Background
Lung cancer still represents a very fatal disease causing the majority of the cancer-related deaths in males worldwide [
1]. Non-small cell lung cancer (NSCLC), including adenocarcinoma, squamous cell carcinoma, and large cell carcinoma, constitutes 80-85% of all lung cancers, whereas small cell lung cancer accounts for 15% to 20%, with most patients having advanced inoperable disease at the time of diagnosis. The current treatment options are surgical resection, platinum-based doublet chemotherapy, and radiation therapy alone or in combination [
2]. However, the prognosis for lung cancer patients still remains poor with an overall 5-year survival rate of only 15% [
3]. Therefore, new, effective, therapeutic approaches for lung cancer are mandatory. One novel strategy represents the targeted therapy. Therein drugs are applied that specifically target genetic mutations and signalling pathways altered in lung cancer [
4]. Up to now four targeted therapies have been FDA-approved for the treatment of lung cancer. Morevoer, the use of oncolytic viruses is a very promising therapeutic approach for the treatment of cancer. These viruses, either naturally occurring or genetically engineered, are replication-competent viruses that are able to selectively infect and destroy cancer cells either after intratumoral or systemic administration [
5]. Among the most intensively investigated oncolytic viruses in preclinical or clinical studies are adenoviruses, herpes simplex virus, Newcastle disease virus, measles virus and Vaccinia virus [
6‐
8]. Vaccinia virus (VACV) is a very promising agent for oncolytic virotherapy, since the use of VACV as a vaccine during the eradication of smallpox clearly demonstrated its safety in human patients. Furthermore, the broad host range, the efficient replication exclusively in the cytoplasm of host cells, the natural tropism for tumor tissues, and its large genome with a huge capacity for the integration of recombinant DNA are further advantages over other oncolytic viruses [
9,
10]. Until now, various oncolytic VACV have been investigated in preclinical and clinical studies [
11]. Recently, it was shown that an oncolytic VACV, GLV-1h68, which was constructed by inserting gene expression cassettes for a Renilla luciferase-
Aequora green fluorescent fusion protein (RUC-GFP), β-galactosidase and β-glucuronidase into the
F14.5 L,
J2R, and
A56R loci, respectively, possesses reduced toxicity and enhanced tumor targeting specificity compared with its parental LIVP strain [
12]. We have demonstrated that treatment with GLV-1h68 or its derivatives led to inhibition of tumor growth in several different xenograft models, including human breast cancer [
12‐
14], anaplastic thyroid carcinoma [
15], malignant pleural mesothelioma xenografts [
16], squamous cell carcinoma [
17], pancreatic carcinoma [
18‐
21], prostate carcinoma [
22‐
24], lung carcinoma [
24], fibrosarcoma [
25], hepatocellular carcinoma [
26] as well as canine mammary adenoma [
27] and carcinoma [
28]. In addition, two clinical trials with GL-ONC1 (clinical grade GLV-1h68) have been successfully started (
http://www.clinicaltrials.gov; references NCT00794131 and NCT01443260).
Due to limitiations of oncolytic tumor therapy observed in several clinical studies, various approaches to enhance the efficiency of oncolytic viruses e.g. by combination with different cancer treatment modalities such as chemotherapy or radiotherapy are currently intensively investigated [
29]. Improved results have been observed for GLV-1h68 using combinations with cisplatin or gemcitabine in pancreatic [
18], mitomycin C in prostate [
23] or the β-galactosidase-activatable prodrug seco-analog of duocarmycin SA in breast tumor xenografts [
14]. Moreover, systemically delivered GLV-1h68 in combination with focal ionizing radiation (IR) resulted in improved tumor growth inhibition and mouse survival in a glioma tumor model [
30].
Cyclophosphamide (CPA) is an alkylating agent that is known to cause crosslinking of DNA and is used for treatment of various tumors. CPA
per se is an inactive prodrug that requires metabolic activation in the liver to become the active compound 4-hydroxycyclophosphamide. CPA has already been applied in combination with oncolytic viruses in various tumor models and its synergistic effects have been observed with Herpes simplex virus in glioma models [
31‐
34], sarcoma [
35] or in lung adenocarcinoma [
36], with adenovirus in a hamster model of renal adenocarcinoma [
37] and in cancer patients [
38], with Reovirus in mouse melanoma [
39,
40], and with VACV in a glioma xenograft model [
41].
In this study, we set out to evaluate the oncolytic activity of GLV-1h68 in the human lung adenocarcinoma cell line PC14PE6 in cell culture, as well as to determine the antitumor potency of GLV-1h68 as monotherapy or in combination with CPA in a mouse model of PC14PE6-RFP lung adenocarcinoma. Our results demonstrate that GLV-1h68 is able to replicate in and kill human PC14PE6-RFP cells in cell culture. Furthermore, GLV-1h68 efficiently colonizes and notably delays the growth of PC14PE6-RFP tumors in a xenograft mouse model. Moreover, combination therapy with CPA and GLV-1h68 significantly improves the antitumoral efficacy of systemically injected GLV-1h68. Higher levels of pro-inflammatory cytokines and chemokines are seen in tumors of GLV-1h68-treated mice, while after CPA or combination treatment factors either expressed by endothelial cells or present in the blood are found to be reduced. Moreover, combination treatment led to a loss of the hemorrhagic phenotype of PC14PE6-RFP tumors. Our results strongly suggest that the enhanced tumor control achieved by combining GLV-1h68 with CPA is due to an action of CPA on the tumor vasculature.
Methods
Cell lines and virus strain
Stably dsRed2-expressing human PC14PE6 cells were engineered and kindly provided by the group of F. Winkler (University of Heidelberg, Neurooncology, Heidelberg, Germany) in 2008 [
42]. This PC14PE6-RFP cells were authenticated by the Leibniz-Institut DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany) to be identical with the parental cell line PC14 (Riken, Japan) in 2012. PC14PE6-RFP cells were maintained in DMEM (PAA Laboratories, Cölbe, Germany) supplemented with 10% fetal bovine serum (FBS), 2 mM GlutaMAX (both from Invitrogen GmbH, Karlsruhe, Germany), 1× non-essential amino acids, 1× penicillin/streptomycin (both from PAA Laboratories) at 37°C under 5% CO
2. African green monkey kidney fibroblasts (CV-1; ATCC number CCL-70) were cultured in growth medium consisting of DMEM with 10% FBS and 1x penicillin/streptomycin at 37°C under 5% CO
2. The attenuated Vaccinia virus strain GLV-1h68 was derived from LIVP (Lister strain from the Institute of Viral preparations, Moscow, Russia), as described previously [
12].
Viral replication assay
For the viral replication assay, PC14PE6-RFP cells grown in 24-well plates were infected with GLV-1h68 at MOI 0.1 or 1.0 in infection medium (DMEM containing 2% FBS and supplements). After one hour of incubation at 37°C with gentle agitation every 20 min, virus-containing supernatants were removed and replaced by fresh growth medium. After 24, 48 or 72 h, cells and supernatants were harvested. Following three freeze-thaw cycles, serial dilutions of the lysates were titered by standard plaque assays on CV-1 cells. All samples were measured in triplicate.
Cell viability assay
To determine viral cytotoxicity, GLV-1h68 infected and uninfected PC14PE6-RFP cells in 24-well plates were analyzed using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Cells were infected with GLV-1h68 at MOI 0.1 or 1.0 or mock-infected with infection medium (DMEM containing 2% FBS and supplements). After one hour of incubation, virus-containing supernatants were removed by aspiration and fresh medium was added. After 24, 48 and 72 h, respectively, media were removed and 500 μl MTT (2.5 mg/ml, Sigma-Aldrich, Germany) solution in DMEM without phenol red (PAA Laboratories, Cölbe, Germany) was added for 2 h at 37°C and 5% CO2. MTT solution was then removed and 400 μl of 1 N HCl in isopropanol was added. Each sample (3×100 μl) was transferred to a 96-well plate and absorbance was measured at 570 nm with a reference wavelength of 650 nm in a Sunrise Microplate reader (Tecan, Austria). The percentage of cell survival was calculated using the following formula: % cell survival = (absorbance value of infected cells/absorbance value of uninfected control cells) × 100%.
Human tumor xenografts and virus injections
All animal experiments were performed in accordance with protocols approved by the government of Unterfranken (Wuerzburg, Germany, protocol number AZ 55.2-2531.01-17/08) or by the Institutional Animal Care and Use Committee (IACUC) of Explora Biolabs (San Diego, USA, protocol number EB11-025). Human PC14PE6-RFP cells (4×105 cells in 100 μl PBS) were implanted subcutaneously into the right flank of six-week-old female nude mice (Hsd/Athymic Nude-Foxn1nu, Harlan Laboratories, Netherlands and Indianapolis). Tumor growth was monitored twice a week using a digital caliper. Tumor volume was calculated with the following formula: [(length × width2) × 0.52]. When tumors reached 100–200 mm3 mice were either injected via the tail vein with 1×107 pfu GLV-1h68 (in 100 μl PBS) or PBS as control and/or combination treatment was started (day 0). For cyclophosphamide (CPA) combination treatment groups of infected or uninfected mice were injected intraperitoneally with 140 mg/kg bodyweight CPA (Sigma-Aldrich, Germany) at day 0, and with 100 mg/kg bodyweight CPA at days 1, 3, 7, 10, 15, 18 and 21.
Virus titration from tumors and organs
To assess viral distribution, PC14PE6-RFP tumor-bearing mice were either infected via tail vein with 1×107 pfu of GLV-1h68 (in 100 μl PBS) or subjected to combination treatment. At indicated time points, mice were sacrificed and tumors and organs were prepared and weighted. Tumors were homogenized in an gentleMACS Dissociator using M-Tubes (both Miltenyi Biotech GmbH, Bergisch Gladbach, Germany) and organs in Precellys tubes (Peqlab, Erlangen, Germany) in a FastPrep™ FP120 (Thermo Electron Corporation, Langenselbold, Germany). After three freeze-thaw cycles, viral titers in homogenates were determined by standard plaque assays on CV-1 cells.
Fluorescence live-animal imaging
Tumor cell growth and viral infection were monitored directly by optical imaging based on dsRed-expression by tumor cells and GFP-expression by virus infected cells respectively and quantified using a Maestro EX imaging system (CRI, Woburn, MA). Mice were anesthetized by isofluran inhalation (induction 4%, maintenance 1%). Images were taken at days 14 and 21 with a Maestro EX imaging system (CRi, Woburn, MA) using appropriate filters for dsRed (tumor, excitation: 503–555 nm, emission: 580 nm cut-in) and GFP (virus, excitation: 445–490 nm, emission: 515 nm cut-in). Images were evaluated and quantified using the Maestro Version 2.10.0 software.
Rodent multi-analyte profile
For preparation of tumor lysates, three mice of each group were sacrificed 7 dpi. Tumors were removed, resuspended in 9 volumes (W/V) lysis buffer (50 mM Tris–HCl (pH 7.4); 2 mM EDTA (pH 7.4), 2 mM PMSF and Complete Mini protease inhibitors (Roche, Mannheim, Germany) and lysed in an gentleMACS Dissociator using M-Tubes (both Miltenyi Biotech GmbH, Bergisch Gladbach, Germany). Samples were then centrifuged at 500 g at 4°C for 5 min and supernatants were submitted to Rules-Based Medicine (Myriad RBM, Austin, USA) for bead-based immunodetection of mouse immune-related protein antigens (RodentMAP® v2.0).
Immunohistochemistry
For immune-histochemistry, tumors were excised and snap-frozen in liquid N2, followed by fixation in 4% paraformaldehyde/PBS pH 7.4 for 16 h at 4°C. Fixed tumors were dehydrated in 10% Sucrose/PBS for 3–4 h followed by 30% sucrose/PBS for 12 h and finally embedded in Tissue-Tek® O.C.T. (Sakura Finetek Europe B.V., Alphen aan den Rijn, Netherlands). Tumor samples were sectioned (15 μm) with the cryostat 2800 Frigocut (Leica Microsystems GmbH, Wetzlar, Germany) and stored at -80°C. Antibody-labeling was performed following fixation in ice-cold aceton. Endothelial cells were labelled with the hamster anti-mouse CD31 antibody (Chemicon, International, Temecula, CA) and Cy3-conjugated secondary donkey anti-hamster antibody obtained from Jackson ImmunoResearch (West Grove, PA). The primary antibody was incubated for 1 h. After washing with PBS, sections were labeled for 30 min with the secondary antibody and finally mounted in Mowiol 4–88.
Fluorescence microscopy
The fluorescence-labelled preparations were examined using the Leica TCS SP2 AOBS confocal laser microscope equipped with an argon, helium-neon and UV laser and the LCS 2.16 software (1024 × 1024 pixel RGB-color images). Digital images were processed with Photoshop 7.0 (Adobe Systems, Mountain View, CA) and merged to yield overlay images.
Measurements of microvessel density
The vascular density was determined in microscopic images (×20 objective, ×10 ocular) of CD31-labelled tumor sections. On the confocal microscope, the CD31 fluorescence was set to a clearly detectable level by adjusting the photo-multiplier before the images were captured. All images were decorated with five horizontal lines at identical positions using Photoshop 7.0 and all vessels which intersected these lines were counted to yield the vascular density. The vascular density was calculated in duplicate for ten images (five images of two different control, GLV-1h68-infected or combination-treated tumors) and presented as mean values with standard deviations.
Statistics
A two-tailed Student’s t test was used for statistical analysis. P values of < 0.05 were considered statistically significant.
Discussion
The efficiency of the oncolytic VACV GLV-1h68 in treating different human or canine cancers in mice has been demonstrated in several preclinical studies [
12‐
28]. Moreover, currently two clinical trials with GLV-1h68 as intervention are carried out (
http://www.clinicaltrials.gov). One great advantage of oncolytic viruses for the treatment of cancers comprises their capability to specifically kill cancerous cells whereas normal cells are not affected. In the current study, we investigated the effect of a combined treatment strategy consisting of oncolytic VACV GLV-1h68 and the alkylating agent CPA on the growth of human lung adenocarcinoma PC14PE6 xenografts. Synergistic effects of diverse oncolytic viruses and CPA have been described in the literature [
31‐
41]. Suggested mechanisms are, if specified, mostly due to the immuno-modulatory properties of CPA. CPA is known to suppress the host innate immunity and inhibition of intratumoral infiltration of mononuclear cells has been described, thus probably extending the survival of the virus within the tumors [
33,
36,
37]. When used in a low- dose metronomic (LDM) schedule, CPA can also, exert antiangiogenic effects [
43‐
45].This seems to be based on the induction of thrombospondin-1 which in turn probably acts antiangiogenic through binding to CD36 [
46] or through binding and sequestering VEGF [
47].
In our study, we could show that combination therapy with GLV-1h68 and CPA led to a significant growth inhibition of PC14PE6-RFP xenografts compared to GLV-1h68 alone from day 21 post injection on (Figure
3B). Analysis of the viral load of combination-treated or GLV-1h68-treated tumors revealed that the better therapeutic effect was not due to an enhanced viral titer in the tumors. Moreover, after combination treatment we did not see enhanced viral titers in the organs of mice suggesting that CPA has no immunosuppressive effect in our experimental model. Immune-related effects of the investigated therapy were not explored in detail in the present study. However, preliminary data obtained by histological analysis of CPA- and non-CPA-treated tumors 7 dpi revealed no significant difference in the content of CD68-positive immune cells, e.g. monocytes/macrophages (Additional file
1). Furthermore depletion of phagocytic cells especially monocytes and macrophages by clodronate liposomes did result in lesser tumor growth retardation than by treatment with CPA alone (Additional file
2). Both observations are again hinting that CPA has no effect on cells of monocytic origin in the PC14PE6-RFP model. By protein profiling of lysates of untreated, CPA-, GLV-1h68- and combination treated tumors, we found the expression of several pro-inflammatory cytokines and chemokines of host origin to be up-regulated in tumor tissues after virus infection and even after combination treatment, although to a lesser extent (Table
3). The majority of these cytokines/chemokines and pro-inflammatory proteins, like eotaxin, MIP-1 beta, MCP-1, MCP-3, MCP-5, TNF-alpha, and MPO are known to be produced by activated macrophages in the inflamed tissue. Similar results also have been obtained in xenograft models of human pancreatic [
18], prostate [
22] and hepatocellular [
26] as well as canine mammary carcinomas [
28] suggesting that it is a general mechanism of GLV-1h68 to induce strong innate host immune responses in tumors of different origins. It is also very interesting to note that after virus infection or combination treatment the CCR5 ligands MIP-1 beta/MIP-3 beta or RANTES, respectively, are overexpressed. The CCR5 ligand chemokine pathway is, among other molecular pathways, known to be consistently activated during immune-mediated cancer rejection as well as other immune-mediated tissue destruction processes, as described in the “Immunologic Constant of Rejection (ICR)” [
48]. The cytokine Oncostatin-M, which is up-regulated after VACV infection, belongs to the IL-6 family and has cytostatic activities on a number of tumor cell lines [
49‐
51]. The only protein that was found to be down-regulated upon viral infection was EGF. Interestingly, mutations of the EGF receptor (EGFR) have been associated with several types of cancers, including lung cancer, and targeted therapy using small molecules disrupting kinase activity of EGFR results in considerable clinical benefit in lung cancer patients [
4]. Moreover, Goswami
et al.[
52] described a paracrine loop between breast carcinoma cells (producing CSF) and macrophages (producing EGF) leading to increased carcinoma cell invasion and they suggested that disabling this loop might result in clinical benefit in the treatment of cancer. So, it is remarkable that by virotherapy with Vaccinia virus endogenous EGF levels are reduced possibly resulting in inhibition of downstream pro-survival signaling pathways or inhibition of tumor cell invasion.
Combination treatment of GLV-1h68 and CPA leads to significant up-regulation of M-CSF-1, MCP-1 and MCP-5, which have pro-inflammatory activities (Table
3B) and again to down-regulation of EGF compared to untreated controls. Concentrations of Apo A-I, VCAM-1, Fibrinogen and vWF are found to be decreased after combination treatment. Apo A-I has anti-inflammatory [
53,
54] but also anti-thrombotic [
55] activities. VCAM-1 is expressed by the cytokine-activated endothelium and mediates leukocyte-endothelial cell adhesion. It is also known that in several types of cancers VCAM-1 is aberrantly expressed on the surface of tumor cells thereby tethering macrophages to tumor cells and generating favorable conditions for tumor angiogenesis, invasion and metastasis [
56,
57]. Recently, a study reported about an EGF-enhanced VCAM-1 expression promoting macrophage and glioblastoma cell interaction and tumor cell invasion [
57]. In another work a VCAM-1-mediated tumor immune evasion has been described [
58,
59]. The authors found that overexpression of VCAM-1 by cancer cells led to decreased apoptosis of tumor cells and a significant decrease in the number of tumor-infiltrating CD8+ T cells expressing VCAM-1. Therefore, reduction of VCAM-1 and EGF by combination treatment might have favorable effects in terms of inhibition of tumor cell invasion and tumor immune evasion. Fibrinogen plays a pivotal role in the coagulation cascade. vWF is synthesized by endothelial cells and megakaryocytes and promotes the adhesion of platelets to sites of vascular injury. Interestingly, the latter four factors, Apo A-I, VCAM-1, Fibrinogen and vWF, which are found to be decreased after combination treatment, are either expressed by endothelial cells or are present in the blood, hinting that combination therapy has an effect on the tumor vasculature in the PC14PE6-RFP model. Compared to CPA treatment alone, infection with GLV-1h68 or combination treatment leads to an up-regulation of several pro-inflammatory cytokines and chemokines (Table
3C and D), like seen in the virus-treated versus untreated tumors, affirming that virus infection leads to a pro-inflammatory status within the tumor. Moreover, compared to monotherapy with CPA, lower levels of VCAM-1 are observed upon combination treatment. Interestingly, after CPA or combination therapy the only protein that was found to be significanlty decreased compared to respective controls was vWF (Table
3E and F). This indicates that the primary effect of CPA in this model is caused by an effect on the vasculature and not by alteration of the immune response. However, it is speculating to assume if the lower levels of these factors are due to a down-regulation on the level of protein expression or are a result of the lower density of blood vessels found in tumors after CPA or combination treatment.
Taken together, these data show that pro-inflammatory cytokines and chemokines are elevated in PC14PE6-RFP tumors of GLV-1h68-treated mice (Table
3A, B, C, and D). But more importantly, levels of factors either expressed by endothelial cells or present in the blood are found to be decreased after CPA or combination treatment (Table
3B, D, E, and F). We therefore suppose that in the PC14PE6-RFP model the enhanced tumor growth inhibition observed after combination therapy is due to an effect on the vasculature rather than an immunosuppressive action of CPA. However, it is conceivable that the modulation of the vasculature may alter the following immunological response to VACV e.g. by hampering immune recruitment to the tumor site.
Especially untreated PC14PE6-RFP tumors, but also GLV-1h68 or CPA-treated tumors have a hemorrhagic phenotype (Figure
5A). This seems to be a hallmark of the PC14PE6 cell line since Yano
et al. already described the development of bloody pleural effusions after
i.v. injection of PC14PE6 cells [
60]. Only recently, Weibel
et al. showed that after subcutaneous implantation of PC14PE6 cells malignant effusion concomitant hematoma formation occurs at the tumor site, and, during oncolytic virotherapy using GLV-1h68 and especially its derivative GLV-1h108, which encodes a single chain antibody against VEGF, tumor-associated hematoma disappeared [
61]. Remarkably, in the present study we could show that after combination treatment PC14PE6-RFP tumors lose their hemorrhagic phenotype. Since at the time point of treatment start, tumors of the combination group already showed a slight blue color, CPA treatment seems not only to prevent the development of the hematoma but to actively rebuild a non-hemorrhagic appearance when used in combination with oncolytic Vaccinia virus GLV-1h68. This additionally indicates that in the PC14PE6-RFP xenograft model combination therapy with GLV-1h68 and CPA acts on the tumor vasculature. To further support this hypothesis, we analyzed the vascular density in CD31-labelled tumor cross-sections 7 dpi. Indeed, there was a decrease in the number of blood vessels after CPA and combination treatment compared to GLV-1h68-treated tumors (Figure
5B), albeit this was not significant. The presence of CD31-positive blood vessels after CPA treatment, further suggest that CPA seems to have no immunosuppressive effect, since CD31 is involved in leukocyte trafficking to sites of inflammation. Previous work in our laboratory demonstrated that GLV-1h68 does not destroy endothelial cells in tumors and that the tumor vasculature in infected tumors is still functional [
62]. In the matter of the significance of tumor vasculature in tumor progression, therapeutic approaches additionally targeting the tumor endothelium may contribute to a better therapeutic outcome. Therefore, combination strategies composed of an agent, which directly kills tumor cells on the one hand (in our case GLV-1h68), and an additional therapeutics, which targets normal cells within the tumors, e.g. tumor vasculature, are particularly advantageous. In this study, such a combination therapy consisting of oncolytic VACV GLV-1h68, destroying tumor cells, and CPA, presumably targeting tumor vasculature, could successfully be established for the PC14PE6-RFP tumor model.
Competing interests
EH and SW are recipients of postdoctoral fellowships of the University of Würzburg from a research service grant provided by Genelux Corporation. AAS is salaried employee of Genelux Corporation and has personal financial interests in Genelux Corporation. The funders had no role in study design, data collection and analysis or decision to publish. This work was supported by a research grant and a service grant from Genelux Corporation (R&D facility in San Diego, CA, USA).
Authors’ contributions
EH and SW designed the study, performed the experiments, statistical analyses and interpretation of data and wrote the manuscript. AAS participated in conceiving the study and writing the manuscript. All authors read and approved the final manuscript.