Control of the cell cycle hinges on the tight regulation of cyclin-dependent kinases (CDKs) and their dynamic interactions with oscillating cyclins, which together orchestrate cellular division (Table 2). Targeting these various functional complexes at different checkpoints restricts aberrant proliferation. When this intricate signaling network is compromised, CDK-cyclin complexes are detached from protective cellular mechanisms to drive uncontrolled growth. Cutaneous malignancies, particularly melanoma, demonstrate substantial dysregulation of the cell cycle network, which contributes to tumor initiation and progression.
2.1. CDKs and p16INK4
The cyclin-dependent kinase-2A gene (CDKN2A) is clustered in the region of chromosome 9p21.3. This tumor suppressor gene encodes p16INK4a, a pivotal regulator of the G1-S checkpoint. The loss of regulation at this checkpoint is critical in the progression of cancer. p16INK4a prevents proliferation by binding to CDK4 and CDK6, inhibiting their interaction with cyclin D and thus preventing the formation of an activated complex.
Undoubtedly, progression of melanoma is closely linked to direct cell cycle regulators. Walker et al., found that 43 of 45 melanoma cell lines exhibited genetic aberrations in
CDK2NA, with deletions being more prevalent than point mutations or methylations [
44]. Over 50% of melanomas were found to have deletions in this locus, implicating
CDKN2A as one of the most common alterations in cutaneous melanoma. The inactivation of this protective gene is a key factor in melanoma susceptibility, particularly in familial cases due to heterozygous germline mutations, and to a lesser extent, in sporadic melanoma. Despite the pivotal role of p16
INK4a in melanoma progression, its expression alone has not proven to be a reliable indicator of tumor recurrence or patient survival [
45]. It is more common to see multiple mutations rather than an isolated p16
INK4a aberration. An
in vivo murine model mimicking human somatic loss of p16
INK4a and activation of
RAS in human melanoma demonstrated rapid growth and development of unpigmented melanomas from adult melanocytes [
46]. This highlighted the potential oncogenicity throughout cell maturity and the synergy of different genetic alterations used by melanocytes to exploit cell cycle regulation.
As noted, binding partners CDK4/6 and cyclin D1 are crucial propellors of the G1-S transition point, via sequential inactivating phosphorylation of the stage-specific tumor suppressor, Retinoblastoma (Rb), ultimately leading to expression downstream transcription factors [
47]. Acting as independent oncogenes, CDK4 and CCND1 amplifications are most common in acral melanomas, where median survival of patients with CDK4 amplification and p16
INK4a loss is significantly decreased [
48]. Activation of CDK4 can result from various mutations, such as loss of CDK4 sensitivity to p16
INK4A or the germline CDK4
R2C mutation that disrupts binding of CDK4-p16
INK4a, both preventing negative regulation of the G1-S transition [
49]. In studies, homozygous mouse knock-in of CDK4
R2C mutation led to widespread formation of tumors within 8 to 10 months, including skin tissue. The dependency of melanoma progression on CDK4 is further emphasized in the work of Zou et al., where CDK4 and cyclin D1 null mice experienced significant reduction in tumorigenic foci relative to wildtype CDK4, and similarly did not lead to tumor production
in vivo [
50]. The team of Sauter et al. demonstrated the oncogenic potential of cyclin D1, using anti-sense therapy to target cyclin D1 in melanoma lines overexpressing this protein. Importantly, targeting of cyclin D1 induced apoptosis
in vitro with significantly decreased tumor burden in mice models observed selectively in the mutated melanocytes [
51,
52]. With upwards of 90% of melanomas demonstrating mutation in various segments of the CDK4/6 pathway, the deep investigation of cyclin-dependent kinase inhibitors (CKIs) in targeted therapy is of no surprise [
53]. Notably, both the independent and concurrent amplification of these cell cycle regulators has been implicated in enhancing therapy resistance in variously mutated melanomas [
54,
55].
Though not as classically defined, deregulation of the G1-S cell cycle transition also contributes to the progression of the non-melanoma cutaneous malignancies. Screening of the
CDKN2A locus for genetic aberrancy in 15 cases of freshly-frozen BCC tissue by Kanellou et al. revealed a previously described G442 (Ala148Thr) polymorphism in three cases that did not hamper the regulatory role of p16
INK4a on CDK4/6-cyclin D1 [
56]. In this same study, a decrease in p16 transcript levels was observed in 14 of 15 samples, implying potential inactivation of the tumor suppressor similar to what is seen in melanoma. In contrast, the team of Eshkoor et al. found that ten samples of paraffin-embedded skin BCCs had significantly increased p16 protein in the nucleus and cytoplasm, with a corresponding significant increase in gene expression [
57]. Additionally, when assessing the role of HPV in BCC, Paolini et al. found that dysregulated keratinocytes overexpressing p16
INK4A in 94% of samples (35 of 37), with 8 samples exhibiting elevated protein levels in >30% of the immunostained cells [
58]. Notably, this variable expression between different research teams may have been a result of differing sample processing and experimental methods.
The mechanism behind the upregulation of p16
INK4a remains undetermined. An alternate study by Eshkoor et al. of ten BCC tissue samples displayed significantly increased protein and mRNA expression of CDK6 and cyclin D1 [
59]. Considering this, BCC appears to be influenced by the p16
INK4a-cyclin D/CDK-pRb signaling pathway to an extent. Accordingly, one might postulate an elevation in p16
INK4a levels as the expected response to increased proliferation within the tumor cells. Upon investigating for regional variability in p16 levels among different subtypes of BCC, Svensson et al. note a correlation of expression with invasiveness, as protein levels were highest in cells along the infiltrating tumor periphery. Persistence of p16
INK4a functionality in this aggressive tumor edge was associated with downregulation of the proliferative marker Ki-67, indicating an inverse relationship between proliferation and infiltration that may be influenced by p16
INK4a [
60].
The risk of BCC, like melanoma, which is augmented by UV radiation-induced DNA damage [
61]. results in augmented p16 expression in comparison to non-sun-exposed skin. Hence, upregulation of p16
INK4a could indicate a broad cellular stress response. An increased expression of p16
INK4a was observed in recurrent BCC lesions relative to non-recurrent, suggesting an association with therapy resistance and/or tumor recurrence [
62]. Most research on BCC has focused upon the role of chronic activation of SHH signaling although the common development of resistance to SMO inhibitors predicts the involvement of other factors such as deregulation of the cell cycle. This possibility remains understudied. An instructive role for aberrant cell cycle regulation in cSCC is also not well investigated although an association of dysregulated cell cycle is broad range of dysplastic cells (10-80%) isolated from samples of actinic keratoses (pre-cancerous lesions), Bowen’s carcinoma (squamous cell carcinoma
in situ) and cutaneous squamous cell carcinoma cells displayed overexpression of p16
INK4a relative to normal tissue [
63]. As well, the involvement of p16
INK4a, cyclin D1 and Rb has been demonstrated in other SCC [
64]. Importantly,
in vitro targeting of cyclin D1 in SCC lines (head and neck, facial and vulvar tissue) and
in vivo immunodeficient mice revealed significant reduction in tumor growth [
52].
Both cBCC and cSCC often bear allelic loss at the
CDKN2a locus [
65]. although the consequence to tumor initiation or progression is not yet clear. Immunohistochemical staining by Zheng et al. suggested a variability in p16
INK4a expression between BCC and SCC. Here, 15% of BCC (47 cases) revealed low levels of positive staining (1+), relative to 80% of cSCC (44 cases) while 20% of cSCC exhibit significantly protein expression [
66]. These results suggest cSCC progression is associated with deregulation of cell cycle mediators but establishing cause and effect requires further study.
2.2. p14ARF and p53
The
CDKN2a locus is alternatively known as ARF-INK4a, due to the alternatively spliced product being p14
ARF, a tumor suppressor with identified inactivating mutations demonstrated in various cancers. Importantly, p14
ARF functions to inhibit the p53-degrading protein MDM2, thereby stabilizing p53 and stabilizing its activity as a crucial cell-cycle and apoptotic regulator [
67]. Moreover, p14
ARF is recognized as a connector of Rb and p53 [
68,
69]. Upon phosphorylation of Rb, associated E2F transcription factors are untethered, which then induce expression of p14
ARF. This increase in p14
ARF activates the p53/p21 pathway, providing an additional layer of control to inhibit cellular proliferation, underscoring the interplay between these regulators in maintaining cellular homeostasis [
70,
71].
There is strong evidence highlighting an inverse relationship of p14
ARF with melanoma progression. In general, exon mutations impacting p16
INK4A function are more frequent than in p14
ARF, although a downregulation of p14
ARF expression can contribute to melanoma oncogenesis [
72,
73]. However, some studies of metastatic melanoma have reported increased mutation of ARF relative to p16
INK4a, which is perhaps sample-dependent [
74]. Regardless, mutations in both genes are significant, and their concurrent inactivation is reported in upwards of 40% of melanoma. [
75] Dobrowolski et al. compared p14
ARF protein levels in increasingly aggressive human melanomas, revealing 11 of 14 benign nevi, 3 of 12 melanomas and 0 of 6 metastatic melanomas showing positive staining for p14
ARF, supporting a role for this protein in melanoma progression [
76]. Moreover, the introduction of FLAG-tagged truncated p14
ARF constructs into the NM39 melanoma cell line resulted in blunted G1 arrest relative to wild-type p14
ARF [
75].
Upon comparison of select tumor suppressor genes in AK and cSCC, Kanellou et al. found downregulation of both p14
ARF and p53 in SCC relative to the pre-cancerous AK lesions [
77] Likewise, investigation of the
CDKN2A locus in 40 human cSCC samples revealed alterations in 76% of samples, with variable point mutations and promoter methylation representing the most common causes of inactivation. [
78] Similarly, Pacifico et al. show a similar inactivation in BCC, with only 1 of 16 sun-exposed patient tumors staining positively for protein expression of p14
ARF [
79]. These studies highlight the potential role of inactivation of p14
ARF in the tumourigenic process for both melanoma and NMSCs.
Nevertheless, mutations in p16 or its loss of expression is likely insufficient for cutaneous tumorigenicity. For example, individuals homozygous for the
CDKN2A germline mutation can remain disease-free. Thus, inactivation of this tumor suppressor in humans is presumed to cooperate with additional driving or spontaneous mutations to orchestrate tumor initiation and progression [
80]. However, experimentally, genomic loss of ARF in mice (murine p19
ARF is equivalent to human p14) led to increased cSCC tumor formation at multiple cutaneous sites when treated with the carcinogen DMBA [
81]. Loss of ARF was found to be independent of p53 signalling, which remained functional via alternate activation pathways, as demonstrated by the continued induction of p21 expression in the context of non-functional p14
ARF [
75,
82].
TP53 is one of the most frequently inactivated regulators in cancer, yet an association of
TP53 function loss in melanoma is controversial. Early studies reported variable
TP53 mutations in melanoma (0-20%). More recent and higher resolution whole exome sequencing analysis have detected inactivating p53 mutation in 15-20% of melanoma samples [
83,
84]. In contrast to NMSCs, which exhibit early mutations of
TP53, these mutations are more often detected as late events associated with advanced melanoma [
85]. The current consensus is that wildtype
TP53 predominates in over 80% of melanomas [
86]. Moreover, elevated p53 expression is detected with increasing tumor progression [
87,
88]. These results imply that wildtype
TP53 function is aberrant in melanoma. This possibility is supported by one study that demonstrated failure of p53 to induce apoptosis. Other studies implicate p53 as a driver of therapy resistance and aggressiveness through expression of shorter isoforms [
86,
89,
90]. Additional mechanisms include disrupted p14
ARF signaling, which causes persistent MDM2-mediated inactivation of p53 [
91]. However, as mentioned earlier, alternative studies have demonstrated p53 can retain its normal function in the presence of inactivated p14
ARF [
81].
In skin cancer, deletions of
CDKN2A are more common than point mutations, the latter more commonly associated with UV exposure [
92]. UV-signature mutations (C to T and CC to TT dipyrimidine sequences) are present in NMSCs and pre-cancerous lesions, likely serving as an early step in carcinogenesis [
79]. Comparison of mutant p53 in aggressive and non-aggressive BCCs revealed detection in 38% and 66% of human samples, respectively [
93]. Moreover, increased tumorigenicity of NMSC was exhibited in heterozygous p53 mutant mice compared to wildtype [
94]. In cSCC, mutant p53 has been reported to be found in greater than 50% of human tumor samples analyzed [
93,
95]. This early involvement implicates p53 in tumor initiation, as noted by the detection of mutations in actinic keratosis, Bowen’s disease and cSCC in situ [
96,
97]. Progression is influenced by additional tumor suppressor mutations at the
CDKN2A locus, especially in the presence of loss-of-function p53 mutants resulting in elevated proliferation, metastatic potential and drug-resistance [
92,
98,
99]. In experimental models of cBCC loss-of-function mutations in
Trp53 is required for sustained tumor transformation. Unlike NMSCs, mutant p53 is less significantly implicated in melanoma.
Collectively these reports suggest that both the initiation and progression of human melanoma and NMSCs may be impacted by dysregulated cell cycle protein functions. Additional studies focusing on specific growth pathways and stemness regulation further implicate the discussed proteins, though are out of the scope of this review.