We have previously described the generation of C-IKKα and N-IKKα transgenic mice which express the human IKKα protein under the control of the Keratin 5 (K5) regulatory elements [20]. As we explained, the modifications that these models carry are either the elimination of the nuclear localization signal (NLS) in the cDNA of the human IKKα by directed point-mutation in C-IKKα mice or the insertion of an additional NLS sequence in the amino terminal in N-IKKα mice, outside of the kinase domain of the human IKKα/CHUK gene. As reported, C-IKKα and N-IKKα mice express the transgenic IKKα in the cytoplasm or in the nucleus of epithelial cells respectively, following the expression pattern of K5 [20]. Thus, they express the transgene in keratinocytes of the basal layer of the epidermis, the outer root sheath (ORS) of hair follicles (HF) and in the immature cells surrounding the sebaceous glands (SG) (Supplementary Fig. 1). Both lines of transgenic mice developed normally after birth. However, while N-IKKα mice remain indistinguishable from their non-transgenic (Control) littermates, transgenic C-IKKα mice showed a characteristic altered fur from very early age (14-day-old) due to a delay in hair growth, and from the age of 3 months they presented diffuse alopecia (Fig. 1A, B).

Fig. 1: External phenotype and histological alterations in the back skin of C-IKKα transgenic mice.

A, B Diffuse alopecia in the back of 14-day-old (left) and 3-month-old (right) C-IKKα mice. H&E staining of back skin sections of 5-day-old Control (C) and C-IKKα (D) mice showing parakeratosis (arrowhead), dyskeratosis (dotted arrow), and epidermal hyperplasia (yellow double-headed arrow) (Gr; dark purple granulated keratinocytes) in C-IKKα mice. E, F Back skin sections from 60-day-old mice. Hyperplasia of the ORS (black arrows) and mature sebocytes of SG (red arrows) in telogenic HF of transgenic mice (F) versus normal ORS of HF and SGs in Control mice (E). G, K Observe the worsening of the skin phenotype of C-IKKα mice from 6 months of age; G Control mouse, note the presence of small SG (arrowhead); H–K C-IKKα mice. H Abundant foci of dyskeratosis and parakeratosis (dotted arrow and arrowheads, respectively); hyperplasia of the ORS of HF (black arrow); numerous orphan and hyperplastic SG (empty arrows). Inset shows parakeratosis at higher magnification. See notable papillomatous hyperplasia in the epidermis of transgenic mice (circle in I) and the formation of sporadic papillomas (black arrow in J). K Formation of follicular cysts (blue arrows) from degenerated HF and severe papillomatous hyperplasia of IFE with dyskeratosis and parakeratosis (bracket). 6–10 animals of each age and genotype were analyzed. Each photo represents a different section of skin from the same mouse or different mice. Scale bars: C, D: 50 µm; E, F, G–I, K: 100 µm; J: 200 µm.

C-IKKα mice exhibit hyperplasia of the epidermis and adnexal organs

Back skin sections of transgenic mice of different ages, from newborns to adults, were analyzed and revealed epidermal alterations that began soon after birth (Fig. 1C–K). Five-day-old C-IKKα mice showed hyperplasia of the interfollicular epidermis (IFE), mainly due to the increased number of cell layers in the stratum granulosum (Fig. 1C, D), as well as foci of dyskeratosis, i.e., abnormal keratinization occurring prematurely below the stratum granulosum. Additionally, the skin of these transgenic mice showed parakeratosis (anomalous retention of nuclei within the stratum corneum) suggesting alterations in epidermal terminal differentiation (Fig. 1D). These changes worsened with age and other alterations also appeared in older mice; thus, 31-and 60-day-old C-IKKα animals showed reduced numbers of HF (Fig. 1F; Supplementary Fig. 2A–C), hyperplasia of the ORS of the HF, as well as hyperplasia of the HF-associated SGs (Fig. 1F; Supplementary Fig. 2D–E). This phenotype was even more notable in C-IKKα mice from 6 months of age, which showed persistent dyskeratosis and parakeratosis of IFE, and papillomatous epidermal hyperplasia (Fig. 1H, I, K) that occasionally led to papilloma development (benign epidermal growths; observed in 3/23 C-IKKα versus 0/28 Control mice) (Fig. 1J). Also, follicular cysts (Fig. 1K), and “orphan” SGs (non-associated to HF) (Fig. 1H, I) were detected in the dermis of transgenic animals.

Similar alterations were also found in the skin of other locations in C-IKKα mice, such as lip, plantar sole, eyelid and tail skin (Supplementary Fig. 3).

The skin of N-IKKα mice exhibits epidermal atrophy, hyperkeratosis, and signs of premature aging

N-IKKα mice showed skin alterations from early ages that became more prominent with aging. Thus, five-day-old N-IKKα mice showed a stretched and thin epidermis with 3–4 layers of keratinocytes, which contrasted with the 6 layers of keratinocytes arranged along soft epidermal folds exhibited by Control animals (Fig. 2A, B). At 31 days of age, N-IKKα mice developed dysplastic HFs and presented foci of epidermal atrophy containing a single layer of keratinocytes (Supplementary Fig. 2F–H). Foci of atrophic epidermis alternated with normal epidermal thickness resembling the histological appearance of wrinkles in aged skin (Supplementary Fig. 2G). When these transgenic mice reached 60 days of age atrophic epidermal foci were more frequent along the IFE, whereas Control mice did not exhibit these changes (Fig. 2C, D). In addition, other alterations associated with premature aging were also detected in the skin of young N-IKKα animals, including a thinner layer of adipose tissue in the hypodermis and increased packing of collagen fibers (sclerosis) in the dermis (Fig. 2D, E; Supplementary Fig. 2G). In addition, 6-month-old and older N-IKKα mice showed altered epidermal differentiation, i.e., the epidermis of N-IKKα mice showed disorders in the maturation of suprabasal keratinocytes, which underwent an abrupt transition from the basal layer of keratinocytes to the stratum corneum, which appeared thickened, parakeratotic and with scales that retained the nucleus (Fig. 2G); this caused the presence in the skin of N-IKKα mice of extensive areas of the IFE that lacked the spinosum and granulosum strata (Fig. 2G, H). Hyperkeratosis and increased desquamation was detected in these transgenic animals (Fig. 2G–I). This altered and accelerated desquamation observed in N-IKKα mice contrasted with the normal shedding of anucleated keratin scales that occurred in Control mice (Fig. 2F). Similar abnormalities were found in the skin of N-IKKα mice in other localizations such as tail and eyelid skin (Supplementary Fig. 4).

Fig. 2: Histological alterations in the back skin of N-IKKα mice.

A, B Back skin sections of 5-day-old mice. A Representative image of the skin of Control mice showing the interfollicular epidermis (IFE) containing 6 layers of keratinocytes and forming soft folds versus the IFE of N-IKKα mice (B) that appears stretched and thinner, with only 3–4 layers of keratinocytes. Back skin of 60-day-old Control (C) and transgenic (D, E) mice. C See in skin sections of Control mice the presence of 2–4 layers of keratinocytes in the IFE, short and well differentiated telogenic HF and abundant adipose panicle (asterisk). D, E The skin of transgenic mice shows wide areas of epidermal atrophy (brackets), sclerosis of the dermis (green star), with foci of parakeratosis (arrowheads, E), and scarce adipose tissue (asterisk). F–I Representative sections of the back skin of 6-12-month-old Control (F) and transgenic (G–I) mice. F See in the skin of Control mice the presence of 2 cell layers of keratinocytes and the stratum corneum composed of anucleated, laminated keratin scales and adipose panicle is present (asterisk). G, H Severe sclerosis of the dermis (green stars), and striking hyperkeratosis (white stars). Arrowheads indicate parakeratotic foci. Note in (G) the extreme epidermal atrophy of the IFE (brackets), the absence of adipose panicle, and the presence of keratinocytes that experiment an abrupt transition from a flat basal layer to nucleated scale (black arrow). I Severe hyperkeratosis (stars), lack of adipose panicle (asterisk) and frequent regions of epidermal atrophy (brackets). 6–10 animals (males and females) of each age and genotype were analyzed. Each photo represents a different section of skin from the same mouse or different mice. Scale bars: A, B: 80 µm; C–E: 100 µm; F–I: 120 µm.

Therefore, our data suggest that IKKα plays different functions in nucleus and cytoplasm of keratinocytes that are necessary for proper skin development, and that alterations in the subcellular localization of IKKα disrupt normal skin homeostasis, leading to the development of different and almost opposite skin phenotypes.

Highly proliferative epidermis with defective terminal differentiation in C-IKKα mice versus increased epidermal terminal differentiation in N-IKKα mice

Given the cutaneous phenotype of C-IKKα and N-IKKα mice that suggested alterations in the balance of proliferation and differentiation of keratinocytes, and considering the key role of IKKα in the regulation of these processes [4], we examined both the proliferation and differentiation properties of the skin of C-and N-IKKα animals. The rate of proliferation was studied by analyzing BrdU incorporation and Ki67 immunostaining in the skin of 18-month-old mice, the age at which these mice exhibit the most intense phenotype. The results showed that, as expected, the epidermis of Control animals was sparsely proliferative, exhibiting a low number of Ki67-positive basal keratinocytes (Fig. 3A, and data not shown). By contrast, a significant number of basal and suprabasal keratinocytes proliferated in the epidermis of C-IKKα mice (Fig. 3B). Although no quantitative differences in the rate of epidermal proliferation were found in N-IKKα mice, suprabasal keratinocyte proliferation was nevertheless observed in isolated foci (Fig. 3C). The analysis of BrdU incorporation confirmed the increased number of proliferative cells only in the epidermis of C-IKKα mice (Fig. 3D). Consistent with these observations, K5 staining showed an increased number of basal and suprabasal keratinocyte layers that expressed this keratin in C-IKKα mice, as opposed to the single basal keratinocyte layer that express K5 in Control mice (Fig. 3E–G).

Fig. 3: Analysis of epidermal proliferation and differentiation in IKKα-transgenic mice.

A–C Anti-Ki67 immunohistochemistry of tail skin sections of 18- month-old mice showing proliferating keratinocytes; n = 6 animals of each genotype were analyzed. A Scarce proliferative basal keratinocytes in the epidermis of Control mice (black arrows). B Increased number of proliferative basal keratinocytes (black arrow) and abundant proliferative suprabasal keratinocytes (red arrow) in C-IKKα mice. C Low numbers of positive basal keratinocytes (black arrows) and foci of proliferative suprabasal keratinocytes (red arrows) in N-IKKα mice. Insets: higher magnification of the section framed by the box. D Graph showing the significant increase in the number of proliferative keratinocytes, positive for BrdU incorporation, in the IFE of the skin of C-IKKα mice (***p < 0.001). Three independent researchers analyzed and counted 4–5 animals of each genotype. Double immunofluorescence of tail skin sections from Control (E) and C-IKKα (F, G) mice with anti-K10 (red) and anti-K5 (green) antibodies showing K5 expression in the basal layer of keratinocytes in Control mice and K10 expression in keratinocytes of the suprabasal layers (E). F, G Note the aberrant expression of K5 in suprabasal layers of keratinocytes from C-IKKα mice, including the most upper epidermal differentiation layers (white arrows in F, G). Note the absence of K10 in several layers of suprabasal keratinocytes in C-IKKα mice. Western blots showing decreased levels of Filaggrin expression in the skin of C-IKKα mice (H) and increased levels in the skin of N-IKKα mice (I). See in (I) two different times of exposition of the same hybridization with the anti-IKKα antibody: the longer exposition allows seeing the endogenous IKKα; the lower exposition was used to quantify levels of IKKα expression. n = 8 mice of each genotype. Anti-IKKα from Novus Biologicals (Supplementary information) that in western blot recognizes mouse and human IKKα was used. Arrows = mIKKα: mouse IKKα; hIKKα: transgenic human IKKα (C-IKKα (H) or N-IKKα (I)). J Day 7 of in vitro differentiation of human HaCaT keratinocytes of the three genotypes. Note the appearance of the domes of stratified HaCaT-Control cells (white arrows). Larger and more abundant domes of differentiation were formed in HaCaT-N-IKKα cells (white arrows) whereas they were not detected in HaCaT-C-IKKα cells. Scale bars: A–C: 150 µm, E–G: 70 µm; J: 50 µm.

Examination of early epidermal differentiation by immunostaining with an anti-K10 antibody showed in the epidermis of C-IKKα mice areas of suprabasal layers of keratinocytes that expressed K5 instead of K10, which was absent (Fig. 3F, G). No changes in the expression of K5 and K10 were observed in the epidermis of N-IKKα mice (data not shown). Analysis of the expression of Filaggrin, a protein marker of epidermal terminal differentiation, showed a lower level of expression in the skin of C-IKKα mice, compared to that detected in Control mice, while it was increased in the skin of N-IKKα mice (Fig. 3H, I). To further confirm that changes in IKKα expression levels in the nucleus or cytoplasm of keratinocytes produced alterations in their capacity to differentiate, we performed in vitro differentiation assays of HaCaT human keratinocytes expressing exogenous IKKα in the nucleus or cytoplasm [19, 20]. After one week growing under serum-deprived conditions, HaCaT-Control-cells showed the first signs of differentiation, i.e., formed small domes of differentiated keratinocytes, while HaCaT-N-IKKα cells exhibited a great number of large differentiation domes and, HaCaT-C-IKKα cells did not yet show any sign of differentiation (Fig. 3J). Therefore, these results in HaCaT keratinocytes confirmed those obtained in vivo in the epidermis of N- and C-IKKα mice, indicating that nuclear expression of IKKα favored keratinocyte differentiation, while in the cytoplasm it caused defects in epidermal differentiation. Interestingly, the hyperplasia of the epidermis, the altered balance between keratinocyte proliferation and differentiation, and the aberrant expression of K5 and K10 in the epidermis of C-IKKα mice were partly reminiscent of the changes described in the skin of IKKα-deficient mice [4,5,6].

The analysis of the apoptotic rate in the epidermis of mice of the three genotypes by immunohistochemical staining and Western blot (WB) with an anti-Activated-Caspase 3 antibody showed no differences (data not shown).

Mechanisms involved in the development of the skin phenotype of the two types of IKKα-transgenic mice

To characterize the mechanisms leading to the development of the cutaneous phenotype in C- and N-IKKα mice, we analyzed the skin transcriptome of 6-month-old mice across the three genotypes using RNAseq. We chose this specific age because it allowed the analysis of changes in gene expression in the skin of mice that did not yet present significant alterations, since one of the objectives of the study was to determine, at the genetic level, if signs of premature aging were detected in N-IKKα animals. Gene set enrichment analysis (GSEA) revealed transcriptome changes characteristics of molecular signatures of aging in N-IKKα animals. Among them, we found reduced expression of genes that are downregulated at advanced ages, such as those involved in fatty acid metabolism, adipogenesis, oxidative phosphorylation and glycolysis (Fig. 4A; Supplementary Fig. 5A-D) [28,29,30]. In addition, we detected the underexpression of genes involved in myogenesis (Fig. 4A; Supplementary Fig. 5E), this inhibition being characteristic in elderly mammals, in which a reduction in muscle mass and its regenerative potential occurs [31]. Taken together, these GSEA results were consistent with histological observations in the skin of N-IKKα mice, which showed a severe decrease in hypodermic adipose tissue along with other signs of accelerated skin aging. In addition, we observed an enrichment of Myc target genes in N-IKKα mice (Fig. 4A; Supplementary Fig. 5F). To further verify these data, we analyzed c-Myc expression by WB and found increased levels in the skin of N-IKKα mice (Fig. 4B), which is in agreement with previous results showing that IKKα increased c-Myc protein stability and therefore c-Myc levels [32, 33], although our data now suggest that this action is only exerted by nuclear IKKα. Both overexpression of c-Myc and the underexpression of lipid-related genes in the skin of N-IKKα mice could explain the hyperkeratosis observed in these animals, since c-Myc is induced as a compensatory mechanism to repair altered terminal epidermal differentiation [34]. In addition, defects in lipid metabolism enzymes are linked to aged skin and could predispose to the development of an aberrant stratum corneum [35].

Fig. 4: Opposite regulation of fatty acid metabolism and c-Myc pathways in N-IKKα and C-IKKα mice.

A GSEA analysis was performed comparing “N-IKKα versus Control skin” or “C-IKKα versus Control skin” and using the Hallmarks gene sets from the MSigDB database. 6 mice of each genotype were analyzed in the RNAseq study. Heatmap represents normalized enrichment scores (NES) of gene sets with significant deregulation (p val < 0.05 and FDR < 0.05). Hatched boxes indicate that the genes are not significantly regulated in the skin of C-IKKα versus the skin of Control mice. Positive values (red color): overexpression in transgenic mice skin (N-IKKα or C-IKKα) versus Control skin; negative values (blue color): underexpression in transgenic mice skin (N-IKKα or C-IKKα) versus Control skin. Results show down- and up-regulation of genes involved in fatty acid metabolism in the skin of N-IKKα mice and C-IKKα mice, respectively. GSEA analysis also shows downregulation of genes of adipogenesis, oxidative phosphorylation, glycolysis and myogenesis pathways in N-IKKα mice. c-Myc target genes are up- and down-regulated in the skin of N-IKKα mice and C-IKKα mice, respectively. B WB showing the increased expression of c-Myc in the skin of N-IKKα mice. Anti-IKKα from Novus Biologicals (Supplementary information) that in western blot recognizes mouse and human IKKα was used. Arrows = mIKKα: mouse IKKα; C-IKKα, N-IKKα: transgenic human IKKα. Actin was used as a loading control.

GSEA also revealed that C-IKKα mouse skin showed increased enrichment of target genes of the canonical NF-κB, and STAT3 pathways (Fig. 5A; Supplementary Fig. 6A, B). We analyzed the activation of these pathways by WB and found that NF-κB signaling was overactivated in the skin of C-IKKα mice, as demonstrated by increased levels of NF-κB activation (measured as P-p65 levels) at basal conditions, in untreated skin (Fig. 5B), and after stimulation with tumor necrosis factor-α (TNF-α), where activation was observed to last for longer times, i.e., up to 40 min post-inoculation of TNF-α (Fig. 5C); IKKβ and IKKγ levels showed no significant changes after TNF-α stimulation (Supplementary Fig. 7). Total levels of p65 were comparable in mice of all three genotypes (Fig. 5B) and did not change with TNF-α stimulation (Fig. 5C); an increase in the expression of IKKβ and IKKγ was observed in C-IKKα mice (Fig. 5B). No changes in the kinetics of p65 phosphorylation were detected in the skin of N-IKKα mice (Supplementary Fig. 8). To determine whether the overactivation of NF-κB in C-IKKα mice was maintained at older ages, we analyzed the kinetics of p65 phosphorylation in the skin of 20-month-old C-IKKα mice and also found higher levels of P-p65 in the skin of these old transgenic mice (Fig. 5D), also verifying that NF-κB hyperactivation increased with age (compare the densitometric data in Fig. 5C, D). Therefore, our data show that there was chronic overactivation of the canonical NF-κB pathway in the skin of C-IKKα mice and that the C-IKKα construct is functional throughout the life of C-IKKα mice.

Fig. 5: Up- and down-regulation of NF-κB and STAT3 pathways and inflammation-related responses in the skin of C-IKKα and N-IKKα mice, respectively.

A GSEA analysis was performed comparing “N-IKKα versus Control skin” or “C-IKKα versus Control skin” and using the Hallmarks gene sets from the MSigDB database. 6 mice of each genotype were analyzed in the RNAseq study. Heatmap represents normalized enrichment scores (NES) of gene sets with significant deregulation (p val < 0.05 and FDR < 0.05). Positive values (red color): overexpression in the skin of C-IKKα mice versus Control skin; negative values (blue color): underexpression in the skin of N-IKKα animals versus Control mice. GSEA shows the up- and down-regulation of genes of the canonical NF-κB pathway, and STAT3 pathway in the skin of C- and N-IKKα mice, respectively. GSEA analysis also shows the up- and down-regulation of genes related to inflammatory responses in the skin of C-IKKα and N-IKKα mice, respectively. B–E WBs of total protein extracts from the back skin of mice of the three genotypes. B Representative WB showing increased activation of NF-κB, measured as P-p65, in the skin of C-IKKα mice, in the basal state, in absence of activation, although no important changes in total p65 levels are detected. Also note the increased levels of IKKβ and IKKγ in the skin of C-IKKα animals. Anti-IKKα from Novus Biologicals (Supplementary information) that in western blot recognizes mouse and human IKKα was used. Arrows = mIKKα: mouse IKKα; C-IKKα, N-IKKα: transgenic human IKKα. C Augmented levels of P-p65 in the skin of 3-day-old C-IKKα mice at the indicated times after TNF-α stimulation. No relevant differences in total levels of p65 were observed. D Increased P-p65 levels in the skin of 20-month-old C-IKKα mice at the indicated times after TNF-α stimulation. E WB showing increased levels of P-STAT3 in the skin of 1-month-old C-IKKα mice. No differences in total levels of STAT3 are detected. GAPDH and Actin: loading controls.

Next, we analyzed the activation of STAT3 signaling and, consistent with the RNAseq data, found that it was increased in the skin of C-IKKα mice (no differences in total levels of STAT3 were found between mice of the three genotypes) (Fig. 5E). In agreement with the chronic activation of these two main inflammatory pathways, NF-κB and STAT3, in the skin of C-IKKα mice, GSEA analysis also showed enrichment in genes involved in the inflammatory response as well as in interferon γ (IFNγ) and IFNα responses, and host versus graft disease (Fig. 5A; Supplementary Fig. 6C, D). Accordingly, we detected infiltration of CD45+ and CD3+ cells in the skin of C-IKKα animals (Fig. 6A–L) and elevated levels of proinflammatory cytokine genes encoding proteins TNF-α, IL6, IL23, IL1β and IL17 (Fig. 6M). These findings were corroborated by a detailed histological analysis that showed infiltration foci of polymorphonuclear neutrophils, mainly intraepidermal or located in the ORS of HFs, in the skin (mainly in the tail and eyelids) of C-IKKα mice (Fig. 6N, O). By contrast, gene sets related to inflammatory and IFNγ and IFNα responses were significantly underexpressed in the skin of N-IKKα mice (Fig. 5A; Supplementary Fig. 6C, D). Overactivation of NF-κB and STAT3 was linked to epidermal hyperplasia and defects in terminal epidermal differentiation [36, 37], so the activation of these pathways could explain the development of the skin abnormalities seen in C-IKKα mice. Interestingly, activation of these signaling pathways as well as increased inflammation were also features of the skin of IKKα–/– mice [24, 38].

Fig. 6: Inflammation features of the skin of C-IKKα mice.

A–F CD45 immunostaining showing inflammatory cell infiltration in the skin of C-IKKα mice (B, E). G–L CD3 immunostaining showing CD3 + T cell infiltration in the skin of C-IKKα mice (H, K). n = 5–6 mice of each genotype were examined (12-month-old). M Graphs showing increased levels of proinflammatory cytokine genes in the eyelid skin of 9 to 12-month-old C-IKKα mice. mRNA levels of expression were analyzed by qPCR analysis. Each circle, square and triangle correspond to a different animal examined. Representative sections of the tail skin of 12-month-old Control (N) and transgenic (O) mice. Arrows in O: infiltration foci of polymorphonuclear neutrophils. Scale bars: A, B, N: 200 µm; C, D, F: 50 µm; E, O: 40 µm; G, H: 100 µm; I–L: 200 µm.

Remarkably, as can be seen, the GSEA showed the opposite enrichment of well-defined biological states (Hallmarks) in each type of transgenic mouse. Thus, some pathways, such as NF-κB and STAT3, and responses to inflammation, IFN-γ and IFN-α, were significantly upregulated in the skin of C-IKKα mice and downregulated in that of N-IKKα mice; and Myc target genes that were increased in the skin of N-IKKα mice appeared significantly downregulated in C-IKKα mice. Other pathways related to adipose metabolism that were significantly downregulated in the skin of N-IKKα mice had a marked tendency to be upregulated in that of C-IKKα animals.

N-IKKα mice develop squamous skin carcinomas

The histological and molecular changes in the skin of N-IKKα mice suggested that these animals could present a greater predisposition to the development of SCCs, i.e., they showed premature signs of skin aging, with aging being the main cause of NMSC development, and, furthermore, they exhibited a marked overexpression of c-Myc in the skin, which plays a causal role in the genesis of several types of murine and human tumors [39], including the development of spontaneous cutaneous SCCs in K5-cMyc transgenic mice [40]. Therefore, we analyzed mice over 18 months old for possible tumor development. Our studies showed that, in addition to the changes explained in the previous paragraphs, preneoplastic changes occurred in the back and tail skin of N-IKKα mice. These alterations included severe atypia of keratinocytes that began within the basal layer of the epidermis (Fig.7B) and extended above, in the suprabasal layers (Fig. 7C, F); these abnormal keratinocytes lost their polarity, and showed small, rounded and pyknotic nuclei arranged parallel to the basement membrane, in sharp contrast to those of Control animals that were elongated and arranged in a palisade, perpendicular to the basement membrane (compare Fig. 7A, B); these regions of severe atypia showed poor definition of the basement membrane profile (compare Fig. 7A with B, C; and 7D with E). In addition, wide areas of severe epidermal hyperplasia with elongated rete ridges were observed in the skin of N-IKKα mice (Fig. 7C). Increased number of dysplastic, severely dedifferentiated, and abortive HFs were also detected (Fig. 7F, H, I).

Fig. 7: Development of premalignant lesions resembling those of patients with actinic keratosis, and cutaneous squamous cell carcinomas in the skin of N-IKKα mice.

Tail skin sections of Control (A, D) and transgenic (B, C, E, F) mice. n = 8 female and male mice of each genotype (18–24 months old). A Note the arrangement of the nucleus of the basal layer of keratinocytes perpendicular to the basal membrane in Control mice. B Representative example of loss of nuclear polarity of keratinocytes with pyknotic nuclei in the basal layer of the epidermis of transgenic mice, and basal cell atypia (bracket). C Indented basal layer of keratinocytes, with pseudocarcinomatous basal hyperplasia forming rete ridges (arrows), showing a clear separation with the dermis (asterisks). Observe the atypia of keratinocytes of basal and suprabasal epidermal layers. D, E Immunostaining with an anti-Laminin antibody. Observe the discontinuous pattern of expression in the basal membrane of N-IKKα mice (arrows). F Dysplastic and hyperkeratotic HF (asterisk) and atypia of keratinocytes which appear pyknotic (arrows). Back skin section of Control (G) and N-IKKα (H, I) mice. H Severely dysplastic HFs (arrows); (I) Abortive, undifferentiated and dysplastic HF (arrow). J–K Carcinomas in situ in the tail skin of 18-and 20-month-old transgenic mice. K’ Amplification of the rectangle in (K) showing the presence of frequent mitosis in basal and suprabasal keratinocytes (red arrows). Note that the basal membrane is not yet interrupted. L–N Malignant tumors developed in N-IKKα mice. L Invasive (infiltrating) squamous cell carcinoma (SCC). Observe the erased limit of the basal membrane and how small undifferentiated tumor cells form cords that infiltrate the dermis (d) (arrows) (ep: epidermis); (L’): amplification of the rectangle in (L) showing how keratinocytes invade the dermis (area delimited by the dashed line). Observe the presence of keratinocytes with loose unions between them at the advancing front. M Well differentiated SCC in the back skin of a 19-month-old N-IKKα mice; note the concentric laminated masses of keratin (keratin pearl). N Undifferentiated SCC developed in the tail skin of a 22-month-old transgenic mouse. d: dermis, ep: epidermis. N’ Amplification of the rectangle in (N) showing the contrast between the differentiated morphology of the epidermis and the dedifferentiated morphology of keratinocytes invading the dermis (area between dashed lines). Scale bars: A, B: 40 µm; C, F, I, J: 50 µm; D, E, G, H, M: 100 µm; K: 40; K’: 20 µm; L: 30 µm; N: 150 µm; L’, N’: 25 µm.

All these changes, together with the areas of epidermal atrophy and parakeratosis, the disruption of the normal sequence of keratinocyte maturation, and the hyperkeratosis described above (Fig. 2), were reminiscent of the characteristics of the skin of patients with actinic keratosis (AK), a disease recognized as a precursor to the development of in situ SCCs (Bowen’s disease) and invasive SCCs [41]. Consistent with these premalignant lesions, N-IKKα mice developed carcinomas in situ (observed in 3/24 N-IKKα versus 1/28 Control mice; p = 0.3) (Fig. 7J–K’), which are characterized by the cellular atypia described before associated with a high mitotic index (Fig. 7K, K’) without breakage of the basal membrane; as well as skin SCCs (identified in 5/24 N-IKKα versus 0/28 Control mice; p = 0.01) with varying degrees of malignancy, ranging from well-differentiated to undifferentiated SCCs (Fig. 7L–N; Supplementary Fig. 9Q, R). Well differentiated SCCs developed in N-IKKα mice (see examples in Fig. 7L, M), were infiltrating, malignant tumors, composed of irregular nests or cords of keratinocytes invading the adjacent dermis (arrows in Fig. 7L) that progressively formed characteristic foci of keratin pearls, consisting of concentric laminated masses of keratin (Fig. 7M). Both carcinomas in situ and infiltrating and well differentiated SCCs retained the expression of keratins, characteristic of keratinocytes (Supplementary Fig. 9A, C, E), although a decrease in the expression of another epithelial marker, E-Cadherin, was observed in invasive SCCs (arrow in Supplementary Fig. 9F). In contrast, undifferentiated SCCs (an example is showed in Fig. 7N) that presented solid masses of anaplastic cells with spindle cell-shaped keratinocytes invading the adjacent dermis, showed minimal or absent squamous differentiation and expressed low levels of keratins or had lost the expression of epithelial markers (Supplementary Fig. 9G, H, P). We verified that the transgene was expressed in the nucleus of epithelial cells of pretumoral and tumoral lesions (Supplementary Fig. 9, I-P).

However, although chronic inflammation and increased activation of STAT3 and NF-κB signaling pathways have been linked to cancer development [