Targeting endothelial cells (EC) that line tumor blood vessels forms the basis for metronomic therapy and is a promising new strategy for the treatment of cancer. Genetic and phenotypic differences between tumor-derived and normal ECs indicate that targeting tumor ECs may be therapeutically useful. In the present study, we examined differences in responses to chemotherapy in microvascular EC lines from tumoral (T-EC) and normal (N-EC) mouse liver tissues. The identity of these cells was confirmed by immunocytochemistry for EC markers, such as vascular endothelial-cadherin and CD31 for both types of ECs, and the tumor-endothelial-specific marker tumor endothelial marker-7 for T-EC. The involvement of Akt in NF-κB-dependent angiogenesis was different between N-EC and T-EC. Chemotherapeutic stress increased angiogenesis in T-EC, but not N-EC via an NF-κB-Akt-dependent manner. Both NF-κB and Akt were involved in enhanced survival and migration in T-EC in response to chemotherapeutic stress. Moreover, Akt was involved in NF-κB-dependent VEGF expression and angiogenesis. These studies, showing differences in cellular responses to chemotherapy in tumor-derived ECs, indicate that specific therapies targeting these cells may be therapeutically useful for liver cancers.
- hepatocellular cancer
- transcription factor
hepatocellular carcinoma (HCC) is a common cancer, representing the fifth most common cancer worldwide (5). The incidence of hepatocellular cancer is increasing in the United States, probably related to the epidemics of chronic viral hepatitis (12). Surgical resection or transplantation can be curative, but few patients present with disease that is localized and amenable to these approaches. Vascular invasion or micrometastases often preclude these approaches, and moreover, are associated with a poor prognosis (3, 35). Structural and functional changes in liver tumor microenvironment have been associated with the development and progression of liver cancers (32). Among these, the induction of angiogenesis, a process that broadly refers to the formation of new capillary blood vessels from preexisting vessels, has been shown to be essential for liver tumor growth as well as tumor invasion to distant sites (17, 24). The development of arterial capillaries and arterioles enables and promotes tumor growth. Although targeting the tumor vasculature promises a new strategy for the treatment of HCC, the mechanisms contributing to angiogenesis and tumor growth are inadequately understood.
Recent studies (7, 16) have shown that tumor-derived endothelial cells are genetically and phenotypically different from normal endothelial cells, and express several specific markers of as-yet unknown function. A comparison of gene expression patterns in endothelial cells from normal and malignant colorectal tissues by serial analysis of gene expression revealed several tumor-derived endothelial cell specific transcripts (36). Tumor-derived endothelial cells have been shown to be genetically unstable and undergo frequent chromosomal alterations compared with normal endothelial cells in normal tissues (2, 7). On the basis of these observations, it appears very likely that the cellular behavior and intracellular responses of tumor-derived endothelial cells will differ from those of non-tumor-derived endothelial cells. Understanding the specific mechanisms involved in growth and survival of tumor-derived endothelial cells may provide potential approaches to therapeutically target these cells, but for the most part these are unknown.
Liver cancers frequently arise in the setting of hepatic inflammation, suggesting an important role of inflammatory responses to tumor growth and progression. The transcription factor nuclear factor-κB (NF-κB) is ubiquitously expressed and plays a pivotal role in malignancies associated with chronic inflammation (11, 21, 29). NF-κB is activated by environmental stress and is a critical mediator of cellular responses to chemotherapy (3). Constitutive expression of NF-κB is frequently observed in different types of cancers. Aberrant activation of NF-κB can influence cellular survival and contributes to the refractoriness of many tumors to conventional chemotherapy induced apoptosis (4, 11, 19, 21, 22, 30). NF-κB can regulate the expression of key mediators of endothelial cell survival and angiogenesis, such as PKB/Akt and VEGF (3, 18, 20, 28, 33). Activation of Akt can promote endothelial cell survival by inhibition of apoptosis and can promote cell migration. Both of these can contribute to neovascularization and angiogenesis. Moreover, Akt can regulate the expression of VEGF, which can also act as a survival factor, in addition to enhancing angiogenesis (6, 7).
The role of NF-κB in proliferative and survival responses in hepatic epithelia has been extensively investigated. In contrast, cellular responses to inflammation in tumor-derived endothelial cells remain very poorly understood, and in particular, the role played by NF-κB-dependent processes in endothelial cells and their relevance for liver tumors are unknown. Our preliminary results indicate that tumor-derived microvascular endothelial cells have enhanced Akt activation and decreased expression of phosphatase and tensin homolog. However, the mechanism by which the NF-κB-Akt interaction contributes to survival in tumor-derived endothelial cells is unknown. Despite the established role of NF-κB in vascular homeostasis, the response of tumor-derived vascular endothelial cells to environmental changes resulting in activation of NF-κB, and the mechanisms of regulation of Akt and VEGF expression by NF-κB are unknown. Because of the potential therapeutic importance of these pathways we sought to evaluate and compare the molecular mechanisms and interactions among NF-κB, Akt, and VEGF in microvascular endothelial cells that were associated with liver cancers and those that were not.
MATERIALS AND METHODS
Antibodies, reagents, and plasmids.
Rabbit polyclonal antibodies against VEGF, phospho-IκB-α, and NF-κB (RelA) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA) and against total Akt, Ser(P)473 Akt, and vascular endothelial (VE)-cadherin were from Cell Signaling (Beverly, MA). Monoclonal antibody to tumor endothelial marker-7 (TEM-7) was obtained from Imgenex (San Diego, CA) and to α-tubulin from Sigma (St. Louis, MO). Akt inhibitor I [1L-6-Hydroxymethyl-chiro-inositol 2-(R)-2-O-methyl-3-O-octadecylcarbonate], Akt Inhibitor V (Triciribine), and SN-50 were purchased from Calbiochem (La Jolla, CA). The plasmids encoding NF-κB RelA-FLAG (CMV-NF-κB RelA) and IκB-α (CMV-I-κB) were originally obtained from Origene (Rockville, MD). The active Akt and dominant negative Akt plasmids were purchased from Upstate Biotechnology (Lake Placid, NY).
Induction of hepatic tumors.
Twelve-day-old male BALB/c mice (Charles River Laboratories, Wilmington, MA) were administered 10 mg/kg body wt diethylnitrosamine (DEN) intraperitoneally for 22 wk. The mice were maintained in accordance with The Ohio State University Institutional Animal Care and Use Committee procedures and guidelines. For in vivo Akt inhibitor studies, animals were then randomized to receive daily 0.2-ml ip doses of either 20% DMSO (vehicle) or 1 mg/kg Akt Inhibitor V in 20% DMSO for 2 wk. For other studies, mice were randomized to receive daily 0.2-ml ip doses of either 5 or 50 mg/kg gemcitabine daily for 1 wk. Subsequently, mice in both groups received 50 mg/kg gemcitabine for another week. Some mice were killed before or after the first week of treatment. All other mice were euthanized at 24 wk of age and individual tumors >1 mm in diameter on the surface of the liver were carefully dissected.
Isolation of microvascular endothelial cells.
Liver tumors or nontumoral liver tissues were cut into 1 mm pieces and incubated in 0.1% collagenase I in MEM for 2 h at 37°C. Microvascular endothelial cells were isolated from tumoral (T-EC) or nontumoral (N-EC) tissues using a magnetic bead immunoaffinity endothelial cell isolation kit from Dynal Biotech (Brown Deer, WI). Briefly, Dynabeads M-450 sheep anti-rat IgG coupled with a monoclonal rat anti-mouse CD31 antibody (30-μl aliquot per 5-ml tube) were incubated in 1 ml of tissue supernatant at 4°C overnight and then washed three times with 10% FBS-DMEM; 1 ml of cell suspension was put into the tube containing the washed beads. After 30 min at 4°C with occasional agitation, the bead-bound cells were recovered, washed and then digested in 1 ml of trypsin/EDTA (GIBCO) to release the beads. The bead-free cells were centrifuged in 10% FBS-DMEM and then resuspended in 7 ml of culture media as described below. The yield of vascular endothelial cells was >98%. The identity of the isolated cells was confirmed by in vitro tube formation in matrigel, Dil-Ac-LDL uptake capacity, and staining for VE-cadherin, an exclusive endothelial cell marker. Because of the potential loss of several signaling pathways and accumulation of genetic aberrations with time in culture, isolated endothelial cells were used within three passages.
Isolated endothelial cells were subcultured in the presence of diphtheria toxin (500 ng/ml; Calbiochem), which was added to T-ECs to kill human tumor cells that might overgrow the culture and to N-ECs for consistency of treatment. The isolated endothelial cells were then maintained in gelatin-coated dishes containing 20% FBS in complete DMEM containing 1 mM sodium pyruvate, 2 mM l-glutamine, 15 mM HEPES, 100 IU/ml penicillin, 100 μg/ml streptomycin, 0.25 μg/ml amphotericin B, and 5 U/ml heparin. Cells grown to confluency on a 60-mm culture plate had a protein content of 120–150 μg by Bradford assay (Bio-Rad, Hercules, CA). AML12, normal mouse liver hepatocytes, and Hepa 1–6 mouse liver hepatoma cell lines were obtained from American Type Culture Collection (Rockville, MD), and were maintained in DMEM medium supplemented with 10% fetal bovine serum (Life Technologies, Rockville, MD).
For VE-cadherin and TEM-7 staining, endothelial cells were grown to confluency on gelatin-coated (1.5%) culture chambers. Cell monolayers were washed twice and fixed in 100% ice-cold acetone for 10 min, then blocked in PBS containing 2% rabbit serum and 5% mouse serum for 1 h at room temperature. Cells were then incubated with primary antibody (anti-VE-cadherin or anti-TEM-7) overnight at 4°C and with secondary FITC-conjugated anti-rabbit or Texas red-conjugated anti-mouse IgG for 1 h at room temperature. After three washes, cells were mounted on glass slides with ProLong antifade mounting medium (Molecular Probes). Images were viewed and captured with the use of a motorized fluorescence microscope imaging system (Axiovert 200; Zeiss MicroImaging, Thornwood, NY).
Endothelial cells were plated at a density of 6 × 105 per 60-mm dish in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Transfections were carried out the following day using TransIT-LT1 reagent (Mirus, Madison, WI) according to the manufacturer's protocols. Briefly, 5–6 μg of plasmid DNA was used per dish, and the transfections were carried out for 4–6 h in Opti-MEM (Invitrogen) medium. The transfection medium was then replaced with serum-containing Dulbecco's modified Eagle's medium for 48 h (unless specified otherwise).
After treatment, confluent cell monolayers in 60-mm dishes were washed twice with ice-cold phosphate-buffered saline and lysed by incubation for 20 min in 1 ml of ice-cold cell lysis buffer (1% Nonidet P-40, 50 mM HEPES, pH 7.4, 150 mM NaCl, 2 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 1 mM sodium vanadate, 1 mM sodium fluoride, and 1× protease mixture) and stored at −70°C. Protein concentrations were measured using the Bradford assay. For Western blot analysis of liver cancer, the tumor tissue was homogenized, and lysates obtained. The protein concentrations of the lysates were measured using a Bradford protein assay kit (Bio-Rad). Equivalent amounts of protein were resolved and mixed with 6× SDS-PAGE sample buffer, followed by electrophoresis in a 4–20% linear gradient Tris·HCl-ready gel (Bio-Rad), and then transferred to nitrocellulose membranes. The membranes were blocked with 5% nonfat dry milk in Tris-buffered saline, pH 7.4, with 0.05% Tween 20 and incubated with primary antibodies and IRDye700 and IRDye800-labeled secondary antibodies (Rockland, Gilbertsville, PA). The protein of interest was visualized and quantitated using an infrared imaging system (Odyssey; LI-COR Bioscience, Lincoln, NE).
Gel electrophoresis mobility shift assay.
Ten micrograms of each nuclear extract sample were incubated with 0.1 pmol of [32P]-labeled double-stranded κB-binding oligonucleotide (5′-GCTGGGGACTTTC-3′) or SP1 binding oligonucleotide (5′-ATTCGATCGGGGCGGGGCGAGC-3′) in buffer containing 1 μg of poly(dI·dC), 1 μg of bovine serum albumin, 10 mM HEPES, pH 7.6, 0.5 mM dithiothreitol, 0.1 mM EDTA, 60 mM KCl, 0.2 mM phenylmethylsulfonyl fluoride, 5 mM MgCl2, and 12% glycerol at room temperature for 30 min before 5% native PAGE, followed by autoradiography. For competition and antibody-mediated supershift experiments, antibodies or oligonucleotides were added to the reaction for 15 min at room temperature before the addition of radiolabeled oligonucleotide probe.
In vitro angiogenesis assay.
Analysis of capillary formation was performed using an in vitro angiogenesis kit (Chemicon, Temecula, CA) according to the manufacturer's instructions. Briefly, 50 μl of gel matrix solution were applied into one well of a 96-well plate and incubated for 1 h at 37°C. Cells were then trypsinized and 5 × 103 cells were suspended in 50 μl of the DMEM containing various concentrations of VEGF and plated onto the gel matrix and incubated for 24 h in a 5% CO2-humidified atmosphere at 37°C. The cell three-dimensional organization was examined under an inverted photomicroscope. Each treatment was performed in triplicate wells.
Endothelial cell migration assay.
Cells seeded on 6-well cell culture plates to confluency. A sterile surgical scalpel was used to make a linear scratch. Endothelial cell sprouting from the edge of the injured monolayer was examined and photographed before and at 48 h after being scratched. The extent of endothelial cell migration adjacent to the scratch was assessed in 10 randomly selected high-power fields and expressed as cells/mm2.
Endothelial cell proliferation.
Cells were seeded into 96-well plates (10,000 cells/well), and incubated in a final volume of 200-μl medium. The cell proliferation index was assessed using a commercially available colorimetric assay (CellTiter 96AQueous; Promega, Madison, WI).
VEGF ELISA assay.
Endothelial cells were washed, trypsinized, counted, and lysed by three sequential freeze/thaw cycles and centrifuged, and the supernatant used for quantification of the levels of VEGF. To determine the VEGF levels in tumor explant cultures, mouse HCC tumors were collected and incubated in 200 μl DMEM containing 0.5% FBS for 24 h at 37°C. VEGF expression was assessed in conditioned medium using a mouse VEGF ELISA kit (Calbiochem, San Diego, CA).
NF-κB binding assay.
Nuclear extracts were obtained from liver tumor tissues or endothelial cells and the DNA binding activity of NF-κB (p65) was determined using the NF-κB p65 transcription factor assay kit following the manufacturer's protocol (Chemicon).
Akt kinase assay.
Endothelial cells or liver tumor tissues were placed on ice and extracted with lysis buffer (Cell Signaling Technology) containing phosphatase inhibitor cocktails I and II (Sigma). Lysates were centrifuged for 15 min at 12,000 g, and Akt was immunoprecipitated from 150 μg of cell extract using anti-Akt monoclonal antibody (Cell Signaling, Beverly, MA) and the Seize immunoprecipitation kit (Pierce, Rockford, IL). Kinase assays were then performed using the kinase assay kit (StressGen, Vancouver, BC, Canada).
Analysis of cell survival.
Microvascular endothelial cells (1 × 105) were seeded into 96-well plates in 1% FBS-containing medium. Cells were collected after 72 h of treatment, washed, and then stained with annexin V-FITC (apoptotic death) or propidium iodide (necrotic death) before analysis using a flow cytometer (model BD-LSR; Becton Dickinson, Franklin Lakes, NJ).
Tumor angiogenesis cDNA macroarrays.
Total RNAs from tumor tissue samples were isolated using the Totally RNA isolation kit (Ambion, Austin, TX), and tumor angiogenesis cDNA macroarray was performed using a Human Cancer/Angiogenesis-1 GEArray Kit (SuperArray, Bethesda, MD) according to the manufacturer's directions. Briefly, total RNA was used as a template for biotinylated probe synthesis. GEArray membrane was prehybridized with GEAhyb hybridization solution at 68°C for 2 h before hybridization in hybridization solution containing denatured cDNA probe of the samples at 68°C overnight. After washing process, the membrane was blocked in GEAblocking solution for 40 min and incubated in the same solution containing alkaline phosphatase-conjugated streptavidin (1:5,000 dilution) for another 40 min at room temperature. The membrane was then incubated with chemiluminescent substrate and exposed to X-ray film. Signals were quantified by scanning the film, and the intensity of the spots was analyzed with the use of NIH Image software. β-Actin and glyceraldehyde-3-phosphate dehydrogenase were used as positive controls, and bacterial plasmid (pUC18) was used as a negative control.
Data are expressed as means ± SE from at least three separate experiments performed in triplicate, unless otherwise noted. The differences between groups were analyzed using a double-sided Student's t-test when only two groups were present and the null hypothesis was rejected at the 0.05 level.
Expression of endothelial cell markers in endothelial cells isolated from tumoral or nontumoral liver tissue.
We began by identifying the expression of endothelial cell markers on microvascular endothelial cells isolated from either tumor or nontumoral liver tissue. With the use of Western blot analysis and immunocytochemistry, the endothelial cell marker VE-cadherin was noted to be highly expressed in both tumor-associated microvascular endothelial cells (T-EC) and the non-tumor-associated endothelial cells (N-EC). Expression levels of VE-cadherin were threefold increased compared with expression levels in normal or malignant mouse hepatocytes (Fig. 1A). Similarly, the expression of platelet endothelial cell adhesion molecule (PECAM-1; CD31), a nonexclusive endothelial cell marker was also increased in both T-EC and N-EC. In contrast, expression of tumor endothelial marker-7 (TEM-7) was threefold increased in T-EC compared with N-EC. TEM-7 is a tumor endothelial marker, which is differentially expressed in tumor endothelial cells, and which has been shown to be expressed in the vasculature of a metastatic liver lesion as well as a variety of other primary and metastatic tumors (36). Moreover, by immunocytochemistry, VE-cadherin was intensely expressed in both N-EC and T-EC, whereas TEM-7 was only expressed in T-ECs (Fig. 1B). Although the expression of mTEM-7 mRNA in mice tumor endothelium can be variable (27, 36, 39), TEM-7 is useful in distinguishing between T-EC from N-EC in our model. Interestingly, the staining pattern for VE-cadherin in T-EC was rather different from that observed in N-EC, which may reflect the oncogenic vascular remodeling properties of these cells. These differences between expression of T-EC and N-EC validate our isolation technique, support the presence of phenotypic differences and suggest the possibility of functional differences between these two cell types.
Chemotherapeutic stress increases angiogenesis in tumor-derived ECs via an NF-κB-Akt-dependent manner.
The presence of distinct phenotypic differences in tumor-derived endothelial cells suggests that these cells may also differ in their response to environmental stress signals such as chemotherapeutic stress. Survival of endothelial cells is essential for angiogenesis and the maintenance of vascular integrity. VEGF has been shown to promote endothelial cell survival, an effect that has been correlated with activation of Akt in endothelial cells (13). Indeed, the ability to target proliferating tumor endothelial cells by the use of conventional cytotoxic agents given frequently at low doses forms the basis of “metronomic” therapy that has shown some promising clinical results for treatment refractory cancers.
Exposure to chemotherapeutic drugs can result in activation of the transcription factor NF-κB, a major mediator of environmental stress that is also involved in survival signaling. NF-κB-dependent increased VEGF expression has been shown to stimulate tumor cell survival and vascular angiogenesis (1). In addition, we have shown that NF-κB can activate Akt in different cell types including vascular endothelial cells (25, 26). Both VEGF and Akt activation can enhance endothelial cell survival. Thus, we assessed the involvement of NF-κB in endothelial cell response to chemotherapy, and specifically the role of NF-κB dependent survival signaling involving either Akt or VEGF. The cytidine analogue gemcitabine is an effective chemotherapeutic drug for several gastrointestinal malignancies such as pancreatic cancer. Gemcitabine has been evaluated for use in HCC, with limited results but with manageable toxicity (15, 38). Thus acquired chemoresistance is a likely cause for the observed clinical responses to this agent. To explore the potential involvement of NF-κB-Akt protein kinase pathways during chemotherapeutic stress, we assessed NF-κB, Akt, and VEGF protein expression after exposure of non-tumor- and tumor-derived endothelial cells to 0.5 μM gemcitabine for 72 h (Fig. 2). In addition, NF-κB binding activity was determined by EMSA. Incubation with gemcitabine dramatically increased NF-κB binding activity compared with DMSO controls in both N-EC and T-EC. However, gemcitabine increased Akt only in T-ECs. Inhibition of Akt activity using an Akt inhibitor blocked gemcitabine induced VEGF expression in T-EC (Fig. 2, A and B). Furthermore, chemotherapeutic stress induced by low concentrations of gemcitabine enhanced the angiogenesis index in T-ECs (Fig. 2, C and D). Moreover, treatment with the NF-κB inhibitor SN-50 or Akt inhibitor reduced the angiogenesis response to gemcitabine.
We next examined the time course of expression of NF-κB, Akt, and VEGF as well as the induction of angiogenesis following exposure of tumor-derived endothelial cells to 0.5 μM gemcitabine for up to 72 h (Fig. 3). An increase in NF-κB was noted within 24 h (173 ± 10%), whereas the increase in Akt occurs later, preceding the change in VEGF expression (Fig. 3A). Moreover, an increase in activation of NF-κB and Akt was also detected (Fig. 3B). Changes in angiogenesis become apparent after 36 h (Fig. 3C). The transcriptional inhibitor Actinomycin D and the translational inhibitor cycloheximide both prevented the increase in Akt phosphorylation by gemcitabine, suggesting the need for new gene expression (Fig. 3D). Taken together, these results suggest that NF-κB dependent activation of Akt is involved in VEGF expression and contributes to tumor endothelial cell survival and angiogenesis.
NF-κB and Akt are involved in inducible resistance to gemcitabine and cell migration in tumor-derived endothelial cells.
We next asked whether these aberrant signaling responses in tumor-derived endothelial cells could contribute to inducible chemoresistance as a consequence of enhanced survival signaling pathways. To evaluate whether induction of survival signaling in response to a chemotherapeutic stress could induce subsequent chemoresistance, we preincubated cells with gemcitabine at a low, noncytotoxic concentration of 0.5 μM for 72 h. There was no significant change in proliferation noted in either T-EC, or N-EC at these concentrations. Cells were then exposed to a higher, cytotoxic concentration (10 μM) of gemcitabine for 48 h, which induces endothelial cell apoptosis and decreases cell viability. NF-κB and Akt inhibitors were added concomitantly with the higher concentration of gemcitabine to assess the involvement of NF-κB-Akt pathways. Preincubation with gemcitabine significantly reduced cell death in T-ECs by 16.2 ± 4.8%, but not in N-ECs, an effect which was reduced by concomitant incubation with either the NF-κB inhibitor SN 50 or an Akt inhibitor (Fig. 4A). In addition, we evaluated the effect of T-EC on cell migration, a phenotypic characteristic of endothelial cells which is essential for tumor angiogenesis. Migration was assessed using a scratch wound model. Exposure to low concentrations of gemcitabine enhanced T-EC migration into the cell-free zone. This effect was decreased by co-incubation with SN-50 or Akt inhibitor (Fig. 4B).
Overexpression of Akt decreases chemotherapy-induced apoptosis in tumor-derived endothelial cells.
We next examined the role of Akt activation on the response of T-EC and N-EC to chemotherapy induced apoptosis. Gemcitabine induced apoptosis in a concentration-dependent manner in both N-EC and T-EC. However, overexpression of Akt decreased gemcitabine induced apoptosis in T-EC (Fig. 5, A and B). In contrast, overexpression of Akt did not significantly alter the sensitivity to gemcitabine-induced apoptosis in N-ECs. In addition, overexpression of IκB-α to decrease NF-κB activation enhanced chemosensitivity of both types of cells to gemcitabine. Taken together, these results show that increased Akt expression in T-EC corresponds to increased resistance of these cells to chemotherapy-induced apoptosis, and suggest the role of Akt dependent VEGF expression as a selective survival factor for tumor-derived endothelium. Furthermore, induction of VEGF as well as angiogenesis were detected in T-ECs but not in N-ECs treated for 72 h with gemcitabine at various concentrations (Fig. 5, C and D).
Akt is involved in NF-kB-dependent VEGF expression and angiogenesis in tumor-derived endothelial cells.
Transfection of T-EC or N-EC with NF-κB increased Akt phosphorylation as well as VEGF expression in both cell types (Fig. 6). Thus nonphysiological overexpression of NF-κB in N-EC can modulate both Akt and VEGF. Moreover, cotransfection with dominant negative Akt blocked the NF-κB dependent upregulation of VEGF in T-EC, but not in N-EC. Meanwhile, overexpression of Akt increases VEGF expression in T-EC, but not in N-EC (Fig. 6, A and B). To evaluate the functional effect of these changes, we assessed NF-κB-dependent angiogenesis in vitro (Fig. 6C). NF-κB transfection enhanced angiogenesis in both T-EC and N-EC. However, overexpression of Akt increased NF-κB dependent angiogenesis by 2.6 ± 0.9-fold while cotransfection with DN-Akt blocked NF-κB-dependent angiogenesis only in T-EC. These findings show that Akt is involved in NF-κB-dependent angiogenesis in T-EC and provide a basis for differential NF-κB dependent signaling in tumor-derived endothelial cells.
Inhibition of Akt decreases VEGF expression in vivo.
Akt is believed to be an attractive target for cancer intervention given the findings from several studies that activation of Akt signaling plays a pivotal role in malignant transformation and chemoresistance by inducing cell survival, growth, migration, and angiogenesis. We thus tested the effect of Akt inhibition in vivo. Mice with DEN induced liver cancers were treated with control diluent or Akt Inhibitor V (Triciribine), which has been shown to potently inhibit tumor cell xenograft growth in nude mice (41). Immunoblots of tumor homogenates showed decreased expression of VEGF expression following treatment with the Akt inhibitor (Fig. 7A). Concomitantly, expression of phospho-IκB-α, a quantitative label of NF-κB activation, was not significantly changed. In addition, expression of several transcripts involved in blood vessel formation and growth was assessed. Of 23 genes studied, the transcripts of only 5 genes were above background (Fig. 7B), and after normalization of expression data for differences in the expression of housekeeping genes and a step-down adjustment of statistical results for multiple comparisons, statistically significant differences were observed only for VEGF and VEGF-D (Table 1).
Effect of gemcitabine on VEGF production in vivo.
We next investigated the effect of pretreatment with low concentrations of gemcitabine on VEGF expression in vivo. Activation of NF-κB and Akt was noted in tumors following daily intraperitoneal administration of gemcitabine (5 mg/kg) for 1 wk. These changes persisted during subsequent treatment with gemcitabine (50 mg/kg/day) for one more week (Fig. 8A). Exposure to low concentrations of gemcitabine also increased VEGF production by 51.2 ± 7.6%. These changes also persisted during subsequent treatment with a higher concentration of gemcitabine. However, a significant change in VEGF expression was not noted in tumors exposed to higher concentrations of gemcitabine from the onset (Fig. 8B).
Our studies provide evidence in support of the concept of the tumor-derived endothelial cell as a mediator of therapeutic resistance in liver cancers. Targeting the endothelial cells that line blood vessels therefore promises to be a potentially useful new therapeutic strategy to improve responses to chemotherapy. Such targeted approaches may be useful to improve sensitivity to chemotherapy in resistant tumors such as HCC. Knowledge of cell survival mechanisms that are either activated by or that can modulate the response to chemotherapeutic stress in tumor endothelial cells is a necessary and essential step to develop targeted therapies. The differences in functional intracellular signaling responses to chemotherapy between tumor-derived and non-tumor-derived endothelial cells that we report herein may provide the basis of selective targeting (Fig. 9). These findings emphasize significant differences between liver tumor-derived endothelial cells and non-tumor-associated endothelial cells that justify the need for cell-type specific studies of signaling pathways involved in cellular responses to chemotherapy.
The activation of an NF-κB-dependent pathway promoting Akt and VEGF expression selectively in tumor-derived cells suggests that these cells have acquired changes that enable resistance to environmental perturbations that may modulate cell survival. These are highly relevant to understanding the molecular basis of metronomic therapy, in which low noncytotoxic concentrations of conventional chemotherapeutic agents have been shown to have an anti-tumoral effect. The success of metronomic therapy has been related to vascular injury, and the effectiveness of such strategies is increased by concomitant use of angiogenesis inhibitors. Our findings identify a mechanism of inducible chemoresistance that may be targeted to improve the efficacy of metronomic therapy.
The growth of hepatocellular cancer requires the development of a new arterial blood supply and maintenance of an adequate blood supply. This requires vascular endothelial cell proliferation and migration, resulting in tube and vascular network formation (35). High microvessel density has been associated with rapid progression and metastasis following surgical resection. Both VEGF and the Akt proto-oncogene are overexpressed in experimental and human HCCs, and overexpression has been associated with more aggressive tumor characteristics (10, 37). Akt is a critical regulator of cell survival as well as cell motility. Constitutive activation of Akt signaling is sufficient to block cell death induced by a variety of apoptotic stimuli and transduction of dominant-negative Akt inhibits growth factor-induced cell survival (9, 14, 23, 34, 40). Moreover activation of Akt increases VEGF expression in vitro as well as angiogenesis in vivo in a matrigel plug assay (31). Upregulation of the phosphatidylinositol 3-kinase/Akt/mammalian target of rapamycin pathway is associated with inhibition of thrombospondin-1 synthesis, which contributes to determining the proangiogenic phenotype of tumor-derived endothelial cells (8). Thus Akt provides a rationale target that can have multiple benefits. The response of the tumor microvasculature to chemotherapy can be manipulated to improve tumor control. Modulation of AKT-derived survival signaling in tumor endothelial cells may enhance responses to radiation therapy or to ablative measures that involve targeted tumor cell killing. Akt plays a pivotal role in tumor formation and growth through its effects on endothelial cell survival, migration, and angiogenesis. This central role makes it an attractive target to manipulate tumor growth and angiogenesis.
Identifying the mechanisms by which aberrant responses of tumor-derived endothelial cells to NF-κB activation result in cell survival may be applicable to other pathological states in which angiogenesis occurs in the setting of environmental perturbations such as inflammation and tissue injury. Approaches to target NF-κB or Akt may be a potentially useful in such conditions by limiting NF-κB activation and expression of Akt. NF-κB and Akt play pivotal roles in regulating genes influencing cell differentiation, growth, and inflammation. Constitutive activation of NF-κB is observed in many cancers including human HCCs and contributes to responses to inflammation. Elucidating the molecular mechanism involved in interaction of NF-κB-Akt-VEGF in tumor-derived endothelial cells, especially after chemotherapy treatment, would enhance our understanding of the relationship between tumor angiogenesis and chemoresistance. Strategies targeting the NF-κB/Akt/VEGF pathway may be useful to improve the effects of chemotherapeutic agents such as gemcitabine on tumor-derived endothelial cells. In the long term, such analysis will identify novel molecular targets for the development of therapeutic approaches for treatment-refractory cancers such as HCC.
This study was supported in part by the Scott and White Hospital Foundation and National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-069370.
We thank Drs. Santosh D'Mello, Li Liu, Paul Chin, and Kyle Johnson from the Department of Molecular and Cell Biology, University of Texas at Dallas; Dr. Dong Chen from Eastern Hepatobiliary Surgical Hospital, Shanghai, China, for kind support and assistance with this work. We also thank Drs. Lih Kuo and Yi Ren in the Department of Systems Biology and Translational Medicine, Texas A&M University System Health Science Center for technical assistance with immunofluorescence studies.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society