Ribosomes and p53 – a new KIT for skin darkening

Authors


  • Coverage on: McGowan et al. (2008) Ribosomal mutations cause p53-mediated dark skin and pleiotropic effects. Nat. Genet. 40: 963–970.

Bert Vogelstein once introduced his audience to the tumor suppressor gene p53 by stating, tongue in cheek: ‘If you think every possible experiment on p53 has been done already, you are mistaken; recent results show that p53 also accumulates in the skins of rats in space!’ Well, space flights are stressful. Think of excess gravity and micro-gravity, ionizing radiation, etc. and it is not surprising that p53 is stabilized and activated. But in retrospect one should have asked, have the space rats shown a tan? The question arises because last year we learned that mice need p53 to tan after UVB exposure, and it was suggested (though not proven) that this is because p53 directly stimulates Pomc, the gene that gives rise to melanocyte stimulating hormone (MSH) (Cui et al., 2007). In a new twist, Kelly McGowan et al. from Greg Barsh’s group now show that ribosomal protein mutations also lead to p53 accumulation and to darker skin, though clearly not by stimulating Pomc. Rather, these mutations stimulate Kit ligand (Kitl, formerly called Steel factor or stem cell factor; Mc Gowan et al., 2008), the growth factor that interacts with its receptor tyrosine kinase, KIT. These are exciting, novel findings that touch on many fundamental questions, including whether distinct cell types respond differently to the same stress and whether the same cell type responds differently to different stresses.

Here are some details: During a continuing effort to analyze N-ethyl-N-nitrosourea-induced mouse dark skin (Dsk) mutants, McGowan et al. found that Dsk3 and Dsk4 have point mutations in Rps19 and Rps20, respectively, each encoding one of the 30 or so proteins that make up the small (40S) subunit of the eukaryotic ribosome. While homozygotes die early during embryonic life, heterozygotes are viable but lighter in weight than their wild-type littermates, and some show a belly spot due to embryonic paucity in melanoblasts. Postnatally, however, melanocytes gradually accumulate in the basal layer of the epidermis, particularly on the tail, feet and ears. So, how might a shortage in a typical house-keeping protein turn melanocyte deficiencies into melanocyte abundance?

McGowan et al. first ask where the mutations act. To answer this question, they make use of a pre-existing line of mice called Rps6lox in which Rps6, encoding yet another 40S ribosomal protein, can be eliminated tissue-specifically by Cre-mediated recombination. The elimination of one of the two copies of Rps6 in melanocytes (achieved by crossing with mice carrying an Mitf-Cre transgene, Alizadeh et al., 2008) leads to mice whose skins are lighter in color than those of Dsk3 or Dsk4 mutants, and lighter still than those of mice that carry the Rps6lox allele but not Cre. In contrast, the elimination of one of the two copies of Rps6 in keratinocytes (using K5-Cre mice, Ramirez et al., 2004) leads to mice that are notably dark, darker still than Dsk3 or Dsk4 mice. Hence, ribosomal protein deficiency in keratinocytes alone can give rise to dark skin, at least in the case of Rps6. And it is likely that the relatively lighter color of Dsk3 or Dsk4 results from a combination of cell-autonomous, negative effects of the mutations operating in melanocytes, and cell non-autonomous, positive effects operating in keratinocytes.

The authors then ask how the mutations act. They approach this question with a series of imaginative guesses. Previous results have shown that Rps6 heterozygosity leads to stimulation of p53, apoptosis, and embryonic death around gastrulation. Indeed, p53 staining was increased in the basal layer of the skins of Rps6lox/+; K5-Cre and of Dsk3 and Dsk4 mice. Moreover, gene expression profiling of Rps6lox/+; K5-Cre skin showed increased mRNA levels of two known p53 targets but not Pomc (that Pomc levels are indeed similar in wild type and mutant was confirmed by RT-PCR). The link to dark skins, however, comes from a gene not identified by expression profiling and not previously known to be a p53 target, Kitl. By RT-PCR, Kitl mRNA levels are increased in the skins of all three Rps mutants and also in the presence of a stabilizing p53 mutation, even without a ribosomal protein mutation in this case. Most importantly, homozygosity for a p53 knock-out allele abolishes Kitl stimulation and skin darkening in Rps6lox/+; K5-Cre mice, and skin darkening is also absent after injection of a neutralizing antibody to KIT.

The above findings clearly establish a pathway leading from Rps mutations to skin darkening. Nevertheless, because KIT antibodies likely neutralize all KIT activity, it is not clear whether only elevated levels of KITL would bring about a darker skin, or whether wild-type levels of KITL might suffice, provided they can cooperate with other melanotropic factors potentially induced by p53. In any event, elevated levels of MSH are not important, as Pomc mRNA levels are not changed along with the increased levels of p53, notwithstanding the observation that p53 stimulates Pomc after UVB (Cui et al., 2007). It is conceivable that distinct stresses induce different p53 isoforms or co-factors that ultimately affect which sets of target genes are stimulated.

Still unclear are the mechanisms by which ribosomal stress induces p53. Although not specifically demonstrated for the Dsk mutants, previous results suggest that many ribosomal protein mutations lead to compromised ribosome biogenesis and nucleolar disruption. One of the links between ribosome deficiencies and p53 may be the double minute protein-2 (MDM2 in mice, HDM2 in humans), an E3 ligase which interacts with p53 and normally leads to p53 ubiquitination and degradation. Interestingly, MDM2/HDM2 proteins also interact with specific 60S ribosomal proteins. Hence, it is conceivable that once ribosome biogenesis is disrupted, ribosomal proteins diffuse, inhibit MDM2/HDM2, and in this way stabilize p53. In fact, it has been speculated that defects in ribosomal biogenesis and nucleolar disruption may be the universal triggers for p53 induction. Nevertheless, much work is needed to identify additional partners in these processes. Given the ease with which one can assess coat color changes in mice, p53-inducing Dsk mutants would seem to be an ideal model for working out the genetics of the underlying mechanisms of induction.

Mere induction of p53, however, does not explain how this protein manages to decide between allowing, say, a genotoxically damaged cell to live on until it is repaired, and killing it when it is beyond repair. For the sake of argument, let us assume that Dsk3 and Dsk4 induce p53 in melanoblasts/melanocytes (a point not yet formally shown), and induce the exact same amounts of p53 in melanoblasts/melanocytes and keratinocytes. We are then faced with the problem of how to explain why melanoblasts/melanocytes are harmed developmentally, while keratinocyte development is seemingly o.k. (although this point, also, needs further exploration). Ribosomal protein synthesis, energetically among the most expensive tasks a cell is faced with, is tightly regulated by energy supply and demand. Hence, it is conceivable that migrating melanoblasts put higher demands on ribosomes than the more stationary keratinocytes, and so their sensitivity to ribosomal stress and p53 induction may be higher than, or different from, that of keratinocytes. It might be interesting in this context to compare the effects of ribosomal stress on melanoblasts/melanocytes with those on the non-migratory retinal pigment epithelium cells.

It is also worth mentioning that melanoblasts/melanocytes are not the only cells that suffer developmentally from ribosomal protein mutations. RPS19, for instance, is mutated in about a fifth of patients with Diamond–Blackfan syndrome (Online Mendelian Inheritance in Man #105650), which is characterized by varying degrees of deficiencies in erythroid precursors, though not by tanning (the latter fact perhaps because RPS19 mutations are more readily compensated in human compared to mouse keratinocytes, perhaps by upregulating the wild type copy of the gene). In fact, McGowan et al. find that their Rps19Dsk3 mutant mice also suffer from a mild erythrocyte deficiency. Interestingly, this deficiency (as well as reduced body size) can be corrected by heterozygosity for the p53 knock-out allele, suggesting that p53 is part of the ribosomal mutant phenotype regardless of how a given cell responds to the mutation.

Still another question remains that concerns the long-term effect of chronically elevated p53 levels. There are several mouse models in which increases in p53 are associated with protection against cancer, but also with premature aging in some cases. The ribosomal dark skin mutants combine elevated p53 levels potentially with impaired translation, at least of some proteins. Reduction in protein synthesis, however, is usually credited with longevity, not premature aging. Hence, the question arises whether the mutant mice respond with longer life spans or age more quickly, or whether each of their cell types reacts in a cell type-specific way.

Finally, one may ask what the above experiments hold in store for understanding melanoma. Loss-of-function mutations in p53 are rare in primary human melanoma. In mice, p53 appears to be somewhat protective in an activated Ras-dependent melanoma (Bardeesy et al., 2001) and less so in a hepatocyte growth factor/scatter factor-dependent melanoma (Ha et al., 2007). Hence, it might be interesting to test whether the combination of ribosomal protein mutations and elevated levels of p53 and KITL would influence melanoma formation in these models. If so, which of the three components, ribosomes, p53, or KIT signaling, might be more important? A test of melanoma formation in ribosomal protein mutants may be all the more interesting because melanoma cells often show changes in ribosomal function. For instance, the mammalian target of rapamycin (mTOR) pathway, which controls many components of protein synthesis including ribosome biogenesis, is activated in many human melanomas; melanomas can carry mutations in the RPS6 kinase RPS6KA2; and ARF, a positive regulator of p53 and frequently mutated in melanoma, is a predominantly nucleolar protein that may be involved in ribosome biogenesis via its interaction with nucleophosmin, a protein that facilitates the transport of ribosomal subunits into the cytoplasm. In addition, some types of human melanomas with infrequent BRAF and NRAS mutations can have copy number changes or activating mutations in KIT. Undoubtedly, the connection between ribosomal function, p53, signaling pathways, and cancer is a complex affair, but Dsk mutants might just help to shovel light into this dark corner. Perhaps we look to a brighter future in which additional Dsk mutants will reveal still other facets of the pathways that McGowan et al. have so elegantly elucidated with Dsk3 and Dsk4.

Ancillary