Engineered measles virus as a novel oncolytic viral therapy system for hepatocellular carcinoma


  • Potential conflict of interest: Nothing to report.


The oncolytic measles virus Edmonston strain (MV-Edm), a nonpathogenic virus targeting cells expressing abundant CD46, selectively destroys neoplastic tissue. Clinical development of MV-Edm would benefit from noninvasive monitoring strategies to determine the speed and extent of the spread of the virus in treated patients and the location of virus-infected cells. We evaluated recombinant MV-Edm expressing carcinoembryonic antigen (CEA) or the human sodium iodide symporter (hNIS) for oncolytic potential in hepatocellular carcinoma (HCC) and efficiency in tracking viruses in vivo by noninvasive monitoring. CD46 expression in human HCC and primary hepatocytes was assessed by flow cytometry and immunohistochemistry. Infectivity, syncytium formation, and cytotoxicity of recombinant MV-Edm in HCC cell lines were evaluated by fluorescence microscopy, crystal violet staining, and the MTS assay. Transgene expression in HCC cell lines after infection with recombinant MV-Edm in vitro and in vivo was assessed by CEA concentration, 125I-uptake, and 123I-imaging studies. Toxicology studies were performed in IfnarKO×CD46 transgenic mice. The CD46 receptor was highly expressed in HCC compared to nonmalignant hepatic tissue. Recombinant MV-Edm efficiently infected HCC cell lines, resulting in extensive syncytium formation followed by cell death. Transduction of HCC cell lines and subcutaneous HCC xenografts with recombinant MV-Edm resulted in high-level expression of transgenes in vitro and in vivo. MV-Edm was nontoxic in susceptible mice. Intratumoral and intravenous therapy with recombinant MV-Edm resulted in inhibition of tumor growth and prolongation of survival with complete tumor regression in up to one third of animals. In conclusion, engineered MV-Edm may be a potent and novel cancer gene therapy system for HCC. MV-Edm expressing CEA or hNIS elicited oncolytic effects in human HCC cell lines in vitro and in vivo, enabling the spread of the virus to be monitored in a noninvasive manner. (HEPATOLOGY 2006;44:1465–1477.)

Hepatocellular carcinoma (HCC) is the most common primary malignancy of the liver, accounting for up to 90% of the incidence of primary hepatic cancer.1 Worldwide, it is the fifth most common cancer and the third most common cause of cancer mortality, causing 550,000 deaths annually. Its incidence is 560,000 per year.1, 2 Its geographic distribution is characterized by wide variability, with most cases in sub-Saharan Africa and Eastern Asia. Its incidence in the United States and Western Europe has more than doubled in the last 2 decades, from 1.3/100,000 in 1978-1980 to 4.1/100,000 in 1998-2000.3 At present, liver cancer is the most rapidly increasing type of cancer in the United States.4

Despite a variety of therapeutic options including surgery, chemotherapy, and local ablative therapies, the prognosis for HCC remains poor, with a median survival of 8 months.3, 5–9 Many therapeutic alternatives are being explored, including hormonal therapy, immunotherapy, and gene therapy.10–12 The current approach to gene therapy for HCC is to use viral and nonviral gene therapy systems. Unfortunately, most of these systems suffer from serious drawbacks such as low efficacy or safety.10 The increasing incidence, the lack of effective therapies, and the devastating prognosis of HCC support the urgent need for new therapeutic agents that are both safe and effective.

Attenuated strains of the measles virus have safely been used for vaccination since 1963. The attenuated Edmonston vaccine strain of the measles virus (MV-Edm) has considerable oncolytic activity that is retained when the virus is genetically modified.13–17 It exerts its cytopathic effect (CPE) by fusing infected cells with the surrounding cells, forming multinucleated syncytia, followed by cell death by apoptotic or nonapoptotic mechanisms.14, 16, 18 The measles virus is a negative-strand RNA paramyxovirus, and it has been shown that MV-Edm preferentially fuses and kills cells overexpressing the CD46 receptor.19 CD46 is a membrane-associated complement regulatory protein ubiquitously expressed on nucleated human cells.20–22 Tumor cells frequently overexpress CD46.23, 24 These mechanisms contribute to the tumor selectivity of MV-Edm.

In the current study, we evaluated MV-Edm as a novel therapeutic agent for HCC. We used recombinant MV-Edm, genetically modified to express either the extracellular domain of the soluble human carcinoembryonic antigen (hCEA) or the human sodium iodide symporter (hNIS).13, 25 hCEA, a soluble marker peptide with no biological activity and normal serum concentrations of less than 5 ng/mL, is commonly used as a tumor marker for human malignancies.26–28 Although 15% of HCC patients show elevated CEA serum concentrations, it is not considered a standard tumor marker for HCC because of its low diagnostic accuracy.29–31 Through the use of hCEA-expressing recombinant MV-Edm and analysis of serum levels of hCEA, viral gene expression and its kinetic profiles can be obtained in a noninvasive manner.25

hNIS is a membrane ion channel mainly expressed in the follicular cells of the thyroid.32 It facilitates iodine uptake in the thyrocytes by a symporter mechanism, transporting 1 iodine ion together with 2 Na ions against the Na gradient maintained by Na-K-ATPase.33 Gene expression of hNIS after MV-NIS infection followed by application of 123I results in increased intracellular uptake of the isotope, allowing noninvasive assessment of hNIS expression and tumor distribution using gamma-camera imaging.34, 35

In the present study we have shown that the use of MV-Edm expressing these marker genes not only has significant therapeutic potential for HCC but also facilitates noninvasive in vivo monitoring, allowing optimization of therapeutic protocols and enhancement of safety.


MV-Edm, measles virus Edmonston strain; CEA, carcinoembryonic antigen; hNIS, human sodium iodide symporter; HCC, hepatocellular carcinoma; CPE, cytopathic effect; GFP, green fluorescent protein; MOI, multiplicity of infection.

Materials and Methods

Cell Culture.

The human HCC cell line Hep-3B was obtained through the American Type Culture Collection (ATCC; Manassas, VA) and maintained in MEM medium supplemented with 10% heat-inactivated FBS and 1% sodium pyruvate. The human HCC cell line HUH-7, a kind gift from Dr. G. Gores, was maintained in DMEM medium supplemented with 10% FBS. The Vero African green monkey kidney cells (ATCC, CCL-81) used for production of the measles virus were maintained in DMEM supplemented with 5% FBS. All media used in this study contained 100 U/mL penicillin-streptomycin. Growth media, sera, and supplements were from Gibco BRL (Grand Island, NY). Primary human hepatocytes were purchased through CellzDirect and maintained in medium as suggested by the manufacturer. All cells used in this study were cultured in a humidified atmosphere of 5% CO2 at 37°C.

Viruses and Infection Assays.

Recombinant MV-Edm encoding the soluble extracellular domain of human CEA (MV-CEA), MV-Edm encoding the green fluorescent marker protein (MV-GFP), and MV-Edm encoding the human NIS protein (MV-NIS) were generated and propagated on Vero cells as described previously.34 Figure 1 shows a schematic diagram of the full-length constructs of MV-eGFP, MV-CEA, and MV-NIS.13, 25, 36 The titers of viral stocks were determined by 50% end-point dilution assays (TCID50) on Vero cells. For virus infection assays, 2 × 105 cells were incubated with recombinant MV-Edm diluted in 1.0 mL of Opti-MEM (Life Technologies, Inc.) for 2 hours at 37°C. At the end of the incubation period, the virus was removed, and the cells were maintained in the standard medium.

Figure 1.

Schematic representation of the genome of recombinant MV-Edm constructs showing the coding regions for viral proteins and the intergenic regions (N, nucleoprotein gene; P, phosphoprotein gene; M, matrix protein gene; F, fusion protein gene; H, hemagglutinin gene; L, polymerase gene). The P cistron encodes 2 additional proteins, C protein and V protein. (A) Overview of the genome of recombinant MV-Edm coding for the human carcinoembryonic antigen (hCEA) as described previously.25 (B) Genome of recombinant MV-Edm encoding enhanced green fluorescent protein (eGFP) as described previously.36 (C) Genome of recombinant MV-Edm coding for the human sodium iodide symporter (NIS), as described previously.13

Flow Cytometry.

Cells were harvested, washed twice in cold 2% BSA-PBS, and incubated with an FITC-labeled monoclonal mouse antihuman CD46 antibody (PharMingen, San Diego, CA) for 1 hour on ice. The cells were then washed twice and run on a Becton-Dickinson FACScan Plus cytometer and analyzed using CellQuest software (Becton-Dickinson, San Jose, CA).


Human liver sections were obtained from the Mayo Tissue Registry according to the IRB protocol. Each parafffin section was treated by being incubated twice in xylene for 5 minutes and then placed through an ethanol gradient (100%, 95%, 80%, 70%, and 60%) for 5 minutes. The sections were then placed in 10 mmol/L of citrate buffer (pH 6.0) and microwaved twice for 2 minutes to improve staining by antigen unmasking. After the sections were washed and quenched of endogenous peroxidases, they were blocked and incubated with CD46 monoclonal antibody (1:750; Immunotech, France) overnight at 4°C. An immunoperoxidase procedure (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA) was carried out to visualize the peroxidase by reaction with diaminobenzidine and hydrogen peroxidase (Vector Laboratories, Burlingame, CA) and counterstain it with hematoxylin. Controls in which the primary antibody was omitted revealed no labeling.

Study of In Vitro Iodine Uptake.

HUH-7 and Hep-3B cells were infected for 2 hours with MV-NIS at multiplicity of infection (MOI) of 0.01, 0.1, and 1.0 in Opti-MEM. The virus suspension was removed, and growth medium was added to the cells. In vitro iodine uptake was studied 24, 48, and 72 hours after infection as previously described.37 Briefly, cells were washed twice in Hanks balanced salt solution (HBSS) and resuspended in HBSS with 10 mmol/L HEPES (N-2-hydroxyethylpiperazine-N′-ethanesulfonic acid, pH 7.3). Potassium perchlorate was added to half the samples to a final concentration of 1 μmol/L. Then 1 × 105 cpm 125I was added to each sample. After 50 minutes of incubation at 37°C, cells were washed twice with cold HBSS and lysed by the addition of 1 mol/L NaOH. Activity in the cell lysate was determined with a gamma counter.

CEA Analysis.

For in vivo experiments, blood samples were collected from mice by retro-orbital bleeding and serum analyzed for CEA concentration. For in vitro experiments, supernatant from MV-CEA-infected and -uninfected HCC cells was collected and analyzed for CEA concentration. CEA levels were measured using the Bayer Centaur Immunoassay System.

Evaluation of CPE In Vitro.

The CPE of recombinant MV-Edm on HCC cell lines was evaluated by crystal violet staining and a CellTiter 96 AQueous nonradioactive cell proliferation assay kit (Promega Corp., Madison, WI). For crystal violet staining, cells were plated in 6-well plates at a density of 2 × 105 cells/well. Twenty-four hours after seeding, cells were infected with MV-GFP as described above. Then 24, 48, 72, 96, and 120 hours after infection, cells were gently washed twice with warm PBS, and the remaining cells were fixed with 2% formaldehyde in PBS for 10 minutes. The wells were washed with PBS and stained with 0.13% crystal violet solubilized in ethanol-formaldehyde (2:1). The stained product was subsequently washed twice with distilled water and air-dried. The CellTiter 96 AQueous nonradioactive cell proliferation assay was performed according to the manufacturer's recommendations. Briefly, cells were plated in 96-well plates at a density of 8 × 103 cells/well. Twenty-four hours after seeding, cells were infected with MV-GFP as already described. Then 24, 48, 72, 96, and 120 hours after infection, cells were incubated with 20 μL of MTS reagent solution for 2 hours at 37°C. Absorbance at 490 nm was recorded using an ELISA plate reader.

In Vivo Experiments.

All procedures involving animals were approved by and performed according to guidelines of the Institutional Animal Care and Use Committee of the Mayo Clinic. A 27-gauge needle was used to subcutaneously inject female SCID and nude mice (4-6 weeks of age; Harlan Laboratories, Indianapolis, IN) with 5 × 105 HUH-7 cells/100 μL of PBS and 5 × 105 Hep-3B cells/100 μL of PBS in the right flank. Mice were examined daily for tumor growth. Tumor length, width, and height were measured with calipers. Tumor volume was calculating according to the formula π/6 × width × length × height.38 When tumors reached a maximum diameter of 0.5 cm, MV-Edm treatment was initiated by either intratumoral or intravenous injection in the tail vein. Animals were routinely bled to obtain serum hCEA levels. Animals were euthanized when a tumor diameter reached 1 cm or when 10% of body weight was lost.

Toxicology Analysis.

Forty-eight IfnarKO × CD46 Ge mice (24 male and 24 female, 6-8 weeks of age; Mayo Clinic, Rochester, MN) were randomly divided into 2 groups (3/sex/group) and treated systemically with 1 × 106 TCID50 MV-NIS or vehicle solution (5% sucrose, 50 mmol/L Tris-HCl, pH 7.4, 2 mmol/L MgCl2) by a single-bolus tail-vein injection. Clinical signs and weights were monitored daily during the first week posttreatment and weekly thereafter. Twelve mice per group were euthanized 2, 5, 22, and 91 days after treatment, and complete necropsies were performed, at which time, blood samples were obtained by retro-orbital plexus blood collection.

Blood samples were analyzed for complete blood count with differential (hemoglobin, red blood cell count, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin concentration, red cell distribution width, white blood cell count, lymphocytes, monocytes, granulocytes, platelet count, mean platelet volume, and platelet distribution width), clinical chemistry (albumin, alkaline phosphatase, alanine aminotransferase, amylase, total bilirubin, blood urea nitrogen, calcium, phosphorus, creatinine, glucose, sodium, potassium, total protein, and globulin), and coagulation (prothrombin and activated partial thromboplastin times). For analysis, the VetScan HMT Hematology System (Union City, CA) and an SCA2000 Veterinary Coagulation Analyzer (San Diego, CA) were used according to the manufacturer's instructions. The calibration was verified before each analysis for quality control, and each sample was analyzed with 3 levels of control.

For analysis of the antimeasles antibodies, an immunofluorescence assay based on BION measles virus antigen substrate slides (BION Enterprises, Ltd., Des Plaines, IL) was used, which was performed according to the manufacturer's instructions. The protocol was modified as described recently by Peng et al. using goat antimouse IgG FITC–conjugated antibodies for detection of murine IgG.25

The organs and tissues harvested were adrenal glands, bone (femur), bone marrow (femur), brain, cecum, colon, duodenum, epididymidis, esophagus, eyes, gall bladder, gonads (testes/ovaries), gross lesions, heart, ileum, jejunum, kidneys, liver, lungs, lymph nodes (mandibular and mesenteric), mammary gland, pituitary gland, salivary glands (mandibular, sublingual, and parotid), sciatic nerve, skeletal muscle, skin (ventral abdomen), spinal cord, spleen, stomach, thymus, thyroid glands, trachea, urinary bladder, and uterine horn. The tissues were preserved in Davidson's solution (eyes only) or 10% neutral buffered formalin embedded in paraffin, sectioned at approximately 5 μm, stained with hematoxylin and eosin, and then examined microscopically by a board-certified veterinary pathologist.

Study of In Vivo Iodine Uptake.

Subcutaneous HCC xenografts were implanted as described in the In Vivo Experiments section. Once the tumors reached a maximum diameter of 0.5 cm, the animals were treated with a single dose of 2 × 106 TCID50 of MV-NIS by intravenous tail-vein injection. Mice were injected intraperitoneally with 500 μCi 123I 3, 7, 10, and 14 days after MV-NIS treatment. 123I instead of 125I was chosen as it facilitates serial iodine uptake studies due to its shorter half-life of 13.2 hours. One hour after injection of the isotope, whole-body imaging to determine 123I activity was performed using a gamma camera (Helix System; Elscint, Haifa, Israel). For quantitative analyses of intratumoral iodine uptake, regions of interest (ROI) within the gamma camera data were selected. ROI 1 was defined as whole-body activity and ROI 2 as intratumoral activity. Activity was corrected for gamma-camera background and unspecific intratumoral iodine uptake. Intratumoral iodine uptake was calculated as a percentage of whole-body activity.

Statistical Methods.

The statistical analysis of the significance of differences in survival between mice treated with recombinant MV-Edm and mice that received UV-inactivated MV-Edm was performed using the log-rank test in the JMP program. P ≤ .05 was defined as indicating significant differences.


Human HCC Overexpresses CD46.

Expression of CD46 in human HCC cell lines HUH-7 and Hep-3B and in primary human hepatocytes was analyzed by flow cytometry. Human HCC cell lines express high levels of CD46 compared to that in primary human hepatocytes. We observed a 2.68-fold increase in the median fluorescence in primary human hepatocytes incubated with an FITC-labeled monoclonal antihuman CD46 antibody compared to that of the isotype control. Median fluorescence was increased 17.47-fold and 16.36-fold in the HUH-7 and Hep-3B cells, respectively (Fig. 2A). Immunochemistry was used to compare CD46 expression in HCC with that in nonmalignant hepatic tissue in human hepatic tissue samples of HCC patients (n = 3). As shown in Fig. 2B, CD46 was overexpressed in HCC tissue compared to that in the surrounding normal hepatic tissue. This result was confirmed in 100% of the human tissue samples analyzed.

Figure 2.

CD46 overexpression in HCC compared to in normal hepatic tissue. (A) High expression of the CD46 receptor in human HCC cell lines HUH-7 and Hep-3B but low expression in primary human hepatocytes. Analysis was performed by flow cytometry using a monoclonal CD46 antibody. The green histogram shows the measured fluorescence of HCC cells incubated with an isotype control and the red histogram of HCC cells labeled with an anti-CD46 FITC antibody. (B) CD46 overexpression in HCC compared to in the surrounding normal hepatic tissue. CD46 expression in human HCC was analyzed by immunohistochemistry (left, section treated only with the secondary antibody; right, the same section treated with primary CD46 antibody and secondary antibody, as described in Material and Methods). CD46 overexpression was confirmed in 100% of analyzed human tissue samples (n = 3).

Recombinant MV-Edm Successfully Infects Human HCC Cell Lines and Causes Syncytium Formation Followed by Significant CPE.

To determine the infectivity of recombinant MV-Edm in human HCC cell lines, HUH-7 and Hep-3B cells were infected with MV-eGFP at MOIs of 0.05 and 0.5. Analysis with light- and fluorescence microscopy was performed every 24 hours over a total of 120 hours. All the human HCC cell lines were successfully infected by MV-eGFP (Fig. 3A). As early as 24 hours after infection, expression of GFP was observed in infected cells even at a low MOI of 0.05. Infection of human HCC cell lines with MV-eGFP resulted in fusion of infected cells with neighboring cells (Fig. 3A). Syncytium formation was more dramatic and showed faster kinetics in Hep-3B cells than in HUH-7 cells. In Hep-3B cells, we observed large syncytia 48 hours after infection, followed by almost complete eradication of the cell layer after 96 hours. In HUH-7 cells, we observed infected cells as well as small-sized syncytia, which expanded over the next 48 hours in size and number.

Figure 3.

Infectivity, induction of syncytia, and CPE of MV-Edm in HCC. (A) Human HCC cell lines HUH-7 and Hep-3B infected with MV-GFP at an MOI of 0.5 showing GFP expression and syncytium formation 48 and 96 hours after infection. Serial fluorescence microscopic analysis was performed every 24 hours. (B) Serial analysis to determine CPE of recombinant MV-Edm performed every 24 hours on HCC cell lines HUH-7 and Hep-3B. Ninety-six hours after infection at MOIs of 0.5 and 0.05, cells were stained with crystal violet representing viable attached cells. (C) Time course of cell viability of HCC cell lines HUH-7 and Hep-3B (n = 8) after infection with recombinant MV-Edm at MOIs of 0.5 and 0.05 analyzed with a CellTiter 96 AQueous nonradioactive cell proliferation assay kit (▪ = negative control, equation image= MOI 0.05, equation image= MOI 0.5).

We further quantified the CPE of recombinant MV-Edm infection on the HCC cells. Hep-3B and HUH-7 cells were infected with MV-eGFP at MOIs of 0.05 and 0.5. Analysis was performed every 24 hours over a total of 120 hours using crystal violet staining and the CellTiter 96 AQueous nonradioactive cell proliferation assay. In both cell lines, recombinant MV-Edm elicited a strong CPE in an MOI-dependent manner (Fig. 3B–C). Reduction of viability was observed in the Hep-3B cells within the first 24 hours. Seventy-two hours after infection, the viability of the Hep-3B cells was reduced to 14.9% at an MOI of 0.5 and to 54% at an MOI of 0.05. In the HUH-7 cells, at an MOI of 0.5, the cytotoxic effect of MV-Edm started 48 hours after infection. One hundred and twenty hours after the HUH-7 cells were infected with MV-eGFP, cell viability had decreased to 41% at an MOI of 0.5 and 77.2% at an MOI of 0.05.

Infection of Human HCC Cell Lines with MV-CEA and MV-NIS Results in High Expression of Marker Genes.

To monitor the spread of the virus and the location of infected cells, we evaluated recombinant measles viruses expressing CEA and hNIS. To do this, 2 × 105 HUH-7 and Hep-3B cells were infected with MV-CEA at MOIs of 0.05 and 0.5. Supernatant was collected every 24 hours over 72 hours and analyzed for CEA concentration. In both cell lines, CEA was detected within the first 24 hours after infection. In the following 48 hours the supernatant concentration of CEA increased 100-fold at both MOIs in Hep-3B cells and 1,017-fold at an MOI of 0.05 and 817-fold at an MOI of 0.5 in HUH-7 cells. In detail, 24 hours after infection the CEA concentration in the Hep-3B cell supernatant was 7.05 ng/mL at an MOI of 0.5 and 0.42 ng/mL at an MOI of 0.05, and 72 hours after infection it had increased to 693.3 ng/mL at an MOI of 0.5 and 44.03 at an MOI of 0.05 (Fig. 4A). In the supernatants of MV-CEA-infected HUH-7 cells, 24 hours after infection CEA concentration was 8.5 ng/mL at an MOI of 0.5 and 0.8 ng/mL at an MOI of 0.05 and 72 hours after infection increased to 6,951.7 ng/mL at an MOI of 0.5 and 691 ng/mL MOI of 0.05 (Fig. 4A).

Figure 4.

Transgene expression in human HCC cell lines after infection with MV-CEA and MV-NIS. Hep-3B and HUH-7 cells were infected with MV-CEA or MV-NIS at different MOIs and analyzed for transgene expression. (A) Time course of CEA concentration in the supernatant of MV-CEA-infected Hep-3B and HUH-7 cells (n = 6). Data points are means with standard deviations (black bars, negative controls; black-and-white bars, MOI 0.05–infected cells; gray bars, MOI 0.5–infected cells). (B) Isotope activity in MV-NIS-infected HCC cells after incubation with 125I showing means with standard deviations (n = 3). To show NIS-mediated 125I-uptake, controls were cotreated with KClO4, which inhibits NIS-mediated iodine uptake.

HUH-7 and Hep-3B cells were infected with MV-NIS at MOIs of 0.01, 0.1 and 1.0. 24, 48, and 72 hours after infection, cells were incubated with 1 × 106 mCi 125I for 1 hour at 37°C and uptake of the isotope was analyzed. In the Hep-3B cells infection with MV-NIS resulted in MOI-dependent uptake of 125I, with a maximum uptake of 1,617 cpm at MOI of 1.0 24 hours postinfection. At later times viral infection had eradicated the cell layer to such an extent that overall uptake was reduced. In the HUH-7 cells, maximum uptake, 1,296.7 cpm, was observed 48 hours after infection at an MOI of 0.01. In the following 48 hours activity decreased at MOIs of 0.1 and 1.0 but increased at MOI of 0.01 (Fig. 4B).

Systemic Treatment of IfnarKO×CD46 Mice With MV-NIS Is Well Tolerated and Nontoxic.

To evaluate the toxicity of recombinant MV-Edm, IfnarKO×CD46 mice were treated systemically with MV-NIS. Mice received a single intravenous bolus of MV-NIS (1 × 106 TCID50/mouse) diluted in 200 μL of the vehicle solution via tail-vein injection or 200 μL of the vehicle solution without virus. Animals were evaluated regularly for clinical symptoms, increase in body weight, and pathological changes by laboratory analysis. On study days 2, 5, 22, and 91 after administration of the virus, 12 mice per group were euthanized, and complete necropsies were performed.

Animals did not develop any pathological clinical signs following treatment with MV-NIS. As shown in Fig. 5A, there was not a significant difference in body weight between MV-NIS-treated and -untreated mice. Between the day the virus was administered and the 84th day after treatment, mean body weight in the treated and negative control groups had increased by factors of 1.40 and 1.43, respectively. Serum analysis showed mild elevation of mean total protein in the MV-NIS-treated group: it was 105% of that in the negative control group (P = .005). Mean BUN was mildly decreased in the MV-NIS-treated group: it was 92% of that in the negative control (P < .001). All other blood parameters including CBC with differential, clinical chemistry for hepatic and renal function markers, electrolytes, glucose and proteins, and coagulation serum markers were not significantly different between the MV-NIS-treated group and the negative control (Fig. 5B–C). Immunological analysis showed antimeasles antibodies developing 22 and 91 days after virus treatment. Histological examination did not show any pathology related to the MV-NIS treatment. Male and female mice did equally well in this study; there were no sex-related differences in the response to MV-NIS.

Figure 5.

Study of toxicity of MV-NIS in IfnarKO×CD46 Ge mice, 24 given 1 × 106 TCID50 MV-NIS intravenously (treated group) and 24 given the vehicle solution (untreated group) by a single tail-vein injection. (A) Increases in body weight of treated (dashed line) and untreated (solid line) mice over time. The graph shows means with standard errors (n = 24). (B) Hepatic and renal serum markers (ALT, alanine aminotransferase, PT, prothrombin time). The columns represent means over time with standard errors (n = 24). Differences between treated (gray column) and untreated (black column) mice were statistically analyzed, and the corresponding P values are indicated in the diagram. (C) CBC in the treatment group (gray column) and the negative control group (black column). Shown are the results of the analysis for hemoglobin (Hb), leukocytes (WBC), and thrombocytes (Plt). Data points represent means over time with standard errors (n = 24). P values of comparison of treated and untreated groups are indicated in the diagram.

Intratumoral Therapy of Human HCC Xenografts with MV-CEA Has a Potent Antitumor Effect and Can Be Monitored by Serial Serum CEA Concentrations.

To evaluate the potential of recombinant MV-Edm for therapy of HCC, we first tested MV-CEA in a subcutaneous human HCC xenograft model. Hep-3B cells (5 × 106 cells/mouse) were implanted in the right flanks of nude mice, and HUH-7 cells (5 × 106 cells/mouse) were implanted in the right flanks of SCID mice. When the maximum tumor diameter measured approximately 0.5 cm, each mouse was treated with a total of 5 doses of MV-hCEA (1 × 107 TCID50) or an equivalent dose of UV-inactivated MV-hCEA for 10 days. Serum CEA concentration, tumor volume, and survival curves are shown in Fig. 6.

Figure 6.

Serum CEA concentrations and tumor suppression after intratumoral MV-CEA therapy of HCC xenografts. Mice bearing s.c. HCC xenografts (HUH-7, Hep-3B) were injected intratumorally with 2 × 106 TCID50 of MV-CEA every other day a total of 5 times (1 × 107 total TCID50/mouse). Treatment groups (n = 9 each) received active MV-CEA; the negative control group (n = 9) received UV-inactivated MV-CEA. (A) Course of serum CEA concentration in HCC xenograft–bearing mice after MV-CEA therapy. (B) Increase in tumor volume after initiation of MV-CEA therapy. The data points are medians with standard error (solid line, tumor volume of MV-CEA-treated mice; dashed line, tumor volume of UV-inactivated MV-CEA-treated mice). The graph ends at the time at which the first mouse had to be euthanized. (C) Survival curves of treated mice (solid line) and negative control mice (dashed line).

CEA was detected in MV-CEA-treated mice but not in UV-inactivated MV-CEA-treated control mice. In both the Hep-3B and HUH-7 HCC xenograft models, serum CEA concentrations could be detected as early as 5 days after initiation of therapy and increased over time. In the Hep-3B xenograft model, the CEA concentration reached its maximum, 312.3 ng/mL, on day 35. In the HUH-7 xenografts a maximum of 10,360 ng/mL was reached on day 12. After reaching maximum concentrations, mean CEA levels started to decrease (Fig. 6A).

In both HCC cell line xenografts, the tumor-suppressive effect of MV-CEA first became apparent on day 10, the last day of therapy, and this therapeutic efficacy then increased over time, resulting in significantly prolonged survival of treated animals (Fig. 6B–C). Among mice bearing Hep-3B tumor xenografts, the median survival of those treated with MV-hCEA and of those treated with UV-inactivated MV-hCEA was 58 and 33 days, respectively. The median survival of the MV-hCEA-treated mice, which was increased 1.75-fold, was significant longer than that of the control group (P = .0015; Fig. 6C). All mice in the control group had to be euthanized on day 48. In the MV-hCEA-treated group, complete tumor regression was observed in one third of treated mice. Among mice bearing HUH-7 tumor xenografts, median survival of those treated with MV-hCEA and of those treated with UV-inactivated MV-hCEA was 34 and 19 days, respectively. This was also a significant, 1.79-fold increase in the median survival of MV-hCEA-treated mice compared to that of the control group (P = .004; Fig. 6C). Histological analysis of tumors demonstrated intratumoral syncytium formation in vivo after MV-Edm treatment (Fig. 7).

Figure 7.

MV-Edm induced syncytium formation in vivo. Mice bearing subcutaneous HUH-7 tumor xenografts received a single i.v. injection of 2 × 106 TCID50 of MV-NIS. Five days after virus treatment, tumors were harvested and paraffin-embedded. Then 5-μm slices were prepared, HE-stained, and analyzed by microscopy. Syncytium formation within the tumor is shown; the inset is a magnification of a syncytium with its cytoplasm stained with eosin and its hematoxylin-stained nuclei in the periphery.

MV-NIS Therapy Results in Highly Increased 123I Uptake in HCC, Allowing Noninvasive Imaging of HCC Tumors.

HUH-7 cells were implanted subcutaneously in SCID mice as described in the Material and Methods section. When the tumors were 0.5 cm in diameter, the mice were injected intravenously via the tail vein with a single dose of 2 × 106 TCID50 of MV-NIS or an equivalent dose of UV-inactivated MV-NIS. Seven and 14 days after MV-NIS treatment, the mice were injected intraperitoneally with 123I (500 μCi/mouse) and analyzed 1 hour later for isotope uptake using whole-body gamma-camera imaging. As shown in Fig. 8A, in MV-NIS-treated animals we observed a strong signal from the HCC tumors, indicating high intratumoral uptake of 123I. In contrast, UV-inactivated MV-NIS-treated mice did not concentrate 123I intratumorally. Viral gene expression and viral replication are tightly coupled. Study of serial iodide uptake can be expected to reflect NIS expression and serve as a surrogate for MV-NIS replication.13 Therefore, we repeated the experiment and performed serial imaging analysis 3, 7, 10, and 14 days after treatment with a single dose of MV-NIS. The serial imaging analysis showed that iodide uptake changed over time, with maximum uptake between days 7 and 10.

Figure 8.

Increased 123I-uptake and tumor suppression in HCC tumors after intravenous MV-NIS therapy. (A) Mice bearing subcutaneous HUH-7 tumor xenografts received a single i.v. injection of 2 × 106 TCID50 of MV-NIS. Then 7 and 14 days after MV-NIS treatment 0.5 mCi 123I was injected intraperitoneally. One hour after application of the isotope, animals were analyzed by whole-body gamma-camera imaging. Tumors, on the right flanks of the mice, show increased activity at the tumor site, indicating 123I uptake, which was not observed in the negative control mouse. Other organs showing activity were thyroid, stomach, and bladder. (B) Mice bearing s.c. HCC xenografts (HUH-7, Hep-3B) were injected intravenously with 2 × 106 TCID50 of MV-NIS every other day for a total of 5 injections (1 × 107 total TCID50/mouse). Each group had 10 mice. Treatment groups received active MV-NIS, and the negative control mice received UV-inactivated MV-NIS. Survival curves are shown for treated mice (solid line) and negative control mice (dashed line).

Quantitative analysis of intratumoral iodine uptake showed that median intratumoral 123I uptake was 8.41% (SE: ±1.43%) of total body 123I uptake on day 3 and 9.13% (SE: ±2.00%) on day 10.

Intravenous Therapy of Human HCC Xenografts With Recombinant MV-Edm Has a Potent Antitumor Effect.

Our next aim was to evaluate the tumor-suppressive potential of recombinant MV-Edm after intravenous therapy with a vector allowing monitoring of viral spread by noninvasive imaging techniques. Therefore, we chose MV-NIS for intravenous therapy of human HCC xenograft–bearing mice. Hep-3B and HUH-7 tumor xenografts were implanted in the Materials and Methods section. Once the tumors reached 0.5 cm in diameter, therapy was started. The therapy regimen consisted of injection via the tail vein of 5 doses of 2 × 106 TCID50 MV-NIS every other day. Mice in the control group were treated with equivalent doses of UV-inactivated MV-NIS. In the Hep-3B xenograft model, median survival of the treated mice was 41 days, 2.7 times the 15 days of the control mice (Fig. 8B), which was a significant increase (P < .0001). On day 22 all the mice in the control group had to be sacrificed. In the MV-NIS-treated group, we observed complete regression in 10% of the mice. In the HUH-7 tumor model, median survival of mice in the control and treated groups was 8 and 12 days, respectively. In HUH-7 xenografts intravenous MV-NIS treatment resulted in increase in the median survival by a factor of 1.5 (P = .0018).


This study showed the tumor-suppressive potential of recombinant MV-Edm as a viral cancer agent for therapy for HCC. Our data showed that: (1) CD46 was overexpressed in HCC compared to in nonmalignant hepatic tissue; (2) MV-Edm successfully infected human HCC cells, resulting in transgene expression, syncytium formation, and tumor cell killing; (3) MV-Edm treatment of IfnarKO×CD46 mice was well tolerated and did not cause any adverse effects; (4) recombinant MV-Edm had a strong tumor-suppressive effect on human HCC xenografts after intratumoral and intravenous therapy (Fig. 6B–C); and (5) recombinant MV-Edm expressing either CEA or hNIS enabled noninvasive monitoring of the kinetics of viral expression and localization of the infected tumor cells.

MV-Edm is a RNA virus characterized by its oncolytic potential and tumor selectivity.13–15, 39 One contributor to its tumor selectivity is CD46 receptor density.19 The pathogenic wild-type measles virus, which is not selectively oncolytic, uses primarily the SLAM receptor.40, 41 The main receptor used by the vaccine strain MV-Edm is CD46.42 Virus entry increases with increasing CD46 receptor density; if the latter exceeds a “threshold,” syncytium formation, and cell killing are induced.19 Our data showed that CD46 is strongly expressed in human HCC cell lines compared to in primary human hepatocytes. Unfortunately, culturing of primary human hepatocytes more than 48 hours resulted in up-regulation of the expression of CD46 receptors and cell fusion after MV-Edm infection (data not shown). Because long-term culturing of human HCC cell lines might also have resulted in up-regulated expression of CD46 receptors in these cell lines, immunohistochemical analysis of human hepatic tissue samples of HCC patients was performed. The results of these studies confirmed overexpression of CD46 in HCC tumor tissue compared to nonmalignant hepatic tissue. Our data are in concordance with an earlier study by Kinugasa et al.,43 which compared CD46 expression in normal human hepatic tissue, HCC, cirrhosis, and chronic hepatitis samples. Their results showed relative CD46 densities of 0.11 + 0.1 in normal hepatic tissue, 0.63 + 0.23 in HCC, 0.21 + 0.07 in cirrhosis, and 0.25 + 0.1 in chronic hepatitis. These data clearly show that CD46 expression is 6 times higher in HCC than in normal hepatic tissue, 3 times higher in HCC than in tissue with cirrhosis, and 2.52 times higher in HCC than in chronic hepatitis without overlap of CD46 density. Therefore, we concluded that HCC fulfills the requirements for viral uptake and selective cell fusion and killing.

It is likely that additional factors contribute to MV-Edm tumor selectivity. Virally infected cells respond with inhibition of viral protein synthesis through IFN-α/β, double-stranded RNA-dependent protein kinase, 2′,5′-oligoadenylate synthetase, and Mx proteins.44–46 In contrast to nonmalignant cells, tumor cells are often defective in IFN-α/β and RNA-dependent protein kinase pathways.19, 47 This impairment in cellular defense is thought to be one of the mechanisms underlying the tumor selectivity of other RNA viruses used in experimental cancer therapy.48–51 It might also be a mechanism contributing to the tumor selectivity of MV-Edm, in addition to other, as yet unknown mechanisms.

Tumor selectivity is of the highest importance in viral therapy in order to avoid damage to nonmalignant tissue. Selectivity is a safety concern, especially for those with HCC, as 60%-80% of HCC patients have reduced liver function because of cirrhosis.52 Therefore, toxicology of IfnarKO×CD46Ge mice was studied. This transgenic mouse strain expresses CD46 at a humanlike tissue specificity and allows systemic virus spread because of a mutation inactivating the type I α/β-interferon receptor.53, 54 IfnarKO×CD46Ge transgenic mice are susceptible to MV-Edm infection and have been accepted by the FDA as an in vivo toxicity model for oncolytic measles viruses (Russell SJ, personal communication). Our data show that systemic treatment of Ifnar×CD46Ge transgenic mice was well tolerated without any adverse effects. The only significant differences between the groups we observed were mildly increased total protein and decreased BUN in MV-NIS-treated mice. We attributed the rise in total protein to the production of antimeasles antibodies. The decrease in BUN was not accompanied by any pathological change in the kidneys. The following provides further support for the safety of MV-Edm as a viral agent: (1) liver toxicity was not observed in CD46-positive mice and squirrel monkeys treated systemically with MV-Edm (S. J. Russell, personal communication); (2) the attenuated Edmonston vaccine strain has successfully been used for vaccination with an excellent safety profile; (3) reports of wild-type measles virus–induced hepatic damage are extremely rare or have failed to show the intrahepatic presence of the measles virus55, 56; and (4) posttransplant measles vaccination of patients who have undergone liver transplantation does not result in significant complications and is considered safe.57, 58 The observation that CD46 expression is increased in cirrhosis, though still 3 times lower than in HCC, indicates clinical phase I toxicity studies are necessary.43, 52 In addition, our previous data showed that intraperitoneal treatment of Ifnar™-CD46Ge transgenic mice with MV-GFP did not result in significant transgene expression, despite the presence of MV-Edm RNA in liver tissue.59 These data support the safety and suggest the tumor selectivity of MV-Edm.

MV-Edm resulted in a strong CPE in vitro and in vivo. We used 2 human HCC cell lines, both of which were susceptible to the cytotoxic effect of MV-Edm but differed in the kinetics of cell death. Hep-3B cells were eliminated very efficiently and quickly, whereas the cytotoxic effect of MV-Edm on HUH-7 cells was observed later. Our in vivo data show that intratumoral MV-Edm therapy results in significant prolongation of survival, almost doubling the median survival time of mice bearing human HCC xenografts and causing complete tumor regression in up to one third of treated animals. Intravenous MV-NIS almost tripled median survival in mice bearing human HCC xenografts and resulted in complete tumor regression in 10% of the Hep-3B xenografts. Intratumoral MV-Edm therapy has been shown to result in limited intratumoral spread of MV-Edm.17 The longer median survival with intravenous therapy than with intratumoral therapy might be explained by better intratumoral MV-Edm distribution after intravenous delivery. Interestingly, the HCC cell lines used in our study showed varied susceptibility to the cytotoxic effect of MV-Edm, most likely resembling the situation in primary human tumors, which are composed of heterogeneous tumor cell populations.60 Both cell lines express comparable levesl of CD46 but differ in other aspects, for example, their p53 status.61, 62 Syncytium formation results in cell death either by apoptosis or a bioenergetic form of cell death with necrosis.16, 18 The exact mechanism of cell death in HCC cells after MV-Edm-induced syncytium formation is currently being investigated by our group. Our preliminary data indicate apoptosis as the mechanism of cell death. Differences in the components of these pathways or other not-yet-identified factors in the process of measles-induced cytotoxicity could explain the differences in susceptibility.

The less susceptible HUH-7 tumors showed higher transgene expression in vitro as well as in vivo than the more susceptible Hep-3B tumors. This observation might be explained by the delayed death of the HUH-7 cells; that is, the increased interval until cell death resulted in an overall increase in viral replication and expression. A similar observation made in an ovarian cancer model was explained by an intratumoral dynamic equilibrium between formation of new tumor cells and death of infected tumor cells.15 The optimal rate of cell killing depends on the growth rate of infected tumor cells relative to uninfected tumor cells.63 In the Hep-3B tumors, the equilibrium was shifted toward the killing of infected cells, resulting in efficient and rapid destruction of infected tumor tissue. In HUH-7 tumors, the virus would have been maintained and viral proteins expressed until cell death occurred. Infected tumor cells fuse with surrounding uninfected tumor cells, resulting in a strong bystander effect.39 Syncytium formation is CD46 dependent, resulting in fusion of tumor cells but not of their normal counterparts.19 Delayed cell death would therefore allow recruitment of more uninfected tumor cells in syncytia. In addition, the increased expression of transgenes in less susceptible cells would be expected to allow further enhancement of the therapeutic effect by the use of therapeutic transgenes, for example, hNIS in combination with 131I,13, 34 a beta-particle-emitting isotope with an average tissue-path length of 0.4 mm, resulting in a bystander effect. HCC is susceptible to the therapeutic effect of 131I.64, 65 One of our future aims is to explore the possibility of further enhancement of the therapeutic effect of MV-Edm by combined radiovirotherapy using MV-NIS and 131I.

Another factor influencing the therapeutic outcome of MV-Edm therapy is the immune response. The CTL response and humoral antibodies influence the dynamics of virus persistence, infection, and cell death.63 It is likely that virally infected tumor cells are recognized by the host's immune system and are eliminated, thereby contributing to the therapeutic effect.

A general concern in systemic viral therapy is the potential for decreased efficacy because of the presence of antibodies. HCC offers the possibility of localized treatments. Most patients with HCC do not have extrahepatic metastasis, and the liver is easily accessible using ultrasound or CT guidance.66 HCC tumor nodules can easily be localized and injected with a therapeutic agent. The liver offers the opportunity for intrahepatic delivery by hepatic artery infusion in the treatment of multilocular HCC. In either case, the presence of anti-MV antibodies is not expected to significantly decrease efficacy. In a mouse model in which mice received passive transfer of anti-MV antibodies, Grote et al. found that intratumoral MV-Edm therapy of human lymphoma xenografts resulted in effective tumor regression without compromise through the presence of anti-MV antibodies.17 These findings were in accordance with the results of a study of intratumoral therapy with a reovirus in immune-competent C3H mice and of a clinical phase II study of a genetically modified adenovirus in patients with advanced head and neck cancer. In both studies, effective tumor regression was achieved despite the presence of antibodies.51, 67 Also, it has been shown that with hepatic artery infusion as the route of administration, repeated infusion of replicating viral agents resulted in sufficient delivery of the virus to elicit therapeutic effects.68

A major drawback of many cancer agents is the lack of convenient methods of monitoring the agent after application to the patient. We have shown in this study that MV-CEA and MV-NIS allow noninvasive tracking of viral gene expression in HCC as well as localization of infected tumor tissue. Our in vivo investigation showed transgene expression up to 58 days after MV-CEA therapy and stable iodine uptake up to 14 days after MV-NIS therapy, which was the end point of the study. The expression profile of the marker gene for MV-CEA was shown to correlate with viral gene expression and the serum concentration of the marker polypeptide to reflect the number of viable cells.25 Expression of viral proteins lasts for the duration of viability of an infected cell. Once the cell has reached the state of cell death, viral gene expression continues through infection of new cells with progeny virus. Only 15% of HCC patients express CEA. Therefore, MV-CEA would be well suited for most of these patients, allowing noninvasive monitoring of viral expression kinetics by serum CEA analysis. MV-NIS combined with 123I enables infected HCC cells to be localized by noninvasive imaging techniques. Analysis of these data will help to determine the pharmacokinetics and pharmacodynamics of measles therapy in HCC; this information is essential for optimizing the safety and efficacy of therapy protocols.

In conclusion, we have shown that trackable MV-Edm is a novel and very potent therapeutic agent for HCC. It has an excellent safety profile and offers the possibility of controlling viral expression kinetics and localizing infected cells by noninvasive methods. The potential of monitoring measles therapy will allow modification of protocols to enhance safety and efficacy. The accessibility of the liver and the relatively low rate of extrahepatic metastasis in HCC qualify HCC as an ideal target. Therefore, recombinant MV-Edm should further be explored as a therapeutic agent for HCC. Hence, we plan to evaluate trackable MV-Edm in a clinical phase I trial.