Cell News 4/2014
20
Research news
ed that deregulation of cell-cell adhesion proteins can give rise
to cancer (Wodarz and Näthke, 2007; Berx and van Roy 2009).
Besides altered adhesion and barrier properties, impaired polarity
protein function may cause altered division orientation and cell
fate. During development or at homeostatic conditions, polarity
proteins of the Par family impinge on several signalling branches
that comprise common proto-oncogenes, e.g. secreted factors of
the Wnt family, growth factor receptor tyrosine kinases such as
EGFR or ErbB2, cytoplasmic signaling mediators like Ras, Raf and
Src, or transcription factors such as c-myc (Martin-Belmonte and
Perez-Moreno 2012). Vice versa, primary tumor suppressor genes
or intercellular adhesion molecules, e.g. VHL, APC, ASPP2, PTEN,
NF2/Merlin, and E-cadherin, are involved in regulation of pola-
rization processes (Martin-Belmonte and Perez-Moreno 2012)
highlighting that regulation of cell-cell adhesion, cyto-architec-
ture and tumorigenesis are intimately linked.
Functional evidence for a role of polarity proteins in
cancer
Mutations in genes of the Scribble complex cause loss of apico-
basal polarity and neoplastic outgrowth when mutated or com-
bined with oncogenic Ras (Bilder, 2004). Such tumor-suppressive
functions have meantime been demonstrated for mammalian
Scribble as well (Zhan et al. 2008; Pearson et al. 2011; Godde
et al. 2014). In human cancers, the expression of polarity prote-
ins is frequently altered. Scribble complex proteins, in line with
their tumor-suppressive function, show mainly reduced protein
levels, whereas the aPKC/Par polarity proteins are often overex-
pressed in human carcinoma, suggesting a causal relation (Huang
& Muthuswamy, 2010; Murray et al., 2011). Mutations in Droso-
phila Par3 complex components alone are not sufficient to drive
tumorigenesis (Pagliarini and Xu 2003). Instead, overexpression
of aPKC cooperates with active Ras (Brumby et al., 2011), sug-
gesting that activation rather than loss of apical polarity prote-
ins facilitates tumorigenesis. Murine tumor models revealed that
loss of aPKCλ, one of the two mammalian aPKC isoforms, strongly
inhibits Ras-driven tumorigenesis in pancreas, colon and lung
(Murray et al., 2009; Murray et al. 2011; Regala et al. 2009; Scotti
et al. 2010; Murray et al. 2012). Moreover, chemical aPKC inhibi-
tion suppresses Smo/Gli-mediated growth of basal cell carcinoma
(Atwood et al. 2013). The function of the closely related aPKC
ζ
isoform in cancer seems less clear. Interestingly, aPKC
ζ
counter-
acts rather than promotes lung tumors through regulation of IL6
upon nutrient deprivation (Galvez et al. 2009). aPKC
ζ
furthermo-
re prevents intestinal tumorigenesis likely through regulation of
metabolic signaling (Ma et al. 2013) and has been reported to
suppress prostate tumors through phosphorylation and regulation
of c-myc (Young et al. 2013). Incongruent with above findings,
aPKC
ζ
is also able to inhibit apoptosis and to reduce sensitivity
of cancer cells towards chemotherapeutic agents, indicating for a
context-specific role of aPKC
ζ
(Rimessi et al. 2012).
Role of Par3 in adult skin homeostasis
To address a function of Par3 in skin, we generated mice with
conditional Par3 deletion in the epidermis and hair follicle (K14-
Cre+;Par3flox/flox; Par3 eKO; Iden et al., 2012b). These mice are
viable and reached adulthood comparable to control litterma-
tes. As we previously demonstrated that aPKC-mediated phos-
phorylation of junctional adhesion molecule JAM-A promotes
tight junction formation (Iden et al., 2012a), we asked for a
potential role of epidermal Par3 in these processes. Interestin-
gly, Par3 deletion in keratinocytes resulted in loss of junctional
recruitment of aPKC and delayed tight junction biogenesis (Iden
et al., 2012b). Our initial analyses further suggest that loss of
Par3 impairs epidermal stem cell homeostasis and differentiati-
on, phenotypes also observed upon epidermal deletion of aPKC
λ
(Niessen et al. 2013). These observations indicate that Par3 co-
operates with aPKC
λ
to regulate epidermal homeostasis.
Par3 can both promote and suppress skin cancers: a
matter of junctional localization?
Despite strong
in vivo
evidence of a role of aPKC in mammalian
cancers, the role of Par3 in tumorigenesis remained unknown.
To elucidate a potential causal link between Par3 dysfunction
and the formation and progression of cancer, we investigated
the consequence of loss of Par3 on DMBA/TPA-induced skin
tumorigenesis. This model is based on induction of oncogenic
mutations of the endogenous Ras locus and subsequent phor-
bolester-mediated promotion of tumor outgrowth (fig. 2). Par3
eKO mice did not form spontaneous tumors. Strikingly, however,
they exhibited strongly reduced formation and growth of Ras-
induced papillomas accompanied by loss of cortical aPKC, redu-
ced proliferation, and increased apoptosis (Iden et al., 2012b)
(fig. 3).
We therefore assessed the relevance of cell-cell contact-loca-
lized Par3 in growth and survival signaling and found that not
only aPKC localization to intercellular junctions required Par3,
but also crucial components of the Ras signaling pathway such
as Sos1/2, Ras and active ERK1/2 were largely lost from cell-cell
contacts in Par3-deficient keratinocytes and epidermis (Iden et
al., 2012b) (fig.4). Intriguingly, reduced ERK activity and growth
of Par3 KO cells was predominantly observed at contacting con-
ditions but not in absence of firm intercellular adhesion. Mo-
reover, impaired ERK and Akt activity upon loss of Par3 were
mimicked by downregulation of aPKC
λ/ι
. Artificial targeting of
aPKC to the membrane (aPKC-CAAX) rescued decreased ERK
activity of Par3-deficient keratinocytes. These data indicate
that Par3 exhibits pro-oncogenic function in papillomagene-
sis likely by correctly localizing Ras signaling components and
aPKC to intercellular adhesions and thereby promoting efficient
downstream mitogenic signaling via MEK/ERK (Iden et al. 2012)
(fig. 4). Interestingly, inactivation of the basolateral polarity
protein Scribble results in increased MAPK signaling (Dow et
al., 2008; Pearson et al. 2011), in line with its tumor-suppressive
role in mammalian and invertebrate tumor models (Humbert et
al. 2008), demonstrating the antagonistic role of apical and ba-
solateral polarity proteins.
Opposing to above phenotypes, however, Par3 mutant mice are
predisposed to the formation of keratoacanthomas (KAs) (fig. 3).
These aggressively growing cutaneous tumors likely originate
