A Novel Strategy for the Generation of Angiostatic Kringle Regions from a Precursor Derived from Plasminogen

Home , Furin

Gene Therapy (2002) 9, 1600–1606  2002 Nature Publishing Group All rights reserved 0969-7128/02 $25.00 www.nature.com/gt RESEARCH ARTICLE A novel strategy for the generation of angiostatic kringle regions from a precursor derived from plasminogen

V Schmitz1,2, L Wang1, M Barajas1, D Peng1, J Prieto1 and C Qian1 1Division of Hepatology and Gene Therapy, Department of Medicine, Medical School, University of Navarra, Pamplona, Spain

In this study we have explored the feasibility of generating AdmuthPlgK3 and AdK1-3 in mice results in significantly angiostatin by incorporating an endoproteolytic furin cleav- decreased endothelial cell infiltration in VEGF-embedded age site into plasminogen to allow conversion of the precur- Matrigel plugs. Treatment with AdmuthPlgK3 and AdK1-3 sor molecule into an angiostatic active K1-3 fragment. We exerts strong antitumoral effect in models of hepatocellular show that secretable angiostatin can be successfully gener- carcinoma and Lewis lung cancer. This antitumor effect was ated from cells infected with adenovirus carrying the furin- associated with decreased microvessel density in the mutated plasminogen (AdmuthPlgK3). Supernatant from tumors. Taken together, our data demonstrate that angiosta- cells transduced with AdmuthPlagK3 inhibits tube formation tin endowed with strong antiangiogenic and antitumor effects and proliferation and migration of human umbilical vein can be released from a furin-mutated plasminogen acting endothelial cells with an efficiency similar to that of super- as a precursor. This strategy may have potential to develop natant from cells infected with adenovirus expressing kringle angiostatic anti-cancer therapies. 1–3 of plasminogen (AdK1-3). Administration of Gene Therapy (2002) 9, 1600–1606. doi:10.1038/sj.gt.3301805

Keywords: gene therapy; cancer; angiogenesis; angiostatin; endoprotease furin

Introduction a process that can restrain metastasis.6,7 Angiostatin can inhibit bFGF-induced endothelial cell proliferation in Tumor angiogenesis plays a key role in the development vitro and bFGF-induced angiogenesis in vivo.5 It has been 1 of a wide variety of malignant tumors. Whereas smaller demonstrated that other fragments of plasminogen, such tumor cell conglomerates can sufficiently be nourished as the first three kringle (K1–3) domains and kringle 5 by diffusion, further tumor growth of nodules with more (K5) exhibit potent inhibitory activities on endothelial cell 3 than 1–2 mm in volume depends on tumor angiogenesis. proliferation.8,9 Inhibition of angiogenesis by treatment Increasing tumor size requires adequate nutrition, oxy- with angiostatin causes significant suppression of tumor genation and metabolization by improved blood supply growth in both murine and human tumor models.5,10–12 that can be achieved by sprouting of new vessel branches Gene therapy strategies against cancer by transfer of a from pre-existing ones. This finding forms the rationale DNA sequence encoding the region of plasminogen cor- behind angiostatic anti-cancer therapies that are aimed at responding to angiostatin have been applied in different pushing the equilibrium between pro- and antiangiog- tumor models.13–16 enic factors in favor of antiangiogenic factors. Antiangi- The ubiquitous endoprotease furin is involved in the ogenic anti-cancer approaches block tumor vasculariz- physiologic conversion of a broad spectrum of precursors ation or tumor angiogenesis by attacking pre-existing into active protein compounds including several growth vessels or inhibiting the sprouting of new branches from factors, hormones, plasma proteins, receptors, and matrix 2–4 those vessels. metalloproteinases.17,18 Furthermore some constituents of Angiostatin, an internal fragment of plasminogen that pathogens, such as human immunodeficiency virus contains the first four of five triple-looped kringle gp160 and diphtheria toxin, are processed in this man- domains of this protein, is an endogenous inhibitor of ner.18 It has been known that the endoprotease furin 5 angiogenesis (Figure 1). Proteolytic enzymes produced recognizes a consensus cleavage site of -ArgXLys/ by some primary tumors have been described to be ArgArg.17 The incorporation into pro-insulin of a furin involved in the physiological generation of angiostatin, cleavage site has been shown to result in the release of active insulin by stably transfected cells. In that study pro-protein to protein conversion reached 90% of the Correspondence: C Qian, Division of Hepatology and Gene Therapy, total amount of expressed pro-protein.19 Also, retroviral Department of Medicine, Medical School, University of Navarra, 31080 gene transfer of rat preproinsulin-1 carrying a furin cleav- Pamplona, Spain 2Present address: Department of Internal Medicine, University Hospital age site leads to the secretion of active insulin capable of 20 Bonn, Germany preventing ketoacidosis effectively. Recently, furin Received 6 March 2002; accepted 11 May 2002 cleavage site was used as a linker sequence to express Generation of angiostatin from a precursor V Schmitz et al 1601 Results

Construction of adenoviruses containing mutated furin- sensitive plasminogen (AdmuthPlgK3) and first three kringles of plasminogen (AdK1-3) We constructed mutated plasminogen carrying a furin restriction site between K3 and K4 by site-directed mutagenesis. The ProValSerThr amino acid sequence between K3 and K4 at position 1069–1072 of wild-type human plasminogen cDNA was changed to four argi- nines. The mutated furin-sensitive sequence was confirmed by sequencing as shown in Figure 1a. Recombi- nant adenoviruses, AdmuthPlgK3 carrying mutated furin-sensitive plasminogen and AdK1-3 carrying the first three kringles of plasminogen, were generated. In order to know whether the mutated plasminogen could be cleaved by furin, the conditioned medium (CM) from A549 cells infected in vitro with either AdmuthPlgK3, AdK1-3 or control vector AdlacZ was analyzed by West- ern blot using a polyclonal antibody against human plasminogen. As shown in Figure 1b, in CM from AdK1-3 infected cells three bands were observed in the MW range of 38 kDa corresponding to different postranscrip- tional glycosylation forms of the K1-3 fragment.7 The same bands were also detected in the CM from AdmuthPlgK3-infected cells. In the CM from these cells, two other bands were observed corresponding to sizes of plasminogen (98 kDa) and of the predicted furin cleavage Figure 1 (a) Schematic structure of plasminogen and nucleotide and amino acid sequences of region between K3 and K4. ProValSerThr in bold product (C-terminal) of the mutated plasminogen (60 at position 1069–1072 of wild-type plasminogen was changed into ArgAr- kDa). These data indicate that the mutated plasminogen gArgArg, a furin restriction site by site-directed mutagenesis. (b) Western can be cleaved to yield an angiostatin-like molecule that blot analysis for angiostatin-like molecules. CM from uninfected and vec- is secreted to the culture medium of the adenovirus- tor-infected cells was subjected to Western blot as described in Materials infected cells. and methods. Lane 1, marker; lane 2, recombinant plasminogen; lane 3, CM from uninfected cells; lane 4, CM from cells infected with AdK1-3; Biological activity of angiostatin like molecules produced lane 5, CM from cells infected with AdmuthPlagK3; lane 6, CM from by AdmuthPlgK3 and AdK1-3 in vitro cells infected with control vector AdlacZ. Angiostatin fragments were detected at three different forms (around 38 kDa) in CM from AdK1-3 To evaluate the anti-proliferative effect of CM from A549 and AdmuthPlagK3-infected cells. Another cleavage product of plasmin- cells infected with AdmuthPlgK3, AdK1-3 or control vec- ogen (C-terminal, around 60 kDa) was detected in sample from tor AdlacZ, HUVE cells were pre-incubated with CM at AdmuthPlgK3 infection. The full length of plasminogen (around 98 kDa) final concentration of 30%. We found that pre-incubation was detected in samples from uninfected and vector-infected cells, with of HUVE cells with CM from AdK1-3 and AdmuthPlgK3- more intensity in sample from AdmutPlgK3 infection due to incomplete cleavage. infected cells inhibited proliferation by 39% and 29%, respectively, as compared with CM from uninfected cells. This inhibitory effect appeared to be dose-dependent, since inhibition was less pronounced in assays with CM at final concentration of 15% (data not shown). CM from AdlacZ-infected cells had no inhibitory effect on prolifer- multiple gene products from a single cistron comprising ation of HUVE cells (data not shown). The effect on several proteins as one single fusion protein from which migration of HUVE cells exerted by CM from AdK1-3-, equimolar amounts of different active products were AdmuthPlgK3- or AdlacZ-infected A549 cells was quant- liberated by furin cleavage.21 ified in a modified double chamber migration assay. We The objective of this study was to generate angiostatic found that the percentage of cell migration was reduced kringles of plasminogen by processing this molecule by 64% and 78%, as compared with the control group through the physiological intracellular pathway for when using CM from AdK1-3 and AdmuthPlgK3, constitutive protein secretion. To this aim we incorpor- respectively (Figure 2a). CM from AdlacZ did not sig- ated the sequences of furin cleavage sites into human nificantly affect cell migration. On the other hand, while plasminogen cDNA between kringle region (K) three the ability to form tube-like structures on a Matrigel (K3) and four (K4). We have used an adenoviral vector membrane in culture dishes was not affected by incu- to transfer furin-mutated plasminogen cDNA bation with CM from non-infected cells or AdlacZ- (AdmuthPlgK3) and to study the antiangiogenic and infected cells (Figure 2b and c), the addition of CM from antitumor effects of the substance thus generated. Our AdK1-3 and AdmuthPlgK3-infected cells strongly data show that cells transduced with AdmuthPlgK3 reduced the formation of tube-like structures (Figure 2d release the first three kringle regions of plasminogen and e). Quantitative analysis showed that AdK1-3 and (K1–3) and this vector causes potent antiangiogenic and AdmuthPlgK3 inhibited tube formation by 43% and 31%, antitumor effects. respectively, in comparison to the control (Figure 2f). CM

Gene Therapy Generation of angiostatin from a precursor V Schmitz et al 1602

Figure 2 (a) Inhibition of HUVE cell migration by angiostatin expressed by adenoviral vectors. HUVE cell migration was determined in a double chamber migration assay. Data are given as mean and s.e.m. of four independent experiments. ∗P Ͻ 0.01 as compared with control vector AdlacZ. (b–f) Inhibition of HUVE cell tube formation by angiostatin expressed by adenoviral vectors. HUVE cells were pre-incubated with CM for 30 min in Matrigel-coated culture wells and were cultured in culture medium supplemented with growth factor for an additional 6 h. At this time-point, HUVE cells started to form capillary-like structures (b). Addition of CM Figure 3 Suppression of VEGF induced angiogenesis in vivo by systemic from AdlacZ-infected cells did not affect tube formation (c). In contrast, administration of adenoviruses. Matrigel containing VEGF was injected addition of CM from AdK1-3 and AdmuthplgK3 infected cells inhibited subcutaneously into mice which had received intravenous administration tube formation (d, e). (f). Intact tubes were quantified in the whole well of different adenoviruses 6 h before. After 14 days, Matrigel plugs were under microscope. Data are given as mean and s.e.m. of two independent removed and stained with HE. (a, b). Representative photomicrograph experiments. ∗P Ͻ 0.01 as compared with control vector AdlacZ. Original showing VEGF-induced neovascularization in Matrigel from animals that magnification of b–e is ×40. received saline or control vector AdlacZ. (c, d) Representative photomicrographs showing inhibition of VEGF-induced neovascularization in Matri- gel from animals treated with AdK1-3 (c) and AdmuthPlgK3 (d). (e) Quantitative analysis of angiogenesis was made by counting the number of endothelium-like cells in an area of mm2. Data are presented expressed from AdlacZ-infected cells had no inhibitory effect on as mean and s.e.m. n = 6 animals per group. P Ͻ 0.01 as compared with control animals receiving AdlacZ or saline. Original magnification tube formation (Figure 2). Taken together, these in vitro × data demonstrate that angiostatin produced from either of panels a–d is 100. AdmuthPlgK3 or AdK1-3 exerts biological activities on endothelial cells. Effect of systemic administration of AdK1-3 and AdmuthPlgK3 on the inhibition of tumor growth in Inhibition of angiogenesis in vivo by systemic athymic mice bearing HCC tumors administration of AdmuthPlgK3 and AdK1-3 The antitumoral potency of AdK1-3 and AdmuthPlgK3 We have also explored whether angiogenesis could be was firstly evaluated in subcutaneous HCC tumors in inhibited in vivo after systemic administration of the athymic mice. Ten days after tumor cell inoculation, mice adenoviral vectors AdmuthPlgK3 and AdK1-3. To this were treated by intravenous injection of AdK1-3 or aim, animals were inoculated s.c. with VEGF-containing AdmuthPlgK3 or AdlacZ at the dose of 1 × 1010 Matrigel 6 h after systemic administration of p.f.u./animal or saline. As shown in Figure 4, treatment AdmuthPlgK3, and AdK1-3 or control vector AdlacZ. with AdK1-3 or AdmuthPlgK3 caused a significant inhi- Fourteen days later, mice were killed and Matrigel plugs bition of tumor growth, as compared with animals were collected for histological examination. As shown in treated with control vector AdlacZ or saline. Figure 3a and b, an intense invasion by endothelium-like cells was observed in Matrigel plugs from animals receiv- Inhibition of tumor growth on established LLC tumor in ing saline and control vector AdlacZ. In contrast, Matri- C57BL/6 and athymic mice by systemic administration gel plugs from mice which received AdK1-3 or of AdK1-3 and AdmuthPlgK3 AdmuthPlgK3 had much less infiltration of endothelium- We have also analyzed the antitumoral effect of AdK1-3 like cells (Figure 3c and d). Quantitative estimation of and AdmuthPlgK3 on well-established subcutaneous endothelium-like cells on Matrigel sections (Figure 3e) LLC tumors in syngenic C57BL/6 mice and athymic showed that administration of AdK1-3 and mice. When tumor size was about 50 mm3 in volume in AdmuthPlgK3 resulted in a reduction of cell infiltration both models, animals were treated by intravenous of 51% and 43%, respectively, in comparison to animals administration of AdK1-3 or AdmuthPlgK3 or AdlacZ at receiving the control vector AdlacZ. a dose of 1 × 1010 p.f.u./animal. Figure 5 showed that

Gene Therapy Generation of angiostatin from a precursor V Schmitz et al 1603

Figure 4 Inhibition of tumor growth in athymic mice with HCC tumors by systemic administration of AdK1-3 or AdmuthPlgK3. Ten days after implantation of Huh-7 cells, mice were treated by intravenous injection of recombinant adenoviruses (1010 p.f.u./mouse) or saline as control. Size of tumor was measured and presented as mean and s.e.m., n = 11–15. ∗P Ͻ 0.05 compared with the control groups receiving AdlacZ or saline. Figure 5 Inhibition of tumor growth on established LLC tumor in C57BL/6 and athymic mice by systemic administration of AdK1-3 or treatment with AdK1-3 or AdmuthPlgK3 caused a sig- AdmuthPlgK3. The established LLC tumor about 50 mm3 in C57BL/6 niﬁcant inhibition of tumor growth in both tumor-bear- mice (a) and in athymic mice (b) were treated by intravenous injection of recombinant adenoviruses (1010 p.f.u./mouse) or saline as control. Size of ing C57BL/6 mice and athymic mice. Administration of = ∗ control vector AdlacZ had no effect on tumor growth tumor was measured and presented as mean and s.e.m., n 11–15. P Ͻ 0.05 compared with the control groups receiving AdlacZ or saline. (Figure 5). In tumor-bearing C57BL/6 mice we observed, at day 2 and 4 after treatment, a reduction in tumor volume of 55% and 65% in AdK1-3- and AdmuthPlgK3- with control vector AdlacZ did not change intratumoral treated animals as compared with control animals receiv- microvessel density (Figure 6b and e). ing control vector AdlacZ (Figure 5a). In athymic mice, at day 2 after therapy, the reduction in tumor volume Discussion was 53% and 37% as compared with control animals, in AdK1-3- and AdmuthPlgK3-treated mice, respectively Here we show the production of a potent antiangiogenic (Figure 5b). factor from an engineered precursor molecule. The method described here is based on the incorporation of Effect of AdmuthPlgK3 and AdK1-3 on tumor an endoproteolytic furin restriction site in between krin- angiogenesis gle region three (K3) and four (K4) of wild-type plasmin- Microvessel density in the tumors was determined by ogen cDNA. When this engineered molecule is expressed immunohistochemistry for von Willebrand factor, as a in eukaryotic cells, it becomes a precursor molecule marker for endothelial cells. Athymic mice with subcut- allowing conversion into the active angiostatic K1-3 plas- aneous LLC tumor of about 50 mm3 in volume were minogen fragment by ubiquitous endoprotease furin treated by intravenous administration of AdK1-3 or the cleavage. The selection of a ubiquitous endoprotease AdmuthPlgK3 at dose of 1 × 1010 p.f.u./animal or control furin for conversion of an engineered precursor would vector AdlacZ at the same dose or saline. Tumors were allow protein processing in practically all tissues.17,18 This removed 7 days after treatment. As shown in Figure 6a is considered to be particularly attractive for angiostatic and b, tumors from animals receiving saline or control anti-cancer strategies, because the anti-angiogenic subst- vector AdlacZ, showed intensive staining for von Willeb- ance would be produced systemically thus obviating the rand factor, indicating the presence of extensive angiog- need for transducing the tumoral tissue directly.2,3 In fact, enesis in this type of tumor. Tumors from AdK1-3 or high systemic levels of the antiangiogenic factor are AdmuthPlgK3 treated animals showed a marked desirable since no relevant side-effects have been reduction of microvessel density (Figure 6c and d). Quan- described for plasminogen-derived angiostatic proteins, titative analysis demonstrated 59% and 61% reduction of so far. the intratumoral microvessel density in AdK1-3- and In this study we have used a ﬁrst generation adenovi- AdmuthPlgK3-treated animals, respectively, as com- ral vector as a vehicle to transfer mutant plasminogen in pared with control animals receiving saline. Treatment vitro and in vivo. A positive control adenovirus encoding

Gene Therapy Generation of angiostatin from a precursor V Schmitz et al 1604 thelial cells. This also holds true for in vivo antiangiogenic effects since both constructs, AdK1-3 and AdmuthPlgK3, had antiangiogenic potency in an in vivo Matrigel angiogenesis assay, when the vectors were administered systemically. These data indicate that mutant plasminogen can be cleaved into an angiostatin-like molecule that exerts antiangiogenic bioactivity in vitro and in vivo conditions. Furthermore, we have demonstrated that both AdmuthPlgK3 and AdK1-3 showed strong antitumoral effect in a subcutaneous HCC model in a similar manner. We also studied the antitumoral efficacy of these constructs in established LLC in immune-competent C57BL/6 and in immune-deficient athymic mice. We found that both vectors could inhibit tumor growth either in C57BL/6 or athymic mouse models. This antitumor effect was mediated by antiangiogenic mechanisms, since intratumoral microvessel density detected by immunohistochemistry is clearly decreased in animals receiving either AdK1-3 or AdmuthPlgK3. Although others and ourselves found that vector-mediated gene transfer of angiostatin could induce significant inhibition of tumor growth, this type of therapy was not sufficient to induce complete tumor regression, even in a prophylactic set- ting.13–16,23 Recent reports have demonstrated that combination of anti-angiogenic therapy (angiostatin and vasostatin) and immunotherapy (interleukin-12) had synergistic effect on tumor regression in different mouse tumor models.24,25 This synergism can be explained by the fact that the antiangiogenic therapy induces tumor Figure 6 Inhibition of tumor angiogenesis by treatment with AdK1-3 or cell death and releases tumor antigens which potentiate AdmuthPlgK3. The established LLC tumor about 50 mm3 in athymic mice were treated by intravenous injection of recombinant adenoviruses (1010 the immune response elicited by immunotherapy. p.f.u./mouse) or saline as control. Animals were killed on day 7 after treat- Another possible explanation could be that the host’s ment and tumor tissue was removed. Microvessel density was determined immune system might get easier access to tumor cells by anti-von Willebrand factor staining. (a, b) Representative photomicro- when angiostatic compounds compromise the tumor vas- graph showing microvessels in tumors from animals that received saline culature. or control vector AdlacZ. (c, d) Representative photomicrographs showing The attractiveness of our strategy for the generation of reduction of microvessels in tumors from animals treated with AdK1-3 (c) and AdmuthPlgK3 (d). (e) Quantitative analysis of microvessel density angiostatic molecules is based on the fact that by using was made by counting the positive stained cells in 10 high power fields this approach it would be feasible to generate several (HPF, ×400 magnification). Data are given as mean of cell number/HPF angiostatic active kringle regions from multiple mutated and s.e.m., n = 3 animals per group. ∗P Ͻ 0.05 compared with the control plasminogen sequences. By this means synergistic × groups receiving AdlacZ or saline. Original magnification of a–dis 400. and/or additive antiangiogenic effects of the plasminogen kringle derivates can be exploited by the enzymatic release of multiple compounds from a single precursor only angiostatin K1–3 and a negative control vector construct. AdlacZ were also used. Western blot analysis of con- In summary, this study demonstrates that active anti- ditioned culture medium of cells infected with the vector, angiogenic kringle regions of plasminogen can be AdmuthPlgK3, demonstrated that plasminogen was cut released from full length furin-mutated plasminogen into the predicted, correct size cleavage products liberat- (muthPlgK3) by enzymatic cleavage and that adenovirus- ing kringle fragments K1–3 together with the remaining mediated gene transfer of this construct exerts a strong C-terminal fragment of plasminogen. Cleavage efficacy antiangiogenic and antitumor effect. did not reach 100% as seen in the Western blot analysis because part of the secreted plasminogen remained uncleaved, an observation which is consistent with data Materials and methods from studies which exploited furin sites as linker sequences to produce multiple gene products from Animals and cell lines mono-cistronic expression of one fusion gene. Cleavage C57BL/6 or athymic mice, 5 to 8 weeks old, were pur- efficacy was described to range between 50% and 90% chased from Harlan (Barcelona, Spain). During the of total amount of secreted protein.19,21,22 As incomplete experimental period, C57BL/6 mice were housed in precursor cleavage might interfere with functional standard conditions and athymic mice received special activity, we studied different functional effects of care and were kept in pathogen-free single ventilated AdmuthPlgK3 on endothelial cell function in comparison cages. All animal procedures were performed according to AdK1-3 and the control vector AdlacZ. Our data to approved protocols and in accordance with rec- showed that both vectors had effect on the inhibition of ommendations for proper care and use of laboratory proliferation, migration and tube formation of endo- animals.

Gene Therapy Generation of angiostatin from a precursor V Schmitz et al 1605 The human and the murine lung cancer cell lines A549 brane was stained with Diff-Quick (DADE BEHRING, and LLC cells, and 293 cell line (embryonic E1 transfor- Du¨ dingen, Switzerland). Five to eight representative med kidney cell line) were obtained from American Type fields in each well were counted at 100× magnification to Culture Collection (ATCC, Rockville, MD, USA). Cells determine the number of migrated cells. The migration were cultured in Dulbecco’s modified Eagle medium was expressed as percentage of maximal migration (DMEM) supplemented with 10% heat-inactivated fetal (defined as migration with VEGF stimulation without bovine serum (FBS). HUVE (human umbilical vein addition of CM). endothelial) cells were obtained from Cascade Biologics (Portland, OR, USA) and were cultured according to the Tube formation assay supplier’s instructions. Human hepatocellular carcinoma The 24-well plates were coated with 320 ␮l Matrigel. cell line Huh-7 was cultured in DMEM supplemented HUVE cells (25 000 cells/75 ␮l) were dispensed into each with 10% FBS. well and were incubated with 75 ␮l of CM for 30 min. After adding 150 ␮l Medium 200 containing 10% low Construction of recombinant adenoviruses encoding serum growth supplement (LSGS; Cascade Biologics, furin-mutated plasminogen and angiostatin K1–3 Portland, OR, USA), cells were incubated for another The human plasminogen cDNA (kindly donated by Pro- 6 h. Tube formation was quantified by counting fessor DeFord, University of NotreDame) was used for intact tubes in the whole well under microscopy at 40× site-directed mutagenesis according to the manufac- magnification. turer’s protocol (GeneEditor in vitro Site-Directed Mutagenesis System; Promega, Madison, WI, USA). The In vivo testing of antiangiogenic effects (Matrigel mutation oligo 5’-GTTTTATTTGAAAGGCGCCGGC angiogenesis assay) GTCTCTCAGAGTGC-3’ was used to introduce a furin The Matrigel assay was performed with modifications as restriction site in between K3 and K4 of plasminogen, described previously.27 Athymic mice were treated by replacing residues 1069–1072 coding for ProValSerThr by intravenous injection of AdmuthPlgK3, AdK1-3 or ArgArgArgArg. The resulting gene fragment muthPlgK3 AdlacZ at dose of 1010 p.f.u. per animal. Six h later 150 was cloned into adenoviral shuttle vector pMV60 with ␮l of VEGF-embedded Matrigel was injected subcutane- CMV immediate–early promoter upstream and polyA ously into the left and right mid-abdominal region of sequence downstream of the insertion site.26 The angios- mice. After 14 days, mice were killed and Matrigel plugs tatin K1–3 was generated by PCR technique using human were removed and sections were stained by HE. Quanti- plasminogen cDNA plasmid as template with primers 5’- tative analysis was done by counting the total number of GCGGCCGCATGGAACATAAGGAAGTGGTTC-3’ and endothelial-like cells in the Matrigel plugs with Indexed 5’-GCGGCCGCTTAGGAGTCACAGGACGGTATCT-3’ and Squared Micrometer, 21mm (ProSciTech, Thuringowa cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, CA, Old, Australia) under microscope. Results were USA). The sequence was confirmed by DNA sequencing. expressed as the mean (± s.e.m.) cell number per square The angiostatin K1–3 sequence was cloned into pMV60. scale (mm2). Recombinant adenoviruses were constructed as described previously.26 In vivo testing of antitumoral effects In the tumor model of hepatocellular carcinoma (HCC) Western blot in athymic mice, 2 × 106 of Huh-7 were implanted subcut- Ten ␮l of conditioned culture medium (CM) of A549 cells aneously into the right hind flank of athymic mice. After which were infected with AdmuthPlgK3, AdK1-3 or 10 days, mice were treated by intravenous injection of AdlacZ for 70 h was used for electrophoresis in 10% SDS- AdmuthPlgK3, AdK1-3 or AdlacZ at dose of 1010 p.f.u. PAGE. Membranes were hybridized with a polyclonal per animal. In models of Lewis lung carcinoma (LLC) in antibody against plasminogen (Biogenesis, Pools, UK) C57BL/6 and athymic mice, 106 LLC cells were injected and a secondary antibody against rabbit IgG (Sigma, St subcutaneously in the right hind flank of mice. When Louis, MO, USA) and were developed according to the tumor reached 50 mm3 in volume, animals were treated protocol (ECL-Plus, Amersham Pharmacia Biotech, Little by intravenous injection of AdmuthPlgK3, AdK1-3 or Chalfont, UK). AdlacZ at dose of 1010 p.f.u. per animal. Tumor growth was monitored every day for the first week after treat- Migration assay ment and every 2 days in the second week by measuring Inserts (8 ␮m pores, Costar) for 24-well culture plates tumor size using calipers. The tumor volume was calcu- were coated with 100 ␮g/ml of rat tail collagen type I lated by the formula: V = length × width2 × 0.52.16,28 (Becton Dickinson, Bedford, MA, USA). HUVE cells with passage number of 4–6 were resuspended in Medium 200 Assessment of microvessel density containing 1% BSA (Sigma) at a concentration of 25 000 The paraffin sections of tumors which received therapy cells/50 ␮l. Fifty ␮l of cell resuspension were seeded into with adenoviral vectors were treated with proteinase K the upper chamber. The lower chamber was filled with (DAKO, Glostrup, Denmark) for 10 min at room tempera- Medium 200 containing 1% BSA. HUVE cells were pre- ture and endogenous peroxidase activity was quenched ␮ ° incubated with 50 lofCMat37C for 30 min before by H2O2-treatment (DAKO) for 5 min. Sections were then adding vascular endothelial growth factor (VEGF) at a blocked with 5% goat serum in Tris buffer containing final concentration of 5 ng/ml to the lower chamber. 0.1% Tween 20, followed by incubation with rabbit anti- These chambers were incubated for 6 h to allow cells to von Willebrand factor (factor VIII-related antigen, migrate through the collagen-coated pore membranes. DAKO) for 20 min at room temperature. Antibody bind- The non-migrated cells were thoroughly scraped off the ing was detected by sequential incubation of the sections upper surface membrane with cotton swabs. The mem- with LSAB2 system (DAKO). Enzymatic activity was

Gene Therapy Generation of angiostatin from a precursor V Schmitz et al 1606 developed using DAB (DAKO) as chromogenic substrate. oendothelioma tumor growth in vivo. Cancer Res 1997; 57: After counterstaining with Mayer hematoxylin, samples 5277–5280. were mounted and examined by light microscopy. The 13 Tanaka T, Cao Y, Folkman J, Fine HA. Viral vector-targeted anti- microvessel density was assessed according to the angiogenic gene therapy utilizing an angiostatin complementary DNA. Cancer Res 1998; 58: 3362–3369. method described by Weidner et al.29 Areas presenting 14 Griscelli F et al. Angiostatin gene transfer: inhibition of tumor highest vascularization were identified and microvessels × growth in vivo by blockage of endothelial cell proliferation asso- counted in high power fields ( 400 magnification). For ciated with a mitosis arrest. Proc Natl Acad Sci USA 1998; 95: each section 10 fields were evaluated. 6367–6372. 15 Indraccolo S et al. Effects of angiostatin gene transfer on functional properties and in vivo growth of Kaposi’s sarcoma cells. Acknowledgements Cancer Res 2001; 61: 5441–5446. 16 Kuo CJ et al. Comparative evaluation of the antitumor activity This work was supported in part by SAF 98-0146 from of antiangiogenic proteins delivered by gene transfer. Proc Natl CICYT and also by J Vidal, MJ Huarte, Dr Cervera, M Acad Sci USA 2001; 98: 4605–4610. Losantos and M Mendez grants for Gene Therapy. VS 17 Molloy SS, Anderson ED, Jean F, Thomas G. Bi-cycling the furin was partly supported by the German Exchange Council pathway: from TGN localization to pathogen activation and (DAAD) and the VERUM Foundation, Munich, Ger- embryogenesis. Trends Cell Biol 1999; 9:28–35. many. We thank Helena Villanueva and Celia Gomar for 18 Nakayama K. Furin: a mammalian subtilisin/Kex2p-like endop- their expert assistance. rotease involved in processing of a wide variety of precursor proteins. Biochem J 1997; 327: 625–635. 19 Gros L et al. Insulin production by engineered muscle cells. Hum References Gene Ther 1999; 10: 1207–1217. 20 Muzzin P, Eisensmith RC, Copeland KC, Woo SL. Hepatic insu- 1 Folkman J. Angiogenesis in cancer, vascular, rheumatoid and lin gene expression as treatment for type 1 diabetes mellitus in other disease. Nat Med 1995; 1:27–31. rats. Mol Endocrinol 1997; 11: 833–837. 2 Gasparini G. The rationale and future potential of angiogenesis 21 Gaken J et al. Fusagene vectors: a novel strategy for the inhibitors in neoplasia. Drugs 1999; 58:17–38. expression of multiple genes from a single cistron. Gene Therapy 3 Griffioen AW, Molema G. Angiogenesis: potentials for pharmac- 2000; 7: 1979–1985. ologic intervention in the treatment of cancer, cardiovascular 22 Groskreutz DJ, Sliwkowski MX, Gorman CM. Genetically engi- diseases, and chronic inflammation. Pharmacol Rev 2000; 52: neered proinsulin constitutively processed and secreted as 237–268. mature, active insulin. J Biol Chem 1994; 269: 6241–6245. 4 Brower V. Tumor angiogenesis - new drugs on the block. Nat 23 Griscelli F et al. Combined effects of radiotherapy and angiosta- Biotechnol 1999; 17: 963–968. tin gene therapy in glioma tumor model. Proc Natl Acad Sci USA 5O’Reilly MS et al. Angiostatin: a novel angiogenesis inhibitor 2000; 97: 6698–6703. that mediates the suppression of metastases by a Lewis lung 24 Yao L et al. Effective targeting of tumor vasculature by the angi- carcinoma. Cell 1994; 79: 315–328. ogenesis inhibitors vasostatin and interleukin-12. Blood 2000; 96: 6 Dong Z, Kumar R, Yang X, Fidler IJ. Macrophage-derived metal- 1900–1905. loelastase is responsible for the generation of angiostatin in 25 Gyorffy S et al. Combined treatment of a murine breast cancer Lewis lung carcinoma. Cell 1997; 88: 801–810. model with type 5 adenovirus vectors expressing murine 7 Gately S et al. Human prostate carcinoma cells express enzy- angiostatin and IL-12: a role for combined anti-angiogenesis and matic activity that converts human plasminogen to the angiog- immunotherapy. J Immunol 2001; 166: 6212–6217. enesis inhibitor, angiostatin. Cancer Res 1996; 56: 4887–4890. 26 Qian C, Bilbao R, Bruna O, Prieto J. Induction of sensitivity to 8 Cao Y et al. Kringle domains of human angiostatin. Characteriz- ganciclovir in human hepatocellular carcinoma cells by aden- ation of the anti-proliferative activity on endothelial cells. J Biol ovirus-mediated gene transfer of herpes simplex virus thymid- Chem 1996; 271: 29461–29467. ine kinase. Hepatology 1995; 22: 118–123. 9 Cao Y et al. Kringle 5 of plasminogen is a novel inhibitor of 27 Yao L et al. Contribution of natural killer cells to inhibition of endothelial cell growth. J Biol Chem 1997; 272: 22924–22928. angiogenesis by interleukin-12. Blood 1999; 93: 1612–1621. 10 O’Reilly MS, Holmgren L, Chen C, Folkman J. Angiostatin 28 Mishima K et al. A peptide derived from the non-receptor-bind- induces and sustains dormancy of human primary tumors in ing region of urokinase plasminogen activator inhibits glioblas- mice. Nat Med 1996; 2: 689–692. toma growth and angiogenesis in vivo in combination with cis- 11 Sim BK et al. A recombinant human angiostatin protein inhibits platin. Proc Natl Acad Sci USA 2000; 97: 8484–8489. experimental primary and metastatic cancer. Cancer Res 1997; 29 Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogen- 57: 1329–1334. esis and metastasis – correlation in invasive breast carcinoma. 12 Lannutti BJ et al. Human angiostatin inhibits murine hemangi- N Engl J Med 1991; 324:1–8.

Gene Therapy