IκB Kinase-α Regulates Endothelial Cell Motility and Tumor Angiogenesis

Angiogenesis is a complex process requiring the regulation of many proangiogenic and antiangiogenic factors, as well as interplay between endothelial cells and other types of cells in the local environment. The activity of the transcription factor nuclear factor-κB (NF-κB) has long been associated with tumor growth and tumor angiogenesis. It plays important roles in cell survival, chemotaxis, and inflammation (1). Conversely, inhibition of NF-κB activity impairs tumor growth in vitro and in vivo (2, 3). Evidence for the roles of NF-κB in most aspects of tumor development and progression has been well-documented. For this reason, there is a worldwide push to develop NF-κB inhibitors for clinical use.

NF-κB is a ubiquitous, heterodimeric transcription factor that is sequestered in the cytosol by the IκB proteins. Phosphorylation of IκB by the inhibitor of κB kinase family of proteins (IKK) leads to proteosome-mediated degradation, allowing NF-κB nuclear translocation and subsequent activation of transcription. The IKK complex is composed of two enzymatic components, IKKα and IKKβ, and one regulatory component, IKKγ. Most studies on NF-κB activation have been on the role of IKKβ in this process, as IKKβ has been shown to be sufficient for NF-κB activation (4, 5). Until recently, IKKα was considered a redundant protein because both IKKα and IKKβ are similar in structure and function, and IKKβ is sufficient to activate NF-κB (4, 5). However, recent studies indicate the role of IKKα in cell signaling is distinct from that of IKKβ (6, 7). IKKα can signal through an alternative NF-κB activation pathway, whereby IKKα is responsible for the processing of p100 to p52 (8). In addition, IKKα has been shown to enter the nucleus upon activation and phosphorylate the H3 protein of the histone complex (9, 10). IKKα also activates different sets of gene expression from IKKβ (6, 7). Therefore, it is quite likely that IKKα could play a separate role from IKKβ.

In this study, we identify a new function of IKKα in tumor angiogenesis. We showed that expression of IKKα in endothelial cells promoted cell motility and vascular tubule formation in vitro through the canonical NF-κB pathway. We also showed that IKKα is up-regulated in tumor endothelium and promotes tumor angiogenesis and tumor growth in vivo.

Materials. Six- to 7-wk-old female Rag1 mice were purchased from Jackson Laboratories. The animals were housed in pathogen-free units at Vanderbilt University Medical Center, in compliance with Institutional Animal Care and Use Committee regulations. Age- and sex-matched mice were used. The adenoviral vectors directing expression of IKKα, IKKβ, and IκBα were kindly provided by Dr. Fiona E. Yull (Vanderbilt University, Nashville, TN) and Dr. David A. Brenner (University of North Carolina at Chapel Hill, Chapel Hill, NC; ref. 11). Adenoviral vectors directing the expression of green fluorescent protein (GFP) and β-galactocidase were used as vector controls. The viral vectors were propagated in 293 cells and purified by CsCl gradients as described (12). Viral titers were determined by optical densitometry, and recombinant viruses were stored in 10% glycerol at −80°C (12). Human lung tumor tissues and surrounding normal lung tissues were collected from patients under surgery in Vanderbilt-Ingram Cancer Center, with written consent.

Cell culture. Cells were maintained in a humidified incubator with 5% CO₂ at 37°C. Human umbilical vein endothelial cells (HUVEC) and human dermal microvascular endothelial cells (HDVEC) were purchased and cultured on gelatin-coated tissue culture dishes in EGM-2 medium (Cambrex). Primary endothelial cells were used between passage number 3 and 7. Lewis lung adenocarcinoma cells (LLC) were obtained from American Type Culture Collection, and grown on tissue culture plates in DMEM plus 10% fetal bovine serum and 1% antibiotics.

shRNAi. Pooled shRNAs (Sigma) for IKKα (CCGGGCATCATAAGGAGTTGGTGTACTCGAGTACACCAACTCCTTATGATGCTTTT, CCGGCCAGATTATGAAGAAGTTGAACTCGAGTTCAACTTCTTCAT AATCTGGTTTTT, CCGGCCAGCCTCTCAATGTGTTCTACTCGAGTAGAACA CATTGAGAGGCTGGTTTTT, CCGGGCAAATGAGGAACAGGGCAATCTCGAG ATTGCCCTGTTCCTCATTTGCTTTTT, CCGGGCGTGCCATTGATCTATATAA CTCGAGTTATATAGATCAATGGCACGCTTTTT) or scrambled control were transfected into endothelial cells using RNAifect system (Qiagen). Protein expression and subsequent experiments were performed 24 h after transfection.

In vitro angiogenesis assays. The assays were performed as described (13). Briefly, for endothelial cell migration, cells were infected with adenoviral vectors expressing the gene of interest for 36 h and then plated on the upper chamber of Transwells (Costar). The chamber was placed in medium containing 25 ng/mL of recombinant vascular endothelial growth factor (VEGF; R & D). Cells were allowed to migrate for 5 h, followed by fixation in 10% buffered formalin and staining with crystal violet. Cells that had migrated to the underside of the filter were counted in 10 randomly selected ×200 fields.

For endothelial cell tubule formation, cells were plated on top of Matrigel and incubated for 18 h at 37°C. Vascular branch crossings were counted in 10 randomly selected fields under microscopy. Each experiment was performed in triplicate and repeated thrice.

Analysis of tumor growth in vivo. The LLC and endothelial cell cotransplantation experiments were performed as described (14). Briefly, HUVECs were infected with adenoviral vectors directing the expression of either GFP or IKKα. Twenty-four hours after infection, endothelial cells and LLC tumor cells were mixed together in growth factor–reduced Matrigel (BD Bioscience) and injected s.c. into the left hind flank of Rag1 mice. Tumor growth was measured every other day with calipers. Tumor volume was calculated as Length × Width²/2.

Histologic analysis. Immunohistochemical and immunofluorescent analyses were performed as described (15). Tumor tissue was harvested, embedded in optimal cutting temperature (OCT), and freshly frozen. Seven-micrometer sections were cut and stained with antibodies to detect IKKα (Santa Cruz), pan CD31 for total vascular density, human specific CD31 antibody, and murine-specific CD31 antibody (Pharmingen). Apoptosis, cell proliferation, and hypoxia were determined by terminal deoxynucleotidyl-transferase–mediated dUTP nick-end labeling (TUNEL) assay (Chemicon), proliferating cell nuclear antigen (PCNA; Santa Cruz), and Hypoxyprobe-1 (Chemicon), respectively, according to manufacturers' protocols. 4′,6-Diamidino-2-phenylindole (DAPI) staining was used to illuminate nuclei. The number of positive cells was counted in 10 randomly selected ×200 fields under microscopy.

Western blot. Expression of IKKα, IKKβ, and IκB in tissue and cell samples was analyzed by Western blot using specific antibodies from Santa Cruz.

Statistics. Results are reported as mean ± SE for each group. Statistical analyses were performed using ANOVA for multiple group comparison and two-tailed Student's t test for two-group comparison. All tests of significance were two sided, and differences were considered statistically significant at a P value of <0.05.

IKKα expression is increased in tumor endothelium. Although it is clear that the NF-κB pathway plays a critical role in tumor development, most of the focus has been placed on the role of IKKβ. To investigate the potential function of IKKα in tumor angiogenesis, we examined IKKα expression in tumor samples. Murine LLC tumor tissues were processed for immunofluorescent staining with antibodies against IKKα and the vascular marker CD31 (Fig. 1A). Surrounding normal tissues were used as controls. We observed a clear increase of IKKα levels in tumor tissues compared with normal tissues. Interestingly, double staining with CD31 antibody indicates that IKKα is highly expressed in the tumor vasculature (Fig. 1A). Similarly, we observed a significant increase of IKKα in human lung cancer biopsies compared with adjacent normal lung tissues (Fig. 1B). These results suggest a potential role of IKKα in tumor angiogenesis.

IKKα induces angiogenesis in vitro. Angiogenesis is a multistep process that includes endothelial cell migration, proliferation, and tubule formation. Therefore, we examined the function of IKKα in each of these areas. We used adenoviral vectors for efficient gene delivery in endothelial cells, as described previously (16). Micro and large endothelial cells (HDVECs or HUVECs) were infected with adenoviral vectors directing the expression of IKKα, IKBα, IKKβ, or IKKα plus IKBα, for 36 hours. A viral vector expressing GFP was used as a vector control, as well as for evaluation of transgene transduction efficiency. We achieved >85% transduction efficiency, determined by counting GFP-positive cells over total cells (data not shown). Transgene expression was also confirmed by Western blot analysis (Fig. 2E).

Next, we analyzed the effects of IKKα on endothelial cell migration. We found that overexpression of IKKα significantly increased cell migration compared with control treated cells (Fig. 2A). Interestingly, this induction of cell migration could be completely blocked with the coexpression of IκBα, an inhibitor of NF-κB, suggesting that IKKα-mediated endothelial motility is dependent on activation of NF-κB signaling. In contrast, IKKβ has no significant effect on endothelial cell migration (Fig. 2A). Conversely, we knocked down IKKα in endothelial cells using shRNA. We achieved >85% IKKα reduction in these cells compared with untreated or scrambled controls (data not shown). Knockdown of IKKα significantly impaired cell motility compared with controls (Fig. 2B). Consistent with the cell migration results, expression of IKKα, but not IKKβ, significantly increased vascular tubule formation in a three-dimensional Matrigel assay compared with control cells, and blocking NF-κB activity by overexpressing IκB completely inhibited IKKα-induced vascular network formation (Fig. 2C and D). Furthermore, there was less vascular network formation in the IκB group than the vector control (Fig. 2D), which supports an intrinsic role for NF-κB in angiogenesis. However, we did not find any difference in endothelial cell proliferation between the IKKα group and vector controls by measuring the amount of BrdUrd incorporation in cells (data not shown). Collectively, these data suggest that IKKα mediates angiogenesis via an effect on endothelial cell motility through the NF-κB canonical pathway.

IKKα promotes tumor angiogenesis and tumor growth in vivo. To investigate the role of IKKα in tumor angiogenesis in vivo, we used a tumor and endothelial cell cotransplantation assay, following a published protocol (14), which allows us to genetically manipulate vascular endothelial cells. LLC-tumor cells and viral vector-infected endothelial cells expressing IKKα or β-gal were mixed with Matrigel and implanted into immune deficient Rag-1 mice. As an additional control, LLC cells alone in Matrigel were used. As expected, expression of IKKα in endothelial cells led to a significant increase in tumor growth compared with the controls of tumor cells alone or tumor cells mixed with endothelial cells expressing β-gal (Fig. 3).

Consistent with the observed increase in tumor growth, histologic examination of tumor tissues showed significantly more blood vessels in tumors mixed with IKKα-expressing endothelial cells compared with the vector control group and tumor cells alone (Fig. 4A and B). Based on the data indicating that IKKα promotes endothelial cell migration and vascular network formation, we postulated that expression of IKKα in endothelial cells might increase tumor vascular formation through direct incorporation into tumor vasculature. To test the hypothesis, we used a human-specific CD31 antibody to detect exogenous human endothelial cells and a murine-specific CD31 antibody to detect the host-derived vasculature. We observed incorporation of exogenous human endothelial cells into tumor vasculature in both groups (Fig. 4C). Importantly, there were significantly more human endothelial cells incorporated into the tumor vasculature in the IKKα group than in the β-gal–expressing control group (Fig. 4C and D), which is in agreement with our in vitro data that IKKα promotes endothelial motility and vascular tubule formation.

Further histologic analysis of tumor tissues showed that there were fewer apoptotic cells present among the tumor cells mixed with IKKα-expressing endothelial cells than in both control groups (Fig. 5A and D). Similarly, there was an increase in cell proliferation of the tumor cells mixed with IKKα-expressing endothelial cells compared with either control (Fig. 5B and E). Consistent with these findings, we observed a decrease of hypoxic regions in tumor cells mixed with IKKα-expressing endothelial cells compared with controls (Fig. 5C). Collectively, these findings indicate that IKKα promotes tumor vascular formation, resulting in better nutrient and oxygen supplies for tumor growth and progression.

Although NF-κB has been linked to tumor development and angiogenesis, most studies are focused on IKKβ. In this study, we show that IKKα is elevated in tumor endothelium compared with normal tissues. Also, overexpression of IKKα in endothelial cells increased cell motility and vascular network formation in vitro, and tumor angiogenesis in vivo. As a result, it reduced hypoxia and apoptosis, increased cell proliferation, as well as promoted tumor growth. In contrast, IKKβ did not affect angiogenic properties in endothelial cells. Interestingly, IKKα-mediated angiogenic responses are dependent on NF-κB activation. Together, these findings reveal a new function of IKKα through the NF-κB canonical pathway in tumor angiogenesis.

The IKK complex contains two catalytic subunits, IKKα and IKKβ,both of which share a similar structure. IKKα and IKKβ differ, however, in their physiologic functions (4, 5). Published data indicate that IKKβ is sufficient to mediate NF-κB activation in response to proinflammatory cytokines and microbial products (4, 5). IKKα seems to be dispensable for these functions. On the other hand, studies have identified essential functions of IKKα in the development of teeth and the epidermis and its derivatives; yet, these functions are mediated through an alternative pathway. independent from NF-κB activation (8, 17–19). Importantly, we identify an IKKα-specific role in angiogenesis though canonical NF-κB signaling in this study. IKKα-induced angiogenic functions could be completely attenuated with coexpression of IκBα, thus indicating that IKKα is acting although the canonical, rather than the IKKα-specific noncanonical pathway. Most surprisingly, IKKα, but not IKKβ, the primary IKK in canonical NF-κB signaling, could induce endothelial cell motility and vascular network formation. This finding adds another level of complexity to NF-κB signaling and shows an added importance for IKKα in this pathway.

NF-κB is implicated in tumor angiogenesis (1). Most studies of this process have focused on the regulation of proangiogenic factors, including VEGF, interleukin (IL)-1β, IL-6, and IL-8. A recent report shows that overexpression of a kinase-dead IKKα in tumor cells strongly inhibits both the constitutive NF-κB–dependent promoter and endogenous gene activation. Targeted array-based gene expression analysis reveals that many of the genes down-regulated upon inhibition of this pathway are involved in tumor angiogenesis (20). Consistent with these published data, molecular profiling in our study showed an elevation of VEGF, IL6, and basic fibroblast growth factor in IKKα-expressing endothelial cells compared with control cells (data not shown). It suggests that IKKα can also promote vascular formation through induction of angiogenic genes, in addition to increased endothelial cell motility observed in this study. Although NF-κB has yet to be directly linked to angiogenesis, it should be noted that IKKα-null mice die 30 minutes after birth and show evidence of cardiovascular and potential angiogenic defects (17). In addition, angiogenic factors such as VEGF and angiopoietin 1 have been shown to activate the Akt-IKKα-NF-κB pathway (21–23). These data support our finding of a role for IKKα in vascular motility and angiogenesis.

According to a recent review, there have been over 750 NF-κB inhibitors developed (24). However, none of the developed inhibitors are currently being used clinically. One reason for this may be the potentially severe side effects of inhibiting the NF-κB pathway in its entirety. Based on our data, drugs targeting IKKα, rather than IKKβ or the entire NF-κB signaling pathway, may prove beneficial in a potential clinical setting.

In summary, this study reveals a new role of IKKα in angiogenesis that is dependent on NF-κB activity. A better understanding of the molecular mechanism of tumor angiogenesis offers the promise for development of novel therapeutic strategies for cancer treatment.

No potential conflicts of interest were disclosed.

Grant support: NIH CA108856, NS45888, and AR053718 (P.C. Lin) and a training grant T32CA009592 (L.M. DeBusk).

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.

We thank Kimberly Boelte at Vanderbilt University Medical Center for editing the manuscript.