Research ArticleCancer Immunotherapy

Oncolytic and Immunotherapeutic Vaccinia Induces Antibody-Mediated Complement-Dependent Cancer Cell Lysis in Humans

See allHide authors and affiliations

Science Translational Medicine  15 May 2013:
Vol. 5, Issue 185, pp. 185ra63
DOI: 10.1126/scitranslmed.3005361


Oncolytic viruses cause direct cytolysis and cancer-specific immunity in preclinical models. The goal of this study was to demonstrate induction of functional anticancer immunity that can lyse target cancer cells in humans. Pexa-Vec (pexastimogene devacirepvec; JX-594) is a targeted oncolytic and immunotherapeutic vaccinia virus engineered to express human granulocyte-macrophage colony-stimulating factor (GM-CSF). Pexa-Vec demonstrated replication, GM-CSF expression, and tumor responses in previous phase 1 trials. We now evaluated whether Pexa-Vec induced functional anticancer immunity both in the rabbit VX2 tumor model and in patients with diverse solid tumor types in phase 1. Antibody-mediated complement-dependent cancer cell cytotoxicity (CDC) was induced by intravenous Pexa-Vec in rabbits; transfer of serum from Pexa-Vec–treated animals to tumor-bearing animals resulted in tumor necrosis and improved survival. In patients with diverse tumor types treated on a phase 1 trial, CDC developed within 4 to 8 weeks in most patients; normal cells were resistant to the cytotoxic effects. T lymphocyte activation in patients was evidenced by antibody class switching. We determined that patients with the longest survival duration had the highest CDC activity, and identified candidate target tumor cell antigens. Thus, we demonstrated that Pexa-Vec induced polyclonal antibody–mediated CDC against multiple tumor antigens both in rabbits and in patients with diverse solid tumor types.


New cancer therapies with complementary mechanisms of action are needed. Active immunotherapy is a promising field with two recent product approvals (1): the anti–CTLA-4 antibody ipilimumab (Yervoy, Bristol-Myers Squibb) (2, 3) and the autologous dendritic cell product sipuleucel-T (Provenge, Dendreon) (4). Numerous hurdles remain, however, including autoimmune toxicities, labor-intensive patient-specific manufacturing, immune evasion, and inability to debulk large established tumors. In addition, broad-based therapeutic cancer vaccines should ideally present multiple tumor antigens, induce multiple cytokines and danger signals within the tumor microenvironment, and trigger immune cell recruitment and activation within tumors.

Targeted oncolytic and immunotherapeutic viruses have the potential to overcome these active immunotherapy hurdles. In contrast to replication-deficient viruses developed for cancer gene therapy (57) and vaccines (810), oncolytic viruses infect, multiply within, and subsequently lyse cancer cells within tumors (1115). First-generation oncolytic viruses were inherently cancer-selective (14) [for example, reovirus (16) and vesicular stomatitis virus (VSV) (17, 18)], whereas second-generation agents were engineered for cancer selectivity [for example, adenovirus (11, 19, 20) and herpes simplex virus (HSV) (21, 22) deletion mutants]. Safety and cancer selectivity were demonstrated clinically, but therapeutic potency and intravenous delivery to metastatic tumors were limited (14, 23, 24). The third-generation transgene-armed HSV expressing human granulocyte-macrophage colony-stimulating factor (hGM-CSF) (T-Vec, Amgen) was associated with an increase in melanoma-associated antigen recognized by T cells (MART-1)–specific T cells in tumors, but functional antitumoral immunity was not assessed (25); additionally, HSV is unable to target tumors after intravenous infusion due to complement inactivation. Preclinical studies have demonstrated that oncolytic viruses can induce functional cancer-specific immunity (2630), but data from clinical trials have been lacking.

Pexa-Vec (pexastimogene devacirepvec; JX-594) is a fourth-generation (31) targeted and armed oncolytic and immunotherapeutic vaccinia virus with disruption of the viral thymidine kinase gene (which makes replication dependent on the high cellular thymidine kinase activity that is a hallmark of cancer cells) and expression of the hGM-CSF and β-galactosidase transgenes under control of the synthetic early-late and p7.5 promoters, respectively (26, 31); this oncolytic virus species can be delivered to tumors intravenously (32, 33). We hypothesized that Pexa-Vec would induce cancer vaccination through simultaneous (i) cancer cell lysis and endogenous tumor antigen release, accompanied by (ii) expression of hGM-CSF to support antigen-presenting cell activation, (iii) Toll-like receptor 2 and 8 activation by vaccinia (3436), and (iv) recruitment of immune effector cells and proinflammatory cytokine induction. The utility of Pexa-Vec and other immunostimulatory viruses, and the design of future oncolytic products, would be improved markedly if the functional anticancer immune responses in treated patients could be elucidated.

We therefore evaluated induction of active cancer immunotherapy initially in the rabbit VX2 tumor model (vaccinia- and hGM-CSF–sensitive) and subsequently in patients with liver tumors of diverse histologies on a phase 1 trial. In phase 1, Pexa-Vec demonstrated cancer-specific replication, hGM-CSF expression, white blood cell stimulation, and objective cancer responses after intratumoral injections; non-injected tumors were also targeted (37). Antibody-mediated complement-dependent cytotoxicity (CDC) is a potent mechanism of cell killing (38), and selective CDC activity against tumor cell lines is a direct measure of functional systemic anticancer immunity. This activity could be studied in archival blood samples from all patients on the phase 1 trial at multiple time points. In addition, CDC can be an important determinant of therapeutic antibody-mediated efficacy in cancer (3945). We therefore assessed Pexa-Vec induction of antibody-mediated CDC in an animal tumor model and in patients’ blood against a panel of tumor cell lines of different histologies over time. Serological analysis of recombinant tumor complementary DNA (cDNA) expression libraries (SEREX) technology was then used to identify potential target tumor antigens.


Pexa-Vec treatment induces antibody-mediated CDC in VX2 tumor–bearing rabbits

We investigated the induction of CDC in the VX2 tumor model (rabbit anaplastic squamous cell carcinoma); oncolytic vaccinia (including Pexa-Vec) is known to replicate in this model (table S1) (26, 46); furthermore, hGM-CSF expressed from Pexa-Vec is active in this model (26, 46). Serum was collected over 28 days from tumor-bearing rabbits treated with Pexa-Vec or phosphate-buffered saline (PBS) control or from non–tumor-bearing controls treated with Pexa-Vec. Serum collected at day 28 after injection was added to VX2 cells or rabbit peripheral blood mononuclear cells (PBMCs) in vitro at a concentration of 3%. A significant decrease in VX2 cell viability was only observed with serum from Pexa-Vec–treated rabbits bearing VX2 tumors (P < 0.0001) (Fig. 1, A and B). Incubation of VX2 cells with serum from either VX2 tumor–bearing rabbits treated with PBS or non–tumor-bearing rabbits treated with Pexa-Vec did not result in decreased cell viability (Fig. 1, A and B). In contrast, PBMC viability did not decrease significantly upon incubation with serum from any treatment group (Fig. 1A). Cancer cell viability was subsequently assessed upon incubation with serum collected at various time points after Pexa-Vec (or PBS) injection. Cell viability was decreased by incubation with serum collected on day 18 or later after initiation of Pexa-Vec treatment (Fig. 1B). The reduction in cell viability was dependent on serum dose, with decreases in cancer cell viability after incubation with serum diluted to 1% in both dividing cells (grown at 50% density) and nondividing cells (grown to 100% density; Fig. 1C). Western blotting was performed with serum from Pexa-Vec–naïve and Pexa-Vec–treated rabbits (week 6) against lysates from VX2 tumor cells (Fig. 1D). Multiple cellular antigens were recognized by antibodies in serum. We next assessed whether immunoglobulin G (IgG) antibody and complement were required for cytolysis in the presence of Pexa-Vec–treated rabbit serum. Purified IgG isolated from serum collected from PBS- or Pexa-Vec–treated rabbits was incubated on target cells in the presence or absence of naïve rabbit complement. Decreased cell viability was only detected in the presence of both complement and IgG collected from Pexa-Vec–treated rabbits (Fig. 1E).

Fig. 1 Antibodies mediating CDC are induced in rabbit VX2 tumor model.

(A) Mean cell viability (±SD) of VX2 cells and rabbit PBMCs upon incubation with 3% pooled serum collected 3 and 6 weeks after PBS or Pexa-Vec treatment of VX2 tumor–bearing rabbits. Serum collected from Pexa-Vec–treated non–tumor-bearing rabbit included as control (n = 3 for each condition tested in triplicate). (B) Mean cell viability (±SD) of VX2 cells upon incubation with 3% rabbit serum. Pexa-Vec was injected on days 0 and 7, and serum was collected at the indicated time points (n = 3 for each condition tested in triplicate). (C) Mean cell viability (±SD) of VX2 cells plated at 50 and 100% cell density after incubation with rabbit serum at concentrations between 0 and 5%. Experiment performed in duplicate. (D) Western blot of VX2 cell lysate probed with serum collected from naïve or Pexa-Vec–treated rabbit (week 6). (E) Viability of A2780 human ovarian cancer cells incubated with IgG isolated from rabbit serum (rabbit treated with PBS or Pexa-Vec). IgG antibody was incubated in the presence or absence of complement, and normal rabbit serum was used as the complement source. n = 2 per group.

Transfer of Pexa-Vec–induced CDC activity increases tumor necrosis and improves survival in rabbits

We next investigated whether CDC-mediating antibodies exhibited antitumor activity in vivo in the absence of Pexa-Vec. Serum was collected from Pexa-Vec–treated VX2 tumor–bearing rabbits 2 weeks after dosing with the virus. Pexa-Vec clearance from serum at this time point was confirmed by quantitative polymerase chain reaction (qPCR) analysis (table S2). Serum was subsequently transferred to naïve VX2 tumor–bearing rabbits. Control animals received serum from PBS-treated VX2-bearing rabbits. Active serum transfer resulted in increased tumor necrosis over time as determined by (i) gross tumor assessment (Fig. 2A), (ii) quantitative histopathology analysis (Fig. 2, B and C), and (iii) computed tomography (CT) scan (Fig. 2, D and E). Serum collected from Pexa-Vec–treated rabbits induced a significant increase in necrosis on histopathology analysis versus control serum (Fig. 2C; P < 0.0001); notably, the baseline necrosis in control tumors was due to outgrowth of tumor blood supply as previously described (47). Consistent with the histopathology findings of necrosis, a substantial reduction in intratumoral enhancement was observed on CT imaging in Pexa-Vec serum-treated tumors; variable enhancement in tumors before serum treatment initiation was attributed to rapid tumor growth. Finally, VX2 tumor–bearing rabbits that received serum from Pexa-Vec–treated rabbits generally exhibited weight gain or maintenance compared to controls who lost weight (Fig. 2F; n = 7 subset) and prolonged survival versus controls (Fig. 2G; P = 0.04). The absence of Pexa-Vec in tumors from serum-recipient rabbits was confirmed by qPCR analysis (table S1).

Fig. 2 Serum from Pexa-Vec–treated rabbits has antitumor activity.

(A) Images of tumors collected 21 days after treatment with serum from PBS- or Pexa-Vec–treated rabbits. (B) Hematoxylin and eosin staining of tumor tissue sections collected from PBS- or Pexa-Vec–treated rabbits. Scale bars, 200 μm. (C) Mean percent necrosis (±SD) in hematoxylin and eosin–stained sections of tumors collected from rabbits receiving serum from PBS- or Pexa-Vec–treated rabbits. (n = 3 rabbits per group, three random fields quantified per tumor). (D) Representative contrast-enhanced CT images of VX2 tumors at baseline, showing tumor dimensions and intratumoral enhancement (corresponding to viable tissue). (E) Representative contrast-enhanced CT images of VX2 tumors collected 23 days after serum transfer from PBS- or Pexa-Vec–treated rabbits. (F) Body weight of VX2 tumor–bearing rabbits after treatment initiation with serum collected from PBS-treated (gray) or Pexa-Vec–treated (black) rabbits, expressed as percentage of baseline weight. (G) Overall survival of VX2 tumor–bearing rabbits receiving serum from PBS-treated (gray) or Pexa-Vec–treated (black) rabbits.

CDC activity in serum from Pexa-Vec–treated patients is cancer-selective

We sought to identify CDC activity in serum collected from Pexa-Vec–treated patients over time. We initially tested serum from three patients with cancers refractory to conventional treatment who were part of a phase 1 trial of Pexa-Vec treatment in patients with primary liver cancer or liver metastases from other cancer types (37). All three of these patients had durable objective tumor responses and long-term survival after Pexa-Vec therapy: patient #301 had clear cell renal cancer and survived for 76+ months; patient #103 had small cell lung cancer and survived for 24.5 months; and patient #304 had melanoma and survived for 12 months (37) (table S3). Incubation of multiple types of cancer cell lines with 5% serum from Pexa-Vec–treated patients decreased the viability of some of the cell lines (Fig. 3, A to E). When cancer cell lines of the same origin as the patients’ cancers were tested, a decrease in cell viability was observed in most cell lines (Fig. 3, A to C); this effect increased over time after Pexa-Vec treatment initiation (Fig. 3, D and E). We also investigated whether serum from Pexa-Vec–treated patients was capable of causing toxicity to normal human cells ex vivo or to cancer cell lines of different histologies (that is, whether shared tumor antigens were being targeted by CDC activity). Normal primary HUVECs (human umbilical vein endothelial cells) and MRC-5 lung fibroblasts did not exhibit decreased cell viability when incubated with serum from any of the three patients tested (comparison of HUVECs or MRC-5 normal cells versus cancer cell with greatest decrease in cell viability, P < 0.005) (Fig. 3, A to C). Generally, tumor cell sensitivity was greatest for cell lines of the same tumor tissue histology as the patients’ tumor type, but isolated tumor cells of different histologies were also sensitive to serum-mediated cytolysis, suggesting shared target antigens (Fig. 3, A to C).

Fig. 3 Antitumor activity is detected in serum from cancer patients treated with Pexa-Vec.

(A) Mean cell viability (±SD) of human tumor and normal cell lines upon incubation with 5% serum from a renal cell cancer (RCC) patient, collected 92 days after treatment with Pexa-Vec on a phase 1 clinical trial (#301). (B) Mean cell viability (±SD) of human tumor and normal cell lines upon incubation with 5% serum from a melanoma patient, collected 92 days after treatment with Pexa-Vec on a phase 1 clinical trial (#304). (C) Mean cell viability (±SD) of human tumor and normal cell lines upon incubation with 5% serum from a non–small cell lung cancer (NSCLC) patient, collected 92 days after treatment with Pexa-Vec on a phase 1 clinical trial (#103). (D) Cell viability of three RCC cell lines after incubation with serum collected from an RCC patient as in (A). (E) Cell viability of four lung cancer cell lines after incubation with serum collected from a lung cancer patient as in (C). (F) Representative confocal microscopy of human RCC SNU-349 cell line after incubation with 5% human serum from RCC patient #301, collected at day 92, after four cycles of Pexa-Vec. Carboxyfluorescein diacetate succinimidyl ester (CFSE) can be seen as green fluorescence in whole live cells, and 7-amino-actinomycin D (7-AAD) is seen as red fluorescence in the nuclei of dead cells. Scale bar, 60 μm. (G) Western blot of tumor cell lysates probed with serum collected from RCC patient in (A) at the indicated time points.

Induction of cell death was visualized by fluorescence imaging of CFSE (which labels intact cells) and 7-AAD (which labels DNA of membrane-compromised dying cells). Fluorescence staining at baseline confirmed cell viability; however, cell death was demonstrated by as early as 30 min after serum exposure (Fig. 3F). Western blotting was performed with serum from patient #301 (renal cell) against lysate from a sensitive renal cancer cell line (Fig. 3G) and showed that multiple cellular antigens were recognized. The timing of induction of this immune reactivity for new antigens was consistent with the timing of functional CDC induction (Fig. 3D).

Decreased cancer cell viability is due to antibody-mediated CDC in patient serum

We next tested the contribution of antibodies and complement to the serum-mediated cancer cell cytotoxicity from Pexa-Vec–treated patients (Fig. 4A). We tested the following five conditions, labeled as serum A to E: (A) baseline serum (before Pexa-Vec treatment), (B) serum obtained 92 days after Pexa-Vec treatment initiation, (C) heat-inactivated serum B (inactivated complement), (D) serum C with fresh untreated serum added (reintroduction of active complement), and (E) serum B that was passed through an IgG removal column. These sera were added to cancer cell line monolayers at a concentration of 5%. Baseline serum (A) did not result in decreased cell viability, whereas serum collected 92 days after Pexa-Vec initiation (B) exhibited potent antitumoral activity (Fig. 4B). Heat inactivation (C) or IgG depletion (E) of serum prevented cell death. Furthermore, restoration of functional complement in serum C resulted in restoration of cytolytic activity (Fig. 4B). Results were similar in serum from all three phase 1 patients evaluated (Fig. 4B). Finally, increasing duration of heat inactivation correlated with decreased CDC activity (fig. S1).

Fig. 4 Pexa-Vec antitumor activity in patient serum is mediated by complement and antibodies and correlates with survival.

(A) Overview of sera used for complement and antibody depletion experiment. B serum was collected 92 days after Pexa-Vec treatment. E serum was generated by filtering B serum through an IgG removal column. (B) Mean cell viability (±SD) of human tumor cell lines upon incubation with different preparations of 5% serum, as outlined in (A). RCC patient (phase 1 #301) serum incubated with SNU-349 cells, lung cancer patient (phase 1 #103) serum incubated with H460 cells, and melanoma patient (phase 1 #304) serum incubated with WM266-4 cells. (C) The graph shows correlation between overall survival (days from Pexa-Vec treatment initiation to death) versus percent viability of A2780 cells after incubation with patient serum collected after Pexa-Vec treatment (Spearman correlation coefficient r = −0.81, P = 0.0005). The results are normalized to cell viability after incubation with patient serum collected at baseline.

CDC induction is associated with prolonged survival in phase 1 trial

Patients’ survival duration on the phase 1 trial was plotted against CDC activity targeting the cancer cell line with the broadest sensitivity to CDC across all patients (A2780 ovarian cancer). Although these patients had diverse cancer types and disease burdens at baseline, the patients who survived the longest had the highest CDC activity against this tumor cell line (Spearman correlation coefficient r = −0.81, P = 0.0005; Fig. 4C). The three patients with the longest survival duration had the best radiographic responses as well (table S3).

Antibodies induced by Pexa-Vec treatment can be used to identify cancer-associated antigens

To determine the feasibility of identifying tumor cell antigens targeted by antibodies in these patients, exploratory SEREX was performed on pooled serum exhibiting CDC activity after Pexa-Vec treatment according to standard methods (48). Serum for SEREX analysis was collected from a uniform patient population with primary hepatocellular carcinoma (n = 7) treated on a phase 2 study, with antigens derived from a hepatocellular carcinoma cell line. Validated tumor surface antigens, as well as intracellular cancer–related proteins, were identified in the screen. Examples included CD24 (cancer stem cell marker) (49) and leptin receptor (overexpressed in some cancers) (50) (Table 1), in addition to intracellular antigens, consistent with recent evidence that suggests that antibodies can recognize intracellular antigens and contribute to anticancer activity (51).

Table 1 Cancer-associated antigens identified in SEREX screen of serum from Pexa-Vec–treated patients.

NCBI, National Center for Biotechnology Information; HCC, hepatocellular carcinoma; GPCR, heterotrimeric guanine nucleotide–binding protein–coupled receptor; GTPase, guanosine triphosphatase.

View this table:


We sought to assess the induction of functional anticancer immunity in humans with diverse tumor types after treatment with an oncolytic and immunotherapeutic virus on a phase 1 trial. We describe here (i) the induction of functional anticancer immunity by an oncolytic virus in humans; (ii) induction of antibody-mediated CDC by a therapeutic virus in humans and in an animal tumor model; (iii) the antitumor efficacy of serum transferred to naïve animals; (iv) the ability of this agent to target a diverse array of tumor types in patients, without reliance on expression of a defined target antigen; and finally, (v) the potential to identify the tumor antigens bound by antibodies from treated patients. These findings are important because they validate the feasibility of active oncolytic immunotherapy in patients with diverse and common solid tumor types. The finding that serum concentrations of only 5% can kill tumor cells ex vivo suggests that this activity is potent and likely is clinically relevant; likewise, the antitumor efficacy of transferred serum in animals supports the therapeutic effect of the induced CDC. In antibody depletion experiments, we demonstrated that antibodies of IgG isotype mediated CDC, providing evidence for engagement of T helper functions (52, 53). In addition, the CDC assay may give investigators a useful pharmacodynamic tool to study this mechanism of action with oncolytic viruses, including Pexa-Vec and T-Vec. Finally, analyses of the binding targets for these antibodies may result in the identification of antibody therapeutics and novel targets for monoclonal antibodies and/or active immunotherapeutics. In summary, these findings have implications for oncolytic virus approaches, cancer vaccines, monoclonal antibodies, and cancer immunotherapies.

Several questions remain and should be addressed in future studies. First, the use of tumor cell lines potentially underestimates the efficacy of CDC because many targeted antigens are likely not shared between patients’ own cancer cells and established cancer cell lines. Ideally, in the future, CDC would be assessed against patients’ own tumor cells grown ex vivo. The relevance of CDC in humans should also be assessed in larger phase 2b to 3 trials. Rituximab (Genentech) has been shown to be recognized by complement and trigger CDC of target cells (3942). Furthermore, ofatumumab (Arzerra), another anti-CD20 monoclonal antibody that has been modified for increased complement activation capacity (43, 44), has demonstrated promising antitumor activity, including in rituximab-refractory patients (45). The importance of this mechanism of action versus the viral oncolytic effect will likely vary depending on the patient population and tumor type being targeted. Notably, the patients on this phase 1 trial had different baseline prognoses and tumor types, and they received varying doses of Pexa-Vec. Therefore, definitive correlations between CDC induction and patient outcome cannot yet be made. In addition, antibody-dependent cellular cytotoxicity (ADCC) and cell-mediated immunity induction by Pexa-Vec should be assessed in future trials. Finally, the role of GM-CSF expression from Pexa-Vec in induction of antibodies mediating CDC must be determined in future studies.

The use of replication-competent, oncolytic vaccinia virus as an anticancer immunotherapy has potential advantages over other active immunotherapy strategies: (i) the induction of potentially diverse anticancer immune responses; (ii) tumor debulking via direct infection and lysis of tumor cells and acute shutdown of tumor vasculature (30, 54); (iii) no ex vivo manipulation steps required for the therapy (unlike autologous cellular therapy); (iv) a patient-specific immune response; and (v) recruitment and activation of immune effector cells within the tumor (55, 56). Future oncolytic and immunotherapeutic vaccinia viruses may encode additional immunomodulatory cytokines, chemokines, and tumor antigens in a single product.

Materials and Methods

Viruses and cell lines

Pexa-Vec, Wyeth strain vaccinia virus (thymidine kinase–inactivated, expressing hGM-CSF), was used throughout this study and was prepared as published previously (26). SNU-349, SNU-482, and SNU-267 [human renal cell carcinoma; Korean Cell Line Bank (KCLB)], SNU-739 (human hepatocellular carcinoma; KCLB), and HT26 (human colon cancer; KCLB) were cultured in RPMI 1640 (HyClone) supplemented with 10% fetal bovine serum (FBS) (HyClone) with penicillin and streptomycin (HyClone). WM266-4, SK-MEL-2, and SK-MEL-5 [human melanoma; American Type Culture Collection (ATCC)] were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (HyClone) supplemented with 10% FBS (HyClone) with penicillin and streptomycin. HOP62, H157, H460, NCI-H23, A427, A549, and PC10 (human lung carcinoma; ATCC), A2780 (human ovarian cancer; ATCC), and MRC-5 nontransformed cells (lung fibroblast; ATCC) were cultured in DMEM containing 10% FBS with penicillin and streptomycin. HUVECs (ATCC) were cultured in endothelial cell medium EBM-2 (Lonza) supplemented with 2% FBS with penicillin and streptomycin. VX2 cells were isolated from VX2 tissues enzymatically (details below) and maintained in vitro in DMEM with 10% FBS for eight passages. The VX2 rabbit tumor cell line originated from Shope virus–induced papilloma-derived squamous cell carcinoma and was established after 72 transplantation passages (57).

Cell viability and CDC assay

All patients gave written informed consent according to guidelines on good clinical practice (NCT00629759; NCT00554372). CDC activity was assessed by measuring cell viability upon incubation with indicated concentrations of serum in 96-well plates. Cell viability in serum after Pexa-Vec administration was normalized to the cell viability of rabbit or patient serum at baseline (before Pexa-Vec treatment). Each cell line was seeded onto 96-well plates and incubated overnight. Cells were subsequently incubated with DMEM (no FBS) and the serum sample at 37°C for 4 hours. Cells were subsequently exposed to PBS and 10 μl of Cell Counting Kit-8 solution (CCK-8 kit, Dojindo Inc.) and incubated at 37°C for 2 hours. Cell viability was measured by optical density (OD) at 450 nm. For complement inhibition, serum was heat-inactivated by incubation at 56°C for 30 min. IgG antibody was purified from rabbit serum with the Melon Gel IgG Spin Purification Kit (Thermo Scientific) and human serum with the ProteoExtract Albumin/IgG Removal Kit (Calbiochem) per the manufacturer’s instructions.

Rabbit VX2 tumor model and isolation of VX2 cells

Rabbits were housed, cared for, and used in experiments as approved by the Ethical Committee for Animal Study at Pusan National University Hospital. VX2 cells were maintained by successive transplants of tumor cells into the muscle in the hindlimbs of carrier inbred New Zealand white rabbits (Samtako). Pexa-Vec [1 × 109 plaque-forming units (PFU)] or PBS was injected intravenously at 3 and 4 weeks after VX2 fragment implantation into skeletal muscle. Serum was collected at baseline and at 3 or 6 weeks after Pexa-Vec or PBS treatment. VX2 cells were isolated as described previously (58). In brief, VX2 cells were enzymatically isolated from VX2 tissues (0.01% collagenase and 0.1% protease overnight at 4°C) and maintained in vitro with DMEM (HyClone) with 10% FBS for eight passages. Fresh VX2 cells were used for each cell viability test. In a parallel study, Pexa-Vec was injected into normal, non–tumor-bearing rabbits, and serum was obtained.

Blood collection and storage

Blood was collected in the Vacuette Z Serum Sep Clot Activator (Greiner Bio One). Tubes were gently inverted several times immediately after blood collection and were subsequently kept upright at room temperature for 30 min to allow for clot formation. Serum was separated by centrifugation at 1800g for 10 min at 4°C, transferred to cryovials in 1-ml aliquots, and then stored at <−70°C until testing.

Western blotting

VX2 cells, rabbit PBMCs, or SNU-349 cells were lysed at ~1 × 106 cells/ml in PRO-PREP protein extraction solution (iNtRON Biotechnology) on ice for 30 min. After centrifugation, 50 μg of protein was separated on SDS–polyacrylamide gel electrophoresis gels and transferred to polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore). Sera were diluted 1:100 in 0.1% TBST (5% skim milk powder, 0.1% Tween 20, 50 mM tris, 150 mM NaCl) and incubated on PVDF membranes for 90 min at room temperature. The membranes were then incubated for 1 hour at room temperature with horseradish peroxidase–conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology) diluted 1:1000 in 0.1% TBST (rabbit serum primary) or anti-human IgG (Sigma, #A1543, 1:5000) (human serum primary) and visualized by enhanced chemiluminescence (ECL kit; Pierce).

VX2 serum transfer experiments

Transfer of serum from naïve/PBS-treated rabbits and rabbits treated with Pexa-Vec was performed two independent times. For both experiments, serum was harvested from rabbits treated with 1 × 109 PFU Pexa-Vec or PBS intravenously weekly for 2 weeks. Two weeks after the second Pexa-Vec treatment, rabbits were euthanized and serum was collected. VX2 tumors were then implanted intramuscularly in naïve rabbits (n = 2). Once tumors were palpable, the animals were treated with 1 ml of serum diluted in 4 ml of PBS intravenously twice a week until death (first experiment) or 2 ml of serum twice a week until death (second experiment). CT scans were performed with a 128-section CT unit (Somatom Definition AS Plus; Siemens Healthcare) with the following parameters: tube voltages of 120 kVp, effective tube current of 90 mA, field of view of 146 mm, and reconstruction thickness of 2 mm at 2-mm intervals. CT scans at baseline and 23 days after serum treatment initiation were performed on a subset of animals. The CT protocol included the acquisition of non-enhanced images and subsequent acquisition of arterial, venous, and delayed-phase image series after the intravenous bolus injection of 8 to 9 ml of nonionic iodinated contrast material [300 mg of iodine per milliliter of iohexol (Omnipaque; GE Healthcare AS), 2 ml/kg, 2.4 to 2.7 g of iodine] at a rate of 2 ml/s via an ear vein. Arterial phase imaging was obtained 10 s after achieving enhancement of the descending aorta to 100 Hounsfield units, as measured with the bolus tracking technique. Venous phase imaging was obtained 10 s after completion of the arterial phase, and delayed-phase imaging was obtained 70 s after venous phase was completed. Rabbits were weighed every 4 days from serum treatment initiation until death.

Tumor necrosis quantification

Histopathology analysis was performed on all tumors harvested. Tumors were collected, formalin-fixed, and paraffin-embedded. Sections (4 μm) were cut and stained with hematoxylin and eosin stains. Percent tumor necrosis was quantified by a trained pathologist on three random fields per section at ×40 magnification.

Fluorescence and confocal microscopy

Each cell line was plated into six-well plates, and cells were incubated for 24 hours to reach 100% cell density. For fluorescence staining, SNU-349 cells were seeded into a coverglass-bottom dish at 3 × 105 cells and left overnight. CFSE and 7-AAD (ACT 1 Assay for CytoToxicity, Cell Technology) were added to stain viable and dying cells, respectively. In live cells, CFSE can be detected as green fluorescence in whole cells, and red fluorescence (7-AAD) can be detected in the nuclei of dead cells. Fluorescence was imaged with the FluoView 1200 confocal microscope (Olympus).

Immunoscreening of the cDNA expression library

Fresh XL1-Blue MRF′ (600 μl) of OD600 = 0.5 was infected with recombinant phages from the primary library (SNU-739 cell line) and diluted in SM buffer [0.1M NaCl, 0.01 M MgSO4, 0.05 M tris-Cl (pH 7.5), 0.1% gelatin]. The mixture was incubated at 37°C for 20 min and then plated with 8 ml of top agar containing kanamycin (15 μg/ml) (Sigma-Aldrich) onto agar plates. Phage plaques appeared after 6 to 8 hours of incubation at 42°C and then transferred onto double nitrocellulose membranes (Millipore), which were presoaked in 10 mM isopropyl-β-D-thiogalactopyranoside (United States Biological) for 30 min. The membrane was oriented by cutting triangles in different places on the membrane and the agar plate. The nitrocellulose membranes were uncovered and then blocked with 5% bovine serum albumin (Santa Cruz Biotechnology). Preadsorbed HCC serum was added as primary antibody for screening, 1:10,000 dilution of alkaline phosphatase–conjugated goat anti-human IgG (Sigma-Aldrich) was the secondary antibody, and the membrane was developed with premixed bromochloroindolyl phosphate–nitro blue tetrazolium solution (Sigma-Aldrich). Only spots that appeared blue on double membranes were considered positive. These plaques were cored from the agar plate corresponding to the membranes and put into SM buffer containing 20 μl of chloroform. All the positive plaques were purified to monoclonality by another two rounds of screening similar to the first round.

Sequencing and sequence analysis

The positive clones were sent to be sequenced by Macrogen Inc. The analysis of the cDNA sequences was performed with BLAST program to search for homologs.

Statistical analysis

The differences in degree of necrosis in VX2 tumors as well as differences in cell viability were compared with the unpaired t test. The Spearman correlation coefficient was calculated in GraphPad Prism 5 software. The statistical analysis for overall survival was performed with Kaplan-Meier estimates and the log-rank test with GraphPad Prism 5 software.

Supplementary Materials

Fig. S1. Cell viability correlates with the extent of complement inactivation.

Table S1. qPCR data confirm the presence of Pexa-Vec in Pexa-Vec–treated VX2 tumors and the absence of Pexa-Vec in serum-treated tumors.

Table S2. qPCR data confirm the absence of Pexa-Vec in immune serum.

Table S3. Phase 1 trial patient outcomes correlate with CDC induction.

References and Notes

  1. Funding: These studies were supported by a grant of the Korean Healthcare Technology R&D project, Ministry for Health, Welfare and Family Affairs, Republic of Korea (A091047). Treatment of patients with Pexa-Vec was funded by Jennerex Inc. and Green Cross Corporation. Author contributions: M.K.K., C.J.B., A.M., J.C.B., D. H. Kirn, and T.-H.H. designed the experiments. M.K.K., Y.K.L., and S.-G.K. performed preclinical experiments. J.H., M.C., and B.H.P. managed Pexa-Vec–treated patients. J.W.L. and D. H. Kang evaluated rabbit CT images. M.K.K., C.J.B., D. H. Kirn, and T.-H.H. drafted the manuscript. Competing interests: C.J.B., A.M., and D. H. Kirn are employees of Jennerex Inc. J.H. and M.C. have received travel awards from Jennerex Inc. M.K.K., Y.K.L., S.-G.K., and T.-H.H. are employees of SillaJen Inc. Data and materials availability: The materials used in this study are available from Jennerex Inc. under a material transfer agreement. A patent application based on this work has been filed.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article