Introduction

Replication competent viruses have been proposed as anticancer agents and a variety of attenuated viruses have been tested for tumor-specific replication and oncolytic activity.¹ The Edmonston vaccine strain of measles virus (MV-Edm) has emerged as one of the more promising candidates for cancer virotherapy,² showing potent and selective oncolytic activity in a preclinical model of lymphoma, multiple myeloma (MM), ovarian cancer, and glioblastoma.^{3, 4, 5, 6} The virus induces formation of giant syncytia by cell–cell fusion, leading to subsequent killing of transformed cells. Repeated intratumoral or systemic intravenous (i.v.) injection of MV-Edm caused complete regression of human xenografts in immunocompromised mice. Recently, clinical activity has been demonstrated after intratumoral administration in patients with cutaneous T-cell lymphoma.⁷ A genetically engineered MV expressing the human thyroidal sodium iodide symporter (MV-NIS) belongs to a new generation of viral vectors combining virotherapy with radiotherapy.⁸ Selective accumulation of radioiodine in MV-NIS-infected tumor cells enhanced the oncolytic activity through radiation-mediated killing of non-infected neighboring cells. This is a new and highly promising approach to the treatment of radiosensitive malignancies such as MM.

A major theoretical impediment to systemic application of MV is pre-existing antiviral immunity. Almost all individuals have circulating anti-measles antibodies and T lymphocytes developed as a result of immunization or natural infection. Rapid neutralization by protective antibodies in the plasma is likely to compromise any systemic application of oncolytic MV by the i.v. route.

MV belongs to the Paramyxoviridae family of viruses, which are characterized by a lipoprotein envelope and a negative-stranded RNA genome.⁹ Attachment and penetration of MV into its host cells is mediated by the two surface glycoproteins: hemagglutinin (H) and fusion protein.^{9, 10} Two cell surface molecules, CD46 and SLAM (signaling lymphocytic activation molecule or CD150), have been identified as MV receptors.^{11, 12, 13} The measles-protective humoral immune response is directed against both H and fusion surface glycoproteins.^{14, 15} H protein-specific antibodies block virus attachment to either of the cellular receptors. Anti-fusion antibodies do not interfere with virus attachment but inhibit subsequent fusion and entry. These antibodies can block cell-to-cell spread of MV infection, which occurs by fusion between infected and neighboring cells.¹⁶ Also, direct virus neutralization (VN) is not the only mechanism of antibody protection against measles. Classical complement pathway activation and antibody-dependent cytotoxicity may also contribute to VN and prevention or recovery from MV infection.^{14, 17, 18, 19} Pre-existing anti-measles antibodies are therefore considered the major barrier for i.v.-injected oncolytic MVs. Also, cytotoxic CD8⁺ T-cell-mediated immunity is absolutely necessary for viral clearance and complete recovery after MV infection.²⁰ However, cellular immunity is often defective in cancer patients, especially in the case of hematological malignancies, and can be further suppressed, if necessary, with a variety of immunosuppressive agents. Here, we introduce a strategy whereby oncolytic MVs are delivered using infected cells as carriers. Infected cells could efficiently deliver MV to tumor cells in the presence of neutralizing titers of anti-measles antibodies. In vivo transfer of infection by heterofusion of cellular carriers into the tumor parenchyma was demonstrated in a diffuse metastatic lymphoma model and in intraperitoneal (i.p.) xenograft models in immunocompromised mice. Repeated injection of infected cells completely protected animals from ascitic tumors in an ovarian cancer model.

Results

In vitro transfer of measles infection by heterofusion

Infected cells of different lineages successfully transferred infection to tumor cells (Supplementary Figures S1 and S2). Optimal infection was achieved at 37°C and at an multiplicity of infection (MOI) of 0.5 in a small-volume Opti-MEM medium, which resulted in >90% infection of delivery cells. As shown in Supplementary Figure S1, green fluorescent protein (GFP)-expressing monocytic U-937 cells infected with MV-Edm successfully transferred infection by heterofusion to lymphoma, MM, and ovarian cancer target cells. Outgrown endothelial cells (OECs) also delivered infection to the targeted cells by heterofusion (Supplementary Figure S2). In all experiments 100 l of the final wash from the virus-infected delivery cells was inoculated on Vero cells and no infection from cell-free virus was detected after 3–5 days of incubation.

Efficiency of heterofusion in the presence of virus-neutralizing antibodies in vitro

Infected delivery cells were overlaid on red fluorescent protein (RFP)—RFP⁺ SKOV3ip.1 cells and cultured for 3 days in the continuous presence of 1:100 neutralizing human serum (Figure 1). Heterofusion triggered by H and fusion proteins expressed on the surface of the infected delivery cells was not inhibited by neutralizing serum antibodies. In contrast to infected cells, cell-free virions were completely neutralized by the same neutralizing serum at a dilution 1:320. Also, the anti-measles antibodies did not block the cell-to-cell spread of infection and giant syncytia formation (Figure 1). In a separate experiment we tested the efficiency of cell–cell heterofusion in the presence of serial dilutions of a highly protective anti-measles human serum. At a dilution of 1:320, which was completely neutralizing for cell-free virus, the number of syncytia seen after overlaying with infected delivery cells was not diminished when compared with the control without antibodies (Figure 1e). Even at a serum dilution of 1:20, infection was successfully transferred, although the efficiency of heterofusion was decreased by 92–96%. In contrast to the absolute number of syncytia, the size of the giant tumor cell syncytia that formed after the initial heterofusion event was not significantly reduced (Figure 1f) indicating that homotypic fusion between tumor cells is more resistant to antibody inhibition than heterofusion between tumor cells and infected delivery cells.

Figure 1.

Heterofusion of measles-infected carrier cells is resistant to antibody neutralization. Transfer of MV-GFP infection (MOI=0.2) from Ug-C811 (U-937 green clone) cells to RFP⁺ovarian cancer cell clones SR-B2 (a) and SR-A3 (b) in the presence of 1:100 diluted highly neutralizing human serum. (c and d) Similar results obtained with infected human OECs (confocal microscopy). As in the previous experiments, cell carriers were treated with antibody and complement to prevent infection by cell-free virions or cell-surface virions. Number of GFP⁺syncitia per well was counted on days 2 and 3 using fluorescent and confocal microscopy. The samples were run in 4–6 wells and the experiment was repeated twice. In (e) infected U-937 were overlaid on SR-B2 cells and cultured in the presence of different dilutions of highly neutralizing serum. Heterofusion was 16-32 times more resistant to neutralization than cell-free virions as determined by the VN test. MV-GFP-infected (MOI=0.5) U-937 (10³) were overlaid on SR-B2 cells and cultured in the presence of different dilutions of highly neutralizing human serum. Control samples of infected delivery cells were cultured in the presence of fusion inhibitory peptide to calculate the number of infected cells by flow cytometry. The efficiency of infection in U-937 was 68%. The VN test with 10³ PFU of cell-free MV-GFP was used as control. At a dilution of 1:20 the same serum antibodies had no effect on the size of syncytia in the target ovarian cancer cells overlaid with (f) MV-GFP-infected U-937 (fluorescent microscopy). MV-GFP absorbed on the surface of J774A.1 mouse macrophages was able to infect Vero cells after overlaying (left) but was completely neutralized (right) by (g) MAb CL48 and complement treatment.

Full figure and legend (312K)

MV nonspecifically absorbed on the cell surface cannot transfer infection in the presence of neutralizing antibodies

MV particles absorbed by a nonspecific mechanism on the surface of mouse J774A.1 macrophages can efficiently deliver infection to tumor cells. In contrast, preincubation with monoclonal antibody (MAb) CL48 (specific for MV H protein) and guinea pig complement before overlaying on Vero monolayer led to complete neutralization of cell surface-associated virus and no infection was observed (Figure 1g).

MV-infected delivery cells efficiently fuse with tumor cells in vivo

We next demonstrated that mixed syncytia could be formed through heterofusion in vivo in an i.p. model of human MM in immunocompromised animals. Mice were injected i.p. with 10⁷ GFP-expressing KAS6/1 MM cells. On days 3 and 7 after cell implantation, 2 10⁶ MV-NIS-infected RFP-expressing U-937 cells were administered by i.p injection. At 48–72 h after injection of the U-937 cells, double-positive (green–red) mixed syncytia were detected in peritoneal lavage fluid and were counted by flow cytometry and fluorescent microscopy (Figure 2). The efficiency of in vivo heterofusion was calculated to be between 0.3 and 5.87% of the recovered MM xenograft cells and varied between individual animals.

Figure 2.

In vivo heterofusion in the peritoneal cavity. MV-NIS-infected UR-D7 cells (RFP⁺clone of U-937 monocytes) were administered i.p. to SCID mice bearing i.p. GFP-expressing KAS6/1 myeloma xenografts. The results show cells isolated from three representative animals (in rows) by peritoneal lavage 2 days after injection of MV delivery cells. The percentage of double-positive (GFP/RFP) cells varied between 0.31 and 5.87% in individual animals (five per group).

Full figure and legend (97K)

MV-infected delivery cells transfer infection to human tumor xenografts in vivo

Systemic delivery in a disseminated lymphoma xenograft model

In a pilot experiment with Raji xenografts in non-obese diabetic/severe combined immunodeficiency disease (SCID) mice, we compared the efficiency of virus delivery using cell carriers injected 3 or 18 h after infection with MV. Lymphoma cells (10⁶) were engrafted by the i.v. route and 2 weeks later 10⁶ MV-GFP (GFP-expressing MV-Edm)-infected U-937 cell carriers were injected by the same route. After five applications, GFP-positive Raji tumor nodules were detected in all animals that had been injected with 18-h-infected U-937 cells. These data demonstrated that injection of cells 18 h after MV infection was more efficient than earlier injection 3 h post-infection (not shown).

In the next experiment non-obese diabetic/SCID mice were engrafted 2 10⁶ RR-E63 (RFP⁺ Raji) cells and were treated by single or three times repeated i.v. injections of MV-infected U-937 carriers. Efficiency of infection of delivery cells was determined to be between 67 and 86%. Double-positive (GFP and RFP) cells/syncytia were found in tumor foci and live MV was recovered from all animals injected with U-937 carriers by overlaying explanted tumor cells on Vero monolayers (Figure 3a and Supplementary Movie S1). In a parallel experiment MV infection was successfully delivered (in 80% of mice) to Raji tumors by three i.v. applications of 2 10⁶ MV-GFP-infected peripheral blood mononuclear cells (PBMCs) (Figure 3b). The efficiency of PBMC's infection was 10.7, 11.5, and 37.0%, respectively. In contrast, no GFP-positive tumor cells were identified after i.v. administration of MV-GFP-infected (efficiency 49–63%) human OECs. In all mice examined, massive meningeal infiltration by Raji cells overlying the occipital lobe of the brain was observed, suggesting that paralysis and death of the engrafted mice were due to central nervous system disease rather than bone marrow infiltration and destruction of the vertebrae. RFP⁺ lymphoma cells were also found in bone marrow and in the circulation (1–2 10⁴/ml) but generally not in other internal organs. In addition to its recovery from the infected tumor lesions, abundant live MV could also be recovered from the lungs on day 1 or day 2 after injection of MV carriers. In contrast, very little virus was recovered from liver or spleen samples (<10 plaque-forming units (PFU) per spleen).

Figure 3.

Systemic and i.p. in vivo delivery of MV-GFP infection using different cell carriers. MV-GFP-infected cells (U-937 or PBMCs) were administered i.v. to mice bearing systemic RR-E63 (RFP⁺Raji cells) lymphoma xenografts. GFP-positive syncytia were seen in meningeal infiltrates 3 days after i.v. injection of infected (a) U-937 cells or (b) PBMCs demonstrated by confocal microscopy. (c) Large multinucleated syncytia were seen in an i.p. SR-B2 (RFP⁺SKOV3ip.1 clone) ovarian cancer tumor model 3 days after injection of 2 10⁶ MV-GFP-infected U-937 cells. Massive MV infection in the tumors was observed in all mice engrafted i.p. with HUH-7 hepatocellular carcinoma cells 3 days after i.p. injection of the carrier cells. Green fluorescent lesions (white arrows) in representative animals from groups treated with MV-GFP-infected (d) U-937 cells or (e) OEC. GFP-positive tumor lesions were also observed at (f) the liver in all mice and large syncytia were demonstrated by (g) confocal microscopy. Samples from the GFP-positive tumors were overlaid on Vero monolayers and MV-GFP was isolated.

Full figure and legend (253K)

Regional delivery to i.p. xenografts

A single i.p. injection of 2 10⁶ MV-GFP-infected U-937 or OEC cells successfully transferred infection to i.p. ovarian and hepatocellular cancer xenografts. Large green fluorescent lesions were easily identified 3 days after the cell delivery injection in all mice with i.p. HUH-7 hepatoma or SR-B2 (RFP⁺SKOV3ip.1) ovarian cancer tumors (Figure 3c–g and Supplementary Movie S2).

Therapeutic activity of MV-GFP-infected cells in an ovarian cancer model

Five repeated therapeutic injections of MV-GFP-infected UR-D7 (RFP⁺U-937) cells significantly improved survival of ovarian cancer engrafted nude mice (experiment A). The therapy was initiated on day 10—the mid-time between tumor engraftment and ascites development. As shown in Figure 4a, the median survivals of mice treated with MV-GFP-infected U-937 cells were more than twice prolonged compared with control mice. However, we did not observe complete protection and treated mice died between 70 and 94 days because of the complications from ascites or subcutaneous tumors at the injection site (Table 1).

Figure 4.

Therapeutic potency of MV-GFP-infected carrier cells in an ovarian cancer model. Nude mice were engrafted i.p. with SKOV3ip.1 cells and treatment was started 10 days later using five repeated i.p. injections of (a) MV-GFP-infected U-937 (RFP⁺clone UR-D7) cells. Median survival was significantly (P<0.001) prolonged for both groups of carrier-cell-treated animals (UR-D7>70 days) compared with the controls (median survival, 24 days). (b) The efficacy of a single injection of cell carriers was compared with multiple applications in SCID mice engrafted i.p. with SKOV3ip.1 tumors. The therapy was started on day 7 using infected UR-D7 monocytic cell carriers. Control mice received non-infected delivery cells on day 7. A single therapeutic injection significantly (P<0.001) improved animal survival more than twice (median survival, 60 days) compared with the control group (median survival, 26 days). All mice treated with repeated injections of MV-infected carrier cells survived without symptoms or site injection tumors for more than 80 days. Ten mice per group were used in both experiments.

Full figure and legend (59K)

Table 1 - Efficiency of MV infection in UR-D7 delivery cells and tumor growth in cell-carrier-treated animals.

Full table

We next compared the efficacy of single versus repeated i.p. injections of 10⁶ MV-GFP-infected monocytic cells using an orthotopic model of ovarian cancer in SCID mice (experiment B). To account for any therapeutic activity of the monocytic cells (e.g., cytokine release), we injected control mice with 10⁶ non-infected UR-D7 cells. All of the control animals developed ascites and died between days 23 and 27 post-tumor engraftment. In contrast, mice treated with single or multiple injections of MV-GFP-infected UR-D7 cells showed greatly prolonged survival. Multiple injections were more effective than a single treatment (Figure 4b). All of the animals in this group were completely protected from relapse of peritoneal tumor and ascites (Table 1). Half of them survived more than 4 months without development of subcutaneous injection site tumors.

Measles-infected delivery cells evade the antibody response

When MV-GFP was administered by i.p. injection in an HUH-7 peritoneal dissemination model of hepatocellular cancer, bright green fluorescent tumor lesions were seen 2–3 days post-treatment. However, when the virus was preincubated in 1:10 diluted neutralizing serum for only 5 min, its infectivity in VN test was reduced by >99.6% (not shown) and infection at the tumor site was completely abolished (Figure 5c–f). When tumors from these MV-GFP-treated animals were harvested, homogenized, and overlaid on Vero cells, virus was identified in two of five animals (only one and three syncytia, respectively, from the whole tumor). In contrast, MV-GFP-infected OEC and U-937 cells transferred infection efficiently in the presence of neutralizing antibodies. Massive infection with large green fluorescent lesions and large syncytia (>500 m) were detected by body fluorescent imaging and confocal microscopy, respectively and MV-GFP was easily recovered by Vero cell overlay (Figure 5g–h). A similar experiment was performed using the Raji model of disseminated lymphoma. A single injection of 10⁶ MV-GFP-infected U-937 cells successfully delivered the virus to multiple tumor sites in four of five passively immunized animals. Double (RFP and GFP)-positive infected Raji cells were identified in bone marrow and menigeal infiltrates by confocal microscopy (Figure 5b). Live MV was recovered from the tumor lesions in the central nervous system and in bones. Serum samples were collected from animals passively protected against MV and all of them were shown to be fully neutralizing for MV-GFP at a dilution of 1:4 (one animal) or 1:8 (four mice). Only two of five animals had identifiable green fluorescent lesions in their tumors on day 3 after a single dose of infected PBMCs (Figure 5a).

Figure 5.

MV-infected cells, but not naked MVs, can bypass measles neutralizing antibodies in i.v. and i.p. delivery models. Using the systemic Raji (RFP⁺RR-E63 cells) tumor model, (a) MV-GFP-infected PBMCs or (b) U-937 successfully delivered the virus to RR-E63 lesions in the central nerve system in the presence of 1:4–1:8 neutralizing human antibodies in the plasma. In the presence of neutralizing antibodies, (c) MV-GFP-infected OEC, or (d) U-937 cells successfully delivered infection to i.p. HUH-7 tumors. In contrast, (e) cell-free MV was completely neutralized and no GFP⁺ was observed compared with (f) control, cell-free MV without antibodies. (g) Large GFP-positive tumor lesions and syncytia were observed in all animals treated with cell-associated MV. (h) MV-GFP could be recovered from all GFP-positive lesions after overlaying the tumor samples on Vero monolayers.

Full figure and legend (346K)

Discussion

The immune response against measles is complex, with innate and acquired immune mechanisms contributing differentially at successive stages after infection.⁹ Innate host defense contributes primarily to early control of infection during the incubation period. Specific cellular and humoral responses are detectable by the time that measles becomes clinically apparent and are required for complete recovery and life-long immunity against reinfection.^{20, 21, 22, 23, 24} Our data showed that antibodies neutralize MV quickly and irreversibly, which underscores the concern that systemically administered MV particles may not survive long enough in the bloodstream to reach the tumor site. Viremia during natural measles infection is exclusively cell-associated, such that live MV can be isolated only from infected PBMC and not as cell-free virions from serum samples.^{14, 25} In addition to antibodies, innate immune mechanisms such as the alternative complement pathway may contribute to MV neutralization in the plasma.^{26, 27} Thus, cell-associated viremia may be a necessary "strategy" for MV to avoid neutralization and to spread infection by cell–cell fusion. Our data show that transfer of infection by cell-to-cell fusion in monolayers is much more resistant to antibody neutralization than naked viruses. This finding is consistent with the published observation that passive immunization with antibodies could not block regression of subcutaneous lymphoma xenografts in mice after intratumoral injection of oncolytic MV.³ However, there are no published data addressing the question of whether heterofusion between measles-infected and uninfected cells is resistant to antibody neutralization.

Monocytes and macrophages are a primary target in natural measles infections, and monocytes have been identified as the main cellular carrier of MV during viremia.^{9, 28, 29, 30, 31} These insights into measles pathogenesis pointed to a new strategy for systemic delivery of oncolytic MV using infected cells instead of purified naked virions. Our in vitro studies clearly demonstrated the efficacy of cell–cell spread of infection by heterofusion between various donor and recipient cells. Direct cell-to-cell transfer of infectivity by monocytic U-937 cells was 16–32 times more resistant to antibody neutralization than infectivity transfer by cell-free viruses. Thus, once infection is successfully transferred to the tumor, it is expected that antibodies will not stop intratumoral virus spread. In this regard it has been established that the contribution of humoral immunity to control and clearance of MV infection is limited.³² It is expected that infected tumor cells will transfer infection both to adjacent tumor cells by homofusion and to intratumoral stromal cells (fibroblasts, endothelial cells, and infiltrating immune cells) by heterofusion. In theory, this will provide a mechanism for transfer of MV infection to distant tumor sites, which is a key issue for the treatment of metastatic cancer. Thus, tumor-infiltrating lymphocytes may play a role as secondary oncolytic MV distributors. However, the question of how individual measles-infected blood cells are able to migrate and distribute MV without fusing to other circulating cells is still obscure. In our studies, using blood samples from four different donors, we found that concanavalin A-activated lymphocytes were infected with very high efficiency. Isolated PBMCs or PBMC-derived cells (monocytes/macrophages, dendritic cells, OECs, or lymphocytes) could be successfully inoculated with MV in vitro and subsequently injected into the same patient. However, detailed additional studies will be required to determine which cells are the ideal carriers for delivery of oncolytic measles to multiple tumor sites.

Recently published data indicate that nonspecific absorption ("hitchhiking") and direct transfer of cell-surface-associated infectious virions ("hand-off") could be a method for delivery of retroviral vectors to sites of tumor growth in cancer gene therapy.³³ However, our data show that MVs nonspecifically absorbed on the surface of carrier cells remain fully susceptible to antibody and complement neutralization. Thus, in all our experiments we treated the measles-infected delivery cells with a strong neutralizing cocktail of antibodies and complement before in vivo application, thereby avoiding any possible confounding effects due to transfer of cell-surface-associated viruses.

Compared with our previous studies of MV therapy in the SKOV3ip.1 ovarian cancer model,^{5, 34} our results of therapy with MV-infected cells showed significantly improved survivals of mice bearing orthotopic ovarian cancer xenografts. When the therapy was initiated 10 days before ascites development, five injections of 10⁶ MV-infected U-937 monocytic cells were sufficient to double the median survival in 100% of treated animals. Earlier treatment (7 days post-engrafment) with infected carriers was even more effective, doubling the median survival after a single dose of cells and leading to complete tumor control in 100% of animals treated with five doses. These data illustrate the feasibility of a promising new strategy for the delivery of oncolytic virotherapy in patients with ovarian cancer. Very recent data demonstrated that mesenchymal progenitor cells can be used as intermediate carriers for targeted delivery of replicative oncolytic adenoviruses.³⁵ We are currently conducting a phase I clinical trial in which a recombinant MV expressing soluble human carcinoembryonic antigen is delivered directly (as naked virus) into the peritoneal cavity of patients with advanced ovarian cancer. Ascites frequently accumulate in patients who have tumor spread in the peritoneal cavity and this fluid is expected to be rich in anti-MV antibodies because the immunoglobulin G content of ascitic fluid is known to reflect that of blood.³⁶ Also, it has been shown that the presence of pre-existing neutralizing antibodies in ascites may prevent initial adenovirus vector delivery in ovarian cancer patients.³⁷ Thus, cell carriers are expected to be useful not only for systemic virus delivery but also for the i.p. route of administration in patients with peritoneal metastases and pre-existing humoral immune response. Injection of infected cells or free virions in the presence of neutralizing antibodies confirmed the superiority of the cell delivery strategy in a peritoneal model of hepatocellular carcinoma. These data illustrated the potential advantage of the cell delivery strategy in immunocompetent, measles-immune individuals.

Systemic administration of infected cells could also bypass a pre-existing anti-MV antibody response in the plasma. Thus, transfer of infection by heterofusion to disseminated Raji tumors was not blocked by a 1:4–1:8 complete neutralizing antibody titer in plasma, achieved by passive immunization with human serum. Our previous studies indicated that anti-MV antibody titers are greatly reduced in myeloma patients with relapsed disease or following stem cell transplantation.³⁸ Also, MV-specific immunity is known to wane rapidly in leukemia and lymphoma patients during treatment such that repeated immunizations are recommended after the treatment is completed. Thus, cellular delivery of oncolytic MVs may be a particularly promising approach for patients with advanced incurable malignancy and a low titer of serum neutralizing antibodies.

The therapeutic potential of attenuated MV therapy was recently demonstrated in patients with cutaneous T-cell lymphoma.⁷ Intratumoral MV gene expression and local tumor regression were observed after intralesional injection of small doses of a commercial measles vaccine. Neither spread of infection nor any effect on distant tumor sites was observed and the anti-measles immune responses of these patients were not presented. Comparing oncolytic virotherapy to natural measles infection, it is not expected that free virions budding from infected tumor cells will have the ability to reach distant tumor metastases. Overall, these observations indicate that cell-associated viremia is a critical factor contributing to the successful systemic spread of MV, whether during natural infection or after administration of oncolytic virotherapy.

In conclusion, this is the first report concerning the systemic administration of replication competent oncolytic measles viruses using infected cell carriers. This strategy allowed oncolytic MV to escape neutralization by antibodies and complement, and subsequently to transfer the virus to tumor cells by in situ cell fusion. Both heterotypic and homotypic cell-to-cell fusion are more resistant to neutralization by anti-MV antibodies than virus-to-cell fusion of naked virions. Survival experiments demonstrated the great potential of MV-infected cell carriers to control ovarian cancer growth. Thus, our data suggest that systemic administration of oncolytic MVs in the form of autologous MV-infected isolated monocytes or whole PBMC population will be the preferred strategy to bypass the pre-existing anti-MV humoral response in cancer patients.

Materials and Methods

MV strains. Recombinant MV strains expressing GFP, human CEA, and NIS were propagated in Vero as described.^{8, 39, 40} Viral stocks were prepared by repeated freezing–thawing procedure and MV titer was calculated by inoculation of Vero cell monolayers with 10-fold serial dilutions. For the preparation of highly purified stocks of MV-NIS and MV expressing soluble human carcinoembryonic antigen, see Supplementary Materials and Methods.

Cell lines primary isolated human cells and culture. Human U-937, Raji, and J774A.1 mouse macrophages were obtained from the American Type Culture Collection (ATCC, Rockville, MD). The MM cell line KAS-6/1 was kindly provided by Dr D Jelinek (Mayo Clinic, Rochester, MN). The SKOV3ip.1 ovarian tumor cells were a kind gift of Dr E Vitetta (University of Texas Southwestern Medical Center). Hepatocellular carcinoma line HUH-7 was provided by Dr N La Russo and Dr G Gores (Mayo Clinic, Rochester MN). GFP- and RFP-expressing cells were generated using pHR-SIN-CSGWdlNotI lentiviral vector,⁴¹ kindly provided by Dr Y Ikeda (Mayo Clinic, Rochester, MN). Hybridoma clone CL48, secreting immunoglobulin M MAb against MV H⁴² was kindly provided by Dr Wild (Institut Pasteur de Lyon, France). (For growth conditions, see Supplementary Materials and Methods). PBMCs were isolated from the peripheral blood of healthy volunteers using a Ficoll gradient centrifugation method. Cells were stimulated with 2.5–5 g/ml concanavalin A (Sigma-Aldrich, St Louis, MO) and were grown for 2–3 days before MV infection. OECs were produced by isolation of endothelial progenitors from peripheral blood as described previously.⁴³ All human blood samples were collected after Institutional Review Board approval.

Antibodies and human serum samples. Human sera were collected from healthy volunteers and patients with MM. All samples were heat inactivated by incubation at 56°C for 30 min and MV neutralization titer was determined by a VN test (see Supplementary Materials and Methods). MAb CL48 was purified from serum-free hybridoma supernatant as described previously.⁴⁴

In vitro cell-mediated transfer of MV infection. Infection of carrier cells (monocytic line or normal cells) was optimized using different MOI=0.1–1 of MV-GFP and incubation for 1–18 h. The cells were cultured in the presence of fusion inhibitory peptide (Bachem California, Torrance, CA) to block syncytia formation.⁴⁵ The number of individual-infected cells was counted by flow cytometry (see Supplementary Materials and Methods). Infected delivery cells were treated with the highly neutralizing combination of 5–20 g/ml of MAb CL48 and 10% guinea pig complement (Calbiochem-Novabiochem, San Diego, CA) for 30–60 min. After repeated washing to remove all traces of unbound antibody, the cells were transferred to monolayers of adherent cells or mixed with suspension target cells at different ratios. To detect any contamination with cell-free virions, 100 l of the last wash was inoculated on Vero monolayers. To demonstrate that MV caused heterofusion between cells, we used a dual-color model. Delivery cells expressing GFP or RFP were infected with MVs without fluorescent marker. After antibody and complement treatment, infected carriers were mixed with the target cells expressing different fluorescent protein. Mixed (double-color) syncytia were demonstrated at different periods of incubation by fluorescent and confocal microscopy. Similar experiments were run in the presence of neutralizing human serum antibodies. MV delivery cells were infected with MOI=0.5 of MV-GFP for 1 h. After MAb CL48 and complement treatment, 10² and 10³ cells per well were overlaid on RFP-expressing SKOV3ip.1 cells in the presence of serially diluted human serum (blood group AB/Rh⁺ with complete MV neutralization titer 1:320). 10³ PFU of MV-GFP were added to the control wells.

Transfer of MV infection using nonspecific absorption of MV on the cell surface. Mouse J774A.1 macrophages (without receptors for MV) were incubated (at 37°C for 2 h) with MOI=0.2 of MV-GFP. Then cells were treated with MAb CL48 in the presence of guinea pig complement and were overlaid on the target Vero monolayers as described above.

Animals and in vivo experiments. All animals were purchased from The Jackson Laboratory (Bar Harbor, ME). All in vivo experimental protocols were approved by the Mayo Foundation Institutional Animal Care and Use Committee. Mice were maintained in the animal barrier facilities of Mayo Clinic (Rochester MN).

In vivo heterofusion between MV-infected delivery cells and tumor cells. Female 5-week-old SCID mice were irradiated (150 cGy) and were engrafted by i.p. injection of 10⁷ KAS6/1 myeloma cells expressing GFP. MV-NIS-infected UR-D7 (RFP⁺clone of U-937) cells (2 10⁶) were injected i.p. on day 3 or day 7. Carrier cells were inoculated with MOI=0.5 of MV-NIS for 2 h in Opti-MEM medium, washed once, and cultured for 18 h. Before injection, cells were treated with antibody/complement cocktail to neutralize cell-free virions. Two or 3 days later, cells were harvested by peritoneal lavage and analyzed by flow cytometry. Cells were diluted and RFP/GFP⁺cells were counted using fluorescent microscopy.

Systemic cell-mediated delivery of MV infection in a Raji lymphoma model. Irradiated non-obese diabetic/SCID mice were engrafted with 10⁶ i.v. injected Raji cells. Two weeks later mice were injected i.v. with 10⁶ MV-infected U-937 cells. Carrier cells were inoculated with MOI=0.5 of MV-GFP for 3 or 18 h and cell-free viruses were neutralized as described. Control mice were injected i.v. with 10⁶ TCID₅₀ of MV-GFP. Three days later the animals were killed, skin was removed, and the body was examined using a GFP mouse imaging system (Lightools Research, Encinitas, CA). MV was recovered by overlaying of tumor cells on Vero monolayers. The efficiency of MV delivery after single or multiple injections of infected cells was tested using 2 10⁶ RR-E63 (RFP⁺Raji cells). MV-GFP (10⁶)-infected (18-h incubation) U-937 carriers were administered i.v. on days 13, 15, and 17. The other animal group received a single injection of 10⁶ delivery cells on day 15 and all animals were killed on day 18. MV-GFP infection was detected by body fluorescence imaging, fluorescent microscopy, confocal microscopy, and virus isolation. A similar experiment with RR-E63-engrafted animals was performed using three repeated injections of MV-infected normal PBMCs or OECs.

Local delivery of MV infection by cell carriers in i.p. xenografts. Female SCID mice were engrafted i.p. with 5 10⁶ HUH-7 heterocellular carcinoma cells. On day 32 animals were treated with a single i.p. injection of MV-GFP-infected OEC or U-937 monocytic cells for 18 h as described above. Groups of five mice were killed on days 2 and 3. GFP+tumor lesions were detected by body imaging and fluorescent microscopy. MV-GFP was recovered by overlaying tumor cells on Vero monolayers. A similar experiment was repeated in nude mice bearing i.p. SR-B2 (RFP⁺SKOV3ip.1) ovarian cancer xenografts. Three weeks later 2 10⁶ MV-GFP-infected U-937 carriers were injected i.p. and tumors were examined for MV-GFP infection.

Survival experiments in an ovarian cancer xenograft model. Nude mice (10 per group) were engrafted i.p. with 5 10⁶ SKOV3ip.1 cells. UR-D7 cells (RFP⁺clone of U-937 monocytes) were infected for 18 h with MOI=0.5 of MV-GFP and were used as carriers. Mice were injected i.p. with 10⁶ infected antibody/complement pretreated cells (in 0.5 ml phosphate-buffered saline) on days 10, 13, 17, 21, and 28. The control group received 0.5 ml phosphate-buffered saline. The therapeutic effect of a single versus multiple administration of MV-infected carrier cells was compared in a similar experiment using an orthotopic ovarian cancer xenograft model in SCID mice. Female 5-week-old SCID mice (10 per group) were irradiated (150 cGy) and on the next day were injected i.p. with 5 10⁶ SKOV3ip.1 cells. One week later, mice were injected i.p. with 10⁶ MV-GFP-infected UR-D7 carrier cells in 0.5 ml phosphate-buffered saline. The second animal group was treated with repeated injections on days 7, 14, 21, 28, and 35. Control mice received non-infected monocytic cells on day 7.

Cell-associated MV delivery in passively immunized animals. HUH-7 cells were engrafted i.p. in SCID mice as described above. On day 32, mice were injected with MV-GFP-infected 10⁶ U-937 cells or 2 10⁶ human PBMCs. Delivery cells were inoculated for 2 h with MOI=0.5 and were additionally cultured for 18 h. After antibody/complement treatment (10 g/ml MAb CL48 and 10% guinea pig complement for 1 h), the cells were resuspended in 0.5 ml 1:10 diluted human AB/Rh⁺serum (VN titer 1:320) and injected by i.p. route. MV-GFP (2 10⁶ PFU) were resuspended in 0.5 ml Opti-MEM with 1:10 diluted serum and were injected i.p. immediately in the control group. Three days later all mice were killed and tumors were harvested for microscopic examination and MV isolation.

For systemic delivery experiment, groups of 5-week-old female non-obese diabetic/SCID mice were irradiated and 10⁶ RR-E63 were injected i.v. on the next day. On day 12 mice were immunized passively with 250 l of undiluted human AB/Rh⁺serum (complete VN titer 1:320) via i.p. route. On the next day, groups of five animals were treated with a single i.v. administration of 10⁶ MV-GFP-infected delivery cells. Carriers were treated with antibody/complement and resuspended before injection in 100 l Opti-MEM with 1:10 diluted serum as described above. Control group received 10⁶ TCID₅₀ of MV-GFP injected in the same way. One day after virotherapy, serum samples were collected and MV neutralization titer was determined by VN test. All animals were killed 3 days after the treatment and were examined for MV-GFP infection and virus isolation.

References

1. Lin, E and Nemunaitis, J (2004). Oncolytic viral therapies. Cancer Gene Ther 11: 643–664. | Article | PubMed | ISI | ChemPort |
2. Nakamura, T and Russell, SJ (2004). Oncolytic measles viruses for cancer therapy. Expert Opin Biol Ther 4: 1685–1692. | Article | PubMed | ISI | ChemPort |
3. Grote, D et al. (2001). Live attenuated measles virus induces regression of human lymphoma xenografts in immunodeficient mice. Blood 97: 3746–3754. | Article | PubMed | ISI | ChemPort |
4. Peng, KW, Ahmann, GJ, Pham, L, Greipp, PR, Cattaneo, R and Russell, SJ (2001). Systemic therapy of myeloma xenografts by an attenuated measles virus. Blood 98: 2002–2007. | Article | PubMed | ISI | ChemPort |
5. Peng, KW, TenEyck, CJ, Galanis, E, Kalli, KR, Hartmann, LC and Russell, SJ (2002). Intraperitoneal therapy of ovarian cancer using an engineered measles virus. Cancer Res 62: 4656–4662. | PubMed | ISI | ChemPort |
6. Phuong, LK et al. (2003). Use of a vaccine strain of measles virus genetically engineered to produce carcinoembryonic antigen as a novel therapeutic agent against glioblastoma multiforme. Cancer Res 63: 2462–2469. | PubMed | ISI | ChemPort |
7. Heinzerling, L, Kunzi, V, Oberholzer, PA, Kundig, T, Naim, H and Dummer, R (2005). Oncolytic measles virus in cutaneous T-cell lymphomas mounts antitumor immune responses in vivo and targets interferon-resistant tumor cells. Blood 106: 2287–2294. | Article | PubMed | ChemPort |
8. Dingli, D et al. (2004). Image-guided radiovirotherapy for multiple myeloma using a recombinant measles virus expressing the thyroidal sodium iodide symporter. Blood 103: 1641–1646. | Article | PubMed | ISI | ChemPort |
9. Griffin, D (2001). Measles virus In: Knipe DM and Howley PM (eds). Fields Virology. Lippincott Williams & Wilkins: Philadelphia, pp 1401–1441.
10. Morrison, TG (2003). Structure and function of a paramyxovirus fusion protein. Biochim Biophys Acta 1614: 73–84. | PubMed | ChemPort |
11. Dorig, RE, Marcil, A, Chopra, A and Richardson, CD (1993). The human CD46 molecule is a receptor for measles virus (Edmonston strain). Cell 75: 295–305. | Article | PubMed | ISI | ChemPort |
12. Naniche, D et al. (1993). Human membrane cofactor protein (CD46) acts as a cellular receptor for measles virus. J Virol 67: 6025–6032. | PubMed | ISI | ChemPort |
13. Tatsuo, H, Ono, N, Tanaka, K and Yanagi, Y (2000). SLAM (CDw150) is a cellular receptor for measles virus. Nature 406: 893–897. | Article | PubMed | ISI | ChemPort |
14. Forthal, DN, Landucci, G, Habis, A, Zartarian, M, Katz, J and Tilles, JG (1994). Measles virus-specific functional antibody responses and viremia during acute measles. J Infect Dis 169: 1377–1380. | PubMed | ChemPort |
15. Bouche, FB, Ertl, OT and Muller, CP (2002). Neutralizing B cell response in measles. Viral Immunol 15: 451–471. | Article | PubMed | ChemPort |
16. Malvoisin, E and Wild, F (1990). Contribution of measles virus fusion protein in protective immunity: anti-F monoclonal antibodies neutralize virus infectivity and protect mice against challenge. J Virol 64: 5160–5162. | PubMed | ISI | ChemPort |
17. Forthal, DN, Landucci, G, Katz, J and Tilles, JG (1993). Comparison of measles virus-specific antibodies with antibody-dependent cellular cytotoxicity and neutralizing functions. J Infect Dis 168: 1020–1023. | PubMed | ChemPort |
18. Forthal, DN and Landucci, G (1998). In vitro reduction of virus infectivity by antibody-dependent cell-mediated immunity. J Immunol Methods 220: 129–138. | Article | PubMed | ChemPort |
19. Perrin, LH, Joseph, BS, Cooper, NR and Oldstone, MB (1976). Mechanism of injury of virus-infected cells by antiviral antibody and complement: participation of IgG, F(ab')2, and the alternative complement pathway. J Exp Med 143: 1027–1041. | Article | PubMed | ChemPort |
20. Permar, SR et al. (2003). Role of CD8(+) lymphocytes in control and clearance of measles virus infection of rhesus monkeys. J Virol 77: 4396–4400. | Article | PubMed | ChemPort |
21. El Mubarak, HS et al. (2004). Measles virus protein-specific IgM, IgA, and IgG subclass responses during the acute and convalescent phase of infection. J Med Virol 72: 290–298. | Article | PubMed | ChemPort |
22. Isa, MB et al. (2001). Measles virus-specific immunoglobulin G isotype immune response in early and late infections. J Clin Microbiol 39: 170–174. | Article | PubMed | ChemPort |
23. Itoh, M, Okuno, Y and Hotta, H (2002). Comparative analysis of titers of antibody against measles virus in sera of vaccinated and naturally infected Japanese individuals of different age groups. J Clin Microbiol 40: 1733–1738. | Article | PubMed |
24. Toptygina, AP, Pukhalsky, AL and Alioshkin, VA (2005). Immunoglobulin G subclass profile of antimeasles response in vaccinated children and in adults with measles history. Clin Diagn Lab Immunol 12: 845–847. | Article | PubMed | ChemPort |
25. Forthal, DN, Aarnaes, S, Blanding, J, de la Maza, L and Tilles, JG (1992). Degree and length of viremia in adults with measles. J Infect Dis 166: 421–424. | PubMed | ChemPort |
26. Devaux, P, Christiansen, D, Plumet, S and Gerlier, D (2004). Cell surface activation of the alternative complement pathway by the fusion protein of measles virus. J Gen Virol 85: 1665–1673. | Article | PubMed | ChemPort |
27. Sissons, JG, Oldstone, MB and Schreiber, RD (1980). Antibody-independent activation of the alternative complement pathway by measles virus-infected cells. Proc Natl Acad Sci USA 77: 559–562. | Article | PubMed | ChemPort |
28. Esolen, LM, Ward, BJ, Moench, TR and Griffin, DE (1993). Infection of monocytes during measles. J Infect Dis 168: 47–52. | PubMed | ChemPort |
29. Mrkic, B et al. (1998). Measles virus spread and pathogenesis in genetically modified mice. J Virol 72: 7420–7427. | PubMed | ISI | ChemPort |
30. Roscic-Mrkic, B et al. (2001). Roles of macrophages in measles virus infection of genetically modified mice. J Virol 75: 3343–3351. | Article | PubMed | ISI | ChemPort |
31. Peng, KW et al. (2003). Biodistribution of oncolytic measles virus after intraperitoneal administration into Ifnar-CD46Ge transgenic mice. Hum Gene Ther 14: 1565–1577. | Article | PubMed | ChemPort |
32. Permar, SR et al. (2004). Limited contribution of humoral immunity to the clearance of measles viremia in rhesus monkeys. J Infect Dis 190: 998–1005. | Article | PubMed |
33. Cole, C et al. (2005). Tumor-targeted, systemic delivery of therapeutic viral vectors using hitchhiking on antigen-specific T cells. Nat Med 11: 1073–1081. | Article | PubMed | ISI | ChemPort |
34. Peng, KW et al. (2006). Pharmacokinetics of oncolytic measles virotherapy: eventual equilibrium between virus and tumor in an ovarian cancer xenograft model. Cancer Gene Ther 13: 732–738. | Article | PubMed | ChemPort |
35. Komarova, S, Kawakami, Y, Stoff-Khalili, MA, Curiel, DT and Pereboeva, L (2006). Mesenchymal progenitor cells as cellular vehicles for delivery of oncolytic adenoviruses. Mol Cancer Ther 5: 755–766. | Article | PubMed | ChemPort |
36. Confino, E, Harlow, L and Gleicher, N (1990). Peritoneal fluid and serum autoantibody levels in patients with endometriosis. Fertil Steril 53: 242–245. | PubMed | ChemPort |
37. Stallwood, Y, Fisher, KD, Gallimore, PH and Mautner, V (2000). Neutralisation of adenovirus infectivity by ascitic fluid from ovarian cancer patients. Gene Ther 7: 637–643. | Article | PubMed | ChemPort |
38. Dingli, D et al. (2005). Interaction of measles virus vectors with Auger electron emitting radioisotopes. Biochem Biophys Res Commun 337: 22–29. | Article | PubMed | ChemPort |
39. Duprex, WP, McQuaid, S, Roscic-Mrkic, B, Cattaneo, R, McCallister, C and Rima, BK (2000). In vitro and in vivo infection of neural cells by a recombinant measles virus expressing enhanced green fluorescent protein. J Virol 74: 7972–7979. | Article | PubMed | ChemPort |
40. Peng, KW, Facteau, S, Wegman, T, O'Kane, D and Russell, SJ (2002). Non-invasive in vivo monitoring of trackable viruses expressing soluble marker peptides. Nat Med 8: 527–531. | Article | PubMed | ISI | ChemPort |
41. Rowe, HM et al. (2006). Immunization with a lentiviral vector stimulates both CD4 and CD8 T cell responses to an ovalbumin transgene. Mol Ther 13: 310–319. | Article | PubMed | ChemPort |
42. Giraudon, P and Wild, TF (1985). Correlation between epitopes on hemagglutinin of measles virus and biological activities: passive protection by monoclonal antibodies is related to their hemagglutination inhibiting activity. Virology 144: 46–58. | Article | PubMed | ISI | ChemPort |
43. Simper, D et al. (2003). Endothelial progenitor cells are decreased in blood of cardiac allograft patients with vasculopathy and endothelial cells of noncardiac origin are enriched in transplant atherosclerosis. Circulation 108: 143–149. | Article | PubMed | ISI |
44. Iankov, ID, Pandey, M, Harvey, M, Griesmann, GE, Federspiel, MJ and Russell, SJ (2006). Immunoglobulin G antibody-mediated enhancement of measles virus infection can bypass the protective antiviral immune response. J Virol 80: 8530–8540. | Article | PubMed | ChemPort |
45. Firsching, R et al. (1999). Measles virus spread by cell-cell contacts: uncoupling of contact-mediated receptor (CD46) downregulation from virus uptake. J Virol 73: 5265–5273. | PubMed | ISI | ChemPort |

Acknowledgements

This work was partially supported by NIH Grant HL66958 and JARI Research Foundation. We thank Dr D Jelinek, Dr N La Russo and Dr G Gores (from Mayo Clinic, Rochester), Dr Vitetta (University of Texas Southwestern Medical Center), and Dr Wild (Institut Pasteur de Lyon, France) for the cell lines. We also thank Dr Ikeda for lentiviral vectors, Dr Guy E Griesmann for MAb CL48 antibody preparation, and Mary Harvey for technical assistance.

Supplementary Material

Figure S1. Heterofusion between U-937 delivery cells (GFP+) and Raji (RFP⁺) lymphoma cells.

Figure S2. MV-GFP infected human outgrown endothelial cells (OEC) successfully transferred infection to RR-E6 (RFP⁺ clone of Raji) cell (A) and SR-A3 (RFP⁺ ovarian SKOV3ip.1) cells (B).

Movie S1. Systemic delivery of GFP-MV infection using cell carriers in Raji lymphoma model.

Movie S2. Delivery of MV using cell carriers in ovarian xenograft model.

Materials and Methods