine arabinoside (AraC), indicated by +AraC, for 1 h at 37°C. After adsorption, the inoculum was removed and cells were washed. At the indicated time points, cells were harvested and lysed in radioimmunoprecipitation assay buffer. Clarified cell lysates were subjected to SDS-PAGE, and Western blotting was performed sequentially using the indicated antibodies, followed by an appropriate horseradish peroxidase-conjugated secondary antibody. (B) Venus/M093 fusion protein is incorporated into MYXV IMVs. IMVs purified through a 36% sucrose cushion were treated with NP-40 lysis buffer in the presence or absence of DTT. The detergent-insoluble fraction (I) was separated from the detergent-soluble fraction (S) or left unfractionated as total (T) input protein. The fractions were subjected to SDS-PAGE for Western blot analyses using the antibodies indicated in panel A.

Fig 3

Venus-tagged MYXV and VACV virions visualized by fluorescence microscopy. (A) Fluorescence microscopy of cells infected with vMyx-Venus/M093 or vVac-Venus/A4. BSC-40 cells grown on glass coverslips were infected with vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 0.1. At 24 h after infection, cells were fixed and glass coverslips were placed on mounting medium containing DAPI. Cells were visualized using a ×63 water-corrected immersion objective on a Leica TCS SP2 confocal microscope. (B) Binding of Venus-tagged MYXV or VACV to BSC-40 cells. BSC-40 cells grown on glass coverslips were washed with ice-cold medium and prechilled on ice. Cells were incubated with vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 10.0 on ice for 1 h. After adsorption, unbound virions were removed by washing cell monolayers with ice-cold medium. Cells were fixed, mounted on DAPI containing mounting medium, and visualized as described for panel A. Yellow, Venus-labeled virions; blue, DNA in nuclei and viral factory. Bars, 10 μm (A) and 20 μm (B).

Fig 4

Venus-tagged MYXV binds to human myeloma cell lines more efficiently than VACV. (A) Binding of vMyx-Venus/M093 or vVac-Venus/A4 to control BSC-40 and HeLa cells compared to U266 and HuNS1 human multiple myeloma cells. Purified vMyx-Venus/M093 or vVac-Venus/A4 was added to prechilled cells at an MOI of 20.0 and incubated on ice for 1 h. After virion binding, cells were washed extensively, fixed with 2% paraformaldehyde, and analyzed by flow cytometry. The percentages of Venus-positive HeLa, U266, and HuNS1 cells detectable by flow cytometry were normalized to the percentage of BSC-40 cells detected. Statistical analyses between the two viruses for HeLa, U266, and HuNS1 cell lines were performed using the Student t test. (B) Venus-tagged MYXV binding to all tested cells, except HuNS1, is more sensitive to inhibition by soluble heparin than VACV binding. Cells were mock treated or treated with Hep I at 37°C for 30 min, washed, and chilled on ice. Purified vMyx-Venus/M093 or vVac-Venus/A4 was mock treated (−) or pretreated with soluble HP for 1 h. Virions were bound to prechilled cells at an MOI of 20.0 on ice for 1 h. After virion binding, cells were washed extensively and fixed with 2% paraformaldehyde. The relative levels of virion binding to cells were assessed by flow cytometry. (C) Soluble laminin blocks the binding of Venus-tagged VACV, but not MYXV, to HeLa cells. HeLa cells were incubated with mock-treated (−) or laminin-treated vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 20.0 on ice for 1 h. After binding, cells were processed as described for panel A and bound virions were detected by flow cytometry. The mean fluorescence intensity was determined and normalized to that for the untreated corresponding virus. Averages of two independent experiments are shown, and the error bars are plotted. Statistical analyses between the untreated and HP- or LN-treated groups for each virus were performed using the Student t test. * and **, P ≤ 0.05 and P ≤ 0.005, respectively, which are considered significant; a, P > 0.05, which is considered insignificant.

Fig 5

Venus-tagged MYXV and VACV bind and infect human CCRF-CEM and Jurkat T lymphoblastoid cells differently. (A) Histograms of the binding of Venus-tagged MYXV or VACV to CCRF-CEM and Jurkat cells at various MOIs. Prechilled CCRF-CEM or Jurkat cells were incubated with vMyx-Venus/M093 or vVac-Venus/A4 at the indicated MOIs for 1 h on ice. After binding, cells were washed and the amount of virions bound to the cells was determined by flow cytometry. Representative histograms of virus binding are shown. Mock infections are shown in filled gray histograms. (B) Differential binding of Venus-tagged MYXV versus VACV to Jurkat cells. Virion binding was performed as described for panel A. The mean fluorescence intensity was determined, and averages of two independent experiments are plotted. The error bars for each plot are shown. (C) Infection of CCRF-CEM and Jurkat cells with Venus-tagged MYXV or VACV. CCRF-CEM and Jurkat cells were infected with vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 20.0 at 37°C for 1 h. After adsorption, unbound virions were removed by washing and cells were incubated at 37°C. At 4 and 24 h postinfection, cells were collected, fixed, and analyzed by flow cytometry. The percentage of Venus-positive cells was determined, and averages of two independent experiments are plotted. The error bars for each plot are shown. Statistical analyses between the two viruses for each cell line were performed using the Student t test. * and **, P ≤ 0.05 and P ≤ 0.005, respectively, which are considered significant; a, P > 0.05, which is considered insignificant.

Fig 6

Binding of Venus-tagged MYXV, but not VACV, to human CCRF-CEM T lymphoblastoid cells is heparan sulfate independent. (A) Soluble heparin does not block Venus-tagged MYXV binding to CCRF-CEM cells. Venus-tagged MYXV or VACV that was mock treated (−) or pretreated with soluble HP was bound to CCRF-CEM or Jurkat cells at an MOI of 100.0 on ice for 1 h. After binding, cells were washed and virion binding was determined by flow cytometry. The mean fluorescence intensity was determined, and averages of two independent experiments are plotted. The error bars for each plot are shown. Statistical analyses between the untreated and HP- or LN-treated groups for each virus were performed using the Student t test. (B) Competitive binding assay. BSC-40, CCRF-CEM, or Jurkat cells were mock adsorbed or incubated with untagged vMyx (Lausanne) or vVac (Western Reserve) at an MOI of 100.0 on ice for 1 h. After binding with primary viruses, cells were washed twice and incubated with vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 10.0 for BSC-40 cells or 100.0 for CCRF-CEM and Jurkat cells on ice for 1 h. Unbound virions were removed by washing, and cells were fixed with 2% paraformaldehyde. The amount of vMyx-Venus/M093 or vVac-Venus/A4 bound to the cells was determined by flow cytometry. The mean fluorescence intensity was determined and normalized to that for the untreated corresponding virus. Averages of two independent experiments are shown, and the error bars are plotted. Statistical analyses between mock-adsorbed cells and cells incubated with untagged vMyx or vVac were performed for each cell line using the Student t test. * and **, P ≤ 0.05 and P ≤ 0.005, respectively, which are considered significant; a, P > 0.05, which is considered insignificant.

Fig 7

Binding of Venus-tagged MYXV or VACV to primary human leukocytes is not mediated by either cell surface heparan sulfate or the extracellular matrix protein laminin. Buffy coat preparations of fresh primary human leukocytes from healthy donors were incubated with mock-, HP-, or LN-treated vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 10.0 on ice for 1 h. (A) For binding, cells were washed after incubation with virus on ice. (B) For infection, cells were further incubated at 37°C for 24 h after virion adsorption. After binding or infection, cells were stained with phycoerythrin-conjugated anti-CD45 and allophycocyanin-conjugated anti-CD14, anti-CD15, anti-CD19, or anti-CD3 lineage markers. The percentage of Venus-positive cells within each population was determined by flow cytometry. A representative of four independent experiments is shown. (C and D) Summary of the types of primary human cells from donor PBMCs susceptible to MYXV or VACV binding and infection. Average percentages of different cell populations bound (C) or infected (D) by vMyx-Venus/M093 or vVac-Venus/A4 after 1 h at 0°C or 24 h at 37°C, respectively, as detected by flow cytometry under various treatments, were determined. Averages of four independent samples and standard deviations are shown. Statistical analyses between the untreated and HP- or LN-treated groups for each virus were performed using the Student t test. * and **, P ≤ 0.05 and P ≤ 0.005, respectively, which are considered significant; a, P > 0.05, which is considered insignificant.

Fig 8

Venus-tagged MYXV or VACV differentially bind CD138⁺ multiple myeloma cells in primary patient bone marrow samples. Human leukocytes were enriched from the primary patient bone marrow samples through a Ficoll gradient. vMyx-Venus/M093 or vVac-Venus/A4 was bound to leukocytes at an MOI of 10.0 on ice for 1 h. After binding for 1 h on ice or infection for 24 h at 37°C, cells were stained with allophycocyanin-conjugated anti-CD138 (A) or anti-CD34 (B), fixed, and analyzed by flow cytometry. Virion binding was determined by flow cytometry.

Fig 9

Venus-tagged MYXV uniquely attaches to HuNS1 cells that have little or no surface integrin β1. (A) Differential susceptibility of human multiple myeloma cell lines U266 and HuNS1 to Venus-tagged MYXV versus VACV. U266 or HuNS1 cells were infected with vMyx-Venus/M093 or vVac-Venus/A4 at an MOI of 20.0 at 37°C for 1 h. Unbound virions were removed by washing after adsorption, and cells were further incubated at 37°C. At 4 or 24 h postinfection, cells were fixed and the percentage of virus-infected cells was determined by flow cytometry. Averages of two independent experiments are shown, and the error bars are plotted. Statistical analyses between the two viruses for each cell line were performed using the Student t test. **, P ≤ 0.005, which is considered significant. (B) HuNS1 cells express very little to no integrin β1 on their surface. HeLa, U266, and HuNS1 cells were stained with allophycocyanin-conjugated anti-CD29 at 4°C or left unstained. Unbound antibodies were removed by washing before they were analyzed by flow cytometry. Expression of integrin β1 was examined by measuring the allophycocyanin fluorescence (histograms shown in red). Unstained cells are shown in filled gray histograms.