Therapeutic Antitumor Potential of Endoglin-Based DNA Vaccine Combined with Immunomodulatory Agents

ORIGINAL ARTICLE Therapeutic antitumor potential of endoglin-based DNA vaccine combined with immunomodulatory agents

M Jarosz, J Jazowiecka-Rakus, T Cichon´ , M Głowala-Kosin´ ska, R Smolarczyk, A Smagur, S Malina, A Sochanik and S Szala

Therapy targeting tumor blood vessels ought to inhibit tumor growth. However, tumors become refractory to antiangiogenic drugs. Therefore, therapeutic solutions should be sought to address cellular resistance to antiangiogenic therapy. In this regard, reversal of the proangiogenic and immunosuppressive phenotype of cancer cells, and the shift of the tumor microenvironment towards more antiangiogenic and immune-stimulating phenotype may hold some promise. In our study, we sought to validate the effects of a combination therapy aimed at reducing tumor blood vessels, coupled with the abrogation of the immunosuppressive state. To achieve this, we developed an oral DNA vaccine against endoglin. This antigen was carried by an attenuated Salmonella Typhimurium and applied before or after tumor cell inoculation into immunocompetent mice. Our results show that this DNA vaccine effectively inhibited tumor growth, in both the prophylactic and therapeutic settings. It also activated both specific and nonspecific immune responses in immunized mice. Activated cytotoxic T-lymphocytes were directed specifically against endothelial and tumor cells overexpressing endoglin. The DNA vaccine inhibited angiogenesis but did not affect wound healing. In combination with interleukin-12-mediated gene therapy, or with cyclophosphamide administration, the DNA vaccine resulted in reduced microvessel density and lowered the level of Treg lymphocytes in the experimental tumors. This effectively inhibited tumor growth and prolonged survival of the treated animals. Polarization of tumor milieu, from proangiogenic and immunosuppressive, towards an immunostimulatory and antiangiogenic profile represents a promising avenue in anticancer therapy.

Gene Therapy (2013) 20, 262--273; doi:10.1038/gt.2012.28; published online 12 April 2012 Keywords: dual targeting; endoglin-based DNA vaccine; immunomodulatory agents; combined antitumor therapy

INTRODUCTION It is this type of hypoxia that is mainly responsible for invasiveness 3,12 -- 14 Despite recent progress in diagnostics, cancer remains a major of cancer cells and formation of metastases. medical challenge. There are 10 million new cancer cases Novel therapeutic solutions must, therefore, take into account worldwide every year. WHO expects the incidence to rise to or omit such resistance. Cellular infiltrates strongly affect tumor 15 million in 2020.1 Therefore, novel prophylactic, diagnostic and progression through formation of the proangiogenic and therapeutic anticancer approaches are urgently needed. An immunosuppressive microenvironment that enables cancer cells 15 important aspect in progression of multiple tumor types is the to escape from immune surveillance, while certain proangio- interplay between genetic alterations within cancer cells and their genic factors (for example, vascular endothelial growth factor surrounding microenvironment.2 To sustain continued growth (VEGF), transforming growth factor-b (TGF-b)) may act as 15 -- 18 beyond 1--2 mm3, tumor mass becomes dependent on genera- immunosuppressants. Thus, it is the complex microenviron- tion of a new vascular network, with the structural support in the mental polarization toward proangiogenic and immunosuppres- form of extracellular matrix, and contribution of several growth sive phenotype (tumor growth-stimulation) that may present factors released by the activated stromal cells.1,3 It was originally the main therapeutic obstacle, and consequently it is of interest to 15,19 believed that an effective therapy targeting tumor blood vessels explore the means of its global reversal. should arrest tumor growth, restrain metastasis and exert a lasting In this study, we tested the effects of a combined immuno- effect without mutational drug resistance that limits the efficacy therapy targeted in such a way as to simultaneously decrease the of traditional anticancer agents.4 In spite of these predictions, number of tumor blood vessels and alleviate tumor immunosup- long-term use of antiangiogenic drugs often leads to therapeutic pression. We relied on the finding that tumor vascular endothelial 2 escape involving a variety of mechanisms dependent on the cells express specific surface proteins, including endoglin properties of blood vessels, cancer cells and regulatory milieu (CD105), which also acts as the co-receptor for TGF-b, known to of bone marrow derived cellular infiltrates.3,5,6 Use of single possess immunomodulating activity. Immune response directed antiangiogenic agents is thus not sufficient.3,7 Moreover, inhibiting against such self antigens overexpressed in the tumor vasculature 20 endothelial cells proliferation may not be effective against tumor has already been contemplated. Endoglin is a unique marker of vasculature formed through such processes as intussusception8 activated vascular endothelial cells and its presence in solid tumor 21 -- 23 or vasculogenic mimicry.9--11 In addition, antiangiogenic therapy blood vasculature is, in some cases, of prognostic value. leads to the so-called post-drug hypoxia within tumor mass. CD105 is also expressed by certain cancer cells (melanoma, Ewing

Center for Translational Research and Molecular Biology of Cancer, Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, Gliwice, Poland. Correspondence: Professor S Szala, Center for Translational Research and Molecular Biology of Cancer, Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, Wybrzez?e Armii Krajowej 15, Gliwice 44-101, Poland. E-mail: [email protected] Received 13 June 2011; revised 23 February 2012; accepted 23 February 2012; published online 12 April 2012 DNA vaccine in combined cancer therapy M Jarosz et al 263 sarcoma and prostate cancer).24,25 As a carrier of the endoglin- restores proliferation of peripheral T lymphocytes and activates encoding gene, we used an attenuated strain of Salmonella nonspecific responses.32,37,38 Moreover, CTX possesses a direct Typhimurium SL7207.26,27 In our study, we used experimental antiangiogenic activity39 achieved, at least in part, by eliminating murine tumor models (B16-F10 melanoma and Renca renal circulating endothelial progenitor cells.32 carcinoma), in which neoplastic cells express endoglin on their We report that anti-endoglin DNA vaccine exerts indeed an surface. The DNA vaccine was therefore directed also against antitumor and antiangiogenic effect in a model of aggressive cancer cells. The antiangiogenic and antitumor effects were murine melanoma. We did not observe any obvious side effects achieved through an oral DNA vaccine carried by Salmonella and when this treatment was combined with IL-12 or CTX durable Typhimurium to secondary lymphoid organs such as Peyer’s antitumor responses were obtained along with repolarization patches (PPs). Immunotherapy with this vaccine breaks immune of the tumor stroma toward an antiangiogenic and more tolerance against self antigens and activates cytotoxic CD8 þ immunocompetent phenotype. In ca. 30% of mice treated with T lymphocytes to destroy tumor vascular endothelial cells, leading our DNA vaccine and IL-12-mediated gene therapy total tumor to inhibition of tumor growth.28 regression was seen. On the other hand, immunotherapy While endoglin-directed antiangiogenic DNA vaccines have combined with chemotherapy (CTX) led to a long-term stabiliza- already been explored to some extent,29,30 we focused on tion of tumor growth. We suggest that targeted immunomodula- antitumor aspect of this approach and reasoned that its effect tion may be the key to successful anticancer treatment with could be markedly enhanced by combination with carefully antiangiogenic DNA vaccines. selected immunomodulating agents. For instance, the immune response in such vaccine-based treatment combinations could be reinforced through inclusion of interleukin-12 (IL-12)31 or cyclophosphamide (CTX).32 IL-12 is a pleiotropic immunomodulatory RESULTS cytokine inducing both nonspecific (NK, NK-T) as well as specific Salmonella Typhimurium transfers the therapeutic gene to PPs (CD4, CD8) immune responses, in addition to the appreciable The construct pcDNA3.1( þ )/mCD105 (Figure 1a) was transfected nonimmune antiangiogenic activity.33 -- 36 Additionally, IL-12 into COS7 cells and the presence of the corresponding protein in eliminates Treg lymphocytes from tumor microenvironment, eukaryotic cells was confirmed by western blotting (Figure 1b). effectively abrogating tumor immunosuppression.36 Similarly, the Next, pcDNA3.1( þ )/mCD105 plasmid was transferred into chemotherapy with CTX lowers the level of Treg lymphocytes, Salmonella Typhimurium SL7207 strain using electroporation.

Peyer’s Patches

KpnI ENDOGLIN XbaI

Control SL7207/EGFP

Figure 1. Vector construction (a), expression of endoglin gene in vitro (b), gene transfer to PPs (c) and endoglin expression by cell lines (d). (a) DNA fragment encoding murine endoglin (mCD105) was obtained by PCR. The reaction product was then inserted into pcDNA3.1( þ ) plasmid using XbaI and KpnI restriction sites. (b) Protein expression of mCD105 was detected by transfecting COS7 cells with the plasmid and performing western blotting with rat anti-murine CD105 antibody. (c) Plasmid-carrying S. Typhimurium bacteria (1 Â 108 per animal) were orally administered to C57Bl/6 mice. After 24 h, the animals were killed, PPs collected and frozen sections obtained. EGFP protein-containing cells were identified using fluorescence microscopy (magnification Â 20). Mouse PPs were stained with hematoxylin/eosin. (d) Expression levels of endoglin (CD105) was examined in B16-F10, Renca and C26 tumor cell lines. As a control, HECa10 endothelial cells were used. Cells were stained with rat PE-conjugated monoclonal antibody recognizing murine endoglin.

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 262 -- 273 DNA vaccine in combined cancer therapy M Jarosz et al 264 Positive SL7207/mCD105 clones were used as a DNA vaccine of immunized mice. Elevation of NK cells (CD49b) was also in prophylactic and therapeutic settings. In our initial experi- detected in these tumors as a function of the immunization ments, we confirmed that the orally administered Salmonella with the DNA vaccine. Collectively, more than twofold increase Typhimurium SL7207 strain, used by us as a delivery vehicle, in the number of CD4 þ T lymphocytes (Figure 5a) and NK cells effectively reached secondary lymphoid organs such as PPs in the (Figure 5c), as well as more than fivefold increase in the level of small intestine. CD8 þ T lymphocytes (Figure 5b) were found in tumors from In order to demonstrate the transfer of a eukaryotic gene to PPs, immunized mice, as compared with control animals. we used a reference gene (EGFP (enhanced green fluorescent protein)). In accordance with our assumptions, in vitro studies DNA vaccine induces endoglin-specific cytotoxicity confirmed first Salmonella Typhimurium SL7207-mediated transfer In order to verify specificity of the immune response induced by the of the eukaryotic gene (EGFP) into murine macrophages (data not studied endoglin-directed DNA vaccine, we examined cytotoxicity of shown). Next, in vivo studies (EGFP fluorescence) confirmed CTLs (cytotoxic T-lymphocytes) activated by this vaccine. To this that the bacteria reached PPs and transferred the gene, which purpose, we used B16-F10 melanoma and HECa10 endothelial cells was subsequently expressed by the host cells (Figure 1c). (both overexpressing endoglin), as well as C26 colorectal carcinoma cells devoid of endoglin (Figure 1d). Cells expressing endoglin Immunization of mice with DNA vaccine inhibits tumor growth in (Figures 6a and b) were much more sensitive to CTLs than controls both prophylactic and therapeutic approach (Figure 6c). DNA vaccine-induced CTLs were more active towards Antitumor properties of the oral endoglin-based DNA vaccine were cells expressing endoglin, as compared with CTLs induced by empty investigated using two experimental murine tumor models (B16-F10 plasmid-carrying bacteria. There were no differences in terms of murine melanoma and Renca renal carcinoma) in which endoglin is cytotoxicity of CTLs against C26 cells devoid of endoglin (that is, overexpressed by cancer cells (Figure 1d). First, in a prophylactic between CTLs induced by DNA vaccine and by control vector). setting, DNA vaccine was administered orally three times before These data show an effective induction of the immune response inoculating mice with either type of cancer cells (B16-F10 or Renca) directed against cells overexpressing endoglin. The response was (Materials and Methods for details). A significant inhibition of tumor directed against both endothelial and cancer cells. growthwasobservedincaseofbothB16-F10(Figure2a)andRenca (Figure 2c) tumors, as compared with control (bacteria carrying Endoglin-based DNA vaccine inhibits angiogenesis in vivo plasmid without insert). Cytotoxicity tests have corroborated the finding that the induced Induction of immune response was tested in a therapeutic immune response is directed against endothelial cells carrying setting (Materials and Methods for details). In brief, mice endoglin on their surface. This glycoprotein is a marker of active pre-inoculated with B16-F10 melanoma cells were administered proliferating endothelial cells participating in the formation of the oral DNA vaccine and tumor growth inhibition was observed new blood vessels. We thus wanted to investigate whether the in the treated mice (Figure 2d). examined DNA vaccine affects inhibition of angiogenesis. To this purpose, we used a Matrigel test and fluorescein isothiocyanate Oral DNA vaccine inhibits formation of experimental metastases (FITC)-labeled Isolectin B4 for staining of endothelium. Macro- Immunization of mice with SL7207/mCD105 bacteria (DNA scopically, sizeable differences were visible in vascularization of vaccine) led to a significant inhibition in the formation Matrigel plugs. Endoglin-directed DNA vaccine inhibits neovascu- of experimental B16-F10 melanoma metastases. Their number in larization in immunized mice, as compared with mice receiving the lungs of immunized mice decreased fourfold compared empty plasmid-carrying bacteria. The level of fluorescein in with mice receiving bacteria with empty plasmid (Figure 2c). supernatants from Matrigel lysates was more than twofold lower This suggests that the DNA vaccine indeed acts against both for mice immunized with endoglin-based DNA vaccine, as endothelial and B16-F10 melanoma cells. compared with Matrigel lysates from controls (Figure 7).

Endoglin-based DNA vaccine activates T lymphocytes Combination of DNA vaccine and IL-12 gene therapy effectively and dendritic cells inhibits tumor growth, decreases the number of blood vessels and the level of tumor Treg lymphocytes The notable antitumor efficacy of the endoglin-directed vaccine As mentioned earlier, we reasoned that the therapeutic effect of suggests the existence of corresponding immune responses. the oral endoglin-based DNA vaccine could be enhanced upon In fact, we observed significantly elevated levels of activated combination with intratumoral administration of antiangiogenic immune cells in immunized mice, as compared with mice immunomodulators, such as IL-12.31 This was found to be the administered bacteria carrying the control plasmid. The number of case, as the intratumoral delivery of the IL-12-encoding plasmid cells bearing T lymphocyte markers (CD28, CD25 and CD69; resulted in a marked extension of antitumor effects associated Figure 3a), and dendritic cells with co-stimulating molecules (CD80 with the DNA vaccine (Figures 8a and b). Following complete and CD86; Figure 3b) was markedly elevated in spleens of the tumor regression, B30% of mice treated with the combined immunized mice, as compared with controls. This clearly suggests regimen showed no signs of tumor re-growth. Mice treated with activation of cellular response by the endoglin-based DNA vaccine. either monotherapy also exhibited tumor growth delay, but their The DNA vaccine triggered stimulation of cytokine expression: survival was shorter than that of the mice treated with the IL-2 (Figure 4a), interferon (IFN)-g (Figure 4b) and tumor necrosis combined approach (data not shown). Sections of tumors factor (TNF)-a (Figure 4c) from splenocytes of immunized mice, as obtained from mice treated with combined therapy (DNA compared with splenocytes from control mice. Higher expression vaccine þ IL-12) showed significant numbers of necrotic areas level of proinflammatory cytokines in T lymphocytes is an and contained infiltrates of immune system cells (Figure 8b). additional proof of our DNA vaccine-caused lymphocyte activation Immunohistochemical analyses demonstrated decreased num- in secondary lymphoid tissues. bers of blood vessels in tumor specimens obtained from mice treated with oral DNA vaccine combined with IL-12-mediated DNA vaccine activates both specific and nonspecific immune gene therapy (Figure 8c) as compared with controls. Combination response therapy involving oral DNA vaccine and IL-12 gene therapy, as The DNA vaccine activated specific cellular immune response, as well as monotherapies, markedly decreased the level of tumor Treg indicated by CD4 þ and CD8 þ T lymphocyte infiltrates in tumors lymphocytes in the treated animals, as compared with controls.

Gene Therapy (2013) 262 -- 273 & 2013 Macmillan Publishers Limited DNA vaccine in combined cancer therapy M Jarosz et al 265

Figure 2. Inhibition of tumor growth and experimental B16-F10 metastases by endoglin-based DNA vaccine. (a, b, c) Prophylactic treatment model. C57Bl/6 mice (n ¼ 6, a; n ¼ 7, c) or BALB/c mice (n ¼ 5, b) were immunized and then challenged with B16-F10 cells (s.c., panel a, or i.v. tail injection, panel c) or Renca cells (b) (Materials and Methods). Marked inhibition of tumor growth was observed. Statistically significant as compared with control vector group: *Po0.002 (panel a, on 22 day of therapy); **Po0.04 (panel b, on 36 day of therapy). On day 23 of the experiment, lungs were removed and fixed in Bouin’s solution. Significantly lower number of experimental metastases was observed. Statistically significant as compared with control vector group: *Po0.008. (d) Therapeutic treatment model. Following challenge with B16-F10 cells, the animals (n ¼ 9) were orally administered endoglin-based DNA vaccine (Materials and Methods). Marked inhibition of tumor growth was also observed. Statistical difference when compared with control vector group (on 21 day of therapy) was ***Po0.003.

This clearly underscores the beneﬁt of a combined therapeutic antiangiogenic effect under certain conditions,39 and it also approach in diminishing immunosuppression in tumor micro- abrogates immunosuppression.38 Considerable inhibition of environment (Figure 8d). B16-F10 murine melanoma tumor growth (Figures 9a and b) was observed in the mice receiving combined therapy (DNA vaccine and CTX), as compared with controls (receiving either Oral vaccine therapy combined with CTX effectively inhibits monotherapy or phosphate-buffered saline (PBSÀ)). Combination tumor growth, decreases the number of blood vessels and of immunotherapy and chemotherapy did not affect body mass in the level of tumor Treg lymphocytes any signiﬁcant way (data not shown). As a result of The oral vaccine was also tested in combination with intraper- this combination, a long-term stabilization of tumor growth was itoneal administration of CTX. This chemotherapeutic shows obtained, but never a complete regression, such as in the case of

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 262 -- 273 DNA vaccine in combined cancer therapy M Jarosz et al 266

Figure 3. Activation of T cells (a) and DCs (b) by endoglin-based DNA vaccine. Flow cytometric analysis of splenocytes from immunized mice. C57Bl/6 mice (n ¼ 4) were immunized and then challenged with B16-F10 cells (Materials and Methods). On day 14 (a) or day 20 (b) of the experiment, mice were killed and cells isolated from spleens were analyzed. For T-cell activation, splenocytes were stained with PE-CD28 and FITC-CD8a; PE-Cy7-CD25, PE-CD69 and FITC-CD3e. For DC analyses used FITC-CD11c, PE-CD80 and PE-CD86 antibodies. Substantially higher level of activated T lymphocytes and dendritic cells was found. Statistical differences when compared with control vector group were *P ¼ 0.2, **Po0.035, ***Po0.02 (see a); *P ¼ 0.051, **Po0.025 (see b).

IL-2/ CD4+ IFN-γ/ CD8+ TNF-α/ CD4+ 4 2 16 * ** 3.5 14 *** 1.6 3 12 2.5 1.2 10 2 8 1.5 0.8 1 6 % fluorescence % fluorescence % fluorescence 0.4 0.5 2 0 0 0 VectorDNA vaccine Vector DNA vaccine Vector DNA vaccine Figure 4. Stimulation of cytokine expression by oral DNA vaccine. Flow cytometric analysis of splenocytes from immunized mice. C57Bl/6 mice (n ¼ 3) were immunized and then challenged with B16-F10 cells (Materials and Methods). After 2 weeks, spleens were collected and the level of cytokines expressed by CD4 þ and CD8 þ lymphocytes was examined. Cells were stained with FITC-CD4 or FITC-CD8a and then ﬁxed, permeabilized and stained with PE-IL-2, PE-IFN-g or PE-TNF-a. Higher levels of expressed cytokines were found. Statistical differences when compared with control vector group were *Po0.005 (a); **P ¼ 0.3 (b), ***Po0.05 (c).

immunotherapy coupled with IL-12-mediated gene therapy. DNA vaccine is safe and whether it does affect other physiological Sections of tumors obtained from mice treated with combined forms of angiogenesis, such as wound healing. therapy (DNA vaccine þ CTX) also showed significant numbers The tests showed that oral administration of Salmonella of necrotic areas and contained infiltrates of immune system cells Typhimurium carrying endoglin-encoding gene does not impede (Figure 9b). regeneration of damaged tissue (data not shown). Furthermore, Compared with controls, combination of the oral DNA vaccine the vaccine did not affect body mass of the tested animals nor with CTX-mediated chemotherapy decreased the number their normal behavior (data not shown). of blood vessels, as well as markedly decreased the level of Treg lymphocytes found in tumor specimens obtained from mice treated with the vaccine and CTX (Figure 9c). This underscores the DISCUSSION benefits of combination therapies in abrogating immunosuppres- Solid tumors larger than 1--2 mm3 require formation of their own sion in tumor microenvironment (Figure 9d). blood vascular network in order to sustain continued growth.1,3 Thus, treatment targeting tumor blood vessels should inhibit tumor progression.4 Therapy attempts with antiangiogenic Endoglin-based DNA vaccine does not affect the rate of wound drugs, such as bevacizumab (avastin), sorafenib or sunitinib, have healing not been overly successful5 in the past; protracted use of We confirmed the antitumor potential of the tested oral DNA antiangiogenic drugs elicits therapeutic resistance.3,5,6 vaccine, expressed as its ability to suppress neovascularization and Novel approaches to anticancer treatment should explore the to decrease the number of tumor blood vessels in mice treated issue of drug resistance shown by endothelial (and/or tumor) with vaccine alone or in combination with immunomodulatory cells. One possible therapeutic solution might be tumor immuno- agents (Figures 7 and 9). This led us to examine whether the oral therapy directed against antigens expressed on the surface of

Gene Therapy (2013) 262 -- 273 & 2013 Macmillan Publishers Limited DNA vaccine in combined cancer therapy M Jarosz et al 267

Figure 5. Induction of specific (a, b) and nonspecific (c) immune response by endoglin-based DNA vaccine. Single-cell suspensions from B16-F10 tumors obtained from vaccinated mice were analyzed by flow cytometry. C57Bl/6 mice (n ¼ 3) were immunized and then challenged with B16-F10 cells (Materials and Methods). After 3 weeks, tumor material was collected and the level of T lymphocytes and NK cells was determined. To identify the subpopulations of T lymphocytes the following antibodies were used: PE-Cyt7-CD3e, PE-CD4 and FITC-CD8a. To identify the level of NK cells, an anti-mouse CD49b (pan-NK cells) antibody was used. Gate dividing negative from positive cells was based on isotype antibody control probes. Significantly higher levels of CD4 þ T lymphocytes (a), CD8 þ lymphocytes (b) as well as NK cells (c) were found. Statistical differences when compared with control vector group were *Po0.03 (a); **Po0.025 (b); ***Po0.015 (c).

Target: HECa10 Target: B16-F10 Target: C26 30 30 * 30 25 Vector Vector Vector * 25 25 DNA DNA 20 20 20 DNA vaccine vaccine vaccine 15 15 15 ** 10 10 10 5 5 5 Specific lysis (%) Specific lysis (%) Specific lysis (%) 0 0 0 100 : 1 12.5 : 1 100 : 1 12.5 : 1 100 : 1 12.5 : 1 E : T Ratio E : T Ratio E : T Ratio Figure 6. Induction of endoglin-speciﬁc immune response (CTLs) by oral DNA vaccine. C57Bl/6 mice (n ¼ 3) were immunized and then challenged with B16-F10 cells (Materials and Methods). After 10 days, splenocytes were isolated, re-stimulated (Materials and Methods) and their cytotoxic activity analyzed by XTT test at various effector cell-to-target cell ratios. Murine endothelial cell line (HECa10), endoglin- expressing tumor cell line (B16-F10) and C26 cell line (negative control) were used as targets. Activated CTLs showed cytotoxicity only towards cells expressing CD105 on their surface. Statistical differences when compared with control vector group were *Po0.025 (a); *Po0.0005, **Po0.0015 (b). There were no differences in terms of cytotoxicity of CTLs against C26 devoid of endoglin (c).

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 262 -- 273 DNA vaccine in combined cancer therapy M Jarosz et al 268 blood vasculature with induction of immune response, in order to restore tumor immune surveillance. We exploited the finding that tumor vascular endothelial cells express specific surface proteins, which can be useful as therapeutic targets.2 Endoglin (CD105) is an antigen known to be highly specific for active tumor endothelial cells.21 -- 23 In this study, we employed attenuated Salmonella Typhimurium SL7207 (aroAÀ) bacterial strain26 as a carrier of therapeutic plasmid construct encoding endoglin. Bacteria carrying endoglin-expressing plasmid were orally administered to mice (the latter either pre- or post-inoculated with cancer cells). According to Darji et al.27, even a single dose (in the range of 108 c.f.u. per mouse) elicits a major cytotoxic response. The reaction is strongest during the fifth week post-immunization and then slowly declines. Orally administered Salmonella Typhimurium, used as a delivery vehicle, targets a therapeutic gene to secondary lymphoid organs such as PPs in the small intestine. Attenuated bacteria are phagocytosed by macrophages and dendritic cells.27,43 During subsequent transport of these cells into the spleen, the bacteria perish (mutation aroAÀ inhibits synthesis of aromatic amino acids)26,44 and release plasmid DNA carrying the therapeutic gene which is subsequently expressed. Proteolysis liberates peptides, which are presented to T lymphocytes mainly by MHC I molecules.28 Such oral DNA vaccine has thus the ability to elicit immune response against own antigen and leads to destruction of endothelial cells lining tumor blood vessels and to tumor growth arrest.28 We indeed observed inhibition of tumor growth in mice immunized with the oral endoglin-based DNA vaccine tested. Also, the number of experimental B16-F10 metastases found in Figure 7. Suppression of neovascularization by endoglin-based oral immunized mice was decreased. Immunization led to the DNA vaccine. C57Bl/6 mice (n ¼ 7) were immunized (Materials and Methods). Ten days after last immunization, Matrigel was activation of immune system, seen as increased expression of implanted into mice. Six days later, FITC-labeled Isolectin B4 for important T lymphocyte markers (CD28, CD25 and CD69), and staining of endothelium was used (Materials and Methods). Sig- co-stimulating molecules (CD80 and CD86) on dendritic cells, as nificantly lower fluorescence levels were detected in supernatants well as intracellular cytokines (IL-2, IFN-g and TNF-a)inT from Matrigel lysates. Statistical difference when compared with lymphocytes. In tumors from immunized mice, we detected control vector group was *Po0.0004. increased infiltration of CD4 þ and CD8 þ T lymphocytes and NK cells. These results point clearly to a link between antitumor activity of our oral DNA vaccine and activation of the immune endothelial cells that co-form tumor blood vasculature.28,40 Such system response. The immune response (CTLs) induced by DNA therapy breaks immune tolerance against own antigen, elicits the vaccine was specifically directed against endoglin present on both immune system to recognize endothelial cells as foreign and endothelial and cancer cells. The vaccine markedly inhibited induces immune reaction (mainly cytotoxic CD8 þ T lymphocytes) angiogenesis in immunized mice. Thus, inhibited growth of directed against these cells.27,28,41,42 In consequence, elimination primary tumors and metastases in immunized mice was caused of endothelial cells should lead to arrest of tumor growth, whereas by suppression in tumor-associated angiogenesis and induction of long-term immune memory might prevent tumor relapse.20,28 endoglin-specific CTLs. Immunotherapy directed against tumor vessel endothelial cells Previous attempts to stimulate the immune system against offers several advantages:40 first, these cells are more easily endoglin included a xenogenic protein45,46 and monoclonal anti- accessible to the immune system effectors, compared with bodies.47,48 Our DNA vaccine offers several benefits: it stimulates a neoplastic cells localized at greater distance from blood capillaries; long-term immune response, is specific and safe and can be second, destruction of a single endothelial cell can inhibit division administered orally.20 The likelihood of side effects of this therapy of 4100 neoplastic cells; third, tumor vessel endothelial cells are is rather small, since it is limited to proliferating endothelial cells,2 more efficiently targeted by CD8 þ -mediated immune response or, as in our study, also to cancer cells overexpressing endoglin. compared with neoplastic cells that show lower MHC I (Major Low cost, noninvasive way of administering, ease of preparation Histocompatibility Complex I) expression levels. and storage, as well as efficient transfer of a therapeutic gene to Tumor progression is affected to a substantial degree by various the target site, all contribute to the usefulness of bacteria as cells forming tumor microenvironment since they participate in carriers of therapeutic constructs.49 Salmonella Typhimurium not de novo formation of tumor vasculature and they play a major role only can harbor eukaryotic expression vectors,28,50 but it can also in establishing an immunosuppressive milieu that enables cancer play the role of a natural adjuvant.43 The bacteria trigger the cells’ escape from immune surveillance.15 Certain proangiogenic release of TNF-a, IFN-g, IL-12 and other proinflammatory factors are also immunosuppressive agents (for example, VEGF, mediators. These agents enhance early nonspecific immune TGF-b).15 -- 18 Thus, the issues of cellular resistance and long-term response and the developed inflammatory condition induces inhibition of tumor growth might be addressed with a therapeutic maturation of dendritic cells and antigen presentation.43 This is strategy that brings about polarization of a proangiogenic mediated by CpG bacterial motifs.51,52 In addition, direct delivery and immunosuppressive phenotype towards antiangiogenic and of an antigen to peripheral lymphatic tissue by a bacterial carrier immunostimulating one.15,19 causes immunogenicity of such antigen to be increased 100- to We sought to verify the effectiveness of a combined therapeutic 1000-fold, thereby inducing a strong and biologically active approach that complements targeted destruction of tumor cytotoxic T lymphocyte response.53

Gene Therapy (2013) 262 -- 273 & 2013 Macmillan Publishers Limited DNA vaccine in combined cancer therapy M Jarosz et al 269

PBS

IL-12

DNA vaccine **

DNA * vaccine + IL-12

** *

Figure 8. Inhibition of B16-F10 tumor growth in response to combination therapy involving endoglin-based DNA vaccine and IL-12-mediated gene therapy. (a, b) Inhibition of tumor growth. Following challenge with B16-F10 cells, C57Bl/6 mice (n ¼ 6--7) were orally administered the endoglin-based DNA vaccine. Additionally, on days 9, 11 and 13 after inoculation with cancer cells, the mice received intratumoral injection of plasmid pBCMGSNeo/mIL-12 (Materials and Methods). Tested regimen of combined therapy proved to be much more effective (as compared with monotherapies) in inhibiting tumor growth. Ca. 30% of mice treated with combined therapy underwent total tumor regression. Statistical differences (ANOVA) when compared with control group (on 22 day of therapy) were *Po0.0002, **Po0.015 (a). Photographs were taken on day 23 of the therapy (b). Considerable necrotic areas and infiltrations of immune system cells could be detected in tumor sections from animals treated with either DNA vaccine, or with immunomodulatory factors, or with one of the combined regimens (day 20 of therapy, see b). Lens magnification Â 20. (c) Reduced density of tumor microvessels. On day 20 of the combined therapy, mice (n ¼ 2) were killed and tumors excised for staining with antibody against CD31. The number of vessels for a given experimental group was counted in five visual fields (four tumor sections). A considerably decreased number of blood vessels was found in tumor tissue sections from mice treated with the combination of drugs. Statistical differences (ANOVA) when compared with control group were *Po0.001, **Po0.03 (c). (d) Reduced amount of Treg lymphocytes. On day 20 of the experiment, tumors (n ¼ 3) were excised and obtained single-cell suspensions used to quantitate Treg lymphocyte levels. Treg lymphocytes were identified using FITC-CD4, APC-CD25 and PE-Foxp3 antibodies. Percentage of Foxp3 þ CD25high þ regulatory lymphocytes (subpopulation of CD4 þ lymphocytes) was determined from lymphocyte population gate. Considerably decreased amounts of tumor Treg lymphocytes were found for both regimens of combined therapies. Statistical differences (ANOVA) when compared with control group were *Po0.0004 (d).

Optimum exploitation of antiangiogenic therapeutic approach Use of DNA vaccine directed against endoglin allows to requires involvement of several pathways engaged in angiogen- circumvent several antiangiogenic drug resistance mechanisms. It esis.54 So far, endoglin-directed vaccines were thought to affect is directed against both endothelial and neoplastic cells. Double only tumor endothelial cells.28,40,45 Some data, however, suggest targeting allows inhibition of angiogenesis and eradication of the presence of this transmembrane protein also on neoplastic neoplastic cells, and might prove sufﬁcient to arrest growth cells.55 CD105 inhibits migration and motility of prostate cancer of certain types of tumors. One of the key advantages provided by cells.25,56 Overexpression of endoglin in squamous esophageal endoglin is its much stronger expression in actively proliferating cancer, as well as breast cancer, decreases invasiveness endothelial cells (43 Â 106 copies per cell), compared with other and tumorigenicity of neoplastic cells.57,58 On the other hand, in angiogenesis-related proteins, that is, VEGFRs (o0.2 Â 106 copies per melanoma or Ewing sarcoma, CD105 strongly stimulates tumor cell).60 The majority of antiangiogenic drugs acts only on proliferating progression.24 The level of endoglin expression in these two types endothelial cells that participate in angiogenesis. Endoglin present of cancer correlates with the ability of tumor cells to form vessel- on the surface of activated endothelial cells and transdifferentiated like structures (vasculogenic mimicry).59 CD105 protein regulates tumor cells also contributes to vasculogenic mimicry. In human expression of integrin b3 and osteopontin, both required for neuroblastoma tumors, up to 70% endothelial cells show both vasculogenic mimicry.24 Elimination of tumor cells featuring expression of CD105 and ampliﬁcation of MYCN oncogene.61 By endoglin on their surface may stimulate aggressiveness (prostate destroying cells featuring endoglin (endothelial cells and transdiffer- cancer) or inhibit progression (melanoma). This is why DNA entiated neoplastic cells), vasculogenic mimicry can be inhibited. vaccine directed against endoglin could be used only in certain Another drug resistance factor, hypoxia, stimulates secretion of types of cancers. proangiogenic factors and tumor progression.62,63 Hypoxia induces

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 262 -- 273 DNA vaccine in combined cancer therapy M Jarosz et al 270

PBS

CTX

* DNA vaccine

* DNA * vaccine + CTX

* * *

Figure 9. Inhibition of B16-F10 tumor growth in response to combination therapy involving endoglin-based DNA vaccine and CTX. (a, b) Inhibition of tumor growth. Following challenge with B16-F10 cells, C57Bl/6 mice (n ¼ 6) were orally administered the endoglin-based DNA vaccine. Additionally, on days 7, 13 and 19 following inoculation with tumor cells, CTX was administered intraperitoneally (Materials and Methods). Tested regimen of combined therapy proved to be much more effective (as compared with monotherapies) in inhibiting tumor growth. Statistical difference (ANOVA) when compared with control group (on 22 day of therapy) was *Po0.0005 (a). Photographs were taken on day 23 of the therapy (b). Considerable necrotic areas and infiltrations of immune system cells could be detected in tumor sections from animals treated with either DNA vaccine, or with immunomodulatory factors, or with one of the combined regimens (day 24 of the therapy, see b). Lens magnification Â 20. (c) Reduced density of tumor microvessels. On day 24 of the combined therapy, mice (n ¼ 2) were killed and tumors excised for staining with antibody against CD31. The number of vessels for a given experimental group was counted in five visual fields (four tumor sections). A considerably decreased number of blood vessels was found in tumor tissue sections from mice treated with the combination of drugs. Statistical difference (ANOVA) when compared with control group was *Po0.0002 (c). (d) Reduced amount of Treg lymphocytes. On day 24 of the experiment, tumors (n ¼ 3) were excised and obtained single-cell suspensions used to quantitate Treg þ high þ lymphocyte levels. Treg lymphocytes were identified using FITC-CD4, APC-CD25 and PE-Foxp3 antibodies. Percentage of Foxp3 CD25 regulatory lymphocytes (subpopulation of CD4 þ lymphocytes) was determined from lymphocyte population gate. Considerably decreased amounts of tumor Treg lymphocytes were found for both regimens of combined therapies. Statistical difference (ANOVA) when compared with control group was **Po0.03 (d).

elevated expression of CD105 in vascular endothelial cells.22 CD105 about a therapeutic effect seen as tumor growth arrest.15 Suitable acts independently of TGF-b counteracting apoptosis in endothelial combinations of our DNA vaccine with IL-12 or CTX brought about cells subjected to hypoxic conditions. Thus, elimination of endoglin long-term stabilization of tumor growth and, in some cases, leads to enhanced cell apoptosis under such conditions.64 complete regression. In order to further enhance therapeutic benefits of our oral DNA The investigated DNA vaccine suppressed neovascularization vaccine, we tested it in combination with IL-12-mediated and decreased the number of tumor blood vessels but it did not gene therapy or with chemotherapy (CTX). IL-12 is a pleiotropic affect other physiological forms of angiogenesis (wound healing). cytokine known to activate immune response.33 -- 36 The benefits Neither did it affect body mass nor cause behavioral changes in of IL-12-mediated gene therapy include local production of this the experimental animals. Side effects were not observed with proinflammatory cytokine within the tumor.35 Intratumoral other vaccines directed against angiogenesis-participating factors administration of the therapeutic construct leads to its increased (such as VEGFA, basic fibroblast growth factor or TEM8).40 concentration at the site of endothelial cells’ proliferation and this To conclude, the oral DNA vaccine developed by us is a variant markedly enhances the antiangiogenic and antitumor effect.65 of anticancer therapies aimed at eliminating tumor blood vessels CTX, besides destroying neoplastic cells, is an immunomodulatory and neoplastic cells. Suitable combinations of this DNA vaccine agent.36 CTX activates a nonspecific response, restores prolifera- with antiangiogenic and immunomodulatory agents, such as IL-12 tion of peripheral T lymphocytes as well as significantly decreases or CTX, caused polarization of tumor microenvironmental the number of regulatory T lymphocytes.37,38 cells towards antiangiogenic, immunosuppression-abrogating Accumulation of Treg in tumors shifts the balance of phenotype, long-term stabilization of tumor growth and even tumor microenvironmental equilibrium (involving regulatory complete tumor eradication. The therapeutic strategy combining T lymphocytes and effector cells) in the direction of immuno- antiangiogenic therapy with immunotherapy proved to be an suppression.66,67 Abrogation of this immunosuppression can bring effective and promising therapeutic approach.

Gene Therapy (2013) 262 -- 273 & 2013 Macmillan Publishers Limited DNA vaccine in combined cancer therapy M Jarosz et al 271 MATERIALS AND METHODS showing green fluorescence (that is, those containing EGFP), the sections Animals, bacterial strains, plasmids and cell lines were viewed using fluorescence microscope (magnification Â 20). Mice (6- to 8-week-old CB7Bl/6 and BALB/c females) were bred in house. The experimental protocol was approved by Local Ethics Commission Oral immunization, tumor cell challenge, treatment with (Medical University of Silesia, Katowice, Poland). Salmonella Typhimurium immunomodulatory factors À þ LB5000 bacterial strain (r m ) was obtained from Professor I Gentschev In prophylactic experiments (induction of immune response before (University of Wu¨rzburg, Germany). The attenuated S. Typhimurium SL7207 inoculation with cancer cells), mice were immunized by oral gavage À (aroA ) attenuated strain was kindly provided by Professor CA Guzmań (repeated three times at 1-week intervals), of SL7207/mCD105 bacteria (German Research Center of Biotechnology, Braunschweig, Germany). Both (1 Â 108 per 100 ml PBSÀ).20,29 Five minutes before administration, mice strains were cultured in LB broth supplemented with 100 mgml-- 1 of were given 100 ml 10% NaHCO3 in order to neutralize gastric acid. In the ampicillin and with a mixture of amino acids, when required. Plasmid fourth week, mice were challenged (subcutaneously or intravenously) pBCMGSNeo carrying a gene encoding murine IL-12 was obtained from with B16-F10 cells (1 Â 105 per 100 ml PBSÀ) or Renca cells (5 Â 105 per Professor H Yamamoto (Osaka University, Japan). Plasmid preparations 100 ml PBSÀ). On day 23 of the experiment, the mice were killed, lungs were isolated using QIAGEN-Endo Free Giga Kit (Qiagen GmbH, Hilden, were excised and fixed in Bouin’s solution and the number of experimental Germany). HECa10 (murine endothelial cells), J774.2 (mouse BALB/c metastases in lungs was determined. In therapeutic experiments, 1 day monocyte macrophage cells), B16-F10 (murine melanoma), C26 (murine after inoculating mice with B16-F10 melanoma cells (1 Â 105 per 100 ml colon carcinoma), Renca (murine renal carcinoma) and COS7 (African green PBSÀ), oral administration of the SL7207/mCD105 DNA vaccine (1 Â 108 monkey SV-40-transferred kidney fibroblast cells) cell lines (ATCC, per 100 ml PBSÀ) was initiated. Bacteria were administered three times, 1 Manassas, VA, USA) were maintained using RPMI 1640 medium week apart.31 Five minutes before each administration, mice were given supplemented with 10% fetal bovine serum. COS7/mCD105 transfected 100 ml 10% NaHCO . Additionally in combined therapy, on days 9, 11 and -- 1 3 cells were grown using medium containing 1 mg ml G-418 (Sigma 13 following inoculation with cancer cells, pBCMGSNeo/mIL-12 plasmid Aldrich, St Louis, MO, USA). Cell cultures were maintained in standard was injected directly into tumors (20 mg DNA per 100 ml PBSÀ).31 CTX 1 conditions (37 C, 5% CO2, 95% humidity). (Endoxan, Baxter, Halle, Germany) was administered intraperitoneally (170 mg kg -- 1 body mass)39 on days 7, 13 and 19 following tumor cell Vector construction and western blotting challenge. Growing tumors were measured with calipers and tumor volumes were determined using the formula: volume ¼ width2 Â The endoglin-encoding cDNA sequence was obtained by PCR. As a length Â 0.52. template, cDNA form murine salivary gland was used (Invitrogen, Carlsbad, CA, USA). Two oligonucleotides were used as starters: 50-AAAAGGTACC ATGGACCGTGGCGTGC-30 and 50-AAAATCTAGACTACGCCATGCTGCTGGT Immunohistochemistry G-30. The PCR product was cloned into pcDNA3.1( þ ) plasmid vector Paraffin sections (5 mm) were prepared from tumor material collected and (Invitrogen) using KpnI and XbaI restriction sites under the control of the fixed on the 20th or 24th day of the experiments testing one of the combined cytomegalovirus promoter. Recombinant plasmid pcDNA3.1( þ )/CD105 therapies (DNA vaccine þ IL-12 or DNA vaccine þ CTX, respectively). The was then confirmed by sequencing to be identical to those in GeneBank sections were then stained immunohistochemically: following incubation at (GI: BC029080). Expression of mCD105-encoding gene was determined by 4 1C for 19 h with rabbit anti-CD31 polyclonal primary antibody (Abcam, western blotting in cell lysates. The recombinant pcDNA3.1( þ )/mCD105 Cambridge, UK), the sections were incubated at room temperature for 30 min plasmid and pcDNA3.1( þ ) control plasmid were transfected into COS7 with secondary antibody labeled with horseradish peroxidase. Finally, the cells using FuGENE transfection reagent (Roche, Basel, Switzerland). sections were incubated with Vector VIP peroxidase substrate (Vector Transfected cells were cultured under selective conditions (1 mg ml -- 1 Laboratories, Burlingame, CA, USA) and, additionally, with hematoxylin. G-418). To identify endoglin protein, a rat monoclonal antibody recognizing murine endoglin was used (R&D Systems, Minneapolis, MN, USA) and goat biotinylated antibody directed against rat IgG (R&D Systems). Flow cytometric analysis Endoglin expression was identified in a murine endothelial cell line HECa10 (positive control) and in three tumor cell lines (B16-F10, Renca and C26). DNA vaccine construct The expression level was determined using PE-CD105 antibody and pcDNA3.1( þ )/mCD105 plasmid with inserted endoglin-coding sequence PE-isotype control. Spleen-derived single-cell suspensions from immunized (1 mg) and pcDNA3.1( þ ) control plasmid (1 mg) were transferred into mice were isolated after 2--3 weeks following inoculation with cancer cells. 29 freshly preparated bacteria Salmonella Typhimurium LB5000 (3 Â 108)by The preparations were checked for the level of activated T lymphocytes electroporation (500 V, 1 impulse, 17 ms) using ECM 830 apparatus using the following antibodies: PE-CD28 and FITC-CD8a (eBioscences, San (BTX Harvard Apparatus, Holliston, MA, USA). Plasmid DNA isolated from Diego, CA, USA); PE-Cy7-CD25, PE-CD69 and FITC-CD3e (BD Pharmingen, single colonies obtained using selection plates was transferred by San Diego, CA, USA). Also, the expression level of co-stimulating molecules electroporation into attenuated Salmonella Typhimurium SL7207 strain on dendritic cells was checked using PE-CD80, PE-CD86 and FITC-CD11c (3 Â 108). Resistant colonies harboring the vaccine vectors were used in antibodies (eBioscences). For intracellular staining of IFN-g, TNF-a i IL-2, subsequent experiments. splenocytes were stimulated with Leukocyte Activation Cocktail (BD Pharmingen) for 4 h. The following antibodies were used to identify cytokines: PE-IFN-g, PE-TNF-a, PE-IL-2, FITC-CD8a and FITC-CD4 (BD In vitro and in vivo gene transfer Pharmingen). Three weeks following inoculation of immunized mice with Plasmid pcDNA3.1( þ )/EGFP (Clontech, Mountain View, CA, USA) was B16-F10 melanoma cells, tumor material was collected for flow cytometric electroporated into Salmonella Typhimurium SL7207 bacteria. For in vitro analysis and single-cell suspension obtained using a digestion mix experiment, J774.2 cells (2 Â 105 per well) were seeded into a 24-well plate. (0.5 mg ml -- 1 collagenase A, Sigma Aldrich; 0.2 mg ml -- 1 hyaluronidase After a 24-h incubation, Salmonella Typhimurium bacteria carrying EGFP type V, Sigma Aldrich; 0.02 mg ml -- 1 DNase I, Roche; per each 0.25 g of gene were added. Gentamycin was further added after 30 min incubation, tumor tissue). Red blood cells were lysed using 0.15 M ammonium chloride followed by tetracycline (after 4 h).68 Cells containing EGFP protein were solution. Dead cells were removed by centrifugation on Lympholyte-M examined at three time points (24, 48 and 72 h) using fluorescence micro- gradients (Cedarlane, Ontario, Canada).36 To identify the subpopulations of scopy (magnification Â 40). For in vivo experiment, mice were orally T lymphocytes, the following antibodies were used: PE-Cy7-CD3e, PE-CD4 8 administered SL7207 bacteria containing the plasmid (1 Â 10 per 100 ml and FITC-CD8a (BD Pharmingen). Treg lymphocytes were identified with PBSÀ). After 24 h, mice were killed and PPs were excised.44,69 Cryostat FITC-CD4, APC-CD25 and PE-Foxp3 antibodies (eBiosciences). Finally, to sections (10 mm thick) were then obtained. In order to identify cells identify the level of NK cells, an anti-mouse CD49b (pan-NK cells) antibody

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 262 -- 273 DNA vaccine in combined cancer therapy M Jarosz et al 272 was used (eBioscences). In all flow cytometry analyses (BD FACSCanto, BD, 6 Loges S, Schmidt T, Carmeliet P. Mechanisms of resistance to anti-angiogenic Franklin Lakes, NJ, USA), gate dividing negative from positive cells was therapy and development of third-generation anti-angiogenic drug candidates. based on isotype antibody control probes. Genes Cancer 2010; 1: 12 -- 25. 7 Carmeliet P, Jain RK. Molecular mechanisms and clinical applications of angiogenesis. Nature 2011; 473: 298 -- 307. Cytotoxicity assay 8Do¨me B, Hendrix MJC, Paku S, To´ va´ri J, T´ıma´r J. Alternative vascularization In all, 10 days after tumor cell challenge, splenocytes were isolated from mechanisms in cancer: pathology and therapeutic implications. Am J Pathol 2007; 170: 1 -- 15. immunized mice and co-cultured (37 1C/5% CO2) with mitomycin C (80 mg per 107 cells;70 Sigma Aldrich)-treated endoglin-expressing target cells 9 Ricci-Vitiani L, Pallini R, Biffoni M, Todaro M, Invernici G, Cenci T et al. Tumour (HECa10) using complete T-STIM culture medium (BD Biosciences, San vascularization via endothelial differentiation of glioblastoma stem-like cells. Nature 2010; 468: 824 -- 828. Jose, CA, USA). Four days later, cytotoxicity of these splenocytes was 10 Wang R, Chadalavada K, Wilshire J, Kowalik U, Hovinga KE, Geber A et al. assessed using In Vitro Toxicology Assay Kit (Sigma Aldrich), an XTT-based Glioblastoma stem-like cells give rise to tumour endothelium. Nature 2010; 468: test against target cell lines: B16-F10, HECa10 cells (both overexpressing 829 -- 833. endoglin) and C26 cells (without endoglin; negative control). 11 Soda Y, Marumoto T, Friedmann-Morvinski D, Soda M, Liu F, Michiue H et al. Transdifferentiation of glioblastoma cells into vascular endothelial cells. Proc Natl Acad Sci USA 2011; 108: 4274 -- 4280. Angiogenesis assay 12 Paèz-Ribes M, Allen E, Hudock J, Takeda T, Okuyama H, Vinãls F et al. Two weeks after last immunization, mice were injected subcutaneously, near Antiangiogenic therapy elicits malignant progression of tumors to increased the abdominal midline, with 500 ml of growth factor reduced Matrigel (BD local invasion and distant metastasis. Cancer Cell 2009; 15: 220 -- 231. Pharmingen) containing 0.4 mgml-- 1 of recombinant mouse fibroblast growth 13 Ebos JML, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG, Kerbel RS. factor-basic (R&D Systems). After 6 days, 200 ml(0.1mgml-- 1)offluorescein- Accelerated metastasis after short-term treatment with a potent inhibitor of labeled Griffonia simplicifolia lectin I, Isolectin B4 (Vector Laboratories) was tumor angiogenesis. Cancer Cell 2009; 15: 232 -- 239. 14 Keunen O, Johansson M, Oudin A, Sanzey M, Rahim SA, Fack F et al. Anti-VEGF injected into tail vein. Then, after 20 min, the mice were killed, Matrigel plugs 69,70 treatment reduces blood supply and increases tumor cell invasion in were excised and homogenized in RIPA lysis buffer. The lysates (n ¼ 7) glioblastoma. Proc Natl Acad Sci USA 2011; 108: 3749 -- 3754. were centrifuged and fluorescein content in supernatants was determined by 15 Szala S, Mitrus I, Sochanik A. Can inhibition of angiogenesis and stimulation of fluorimetry at 515 nm. In each case, background fluorescence was immune response be combined into a more effective antitumor therapy? determined in noninjected control mice and subtracted. Cancer Immunol Immunother 2010; 59: 1449 -- 1455. 16 Ahmad M, Rees RC, Ali SA. Escape from immunotherapy: possible mechanisms that influence tumor regression/progression. Cancer Immunol Immunother 2004; Assessment of possible side effects 53: 844 -- 854. Four full-thickness circular excisional wounds (4--5 mm diameter) per 17 Kim R, Emi M, Tanabe K, Arihiro K. Tumor-driven evolution of immuno- mouse were made 1 week after the last immunization and time elapsed suppressive networks during malignant progression. Cancer Res 2006; 66: until complete wound closure was measured.20,71 Animal behavior and 5527 -- 5536. changes in body mass of the treated and control mice were evaluated. 18 Disis ML. Enhancing cancer vaccine efficacy via modulation of the tumor microenvironment. Clin Cancer Res 2009; 15: 6476 -- 6478. 19 Ostrand-Rosenberg S. Immune surveillance: a balance between protumor and Statistical analysis antitumor immunity. Curr Opin Genet Dev 2008; 18:11--18. Tumor growth kinetics and statistical significance of differences between 20 Niethammer AG, Xiang R, Becker JC, Wodrich H, Pertl U, Karsten G et al. A DNA vaccine against VEGF receptor 2 prevents effective angiogenesis and inhibits the experimental and control groups in immunological and immuno- tumor growth. Nat Med 2002; 8: 1369 -- 1375. histochemical analyses were evaluated by ANOVA (analysis of variance). 21 Duff SE, Li C, Garland JM, Kumar S. CD105 is important for angiogenesis: evidence The comparisons between the two groups were analyzed by Student’s and potential applications. FASEB J 2003; 17: 984 -- 992. t-test. P-values o0.05 were considered statistically significant. 22 Dallas NA, Samuel S, Xia L, Fan F, Gray MJ, Lim SJ et al. Endoglin (CD105): a marker of tumor vasculature and potential target for therapy. Clin Cancer Res 2008; 14: 1931 -- 1937. CONFLICT OF INTEREST 23 Fonsatti E, Nicolay HJM, Altomonte M, Covre A, Maio M. Targeting cancer vasculature via endoglin/CD105: a novel antibody-based diagnostic and The authors declare no conflict of interest. therapeutic strategy in solid tumours. Cardiovasc Res 2010; 86:12--19. 24 Pardali E, van der Schaft DW, Wiercinska E, Gorter A, Hogendoorn PC, Griffioen AW et al. Critical role of endoglin in tumor cell plasticity of Ewing sarcoma and ACKNOWLEDGEMENTS melanoma. Oncogene 2010; 30: 334 -- 345. We thank Professor J Rak for editorial help with manuscript preparation. We thank 25 Romero D, O’Neill C, Terzic A, Contois L, Young K, Conley BA et al. Professor I Gentschev, who supplied the Salmonella Typhimurium LB5000 bacterial Endoglin regulates cancer-stromal cell interactions in prostate tumors. strain, and Professor CA Guzmań, who provided the Salmonella Typhimurium SL7207 Cancer Res 2011; 71: 3482 -- 3493. and Professor H Yamamoto for the plasmid pBCMGSNeo/IL-12. This study was 26 Stocker BAD. Aromatic-dependent Salmonella as anti-bacterial vaccines and as supported by a Grant No. PBZ-MNiI-1/1/2005 from Pharmaceutical Research Institute presenters of heterologous antigens or of DNA encoding them. J Biotechnol 2000; (Warsaw, Poland) and a Grant No. NN 401 587 540 from the Ministry of Science and 83: 45 -- 50. Higher Education (Poland). 27 Darji A, Guzmań CA, Gerstel B, Wachholz P, Timmis KN, Wehland J et al. Oral somatic transgene vaccination using attenuated S. typhimurium. Cell 1997; 91: 765 -- 775. 28 Reisfeld RA, Niethammer AG, Luo Y, Xiang R. DNA vaccines suppress tumor REFERENCES growth and metastases by the induction of anti-angiogenesis. Immunol Rev 2004; 1 Tabi Z, Man S. Challenges for cancer vaccine development. Adv Drug Deliv Rev 199: 181 -- 190. 2006; 58: 902 -- 915. 29 Lee S-H, Mizutani N, Mizutani M, Luo Y, Zhou H, Kaplan C et al. Endoglin (CD105) is 2 Hofmeister V, Vetter C, Schrama D, Bro¨cker E-B, Becker JC. Tumor stroma- a target for an oral DNA vaccine against breast cancer. Cancer Immunol associated antigens for anti-cancer immunotherapy. Cancer Immunol Immunother Immunother 2006; 55: 1565 -- 1574. 2006; 55: 481 -- 494. 30 Jiao J-G, Li Y-N, Wang H, Liu Q, Cao J-X, Bai R-Z et al. A plasmid DNA vaccine 3 Azam F, Mehta S, Harris AL. Mechanisms of resistance to antiangiogenesis encoding the extracellular domain of porcine endoglin induces anti-tumour therapy. Eur J Cancer 2010; 46: 1323 -- 1332. immune response against self-endoglin-related angiogenesis in two liver cancer 4 Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 1971; 285: models. Dig Liver Dis 2006; 38: 578 -- 587. 1182 -- 1186. 31 Feng K-K, Zhao H-Y, Qiu H, Liu J-X, Chen J. Combined therapy with flk1-based 5 Ebos JML, Kerbel RS. Antiangiogenic therapy: impact on invasion, disease DNA vaccine and interleukin-12 results in enhanced antiangiogenic and progression, and metastasis. Nat Rev Clin Oncol 2011; 8: 210 -- 221. antitumor effects. Cancer Lett 2005; 221:41--47.

Gene Therapy (2013) 262 -- 273 & 2013 Macmillan Publishers Limited DNA vaccine in combined cancer therapy M Jarosz et al 273 32 Loeffler M, Kru¨ger JA, Reisfeld RA. Immunostimulatory effects of low-dose 51 Tascon RE, Ragno S, Lowrie DB, Colston MJ. Immunostimulatory bacterial DNA cyclophosphamide are controlled by inducible nitric oxide synthase. Cancer Res sequences activate dendritic cells and promote priming and differentiation of 2005; 65: 5027 -- 5030. CD8+ T cells. Immunology 2000; 99:1--7. 33 Portielje JEA, Gratama JW, van Ojik HH, Stoter G, Kruit WHJ. IL-12: a promising 52 Fioretti D, Iurescia S, Fazio VM, Rinaldi M. DNA vaccines: developing new adjuvant for cancer vaccination. Cancer Immunol Immunother 2003; 52: 133 -- 144. strategies against cancer. J Biomed Biotechnol 2010; 2010: 174378. 34 Del Vecchio M, Bajetta E, Canova S, Lotze MT, Wesa A, Parmiani G et al. 53 Maloy KJ, Erdmann I, Basch V, Sierro S, Kramps TA, Zinkernagel RM et al. Interleukin-12: biological properties and clinical application. Clin Cancer Res 2007; Intralymphatic immunization enhances DNA vaccination. Proc Natl Acad Sci USA 13: 4677 -- 4685. 2001; 98: 3299 -- 3303. 35 Uemura A, Takehara T, Miyagi T, Suzuki T, Tatsumi T, Ohkawa K et al. Natural killer 54 Jain RK. Normalization of tumor vasculature: an emerging concept in cell is a major producer of interferon c that is critical for the IL-12-induced antiangiogenic therapy. Science 2005; 307:58--62. anti-tumor effect in mice. Cancer Immunol Immunother 2010; 59: 453 -- 463. 55 ten Dijke P, Goumans MJ, Pardali E. Endoglin in angiogenesis and vascular 36 Kilinc MO, Aulakh KS, Nair RE, Jones SA, Alard P, Kosiewicz MM et al. Reversing diseases. Angiogenesis 2008; 11: 79 -- 89. tumor immune suppression with intratumoral IL-12: activation of tumor- 56 Liu Y, Jovanovic B, Pins M, Lee C, Bergan RC. Over expression of endoglin in associated T effector/memory cells, induction of T suppressor apoptosis, and human prostate cancer suppresses cell detachment, migration and invasion. infiltration of CD8+ T effectors. J Immunol 2006; 177: 6962 -- 6973. Oncogene 2002; 21: 8272 -- 8281. 37 Ghiringhelli F, Menard C, Puig PE, Ladoire S, Roux S, Martin F et al. Metronomic 57 Wong VC, Chan PL, Bernabeu C, Law S, Wang LD, Li JL et al. Endoglin (ENG), cyclophosphamide regimen selectively depletes CD4+CD25+ regulatory T cells silenced by both epigenetic inactivation and allelic loss in esophageal squamous and restores T and NK effector functions in end stage cancer patients. cell carcinoma. Int J Cancer 2008; 123: 2816 -- 2823. Cancer Immunol Immunother 2007; 56: 641 -- 648. 58 Henry LA, Johnson DA, Sarrio´ D, Lee S, Quinlan PR, Crook T et al. Endoglin 38 Zitvogel L, Apetoh L, Ghiringhelli F, Kroemer G. Immunological aspects of cancer expression in breast tumor cells suppresses invasion and metastasis and chemotherapy. Nat Rev Immunol 2008; 8:59--73. correlates with improved clinical outcome. Oncogene 2011; 30: 1046 -- 1058. 39 Browder T, Butterfield CE, Kra¨ling BM, Shi B, Marshall B, O’Reilly MS et al. 59 Pe´rez-Go´ mez E, Del Castillo G, Juan Francisco S, Lo´pez-Novoa JM, BernabeúC, Antiangiogenic scheduling of chemotherapy improves efficacy against Quintanilla M. The role of the TGF-b coreceptor endoglin in cancer. experimental drug-resistant cancer. Cancer Res 2000; 60: 1878 -- 1886. ScientificWorldJournal 2010; 10: 2367 -- 2384. 40 Matejuk A, Leng Q, Chou S-T, Mixson AJ. Vaccines targeting the neovasculature of 60 Zhang Y, Yang Y, Hong H, Cai W. Multimodality molecular imaging of CD105 tumors. Vasc Cell 2011; 3:7. (Endoglin) expression. Int J Clin Exp Med 2011; 4:32--42. 41 Okaji Y, Tsuno NH, Kitayama J, Saito S, Takahashi T, Kawai K et al. Vaccination with 61 Pezzolo A, Parodi F, Corrias MV, Cinti R, Gambini C, Pistoia V. Tumor origin of autologous endothelium inhibits angiogenesis and metastasis of colon endothelial cells in human neuroblastoma. J Clin Oncol 2007; 25: 376 -- 383. cancer through autoimmunity. Cancer Sci 2004; 95:85--90. 62 Abdollahi A, Folkman J. Evading tumor evasion: current concepts and 42 Gardlik R, Behuliak M, Palffy R, Celec P, Li CJ. Gene therapy for cancer: perspectives of anti-angiogenic cancer therapy. Drug Resist Updat 2010; 13: bacteria-mediated anti-angiogenesis therapy. Gene Therapy 2011; 18: 425 -- 431. 16 -- 28. 43 Paglia P, Medina E, Arioli I, Guzman CA, Colombo MP. Gene transfer in dendritic 63 De Bock K, Mazzone M, Carmeliet P. Antiangiogenic therapy, hypoxia, and cells, induced by oral DNA vaccination with Salmonella typhimurium, metastasis: risky liaisons, or not? Nat Rev Clin Oncol 2011; 8: 393 -- 404. results in protective immunity against a murine fibrosarcoma. Blood 1998; 92: 64 Li C, Issa R, Kumar P, Hampson IN, Lopez-Novoa JM, Bernabeu C et al. CD105 3172 -- 3176. prevents apoptosis in hypoxic endothelial cells. J Cell Sci 2003; 116: 2677 -- 2685. 44 Luo Y, Zhou H, Mizutani M, Mizutani N, Liu C, Xiang R et al. A DNA vaccine 65 Dickerson EB, Akhtar N, Steinberg H, Wang Z-Y, Lindstrom MJ, Padilla ML et al. targeting Fos-related antigen 1 enhanced by IL-18 induces long-lived T-cell Enhancement of the antiangiogenic activity of interleukin-12 by peptide targeted

memory against tumor recurrence. Cancer Res 2005; 65: 3419 -- 3427. delivery of the cytokine to avb3 integrin. Mol Cancer Res 2004; 2: 663 -- 673. 45 Tan GH, Tian L, Wei YQ, Zhao X, Li J, Wu Y et al. Combination of low-dose cisplatin 66 Curiel TJ, Coukos G, Zou L, Alvarez X, Cheng P, Mottram P et al. Specific and recombinant xenogeneic endoglin as a vaccine induces synergistic antitumor recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege activities. Int J Cancer 2004; 112: 701 -- 706. and predicts reduced survival. Nat Med 2004; 10: 942 -- 949. 46 Tan G-H, Li Y-N, Huang F-Y, Wang H, Bai R-Z, Jang J. Combination of recombinant 67 Zou W. Immunosuppressive networks in the tumour environment and their xenogeneic endoglin DNA and protein vaccination enhances anti-tumor effects. therapeutic relevance. Nat Rev Cancer 2005; 5: 263 -- 274. Immunol Invest 2007; 36: 423 -- 440. 68 Michael A, Stratford R, Khan S, Dalgleish A, Pandha H. Novel strains of Salmonella 47 Tsujie M, Tsujie T, Toi H, Uneda S, Shiozaki K, Tsai H et al. Anti-tumor activity of typhimurium as potential vectors for gene delivery. FEMS Microbiol Lett 2004; 238: an anti-endoglin monoclonal antibody is enhanced in immunocompetent mice. 345 -- 351. Int J Cancer 2008; 122: 2266 -- 2273. 69 Xiang R, Mizutani N, Luo Y, Chiodoni C, Zhou H, Mizutani M et al. A DNA vaccine 48 Uneda S, Toi H, Tsujie T, Tsujie M, Harada N, Tsai H et al. Anti-endoglin monoclonal targeting survivin combines apoptosis with suppression of angiogenesis in lung antibodies are effective for suppressing metastasis and the primary tumors by tumor eradication. Cancer Res 2005; 65: 553 -- 561. targeting tumor vasculature. Int J Cancer 2009; 125: 1446 -- 1453. 70 Kaplan CD, Kru¨ger JA, Zhou H, Luo Y, Xiang R, Reisfeld RA. A novel DNA vaccine 49 Paglia P, Terrazzini N, Schulze K, Guzmań CA, Colombo MP. In vivo correction of encoding PDGFRbeta suppresses growth and dissemination of murine colon, lung genetic defects of monocyte/macrophages using attenuated Salmonella as oral and breast carcinoma. Vaccine 2006; 24: 6994 -- 7002. vectors for targeted gene delivery. Gene Therapy 2000; 7: 1725 -- 1730. 71 Loeffler M, Kru¨ger JA, Niethammer AG, Reisfeld RA. Targeting tumor-associated 50 Sirard J-C, Niedergang F, Kraehenbuhl J-P. Live attenuated Salmonella: a fibroblasts improves cancer chemotherapy by increasing intratumoral drug paradigm of mucosal vaccines. Immunol Rev 1999; 171: 5 -- 26. uptake. J Clin Invest 2006; 116: 1955 -- 1962.

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 262 -- 273