Adeno-associated virus-mediated expression of apolipoprotein (a) kringles suppresses hepatocellular carcinoma growth in mice

Authors

  • Kyuhyun Lee,

    1. Gene Therapy Laboratory, Biomolecular Engineering Division, MOGAM Biotechnology Research Institute, Yongin, Kyonggi-Do, Republic of Korea
    Search for more papers by this author
  • Sung-Tae Yun,

    1. Gene Therapy Laboratory, Biomolecular Engineering Division, MOGAM Biotechnology Research Institute, Yongin, Kyonggi-Do, Republic of Korea
    Search for more papers by this author
  • Young-Gun Kim,

    1. Gene Therapy Laboratory, Biomolecular Engineering Division, MOGAM Biotechnology Research Institute, Yongin, Kyonggi-Do, Republic of Korea
    Search for more papers by this author
  • Yeup Yoon,

    1. New Biologics Division, MOGAM Biotechnology Research Institute, Yongin, Kyonggi-Do, Republic of Korea
    Search for more papers by this author
  • Eui-Cheol Jo

    Corresponding author
    1. Gene Therapy Laboratory, Biomolecular Engineering Division, MOGAM Biotechnology Research Institute, Yongin, Kyonggi-Do, Republic of Korea
    • Biomolecular Engineering Division, MOGAM Biotechnology Research Institute, 341 Pojung-Dong Kiheung-Ku, Yongin, Kyonggi-Do 446-799, Republic of Korea
    Search for more papers by this author
    • fax: (82) 31 260 9808


  • Presented in part in abstract form at the 8th annual meeting of the American Society for Gene Therapy, St. Louis, MO, 2005.

  • Potential conflict of interest: Nothing to report.

Abstract

Hepatocellular carcinoma (HCC) constitutes more than 90% of all primary liver cancers. HCC is a hypervascular tumor that develops from dedifferentiation of small avascular HCC and is therefore a good target for anti-angiogenic gene therapy. Recent studies have identified apolipoprotein(a) [apo(a)] kringles LK68 and LK8 (LKs) as having a potential anti-angiogenic and anti-tumor activity, and the current study evaluates the therapeutic potential of gene therapy with recombinant adeno-associated virus carrying genes encoding LKs (rAAV-LK) in the treatment of hypervascular HCC. We generated rAAV-LK to obtain persistent transgene expression in vivo, which is essential for anti-angiogenic therapy. The rAAV-produced LKs substantially inhibited proliferation and migration of human umbilical vein endothelial cells (HUVECs) in vitro, validating their anti-angiogenic potential. Intramuscular administration of rAAV-LK gave 60% to 84% suppression (P < .05) of tumor growth in mice bearing subcutaneously transplanted HCC derived from Huh-7 and Hep3B cells, respectively. Histological and immunohistochemical analyses of HCC tumor sections showed that a single administration of rAAV-LK gave rise to persistent expression of LKs that inhibited tumor angiogenesis and triggered tumor apoptosis, and, thus, significantly suppressed tumor growth. The administration of rAAV-LK provided a significant survival benefit (P < .05), and 3 of 10 rAAV-LK–treated mice were still alive without visible tumors and without clinical symptoms 188 days after treatment. In conclusion, rAAV-LK is a potential candidate for anti-angiogenic gene therapy in the treatment of HCC. (HEPATOLOGY 2006;43:1063–1073.)

Hepatocellular carcinoma (HCC) is a hypervascular tumor that progresses from a small well-differentiated avascular HCC to a larger poorly differentiated hypervascular HCC as the tumor undergoes dedifferentiation.1 An avascular tumor rarely grows larger than 2 to 3 mm3. However, once vascularized, tumors progressively grow and disseminate to distant sites.2–4 Angiogenesis is therefore a prime target for anticancer drug development.2–4 More than 20 endogenous anti-angiogenic inhibitors have been characterized, and evidently, many of these factors are fragments of larger precursor molecules that contain kringle domains.5 Apo(a) is a glycoprotein composed of multiple kringle domains and a protease-like domain homologous to plasminogen.6 Ten types of kringles homologous to plasminogen kringle 4, designated KIV-1 to KIV-10, and one kringle homologous to plasminogen kringle 5, KV, are found in apo(a). Each KIV domain is present as a single copy, with the exception of KIV-2, which varies from 3 to 42 copies in different apo(a) alleles.7 Lipoprotein(a), a highly atherogenic lipoprotein that is correlated with an increased risk of cardiovascular diseases and stroke, is composed of apo(a) disulfide linked to apo B-100.8 Earlier studies showed an anti-angiogenic activity of apo(a), suggesting a possible role for apo(a) in bridging atherogenesis and angiogenesis in vivo.8, 9

Recently, Kim et al.10 have demonstrated that apo(a) kringle LK68, consisting of KIV-9-KIV-10-KV, inhibits the proliferation and migration of endothelial cells in vitro and suppresses angiogenesis-dependent growth of human colon (HCT-15) and lung (A549) tumors in vivo when administered systemically. LK68-expressing murine colorectal cancer cells (CT26) restricted hepatic metastases to smaller sizes, prevented peritoneal dissemination, and significantly suppressed colon tumor growth in different experimental models.11, 12 These investigators also demonstrated that apo(a) kringle KV, also referred to as LK8, inhibits the migration of HUVECs in a dose-dependent manner, and inhibits new capillary formation in vivo.13

However, anti-angiogenic cancer therapy requires long-term administration of relatively high concentrations of anti-angiogenic proteins to ensure tumor growth suppression in vivo.14, 15 The relatively short half-life of these proteins leads to oscillations in protein levels in vivo and brings into question whether prolonged therapeutic levels could be sustained in the tumor mass. Together with estimates of the relatively high therapeutic doses required, there may be significant economic constraints on widespread clinical use of these agents.16, 17 Gene therapy approaches that directly deliver therapeutic genes in vivo could be an ideal solution for this situation. A vector derived from the adeno-associated virus (AAV) is well known for mediating long-term gene expression, is not pathogenic in humans, and can transduce a wide spectrum of cells from multiple tissue origins.18 AAV-mediated anti-angiogenic gene therapy therefore constitutes an attractive solution to overcome the constraints found by previous clinical evaluations of anti-angiogenic proteins.

The current study was performed to evaluate the therapeutic potential of gene therapy with recombinant AAV carrying LK8 and LK68 genes (rAAV-LK) in the treatment of hypervascular HCC. Our data show that a single administration of rAAV-LK allows persistent expression of LKs in an experimental animal model of HCC. This treatment significantly suppressed HCC by inhibiting tumor angiogenesis and inducing tumor apoptosis. As a consequence, host viability was significantly improved.

Abbreviations

HCC, hepatocellular carcinoma; AAV, adeno-associated virus; rAAV-LK, recombinant AAV carrying genes encoding LKs; PBS, phosphate-buffered saline; PCNA, proliferating cell nuclear antigen; HE, hematoxylin-eosin; MVDmean, mean microvessel density; TUNEL, terminal deoxynucleotidyl transferase nick end labeling; VEGF, vascular endothelial growth factor; bFGF, basic fibroblast growth factor.

Materials and Methods

Plasmids and rAAV Production.

LK68 or LK8 cDNA was amplified by polymerase chain reaction using pcDNA-LK68 or -LK810, 13 as a template with the forward primer 5′-GgaattcAAGCTGGCTAGCCA-3′ and the reverse primer 5′-ATctcgagTTActtgtcatcgtcgtccttgtagtcAGAGGATGCACAGAG-3′. Lowercase bases represent overhangs that include sequences complementary to cloning recognition site for EcoRI and XhoI restriction endonucleases and to a FLAG tag. The polymerase chain reaction products were cloned into linearized pAAV plasmids (derived from AAV serotype 2), to generate the pAAV-LK68 or -LK8 vectors. pAAV-GFP and -LacZ were obtained from Stratagene (La Jolla, CA), and pDG was generously provided by Professor Kleinschmidt (DKFZ, Heidelberg, Germany).

rAAV was produced by the dual transfection method employing one of the vector plasmids (pAAV-LK68, -LK8, -GFP, or -LacZ) and the helper plasmid pDG following a procedure described in detail elsewhere.19–22 Infectious virus particles (IP) of rAAV were quantified using the replication center assay method described in detail elswhere.23

Identification of LKs.

For Western blot analysis, all samples were resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis on a 15% Tris-glycine gel. Rabbit anti-FLAG polyclonal antibody (Sigma, St. Louis, MO) was used as primary antibody (dilution 1:3,000). An anti-rabbit horseradish peroxidase–conjugated antibody (KPL, Gaithersburg, MD) was used to detect FLAG-positive bands. Proliferation inhibition and wound migration assays were performed as previously described in detail elswhere.11, 24

Animals.

Five-week-old male BALB/c nu/nu mice (Charles River, Yokohama, Japan) were housed in a sterile environment. Animals were maintained under specific pathogen-free conditions in the Animal Laboratory Unit of the MOGAM Biotechnology Research Institute. All mice were fed a commercial diet, given water ad libitum, and subjected to a 12-hour light/12-hour dark cycle. All the mice were cared for according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences.

In Vivo Expression.

Ten BALB/c nu/nu mice (Charles River) were injected with 1 × 109 IP rAAV-LK68, or -LK8 in the anterior muscle of the hind leg (n = 5/group). After the injection, blood was taken from ocular venous plexus at regular intervals for 4 weeks. Western blot analysis was carried out by the procedure described previously. For quantification of LK proteins, a sandwich ELISA protocol was established using two in-house antibodies; a rabbit anti-LK8 antibody for coating (0.25 μg/well), a rat anti-LK8 antibody (dilution 1:1,000) as primary antibody. Serially diluted plasma or calibrators (purified LK8 or LK68 proteins) were loaded, and LKs bound to the plate were detected with an anti-rat horseradish peroxidase–conjugated antibody (KPL).

Subcutaneous HCC Tumor Xenograft Models Derived From Huh-7 and Hep3B Cells.

1 × 107 Huh-7 cells (JCRB, Osaka, Japan) or Hep3B cells (HB-8064, ATCC, Manassas, VA) were subcutaneously injected into the flank of the mice (30 mice for Huh-7, and 40 mice for Hep3B). The tumor size was measured every 2 to 3 days to determine the tumor volume using the following formula: tumor volume = πab2/6 (a, long diameter; b, short diameter). When the resulting tumors reached a volume of 50 to 100 mm3, the mice were randomized into four groups and injected with 1 × 109 IP of rAAV-LK68 (n = 7 for Huh-7, n = 8 for Hep3B), rAAV-LK8 (n = 7 for Huh-7, n = 8 for Hep3B), rAAV-LacZ (n = 7 for Huh-7, n = 10 for Hep3B), or with phosphate-buffered saline (PBS) (n = 6 for Huh-7, n = 10 for Hep3B) in the anterior muscle of the hind leg. After the injection, blood was taken from the ocular venous plexus to examine expression of LKs in the blood. In an independent survival study, 50 mice bearing tumors of Huh-7 cells were randomized into five groups and injected with 1 × 109 IP rAAV-LK68 (n = 10), rAAV-LK8 (n = 10), or rAAV–LacZ (n = 8), rAAV–GFP (n = 8), or with PBS (n = 8) in the anterior muscle of the hind leg. Moribund mice were killed when experimental end points (tumor size exceeding 2 cm) were reached.

Histology and Immunohistochemistry.

Tumor specimens were dissected from mice and fixed in 10% buffered formalin solution or in a zinc fixative (BD Biosciences, San Jose, CA) overnight. Specimens were embedded in paraffin, and 4-μm sections were cut. The paraffin sections were de-paraffinized with xylene and stained with hematoxylin-eosin (HE) for histopathological analysis or treated with monoclonal antibody against proliferating cell nuclear antigen (PCNA; DAKO, Glostrup, Denmark). The relative area occupied by live HCC cells in HE-stained tumor sections were quantified from microscopic images using SigmaScan® Pro 5.0 software (Systat Software, Point Richmond, CA), and the mean areas per microscopic field were determined. The cell proliferative index was measured by counting cells stained with anti-PCNA antibody and expressed as a percentage relative to the total number of cells. Five random fields at 200× magnification were captured for each tumor.

For CD 31 immunocytochemistry, zinc-fixed paraffin sections were incubated overnight at 4°C with 0.5 mg/mL rat monoclonal anti-mouse CD31 (Pharmingen, San Diego, CA). A biotin-conjugated anti-rat antibody (eBioscience, San Diego, CA) was used to detect anti-CD31 and visualized with streptavidin-biotin using a Histostain SP kit (Zymed Laboratories, San Francisco, CA). Slides were then counterstained with Mayer's hematoxylin (Fisher Scientific, Fair Lawn, NJ), dehydrated in alcohol, cleared in xylene, and mounted. CD31-positive microvessels within or surrounding the HCC were counted in five random fields at 400× magnification for each of five tissue sections per group. Mean microvessel density (MVDmean) in all the fields examined was then recorded and expressed as counts/field.

TUNEL staining was performed using a fluorescein-labeled in situ cell death detection kit (Roche, Mannheim, Germany). The formalin-fixed paraffin sections were deparaffinized and digested with proteinase K, and incubated at 37°C for 1 hour after the addition of the TUNEL reaction mixture. Samples were embedded with antifade and analyzed under a fluorescence microscope (Olympus, Tokyo, Japan). An apoptosis index was calculated as the percentage of TUNEL-positive HCC cells with respect to the total number of HCC cells evaluated, using five random fields for each of the five tissue sections per group. Approximately 2,000 non-necrotic HCC cells were evaluated for each tissue section at 200× magnification.

Statistical Analysis.

Results were expressed as mean ± standard deviation, and a Student t test was used to evaluate statistical significance. P < .05 was considered statistically significant. For the survival study, log rank tests of Kaplan-Meier survival curves were performed to examine the significance of the treatment effect.

Results

rAAV-Mediated Expression of LKs In Vitro.

The pAAV-LK68 and -LK8 vectors were transcriptionally active as determined by transient transfection of HEK 293 cells, and generated rAAV-LK68 or rAAV-LK8 that induced expression of LK68 or LK8 (Fig. 1A) in HEK 293 cells (Fig. 1B-C). LK68 migrated as a single approximately 55-kd band and LK8 as a single approximately 15-kd band on sodium dodecyl sulfate polyacrylamide gel electrophoresis under reducing conditions. The molecular mass of LK68 produced in HEK 293 cells was almost identical to that of LK68 from CHO cells and from murine CT26 cells.11, 15

Figure 1.

AAV-mediated expression of LK68 and LK8 in vitro. (A) A schematic picture of apolipoprotein (a) and its kringle domains (LK8 and LK68). (B) Transfection of HEK 293 cells (ATCC, CRL-1573) with pAAV-LK8, -LK68, or -LacZ. The plasmid vectors were transfected into HEK 293 cells. After a 24-hour incubation, the cell pellet and the supernatant were harvested by centrifugation, and samples were prepared for Western blot analysis as described in Materials and Methods. Lanes 1-3: cell pellet, lanes 4-6: supernatant. Lanes 1 and 4: pAAV-LK8-transfected samples, lanes 2 and 5: pAAV-LK68–transfected samples, lanes 3 and 6: pAAV-LacZ–transfected samples. (C) Infection of HEK 293 cells with rAAV-LK68, -LK8, or -LacZ. The vector plasmids (pAAV-LK8, -LK68, or -LacZ) and a helper plasmid pDG were co-transfected into HEK 293 cells to prepare rAAV, which was then used to infect HEK 293 cells to activate expression of LK68 and LK8 proteins. LK68 and LK8 expression was detected by Western blot analysis using a rabbit anti-FLAG antibody (Sigma, St Louis, MO). Lanes 1-3: cell pellet, lanes 4-6: supernatant. Lanes 1 and 4: rAAV-LK8–infected samples, lanes 2 and 5: rAAV-LK68–infected samples, lanes 3 and 6: rAAV-LacZ–infected samples. rAAV, recombinant adeno-associated virus carrying genes encoding LKs.

Inhibition of Endothelial Cell Proliferation and Migration In Vitro by LKs.

We first examined the ability of rAAV-produced LKs to inhibit proliferation of HUVECs. HUVECs pulsed with rAAV-produced LKs showed a 70% to 90% decline in cell proliferation from day 4 compared with the negative controls (Fig. 2A) and showed an approximately 70% to 90% reduction in their migration, compared with controls (P < .0000003 and P < .00007 versus the positive control; Fig. 2B-C). Cell migration recovered to control levels after pretreatment with an anti-LK8 antibody, which interacts with both LK68 and LK8 (P < .0000003 and P < .00007; Fig. 2C), indicating that rAAV-produced LK68 and LK8 are responsible for the observed anti-migratory effects.

Figure 2.

rAAV-LK68 and -LK8 mediate inhibition of HUVEC proliferation and migration. (A) Proliferation inhibition assay. HUVECs (ATCC, CRL 1730) were pulsed every 2 days with culture supernatant of HEK 293 cell culture infected with rAAV-LK68 (white circle), -LK8 (black circle), and -LacZ (black triangle) at a multiplicity of infection (MOI) of 5. A mock control culture (white triangle) was pulsed with fresh culture medium. The analysis was carried out in triplicate, and the corresponding results are presented as means ± SD. *P < .02 for rAAV-LK68 versus both negative control groups (the rAAV-LacZ–treated group and the mock HUVEC control group) **P < .02 for rAAV-LK8 versus both negative control groups. (B) Representative images of migrating cells in a wound migration assay. Original magnifications 100×. The area between the dotted lines indicates the area occupied by the initial wound. Migration by HUVECs in EBM-2 medium (Clonetics, San Diego, CA) containing bFGF (3 ng/mL; Roche, Mannheim, Germany) was used as a positive control. (C) Cells that migrated into the denuded area were counted and expressed as percentage of cells that migrated in the positive control. The number of cells migrating in EBM-2 medium containing bFGF (the positive control) was considered 100%, and migration of cells in response to EBM-2 medium (without bFGF) was considered the basal level (0%). The analysis was carried out in triplicate, and the corresponding results were presented as means ± SD. *P < .00007 vs. the positive control and **P < .0000003 vs. HUVECs treated with rAAV-LK68 or -LK8 in the presence of anti-LK8 antibody. rAAV-LK, recombinant adeno-associated virus carrying genes encoding LKs; bFGF, basic fibroblast growth factor.

In Vivo Expression of LK68 and LK8 Genes in BALB/c Nude Mice.

The mice receiving 1 × 109 IP of rAAV-LK68 or –LK8 started to produce LK68 or LK8 in their blood circulation by 1 week post-administration, and levels gradually increased over time (Fig. 3A), and plateaued at 87 ± 36 ng/mL or 23 ± 12 ng/mL, respectively (Fig. 3B). The calculated molecular mass of LK68-FLAG was 3.8-fold higher than that of LK8-FLAG; therefore, the molar concentrations of these molecules were quite similar.

Figure 3.

Expression profiles of LK8 and LK68 after AAV-mediated transfer of the corresponding genes into BALB/c nude mice. (A) Identification of LK proteins in plasma samples using Western blot analysis. Dpi indicates days post-injection of rAAV into mice. (B) Time-course profile of LK68 (white circle) and LK8 (black circle) titers in plasma samples after rAAV-LK68 or -LK8 treatment. Titers were quantified using a sandwich ELISA (see Methods). Error bars denote standard deviation (n = 5). rAAV-LK, recombinant adeno-associated virus carrying genes encoding LKs.

Suppression of Subcutaneous HCC Tumors Derived From Huh-7 Human HCC.

rAAV-produced LKs act as typical anti-angiogenic molecules in vitro and were expressed for several weeks in rAAV-LK-treated BALB/c nu/nu mice. We therefore evaluated the anti-tumor efficacy of rAAV-LK gene therapy for established HCC tumors derived from Huh-7 cells. Expression levels of LK8 and LK68 on day 12 after rAAV treatment was 25.3 ± 21.2 ng/mL and 97.3 ± 77.8 ng/mL, respectively. The mice were killed on day 12, and their tumors were resected. The tumors in mice receiving rAAV-LK8, and -LK68 remained small compared with the negative control groups (the PBS-treated group or the rAAV-LacZ-treated group). rAAV-LK administration appeared to extend the latent period of the tumors and reduced the rate of tumor growth, compared with the negative control groups (Fig. 4). rAAV-LK68 and -LK8 inhibited tumor growth by 84% and 73%, respectively (P < .05 vs. each negative control group), at day 12 after treatment.

Figure 4.

rAAV-LK68 and -LK8 suppress growth of subcutaneous HCC tumors derived from Huh-7 cells xenografted into BALB/c nude mice. The mice bearing tumors of Huh-7 HCC cells were randomized into four groups, and injected with 1 × 109 IP rAAV–LK8 (black circles, n = 7), -LK68 (white circles, n = 7), -LacZ (black triangle, n = 7), or PBS (white triangle, n = 6) in the anterior muscle of the hind leg. The average time between tumor implantation and vector application was 12 days. The statistical significance was examined by Student t test. *P < .05 and ** P < .05 versus both negative control groups. rAAV-LK, recombinant adeno-associated virus carrying genes encoding LKs; PBS, phosphate-buffered saline.

rAAV-LK Reduced Tumor Angiogenesis of Huh-7–Derived Tumors In Vivo.

HE-stained tumor sections showed that viable Huh-7 cells were substantially reduced in the rAAV-LK68 or -LK8–treated group (16.1% ± 2.4%, P < .04 or 12.8% ± 2.5%, P < .001), compared with the control groups (61.8% ± 11.8% for the PBS-treated group, or 57.8% ± 3.8% for the rAAV-LacZ–treated group) (Fig. 5). Mean microvessel density (MVDmean) within tumor sections of the rAAV-LK68 or -LK8–treated group was reduced (40.9 ± 16.6/field, P < .03 or 33.1 ± 11.3/field, P < .02), compared with the control groups (104.7 ± 17.1/field for the PBS-treated group, and 93.3 ± 29.2/field for the rAAV-LacZ–treated group; Fig. 5). Consistent with these MVDmean profiles, the TUNEL-stained HCC sections showed more apoptotic Huh-7 cells in the rAAV-LK–treated groups, compared with the control groups (Fig. 5). The apoptosis index determined by TUNEL staining was twofold to threefold higher in the rAAV-LK68 or -LK8–treated group (8.43% ± 3.51%, P < .0002 or 4.23% ± 3.20%, P < .03) than in the control groups (1.23% ± 0.48% for the PBS-treated group and 1.72% ± 0.83% for the rAAV-LacZ–treated group), whereas the proliferative index determined by PCNA staining was similar in the rAAV-LK68 or -LK8–treated group (86.8% ± 1.4%, P < .83 or 83.9% ± 6.1%, P < .61) and in the control groups (86.0% ± 1.8% for the PBS-treated group or 87.1% ± 2.2% for the rAAV-LacZ–treated group). These results suggest that rAAV-produced LKs reduced tumor angiogenesis and led to apoptosis of tumor cells, which, consequently, suppressed tumor growth.

Figure 5.

Histology and immunohistochemistry of subcutaneous HCC tumors derived from Huh-7 xenografted into BALB/c nude mice. HE-, anti-PCNA–, anti-CD31-, and TUNEL-stained tumor sections. Magnifications 200× for HE-, PCNA-, and TUNEL-stained sections, and 400× for CD31-stained sections. HCC, hepatocellular carcinoma; HE, hematoxylin-eosin; PCNA, proliferating cell nuclear antigen; TUNEL, terminal deoxynucleotidyl transferase-mediated nick-end labeling.

Suppression of HCC Tumors Derived From Hep3B Human Subcutaneous HCC.

We next explored the use of rAAV-LK as a potential treatment for established subcutaneous HCC tumors derived from Hep3B cells. rAAV-LK administration allowed persistent expression of each transgene during the entire period of experiment (10-25 ng/mL LK68 and 60-90 ng/mL LK8 for 4 weeks). The tumors in mice receiving rAAV-LK remained small compared with the negative control groups (Fig. 6). rAAV-LK68 and -LK8 inhibited tumor growth by 60% to 66% and 68% to 73%, respectively (P < .03 and P < .009 vs. the negative control groups), at day 26 after treatment.

Figure 6.

rAAV-LK68 and -LK8 suppress growth of subcutaneous HCC tumors derived from Hep3B cells xenografted into BALB/c nude mice. The mice bearing tumors of Hep3B HCC cells were randomized into four groups and injected with 1 × 109 IP rAAV–LK8 (black circle, n = 8), -LK68 (white circle, n = 8), -LacZ (black triangle, n = 10), or PBS (white triangle, n = 10) in the anterior muscle of the hind leg. The average time between tumor implantation and vector application was 23 days. The statistical significance was examined by Student t test. *P < .03 and **P < .009 versus both negative control groups. HCC, hepatocellular carcinoma; rAAV-LK, recombinant adeno-associated virus carrying genes encoding LKs; PBS, phosphate-buffered saline.

rAAV-LK–Reduced Tumor Angiogenesis In Vivo in Hep3B-Derived HCC.

The population of viable Hep3B cells in HE-stained tumor sections was greatly reduced in the rAAV-LK68 or -LK8–treated group (13.8% ± 4.5%, P < .004 or 14.3% ± 3.1%, P < .013), compared with the control groups (67.3% ± 2.1% for the PBS-treated group or 56.6% ± 7.4% for the rAAV-LacZ–treated group) (Fig. 7). MVDmean within-tumor sections of the rAAV-LK68 or -LK8–treated group (37.3 ± 27.4/field, P < .04 or 34.3 ± 22.9/field, P < .03) was also reduced, compared with the control groups (153.6 ± 84.1/field for the PBS-treated group, and 147.6 ± 60.7/field for the rAAV-LacZ–treated group; Fig. 7). TUNEL staining of the HCC sections showed more apoptotic Hep3B cells in the rAAV-LK–treated groups than found in the control groups (Fig. 7). The apoptosis index in the rAAV-LK68 or -LK8–treated group (5.83% ± 2.53%, P < .02 or 5.63% ± 2.60%, P < .02) was fourfold to fivefold higher than in the control groups (1.20% ± 0.29% for the PBS-treated group and 1.24% ± 0.36% for the rAAV-LacZ–treated group). The proliferative index determined by PCNA staining was similar in the rAAV-LK68 or -LK8–treated group (81.6% ± 2.1%, P < .51 or 81.7% ± 0.7%, P < .14) and in the control groups (85.9% ± 1.5% for the PBS-treated group or 80.7% ± 0.5% for the rAAV-LacZ–treated group). These results confirmed that rAAV-produced LKs reduced tumor angiogenesis and led to increased apoptosis and reduced tumor growth, irrespective of the HCC tumor origin.

Figure 7.

Histology and immunohistochemistry of subcutaneous HCC tumors derived from Hep3B xenografted into BALB/c nude mice. HE-, anti-PCNA–, anti-CD31-, and TUNEL-stained tumor sections. Magnifications 200× for HE-, PCNA-, and TUNEL-stained sections, and 400× for CD31-stained sections. PCNA, proliferating cell nuclear antigen; TUNEL, terminal deoxynucleotidyl transferase-mediated nick-end labeling.

Improved Survival Rate in Subcutaneous Tumor Model of Huh-7 HCC.

The potent anti-angiogenic activity of rAAV-produced LKs and the antitumor efficacy of rAAV-LK gene therapy suggested that LKs could have a positive effect on long-term host viability. We tested this hypothesis in a subcutaneous Huh-7 HCC tumor xenograft model. The mice in the negative control groups began to die 23 days after treatment and none survived 62 days. In contrast, in the rAAV-LK–treated groups, animals began to die 37 or 39 days post-treatment and survived significantly longer than mice in the negative control groups (P < .05 for rAAV-LK8 or P < .002 for rAAV-LK68; Fig. 8A). Three of 10 mice receiving rAAV-LK68 or rAAV-LK8 were still alive without any visible tumors and without clinical symptoms 188 days after treatment (Fig. 8B-D). These mice developed tumors that began to regress at day 13 post-treatment (Fig. 8C), probably due to apparent expression of LKs around this time (Fig. 2). The tumors became and remained undetectable from day 27 post-treatment until the end of the experiment. The relapse pattern of the tumors was in good agreement with that from others that reported tumor relapse within a few weeks after anti-angiogenic gene treatment, leading to tumor-free survival of the mice for 200 days.25 These results suggest that a single administration of rAAV-LK significantly inhibited the growth of a solid tumor, leading to the increase in host viability.

Figure 8.

rAAV-LK prolonged the survival of mice with subcutaneous HCC tumors derived from Huh-7 cells. (A) The Kaplan-Meier survival curve of the mice. Fifty mice bearing tumors of Huh-7 HCC cells were randomized into five groups, and injected with 1 × 109 IP rAAV-LK68 (white circle, n = 10), rAAV-LK8 (black circle, n = 10), or rAAV–LacZ (black triangle, n = 8), rAAV-GFP (black square, n = 8), or with PBS (white triangle, n = 8) in the anterior muscle of the hind leg. The average time between tumor implantation and vector application was 12 days. The statistical significance was evaluated by log-rank test of the Kaplan-Meier survival curve. *P < .002 and **P < .05 versus the negative control groups. (B) Photographs of the survivors taken at 188 days post-treatment. (C) Tumor growth profile of the survivors treated with rAAV-LK68 or –LK8. (D) HE-stained tissue sections of various organs dissected from each survivor treated with rAAV-LK68 or –LK8 188 days after treatment. Magnifications 400×. rAAV-LK, recombinant adeno-associated virus carrying genes encoding LKs; PBS, phosphate-buffered saline.

Discussion

Solid tumors account for more than 85% of cancer mortality.26 Because cancer cells in these tumors require access to blood vessels for growth and metastasis, inhibiting vessel formation offers hope for reducing the mortality and morbidity from these tumors.3 Both the preclinical and clinical studies reported to date, however, show that tumor regression by anti-angiogenic treatment is slow and only modestly reduces tumor growth, presumably because of the dynamic nature of the microenvironment surrounding tumor vasculature.26 Therefore, gene therapy that can provide persistent expression of native anti-angiogenic molecules in vivo offers an ideal means of anti-angiogenic treatment. Numerous anti-angiogenic gene therapies have been performed for various cancer targets in experimental animal models,27 some agents were restricted to local or ex vivo routes of delivery, and others were provided systemically.

In this study, we carried out a systemic anti-angiogenic therapy employing transgenes encoding LK68 and LK8, to determine their gross effect on tumor suppression, targeting mainly HCC. HCC is a typical hypervascular tumor, distinguished from benign lesions by the characteristic abundant and tortuous blood vessels visible by angiography.3, 28 Starvation of the tumor by cutting off feeding vessels near the HCC has been used as a method to control tumor growth.27 Furthermore, patients with higher microvessel density in tumors have a poor prognosis even after curative resection, indicating that angiogenesis plays an important role in HCC.28 Microvessel density directly reflects the angiogenicity of tumors, and MVDmean determined by CD31 staining could be an important criterion for tumor regression.28 A recent clinical study using the vascular endothelial growth factor (VEGF)-specific antibody bevacizumab showed that the addition of this anti-angiogenic agent to fluorouracil-based combination chemotherapy resulted in statistically significant and clinically meaningful improvement in survival among patients with metastatic colorectal cancer,29 by decreasing tumor perfusion, vascular volume, microvascular density, interstitial fluid pressure, and the number of viable, circulating endothelial and progenitor cells, and also by increasing the fraction of vessels with pericyte coverage.30 In concert with these clinical results, tumor sections from our xenograft models showed high vascularity when stained with a CD31 antibody, and the density was reduced by twofold to threefold on rAAV-mediated gene transfer of LKs (Figs. 5 and 7).

At the protein level, Kim et al.10 and Yu et al.11, 12 demonstrated that the vascularity in LK68-treated tumor tissues was greatly reduced, and the expression of angiogenin, VEGF, and basic fibroblast growth factor (bFGF), which switch on the angiogenic phenotype in tumor tissues, was dramatically decreased at peripheral regions of the tumors. Similarly, LK8 was found to inhibit the migration of HUVECs in a dose-dependent manner in vitro and new capillary formation in vivo.13 Our findings that rAAV-mediated gene transfer gives rise to LK68 and LK8 gene expression both in vitro and in vivo (Figs. 1 and 3), and that the rAAV-produced LK68 and LK8 substantially inhibit HUVEC proliferation and migration in vitro (Fig. 2), is in good agreement with these previous characterizations of the anti-angiogenicity of recombinant LK8 and LK68 proteins.

The molecular mechanism of LK68 and LK8 anti-angiogenicity leading to tumor apoptosis is not fully understood, and the cognate receptors for LK68 and LK8 have not been identified. At the protein level, Ahn et al.31 demonstrated that LK68 selectively inhibited ERK1/2 phosphorylation induced by bFGF, VEGF, or hepatocyte growth factor (HGF), without influencing the phosphorylation of Akt. Although the LK8 sequence is present in LK68, LK8 alone appears to interfere with the formation of focal adhesion complexes and the consequent organization of actin stress fibers, instead of targeting the ERK and PI3-K/Akt pathways to inhibit endothelial cell migration.13 Fragments of plasminogen display differential effects on suppression of endothelial cell growth. KV is most similar to kringle 1 (KI), which shows the most potent inhibition of endothelial cell proliferation, whereas fragments containing KIV alone are comparatively inefficient in suppressing endothelial cell growth.32 These findings appear to explain why anti-angiogenic and anti-tumor effects exhibited by LK68 (KIV-9-KIV-10-KV) and LK8 (KV), respectively, were similar.

Anti-angiogenic therapy is generally recognized as tumorostatic instead of tumoricidal, because it exerts its function directly on the endothelial cells and indirectly on the cancer cells.33 Nevertheless, anti-angiogenic gene therapy not only decreases tumor growth but also shows survival benefit, which is attributed mainly to increased tumor cell apoptosis rather than decreased proliferation.24, 25, 33–37 Recently, Perri et al.37 claimed that plasminogen KV possesses additional anti-macrophage property, and thus, the two-pronged activity of KV promotes significant long-term survival of test animals. Human angiostatin, consisting of plasminogen KI-KII-KIII-KIV, inhibits endothelial cell growth (median effective dose = 135 nmol/L)38 and migration in a specific manner.39 The administration of rAAV-angiostatin alone demonstrated significant survival benefit (40%-60%) in animal models of xenografted human glioma24, 33 and ovarian cancer cells,35 when tumor cells were implanted 3 to 4 weeks after the virus administration. KV alone appears to be even more potent (median effective dose = 50 nmol/L) than angiostatin,32, 40, 41 and the inclusion of the KV domain in the angiostatin molecule gave rise to more than 50-fold increase in its inhibitory effect on endothelial cell proliferation, which then gave more salient anti-tumor effects than angiostatin.41 Furthermore, KIV and KV are presumed to possess the functional elements for the inhibition of endothelial cell migration.39, 40 LKs, consisting of apo(a) KIV and/or KV, therefore, may have additional advantage over angiostatin, which could help induce a more pronounced anti-tumor effect. Further molecular mechanism of the indirect tumoricidal effect of the anti-angiogenic therapy remains to be elucidated in future studies.

One of the limiting factors in successful anti-angiogenic monotherapy may be the already shortened life-expectancy of the patients that are the focus of most clinical studies, those with advanced or end-stage HCC.15, 16, 29 Another critical issue is the discrepancy in directly translating preclinical efficacy to the clinical setting, where tumors have a more established vasculature and are less sensitive to anti-angiogenic therapy.15, 16 Investigators of recent clinical studies have thus suggested undertaking a rational design of clinical studies to achieve an optimal effect or including less severely compromised patients in trials.15, 16 A growing collection of papers suggests that combined administration of anti-angiogenic and cytotoxic therapies or adjuvant anti-angiogenic treatment could yield a maximal benefit, because anti-angiogenic treatment would enhance the delivery and therapeutic efficacy of cytotoxic agents by normalizing the abnormal tumor vasculature.27 Anti-angiogenic gene therapy after curative resection or ablation of HCC, or in combination with chemotherapeutics, may serve as a clinically accessible therapeutic strategy for HCC.

In this study, we were able to show that a single administration of rAAV-LK gave rise to persistent expression of LKs, inhibited tumor angiogenesis, triggered tumor apoptosis (Figs. 5 and 7), and substantially suppressed tumor growth in the experimental animal models of HCC (Figs. 4 and 6), thereby leading to improved host viability (Fig. 8). These results imply that rAAV-LK is a potential candidate for anti-angiogenic gene therapy in the treatment of hypervascular tumors, such as HCC, in clinical settings.

Acknowledgements

The authors thank Dr. Jang-Seong Kim for extensive discussions of this work, Joon-Young Park and Byung-Pil Lim for preparing AAV vectors, Ji-Hyun Lee and Hak-Kyu Joo for their skillful technical assistance in the animal studies, and Dr. Sung-Nam Kim at Green Cross Reference Lab (Yongin, Korea) for her contribution to histological and immunohistochemical analysis of tumor sections. We also thank Jürgen Kleinschmidt (DKFZ, Heidelberg, Germany) for his kind gift of pDG.

Ancillary