Hypoxia-inducible factor-1α (HIF-1α) is an important transcriptional factor that is activated when mammalian cells experience hypoxia, a tumor microenvironmental condition that plays pivotal roles in tumor progression and treatment. In this study, we examined the idea of down-regulating HIF-1α in tumor cells for therapeutic gain. We show that the expression levels of HIF-1α can be significantly attenuated by use of the recently established small interfering RNA technology in combination with adenovirus-mediated gene transfer. Down-regulation of the HIF-1α protein enhanced hypoxia-mediated tumor cell apoptosis in vitro. Subcutaneous tumor growth was also prevented from cells with attenuated HIF-1α expression. In addition, intratumoral injection of adenovirus encoding the HIF-1α-targeted small interfering RNA had a small but significant effect on tumor growth when combined with ionizing radiation. Therefore, our results provide proof of HIF-1α as an effective target for anticancer therapy. They also suggest that an adenovirus-based small interfering RNA gene transfer approach may be a potentially effective adjuvant strategy for cancer treatment.
Hypoxia, the reduction in normal oxygen tension, has been implicated in a host of human diseases (1). Hypoxia is prominently involved tumor growth and tumor development (2, 3). Specifically, it plays a critical role in promoting mutagenesis and selecting for malignant tumor growth (4). It has also been involved in promoting tumor angiogenesis (5), the process of neovascular formation that is critical in tumor development.
At the molecular level, regulation of cellular response is largely mediated by the hypoxia-inducible factor-1α (HIF-1α) proteins, which are transcriptional factors that activate a large assortment of downstream genes. The regulation of these α subunits is achieved post-translationally through the ubiquitin-proteosome system. Under normoxic conditions, the α subunits are hydroxylated by the proline hydroxylase (6). The hydroxylated HIF-α proteins are usually bound by the von Hippel-Lindau (VHL) protein (7). This binding targets them for degradation by the ubiquitin-proteosome system (7). Under hypoxic conditions, the proline residues are not hydroxylated, and HIF-α protein levels increase. Among the HIF-α proteins, HIF-1α is the best understood thus far. HIF-1α heterodimerizes with aryl hydrocarbon receptor nuclear translocator and interacts with cofactors such as CBP/p300 and the DNA polymerase II complex to bind to the hypoxia-response elements. This process activates the expression of numerous hypoxia-responsive genes. These genes are related to erythropoiesis, angiogenesis, glycolysis, and apoptosis (8).
Clinically, tumor hypoxia is a well-recognized prognostic factor in radiation therapy and chemotherapy (9). Histologic analyses indicate that increased levels of intracellular HIF-1α are associated with poor prognosis and resistance to therapy in head and neck, ovarian, and esophageal cancer (10). It has also been demonstrated that HIF-1α is overexpressed in breast, lung, skin, colon, ovarian, gastric, pancreatic, prostate, and renal carcinomas, and is associated with tumor progression (11).
Because of the prominent involvement of HIF-1α in cancer development and prognosis, many efforts are under way to target HIF-1α or proteins that interact with it. Recent data indicate that in vivo delivery of antisense to HIF-1α can enhance cancer immunotherapy (12). A gene therapy strategy to block the interaction between HIF-1α and its transcriptional coactivator CBP/p300 led to attenuation of hypoxia-inducible gene expression and inhibition of tumor growth in a mouse xenograft model (13).
In this study, we used the newly established small interfering RNA technology to target the HIF-1α protein in tumor cells. Small interfering RNA are short, double-stranded RNA molecules that can target specific complementary mRNAs for degradation via a cellular process termed RNA interference. This novel strategy for regulating gene expression was discovered initially in Caenorhabditis elegans (14). The small interfering RNA technology has been widely and rapidly adopted in the scientific community after the discovery of its ability to silence genes in mammalian cells (15). We investigated whether the HIF-1α protein can be targeted for therapeutic purposes via adenovirus delivered small interfering RNAs.
Materials and Methods
We have used the following cell lines in this study: (1) HEK 293, an adenovirus E1 gene transduced human embryonic kidney cell line for adenovirus packaging and expansion, (2) HeLa, a human cervical adenocarcinoma cell line obtained from American Type Culture Collection (Manassas, Virginia), and (3) HCT116, a human colon cancer cell line obtained from the Tissue Culture Facility at Duke University Medical Center. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen Inc., Carlsbad, CA) with 10% fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin at 37°C, 5%CO2.
Designing Small Interfering RNA-Encoding Minigenes Targeted at HIF-1α.
To design the small interfering RNA-encoding minigenes, we used a Internet-based program available at the website of Ambion Inc. (Austin, TX). Oligonucleotide DNA sequences based on these targeting sequences were then synthesized by commercial sources. These oligos contain two 19-mer complimentary targeting sequences with a loop sequence separating them and a polythymidine tract to terminate transcription. In addition, they were engineered to possess BamHI- and HindIII-compatible overhangs that facilitate their ligation into the expression vector pSilencer-2.0 (Ambion Inc.), which is a plasmid with a human U6 gene-based RNA polymerase III promoter. The derived HIF-1α–targeted minigene-encoding plasmid was pSilencer-siHIF-1α. The control plasmid was pSilencer-siNT (for nontargeted, obtained from Ambion), which is a plasmid with a similar structure but encoding a nonsense minigene with no homology to any known sequences in the human genome. The sequence of the scrambled minigene is as follows: AAT TCT CCG AAC GTG TCA CGT.
The AdEasy system of adenovirus packaging, including plasmid pAdtrack, pAdeasy-1, and the packaging Escheria coli BJ5183 cells was commercially purchased from Stratagen Corp. (La Jolla, CA) The small interfering RNA-encoding gene expression cassette (with the U6 gene promoter) was then excised by PvuII/HindIII from pSilencer-siHIF-1α and subcloned into the EcoRV/HindIII sites of pAdTrack. The resulting plasmid was pAdTrack-siHIF-1α. Packaging and production of the adenovirus that carries the HIF-1a–targeted small interfering RNA gene was carried out by following the manufacturer’s protocol. Briefly, the pAdtrack-U6-siHIF-1a plasmid was linearized by PmeI and then recombined with pAdeasy-1 plasmid in recA+ bacteria BJ5183. The resultant pAdeasy-siHIF-1α was then transfected into relatively low passage (passage no. <30) 293 cells after linearization by PacI. After 7 to 10 days, we obtained infectious adenovirus vector, AdsiHIF-1α. Large-scale preparation of the particles was carried out subsequently following established protocols.
Quantitative PCR Assessment of HIF-1α mRNA Level.
To measure the level of HIF-1α mRNA in cells that have been infected with the small interfering RNA-encoding adenovirus vectors, quantitative PCR technology was used. Twenty-four hours after infection by adenovirus vectors (AdsiNT or AdsiHIF-1α), total RNA from the infected cells were extracted by use of the Rneasy kit (Qiagen, Valencia, CA). Afterward, cDNA from the mRNA were synthesized by use of the SuperScript first-strand synthesis system for reverse transcription-PCR (Invitrogen). The cDNA was then used as templates in quantitative PCR reactions. The quantitative PCR reactions were carried out by use of the QuantiTech SYBR Green PCR kit (Qiagen) in an ABI PRISM 7900 apparatus. Relative quantification of HIF-1α was done by comparative CT method. The relative amount of target (HIF-1α), normalized to an endogenous sequence, is given by 2−ΔΔ CT.
The primers used for the amplification of β-actin are as follows: (forward) 5′- TCAAGATCATTGCTCCTCCTG-3′ and (reverse) 5′-CTGCTTGCTGATCCACATCTG-3′. The primers used for the amplification of the HIF1α gene are as follows: (forward) 5′-CTGATCATCTGACCAAAACTC-3′ and (reverse) 5′-GTTTCAACCCAGACATATCCAC-3′.
Hypoxia treatment of cells was achieved by incubating cells in a Bactron Anaerobic/Environmental Chamber (Sheldon Manufacturing, Corvallis, OR). During incubation, a humidified environment at 37°C was maintained. In addition, the atmosphere was maintained at 5% CO2 and 0.5% O2.
Western Blot Analysis.
The antibodies to caspase-3, Bcl-XL, and HIF-1α were purchased from Cell Signaling Technology (Beverly, MA), BD Phar-Mingen (Palo Alto, CA), and Santa Cruz Biotechnology (Santa Cruz, CA). HeLa cells were infected by adenovirus and subjected to hypoxia treatment. Treated cells were collected and lysed. About 0.6 μg to 2 μg of total protein was electrophoreses on a 6% SDS-PAGE gel. The proteins were then transferred to a nitrocellulose membrane by use of an electroblotting device. The membranes were then blocked with 5% nonfat milk in PBS +0.1% Tween 20 overnight at 4°C. Afterward they were incubated with the primary antibody for 2 ours, washed with PBS +0.1% Tween 20 three times (15 minutes each time), and then incubated with horseradish peroxidase-conjugated secondary antibody (IgG).
After washing with PBS +0.1% Tween 20 three times, the signal was visualized by use of the ECL kit (Amersham, Arlington Heights, IL).
Hochest33342 Staining for Apoptotic Cells.
HeLa cells were cultured in 12-well plates to 60% to 70% confluence, infected with AdsiHIF-1a or AdsiNC (10 multiplicity of infection) for 24 hours, and then subjected to hypoxia (0.5% O2) for 24 hours. Hypoxia was induced by placing the cells in an anaerobic tissue culture hood (Sheldon Manufacturing) maintained at 37°C. At the end of the exposure the cells were fixed (methanol to acetic acid; 3:1) for 5 minutes at 4°C and washed with sterilized, distilled H2O three times. Subsequently the cells were stained with Hoechst 33342 (5 μg/mL; Calbiochem, La Jolla, CA) for 10 minutes at room temperature. The cells were then washed three times with sterilized, distilled H2O. The fraction of apoptotic nonapoptotic cells was derived by counting under a fluorescence microscope. Four randomly chosen areas were counted and averaged to derive the value of apoptotic cell fraction. Counting was carried out by two independent investigators.
Tumor Growth Delay Studies.
Two series of experiments were conducted. In the first series of experiments, ∼5 × 106 of HeLa or HCT116 cells infected with AdsiHIF-1α or control AdsiNC viruses were transplanted s.c. in 50 μL of PBS in the right hind limbs of Balb/C nude mice 24 hours after virus infection. Each treatment group consisted of 8 to 10 animals. Growth curves are plotted as the mean relative treatment group tumor volume ± SEM. The following formula was used to calculate tumor volume (16): V = (1/2) W2 × L (W, the shortest dimension; L, the longest dimension).
In the second series of experiments, ∼5 × 106 HCT tumor cells were injected s.c. (in 50 μL of PBS) into the hind leg of nude mice. When tumors grew to sizes of 7 to 8 mm in diameter, adenovirus vectors were injected into the tumor mass (1 × 108 plaque-forming units in 30 μL). About 24 hours later, the tumors were irradiated with a 4 MeV linear accelerator (Varian, Palo Alto, CA) at a dose rate of 2 Gy/min. Three doses were given at 6 Gy each. Adenovirus vectors were given 24 hours before each dose. Tumor growth was then followed by daily measurement. Growth curves are plotted as the mean relative treatment group tumor volume ± SEM. Mean times to reaching three times initial tumor volumes (phase of exponential regrowth) for each group were calculated and compared using the Kruskal-Wallis and the two-sided Mann Whitney test (nonparametric).
Results and Discussion
Down-regulation of HIF-1α in HeLa Cells by Adenovirus-delivered Small Interfering RNA.
To target HIF-1α by the small interfering RNA-based approach, we designed several small interfering RNA targeting sequences by use of a web-based program from Ambion (Austin, TX). Each small interfering RNA was synthesized as complimentary oligonucleotides and cloned into pSilencer-2.0 vector following a published approach (17). The resulting constructs were verified by sequencing and screened for their ability to down-regulate HIF-1α expression. The small interfering RNA-encoding vector that was most effective appeared to be one targeted to 5′-ATGACATGAAAGCACAGAT-3′ that corresponds to nucleotides 244 to 262 downstream of the AUG start codon of the HIF-1α gene (GenBank accession no. NM-001530).
To study the efficacy of the HIF-1α–targeted small interfering RNA, we transferred the above small interfering RNA encoding gene expression cassette into an adenovirus vector. This vector was then tested for its capacity to down-regulate HIF-1α in HeLa cells. After AdsiHIF-1a and AdsiNT (a control vector with a scramble small interfering RNA sequence) infection for 24 hours, HeLa cells were subjected to hypoxia (0.5%) for 24 hours. The cells were then harvested, and Western blot analysis was conducted. Results show that HIF-1α was greatly down-regulated in AdsiHIF-1α–infected hypoxic HeLa (>90%) but not in the control cells (Fig. 1). In addition, quantitative PCR was used to measure the level of mRNA in AdsiNT and AdsiHIF-1α–infected HeLa cells (under normoxic condition). The results indicate that mRNA level of HIF-1α was reduced to 90% to 91% of the normal level. In comparison, the mRNA level of the β-actin remained unchanged. Therefore, these results indicated that adenovirus is a useful tool for delivery of the small interfering RNA into tumor cells. This is consistent with several earlier reports (18).
Sensitization of Tumor Cells to Hypoxia-Induced Cell Apoptosis by Adenovirus-delivered HIF-1α–Targeted Small Interfering RNA.
Exposure to hypoxic conditions is known to induce apoptosis in many cells. However, the role of HIF-1α is controversial in this process. Some studies suggested that HIF-1α is a mediator of hypoxia-induced cell death. In support of this viewpoint is the observation that HIF-1α can activate the transcription of many proapoptotic genes such as NIX and NIP3 (19). Other reports suggested that elevated HIF-1α expression could render tumor cells resistant to hypoxic exposure. This is a key question, because there are efforts aimed at developing HIF-1α inhibitors for cancer therapy.
We evaluated the effects of HIF-1α down-regulation on HeLa cells. We infected HeLa cells with AdsiHIF-1α. This has been shown to cause >90% down-regulation of HIF-1α at the mRNA and protein level (Fig. 1). After infecting HeLa cells with AdsiHIF-1a or AdsiNT vectors for 24 hours, the cells were exposed to hypoxic conditions (0.5% O2) for 24 hours. Hoechst 33342 nuclear staining was then used to quantify apoptosis in HeLa cells. Apoptotic cells were typically identified as those cells that possess significantly smaller, condensed, and fragmented nuclei under a fluorescence microscope (Fig. 2 A). AdsiHIF-1α (10 multiplicities of infection) –infected HeLa cells showed that 87.3 ± 9.7% cells were undergoing apoptotic cell death. This is compared with a 12.7 ± 4.3% death rate in the control virus-infected cells. The difference is statistically significant (P < 0.01). Under normoxic conditions, negligible cell death was observed in either AdsiHIF-1a or control virus-infected HeLa cells. These results indicated to us that the down-regulation of HIF-1α can significantly enhance apoptosis in HeLa cells exposed to hypoxic conditions.
To determine the molecular mechanism underlying hypoxia-induced apoptotic cell death in HeLa cells, we analyzed the levels of two proteins known to be involved in cellular apoptosis. These proteins include cleaved caspase-3, an effector of apoptosis, and Bcl-XL, a negative regulator of apoptosis. In AdsiHIF-1α–infected HeLa cells, hypoxia treatment caused a significant increase in the cleavage of caspase-3, indicating its activation. At the same time, hypoxia significantly reduced the expression of Bcl-XL (Fig. 2 B), a cellular survival factor. Therefore, our results provided strong evidence that inhibition of HIF-1α gene expression levels sensitized tumor to hypoxia via activation of cellular apoptotic pathways.
Antitumor Effect of Silencing HIF-1α Expression.
We subsequently evaluated the effects of HIF-1α down-regulation on tumor growth. To achieve efficient suppression of HIF-1α, cells from the cervical cancer cell line HeLa and colon cancer cell line HCT116 were preinfected with AdsiHIF-1α and AdsiNT (at multiplicity of infection of 10) for 24 hours and were implanted s.c. into nude mice. Tumor growth was subsequently measured. As can be seen in Fig. 3 (top panels), in both HeLa and HCT116 cells, cells infected with AdsiHIF-1α grew significantly slower than cells infected with control AdsiNT. In addition, immunohistochemistry analysis of resected tumors indicated that AdsiHIF-1α indeed suppressed the expression of the HIF-1α significantly (Fig. 3, bottom panel). Therefore, our results indicated that down-regulation of HIF-1α significantly inhibits tumor cell growth in vivo as well as in vitro under hypoxic conditions. These data also indicate that HIF-1α is important in the initial phase of tumor growth. Our results are consistent with previous studies (13).
To examine the efficacy of AdsiHIF-1α as a gene therapy vector, experiments with established HCT116 tumors were conducted. AdsiNT and AdsiHIF-1α were injected into HCT116 tumors that were established in Balb/C nude mice. The mice were then irradiated with 3 × 6 Gy (see Materials and Methods for details), with each irradiation preceded by an adenovirus vector injection 24 hours earlier. Fig. 4 shows the results. In contrast to the preinfection experiments shown in Fig. 3, injection of AdsiHIF-1α itself had no effect on tumor growth when compared with injection of AdsiNT. However, when injection of AdsiHIF1-α was combined with irradiation, a smaller but significant growth delay (∼6 days) was observed. This difference, although small, is nonetheless significant (P < 0.05).
The exact mechanisms for the lack of effect on tumor growth in established tumors versus the significant effect on tumor growth from preinfected cells are not clear. It is likely to be because only a small fraction (5–15%, data not shown) of the whole tumor mass was infected after injection of the adenovirus vectors.
Taken together, our study suggests that small interfering RNA-mediated down-regulation of the expression of the HIF-1α gene can effectively sensitize tumor cells to hypoxic conditions. It can also significantly slow down early tumor growth if the majority of HIF1-α in most tumor cells is suppressed. Therefore, we provide additional proof that HIF-1α is a prime target for anticancer therapeutics development. Our results also indicated that adenovirus-mediated delivery of small interfering RNA is can be a potentially effective approach to silence gene expression in tumor cells for gene therapy or gene function studies if a more efficient approaches for the delivery of adenovirus vectors to the tumors can be adopted. In this respect, conditionally replicative viruses that selectively replicate in tumor cells (20) may provide an attractive delivery vehicle.
Grant support: Grant CA81512 from the National Cancer Institute (C-Y. Li). X. Zhang was a W. Osborn Lee Fellow at the Duke University Comprehensive Cancer Center.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Chuan-Yuan Li, Department of Radiation Oncology, Duke University Medical Center, Durham, NC 27710. Phone: 919-681-4897; Fax: 919-684-8718; E-mail: email@example.com