Inhibitor-κB Kinase in Tumor Promotion and Suppression During Progression of Squamous Cell Carcinoma

Previous studies have provided evidence that activation of inhibitor-κB kinases (IKK) leads to nuclear translocation and activation of nuclear factor-κB (NF-κB), a transcription factor that can promote cell survival and malignant phenotypic changes important in development and progression of squamous cell carcinomas (SCC) and other cancers (Fig. 1; refs. 1, 2). Recent findings in this issue of Clinical Cancer Research (3) and elsewhere (4) provide evidence for an important new role for one of the IKK subunits, IKKα, as a tumor suppressor in SCC. In this issue, Maeda et al. (3) report that loss or decreased nuclear expression of IKKα occurs in 20 of 64 (∼33%) of oral SCCs and is significantly associated with decreased histologic differentiation and prognosis. Using HaCaT immortalized keratinocytes, they show that decreased expression of IKKα, involucrin, and keratins are functionally linked during dedifferentiation, which occurs with calcium deprivation. Based on the observation that the gene encoding IKKα is located near 10q24, a locus altered in a subset of SCC, they sought evidence for genetic or epigenetic alterations of the IKKα gene. Although in oral SCC allelic/biallelic loss of the gene was found in only a few cases, they showed that microsatellite instability and epigenetic inactivation by promoter hypermethylation occurs with high frequency in those SCCs that exhibit loss of nuclear IKKα.These results are consistent with the recent demonstration by Liu et al. (4) that marked reduction of expression and mutation of the IKKα locus is associated with poorly differentiated human SCC of the skin and that transgenic mice overexpressing normal IKKα in the epidermis develop fewer SCCs and metastases when subjected to a standard two-stage 7,12-dimethylbenz(a)anthraacene/12-O-tetradecanoylphorbol-13-acetate tumor induction protocol. In their murine model, expression of IKKα promoted terminal differentiation and reduced proliferation, angiogenesis, and metastasis. Furthermore, SCC that appeared in IKKα+/− mice showed loss of heterozygosity, a clear indication of a tumor suppressor function.33Y. Hu, personal communication. Down-regulation of nuclear IKKα has also been observed in association with decreased differentiation and invasive and metastatic progression of SCC of the skin, lung, and head and neck.44A. Costanzo and M. Karin, unpublished data. These data emerging from independent laboratories indicate that nuclear IKKα is a major SCC tumor suppressor in mice and men. Such loss of nuclear IKKα is associated with decreased differentiation and epithelial to mesenchymal transition, a phenotype linked with SCC progression, invasion, and worse prognosis.Early studies revealed that IKKα together with IKKβ and IKKγ forms a cytoplasmic complex that integrates numerous signaling pathways important in activation of NF-κB transcription factors (Fig. 1A and B), which together have been implicated in protective responses to injury, promotion of inflammation, and oncogenesis (1, 2). This IKK-NF-κB pathway has become known as the canonical pathway (Fig. 1B). This tripartite IKK complex has been shown to phosphorylate via IKKβ Ser32 and Ser36 of inhibitor of κBα (IκBα), leading to its polyubiquitylation and proteasomal degradation, releasing NF-κB for nuclear translocation and binding to target gene promoters. This IKK complex was also shown to phosphorylate Ser536 of the RELA subunit of the NF-κB1/RELA (p50/p65) heterodimer (5), important for transactivation of genes by bound NF-κB, as well as for termination of the induction response (6). Genes controlled by NF-κB execute a diverse program of cellular functions, including survival, proliferation, inflammation, angiogenesis, and innate and adaptive immunity, many of the processes altered during development of cancer (7).Several studies have provided evidence that the canonical IKK and NF-κB pathway plays an important role in the pathogenesis of SCC. One of our laboratories discovered that NF-κB1/RELA (p50/p65), the NF-κB heterodimer activated by the canonical pathway, is aberrantly activated in murine and human SCC lines derived from advanced tumors (8, 9). Zhang et al. (10) confirmed that increased nuclear activation of RELA/p65 frequently occurs in development of squamous dysplasias and carcinomas and is associated with malignant progression and decreased prognosis. Increased expression of NF-κB–related gene signatures is observed in SCC lines and tumor specimens in association with tumor progression, metastasis, epithelial-mesenchymal transition, and decreased prognosis (11–13). Consistent with a functional role for the IKK and NF-κB canonical pathway in aberrant gene expression and the malignant phenotype, we showed that forced expression of IκBα mutated at the IKK phospho-acceptor sites inhibits altered gene expression together with cell survival, proliferation, angiogenesis, tumorigenesis, and therapeutic resistance by SCC (12, 14–16). Tamatani et al. (17) provided further evidence that IKK complex immunoprecipitated by anti-IKKα antibody contributes to phosphorylation of IκBα activation in head and neck SCC. Consistent with a role of IKK in NF-κB activation, we found that expression of either IKKα- or IKKβ K44A–kinase deficient mutants can attenuate NF-κB reporter gene activation in SCC (18).IKKα has been shown to have critical function distinct from its role as part of the canonical IKK pathway. Besides the trimeric IKK complex involved in the canonical pathway, a cytosolic homodimeric IKKα complex acts as the lynchpin in an alternative pathway that activates NF-κB2/RELB dimers important in development of B lymphocytes and secondary lymphoid organs (Fig. 1C; refs. 1, 2, 19). This pathway has been shown to mediate processing of NF-κB2 p100 to p52, which has also been shown to be deregulated and activate NF-κB target genes in some cancers (19). Although we have detected nuclear localization of NF-κB2 in head and neck SCC lines (8),55L. Nottingham, unpublished results. its functional role has not yet been defined, and these lines may represent the subset of well-differentiated SCC that have not lost nuclear IKKα (3, 4).IKKα has been shown recently to function in the nucleus both as transcriptional regulator and protein kinase (Fig. 1D). Epidermal growth factor has been shown to induce the recruitment of IKKα to the c-fos promoter to regulate promoter-specific histone H3 Ser(10) phosphorylation. This occurs in a manner that is independent of p65/RELA, but together with p65/RELA, IKKα up-regulates c-fos expression (20). Park et al. (21) detected similar bacterial lipopolysaccharide-induced IKKα binding and chromatin modification in association with RELA in promoting expression of a variety of NF-κB modulated genes in macrophage lines. Although this mechanism has not been shown in keratinocytes or SCC, epidermal growth factor ligand and receptor overexpression and lipopolysaccharide exposure are common in head and neck SCC. It should be noted that the study by Maeda et al. in this issue did not examine whether the loss of nuclear IKKα affects nuclear localization of NF-κB or transactivation of its target genes, information that would be helpful in determining if the alterations associated with IKKα loss may affect and be mediated via NF-κB–dependent or NF-κB–independent mechanisms.Now, nuclear IKKα has also been shown to function as a tumor suppressor gene (Fig. 1E; refs. 3, 4). Perhaps most critical and relevant to its role as a tumor suppressor in SCC is the observation that IKKα controls the proliferation and differentiation of epidermal keratinocytes (22, 23). IKKα-deficient mice show dysregulated hyperproliferation and defective differentiation of the epidermis, as well as increased NF-κB activity due to loss of an inhibitor function (23). This augmentation of NF-κB activation was shown to occur in keratinocytes stimulated with TNFα, a cytokine found previously to play an important role in promoting SCC formation in mice (24). The increased activation of NF-κB is most likely due to formation of IKK complexes containing only IKKβ as the kinase subunit, with IKKβ being more active than IKKα (23). The control of keratinocyte proliferation and differentiation, however, is not dependent on IKKα protein kinase activity but requires its nuclear translocation (25). Thus, nuclear IKKα is critical for keratinocyte differentiation, and its loss promotes proliferation while preventing the differentiation of the very cell type that causes SCC.One of the major questions that pertain to the function of nuclear IKKα in keratinocytes and its loss in SCC is how it controls cell differentiation and proliferation. It was shown that reintroduction of IKKα into IKKα-deficient keratinocytes induces cell cycle arrest (23), but the mechanism by which IKKα controls cell cycle progression is still being unraveled. IKKα seems to promote expression of a keratinocyte differentiation-inducing factor, independent of its role in activation of NF-κB (23, 26). Keratinocytes with targeted deletion of IKKα were also reported to be defective in calcium-induced differentiation and barrier function, and this was shown to be mediated via effects on retinoic acid receptor gene expression (26). These functions of IKKα in normal squamous epithelia seem most consistent with the associated loss of IKKα expression and differentiation observed in the subset of cutaneous and head and neck SCC described by Maeda et al. (3) and Liu et al. (4). These results suggest that in addition to its cytoplasmic role in modulation of NF-κB activation, nuclear IKKα may be a key mediator of another mechanism or pathway that represses more pathogenic phenotypic changes along the spectrum of SCC.At this writing, this mechanism is unknown, but based on previous reports, there are several intriguing candidates. One of our laboratories reported previously that stepwise progression in a murine SCC model is associated with increased activation of NF-κB and target genes (9, 11), including cytokines, such as Gro-1, which promote inflammation, angiogenesis, and increased tumorigenesis (12, 27). In a subsequent study, we found that NF-κB activation and Gro-1 expression could be repressed by transforming growth factor β1 (TGFβ1) via NF-κB–dependent and NF-κB–independent mechanisms in early-stage SCC, but not in the more advanced metastatic SCC (28). This occurred without apparent loss of TGFβ receptor, suggesting that dysregulation of other intermediates downstream of the receptor could be involved in disruption of TGFβ signaling and malignant progression in SCC. Li et al. (29) report that dysregulation of the TGFβ signaling pathway at the level of TGFβ receptor 2 and SMADs is common in SCC and, as with loss of IKKα, is associated with increased inflammation, angiogenesis, proliferation, and increased skin-derived SCC formation. Additionally, Kobielak et al. (30) have reported that targeted knockout of α-catenin, another common alteration in SCC, results in increased NF-κB activation and increased SCC carcinogenesis of the skin. Interestingly, dysregulation of TGFβ signaling has been linked to cadherin/catenin expression in squamous epithelia. It will be interesting if malignant progression is linked to NF-κB–dependent and NF-κB–independent pathways altered by dysregulation of IKKα or other components of the TGFβ pathway. Thus, loss of IKKα can result in defective differentiation and aberrant proliferation as well as up-regulation of NF-κB signaling due to loss of several inhibitory loops. Together, IKKα loss could result in increased cell survival and angiogenesis, loss of differentiation, epithelial to mesenchymal transition, and malignant progression. Further studies of the mechanisms regulating IKKα nuclear localization, loss, and function will be important to understanding the potential for therapeutic targeting of these mechanisms.There are potential therapeutic implications suggested by the studies of Maeda et al. The current results indicating hypermethylation may be a common mechanism for inactivation of IKKα in head and neck SCC raise the potential that IKKα expression and at least part of its tumor suppressor function could be restored by treatment with demethylating agents, which have been shown to reactivate other hypermethylated tumor suppressor genes. Such an approach could be tested alone, and together with TGFβ, in the subset of SCC found to be deficient for IKKα expression.