Abstract
The components of the nuclear factor-κB (NF-κB) family of transcription factors are critical for regulating the response to immune challenges. Recently, a role for NF-κB in skin biology has been revealed. Within the cascade of proteins whose activities impinge upon the activation of NF-κB, the NEMO (NF-κB essential modulator)/IKKγ protein is required for the activation of the IκB kinases, which in turn, promote the degradation of IκB proteins, leading to the derepression of NF-κB activity. Courtois and Israël discuss the role of NEMO/IKKγ in normal physiological activation of NF-κB and the consequences of defective NF-κB activation, as an effect of NEMO/IKKγ mutations, which can lead to incontinentia pigmenti, a disease marked by alopecia, tooth eruption, skin lesions, and changes in skin pigmentation.
Members of the Rel/NF-κB family of transcription factors play important roles in immune, inflammatory, and apoptotic responses by inducing the expression of numerous cellular and viral genes (1, 2). The nuclear factor-κB (NF-κB) activity consists of homo- or heterodimers of related proteins that share a conserved DNA binding and dimerization domain called the Rel homology domain. In most cell types, NF-κB is sequestered in the cytoplasm bound to inhibitory proteins called IκBα, IκBβ, IκBϵ, p105, and p100. In response to diverse stimuli, including inflammatory cytokines and mitogens, as well as several viral proteins, active NF-κB translocates to the nucleus as a result of the complete proteolytic degradation of the IκB proteins or the partial degradation of the p105 and p100 precursors. This mechanism has been best studied with the inhibitor IκBα and has been demonstrated to involve phosphorylation on two specific serine residues followed by polyubiquitination and degradation of IκBα by the 26S proteasome (3). More recently, a specific serine protein kinase activity responsible for IκBα phosphorylation has been identified within a large cytoplasmic multisubunit complex (700 to 900 kD), and two kinase subunits (IKK1/IKKα and IKK2/IKKβ) and a structural/regulatory component [NF-κB essential modulator (NEMO)/IKKγ/Fip-3/IKKAP] have been cloned (4). The essential role of NEMO was demonstrated during its cloning on the basis of genetic complementation of a fibroblast cell line defective in NF-κB response to multiple stimuli (5). The physiological role of the NF-κB family has been mostly studied by mutational targeting of the various activating subunits or the individual IκB subunits in the mouse (6) (Fig. 1). Specific inactivation of individual NF-κB proteins leads to multiple immune dysfunctions, except for the inactivation of p65, which resulted in death of the animals around day 16 of embryonic development (7). This has been attributed to extensive apoptosis in the liver, as mediated by tumor necrosis factor-α (TNF-α) (8), a cytokine with both pro- and anti-apoptotic roles. In p65-/- mice, the strong reduction of NF-κB activity leads to apoptosis. The inactivation of IKK2 (9-11), IKK1/IKK2 (12), or NEMO (13-15) leads to a similar phenotype, although the age of death varies slightly--the more extensively NF-κB activation is inhibited, the earlier the death.
Gene targeting of IKK and NF-κB subunits. Only major features of single knockout are indicated. To simplify the picture, c-rel, relB, p52, and IκBϵ gene targeting phenotypes are not shown. In each case, immune defects were observed (6, 27). IP, incontinentia pigmenti; E, embryonic day; and P, postnatal day.
More recently, a role for NF-κB in the biology of epidermis has been proposed (16). An original observation indicated that NF-κB becomes activated when epidermal cells exit the basal layer and undergo terminal differentiation (17). The role of NF-κB was therefore addressed by expressing in the skin a mutant IκBα that is resistant to degradation and therefore prevents NF-κB activation. Overexpression of this molecule in the basal layer resulted in a hyperplastic epithelium and lack of hair formation (17, 18). Epidermal hyperplasia could also be induced by pharmacological inhibition of NF-κB. Another group confirmed these observations and observed, in addition, an increased basal apoptosis in the skin of these manipulated mice, as well as the spontaneous development of squamous cell carcinomas (19). As expected, overexpression of constitutive NF-κB in the skin led to the formation of a thin hypoproliferative epidermis. Nonetheless, manipulating the NF-κB response did not alter the expression pattern of various differentiation markers. A model suggested by these data is that NF-κB activation is required to cause or maintain cell-cycle arrest when keratinocytes leave the basal layer to undergo terminal differentiation (16).
Another hint as to a possible role for NF-κB in skin biology was suggested by the phenotype of IKK1-deficient mice. The observed phenotype was unexpected, especially in view of the phenotype of an IKK2 knockout (20-22). The IKK1-deficient mice die just after birth, exhibit multiple skeleton abnormalities, and seem to have no limbs, tail, or ears; in fact, they have a very thick skin that prevents these structures from extending properly out of the body trunk. Most important, the epidermis of the mutant mice is 5 to 10 times as thick as that of normal mice, because of an excessive proliferation of the basal layer, and epidermal differentiation is almost completely absent (the expression of late terminal differentiation markers was strongly reduced). A thickened epidermis as a result of the inactivation of a component of the NF-κB signaling pathway is consistent with the model proposed above; however, the alteration in terminal differentiation was unexpected and suggests that NF-κB might respond to an as-yet-unknown developmental signal in the skin. These results point to a specific role of IKK1 in the skin, which does not necessarily correlate with NF-κB activation. In any case, the data summarized above suggest that NF-κB probably plays a dual role at the level of epidermis: It protects keratinocytes from apoptosis [see also (23)], and at the same time, it controls their proliferation and possibly their differentiation.
The most spectacular demonstration of the involvement of NF-κB in the biology of epidermis, however, has been the recent demonstration of the involvement of the NEMO molecule in a human X-linked dominant genodermatosis called incontinentia pigmenti (IP) (24, 25). This disease, with an incidence of between 1/10,000 and 1/100,000 presents almost exclusively in females, as afflicted males die in utero. In affected females, the disorder is highly variable in presentation. Typically, IP has four dermatological stages that begin neonatally with blisters and an inflammatory response. A second stage involves the appearance of wartlike lesions, and this is followed by streaks and whorls of pigmented skin that follow Blaschko's lines. These generally disappear by the second decade, and adults may show areas of dermal scarring. Problems with tooth eruption and alopecia are common in IP, and more serious complications include retinal dysplasia and neurological signs that occur in a small proportion of cases. IP females exhibit extremely skewed X-chromosome inactivation in blood. For some unknown reasons, X inactivation skewing is less complete in the skin, and selective elimination of cells bearing a mutated X chromosome only starts at or around birth. The nature of the signal that triggers IP cells' elimination at this particular time is itself unknown.
The gene responsible for IP was mapped to an interval of about 2 Mb distal to the color vision locus in Xq28. Last year, the gene encoding NEMO was also mapped to Xq28 and found to overlap with the glucose-6-phosphate dehydrogenase (G6PD) locus (26). This localization, as well as the exquisite sensitivity to apoptosis of cells derived from IP patients, suggested that NEMO might be involved in this pathology. The analysis of a large collection of patients indeed showed that 85% of them carried a complex rearrangement of the NEMO gene (25). This rearrangement involves excision of the region between two repeated sequences located upstream of exon 4 and downstream of exon 10, respectively, resulting in the synthesis of a truncated NEMO molecule consisting of 133 amino acids (corresponding to exons 1 to 3), which is devoid of activity. In IP patients carrying this rearrangement, a lack of NF-κB activation was demonstrated by studying fetus-derived primary fibroblasts: these cells are unresponsive to all tested NF-κB-activating stimuli, they do not show degradation of the IκB molecules when stimulated, and they are very sensitive to TNF-α-induced apoptosis.
The high frequency of the rearrangement affecting the NEMO locus will allow IP to be easily diagnosed in the vast majority of cases by a simple polymerase chain reaction assay. On the treatment side, generating an animal model would be helpful. Recently two papers have described the phenotype of nemo knockout mice (14, 15). As previously described in knockout mice devoid of NF-κB activity, males lacking nemo (nemo-) die from liver apoptosis around day 13. It is also interesting that heterozygous female mice develop a transient dermatosis at birth, characterized by patchy skin lesions with massive granulocyte infiltration, hyperproliferation, and increased apoptosis of keratinocytes, reminiscent of what is observed in human IP. Hyperpigmentation, a hallmark of IP owing to the presence in the dermis of phagocytes containing melanosome complexes, was also observed in skin sections from nemo+/- mice. Ultrastructural analysis of these mice indicated that, contrary to what is observed in wild-type animals, basal layer keratinocytes are in loose contact with each other, forming filopodia that extend into the dilated intercellular spaces. This phenomenon was also observed in the skin of IP patients, and a similar picture has been described in the basal layer of the epidermis of IKK1-/- mice. A possible explanation for the progressive, but transient, nature of the skin pathology observed in these female mice, as well as in IP patients (14), is as follows. First, nemo- cells hyperproliferate and differentiate normally because of a lack of NF-κB activation, and second, some of these cells die and induce an inflammatory response involving cytokine and chemokine release by adjacent normal cells, which in turn increases the death rate of mutant cells. Eventually most of the nemo- cells are eliminated and replaced by normal cells; however, any remaining nemo- cells will eventually lead to another cycle of dermatosis, as has actually been observed in human patients.
Severe skewing of X-chromosome inactivation in the lymphocyte compartment suggests that, as is observed in IP patients, mouse lymphocytes carrying an active copy of the mutated X chromosome are specifically eliminated after birth. However, in contrast to what is observed in humans, these female mice exhibit a significant rate of early lethality, although this seems to vary depending on the murine genetic background. Therefore, these mice represent a valuable animal model for IP, which will allow therapeutic approaches to be tested.
Study of IP patients should provide a unique view of the in vivo role of NF-κB in humans. As pointed above, a total inactivation of NEMO results in male lethality during embryogenesis. It remains to be determined whether death is caused by liver apoptosis, as in mice. As for IP females, their phenotype may be more difficult to assess, because of their mosaic status. Their skin syndrome already suggests a complex interplay between mutated and wild-type cells. In other tissues, much less is known, but similar events are likely to occur. Rare patients carrying missense mutations of the NEMO gene may be more informative. One can predict that some of these mutations will only partially affect NEMO function and, consequently, generate a less morbid phenotype in males. Two such cases have already been described, and one of them exhibited severe innate immune deficiency together with osteopetrosis and lymphodema. This suggests that NEMO mutations may produce, in males, a phenotype quite distinct from IP and characterized by immune deficiency. Future studies should reveal how frequently patients have this condition and whether NEMO dysfunction is indeed the cause of other human syndromes in addition to IP.
